1 2013-11-05 Richard Biener <rguenther@suse.de>
4 * tree-dfa.c (get_ref_base_and_extent): Merge common code
5 in MEM_REF and TARGET_MEM_REF handling. Make sure to
6 process trailing array detection before diving into the
7 view-converted object (and possibly apply some extra offset).
9 2013-11-05 Joseph Myers <joseph@codesourcery.com>
11 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
13 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
15 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
17 PR tree-optimization/58958
18 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
19 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
21 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
23 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
24 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
26 2013-11-05 Jakub Jelinek <jakub@redhat.com>
28 PR tree-optimization/58984
29 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
30 set *SIZE_P if non-NULL on success.
31 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
33 (ipcp_transform_function): Likewise. Punt if size of access
34 is different from TYPE_SIZE on v->value's type.
36 2013-11-05 Tobias Burnus <burnus@net-b.de>
38 * doc/invoke.texi (-fopenmp-simd): Document new option.
39 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
40 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
41 * tree.c (attribute_value_equal): Ditto.
43 2013-11-04 Wei Mi <wmi@google.com>
45 * sched-rgn.c (add_branch_dependences): Keep insns in
46 a SCHED_GROUP at the end of BB to remain their location.
48 2013-11-04 Wei Mi <wmi@google.com>
50 * gcc/config/i386/i386.c (memory_address_length): Extract a part
51 of code to rip_relative_addr_p.
52 (rip_relative_addr_p): New Function.
53 (ix86_macro_fusion_p): Ditto.
54 (ix86_macro_fusion_pair_p): Ditto.
55 * gcc/config/i386/i386.h: Add new tune features about macro-fusion.
56 * gcc/config/i386/x86-tune.def (DEF_TUNE): Ditto.
57 * gcc/doc/tm.texi: Generated.
58 * gcc/doc/tm.texi.in: Ditto.
59 * gcc/haifa-sched.c (try_group_insn): New Function.
60 (group_insns_for_macro_fusion): Ditto.
61 (sched_init): Call group_insns_for_macro_fusion.
62 * gcc/target.def: Add two hooks: macro_fusion_p and
65 2013-11-04 Kostya Serebryany <kcc@google.com>
67 Update to match the changed asan API.
68 * asan.c (asan_function_start): New function.
69 (asan_emit_stack_protection): Update the string stored in the
70 stack red zone to match new API. Store the PC of the current
71 function in the red zone.
72 (asan_global_struct): Update the __asan_global definition to match
74 (asan_add_global): Ditto.
75 * asan.h (asan_function_start): New prototype.
76 * final.c (final_start_function): Call asan_function_start.
77 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1 to __asan_init_v3.
79 2013-11-04 Wei Mi <wmi@google.com>
81 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): Separate
82 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
83 * gcc/config/i386/i386.c (ix86_option_override_internal): Ditto.
84 (ix86_issue_rate): Ditto.
85 (ix86_adjust_cost): Ditto.
86 (ia32_multipass_dfa_lookahead): Ditto.
87 (ix86_sched_init_global): Ditto.
88 (get_builtin_code_for_version): Ditto.
89 * gcc/config/i386/i386.h (enum target_cpu_default): Ditto.
90 (enum processor_type): Ditto.
91 * gcc/config/i386/x86-tune.def (DEF_TUNE): Ditto.
93 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
95 PR rtl-optimization/58967
96 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
97 !lra_in_progress for mode sizes bigger word.
99 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
101 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
102 arguments to merge instruction for little endian.
103 (vec_widen_umult_lo_v16qi): Likewise.
104 (vec_widen_smult_hi_v16qi): Likewise.
105 (vec_widen_smult_lo_v16qi): Likewise.
106 (vec_widen_umult_hi_v8hi): Likewise.
107 (vec_widen_umult_lo_v8hi): Likewise.
108 (vec_widen_smult_hi_v8hi): Likewise.
109 (vec_widen_smult_lo_v8hi): Likewise.
111 2013-11-04 Ian Lance Taylor <iant@google.com>
113 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
114 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
115 builtins that take pointers.
116 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions
118 * lto-wrapper.c (merge_and_complain): Collect
119 OPT_fnon_call_exceptions.
120 (run_gcc): Pass -fnon-call-exceptions.
122 2013-11-04 Jakub Jelinek <jakub@redhat.com>
124 * optabs.c (expand_vec_perm): Revert one incorrect line from
127 PR tree-optimization/58978
128 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
129 use_stmt by single_imm_use directly. Only call single_imm_use
132 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
134 PR rtl-optimization/58968
135 * lra-spills.c (return_regno_p): New function.
136 (lra_final_code_change): Use it.
138 2013-11-04 Joseph Myers <joseph@codesourcery.com>
140 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document
142 * target.def (float_exceptions_rounding_supported_p): New hook.
143 * targhooks.c (default_float_exceptions_rounding_supported_p): New
145 * targhooks.h (default_float_exceptions_rounding_supported_p):
147 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
149 * doc/tm.texi: Regenerate.
150 * config.gcc (powerpc*-*-linux*): Set extra_objs.
151 * config/rs6000/rs6000-linux.c: New file.
152 * config/rs6000/rs6000-protos.h
153 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
154 * config/rs6000/linux.h
155 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
156 * config/rs6000/linux64.h
157 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
158 * config/rs6000/t-linux (rs6000-linux.o): New rule.
159 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
161 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
163 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
164 Replace the define_insn_and_split with a define_insn and two
165 define_splits, with the split after reload re-permuting the source
166 register to its original value.
167 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
168 (*vsx_le_perm_store_v8hi): Likewise.
169 (*vsx_le_perm_store_v16qi): Likewise.
171 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
173 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
176 2013-11-04 Jakub Jelinek <jakub@redhat.com>
178 PR tree-optimization/58946
179 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
180 bbs with bbinfo[idx].op != NULL before all blocks with
181 bbinfo[idx].op == NULL.
183 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
185 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
186 (avr_dump_double_int_hex): Likewise.
187 (avr_log_vadump): Remove %D and %X handling.
188 * config/avr/avr.c (avr_double_int_push_digit): Delete.
189 (avr_map_op_t): Change map from double_int to unsigned int.
190 (avr_map_op): Update accordingly.
191 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
192 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
193 unsigned ints rather than double_ints.
195 2013-11-03 Marek Polacek <polacek@redhat.com>
197 Implement -fsanitize=vla-bound.
198 * opts.c (common_handle_option): Handle vla-bound.
199 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE):
201 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
202 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
204 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
206 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
209 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
211 * config/i386/constraints.md (Ts, Tv): New address constrains.
212 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
213 constraint for address_no_seg_operand.
214 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
215 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
216 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
217 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
218 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
219 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
220 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
221 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
222 constraint for vsib_address_operand.
224 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
226 * gcse.c (pre_delete): Remove references to regmove from comments.
227 * recog.c: (validate_replace_rtx_1): Likewise.
228 * config/rl78/rl78.c: Likewise.
229 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
230 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
231 * common/config/mmix/mmix-common.c: Likewise.
233 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
235 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
236 * gimplify.c (gimplify_compound_lval): Likewise.
237 * graphite-clast-to-gimple.c (gloog): Likewise.
238 * graphite-dependences.c (loop_is_parallel_p): Likewise.
239 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
240 (limit_scop); Likewise.
241 (build_scops): Likewise.
242 (dot_scop): Likewise.
243 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
244 (build_scop_drs): Likewise.
245 (insert_stmts): Likewise.
246 (insert_out_of_ssa_copy): Likewise.
247 (remove_phi): Likewise.
248 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
249 * hw-doloop.c (discover_loop): Likewise.
250 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
251 * tree-dfa.c (dump_enumerated_decls): Likewise.
252 * tree-if-conv.c (if_convertable_loop_p): Likewise.
253 * tree-inline.c (tree_function_versioning): Likewise.
254 * tree-loop-distribution.c (build_rdg): Likewise.
255 (rdg_flag_vertex_and_dependent): Likewise.
256 (distribute_loop): Likewise.
257 * tree-parloops.c (loop_parallel_p): Likewise.
258 (eliminate_local_variables): Likewise.
259 (separate_decls_in_region): Likewise.
260 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
261 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
262 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
263 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
264 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
265 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
266 (vectorizable_condition): Likewise.
268 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
270 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
271 * configure: Regenerate.
272 * config/i386/i386.md (*movdi_internal): Change
273 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
274 (*movdf_internal): Ditto.
275 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
276 * config/i386/sse.md (vec_concatv2di): Output interunit movq
277 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
279 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
281 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
284 2013-10-31 David Malcolm <dmalcolm@redhat.com>
286 Automated part of renaming of symtab_node_base to symtab_node.
288 Patch autogenerated by rename_symtab.py from
289 https://github.com/davidmalcolm/gcc-refactoring-scripts
290 revision 58bb219cc090b2f4516a9297d868c245495ee622
291 with ChangeLog entry fixed up by hand.
293 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
295 (cgraph_node_for_asm): Likewise.
296 * cgraph.h (symtab_node_base): Likewise.
297 (cgraph_node): Likewise.
298 (varpool_node): Likewise.
299 (is_a_helper <cgraph_node>::test): Likewise.
300 (is_a_helper <varpool_node>::test): Likewise.
301 (symtab_nodes): Likewise.
302 (symtab_register_node): Likewise.
303 (symtab_unregister_node): Likewise.
304 (symtab_remove_node): Likewise.
305 (symtab_get_node): Likewise.
306 (symtab_node_for_asm): Likewise.
307 (symtab_node_asm_name): Likewise.
308 (symtab_node_name): Likewise.
309 (symtab_insert_node_to_hashtable): Likewise.
310 (symtab_add_to_same_comdat_group): Likewise.
311 (symtab_dissolve_same_comdat_group_list): Likewise.
312 (dump_symtab_node): Likewise.
313 (debug_symtab_node): Likewise.
314 (dump_symtab_base): Likewise.
315 (verify_symtab_node): Likewise.
316 (verify_symtab_base): Likewise.
317 (symtab_used_from_object_file_p): Likewise.
318 (symtab_alias_ultimate_target): Likewise.
319 (symtab_resolve_alias): Likewise.
320 (fixup_same_cpp_alias_visibility): Likewise.
321 (symtab_for_node_and_aliases): Likewise.
322 (symtab_nonoverwritable_alias): Likewise.
323 (availability symtab_node_availability): Likewise.
324 (symtab_semantically_equivalent_p): Likewise.
325 (fixup_same_cpp_alias_visibility): Likewise.
326 (symtab_prevail_in_asm_name_hash): Likewise.
329 (varpool_first_variable): Likewise.
330 (varpool_next_variable): Likewise.
331 (varpool_first_static_initializer): Likewise.
332 (varpool_next_static_initializer): Likewise.
333 (varpool_first_defined_variable): Likewise.
334 (varpool_next_defined_variable): Likewise.
335 (cgraph_first_defined_function): Likewise.
336 (cgraph_next_defined_function): Likewise.
337 (cgraph_first_function): Likewise.
338 (cgraph_next_function): Likewise.
339 (cgraph_first_function_with_gimple_body): Likewise.
340 (cgraph_next_function_with_gimple_body): Likewise.
341 (symtab_alias_target): Likewise.
342 (symtab_real_symbol_p): Likewise.
343 (symtab_can_be_discarded): Likewise.
344 * cgraphbuild.c (mark_address): Likewise.
345 (mark_load): Likewise.
346 (mark_store): Likewise.
347 * cgraphunit.c (decide_is_symbol_needed): Likewise.
349 (enqueue_node): Likewise.
350 (referred_to_p): Likewise.
351 (cgraph_process_same_body_aliases): Likewise.
352 (analyze_functions): Likewise.
353 (handle_alias_pairs): Likewise.
354 (output_weakrefs): Likewise.
356 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
357 * ipa-inline-analysis.c (inline_write_summary): Likewise.
358 * ipa-prop.c (remove_described_reference): Likewise.
359 (try_decrement_rdesc_refcount): Likewise.
360 (ipa_edge_duplication_hook): Likewise.
361 * ipa-ref.c (ipa_record_reference): Likewise.
362 (ipa_maybe_record_reference): Likewise.
363 (ipa_clone_ref): Likewise.
364 (ipa_clone_references): Likewise.
365 (ipa_clone_referring): Likewise.
366 (ipa_find_reference): Likewise.
367 (ipa_remove_stmt_references): Likewise.
368 (ipa_clear_stmts_in_references): Likewise.
369 * ipa-ref.h (symtab_node_base): Likewise.
371 (ipa_record_reference): Likewise.
372 (ipa_maybe_record_reference): Likewise.
373 (ipa_clone_references): Likewise.
374 (ipa_clone_referring): Likewise.
375 (ipa_clone_ref): Likewise.
376 (ipa_find_reference): Likewise.
377 (ipa_remove_stmt_references): Likewise.
378 (ipa_clear_stmts_in_references): Likewise.
379 * ipa-reference.c (ipa_reference_write_optimization_summary):
381 * ipa.c (enqueue_node): Likewise.
382 (process_references): Likewise.
383 (walk_polymorphic_call_targets): Likewise.
384 (symtab_remove_unreachable_nodes): Likewise.
385 (address_taken_from_non_vtable_p): Likewise.
386 (comdat_can_be_unshared_p_1): Likewise.
387 (comdat_can_be_unshared_p): Likewise.
388 (can_replace_by_local_alias): Likewise.
389 (function_and_variable_visibility): Likewise.
390 * is-a.h: Likewise (within example in comment).
391 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
392 (lto_symtab_encoder_encode): Likewise.
393 (lto_symtab_encoder_delete_node): Likewise.
394 (lto_symtab_encoder_in_partition_p): Likewise.
395 (lto_set_symtab_encoder_in_partition): Likewise.
396 (output_refs): Likewise.
397 (compute_ltrans_boundary): Likewise.
398 (output_symtab): Likewise.
399 (input_node): Likewise.
400 (input_ref): Likewise.
401 (input_edge): Likewise.
402 (input_cgraph_1): Likewise.
403 (input_refs): Likewise.
404 (output_cgraph_opt_summary): Likewise.
405 (input_node_opt_summary): Likewise.
406 (input_cgraph_opt_section): Likewise.
407 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
409 * lto-streamer-out.c (lto_output): Likewise.
410 (output_symbol_p): Likewise.
411 (produce_symtab): Likewise.
412 * lto-streamer.h (lto_encoder_entry): Likewise.
413 (lto_free_function_in_decl_state_for_node): Likewise.
414 (lto_symtab_encoder_encode): Likewise.
415 (lto_symtab_encoder_delete_node): Likewise.
416 (lto_symtab_encoder_in_partition_p): Likewise.
417 (lto_set_symtab_encoder_in_partition): Likewise.
418 (lto_symtab_encoder_lookup): Likewise.
419 (lsei_node): Likewise.
420 (lto_symtab_encoder_deref): Likewise.
421 * symtab.c (symtab_hash): Likewise.
422 (assembler_name_hash): Likewise.
423 (symtab_nodes): Likewise.
424 (hash_node): Likewise.
426 (hash_node_by_assembler_name): Likewise.
427 (eq_assembler_name): Likewise.
428 (insert_to_assembler_name_hash): Likewise.
429 (unlink_from_assembler_name_hash): Likewise.
430 (symtab_prevail_in_asm_name_hash): Likewise.
431 (symtab_register_node): Likewise.
432 (symtab_insert_node_to_hashtable): Likewise.
433 (symtab_unregister_node): Likewise.
434 (symtab_get_node): Likewise.
435 (symtab_remove_node): Likewise.
436 (symtab_initialize_asm_name_hash): Likewise.
437 (symtab_node_for_asm): Likewise.
438 (symtab_add_to_same_comdat_group): Likewise.
439 (symtab_dissolve_same_comdat_group_list): Likewise.
440 (symtab_node_asm_name): Likewise.
441 (symtab_node_name): Likewise.
442 (dump_symtab_base): Likewise.
443 (dump_symtab_node): Likewise.
444 (dump_symtab): Likewise.
445 (debug_symtab_node): Likewise.
446 (verify_symtab_base): Likewise.
447 (verify_symtab_node): Likewise.
448 (verify_symtab): Likewise.
449 (symtab_used_from_object_file_p): Likewise.
450 (symtab_node_availability): Likewise.
451 (symtab_alias_ultimate_target): Likewise.
452 (fixup_same_cpp_alias_visibility): Likewise.
453 (symtab_resolve_alias): Likewise.
454 (symtab_for_node_and_aliases): Likewise.
455 (symtab_for_node_and_aliases): Likewise.
456 (symtab_nonoverwritable_alias_1): Likewise.
457 (symtab_nonoverwritable_alias): Likewise.
458 (symtab_semantically_equivalent_p): Likewise.
459 * value-prof.c (init_node_map): Likewise.
460 * varasm.c (find_decl): Likewise.
461 * varpool.c (varpool_node_for_asm): Likewise.
462 (varpool_remove_unreferenced_decls): Likewise.
464 2013-10-31 David Malcolm <dmalcolm@redhat.com>
466 Manual part of renaming of symtab_node_base to symtab_node.
468 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
469 clashes with the preferred name for the base class.
470 (const_symtab_node): Remove redundant typedef.
472 2013-10-31 Jakub Jelinek <jakub@redhat.com>
474 * optabs.c (expand_vec_perm): Avoid vector mode punning
476 * expmed.c (store_bit_field_1): Likewise.
477 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
478 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
479 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
480 * config/i386/i386.c (ix86_expand_vector_move_misalign,
481 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
482 ix86_expand_sse_unpack, ix86_expand_args_builtin,
483 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
484 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
485 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
486 expand_vec_perm_vpshufb2_vpermq,
487 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
488 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
489 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
490 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
493 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
495 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
496 Use gcc_unreachable for unhandled modes. Do not check results of
497 expand_simple_binop. If not expanded to target, move the result.
499 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
500 Shiva Chen <shiva0217@gmail.com>
502 * config.gcc (nds32*-*-*): Add nds32 target.
503 * config/nds32/nds32.c: New file.
504 * config/nds32/nds32.h: New file.
505 * config/nds32/nds32.md: New file.
506 * config/nds32/constants.md: New file.
507 * config/nds32/constraints.md: New file.
508 * config/nds32/iterators.md: New file.
509 * config/nds32/nds32-doubleword.md: New file.
510 * config/nds32/nds32-intrinsic.md: New file.
511 * config/nds32/nds32_intrinsic.h: New file.
512 * config/nds32/nds32-modes.def: New file.
513 * config/nds32/nds32-multiple.md: New file.
514 * config/nds32/nds32.opt: New file.
515 * config/nds32/nds32-opts.h: New file.
516 * config/nds32/nds32-protos.h: New file.
517 * config/nds32/nds32-peephole2.md: New file.
518 * config/nds32/pipelines.md: New file.
519 * config/nds32/predicates.md: New file.
520 * config/nds32/t-mlibs: New file.
521 * common/config/nds32: New directory and files.
523 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
524 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
525 * doc/install.texi (Cross-Compiler-Specific Options): Document
526 --with-nds32-lib for nds32 target.
527 * doc/extend.texi (Function Attributes, Target Builtins): Document
528 nds32 specific attributes.
530 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
532 * lra-constraints (process_alt_operands): Use the result
533 elimination register for operand when matching constraints.
535 2013-10-31 Jakub Jelinek <jakub@redhat.com>
537 * tree-vrp.c (maybe_set_nonzero_bits): New function.
538 (remove_range_assertions): Call it.
540 * tree.c (tree_ctz): New function.
541 * tree.h (tree_ctz): New prototype.
542 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
543 first argument from tree to const_tree.
544 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
545 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
546 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
548 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
549 * tree-vect-loop.c (vect_analyze_loop_operations,
550 vect_transform_loop): Don't force scalar loop for bound just because
551 number of iterations is unknown, only do it if it is not known to be
552 a multiple of vectorization_factor.
553 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
555 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
556 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
557 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
558 (set_nonzero_bits, get_nonzero_bits): New prototypes.
559 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
560 a default def isn't partially constant.
561 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
562 is known to be partially zero.
563 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
564 to see if a default def isn't partially constant.
565 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
566 creation of a range, if VR_RANGE, try to improve nonzero_bits from
568 (set_nonzero_bits, get_nonzero_bits): New functions.
570 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
571 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
572 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
573 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
574 uses but in the condition and ASSERT_EXPR and the other successor of
575 the predecessor bb is __builtin_unreachable (), set_range_info of the
576 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
578 2013-10-31 Martin Jambor <mjambor@suse.cz>
580 PR rtl-optimization/58934
582 2013-10-30 Martin Jambor <mjambor@suse.cz>
583 PR rtl-optimization/10474
584 * ira.c (find_moveable_pseudos): Do not calculate dominance info
586 (interesting_dest_for_shprep): New function.
587 (split_live_ranges_for_shrink_wrap): Likewise.
588 (ira): Calculate dominance info and df analysis. Call
589 split_live_ranges_for_shrink_wrap.
591 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
592 Yury Gribov <y.gribov@samsung.com>
595 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
596 shadow pointer to avoid clobbering the main one.
598 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
600 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
602 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
605 * ira-color.c (update_costs_from_copies): Add new parameter. Use
606 it for calling update_costs_from_allocno.
607 (assign_hard_reg): Call restore_costs_from_copies only for
608 !retry_p. Pass new argument to update_costs_from_copies.
609 (color_pass): Pass new argument to update_costs_from_copies.
610 (ira_mark_allocation_change): Ditto.
612 2013-10-30 Sharad Singhai <singhai@google.com>
615 * opts.c (common_handle_option): Remove deprecated option
616 -ftree-vectorizer-verbose.
617 * doc/invoke.texi (Debugging Options): Ditto.
618 * opts-global.c (handle_common_deferred_options): Ditto.
619 (dump_remap_tree_vectorizer_verbose): Delete.
620 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
622 2013-10-30 DJ Delorie <dj@redhat.com>
624 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
625 that take no arguments.
627 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
630 * reload1.c (update_eliminables_and_spill): New function, broken
632 (reload): Use it. Check for frame size change after frame size
633 alignment, and call update_eliminables_and_spill first if continue-ing.
635 2013-10-30 Cong Hou <congh@google.com>
638 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
639 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
640 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
642 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
644 PR rtl-optimization/58369
645 * reload1.c (compute_reload_subreg_offset): New function.
646 (choose_reload_regs): Use it to pass endian-correct
647 offset to subreg_regno_offset.
649 2013-10-30 Tobias Burnus <burnus@net-b.de>
652 * gcc/tree-cfg.c (replace_loop_annotate): Replace warning by
655 2013-10-30 Jason Merrill <jason@redhat.com>
657 * configure.ac (loose_warn): Add -Wno-format if
658 --disable-build-format-warnings.
660 2013-10-30 David Malcolm <dmalcolm@redhat.com>
662 * cgraphunit.c (analyze_functions): Split symtab_node declarations
663 onto multiple lines to make things easier for rename_symtab.py.
665 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
666 (symtab_semantically_equivalent_p): Likewise.
668 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
671 * lra.c (check_rtl): Remove address check before LRA work.
673 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
675 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
676 POINTER_PLUS_EXPR in the defining statement.
678 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
681 * tree-pass.h (make_pass_regmove): Remove.
682 * timevar.def (TV_REGMOVE): Remove.
683 * passes.def (pass_regmove): Remove.
684 * opts.c (default_options_table): Remove entry for regmove.
685 * doc/passes.texi: Remove regmove pass description.
686 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
688 (-fdump-rtl-regmove): Ditto.
689 * common.opt (foptimize-register-move, fregmove): Ignore.
690 * Makefile.in (OBJS): Remove regmove.o.
692 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
694 (struct ira_allocno) New member allocno_prefs.
695 (ALLOCNO_PREFS): New macro.
696 (ira_prefs, ira_prefs_num): New external vars.
697 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
698 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
699 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
701 (ira_add_allocno_copy_to_list): Remove prototype.
702 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
703 (ira_pref_iterator): New type.
704 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
705 (FOR_EACH_PREF): New macro.
706 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
707 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
709 (ira_setup_alts): New function.
710 (decrease_live_ranges_number): New function.
711 (ira): Call the above function.
712 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
713 (ira_create_allocno): Initialize allocno prefs.
714 (pref_pool, pref_vec): New static vars.
715 (initiate_prefs, find_allocno_pref, ira_create_pref): New
717 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
718 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
719 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
720 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
721 (ira_add_allocno_copy_to_list): Make static. Rename to
722 add_allocno_copy_to_list.
723 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
724 swap_allocno_copy_ends_if_necessary.
725 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
726 ira_remove_allocno_prefs.
727 (ira_flattening): Ditto.
728 (ira_build): Call initiate_prefs, print_prefs.
729 (ira_destroy): Call finish_prefs.
730 * ira-color.c (struct update_cost_record): New.
731 (struct allocno_color_data): Add new member update_cost_records.
732 (update_cost_record_pool): New static var.
733 (init_update_cost_records, get_update_cost_record): New functions.
734 (free_update_cost_record_list, finish_update_cost_records): Ditto.
735 (struct update_cost_queue_elem): Add member from.
736 (initiate_cost_update): Call init_update_cost_records.
737 (finish_cost_update): Call finish_update_cost_records.
738 (queue_update_cost, get_next_update_cost): Add new param from.
739 (Update_allocno_cost, update_costs_from_allocno): New functions.
740 (update_costs_from_prefs): Ditto.
741 (update_copy_costs): Rename to update_costs_from_copies.
742 (restore_costs_from_copies): New function.
743 (update_conflict_hard_regno_costs): Don't go back.
744 (assign_hard_reg): Call restore_costs_from_copies. Add printing
746 (pop_allocnos): Add priniting more debug info.
747 (color_allocnos): Remove prefs for conflicting hard regs.
748 Call update_costs_from_prefs.
749 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
750 (get_dup_num): Rename, modify, and move to ira.c
751 (process_regs_for_copy): Add prefs.
752 (add_insn_allocno_copies): Put src as first arg of
753 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
754 * ira-costs.c (record_reg_classes): Modify and move code into
755 record_operands_costs.
756 (find_costs_and_classes): Create prefs for the hard reg of small
758 (process_bb_node_for_hard_reg_moves): Add prefs.
760 2013-10-30 Richard Biener <rguenther@suse.de>
763 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
764 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
766 (pre_and_rev_post_order_compute): ... which now wraps it.
767 * graph.c (draw_cfg_nodes_no_loops): Use
768 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
771 2013-10-30 Christian Bruel <christian.bruel@st.com>
773 * gcc/config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
774 (sh_expand_cmpstr): Handle known align and schedule improvements.
775 * gcc/config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
776 * gcc/config/sh/sh.md (cmpstrnsi): New pattern.
778 2013-10-30 Martin Jambor <mjambor@suse.cz>
780 PR rtl-optimization/10474
781 * ira.c (find_moveable_pseudos): Do not calculate dominance info
783 (interesting_dest_for_shprep): New function.
784 (split_live_ranges_for_shrink_wrap): Likewise.
785 (ira): Calculate dominance info and df analysis. Call
786 split_live_ranges_for_shrink_wrap.
788 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
791 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
793 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
795 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
796 * tree.h (POINTER_BOUNDS_P): New.
797 (BOUNDED_TYPE_P): New.
799 (pointer_bounds_type_node): New.
800 * tree.c (build_common_tree_nodes): Initialize
801 pointer_bounds_type_node.
802 * gimple.h (gimple_call_get_nobnd_arg_index): New.
803 (gimple_call_num_nobnd_args): New.
804 (gimple_call_nobnd_arg): New.
805 (gimple_return_retbnd): New.
806 (gimple_return_set_retbnd): New
807 * gimple.c (gimple_build_return): Increase number of ops
808 for return statement.
809 (gimple_call_get_nobnd_arg_index): New.
810 * gimple-pretty-print.c (dump_gimple_return): Print second op.
812 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
814 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
815 with "chkp ctor" and "bnd_legacy" attributes.
816 * gimplify.c (gimplify_init_constructor): Avoid infinite
817 loop during gimplification of bounds initializer.
819 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
821 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
822 (handle_bnd_legacy): New.
823 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
824 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
827 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
829 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
830 (BT_FN_PTR_CONST_PTR): New.
831 (BT_FN_CONST_PTR_CONST_PTR): New.
832 (BT_FN_PTR_CONST_PTR_SIZE): New.
833 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
834 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
835 (BT_FN_VOID_CONST_PTR_SIZE): New.
836 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
837 * chkp-builtins.def: New.
838 * builtins.def: include chkp-builtins.def.
839 (DEF_CHKP_BUILTIN): New.
840 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
841 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
842 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
843 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
844 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
845 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
846 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
847 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
848 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
849 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
850 * common.opt (fcheck-pointer-bounds): New.
851 * toplev.c (process_options): Check Pointer Bounds Checker is
853 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
855 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
857 * target.def (builtin_chkp_function): New.
858 (chkp_bound_type): New.
859 (chkp_bound_mode): New.
860 (fn_abi_va_list_bounds_size): New.
861 (load_bounds_for_arg): New.
862 (store_bounds_for_arg): New.
863 * targhooks.h (default_load_bounds_for_arg): New.
864 (default_store_bounds_for_arg): New.
865 (default_fn_abi_va_list_bounds_size): New.
866 (default_chkp_bound_type): New.
867 (default_chkp_bound_mode): New.
868 (default_builtin_chkp_function): New.
869 * targhooks.c (default_load_bounds_for_arg): New.
870 (default_store_bounds_for_arg): New.
871 (default_fn_abi_va_list_bounds_size): New.
872 (default_chkp_bound_type): New.
873 (default_chkp_bound_mode); New.
874 (default_builtin_chkp_function): New.
875 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
876 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
877 (TARGET_STORE_BOUNDS_FOR_ARG): New.
878 (TARGET_BUILTIN_CHKP_FUNCTION): New.
879 (TARGET_CHKP_BOUND_TYPE): New.
880 (TARGET_CHKP_BOUND_MODE): New.
881 * doc/tm.texi: Regenerated.
882 * langhooks.h (lang_hooks): Add chkp_supported field.
883 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
884 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
886 2013-10-29 Andrew Pinski <apinski@cavium.com>
888 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
889 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
890 fails, combining the branches anyways.
891 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
892 increases the number of combinings.
893 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
895 2013-10-29 Mike Stump <mikestump@comcast.net>
897 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
898 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
899 (make_vector_mode): Increase namebuf to 16.
900 (emit_insn_modes_h): When processing BImode, don't
901 also match partial int modes.
902 (emit_class_narrowest_mode): Likewise.
904 * config/bfin/bfin-modes.def: Add precision to PDI.
905 * config/m32c/m32c-modes.def: Add precision to PSI.
906 * config/msp430/msp430-modes.def: Add precision to PSI.
907 * config/rs6000/rs6000-modes.def: Add precision to PTI.
908 * config/sh/sh-modes.def: Add precision to PSI and PDI.
910 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
913 * config/sh/sh.md (*addc): Rename existing variations to ...
914 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
915 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
916 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
917 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
919 2013-10-29 Teresa Johnson <tejohnson@google.com>
922 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
923 insanities when updating probabilities.
925 2013-10-29 David Malcolm <dmalcolm@redhat.com>
927 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
928 prettyprinter for cgraph_node to reflect the conversion of the
929 symtable types to a C++ class hierarchy: it now *is* a
930 symtab_node_base, rather than having one (named "symbol").
932 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
934 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
935 __cilkrts_pop_frame. If matched, then return true for built-in
937 (expand_builtin): Added BUILT_IN_CILK_DETACH and
938 BUILT_IN_CILK_POP_FRAME case.
939 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
940 (lhs_cilk_detect_spawn): Likewise.
941 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
942 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
943 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
944 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
945 (LANG_HOOKS_CILKPLUS): Likewise.
946 * tree.h (CILK_SPAWN_FN): Likewise.
947 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
948 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
949 (OBJS): Added cilk-common.o.
950 (BUILTINS_DEF): Added cilk-builtins.def.
951 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
952 (lhd_cilk_detect_spawn): Likewise.
953 * langhooks.h (lang_hooks_for_cilkplus): New struct.
954 (struct lang_hooks): Added new field called "cilkplus."
955 * cilk-common.c: New file.
957 * cilk-builtins.def: Likewise.
958 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
959 "__cilk" macro and set it to 200.
960 * function.h (struct function::cilk_frame_decl): New field.
961 (struct function::is_cilk_function): Likewise.
962 (struct function::calls_cilk_spawn): Likewise.
963 * gimplify.c (gimplify_call_expr): Added a check if the function call
964 being gimplified is a spawn detach point. If so, then add pop_frame
965 and detach function calls.
966 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
967 for gimplifying _Cilk_spawn and _Cilk_sync statements.
968 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
969 function. If so, added a _Cilk_sync and gimplified it.
970 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
971 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
972 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
974 (can_inline_edge_p): Prevent inling of spawnee function.
975 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
976 for functions that use Cilk keywords.
977 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
978 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
979 CILK_SYNC_STMT cases.
980 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
982 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
983 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
984 * passes.texi (Cilk Keywords): New section that describes the compiler
985 code changes for handling Cilk Keywords.
987 2013-10-29 David Malcolm <dmalcolm@redhat.com>
989 Patch autogenerated by refactor_symtab.py from
990 https://github.com/davidmalcolm/gcc-refactoring-scripts
991 revision 58bb219cc090b2f4516a9297d868c245495ee622
993 * asan.c (asan_finish_file): Update for conversion of symtab types to
994 a true class hierarchy.
995 * cfgexpand.c (estimated_stack_frame_size): Likewise.
996 * cgraph.c (cgraph_get_body): Likewise.
997 (cgraph_get_create_real_symbol_node): Likewise.
998 (verify_cgraph_node): Likewise.
999 (verify_edge_corresponds_to_fndecl): Likewise.
1000 (verify_edge_count_and_frequency): Likewise.
1001 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
1002 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1003 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
1004 (cgraph_node_cannot_return): Likewise.
1005 (cgraph_set_pure_flag_1): Likewise.
1006 (cgraph_set_const_flag_1): Likewise.
1007 (cgraph_set_nothrow_flag_1): Likewise.
1008 (cgraph_make_node_local_1): Likewise.
1009 (cgraph_for_node_and_aliases): Likewise.
1010 (cgraph_for_node_thunks_and_aliases): Likewise.
1011 (cgraph_node_can_be_local_p): Likewise.
1012 (cgraph_node_cannot_be_local_p_1): Likewise.
1013 (cgraph_function_body_availability): Likewise.
1014 (dump_cgraph_node): Likewise.
1015 (cgraph_rtl_info): Likewise.
1016 (cgraph_mark_address_taken_node): Likewise.
1017 (cgraph_remove_node): Likewise.
1018 (cgraph_release_function_body): Likewise.
1019 (cgraph_update_edges_for_call_stmt_node): Likewise.
1020 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
1021 (cgraph_make_edge_direct): Likewise.
1022 (cgraph_resolve_speculation): Likewise.
1023 (cgraph_speculative_call_info): Likewise.
1024 (cgraph_turn_edge_to_speculative): Likewise.
1025 (cgraph_create_edge_1): Likewise.
1026 (cgraph_set_call_stmt): Likewise.
1027 (cgraph_node_for_asm): Likewise.
1028 (cgraph_add_thunk): Likewise.
1029 (cgraph_same_body_alias): Likewise.
1030 (cgraph_create_function_alias): Likewise.
1031 (cgraph_create_node): Likewise.
1032 (cgraph_create_empty_node): Likewise.
1033 (record_function_versions): Likewise.
1034 (used_from_object_file_p): Likewise.
1035 * cgraph.h (symtab_can_be_discarded): Likewise.
1036 (symtab_real_symbol_p): Likewise.
1037 (cgraph_mark_force_output_node): Likewise.
1038 (cgraph_edge_recursive_p): Likewise.
1039 (symtab_alias_target): Likewise.
1040 (varpool_all_refs_explicit_p): Likewise.
1041 (varpool_can_remove_if_no_refs): Likewise.
1042 (cgraph_only_called_directly_or_aliased_p): Likewise.
1043 (cgraph_next_function_with_gimple_body): Likewise.
1044 (cgraph_first_function_with_gimple_body): Likewise.
1045 (cgraph_function_with_gimple_body_p): Likewise.
1046 (cgraph_next_function): Likewise.
1047 (cgraph_first_function): Likewise.
1048 (cgraph_next_defined_function): Likewise.
1049 (cgraph_first_defined_function): Likewise.
1050 (varpool_next_defined_variable): Likewise.
1051 (varpool_first_defined_variable): Likewise.
1052 (varpool_next_static_initializer): Likewise.
1053 (varpool_first_static_initializer): Likewise.
1054 (varpool_next_variable): Likewise.
1055 (varpool_first_variable): Likewise.
1056 (varpool_node_name): Likewise.
1057 (varpool): Likewise.
1059 (is_a_helper <varpool_node>::test): Likewise.
1060 (is_a_helper <cgraph_node>::test): Likewise.
1061 (varpool_variable_node): Likewise.
1062 (cgraph_function_or_thunk_node): Likewise.
1063 (varpool_alias_target): Likewise.
1064 (cgraph_alias_target): Likewise.
1065 (cgraph_node_name): Likewise.
1066 (varpool_node_asm_name): Likewise.
1067 (cgraph_node_asm_name): Likewise.
1068 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
1069 (cgraph_rebuild_references): Likewise.
1070 (rebuild_cgraph_edges): Likewise.
1071 (record_eh_tables): Likewise.
1072 (build_cgraph_edges): Likewise.
1073 (mark_store): Likewise.
1074 (mark_load): Likewise.
1075 (mark_address): Likewise.
1076 (record_type_list): Likewise.
1077 (record_reference): Likewise.
1078 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
1079 (cgraph_materialize_clone): Likewise.
1080 (cgraph_function_versioning): Likewise.
1081 (cgraph_copy_node_for_versioning): Likewise.
1082 (update_call_expr): Likewise.
1083 (cgraph_find_replacement_node): Likewise.
1084 (cgraph_create_virtual_clone): Likewise.
1085 (cgraph_clone_node): Likewise.
1086 * cgraphunit.c (compile): Likewise.
1087 (output_weakrefs): Likewise.
1088 (output_in_order): Likewise.
1089 (expand_function): Likewise.
1090 (assemble_thunks_and_aliases): Likewise.
1091 (expand_thunk): Likewise.
1092 (mark_functions_to_output): Likewise.
1093 (handle_alias_pairs): Likewise.
1094 (analyze_functions): Likewise.
1095 (walk_polymorphic_call_targets): Likewise.
1096 (varpool_finalize_decl): Likewise.
1097 (process_function_and_variable_attributes): Likewise.
1098 (cgraph_process_same_body_aliases): Likewise.
1099 (analyze_function): Likewise.
1100 (cgraph_add_new_function): Likewise.
1101 (cgraph_finalize_function): Likewise.
1102 (referred_to_p): Likewise.
1103 (cgraph_reset_node): Likewise.
1104 (cgraph_process_new_functions): Likewise.
1105 (enqueue_node): Likewise.
1106 (decide_is_symbol_needed): Likewise.
1107 * coverage.c (coverage_compute_profile_id): Likewise.
1108 * dbxout.c (dbxout_expand_expr): Likewise.
1109 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
1110 (reference_to_unused): Likewise.
1111 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1112 * gimplify.c (unvisit_body): Likewise.
1113 (unshare_body): Likewise.
1114 * ipa-cp.c (ipcp_generate_summary): Likewise.
1115 (ipcp_decision_stage): Likewise.
1116 (identify_dead_nodes): Likewise.
1117 (decide_whether_version_node): Likewise.
1118 (decide_about_value): Likewise.
1119 (perhaps_add_new_callers): Likewise.
1120 (create_specialized_node): Likewise.
1121 (update_profiling_info): Likewise.
1122 (ipcp_propagate_stage): Likewise.
1123 (estimate_local_effects): Likewise.
1124 (good_cloning_opportunity_p): Likewise.
1125 (devirtualization_time_bonus): Likewise.
1126 (propagate_constants_accross_call): Likewise.
1127 (initialize_node_lattices): Likewise.
1128 (ipcp_cloning_candidate_p): Likewise.
1129 (determine_versionability): Likewise.
1130 (print_all_lattices): Likewise.
1131 (print_lattice): Likewise.
1132 (ipcp_discover_new_direct_edges): Likewise.
1133 * ipa-devirt.c (ipa_devirt): Likewise.
1134 (likely_target_p): Likewise.
1135 (update_type_inheritance_graph): Likewise.
1136 (possible_polymorphic_call_target_p): Likewise.
1137 (dump_possible_polymorphic_call_targets): Likewise.
1138 (devirt_variable_node_removal_hook): Likewise.
1139 (record_binfo): Likewise.
1140 (maybe_record_node): Likewise.
1141 (build_type_inheritance_graph): Likewise.
1142 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1143 (inline_generate_summary): Likewise.
1144 (inline_analyze_function): Likewise.
1145 (do_estimate_growth): Likewise.
1146 (simple_edge_hints): Likewise.
1147 (estimate_node_size_and_time): Likewise.
1148 (estimate_edge_devirt_benefit): Likewise.
1149 (compute_inline_parameters): Likewise.
1150 (estimate_function_body_sizes): Likewise.
1151 (compute_bb_predicates): Likewise.
1152 (initialize_inline_failed): Likewise.
1153 (dump_inline_summary): Likewise.
1154 (dump_inline_edge_summary): Likewise.
1155 * ipa-inline-transform.c (inline_transform): Likewise.
1156 (preserve_function_body_p): Likewise.
1157 (save_inline_function_body): Likewise.
1158 (inline_call): Likewise.
1159 (clone_inlined_nodes): Likewise.
1160 (can_remove_node_now_p): Likewise.
1161 (can_remove_node_now_p_1): Likewise.
1162 * ipa-inline.c (early_inliner): Likewise.
1163 (early_inline_small_functions): Likewise.
1164 (inline_always_inline_functions): Likewise.
1165 (ipa_inline): Likewise.
1166 (flatten_function): Likewise.
1167 (inline_small_functions): Likewise.
1168 (speculation_useful_p): Likewise.
1169 (recursive_inlining): Likewise.
1170 (update_caller_keys): Likewise.
1171 (reset_edge_caches): Likewise.
1172 (update_edge_key): Likewise.
1173 (edge_badness): Likewise.
1174 (relative_time_benefit): Likewise.
1175 (want_inline_self_recursive_call_p): Likewise.
1176 (want_inline_small_function_p): Likewise.
1177 (want_early_inline_function_p): Likewise.
1178 (num_calls): Likewise.
1179 (can_early_inline_edge_p): Likewise.
1180 (can_inline_edge_p): Likewise.
1181 (report_inline_failed_reason): Likewise.
1182 * ipa-profile.c (ipa_profile): Likewise.
1183 (ipa_propagate_frequency): Likewise.
1184 (ipa_propagate_frequency_1): Likewise.
1185 (ipa_profile_generate_summary): Likewise.
1186 * ipa-prop.c (ipcp_transform_function): Likewise.
1187 (read_replacements_section): Likewise.
1188 (ipa_prop_read_section): Likewise.
1189 (ipa_modify_call_arguments): Likewise.
1190 (ipa_print_node_params): Likewise.
1191 (propagate_controlled_uses): Likewise.
1192 (update_indirect_edges_after_inlining): Likewise.
1193 (remove_described_reference): Likewise.
1194 (ipa_make_edge_direct_to_target): Likewise.
1195 (ipa_analyze_node): Likewise.
1196 (ipa_analyze_params_uses): Likewise.
1197 (ipa_compute_jump_functions): Likewise.
1198 (ipa_get_callee_param_type): Likewise.
1199 (ipa_print_node_jump_functions): Likewise.
1200 (ipa_initialize_node_params): Likewise.
1201 (ipa_populate_param_decls): Likewise.
1202 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
1203 (write_agg_replacement_chain): Likewise.
1204 (ipa_write_node_info): Likewise.
1205 (ipa_edge_duplication_hook): Likewise.
1206 (try_decrement_rdesc_refcount): Likewise.
1207 * ipa-pure-const.c (propagate_nothrow): Likewise.
1208 (propagate_pure_const): Likewise.
1209 (pure_const_read_summary): Likewise.
1210 (pure_const_write_summary): Likewise.
1211 (analyze_function): Likewise.
1212 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
1213 (ipa_ref_referring_ref_list): Likewise.
1214 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
1215 (ipa_remove_stmt_references): Likewise.
1216 (ipa_find_reference): Likewise.
1217 (ipa_dump_referring): Likewise.
1218 (ipa_dump_references): Likewise.
1219 (ipa_record_reference): Likewise.
1220 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
1221 (ipa_reference_write_optimization_summary): Likewise.
1222 (write_node_summary_p): Likewise.
1223 (propagate): Likewise.
1224 (read_write_all_from_decl): Likewise.
1225 (generate_summary): Likewise.
1226 (analyze_function): Likewise.
1227 (propagate_bits): Likewise.
1228 (ipa_reference_get_not_written_global): Likewise.
1229 (ipa_reference_get_not_read_global): Likewise.
1230 * ipa-split.c (execute_split_functions): Likewise.
1231 (split_function): Likewise.
1232 * ipa-utils.c (ipa_merge_profiles): Likewise.
1233 (dump_cgraph_node_set): Likewise.
1234 (ipa_reverse_postorder): Likewise.
1235 (ipa_edge_within_scc): Likewise.
1236 (ipa_get_nodes_in_cycle): Likewise.
1237 (ipa_free_postorder_info): Likewise.
1238 (ipa_reduced_postorder): Likewise.
1239 (searchc): Likewise.
1240 (recursive_call_p): Likewise.
1241 * ipa.c (ipa_cdtor_merge): Likewise.
1242 (record_cdtor_fn): Likewise.
1243 (function_and_variable_visibility): Likewise.
1244 (varpool_externally_visible_p): Likewise.
1245 (cgraph_externally_visible_p): Likewise.
1246 (comdat_can_be_unshared_p): Likewise.
1247 (comdat_can_be_unshared_p_1): Likewise.
1248 (address_taken_from_non_vtable_p): Likewise.
1249 (ipa_discover_readonly_nonaddressable_vars): Likewise.
1250 (symtab_remove_unreachable_nodes): Likewise.
1251 (walk_polymorphic_call_targets): Likewise.
1252 (process_references): Likewise.
1253 (enqueue_node): Likewise.
1254 (has_addr_references_p): Likewise.
1255 (cgraph_non_local_node_p_1): Likewise.
1256 * is-a.h (varpool_analyze_node): Likewise.
1257 * lto-cgraph.c (input_symtab): Likewise.
1258 (merge_profile_summaries): Likewise.
1259 (input_cgraph_1): Likewise.
1260 (input_edge): Likewise.
1261 (input_varpool_node): Likewise.
1262 (input_node): Likewise.
1263 (input_overwrite_node): Likewise.
1264 (compute_ltrans_boundary): Likewise.
1265 (output_refs): Likewise.
1266 (lto_output_varpool_node): Likewise.
1267 (lto_output_node): Likewise.
1268 (reachable_from_other_partition_p): Likewise.
1269 (referenced_from_other_partition_p): Likewise.
1270 (lto_output_edge): Likewise.
1271 (output_node_opt_summary): Likewise.
1272 (add_node_to): Likewise.
1273 (reachable_from_this_partition_p): Likewise.
1274 (lto_set_symtab_encoder_in_partition): Likewise.
1275 (lto_symtab_encoder_in_partition_p): Likewise.
1276 (lto_set_symtab_encoder_encode_initializer): Likewise.
1277 (lto_symtab_encoder_encode_initializer_p): Likewise.
1278 (lto_set_symtab_encoder_encode_body): Likewise.
1279 (lto_symtab_encoder_encode_body_p): Likewise.
1280 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1282 * lto-streamer-in.c (lto_read_body): Likewise.
1283 (fixup_call_stmt_edges): Likewise.
1284 (fixup_call_stmt_edges_1): Likewise.
1285 * lto-streamer-out.c (produce_symtab): Likewise.
1286 (output_symbol_p): Likewise.
1287 (write_symbol): Likewise.
1288 (lto_output): Likewise.
1289 (copy_function): Likewise.
1290 (output_function): Likewise.
1291 * passes.c (function_called_by_processed_nodes_p): Likewise.
1292 (ipa_write_optimization_summaries): Likewise.
1293 (ipa_write_summaries): Likewise.
1294 (do_per_function_toporder): Likewise.
1295 (do_per_function): Likewise.
1296 (dump_passes): Likewise.
1297 * symtab.c (symtab_semantically_equivalent_p): Likewise.
1298 (symtab_nonoverwritable_alias): Likewise.
1299 (symtab_nonoverwritable_alias_1): Likewise.
1300 (symtab_for_node_and_aliases): Likewise.
1301 (symtab_resolve_alias): Likewise.
1302 (fixup_same_cpp_alias_visibility): Likewise.
1303 (symtab_alias_ultimate_target): Likewise.
1304 (symtab_used_from_object_file_p): Likewise.
1305 (verify_symtab_base): Likewise.
1306 (dump_symtab_base): Likewise.
1307 (symtab_node_name): Likewise.
1308 (symtab_node_asm_name): Likewise.
1309 (symtab_dissolve_same_comdat_group_list): Likewise.
1310 (symtab_add_to_same_comdat_group): Likewise.
1311 (symtab_unregister_node): Likewise.
1312 (symtab_insert_node_to_hashtable): Likewise.
1313 (symtab_register_node): Likewise.
1314 (unlink_from_assembler_name_hash): Likewise.
1315 (insert_to_assembler_name_hash): Likewise.
1316 (eq_assembler_name): Likewise.
1317 (hash_node_by_assembler_name): Likewise.
1318 (eq_node): Likewise.
1319 (hash_node): Likewise.
1320 * toplev.c (wrapup_global_declaration_2): Likewise.
1321 * trans-mem.c (ipa_tm_execute): Likewise.
1322 (ipa_tm_transform_clone): Likewise.
1323 (ipa_tm_transform_transaction): Likewise.
1324 (ipa_tm_transform_calls_redirect): Likewise.
1325 (ipa_tm_insert_gettmclone_call): Likewise.
1326 (ipa_tm_insert_irr_call): Likewise.
1327 (ipa_tm_create_version): Likewise.
1328 (ipa_tm_create_version_alias): Likewise.
1329 (ipa_tm_mark_forced_by_abi_node): Likewise.
1330 (ipa_tm_mark_force_output_node): Likewise.
1331 (ipa_tm_diagnose_tm_safe): Likewise.
1332 (ipa_tm_mayenterirr_function): Likewise.
1333 (ipa_tm_scan_irr_function): Likewise.
1334 (ipa_tm_note_irrevocable): Likewise.
1335 (ipa_tm_scan_calls_clone): Likewise.
1336 (get_cg_data): Likewise.
1337 * tree-eh.c (tree_could_trap_p): Likewise.
1338 * tree-emutls.c (ipa_lower_emutls): Likewise.
1339 (create_emultls_var): Likewise.
1340 (lower_emutls_function_body): Likewise.
1341 (gen_emutls_addr): Likewise.
1342 (emutls_decl): Likewise.
1343 (new_emutls_decl): Likewise.
1344 * tree-inline.c (tree_function_versioning): Likewise.
1345 (optimize_inline_calls): Likewise.
1346 (expand_call_inline): Likewise.
1347 (estimate_num_insns): Likewise.
1348 (copy_bb): Likewise.
1349 (delete_unreachable_blocks_update_callgraph): Likewise.
1350 * tree-nested.c (gimplify_all_functions): Likewise.
1351 (create_nesting_tree): Likewise.
1352 (check_for_nested_with_variably_modified): Likewise.
1353 * tree-pretty-print.c (dump_function_header): Likewise.
1354 * tree-profile.c (tree_profiling): Likewise.
1355 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
1356 (modify_function): Likewise.
1357 (convert_callers): Likewise.
1358 (convert_callers_for_node): Likewise.
1359 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
1360 (associate_varinfo_to_alias): Likewise.
1361 (create_variable_info_for): Likewise.
1362 (get_constraint_for_ssa_var): Likewise.
1363 * tree-vectorizer.c (increase_alignment): Likewise.
1364 * tree.c (find_decls_types_in_var): Likewise.
1365 (find_decls_types_in_node): Likewise.
1366 (free_lang_data_in_decl): Likewise.
1367 * value-prof.c (gimple_ic_transform): Likewise.
1368 (gimple_ic): Likewise.
1369 (check_ic_target): Likewise.
1370 (init_node_map): Likewise.
1371 * varasm.c (decl_binds_to_current_def_p): Likewise.
1372 (default_binds_local_p_1): Likewise.
1373 (dump_tm_clone_pairs): Likewise.
1374 (assemble_alias): Likewise.
1375 (find_decl): Likewise.
1376 (mark_decl_referenced): Likewise.
1377 * varpool.c (varpool_for_node_and_aliases): Likewise.
1378 (varpool_extra_name_alias): Likewise.
1379 (varpool_create_variable_alias): Likewise.
1380 (add_new_static_var): Likewise.
1381 (varpool_finalize_named_section_flags): Likewise.
1382 (varpool_remove_unreferenced_decls): Likewise.
1383 (enqueue_node): Likewise.
1384 (varpool_assemble_decl): Likewise.
1385 (assemble_aliases): Likewise.
1386 (varpool_analyze_node): Likewise.
1387 (cgraph_variable_initializer_availability): Likewise.
1388 (varpool_add_new_variable): Likewise.
1389 (ctor_for_folding): Likewise.
1390 (dump_varpool_node): Likewise.
1391 (varpool_remove_initializer): Likewise.
1392 (varpool_remove_node): Likewise.
1393 (varpool_node_for_decl): Likewise.
1394 (varpool_create_empty_node): Likewise.
1395 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
1396 (ix86_get_function_versions_dispatcher): Likewise.
1398 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1400 * cgraph.h (symtab_node_base): Convert to a class;
1401 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
1402 chain_next/prev from symtab_node_def.
1403 (cgraph_node): Inherit from symtab_node; add GTY option
1404 tag ("SYMTAB_FUNCTION").
1405 (varpool_node): Inherit from symtab_node; add GTY option
1406 tag ("SYMTAB_VARIABLE").
1407 (symtab_node_def): Remove.
1408 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
1409 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
1410 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
1411 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
1413 * ipa-ref.h (symtab_node_def): Drop.
1414 (symtab_node): Change underlying type from symtab_node_def to
1416 (const_symtab_node): Likwise.
1418 * is-a.h: Update examples in comment.
1420 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
1421 (assembler_name_hash): Likewise.
1423 2013-10-29 Martin Liska <marxin.liska@gmail.com>
1425 * doc/tree-ssa.texi (gimple_phi_result): Document.
1426 (gimple_phi_num_args, gimple_phi_arg): Likewise.
1427 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
1428 (PHI_RESULT, PHI_NUM_ARGS): Remove.
1429 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
1431 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1433 * expr.h: Revert change and include tree-core.h.
1434 * rtl.h: Revert change and don't include tree-core.h.
1436 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1438 * config/darwin.c: Include gimple.h.
1439 * config/i386/winnt.c: Likewise.
1441 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
1443 PR tree-optimization/19831
1444 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
1446 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1448 * tree-outof-ssa.h: Remove include files.
1449 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
1451 * tree-ssa-coalesce.c: Likewise.
1452 * cfgexpand.c: Likewise.
1453 * tree-ssa-ter.c: Likewise.
1454 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
1455 * lto-streamer.h: Likewise.
1456 * cgraphbuild.c: Add gimple.h to include list.
1457 * data-streamer-in.c: Likewise.
1458 * ipa-cp.c: Likewise.
1459 * tree-streamer.c: Likewise.
1460 * lto-compress.c: Likewise.
1461 * ipa-reference.c: Likewise.
1462 * data-streamer-out.c: Likewise.
1463 * lto-cgraph.c: Likewise.
1464 * cgraphclones.c: Likewise.
1465 * ipa-utils.c: Likewise.
1466 * data-streamer.c: Likewise.
1467 * ipa-split.c: Likewise.
1468 * lto-section-in.c: Likewise.
1469 * tree-streamer-out.c: Likewise.
1470 * ipa-prop.c: Likewise.
1471 * tree-streamer-in.c: Likewise.
1472 * symtab.c: Likewise.
1473 * opts-global.c: Likewise.
1474 * lto-opts.c: Likewise.
1475 * lto-section-out.c: Likewise.
1476 * lto-streamer.c: Likewise.
1477 * rtl.h: Add tree-core.h to include list.
1478 * expr.h: Remove tree-core.h from include list.
1479 * gimple.h: Likewise.
1480 * ipa-utils.h: Likewise.
1481 * streamer-hooks.h: Likewise.
1482 * streamer-hooks.c: Include input.h.
1484 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1486 * config/arm/arm.c (cortexa7_extra_costs): New table.
1487 (arm_cortex_a7_tune): New.
1488 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
1490 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
1492 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
1494 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1496 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
1497 to use autogenerated markers for a class-hierarchy, every class
1498 must have a GTY marker.
1499 * gengtype.h (struct type): Add linked list of subclasses to
1500 the "s" member of the union.
1501 (add_subclass): New decl.
1502 * gengtype-state.c (read_state_struct_type): Set up subclass
1504 * gengtype.c (get_ultimate_base_class): New.
1505 (add_subclass): New.
1506 (new_structure): Set up subclass linked list.
1507 (set_gc_used_type): Propagate usage information to subclasses.
1508 (output_mangled_typename): Use get_ultimate_base_class.
1509 (walk_subclasses): Use the subclass linked list, avoiding an
1510 O(N^2) when writing out all types.
1511 (walk_type): Issue an error if the base class is missing a tag,
1512 rather than generating bogus C code. Add a gcc_unreachable
1513 default case, in case people omit tags from concrete subclasses,
1514 or get the values wrong.
1515 (write_func_for_structure): Issue an error for subclasses for
1516 which the base doesn't have a "desc", since otherwise the
1517 autogenerated routines for the base would silently fail to visit
1518 any subclass fields.
1519 (write_root): Use get_ultimate_base_class, tweaking constness of
1520 tp to match that function's signature.
1522 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1524 * doc/gty.texi (GTY Options): Add note about inheritance to
1525 description of desc and tag.
1526 (Inheritance and GTY): New.
1528 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1530 * gengtype-parse.c (opts_have): Drop "static" so that
1531 we can use this from gengtype.c.
1532 * gengtype.c (set_gc_used_type): Mark any base class as used;
1533 update field traversal to visit inherited fields.
1534 (output_mangled_typename): Convert references to classes within
1535 an inheritance hierarchy to reference the ultimate base class,
1536 since only it will have gt_ functions.
1537 (get_string_option): New.
1538 (walk_subclasses): New.
1539 (walk_type): Treat GTY structs that have a "desc" as being the
1540 root of an inheritance hierarchy. Generate a switch on it
1541 within the marking function which walks all subclasses, adding
1542 cases for them via walk_subclasses. For subclasses, visit all
1543 fields of the type (including inherited ones).
1544 (write_func_for_structure): Don't write fns for subclasses, only
1545 for the ultimate base class within an inheritance hierarchy.
1546 Subclasses-marking will be handled by the base class marking functions.
1547 (write_types): Likewise.
1548 (write_local_func_for_structure): Likewise.
1549 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
1550 a "tag" option (and are thus concrete subclasses).
1551 (write_root): Use the marker function for the ultimate base class.
1552 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
1553 (opts_have): Add declaration.
1555 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
1557 * lra-spills.c (lra_final_code_change): Remove useless move insns
1558 originated from moves of pseudos.
1560 2013-10-28 Jeff Law <law@redhat.com>
1562 * lower-subreg.c (resolve_simple_move): Fix comment typo.
1564 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
1566 * df-scan.c (df_collection_rec): Adjust.
1567 (copy_defs): New constant.
1568 (copy_uses): Likewise.
1569 (copy_eq_uses): Likewise.
1570 (copy_mw): Likewise.
1571 (copy_all): Likewise.
1572 (df_insn_rescan): Adjust.
1573 (df_notes_rescan): Likewise.
1574 (df_swap_refs): Likewise.
1575 (df_sort_and_compress_refs): Likewise.
1576 (df_sort_and_compress_mws): Likewise.
1577 (df_install_refs): Likewise.
1578 (df_install_mws): Likewise.
1579 (df_refs_add_to_chains): Add flags parameter controlling which vectors
1581 (df_bb_refs_record): Adjust.
1582 (df_record_entry_block_defs): Likewise.
1583 (df_record_exit_block_defs): Likewise.
1584 (df_refs_verify): Likewise.
1585 (df_mws_verify): Likewise.
1586 (df_insn_refs_verify): Likewise.
1587 (df_bb_verify): Likewise.
1588 * ipa-pure-const.c (finish_state): Remove.
1589 (propagate): Adjust.
1590 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
1591 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
1592 var-tracking.c: Adjust.
1593 * vec.c (stack_vecs): Remove.
1594 (register_stack_vec): Likewise.
1595 (stack_vec_register_index): Likewise.
1596 (unregister_stack_vec): Likewise.
1597 * vec.h (struct va_stack): Remove.
1598 (struct vec<T, A, vl_ptr>): Specialize as
1599 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
1600 allocation strategy compatable with the vl_ptr layout.
1601 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
1602 specialization anyway.
1603 (class stack_vec): New class.
1604 (vec_stack_alloc): Remove.
1605 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
1607 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1608 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1609 Sergey Lega <sergey.s.lega@intel.com>
1610 Anna Tikhonova <anna.tikhonova@intel.com>
1611 Ilya Tocar <ilya.tocar@intel.com>
1612 Andrey Turetskiy <andrey.turetskiy@intel.com>
1613 Ilya Verbin <ilya.verbin@intel.com>
1614 Kirill Yukhin <kirill.yukhin@intel.com>
1615 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1617 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
1618 (*prefetch_avx512pf_<mode>): New.
1619 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
1620 (avx512f_maskcmp<mode>3): Ditto.
1621 (vashrv16si3): Ditto.
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>
1633 * config/i386/i386.md (any_truncate): New.
1634 (trunsuffix): Ditto.
1635 * config/i386/predicates.md (const_8_to_9_operand): New.
1636 (const_10_to_11_operand): Ditto.
1637 (const_12_to_13_operand): Ditto.
1638 (const_14_to_15_operand): Ditto.
1639 (const_16_to_19_operand): Ditto.
1640 (const_20_to_23_operand): Ditto.
1641 (const_24_to_27_operand): Ditto.
1642 (const_28_to_31_operand): Ditto.
1643 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
1644 (cvtusi2<ssescalarmodesuffix>32): New.
1645 (cvtusi2<ssescalarmodesuffix>64): Ditto.
1646 (ufloatv16siv16sf2): Ditto.
1647 (avx512f_fix_notruncv16sfv16si): Ditto.
1648 (avx512f_ufix_notruncv16sfv16si): Ditto.
1649 (avx512f_vcvtss2usi): Ditto.
1650 (avx512f_vcvtss2usiq): Ditto.
1651 (avx512f_vcvttss2usi): Ditto.
1652 (avx512f_vcvttss2usiq): Ditto.
1653 (avx512f_vcvtsd2usi): Ditto.
1654 (avx512f_vcvtsd2usiq): Ditto.
1655 (avx512f_vcvttsd2usi): Ditto.
1656 (avx512f_vcvttsd2usiq): Ditto.
1657 (ufloatv8siv8df): Ditto.
1658 (avx512f_cvtdq2pd512_2): Ditto.
1659 (avx512f_cvtpd2dq512): Ditto.
1660 (avx512f_ufix_notruncv8dfv8si): Ditto.
1661 (avx512f_cvtpd2ps512): Ditto.
1662 (vec_unpacks_lo_v16sf): Ditto.
1663 (vec_unpacks_hi_v16sf): Ditto.
1664 (vec_unpacks_float_hi_v16si): Ditto.
1665 (vec_unpacks_float_lo_v16si): Ditto.
1666 (avx512f_unpckhps512): Ditto.
1667 (avx512f_unpcklps512): Ditto.
1668 (avx512f_movshdup512): Ditto.
1669 (avx512f_movsldup512): Ditto.
1670 (vec_extract_lo_v32hi): Ditto.
1671 (vec_extract_hi_v32hi): Ditto.
1672 (vec_extract_lo_v64qi): Ditto.
1673 (vec_extract_hi_v64qi): Ditto.
1674 (avx512f_unpckhpd512): Ditto.
1675 (avx512f_movddup512): Ditto.
1676 (avx512f_unpcklpd512): Ditto.
1677 (*avx512f_unpcklpd512): Ditto.
1678 (avx512f_shufps512_1): Ditto.
1679 (avx512f_shufpd512_1): Ditto.
1680 (avx512f_interleave_highv8di): Ditto.
1681 (avx512f_interleave_lowv8di): Ditto.
1682 (PMOV_DST_MODE): Ditto.
1683 (pmov_src_mode): Ditto.
1684 (pmov_src_lower): Ditto.
1686 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
1687 (*avx512f_<code>v8div16qi2): Ditto.
1688 (*avx512f_<code>v8div16qi2_store): Ditto.
1689 (vec_widen_umult_even_v16si): Ditto.
1690 (*vec_widen_umult_even_v16si): Ditto.
1691 (vec_widen_smult_even_v16si): Ditto.
1692 (*vec_widen_smult_even_v16si): Ditto.
1693 (avx512f_interleave_highv16si): Ditto.
1694 (avx512f_interleave_lowv16si): Ditto.
1695 (avx512f_<code>v16qiv16si2): Ditto.
1696 (avx512f_<code>v16hiv16si2): Ditto.
1697 (avx512f_<code>v8qiv8di2): Ditto.
1698 (avx512f_<code>v8hiv8di2): Ditto.
1699 (avx512f_<code>v8siv8di2): Ditto.
1700 (avx512cd_maskb_vec_dupv8di): Ditto.
1701 (avx512cd_maskw_vec_dupv16si): Ditto.
1702 (avx512f_vcvtph2ps512): Ditto.
1703 (avx512f_vcvtps2ph512): Ditto.
1704 (VEC_EXTRACT_MODE): Extened with wider modes.
1705 (VEC_PERM_AVX2): Ditto.
1706 (VEC_PERM_CONST): Ditto.
1708 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
1710 * config/arc/arc.c (arc_ccfsm_post_advance):
1711 Add comment about TYPE_RETURN.
1713 2013-10-28 Bin Cheng <bin.cheng@arm.com>
1715 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
1716 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
1717 (strip_offset): Convert offset to unsigned number.
1719 2013-10-27 Tom de Vries <tom@codesourcery.com>
1721 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
1722 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
1724 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
1726 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
1727 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
1728 (TARGET_BIG_ENDIAN): New macro.
1729 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
1730 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
1731 * config/sh/sh.c: Likewise.
1732 * config/sh/sh.h: Likewise.
1734 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
1736 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
1737 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
1739 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1741 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
1743 (cmpstr_t, cmpstrsi): Fix formatting.
1745 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1748 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
1749 addressing, do not use general_operand for memory operands.
1751 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
1754 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1755 * lra-spills.c (lra_final_code_change): Remove useless move insns.
1757 2013-10-26 Jeff Law <law@redhat.com>
1759 * predict.c (PRED_MUDFLAP): Remove.
1760 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
1762 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
1763 (OBJS): Remove tree-nomudflap.o
1764 (GTFILES): Remove tree-mudflap.c
1765 * builtins.c (expand_builtin_alloc): Remove mudflap support.
1766 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
1767 (mfwrap_spec, mflib_spec): Likewise.
1768 (cpp_unique_options, cc1_options, static_specs): Likewise.
1769 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
1770 * passes.def: Likewise.
1771 * toplev.c (compile_file, process_options): Likewise.
1772 * tree-inline.c (copy_tree_r): Likewise.
1773 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
1774 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
1775 (build_constant_desc, output_constant_def_contents): Likewise.
1776 (categorize_decl_for_section): Likewise.
1777 * tree-mudflap.c: Removed.
1778 * tree-mudflap.h: Removed.
1779 * tree-nomudflap.c: Removed.
1780 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
1781 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
1782 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
1783 * config/sol2.h (MFLIB_SPEC): Likewise.
1784 * doc/install.texi: Remove mudflap references.
1785 * doc/passes.texi: Similarly.
1786 * doc/sourcebuild.texi: Similarly.
1787 * doc/invoke.texi: Remove mudlfap related options.
1789 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1791 PR rtl-optimization/58759
1792 * lra-constraints.c (lra_constraints): Remove wrong condition to
1793 remove insn setting up an equivalent pseudo.
1795 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1797 * config/rs6000/rs6000-protos.h
1798 (rs6000_secondary_memory_needed_mode): New prototype.
1799 * config/rs6000/rs6000.c: Include ira.h.
1800 (TARGET_LRA_P): Redefine.
1801 (rs6000_legitimate_offset_address_p): Call
1802 legitimate_constant_pool_address_p in strict mode for LRA.
1803 (rs6000_legitimate_address_p): Ditto.
1804 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
1805 (rs6000_emit_move): Add LRA version of code to generate load/store
1807 (rs6000_secondary_memory_needed_mode): New.
1808 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
1809 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
1810 constants, memory, and FP registers.
1811 (rs6000_lra_p): New.
1812 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
1813 * config/rs6000/rs6000.opt (mlra): New option.
1814 * lra-spills.c (lra_final_code_change): Remove useless move insns.
1816 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
1818 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
1819 has_single_use () and not do the conversion if has_single_use ()
1820 returns false for the multiplication result.
1822 2013-10-25 David Malcolm <dmalcolm@redhat.com>
1824 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
1825 to "NODE", since this works on a "tree", not an
1827 (CONSTANT_CLASS_P): Likewise.
1830 (INDIRECT_REF_P): Likewise.
1831 (REFERENCE_CLASS_P): Likewise.
1832 (COMPARISON_CLASS_P): Likewise.
1833 (UNARY_CLASS_P): Likewise.
1834 (BINARY_CLASS_P): Likewise.
1835 (STATEMENT_CLASS_P): Likewise.
1836 (VL_EXP_CLASS_P): Likewise.
1837 (EXPRESSION_CLASS_P): Likewise.
1838 (IS_TYPE_OR_DECL_P): Likewise.
1840 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
1842 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
1843 ADDR_EXPR in the defining statement.
1845 2013-10-25 Richard Biener <rguenther@suse.de>
1847 PR tree-optimization/58626
1848 * tree-loop-distribution.c (enum rdg_dep_type): Remove
1849 anti_dd, output_dd and input_dd.
1850 (struct rdg_edge): Remove level and relation members.
1851 (RDGE_LEVEL, RDGE_RELATION): Remove.
1852 (dot_rdg_1): Adjust.
1853 (create_rdg_edge_for_ddr): Remove.
1854 (create_rdg_edges_for_scalar): Adjust.
1855 (create_edge_for_control_dependence): Likewise.
1856 (create_rdg_edges): Split into ...
1857 (create_rdg_flow_edges): ... this
1858 (create_rdg_cd_edges): ... and this.
1860 (build_rdg): Likewise, do not compute data dependences or
1862 (pg_add_dependence_edges): New function.
1864 (distribute_loop): First apply all non-dependence based
1865 partition mergings. Then compute dependences between partitions
1866 and merge and order partitions according to them.
1868 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
1870 PR rtl-optimization/58831
1871 * alias.c (init_alias_analysis): At the beginning of each iteration,
1872 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
1874 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
1876 * recog.c (search_ofs): New static variable moved from...
1877 (peep2_find_free_register): ...here.
1878 (peephole2_optimize): Initialize it.
1880 2013-10-25 Tobias Burnus <burnus@net-b.de>
1882 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
1884 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
1886 * config/i386/i386.h (TARGET_MPX): New define.
1887 (TARGET_MPX_P): Ditto.
1889 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
1891 * config/i386/constraints.md (B): New.
1894 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
1895 * config/i386/i386-modes.def (BND32): New.
1897 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
1898 * config/i386/i386.c (isa_opts): Add mmpx.
1899 (regclass_map): Add bound registers.
1900 (dbx_register_map): Likewise.
1901 (dbx64_register_map): Likewise.
1902 (svr4_dbx_register_map): Likewise.
1904 (ix86_option_override_internal): Support MPX ISA.
1905 (ix86_conditional_register_usage): Support bound registers.
1906 (print_reg): Likewise.
1907 (ix86_code_end): Add MPX bnd prefix.
1908 (output_set_got): Likewise.
1909 (ix86_output_call_insn): Likewise.
1910 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
1911 (ix86_print_operand_punct_valid_p): Likewise.
1912 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
1914 (ix86_class_likely_spilled_p): Add bound regs support.
1915 (ix86_hard_regno_mode_ok): Likewise.
1916 (x86_order_regs_for_local_alloc): Likewise.
1917 (ix86_bnd_prefixed_insn_p): New.
1918 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
1919 (FIXED_REGISTERS): Add bound registers.
1920 (CALL_USED_REGISTERS): Likewise.
1921 (REG_ALLOC_ORDER): Likewise.
1922 (HARD_REGNO_NREGS): Likewise.
1924 (VALID_BND_REG_MODE): New.
1925 (FIRST_BND_REG): New.
1926 (LAST_BND_REG): New.
1927 (reg_class): Add BND_REGS.
1928 (REG_CLASS_NAMES): Likewise.
1929 (REG_CLASS_CONTENTS): Likewise.
1931 (ANY_BND_REG_P): New.
1933 (HI_REGISTER_NAMES): Add bound registers.
1934 * config/i386/i386.md (UNSPEC_BNDMK): New.
1935 (UNSPEC_BNDMK_ADDR): New.
1936 (UNSPEC_BNDSTX): New.
1937 (UNSPEC_BNDLDX): New.
1938 (UNSPEC_BNDLDX_ADDR): New.
1939 (UNSPEC_BNDCL): New.
1940 (UNSPEC_BNDCU): New.
1941 (UNSPEC_BNDCN): New.
1942 (UNSPEC_MPX_FENCE): New.
1945 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
1946 (length_immediate): Likewise.
1947 (prefix_0f): Likewise.
1949 (prefix_rep): Check for bnd prefix.
1950 (length_nobnd): New.
1951 (length): Use length_nobnd if specified.
1956 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
1959 (simple_return_internal): Likewise.
1960 (simple_return_pop_internal): Likewise.
1961 (*indirect_jump): Add MPX bnd prefix.
1962 (*tablejump_1): Likewise.
1963 (simple_return_internal_long): Likewise.
1964 (simple_return_indirect_internal): Likewise.
1968 (*mov<mode>_internal_mpx): New.
1969 (<mode>_<bndcheck>): New.
1970 (*<mode>_<bndcheck>): New.
1975 * config/i386/predicates.md (lea_address_operand): Rename to...
1976 (address_no_seg_operand): ... this.
1977 (address_mpx_no_base_operand): New.
1978 (address_mpx_no_index_operand): New.
1979 (bnd_mem_operator): New.
1980 * config/i386/i386.opt (mmpx): New.
1981 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
1982 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
1984 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
1986 * mode-classes.def (MODE_POINTER_BOUNDS): New.
1987 * tree.def (POINTER_BOUNDS_TYPE): New.
1988 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
1989 (POINTER_BOUNDS_MODE): New.
1990 (make_pointer_bounds_mode): New.
1991 * machmode.h (POINTER_BOUNDS_MODE_P): New.
1992 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
1993 (layout_type): Support POINTER_BOUNDS_TYPE.
1994 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
1995 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
1996 (type_contains_placeholder_1): Likewise.
1997 * tree.h (POINTER_BOUNDS_TYPE_P): New.
1998 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
1999 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
2001 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2003 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
2004 mode of address computation when calling memory_address_addr_space.
2006 2013-08-24 Richard Henderson <rth@twiddle.net>
2009 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
2010 instead of create_convert_operand_to.
2011 (maybe_emit_sync_lock_test_and_set): Likewise.
2012 (expand_atomic_compare_and_swap): Likewise.
2013 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
2015 2013-08-24 Sriraman Tallam <tmsriram@google.com>
2017 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
2018 Change param_is to use the struct and not the pointer to the struct.
2020 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
2022 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
2023 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2024 Move to targhooks.c.
2025 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
2026 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2027 Relocate from gimplify.c.
2028 * targhooks.h: Add 2 prototypes.
2029 * tree.h: Delete 2 prototypes.
2031 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2033 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
2034 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
2036 2013-10-24 Cong Hou <congh@google.com>
2038 * convert.c (convert_to_real): Guard those unsafe math function
2039 convertions with flag_unsafe_math_optimizations. Handle sqrt()
2042 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
2045 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
2047 (cgraph_create_edge): Use the new argument.
2048 (cgraph_create_indirect_edge): Likewise.
2050 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
2052 * gcc/config/arc/arc.c (arc_ccfsm_post_advance): Also handle
2054 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
2055 changes statep->state.
2057 2013-10-24 Tobias Burnus <burnus@net-b.de>
2060 * tree-cfg.c (replace_loop_annotate): New function.
2061 (execute_build_cfg): Call it.
2062 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
2063 * internal-fn.c (expand_ANNOTATE): New function.
2064 * internal-fn.def (ANNOTATE): Define as new internal function.
2065 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
2066 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
2067 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
2068 * doc/extend.texi (Pragmas): Document #pragma ivdep.
2069 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
2071 2013-10-17 Ian Bolton <ian.bolton@arm.com>
2072 Marcus Shawcroft <marcus.shawcroft@arm.com>
2074 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
2075 Special case reload SP+C into none GENERAL_REGS.
2077 2013-10-24 Michael Matz <matz@suse.de>
2079 * gengtype.c (is_file_equal): Check that files will be same length.
2081 2013-10-25 Christian Bruel <christian.bruel@st.com>
2083 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
2084 * gcc/config/sh/t-sh (sh-mem.o): New rule.
2085 * gcc/config/sh/sh-mem.cc (expand_block_move): Moved here.
2086 (sh_expand_cmpstr): New function.
2087 * gcc/config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
2088 * gcc/config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
2089 * gcc/config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
2090 (rotlhi3_8): Rename.
2092 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2094 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
2095 * Makefile.in (CXXDEPMODE): Assign and change users.
2096 (CCDEPMODE): Delete.
2097 * configure: Regenerate.
2099 2013-10-23 David Malcolm <dmalcolm@redhat.com>
2101 * gengtype-parse.c (require_without_advance): New.
2102 (type): For GTY-marked types that are not GTY((user)), parse any
2103 base classes, requiring them to be single-inheritance, and not
2104 be templates. For non-GTY-marked types and GTY((user)),
2105 continue to skip over any C++ inheritance specification.
2106 * gengtype-state.c (state_writer::write_state_struct_type):
2107 Write base class of type (if any).
2108 (read_state_struct_type): Read base class of type (if any).
2109 * gengtype.c (new_structure): Add a "base_class" parameter.
2110 (create_optional_field_): Update for new parameter to new_structure.
2111 (adjust_field_rtx_def): Likewise.
2112 (adjust_field_tree_exp): Likewise.
2113 * gengtype.h (struct type): Add "base_class" field to the s
2115 (new_structure): Add "base" parameter.
2117 2013-10-23 Sriraman Tallam <tmsriram@google.com>
2120 * config/i386/i386.c (ix86_option_override_internal):
2121 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
2122 (ix86_valid_target_attribute_tree):
2123 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
2124 Change TARGET_SSE to TARGET_SSE_P (opts->...)
2126 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2128 * tree-ssa-loop.h: Remove include files.
2129 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2130 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
2131 * graphite-clast-to-gimple.c: Likewise.
2132 * graphite-scop-detection.c: Likewise.
2133 * graphite-sese-to-poly.c: Likewise.
2134 * ipa-inline-analysis.c: Likewise.
2135 * ipa-pure-const.c: Likewise.
2136 * loop-init.c: Likewise.
2137 * passes.c: Likewise.
2138 * predict.c: Likewise.
2139 * tree-cfg.c: Likewise.
2140 * tree-cfgcleanup.c: Likewise.
2141 * tree-chrec.c: Likewise.
2142 * tree-data-ref.c: Likewise.
2143 * tree-loop-distribution.c: Likewise.
2144 * tree-parloops.c: Likewise.
2145 * tree-predcom.c: Likewise.
2146 * tree-scalar-evolution.c: Likewise.
2147 * tree-ssa-address.c: Likewise.
2148 * tree-ssa.c: Likewise.
2149 * tree-ssa-dce.c: Likewise.
2150 * tree-ssa-loop.c: Likewise.
2151 * tree-ssa-loop-im.c: Likewise.
2152 * tree-ssa-loop-ivcanon.c: Likewise.
2153 * tree-ssa-loop-ivopts.c: Likewise.
2154 * tree-ssa-loop-manip.c: Likewise.
2155 * tree-ssa-loop-niter.c: Likewise.
2156 * tree-ssa-loop-prefetch.c: Likewise.
2157 * tree-ssa-loop-unswitch.c: Likewise.
2158 * tree-ssa-reassoc.c: Likewise.
2159 * tree-vect-data-refs.c: Likewise.
2160 * tree-vect-loop.c: Likewise.
2161 * tree-vect-loop-manip.c: Likewise.
2162 * tree-vectorizer.c: Likewise.
2163 * tree-vect-stmts.c: Likewise.
2164 * tree-vrp.c: Likewise.
2166 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2168 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
2170 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2172 PR tree-optimization/58775
2173 PR tree-optimization/58791
2174 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
2175 (insert_stmt_after): Rewritten, don't move the stmt, but really
2177 (get_stmt_uid_with_default): Remove.
2178 (build_and_add_sum): Use insert_stmt_after and
2179 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
2180 (update_range_test): Set uid on stmts added by
2181 force_gimple_operand_gsi. Don't immediately modify statements
2182 in inter-bb optimization, just update oe->op values.
2183 (optimize_range_tests): Return bool whether any changed have been made.
2184 (update_ops): New function.
2185 (struct inter_bb_range_test_entry): New type.
2186 (maybe_optimize_range_tests): Perform statement changes here.
2187 (not_dominated_by, appears_later_in_bb, get_def_stmt,
2188 ensure_ops_are_available): Remove.
2189 (find_insert_point): Rewritten.
2190 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
2191 return LHS of the (new resp. old) stmt. Don't call
2192 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
2193 instead of last, move new stmt at the right place.
2194 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
2195 (negate_value): Likewise. Set uids.
2196 (break_up_subtract_bb): Initialize uids.
2197 (reassociate_bb): Adjust rewrite_expr_tree caller.
2198 (do_reassoc): Don't call renumber_gimple_stmt_uids.
2200 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
2203 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
2204 TARGET_32BIT final condition.
2205 (mulsi3_internal2 and splitter): Same.
2207 2013-10-23 Jeff Law <law@redhat.com>
2209 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
2210 through joiner blocks with abnormal outgoing edges.
2212 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
2213 Add parameter JOINERS, to allow/disallow threading through joiner
2215 (thread_block): New. Call thread_block_1.
2216 (mark_threaded_blocks): Remove code to filter out certain cases
2217 of threading through joiner blocks.
2218 (thread_through_all_blocks): Document how we can have a dangling
2219 edge AUX field and clear it.
2221 2013-10-23 Ian Lance Taylor <iant@google.com>
2223 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
2224 (C++ Dialect Options): Likewise.
2225 (Optimize Options): Likewise.
2227 2013-10-23 Tom de Vries <tom@codesourcery.com>
2229 PR tree-optimization/58805
2230 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
2232 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2234 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
2235 sequence based on get_range_info returned range.
2237 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2239 * tree-ssa.h: Remove all #include's
2240 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2241 * alias.c: Move required includes from tree-ssa.h.
2243 * builtins.c: Likewise.
2244 * calls.c: Likewise.
2245 * cfgexpand.c: Likewise.
2246 * cfghooks.c: Likewise.
2247 * cfgloop.c: Likewise.
2248 * cfgloopmanip.c: Likewise.
2249 * cgraph.c: Likewise.
2250 * cgraphbuild.c: Likewise.
2251 * cgraphclones.c: Likewise.
2252 * cgraphunit.c: Likewise.
2254 * except.c: Likewise.
2256 * final.c: Likewise.
2257 * fold-const.c: Likewise.
2258 * ggc-page.c: Likewise.
2259 * gimple-builder.c: Likewise.
2260 * gimple-fold.c: Likewise.
2261 * gimple-iterator.c: Likewise.
2262 * gimple-low.c: Likewise.
2263 * gimple-pretty-print.c: Likewise.
2264 * gimple-ssa-strength-reduction.c: Likewise.
2265 * gimple-streamer-in.c: Likewise.
2266 * gimple-streamer-out.c: Likewise.
2267 * gimplify.c: Likewise.
2268 * graphite-blocking.c: Likewise.
2269 * graphite-clast-to-gimple.c: Likewise.
2270 * graphite-dependences.c: Likewise.
2271 * graphite-interchange.c: Likewise.
2272 * graphite-optimize-isl.c: Likewise.
2273 * graphite-poly.c: Likewise.
2274 * graphite-scop-detection.c: Likewise.
2275 * graphite-sese-to-poly.c: Likewise.
2276 * graphite.c: Likewise.
2277 * ipa-cp.c: Likewise.
2278 * ipa-inline-analysis.c: Likewise.
2279 * ipa-inline-transform.c: Likewise.
2280 * ipa-inline.c: Likewise.
2281 * ipa-prop.c: Likewise.
2282 * ipa-pure-const.c: Likewise.
2283 * ipa-reference.c: Likewise.
2284 * ipa-split.c: Likewise.
2285 * ipa-utils.c: Likewise.
2286 * loop-init.c: Likewise.
2287 * lto-cgraph.c: Likewise.
2288 * lto-section-in.c: Likewise.
2289 * lto-section-out.c: Likewise.
2290 * lto-streamer-in.c: Likewise.
2291 * lto-streamer-out.c: Likewise.
2292 * lto-streamer.c: Likewise.
2293 * omp-low.c: Likewise.
2294 * passes.c: Likewise.
2295 * predict.c: Likewise.
2296 * print-tree.c: Likewise.
2297 * profile.c: Likewise.
2299 * targhooks.c: Likewise.
2300 * tracer.c: Likewise.
2301 * trans-mem.c: Likewise.
2302 * tree-call-cdce.c: Likewise.
2303 * tree-cfg.c: Likewise.
2304 * tree-cfgcleanup.c: Likewise.
2305 * tree-chrec.c: Likewise.
2306 * tree-complex.c: Likewise.
2307 * tree-data-ref.c: Likewise.
2308 * tree-dfa.c: Likewise.
2309 * tree-eh.c: Likewise.
2310 * tree-emutls.c: Likewise.
2311 * tree-if-conv.c: Likewise.
2312 * tree-inline.c: Likewise.
2313 * tree-into-ssa.c: Likewise.
2314 * tree-loop-distribution.c: Likewise.
2315 * tree-mudflap.c: Likewise.
2316 * tree-nested.c: Likewise.
2317 * tree-nrv.c: Likewise.
2318 * tree-object-size.c: Likewise.
2319 * tree-outof-ssa.c: Likewise.
2320 * tree-parloops.c: Likewise.
2321 * tree-phinodes.c: Likewise.
2322 * tree-predcom.c: Likewise.
2323 * tree-pretty-print.c: Likewise.
2324 * tree-profile.c: Likewise.
2325 * tree-scalar-evolution.c: Likewise.
2326 * tree-sra.c: Likewise.
2327 * tree-ssa-address.c: Likewise.
2328 * tree-ssa-alias.c: Likewise.
2329 * tree-ssa-ccp.c: Likewise.
2330 * tree-ssa-coalesce.c: Likewise.
2331 * tree-ssa-copy.c: Likewise.
2332 * tree-ssa-copyrename.c: Likewise.
2333 * tree-ssa-dce.c: Likewise.
2334 * tree-ssa-dom.c: Likewise.
2335 * tree-ssa-dse.c: Likewise.
2336 * tree-ssa-forwprop.c: Likewise.
2337 * tree-ssa-ifcombine.c: Likewise.
2338 * tree-ssa-live.c: Likewise.
2339 * tree-ssa-loop-ch.c: Likewise.
2340 * tree-ssa-loop-im.c: Likewise.
2341 * tree-ssa-loop-ivcanon.c: Likewise.
2342 * tree-ssa-loop-ivopts.c: Likewise.
2343 * tree-ssa-loop-manip.c: Likewise.
2344 * tree-ssa-loop-niter.c: Likewise.
2345 * tree-ssa-loop-prefetch.c: Likewise.
2346 * tree-ssa-loop-unswitch.c: Likewise.
2347 * tree-ssa-loop.c: Likewise.
2348 * tree-ssa-math-opts.c: Likewise.
2349 * tree-ssa-operands.c: Likewise.
2350 * tree-ssa-phiopt.c: Likewise.
2351 * tree-ssa-phiprop.c: Likewise.
2352 * tree-ssa-pre.c: Likewise.
2353 * tree-ssa-propagate.c: Likewise.
2354 * tree-ssa-reassoc.c: Likewise.
2355 * tree-ssa-sccvn.c: Likewise.
2356 * tree-ssa-sink.c: Likewise.
2357 * tree-ssa-strlen.c: Likewise.
2358 * tree-ssa-structalias.c: Likewise.
2359 * tree-ssa-tail-merge.c: Likewise.
2360 * tree-ssa-ter.c: Likewise.
2361 * tree-ssa-threadedge.c: Likewise.
2362 * tree-ssa-threadupdate.c: Likewise.
2363 * tree-ssa-uncprop.c: Likewise.
2364 * tree-ssa-uninit.c: Likewise.
2365 * tree-ssa.c: Likewise.
2366 * tree-ssanames.c: Likewise.
2367 * tree-stdarg.c: Likewise.
2368 * tree-streamer-in.c: Likewise.
2369 * tree-switch-conversion.c: Likewise.
2370 * tree-tailcall.c: Likewise.
2371 * tree-vect-data-refs.c: Likewise.
2372 * tree-vect-generic.c: Likewise.
2373 * tree-vect-loop-manip.c: Likewise.
2374 * tree-vect-loop.c: Likewise.
2375 * tree-vect-patterns.c: Likewise.
2376 * tree-vect-slp.c: Likewise.
2377 * tree-vect-stmts.c: Likewise.
2378 * tree-vectorizer.c: Likewise.
2379 * tree-vrp.c: Likewise.
2382 * value-prof.c: Likewise.
2383 * var-tracking.c: Likewise.
2384 * varpool.c: Likewise.
2385 * vtable-verify.c: Likewise.
2387 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2389 * config/tilegx/tilegx.c: Include "tree.h".
2391 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2393 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
2394 the info, not after it.
2395 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
2396 dump_ssaname_info call.
2397 (pp_gimple_stmt_1): Adjust caller.
2398 (dump_phi_nodes): Likewise. Don't print "# " here.
2400 2013-10-22 Jan Hubicka <jh@suse.cz>
2402 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
2404 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
2405 * i386.c (expand_small_movmem_or_setmem): New function.
2406 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
2408 (alg_usable_p): Add support for value ranges; cleanup.
2409 (ix86_expand_set_or_movmem): Add support for misaligned moves.
2411 2013-10-22 Sterling Augustine <saugustine@google.com>
2413 * doc/invoke.texi: Document -ggnu-pubnames.
2414 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
2416 * dwarf2out.c: Include gdb/gdb-index.h.
2417 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
2418 debug_generate_pub_sections.
2419 (is_java, output_pubtables, output_pubname): New functions.
2420 (include_pubname_in_output): Handle debug_generate_pub_sections at
2422 (size_of_pubnames): Use new local space_for_flags based on
2423 debug_generate_pub_sections.
2424 (output_pubnames): Unify pubnames and pubtypes output logic.
2425 Genericize comments. Call output_pubname.
2426 (dwarf2out_finish): Move logic to output_pubnames and call it.
2428 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
2431 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
2432 Remove CCCmode handling.
2433 <case LTU>: Return 'c' suffix for CCCmode.
2434 <case GEU>: Return 'nc' suffix for CCCmode.
2435 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
2436 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
2437 (*sub<mode>3_cc_overflow): Ditto.
2438 (*subsi3_zext_cc_overflow): Ditto.
2440 2013-10-22 Steve Ellcey <sellcey@mips.com>
2442 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
2443 (AND (NOT OP1) (NOT OP2)).
2445 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
2447 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2448 meaning of merge-high and merge-low masks for little endian; avoid
2449 use of vector-pack masks for little endian for mismatched modes.
2451 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2453 * config/tilepro/tilepro.c: Include "tree.h".
2455 2013-10-22 Andreas Schwab <schwab@suse.de>
2457 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
2460 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
2462 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
2463 [Ed Smith-Rowland]): New entries.
2464 ([Stephen M. Webb]): Update.
2466 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
2468 * tree-ssa-ter.h: Remove duplicate copy of file contents.
2470 2013-10-21 Marek Polacek <polacek@redhat.com>
2473 * fold-const.c (fold_range_test): Return 0 if the type is not
2476 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
2478 * system.h: Move hwint.h include further down.
2479 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
2480 gcc_checking_asserts.
2481 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
2483 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2485 Fix volatile issues in optimize_bit_field_compare.
2486 * fold-const.c (optimize_bit_field_compare): Bail out if
2487 lvolatilep or rvolatilep.
2489 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2491 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
2492 and get_inner_reference returning different pmode for non-volatile
2493 bit-field members dependent on flag_strict_volatile_bitfields.
2494 * stor-layout.c (layout_decl): Remove special handling of
2495 flag_strict_volatile_bitfields.
2496 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
2497 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
2499 2013-10-21 Paulo Matos <pmatos@broadcom.com>
2501 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
2504 2013-10-21 Jeff Law <law@redhat.com>
2506 * tree-ssa-threadedge.c (thread_through_normal_block): New
2507 argument VISITED. Remove VISISTED as a local variable. When we
2508 have a threadable jump, verify the destination of the jump has not
2510 (thread_across_edge): Allocate VISITED bitmap once at function
2511 scope and use it throughout. Make sure to set appropriate bits in
2512 VISITED for E (start of jump thread path).
2513 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
2514 through a joiner if any edge on the path has a recorded jump thread.
2516 2013-10-21 Ian Lance Taylor <iant@google.com>
2518 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
2519 don't imply that attributes can solve all problems.
2520 (Directory Options): Fix typo.
2522 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2524 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
2525 extend and extend_add.
2527 2013-10-21 Richard Biener <rguenther@suse.de>
2529 PR tree-optimization/58794
2530 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
2531 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
2533 2013-10-21 Richard Biener <rguenther@suse.de>
2536 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
2537 to (T) X for sign-changing conversions (or no conversion).
2539 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
2541 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
2543 2013-10-20 Jan Hubicka <jh@suse.cz>
2545 * config/i386/i386-tune.def: Add comment; organize into categories
2547 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2549 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
2550 argument. Update function comment.
2551 (expand_set_or_movmem_via_rep): New function combining
2552 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
2553 (expand_movmem_via_rep_mov): Remove.
2554 expand_setmem_via_rep_stos): Remove.
2555 (expand_movmem_epilogue): Update calls correspondingly.
2556 (expand_setmem_epilogue_via_loop): Likewise.
2558 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
2559 (expand_set_or_movmem_prologue): New function combining
2560 expand_movmem_prologue and expand_setmem_prologue.
2561 (expand_movmem_prologue): Remove.
2562 (expand_setmem_prologue): Remove.
2563 (expand_set_or_movmem_constant_prologue): New function combining
2564 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
2565 (expand_constant_movmem_prologue): Remove.
2566 (expand_constant_setmem_prologue): Remove.
2567 (promote_duplicated_reg): Allow vector-const0 value.
2568 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
2569 and ix86_expand_setmem.
2570 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
2571 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
2573 2013-10-21 Diego Novillo <dnovillo@google.com>
2575 * asan.c: Include tree.h
2576 * bb-reorder.c: Likewise.
2577 * cfgcleanup.c: Likewise.
2578 * cfgloopmanip.c: Likewise.
2579 * data-streamer-in.c: Likewise.
2580 * data-streamer-out.c: Likewise.
2581 * data-streamer.c: Likewise.
2582 * dwarf2cfi.c: Likewise.
2583 * graphite-blocking.c: Likewise.
2584 * graphite-clast-to-gimple.c: Likewise.
2585 * graphite-dependences.c: Likewise.
2586 * graphite-interchange.c: Likewise.
2587 * graphite-optimize-isl.c: Likewise.
2588 * graphite-poly.c: Likewise.
2589 * graphite-scop-detection.c: Likewise.
2590 * graphite-sese-to-poly.c: Likewise.
2591 * graphite.c: Likewise.
2592 * ipa-devirt.c: Likewise.
2593 * ipa-profile.c: Likewise.
2596 * loop-init.c: Likewise.
2597 * loop-unroll.c: Likewise.
2598 * lower-subreg.c: Likewise.
2599 * lto/lto-object.c: Likewise.
2600 * recog.c: Likewise.
2601 * reginfo.c: Likewise.
2602 * tree-loop-distribution.c: Likewise.
2603 * tree-parloops.c: Likewise.
2604 * tree-ssa-strlen.c: Likewise.
2605 * tree-streamer.c: Likewise.
2606 * value-prof.c: Likewise.
2607 * target-globals.c: Likewise.
2608 * expr.h: Include tree-core.h instead of tree.h.
2609 * gimple.h: Likewise.
2610 * ipa-prop.h: Likewise.
2611 * ipa-utils.h: Likewise.
2612 * lto-streamer.h: Likewise.
2613 * streamer-hooks.h: Likewise.
2614 * ipa-reference.h: Include cgraph.h instead of tree.h.
2615 * cgraph.h: Include basic-block.h instead of tree.h.
2616 * tree-streamer.h: Do not include tree.h.
2617 * genattrtab.c (write_header): Generate inclusion of
2619 * genautomata.c (main): Likewise.
2620 * genemit.c: Likewise.
2621 * genopinit.c: Likewise.
2622 * genoutput.c (output_prologue): Likewise.
2623 * genpeep.c: Likewise.
2625 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2627 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
2629 (vec_unpacku_hi_v8hi): Likewise.
2630 (vec_unpacku_lo_v16qi): Likewise.
2631 (vec_unpacku_lo_v8hi): Likewise.
2633 2013-10-20 Jan Hubicka <jh@suse.cz>
2635 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
2636 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
2637 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
2639 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
2641 * config/mips/mips.h (ISA_HAS_WSBH): Define.
2642 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
2644 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
2646 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
2649 * system.h: Undef m_slot.
2651 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2653 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
2654 all elements for both endian flavors.
2656 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2659 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
2660 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
2661 and SI_REG for 64bit SYSV ABI targets.
2663 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2665 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
2666 to multi_reg_return. Clarify that we are skipping USEs of multiple
2667 return registers. Use bool type where appropriate.
2669 2013-10-18 Jan Hubicka <jh@suse.cz>
2671 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
2673 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
2674 (X86_TUNE_PUSH_MEMORY): Likewise.
2675 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
2676 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
2677 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
2679 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
2680 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
2682 (ix86_option_override_internal): Update to use tune features instead
2685 2013-10-18 Cong Hou <congh@google.com>
2687 PR tree-optimization/58508
2688 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
2689 statement that contains data refs with zero-step.
2691 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2693 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
2694 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
2695 * gimple-low.c (gimple_check_call_arg,
2696 gimple_check_call_matching_types): Move to cgraph.c.
2697 * gimple-low.h: Remove prototype.
2698 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
2699 Relocate from gimple-low.c.
2700 * cgraph.h: Add prototype. Don't include basic-block.h.
2701 * gimplify.c: Add gimple-low to include list.
2702 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
2703 * tree-eh.c: Add gimple-low to include list.
2704 * tree-nested.c: Likewise.
2705 * cfgexpand.c: Add tree-ssa-address.h to include list.
2707 * gimple-fold.c: Likewise.
2708 * gimple-ssa-strength-reduction.c: Likewise.
2709 * trans-mem.c: Likewise.
2710 * tree-mudflap.c: Likewise.
2711 * tree-ssa-loop-ivopts.c: Likewise.
2712 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
2713 (degenerate_phi_result): Move to tree-phinodes.c.
2714 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
2715 * tree-ssa-threadedge.c: Likewise.
2716 * tree-vrp.c: Likewise.
2717 * tree-phinodes.c (degenerate_phi_result): Relocate here.
2718 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
2719 * tree-phinodes.h (degenerate_phi_result): Add prototype.
2720 * tree-ssa-copy.c: Include tree-ssa-dom.h.
2721 * tree-ssa-forwprop.c: Likewise.
2722 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
2723 pass_data_cleanup_cfg_post_optimizing,
2724 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
2725 * tree-optimize.c: Delete File.
2726 * graphite.c: Include tree-cfgcleanup.h.
2727 * passes.c: Likewise.
2728 * tree-cfg.c: Likewise.
2729 * tree-profile.c: Likewise.
2730 * tree-ssa-dse.c: Likewise.
2731 * tree-ssa-loop-ivcanon.c: Likewise.
2732 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
2733 * tree-outof-ssa.c: Include sbitmap.h.
2734 * tree-ssa-live.c: Likewise.
2735 * tree-ssa-propagate.c: Likewise.
2736 * tree-ssa-structalias.c: Likewise.
2737 * tree-stdarg.c: Likewise.
2738 * Makefile.in (OBJS): Delete tree-optimize.o.
2739 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
2740 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
2741 * varasm.c: Include basic-block.h.
2742 * cfgloop.h: Include function.h instead of basic-block.h
2743 (bb_loop_depth): Move to cfgloop.c.
2744 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
2746 2013-10-18 Teresa Johnson <tejohnson@google.com>
2748 * predict.c (probably_never_executed): Compare frequency-based
2749 count to number of training runs.
2750 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
2752 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2754 * config/arm/arm.c (cortexa9_extra_costs): New table.
2755 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
2757 2013-10-18 Jeff Law <law@redhat.com>
2759 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
2761 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
2762 (register_jump_thread): Add "registered_jump_thread" debug
2764 * dbgcnt.def (registered_jump_thread): New debug counter.
2766 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2768 * config/rs6000/rs6000.c: Include cgraph.h.
2770 2013-10-18 Teresa Johnson <tejohnson@google.com>
2772 * tree-ssa-tail-merge.c (replace_block_by): Update edge
2773 weights during merging.
2775 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2777 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
2778 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
2779 * tree-ssa-operands.h: Remove prototype.
2780 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
2781 * gimple.c (virtual_operand_p): Relocate from gimple.c.
2782 * gimple.h: Add prototype.
2783 * gimple-ssa.h: Include tree-ssa-operands.h.
2784 * tree-dump.c: Add tree-cfg.h to include list.
2785 * tree-ssa-alias.c: Add ipa-reference.h to include list.
2786 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
2787 * config/i386/i386.c: Don't include tree-flow.h.
2788 * config/rs6000/rs6000.c: Likewise.
2790 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2792 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
2794 2013-10-18 Richard Biener <rguenther@suse.de>
2796 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
2797 or TYPE_UNSIGNED of integral types.
2798 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
2799 NULL_TREE for zero-precision integral types.
2801 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
2803 * config/aarch64/arm_neon.h
2804 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
2806 2013-10-17 Sriraman Tallam <tmsriram@google.com>
2809 * opth-gen.awk: Define target_flags_explicit.
2811 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2813 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
2814 fields to the reg_addr array that describes the valid addressing
2815 mode for any register, general purpose registers, floating point
2816 registers, and Altivec registers.
2817 (FIRST_RELOAD_REG_CLASS): Likewise.
2818 (LAST_RELOAD_REG_CLASS): Likewise.
2819 (struct reload_reg_map_type): Likewise.
2820 (reload_reg_map_type): Likewise.
2821 (RELOAD_REG_VALID): Likewise.
2822 (RELOAD_REG_MULTIPLE): Likewise.
2823 (RELOAD_REG_INDEXED): Likewise.
2824 (RELOAD_REG_OFFSET): Likewise.
2825 (RELOAD_REG_PRE_INCDEC): Likewise.
2826 (RELOAD_REG_PRE_MODIFY): Likewise.
2827 (reg_addr): Likewise.
2828 (mode_supports_pre_incdec_p): New helper functions to say whether
2829 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
2830 (mode_supports_pre_modify_p): Likewise.
2831 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
2832 print the valid address mode bits for each mode.
2833 (rs6000_debug_print_mode): Likewise.
2834 (rs6000_debug_reg_global): Likewise.
2835 (rs6000_setup_reg_addr_masks): New function to set up the address
2836 mask bits for each type.
2837 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
2838 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
2839 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
2840 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
2841 PRE_MODIFY are supported.
2842 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
2843 registers, instead of {src,dest}_av_p.
2844 (rs6000_print_options_internal): Tweak the debug output slightly.
2846 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
2848 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
2851 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
2853 * tree-flow.h (struct omp_region): Move to omp-low.c.
2854 Remove omp_ prototypes and variables.
2855 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
2856 (copy_var_decl): Relocate prototype from tree-flow.h.
2857 * gimple.c (copy_var_decl): Relocate from omp-low.c.
2858 * tree.h: Move prototype to omp-low.h.
2859 * omp-low.h: New File. Relocate prototypes here.
2860 * omp-low.c (struct omp_region): Make local here.
2861 (root_omp_region): Make static.
2862 (copy_var_decl) Move to gimple.c.
2863 (new_omp_region): Make static.
2864 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
2865 * tree-cfg.c: Include omp-low.h.
2866 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
2867 * gimplify.c: Include omp-low.h.
2868 * tree-parloops.c: Likewise.
2870 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
2872 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
2873 operands are in memory, prefer to force non-matched operand 1 to
2876 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2879 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
2880 restrict TImode addresses to single indirect registers if both
2881 -mquad-memory and -mvsx-timode are used.
2882 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
2883 we should emit load/store quad. Remove using %y for quad memory
2886 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
2887 constraints to allow load/store quad on machines where TImode is
2888 not allowed in VSX registers. Use 'n' instead of 'F' constraint
2889 for TImode to load integer constants.
2891 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2893 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
2895 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
2897 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
2898 handling of STACK_REG.
2900 2013-10-17 Richard Biener <rguenther@suse.de>
2902 PR tree-optimization/58143
2903 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
2905 (rewrite_to_defined_overflow): Likewise.
2906 (move_computations_dom_walker::before_dom): Rewrite stmts
2907 with undefined signed overflow that are not always executed
2908 into unsigned arithmetic.
2910 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
2913 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
2914 explicit isa flag to be an options variable, instead of using
2915 global_options_set. Remove define from rs6000.h.
2916 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
2918 * config/rs6000/rs6000.c (rs6000_option_override_internal):
2919 Initialize rs6000_isa_flags_explicit.
2920 (rs6000_function_specific_save): Add gcc_options* parameter, so
2921 that the powerpc builds after the 2013-10-15 changes.
2922 (rs6000_function_specific_restore): Likewise.
2924 2013-10-16 DJ Delorie <dj@redhat.com>
2926 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
2927 op is a REG before checking REGNO.
2928 (rl78_alloc_physical_registers): Verify pattern is a SET before
2931 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2933 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
2935 (vec_unpacks_lo_v4sf): Likewise.
2936 (vec_unpacks_float_hi_v4si): Likewise.
2937 (vec_unpacks_float_lo_v4si): Likewise.
2938 (vec_unpacku_float_hi_v4si): Likewise.
2939 (vec_unpacku_float_lo_v4si): Likewise.
2941 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2943 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
2944 (vsx_concat_v2sf): Likewise.
2946 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
2948 * config/aarch64/aarch64.md
2949 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
2951 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
2953 PR tree-optimization/58697
2954 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
2955 estimated_loop_iterations_int.
2956 (max_stmt_executions_int): Call get_max_loop_iterations_int.
2957 (get_max_loop_iterations_int): New. HWINT version of
2958 get_max_loop_iterations.
2959 * cfgloop.h: Add prototypes.
2960 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
2961 * loop-unroll.c (decide_peel_once_rolling): Call
2962 get_estimated_loop_iterations_int.
2963 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
2964 * tree-ssa-loop-niter.h: Tweak prototypes.
2966 2013-10-16 David Malcolm <dmalcolm@redhat.com>
2968 * gengtype-parse.c (struct_field_seq): Ignore access-control
2969 keywords ("public:" etc).
2971 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
2973 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
2974 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
2976 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
2978 * config/arm/arm.opt (mlra): New option.
2979 * config/arm/arm.c (arm_lra_p): New function.
2980 (TARGET_LRA_P): Define.
2982 2013-10-16 Paulo Matos <pmatos@broadcom.com>
2984 * tree-core.h (tree_code_name): Remove.
2985 * tree.h (get_tree_code_name): New prototype.
2986 * tree.c (tree_code_name): Make static.
2987 (get_tree_code_name): New function.
2988 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
2989 tree_class_check_failed, tree_range_check_failed,
2990 tree_not_class_check_failed, omp_clause_check_failed,
2991 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
2992 wrapper get_tree_code_name instead of calling tree_code_name directly.
2993 * tree-vrp.c (dump_asserts_for): Likewise.
2994 * tree-dump.c (dequeue_and_dump): Likewise.
2995 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
2996 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
2997 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
2998 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
2999 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
3000 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
3001 dump_gimple_omp_for): Likewise.
3002 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
3003 * tree-ssa-pre.c (print_pre_expr): Likewise.
3004 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
3005 * print-tree.c (print_node_brief, print_node): Likewise.
3006 * gimple.c (gimple_check_failed): Likewise.
3007 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
3008 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
3009 * config/mep/mep.c (mep_validate_vliw): Likewise.
3010 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
3011 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
3013 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
3015 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
3018 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
3020 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
3021 (MULTILIB_MATCHES): Add multilib for -march=v8.
3023 2013-10-15 Sriraman Tallam <tmsriram@google.com>
3026 * optc-save-gen.awk: Add extra parameter to the save and restore
3028 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
3029 * tree.c (build_optimization_node): New parameter. Add extra parameter
3030 to call to cl_optimization_save.
3031 (build_target_option_node): New parameter. Add extra parameter
3032 to call to cl_target_option_save.
3033 * tree.h (build_optimization_node): New parameter.
3034 (build_target_option_node): New parameter.
3035 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
3036 build_optimization_node and build_target_option_node.
3037 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
3038 (handle_pragma_push_options): Ditto.
3039 * toplev.c (process_options): Ditto.
3040 * opts.c (init_options_struct): Check for opts_set non-null.
3041 * target.def (target_option.save): New parameter.
3042 (target_option.restore): New parameter.
3043 * tm.texi: Generate.
3044 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
3045 (ix86_pragma_target_parse): Ditto.
3046 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
3048 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
3049 to build_optimization_node and build_target_option_node.
3050 (rs6000_valid_attribute_p): Ditto.
3051 (rs6000_pragma_target_parse): Ditto.
3052 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
3054 * config/i386/i386.h:
3055 TARGET_64BIT_P: New Macro
3056 TARGET_MMX_P: New Macro.
3057 TARGET_3DNOW_P: New Macro.
3058 TARGET_3DNOW_A_P: New Macro.
3059 TARGET_SSE_P: New Macro.
3060 TARGET_SSE2_P: New Macro.
3061 TARGET_SSE3_P: New Macro.
3062 TARGET_SSSE3_P: New Macro.
3063 TARGET_SSE4_1_P: New Macro.
3064 TARGET_SSE4_2_P: New Macro.
3065 TARGET_AVX_P: New Macro.
3066 TARGET_AVX2_P: New Macro.
3067 TARGET_AVX512F_P: New Macro.
3068 TARGET_AVX512PF_P: New Macro.
3069 TARGET_AVX512ER_P: New Macro.
3070 TARGET_AVX512CD_P: New Macro.
3071 TARGET_FMA_P: New Macro.
3072 TARGET_SSE4A_P: New Macro.
3073 TARGET_FMA4_P: New Macro.
3074 TARGET_XOP_P: New Macro.
3075 TARGET_LWP_P: New Macro.
3076 TARGET_ABM_P: New Macro.
3077 TARGET_BMI_P: New Macro.
3078 TARGET_BMI2_P: New Macro.
3079 TARGET_LZCNT_P: New Macro.
3080 TARGET_TBM_P: New Macro.
3081 TARGET_POPCNT_P: New Macro.
3082 TARGET_SAHF_P: New Macro.
3083 TARGET_MOVBE_P: New Macro.
3084 TARGET_CRC32_P: New Macro.
3085 TARGET_AES_P: New Macro.
3086 TARGET_PCLMUL_P: New Macro.
3087 TARGET_CMPXCHG16B_P: New Macro.
3088 TARGET_FSGSBASE_P: New Macro.
3089 TARGET_RDRND_P: New Macro.
3090 TARGET_F16C_P: New Macro.
3091 TARGET_RTM_P: New Macro.
3092 TARGET_HLE_P: New Macro.
3093 TARGET_RDSEED_P: New Macro.
3094 TARGET_PRFCHW_P: New Macro.
3095 TARGET_ADX_P: New Macro.
3096 TARGET_FXSR_P: New Macro.
3097 TARGET_XSAVE_P: New Macro.
3098 TARGET_XSAVEOPT_P: New Macro.
3099 TARGET_LP64_P: New Macro.
3100 TARGET_X32_P: New Macro.
3101 TARGET_FPMATH_DEFAULT_P: New Macro.
3102 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
3103 * config/i386/i386.c (ix86_option_override_internal): New parameters.
3105 Change ix86_tune_string to access opts->x_ix86_tune_string.
3106 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
3107 Change ix86_arch_string to access opts->x_ix86_arch_string.
3108 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
3109 Change ix86_pmode to access opts->x_ix86_pmode.
3110 Change ix86_abi to access opts->x_ix86_abi.
3111 Change ix86_cmodel to access opts->x_ix86_cmodel.
3112 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
3113 Change ix86_isa_flags_explicit to access
3114 opts->x_ix86_isa_flags_explicit.
3115 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
3116 Change ix86_regparm to access opts->x_ix86_regparm.
3117 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
3118 Change ix86_preferred_stack_boundary_arg to access
3119 opts->x_ix86_preferred_stack_boundary_arg.
3120 Change ix86_force_align_arg_pointer to access
3121 opts->x_ix86_force_align_arg_pointer.
3122 Change ix86_incoming_stack_boundar_arg to access
3123 opts->x_ix86_incoming_stack_boundar_arg.
3124 Change ix86_fpmath to access opts->x_ix86_fpmath.
3125 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
3126 Change ix86_recip_name to access opts->x_ix86_recip_name.
3127 Change ix86_stack_protector_guard to access
3128 opts->x_ix86_stack_protector_guard.
3129 Change ix86_tune_memcpy_strategy to access
3130 opts->x_ix86_tune_memcpy_strategy.
3131 Change ix86_tune_memset_strategy to access
3132 opts->x_ix86_tune_memset_strategy.
3133 Change global_options to access opts.
3134 Change global_options_set to access opts_set.
3135 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
3136 Change TARGET_MMX to TARGET_MMX_P (opts->...).
3137 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
3138 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
3139 Change TARGET_SSE to TARGET_SSE_P (opts->...).
3140 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
3141 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
3142 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
3143 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
3144 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
3145 Change TARGET_AVX to TARGET_AVX_P (opts->...).
3146 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
3147 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
3148 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
3149 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
3150 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
3151 Change TARGET_FMA to TARGET_FMA_P (opts->...).
3152 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
3153 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
3154 Change TARGET_XOP to TARGET_XOP_P (opts->...).
3155 Change TARGET_LWP to TARGET_LWP_P (opts->...).
3156 Change TARGET_ABM to TARGET_ABM_P (opts->...).
3157 Change TARGET_BMI to TARGET_BMI_P (opts->...).
3158 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
3159 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
3160 Change TARGET_TBM to TARGET_TBM_P (opts->...).
3161 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
3162 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
3163 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
3164 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
3165 Change TARGET_AES to TARGET_AES_P (opts->...).
3166 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
3167 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
3168 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
3169 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
3170 Change TARGET_F16C to TARGET_F16C_P (opts->...).
3171 Change TARGET_RTM to TARGET_RTM_P (opts->...).
3172 Change TARGET_HLE to TARGET_HLE_P (opts->...).
3173 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
3174 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
3175 Change TARGET_ADX to TARGET_ADX_P (opts->...).
3176 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
3177 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
3178 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
3179 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
3180 Change TARGET_X32 to TARGET_X32_P (opts->...).
3181 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
3182 Change TARGET_FLOAT_RETURNS_IN_80387 to
3183 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
3184 (ix86_function_specific_save): New parameter. Use opts-> fields
3185 to replace global fields.
3186 (ix86_function_specific_restore): Ditto.
3187 (ix86_valid_target_attribute_inner_p): New parameters.
3189 Fix call to ix86_handle_option and set_option.
3190 (ix86_valid_target_attribute_tree): New parameters.
3191 Change global_options to access opts.
3192 Change global_options_set to access opts_set.
3193 Fix call to ix86_valid_target_attribute_inner_p.
3194 Change ix86_tune_string to access opts->x_ix86_tune_string.
3195 Change ix86_arch_string to access opts->x_ix86_arch_string.
3196 Change ix86_fpmath to access opts->x_ix86_fpmath
3197 Fix call to ix86_option_override_internal.
3198 Fix call to ix86_add_new_builtins.
3199 Fix calls to build_optimization_node and build_target_option_node.
3200 (ix86_valid_target_attribute_p): Remove access to global_options.
3201 Use new gcc_options structure func_options.
3202 Fix call to ix86_valid_target_attribute_tree.
3203 Fix call to build_optimization_node.
3204 (get_builtin_code_for_version): Fix call to
3205 ix86_valid_target_attribute_tree.
3207 2013-10-15 David Malcolm <dmalcolm@redhat.com>
3209 * Makefile.in (PICFLAG): New.
3210 (enable_host_shared): New.
3211 (INTERNAL_CFLAGS): Use PICFLAG.
3212 (LIBIBERTY): Use pic build of libiberty.a if configured with
3213 --enable-host-shared.
3214 * configure.ac: Add --enable-host-shared, setting up new
3216 * configure: Regenerate.
3217 * doc/install.texi (--enable-shared): Add note contrasting it with ...
3218 (--enable-host-shared): New option.
3220 2013-10-15 Richard Biener <rguenther@suse.de>
3222 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
3223 for built-in functions.
3225 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
3227 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
3228 (optimize_range_tests_1): New function,
3229 extracted from optimize_range_tests.
3230 (optimize_range_tests_xor): Similarly.
3231 (optimize_range_tests_diff): New function.
3232 (optimize_range_tests): Use optimize_range_tests_1.
3234 2013-10-15 Cong Hou <congh@google.com>
3236 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
3237 requirement of the reduction pattern so that one operand of the
3238 reduction operation can come from outside of the loop.
3240 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3242 * config/arm/neon-schedgen.ml: Remove.
3243 * config/arm/cortex-a9-neon.md: Remove comment regarding
3246 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3248 * config/arm/types: Remove old neon types.
3250 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3252 * config/arm/cortex-a7.md
3253 (cortex_a7_neon_type): New.
3254 (cortex_a7_neon_mul): Update for new types.
3255 (cortex_a7_neon_mla): Likewise.
3256 (cortex_a7_neon): Likewise.
3258 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3260 * config/arm/cortex-a15-neon.md
3261 (cortex_a15_neon_type): New,
3263 (cortex_a15_neon_int_1): Remove.
3264 (cortex_a15_neon_int_2): Likewise.
3265 (cortex_a15_neon_int_3): Likewise.
3266 (cortex_a15_neon_int_4): Likewise.
3267 (cortex_a15_neon_int_5): Likewise.
3268 (cortex_a15_neon_vqneg_vqabs): Likewise.
3269 (cortex_a15_neon_vmov): Likewise.
3270 (cortex_a15_neon_vaba): Likewise.
3271 (cortex_a15_neon_vaba_qqq): Likewise.
3272 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3273 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3274 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3276 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3277 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3278 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
3279 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3280 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3281 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3282 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3283 (cortex_a15_neon_shift_1): Likewise.
3284 (cortex_a15_neon_shift_2): Likewise.
3285 (cortex_a15_neon_shift_3): Likewise.
3286 (cortex_a15_neon_vshl_ddd): Likewise.
3287 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3288 (cortex_a15_neon_vsra_vrsra): Likewise.
3289 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3290 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3291 (cortex_a15_neon_bp_3cycle): Likewise.
3292 (cortex_a15_neon_ldm_2): Likewise.
3293 (cortex_a15_neon_stm_2): Likewise.
3294 (cortex_a15_neon_mcr): Likewise.
3295 (cortex_a15_neon_mrc): Likewise.
3296 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3297 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3298 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3299 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3300 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3301 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3302 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3303 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3304 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3305 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3306 (cortex_a15_neon_bp_simple): Likewise.
3307 (cortex_a15_neon_bp_2cycle): Likewise.
3308 (cortex_a15_neon_bp_3cycle): Likewise.
3309 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3310 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3311 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3312 (cortex_a15_neon_vld2_4_regs): Likewise.
3313 (cortex_a15_neon_vld3_vld4): Likewise.
3314 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3315 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3316 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
3317 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
3318 (cortex_a15_neon_vst3_vst4): Rename to...
3319 (cortex_a15_neon_vst4): This, update for new attributes.
3320 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
3321 (cortex_a15_neon_vld3_vld4_lane): Likewise.
3322 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3323 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3324 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3325 (cortex_a15_neon_ldm_2): Likewise.
3326 (cortex_a15_neon_stm_2): Likewise.
3327 (cortex_a15_neon_mcr): Likewise.
3328 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3329 (cortex_a15_neon_mrc): Likewise.
3330 (cortex_a15_neon_mrrc): Likewise.
3332 (cortex_a15_neon_abd): New.
3333 (cortex_a15_neon_abd_q): Likewise.
3334 (cortex_a15_neon_aba): Likewise.
3335 (cortex_a15_neon_aba_q): Likewise.
3336 (cortex_a15_neon_acc): Likewise.
3337 (cortex_a15_neon_acc_q): Likewise.
3338 (cortex_a15_neon_arith_basic): Likewise.
3339 (cortex_a15_neon_arith_complex): Likewise.
3340 (cortex_a15_neon_multiply): Likewise.
3341 (cortex_a15_neon_multiply_q): Likewise.
3342 (cortex_a15_neon_mla): Likewise.
3343 (cortex_a15_neon_mla_q): Likewise.
3344 (cortex_a15_neon_sat_mla_long): Likewise.
3345 (cortex_a15_neon_shift_acc): Likewise.
3346 (cortex_a15_neon_shift_imm_basic): Likewise.
3347 (cortex_a15_neon_shift_imm_complex): Likewise.
3348 (cortex_a15_neon_shift_reg_basic): Likewise.
3349 (cortex_a15_neon_shift_reg_basic_q): Likewise.
3350 (cortex_a15_neon_shift_reg_complex): Likewise.
3351 (cortex_a15_neon_shift_reg_complex_q): Likewise.
3352 (cortex_a15_neon_fp_negabs): Likewise
3353 (cortex_a15_neon_fp_arith): Likewise
3354 (cortex_a15_neon_fp_arith_q): Likewise
3355 (cortex_a15_neon_fp_cvt_int): Likewise
3356 (cortex_a15_neon_fp_cvt_int_q): Likewise
3357 (cortex_a15_neon_fp_cvt_16): Likewise
3358 (cortex_a15_neon_fp_mul): Likewise
3359 (cortex_a15_neon_fp_mul_q): Likewise
3360 (cortex_a15_neon_fp_mla): Likewise
3361 (cortex_a15_neon_fp_mla_q): Likewise
3362 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
3363 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
3364 (cortex_a15_neon_bitops): Likewise.
3365 (cortex_a15_neon_bitops_q): Likewise.
3366 (cortex_a15_neon_from_gp): Likewise.
3367 (cortex_a15_neon_from_gp_q): Likewise.
3368 (cortex_a15_neon_tbl3_tbl4): Likewise.
3369 (cortex_a15_neon_zip_q): Likewise.
3370 (cortex_a15_neon_to_gp): Likewise.
3371 (cortex_a15_neon_load_a): Likewise.
3372 (cortex_a15_neon_load_b): Likewise.
3373 (cortex_a15_neon_load_c): Likewise.
3374 (cortex_a15_neon_load_d): Likewise.
3375 (cortex_a15_neon_load_e): Likewise.
3376 (cortex_a15_neon_load_f): Likewise.
3377 (cortex_a15_neon_store_a): Likewise.
3378 (cortex_a15_neon_store_b): Likewise.
3379 (cortex_a15_neon_store_c): Likewise.
3380 (cortex_a15_neon_store_d): Likewise.
3381 (cortex_a15_neon_store_e): Likewise.
3382 (cortex_a15_neon_store_f): Likewise.
3383 (cortex_a15_neon_store_g): Likewise.
3384 (cortex_a15_neon_store_h): Likewise.
3385 (cortex_a15_vfp_to_from_gp): Likewise.
3387 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3389 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
3391 (cortex_a9_neon_vshl_ddd): Remove.
3392 (cortex_a9_neon_vst3_vst4): Likewise.
3393 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3395 (cortex_a9_neon_bit_ops_q): New.
3397 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
3398 (cortex_a9_neon_int_2): Likewise.
3399 (cortex_a9_neon_int_3): Likewise.
3400 (cortex_a9_neon_int_4): Likewise.
3401 (cortex_a9_neon_int_5): Likewise.
3402 (cortex_a9_neon_vqneg_vqabs): Likewise.
3403 (cortex_a9_neon_vmov): Likewise.
3404 (cortex_a9_neon_vaba): Likewise.
3405 (cortex_a9_neon_vaba_qqq): Likewise.
3406 (cortex_a9_neon_shift_1): Likewise.
3407 (cortex_a9_neon_shift_2): Likewise.
3408 (cortex_a9_neon_shift_3): Likewise.
3409 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3410 (cortex_a9_neon_vsra_vrsra): Likewise.
3411 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3412 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3413 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3415 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3416 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3417 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3419 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3420 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3421 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3422 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3423 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3424 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3425 (cortex_a9_neon_fp_vsum): Likewise.
3426 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3427 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3428 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3429 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3430 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3431 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3432 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3433 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3434 (cortex_a9_neon_bp_simple): Likewise.
3435 (cortex_a9_neon_bp_2cycle): Likewise.
3436 (cortex_a9_neon_bp_3cycle): Likewise.
3437 (cortex_a9_neon_ldr): Likewise.
3438 (cortex_a9_neon_str): Likewise.
3439 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3440 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3441 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3442 (cortex_a9_neon_vld2_4_regs): Likewise.
3443 (cortex_a9_neon_vld3_vld4): Likewise.
3444 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3445 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3446 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3447 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3448 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3449 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3450 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3451 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3452 (cortex_a9_neon_mcr): Likewise.
3453 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3454 (cortex_a9_neon_mrc): Likewise.
3455 (cortex_a9_neon_mrrc): Likewise.
3457 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3459 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
3461 (cortex_a8_neon_vshl_ddd): Remove.
3462 (cortex_a8_neon_vst3_vst4): Likewise.
3463 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3465 (cortex_a8_neon_bit_ops_q): New.
3467 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
3468 (cortex_a8_neon_int_2): Likewise..
3469 (cortex_a8_neon_int_3): Likewise.
3470 (cortex_a8_neon_int_5): Likewise.
3471 (cortex_a8_neon_vqneg_vqabs): Likewise.
3472 (cortex_a8_neon_int_4): Likewise.
3473 (cortex_a8_neon_vaba): Likewise.
3474 (cortex_a8_neon_vaba_qqq): Likewise.
3475 (cortex_a8_neon_shift_1): Likewise.
3476 (cortex_a8_neon_shift_2): Likewise.
3477 (cortex_a8_neon_shift_3): Likewise.
3478 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3479 (cortex_a8_neon_vsra_vrsra): Likewise.
3480 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3481 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3482 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3484 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3485 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3486 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3488 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3489 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3490 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3491 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3492 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3493 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3494 (cortex_a8_neon_fp_vsum): Likewise.
3495 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3496 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3497 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3498 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3499 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3500 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3501 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3502 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3503 (cortex_a8_neon_bp_simple): Likewise.
3504 (cortex_a8_neon_bp_2cycle): Likewise.
3505 (cortex_a8_neon_bp_3cycle): Likewise.
3506 (cortex_a8_neon_ldr): Likewise.
3507 (cortex_a8_neon_str): Likewise.
3508 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3509 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3510 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3511 (cortex_a8_neon_vld2_4_regs): Likewise.
3512 (cortex_a8_neon_vld3_vld4): Likewise.
3513 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3514 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3515 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3516 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3517 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3518 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3519 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3520 (cortex_a8_neon_mcr): Likewise.
3521 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3522 (cortex_a8_neon_mrc): Likewise.
3523 (cortex_a8_neon_mrrc): Likewise.
3525 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3527 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
3529 * config/aarch64/aarch64-simd.md (neon_type): Remove.
3530 (aarch64_simd_dup<mode>): Add "type" attribute.
3531 (aarch64_dup_lane<mode>): Likewise.
3532 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
3533 (*aarch64_simd_mov<mode>): Likewise.
3534 (aarch64_simd_mov_from_<mode>low): Likewise.
3535 (aarch64_simd_mov_from_<mode>high): Likewise.
3536 (orn<mode>3): Likewise.
3537 (bic<mode>3): Likewise.
3538 (add<mode>3): Likewise.
3539 (sub<mode>3): Likewise.
3540 (mul<mode>3): Likewise.
3541 (*aarch64_mul3_elt<mode>): Likewise.
3542 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
3543 (*aarch64_mul3_elt_to_128df): Likewise.
3544 (*aarch64_mul3_elt_to_64v2df): Likewise.
3545 (neg<mode>2): Likewise.
3546 (abs<mode>2): Likewise.
3547 (abd<mode>_3): Likewise.
3548 (aba<mode>_3): Likewise.
3549 (fabd<mode>_3): Likewise.
3550 (*fabd_scalar<mode>3): Likewise.
3551 (and<mode>3): Likewise.
3552 (ior<mode>3): Likewise.
3553 (xor<mode>3): Likewise.
3554 (one_cmpl<mode>2): Likewise.
3555 (aarch64_simd_vec_set<mode>): Likewise.
3556 (aarch64_simd_lshr<mode>): Likewise.
3557 (aarch64_simd_ashr<mode>): Likewise.
3558 (aarch64_simd_imm_shl<mode>): Likewise.
3559 (aarch64_simd_reg_sshl<mode): Likewise.
3560 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
3561 (aarch64_simd_reg_shl<mode>_signed): Likewise.
3562 (aarch64_simd_vec_setv2di): Likewise.
3563 (aarch64_simd_vec_set<mode>): Likewise.
3564 (aarch64_mla<mode>): Likewise.
3565 (*aarch64_mla_elt<mode>): Likewise.
3566 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
3567 (aarch64_mls<mode>): Likewise.
3568 (*aarch64_mls_elt<mode>): Likewise.
3569 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
3570 (<su><maxmin><mode>3): Likewise.
3571 (move_lo_quad_<mode>): Likewise.
3572 (aarch64_simd_move_hi_quad_<mode>): Likewise.
3573 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
3574 (vec_pack_trunc_<mode>): Likewise.
3575 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
3576 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
3577 (*aarch64_<su>mlal_lo<mode>): Likewise.
3578 (*aarch64_<su>mlal_hi<mode>): Likewise.
3579 (*aarch64_<su>mlsl_lo<mode>): Likewise.
3580 (*aarch64_<su>mlsl_hi<mode>): Likewise.
3581 (*aarch64_<su>mlal<mode>): Likewise.
3582 (*aarch64_<su>mlsl<mode>): Likewise.
3583 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
3584 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
3585 (add<mode>3): Likewise.
3586 (sub<mode>3): Likewise.
3587 (mul<mode>3): Likewise.
3588 (div<mode>3): Likewise.
3589 (neg<mode>2): Likewise.
3590 (abs<mode>2): Likewise.
3591 (fma<mode>4): Likewise.
3592 (*aarch64_fma4_elt<mode>): Likewise.
3593 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
3594 (*aarch64_fma4_elt_to_128df): Likewise.
3595 (*aarch64_fma4_elt_to_64v2df): Likewise.
3596 (fnma<mode>4): Likewise.
3597 (*aarch64_fnma4_elt<mode>): Likewise.
3598 (*aarch64_fnma4_elt_<vswap_width_name><mode>
3599 (*aarch64_fnma4_elt_to_128df): Likewise.
3600 (*aarch64_fnma4_elt_to_64v2df): Likewise.
3601 (<frint_pattern><mode>2): Likewise.
3602 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
3603 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
3604 (vec_unpacks_lo_v4sf): Likewise.
3605 (aarch64_float_extend_lo_v2df): Likewise.
3606 (vec_unpacks_hi_v4sf): Likewise.
3607 (aarch64_float_truncate_lo_v2sf): Likewise.
3608 (aarch64_float_truncate_hi_v4sf): Likewise.
3609 (aarch64_vmls<mode>): Likewise.
3610 (<su><maxmin><mode>3): Likewise.
3611 (<maxmin_uns><mode>3): Likewise.
3612 (reduc_<sur>plus_<mode>): Likewise.
3613 (reduc_<sur>plus_v2di): Likewise.
3614 (reduc_<sur>plus_v2si): Likewise.
3615 (reduc_<sur>plus_<mode>): Likewise.
3616 (aarch64_addpv4sf): Likewise.
3617 (clz<mode>2): Likewise.
3618 (reduc_<maxmin_uns>_<mode>): Likewise.
3619 (reduc_<maxmin_uns>_v2di): Likewise.
3620 (reduc_<maxmin_uns>_v2si): Likewise.
3621 (reduc_<maxmin_uns>_<mode>): Likewise.
3622 (reduc_<maxmin_uns>_v4sf): Likewise.
3623 (aarch64_simd_bsl<mode>_internal): Likewise.
3624 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
3625 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
3626 (aarch64_get_lane<mode>): Likewise.
3627 (*aarch64_combinez<mode>): Likewise.
3628 (aarch64_combine<mode>): Likewise.
3629 (aarch64_simd_combine<mode>): Likewise.
3630 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
3631 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
3632 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
3633 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
3634 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
3635 (aarch64_<sur>h<addsub><mode>): Likewise.
3636 (aarch64_<sur><addsub>hn<mode>): Likewise.
3637 (aarch64_<sur><addsub>hn2<mode>): Likewise.
3638 (aarch64_pmul<mode>): Likewise.
3639 (aarch64_<su_optab><optab><mode>): Likewise.
3640 (aarch64_<sur>qadd<mode>): Likewise.
3641 (aarch64_sqmovun<mode>): Likewise.
3642 (aarch64_<sur>qmovn<mode>): Likewise.
3643 (aarch64_s<optab><mode>): Likewise.
3644 (aarch64_sq<r>dmulh<mode>): Likewise.
3645 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3646 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
3647 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3648 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
3649 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3650 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3651 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
3652 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
3653 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
3654 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
3655 (aarch64_sqdmull<mode>): Likewise.
3656 (aarch64_sqdmull_lane<mode>_internal): Likewise.
3657 (aarch64_sqdmull_n<mode>): Likewise.
3658 (aarch64_sqdmull2<mode>_internal): Likewise.
3659 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
3660 (aarch64_sqdmull2_n<mode>_internal): Likewise.
3661 (aarch64_<sur>shl<mode>): Likewise.
3662 (aarch64_<sur>q<r>shl<mode>
3663 (aarch64_<sur>shll_n<mode>): Likewise.
3664 (aarch64_<sur>shll2_n<mode>): Likewise.
3665 (aarch64_<sur>shr_n<mode>): Likewise.
3666 (aarch64_<sur>sra_n<mode>): Likewise.
3667 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
3668 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
3669 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
3670 (aarch64_cm<optab><mode>): Likewise.
3671 (aarch64_cm<optab>di): Likewise.
3672 (aarch64_cm<optab><mode>): Likewise.
3673 (aarch64_cm<optab>di): Likewise.
3674 (aarch64_cmtst<mode>): Likewise.
3675 (aarch64_cmtstdi): Likewise.
3676 (aarch64_cm<optab><mode>): Likewise.
3677 (*aarch64_fac<optab><mode>): Likewise.
3678 (aarch64_addp<mode>): Likewise.
3679 (aarch64_addpdi): Likewise.
3680 (sqrt<mode>2): Likewise.
3681 (vec_load_lanesoi<mode>): Likewise.
3682 (vec_store_lanesoi<mode>): Likewise.
3683 (vec_load_lanesci<mode>): Likewise.
3684 (vec_store_lanesci<mode>): Likewise.
3685 (vec_load_lanesxi<mode>): Likewise.
3686 (vec_store_lanesxi<mode>): Likewise.
3687 (*aarch64_mov<mode>): Likewise.
3688 (aarch64_ld2<mode>_dreg): Likewise.
3689 (aarch64_ld2<mode>_dreg): Likewise.
3690 (aarch64_ld3<mode>_dreg): Likewise.
3691 (aarch64_ld3<mode>_dreg): Likewise.
3692 (aarch64_ld4<mode>_dreg): Likewise.
3693 (aarch64_ld4<mode>_dreg): Likewise.
3694 (aarch64_tbl1<mode>): Likewise.
3695 (aarch64_tbl2v16qi): Likewise.
3696 (aarch64_combinev16qi): Likewise.
3697 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
3698 (aarch64_st2<mode>_dreg): Likewise.
3699 (aarch64_st2<mode>_dreg): Likewise.
3700 (aarch64_st3<mode>_dreg): Likewise.
3701 (aarch64_st3<mode>_dreg): Likewise.
3702 (aarch64_st4<mode>_dreg): Likewise.
3703 (aarch64_st4<mode>_dreg): Likewise.
3704 (*aarch64_simd_ld1r<mode>): Likewise.
3705 (aarch64_frecpe<mode>): Likewise.
3706 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
3707 (aarch64_frecps<mode>): Likewise.
3709 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3711 * config/arm/iterators.md (V_elem_ch): New.
3713 (VQH_type): Likewise.
3714 * config/arm/arm.md (is_neon_type): New.
3715 (conds): Use is_neon_type.
3716 (anddi3_insn): Update type attribute.
3717 (xordi3_insn): Likewise.
3718 (one_cmpldi2): Likewise.
3719 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
3720 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
3721 (*movmisalign<mode>_neon_store): Likewise.
3722 (*movmisalign<mode>_neon_load): Likewise.
3723 (vec_set<mode>_internal): Likewise.
3724 (vec_set<mode>_internal): Likewise.
3725 (vec_setv2di_internal): Likewise.
3726 (vec_extract<mode>): Likewise.
3727 (vec_extract<mode>): Likewise.
3728 (vec_extractv2di): Likewise.
3729 (*add<mode>3_neon): Likewise.
3730 (adddi3_neon): Likewise.
3731 (*sub<mode>3_neon): Likewise.
3732 (subdi3_neon): Likewise.
3733 (fma<VCVTF:mode>4): Likewise.
3734 (fma<VCVTF:mode>4_intrinsic): Likewise.
3735 (*fmsub<VCVTF:mode>4): Likewise.
3736 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3737 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3738 (ior<mode>3): Likewise.
3739 (and<mode>3): Likewise.
3740 (orn<mode>3_neon): Likewise.
3741 (orndi3_neon): Likewise.
3742 (bic<mode>3_neon): Likewise.
3743 (bicdi3_neon): Likewise.
3744 (xor<mode>3): Likewise.
3745 (one_cmpl<mode>2): Likewise.
3746 (abs<mode>2): Likewise.
3747 (neg<mode>2): Likewise.
3748 (negdi2_neon): Likewise.
3749 (*umin<mode>3_neon): Likewise.
3750 (*umax<mode>3_neon): Likewise.
3751 (*smin<mode>3_neon): Likewise.
3752 (*smax<mode>3_neon): Likewise.
3753 (vashl<mode>3): Likewise.
3754 (vashr<mode>3_imm): Likewise.
3755 (vlshr<mode>3_imm): Likewise.
3756 (ashl<mode>3_signed): Likewise.
3757 (ashl<mode>3_unsigned): Likewise.
3758 (neon_load_count): Likewise.
3759 (ashldi3_neon_noclobber): Likewise.
3760 (ashldi3_neon): Likewise.
3761 (signed_shift_di3_neon): Likewise.
3762 (unsigned_shift_di3_neon): Likewise.
3763 (ashrdi3_neon_imm_noclobber): Likewise.
3764 (lshrdi3_neon_imm_noclobber): Likewise.
3765 (<shift>di3_neon): Likewise.
3766 (widen_ssum<mode>3): Likewise.
3767 (widen_usum<mode>3): Likewise.
3768 (quad_halves_<code>v4si): Likewise.
3769 (quad_halves_<code>v4sf): Likewise.
3770 (quad_halves_<code>v8hi): Likewise.
3771 (quad_halves_<code>v16qi): Likewise.
3772 (reduc_splus_v2di): Likewise.
3773 (neon_vpadd_internal<mode>): Likewise.
3774 (neon_vpsmin<mode>): Likewise.
3775 (neon_vpsmax<mode>): Likewise.
3776 (neon_vpumin<mode>): Likewise.
3777 (neon_vpumax<mode>): Likewise.
3778 (*ss_add<mode>_neon): Likewise.
3779 (*us_add<mode>_neon): Likewise.
3780 (*ss_sub<mode>_neon): Likewise.
3781 (*us_sub<mode>_neon): Likewise.
3782 (neon_vadd<mode>_unspec): Likewise.
3783 (neon_vaddl<mode>): Likewise.
3784 (neon_vaddw<mode>): Likewise.
3785 (neon_vhadd<mode>): Likewise.
3786 (neon_vqadd<mode>): Likewise.
3787 (neon_vaddhn<mode>): Likewise.
3788 (neon_vmul<mode>): Likewise.
3789 (neon_vfms<VCVTF:mode>): Likewise.
3790 (neon_vmlal<mode>): Likewise.
3791 (neon_vmls<mode>): Likewise.
3792 (neon_vmlsl<mode>): Likewise.
3793 (neon_vqdmulh<mode>): Likewise.
3794 (neon_vqdmlal<mode>): Likewise.
3795 (neon_vqdmlsl<mode>): Likewise.
3796 (neon_vmull<mode>): Likewise.
3797 (neon_vqdmull<mode>): Likewise.
3798 (neon_vsub<mode>_unspec): Likewise.
3799 (neon_vsubl<mode>): Likewise.
3800 (neon_vsubw<mode>): Likewise.
3801 (neon_vqsub<mode>): Likewise.
3802 (neon_vhsub<mode>): Likewise.
3803 (neon_vsubhn<mode>): Likewise.
3804 (neon_vceq<mode>): Likewise.
3805 (neon_vcge<mode>): Likewise.
3806 (neon_vcgeu<mode>): Likewise.
3807 (neon_vcgt<mode>): Likewise.
3808 (neon_vcgtu<mode>): Likewise.
3809 (neon_vcle<mode>): Likewise.
3810 (neon_vclt<mode>): Likewise.
3811 (neon_vcage<mode>): Likewise.
3812 (neon_vcagt<mode>): Likewise.
3813 (neon_vtst<mode>): Likewise.
3814 (neon_vabd<mode>): Likewise.
3815 (neon_vabdl<mode>): Likewise.
3816 (neon_vaba<mode>): Likewise.
3817 (neon_vabal<mode>): Likewise.
3818 (neon_vmax<mode>): Likewise.
3819 (neon_vmin<mode>): Likewise.
3820 (neon_vpaddl<mode>): Likewise.
3821 (neon_vpadal<mode>): Likewise.
3822 (neon_vpmax<mode>): Likewise.
3823 (neon_vpmin<mode>): Likewise.
3824 (neon_vrecps<mode>): Likewise.
3825 (neon_vrsqrts<mode>): Likewise.
3826 (neon_vqabs<mode>): Likewise.
3827 (neon_vqneg<mode>): Likewise.
3828 (neon_vcls<mode>): Likewise.
3829 (clz<mode>2): Likewise.
3830 (popcount<mode>2): Likewise.
3831 (neon_vrecpe<mode>): Likewise.
3832 (neon_vrsqrte<mode>): Likewise.
3833 (neon_vget_lane<mode>_sext_internal): Likewise.
3834 (neon_vget_lane<mode>_zext_internal): Likewise.
3835 (neon_vdup_n<mode>): Likewise.
3836 (neon_vdup_n<mode>): Likewise.
3837 (neon_vdup_nv2di): Likewise.
3838 (neon_vdup_lane<mode>_interal): Likewise.
3839 (*neon_vswp<mode>): Likewise.
3840 (neon_vcombine<mode>): Likewise.
3841 (float<mode><V_cvtto>2): Likewise.
3842 (floatuns<mode><V_cvtto>2): Likewise.
3843 (fix_trunc<mode><V_cvtto>2): Likewise.
3844 (fixuns_trunc<mode><V_cvtto>2
3845 (neon_vcvt<mode>): Likewise.
3846 (neon_vcvt<mode>): Likewise.
3847 (neon_vcvtv4sfv4hf): Likewise.
3848 (neon_vcvtv4hfv4sf): Likewise.
3849 (neon_vcvt_n<mode>): Likewise.
3850 (neon_vcvt_n<mode>): Likewise.
3851 (neon_vmovn<mode>): Likewise.
3852 (neon_vqmovn<mode>): Likewise.
3853 (neon_vqmovun<mode>): Likewise.
3854 (neon_vmovl<mode>): Likewise.
3855 (neon_vmul_lane<mode>): Likewise.
3856 (neon_vmul_lane<mode>): Likewise.
3857 (neon_vmull_lane<mode>): Likewise.
3858 (neon_vqdmull_lane<mode>): Likewise.
3859 (neon_vqdmulh_lane<mode>): Likewise.
3860 (neon_vqdmulh_lane<mode>): Likewise.
3861 (neon_vmla_lane<mode>): Likewise.
3862 (neon_vmla_lane<mode>): Likewise.
3863 (neon_vmlal_lane<mode>): Likewise.
3864 (neon_vqdmlal_lane<mode>): Likewise.
3865 (neon_vmls_lane<mode>): Likewise.
3866 (neon_vmls_lane<mode>): Likewise.
3867 (neon_vmlsl_lane<mode>): Likewise.
3868 (neon_vqdmlsl_lane<mode>): Likewise.
3869 (neon_vext<mode>): Likewise.
3870 (neon_vrev64<mode>): Likewise.
3871 (neon_vrev32<mode>): Likewise.
3872 (neon_vrev16<mode>): Likewise.
3873 (neon_vbsl<mode>_internal): Likewise.
3874 (neon_vshl<mode>): Likewise.
3875 (neon_vqshl<mode>): Likewise.
3876 (neon_vshr_n<mode>): Likewise.
3877 (neon_vshrn_n<mode>): Likewise.
3878 (neon_vqshrn_n<mode>): Likewise.
3879 (neon_vqshrun_n<mode>): Likewise.
3880 (neon_vshl_n<mode>): Likewise.
3881 (neon_vqshl_n<mode>): Likewise.
3882 (neon_vqshlu_n<mode>): Likewise.
3883 (neon_vshll_n<mode>): Likewise.
3884 (neon_vsra_n<mode>): Likewise.
3885 (neon_vsri_n<mode>): Likewise.
3886 (neon_vsli_n<mode>): Likewise.
3887 (neon_vtbl1v8qi): Likewise.
3888 (neon_vtbl2v8qi): Likewise.
3889 (neon_vtbl3v8qi): Likewise.
3890 (neon_vtbl4v8qi): Likewise.
3891 (neon_vtbl1v16qi): Likewise.
3892 (neon_vtbl2v16qi): Likewise.
3893 (neon_vcombinev16qi): Likewise.
3894 (neon_vtbx1v8qi): Likewise.
3895 (neon_vtbx2v8qi): Likewise.
3896 (neon_vtbx3v8qi): Likewise.
3897 (neon_vtbx4v8qi): Likewise.
3898 (*neon_vtrn<mode>_insn): Likewise.
3899 (*neon_vzip<mode>_insn): Likewise.
3900 (*neon_vuzp<mode>_insn): Likewise.
3901 (neon_vld1<mode>): Likewise.
3902 (neon_vld1_lane<mode>): Likewise.
3903 (neon_vld1_lane<mode>): Likewise.
3904 (neon_vld1_dup<mode>): Likewise.
3905 (neon_vld1_dup<mode>): Likewise.
3906 (neon_vld1_dupv2di): Likewise.
3907 (neon_vst1<mode>): Likewise.
3908 (neon_vst1_lane<mode>): Likewise.
3909 (neon_vst1_lane<mode>): Likewise.
3910 (neon_vld2<mode>): Likewise.
3911 (neon_vld2<mode>): Likewise.
3912 (neon_vld2_lane<mode>): Likewise.
3913 (neon_vld2_lane<mode>): Likewise.
3914 (neon_vld2_dup<mode>): Likewise.
3915 (neon_vst2<mode>): Likewise.
3916 (neon_vst2<mode>): Likewise.
3917 (neon_vst2_lane<mode>): Likewise.
3918 (neon_vst2_lane<mode>): Likewise.
3919 (neon_vld3<mode>): Likewise.
3920 (neon_vld3qa<mode>): Likewise.
3921 (neon_vld3qb<mode>): Likewise.
3922 (neon_vld3_lane<mode>): Likewise.
3923 (neon_vld3_lane<mode>): Likewise.
3924 (neon_vld3_dup<mode>): Likewise.
3925 (neon_vst3<mode>): Likewise.
3926 (neon_vst3qa<mode>): Likewise.
3927 (neon_vst3qb<mode>): Likewise.
3928 (neon_vst3_lane<mode>): Likewise.
3929 (neon_vst3_lane<mode>): Likewise.
3930 (neon_vld4<mode>): Likewise.
3931 (neon_vld4qa<mode>): Likewise.
3932 (neon_vld4qb<mode>): Likewise.
3933 (neon_vld4_lane<mode>): Likewise.
3934 (neon_vld4_lane<mode>): Likewise.
3935 (neon_vld4_dup<mode>): Likewise.
3936 (neon_vst4<mode>): Likewise.
3937 (neon_vst4qa<mode>): Likewise.
3938 (neon_vst4qb<mode>): Likewise.
3939 (neon_vst4_lane<mode>): Likewise.
3940 (neon_vst4_lane<mode>): Likewise.
3941 (neon_vec_unpack<US>_lo_<mode>): Likewise.
3942 (neon_vec_unpack<US>_hi_<mode>): Likewise.
3943 (neon_vec_<US>mult_lo_<mode>): Likewise.
3944 (neon_vec_<US>mult_hi_<mode>): Likewise.
3945 (neon_vec_<US>shiftl_<mode>): Likewise.
3946 (neon_unpack<US>_<mode>): Likewise.
3947 (neon_vec_<US>mult_<mode>): Likewise.
3948 (vec_pack_trunc_<mode>): Likewise.
3949 (neon_vec_pack_trunc_<mode>): Likewise.
3950 (neon_vabd<mode>_2): Likewise.
3951 (neon_vabd<mode>_3): Likewise.
3953 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3955 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
3956 (load_pair): Update type attribute.
3957 (store_pair): Update type attribute.
3958 * config/aarch64/iterators.md (q): New.
3960 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3962 * config/arm/types.md: Add new types for Neon insns.
3964 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
3965 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3966 Sergey Lega <sergey.s.lega@intel.com>
3967 Anna Tikhonova <anna.tikhonova@intel.com>
3968 Ilya Tocar <ilya.tocar@intel.com>
3969 Andrey Turetskiy <andrey.turetskiy@intel.com>
3970 Ilya Verbin <ilya.verbin@intel.com>
3971 Kirill Yukhin <kirill.yukhin@intel.com>
3972 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3974 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
3975 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
3976 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
3978 (srcp14<mode>): Ditto.
3979 (rsqrt14<mode>): Ditto.
3980 (rsqrt14<mode>): Ditto.
3981 (avx512f_vmscalef<mode>): Ditto.
3982 (avx512f_scalef<mode>): Ditto.
3983 (avx512f_getexp<mode>): Ditto.
3984 (avx512f_sgetexp<mode>): Ditto.
3985 (avx512f_fixupimm<mode>): Ditto.
3986 (avx512f_sfixupimm<mode>): Ditto.
3987 (avx512f_rndscale<mode>): Ditto.
3988 (*avx512er_exp2<mode>): Ditto.
3989 (*avx512er_rcp28<mode>): Ditto.
3990 (avx512er_rsqrt28<mode>): Ditto.
3991 (avx512f_getmant<mode>): Ditto.
3992 (avx512f_getmant<mode>): Ditto.
3993 (avx512f_rndscale<mode>): Fix formatting.
3995 2013-10-15 Martin Jambor <mjambor@suse.cz>
3997 * ipa-utils.h (ipa_edge_within_scc): Declare.
3998 * ipa-cp.c (edge_within_scc): Moved...
3999 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
4001 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4002 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4003 Sergey Lega <sergey.s.lega@intel.com>
4004 Anna Tikhonova <anna.tikhonova@intel.com>
4005 Ilya Tocar <ilya.tocar@intel.com>
4006 Andrey Turetskiy <andrey.turetskiy@intel.com>
4007 Ilya Verbin <ilya.verbin@intel.com>
4008 Kirill Yukhin <kirill.yukhin@intel.com>
4009 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4011 * config/i386/predicates.md (const_8_to_15_operand): New.
4012 (const_16_to_31_operand): Ditto.
4013 * config/i386/sse.md (V8FI): New.
4015 (reduc_splus_v8df): Ditto.
4016 (reduc_splus_v16sf): Ditto.
4017 (avx512f_vextract<shuffletype>32x4_1): Ditto.
4018 (vec_extract_hi_<mode>): Ditto.
4019 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
4020 (vec_set_lo_<mode>): Ditto.
4021 (vec_set_hi_<mode>): Ditto.
4022 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
4023 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
4024 (avx512f_pshufd_1): Ditto.
4025 (avx512f_broadcast<mode>): Ditto.
4026 (avx512f_broadcast<mode>): Ditto.
4027 (define_split): Split vec_extract_lo into move.
4028 (ssequartermode): Ditto.
4029 (ssedoublemode): Extened with wider modes.
4030 (vec_extract_lo_<mode>): Ditto.
4032 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4033 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4034 Sergey Lega <sergey.s.lega@intel.com>
4035 Anna Tikhonova <anna.tikhonova@intel.com>
4036 Ilya Tocar <ilya.tocar@intel.com>
4037 Andrey Turetskiy <andrey.turetskiy@intel.com>
4038 Ilya Verbin <ilya.verbin@intel.com>
4039 Kirill Yukhin <kirill.yukhin@intel.com>
4040 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4042 * config/i386/predicates.md (register_or_constm1_operand): New.
4043 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
4044 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
4045 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
4046 UNSPEC_SCATTER_PREFETCH
4048 (avx512f_ucmp<mode>3): Ditto.
4049 (avx512f_vternlog<mode>): Ditto.
4050 (avx512f_align<mode>): Ditto.
4051 (<shift_insn><mode>3): Ditto.
4052 (avx512f_<rotate>v<mode>): Ditto.
4053 (avx512f_<rotate><mode>): Ditto.
4054 (avx512f_eq<mode>3): Ditto.
4055 (avx512f_eq<mode>3_1): Ditto.
4056 (avx512f_gt<mode>3): Ditto.
4057 (avx512f_testm<mode>3): Ditto.
4058 (avx512f_testnm<mode>3): Ditto.
4059 (avx512pf_gatherpf<mode>): Ditto.
4060 (*avx512pf_gatherpf<mode>_mask): Ditto.
4061 (*avx512pf_gatherpf<mode>): Ditto.
4062 (avx512pf_scatterpf<mode>): Ditto.
4063 (*avx512pf_scatterpf<mode>_mask): Ditto.
4064 (*avx512pf_scatterpf<mode>): Ditto.
4065 (avx512f_vec_dup_gpr<mode>): Ditto.
4066 (clz<mode>2): Ditto.
4067 (conflict<mode>): Ditto.
4068 (REDUC_SMINMAX_MODE): Extened with wider modes.
4069 (reduc_<code>_<mode>): Ditto.
4070 (vlshr<mode>3): Ditto.
4071 (vashl<mode>3): Ditto.
4073 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4074 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4075 Sergey Lega <sergey.s.lega@intel.com>
4076 Anna Tikhonova <anna.tikhonova@intel.com>
4077 Ilya Tocar <ilya.tocar@intel.com>
4078 Andrey Turetskiy <andrey.turetskiy@intel.com>
4079 Ilya Verbin <ilya.verbin@intel.com>
4080 Kirill Yukhin <kirill.yukhin@intel.com>
4081 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4083 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
4086 (avx512fmaskmode): Ditto.
4087 (bcstscalarsuff): Ditto.
4088 (avx512f_blendm<mode>): Ditto.
4089 (cmp_imm_predicate): Ditto.
4090 (avx512f_cmp<mode>3): Ditto.
4091 (avx512f_vec_dup<mode>): Ditto.
4092 (avx512f_vec_dup_mem<mode>): Ditto.
4093 (avx512f_vpermi2var<mode>3): Ditto.
4094 (avx512f_vpermt2var<mode>3): Ditto.
4095 (vec_init<mode>): Ditto.
4096 (avx512f_gathersi<mode>): Ditto.
4097 (*avx512f_gathersi<mode>): Ditto.
4098 (*avx512f_gathersi<mode>_2): Ditto.
4099 (avx512f_gatherdi<mode>): Ditto.
4100 (*avx512f_gatherdi<mode>): Ditto.
4101 (*avx512f_gatherdi<mode>_2): Ditto.
4102 (avx512f_scattersi<mode>): Ditto.
4103 (*avx512f_scattersi<mode>): Ditto.
4104 (avx512f_scatterdi<mode>): Ditto.
4105 (*avx512f_scatterdi<mode>): Ditto.
4106 (sseintprefix): Extened with wider modes.
4107 (VEC_GATHER_IDXSI): Ditto.
4108 (VEC_GATHER_IDXDI): Ditto.
4109 (VEC_GATHER_SRCDI): Ditto.
4111 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
4112 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4114 * config/arm/t-aprofile: New file.
4115 * config.gcc: Handle --with-multilib-list option.
4117 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
4119 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
4122 2013-10-15 Richard Biener <rguenther@suse.de>
4124 * tree-loop-distribution.c (build_empty_rdg): Inline into
4126 (rdg_flag_vertex): Inline into single user.
4127 (rdg_flag_vertex_and_dependent): Likewise.
4128 (build_rdg_partition_for_vertex): Remove processed bitmap.
4129 (rdg_build_partitions): Simplify.
4131 2013-10-15 Richard Biener <rguenther@suse.de>
4133 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
4134 Restructure forwarding through conversions and copies to
4135 avoid performing copy-propagation the wrong way. Adjust
4136 recursion invocations.
4137 (forward_propagate_addr_expr): Add argument stating if we
4138 are recursing from a single-use.
4139 (ssa_forward_propagate_and_combine): Adjust.
4141 2013-10-14 David Malcolm <dmalcolm@redhat.com>
4143 * dumpfile.h (gcc::dump_manager): New class, to hold state
4144 relating to dumpfile management.
4145 (get_dump_file_name): Remove in favor of method of dump_manager.
4146 (dump_initialized_p): Likewise.
4147 (dump_start): Likewise.
4148 (dump_finish): Likewise.
4149 (dump_switch_p): Likewise.
4150 (dump_register): Likewise.
4151 (get_dump_file_info): Likewise.
4152 * context.c (gcc::context::context): Construct the dump_manager
4154 * context.h (gcc::context::get_dumps): New.
4155 (gcc::context::m_dumps): New.
4156 * coverage.c (coverage_init): Port to dump_manager API.
4157 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
4158 (extra_dump_files_in_use): Likewise.
4159 (extra_dump_files_alloced): Likewise.
4160 (gcc::dump_manager::dump_manager): New.
4161 (dump_register): Convert to...
4162 (gcc::dump_manager::dump_register): ...method, replacing
4163 function-static next_dump with m_next_dump field.
4164 (get_dump_file_info): Convert to...
4165 (gcc::dump_manager::get_dump_file_info): ...method.
4166 (get_dump_file_name): Convert to...
4167 (gcc::dump_manager::get_dump_file_name): ...method.
4168 (dump_start): Convert to...
4169 (gcc::dump_manager::dump_start): ...method.
4170 (dump_finish): Convert to...
4171 (gcc::dump_manager::dump_finish): ...method.
4172 (dump_begin): Replace body with...
4173 (gcc::dump_manager::dump_begin): ...new method.
4174 (dump_phase_enabled_p): Convert to...
4175 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4176 (dump_phase_enabled_p): Convert to...
4177 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4178 (dump_initialized_p): Convert to...
4179 (gcc::dump_manager::dump_initialized_p): ...method.
4180 (dump_flag_name): Replace body with...
4181 (gcc::dump_manager::dump_flag_name): ...new method.
4182 (dump_enable_all): Convert to...
4183 (gcc::dump_manager::dump_enable_all): ...new method.
4184 (opt_info_enable_passes): Convert to...
4185 (gcc::dump_manager::opt_info_enable_passes): ...new method.
4186 (dump_switch_p_1): Convert to...
4187 (gcc::dump_manager::dump_switch_p_1): ...new method.
4188 (dump_switch_p): Convert to...
4189 (gcc::dump_manager::dump_switch_p): ...new method.
4190 (opt_info_switch_p): Port to dump_manager API.
4191 (enable_rtl_dump_file): Likewise.
4192 * opts-global.c (handle_common_deferred_options): Port to new
4194 * passes.c (pass_manager::finish_optimization_passes): Likewise.
4195 (pass_manager::register_one_dump_file): Likewise.
4196 (pass_manager::register_pass): Likewise.
4197 (pass_init_dump_file): Likewise.
4198 (pass_fini_dump_file): Likewise.
4199 * statistics.c (statistics_early_init): Likewise.
4201 2013-10-14 Richard Biener <rguenther@suse.de>
4203 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
4204 iterative_hash_canonical_type, gimple_canonical_type_hash,
4205 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
4206 gimple_register_canonical_type, print_gimple_types_stats,
4207 free_gimple_type_tables): Move to lto/lto.c
4208 (gt-gimple.h): Do not include.
4209 * gimple.h (gimple_register_canonical_type,
4210 print_gimple_types_stats, free_gimple_type_tables): Remove.
4211 * Makefile.in (GTFILES): Remove gimple.c.
4213 2013-10-14 Travis Snoozy <quandary@remstate.com>
4216 * config/msp430/msp430.c (msp430_option_override): Correct thinko
4217 scanning for msp430x targets.
4219 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
4222 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
4223 (registers_ok_for_ldd_peep): Move around.
4224 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
4225 * config/sparc/sparc.md (widening peepholes): Use it.
4227 2013-10-14 Richard Biener <rguenther@suse.de>
4231 * gimple.c (iterative_hash_canonical_type): Make sure to
4232 record the hash into the correct hashtable slot.
4234 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
4236 PR rtl-optimization/58662
4237 * combine.c (try_combine): Take into account death nodes on I2 when
4238 splitting a PARALLEL of two independent SETs. Fix dump message.
4240 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
4243 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
4244 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
4245 * config/sh/t-sh (sh_treg_combine.o): New entry.
4246 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
4247 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
4248 (register_sh_passes): New function. Register sh_treg_combine pass.
4249 (sh_option_override): Invoke it.
4250 (sh_canonicalize_comparison): Handle op0_preserve_value.
4251 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
4252 opportunities. Canonicalize branch condition.
4253 (nott): Allow only if pseudos can be created for non-SH2A.
4255 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
4258 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
4259 (ix86_expand_movmem): Replace copy_addr_to_reg with
4260 ix86_copy_addr_to_reg.
4261 (ix86_expand_setmem): Likewise.
4263 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
4265 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
4266 provided by ix86_fp_compare_mode instead of CCFPUmode.
4268 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
4270 * config/aarch64/arm_neon.h
4271 (vtbx<1,3>_<psu>8): Fix register constriants.
4273 2013-10-11 Jeff Law <law@redhat.com>
4275 PR tree-optimization/58640
4276 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
4277 threading paths that cross over two loop entry points.
4279 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4281 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
4282 handle vector float as well.
4283 (*vsx_le_perm_load_v4si): Likewise.
4284 (*vsx_le_perm_store_v2di): Likewise.
4285 (*vsx_le_perm_store_v4si): Likewise.
4287 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4289 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
4290 directly to circumvent subtract from splat{31} workaround.
4291 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
4293 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
4294 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
4295 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
4296 (altivec_vperm_<mode>): Convert to define_insn_and_split to
4297 separate big and little endian logic.
4298 (*altivec_vperm_<mode>_internal): New define_insn.
4299 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
4300 separate big and little endian logic.
4301 (*altivec_vperm_<mode>_uns_internal): New define_insn.
4302 (vec_permv16qi): Add little endian logic.
4304 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
4306 * doc/extend.texi (returns_nonnull): Remove arguments.
4308 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4309 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4310 Sergey Lega <sergey.s.lega@intel.com>
4311 Anna Tikhonova <anna.tikhonova@intel.com>
4312 Ilya Tocar <ilya.tocar@intel.com>
4313 Andrey Turetskiy <andrey.turetskiy@intel.com>
4314 Ilya Verbin <ilya.verbin@intel.com>
4315 Kirill Yukhin <kirill.yukhin@intel.com>
4316 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4318 * config/i386/sse.md (VI48F_256_512): New.
4319 (avx2_permvar<mode>): Change to ...
4320 (<avx2_avx512f>_permvar<mode>): This.
4322 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4323 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4324 Sergey Lega <sergey.s.lega@intel.com>
4325 Anna Tikhonova <anna.tikhonova@intel.com>
4326 Ilya Tocar <ilya.tocar@intel.com>
4327 Andrey Turetskiy <andrey.turetskiy@intel.com>
4328 Ilya Verbin <ilya.verbin@intel.com>
4329 Kirill Yukhin <kirill.yukhin@intel.com>
4330 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4332 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
4333 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
4334 * config/i386/sse.md (VI4_AVX): New.
4335 (sf2simodelower): Ditto.
4336 (sse2_cvtps2dq): Change to ...
4337 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
4339 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4340 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4341 Sergey Lega <sergey.s.lega@intel.com>
4342 Anna Tikhonova <anna.tikhonova@intel.com>
4343 Ilya Tocar <ilya.tocar@intel.com>
4344 Andrey Turetskiy <andrey.turetskiy@intel.com>
4345 Ilya Verbin <ilya.verbin@intel.com>
4346 Kirill Yukhin <kirill.yukhin@intel.com>
4347 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4349 * config/i386/sse.md (V_512): New.
4351 (vcond<V_512:mode><VF_512:mode>): Ditto.
4352 (vcond<V_512:mode><VI_512:mode>): Ditto.
4353 (vcondu<V_512:mode><VI_512:mode>): Ditto.
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>
4365 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
4366 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
4368 (fma<mode>4): Removed condition.
4369 (fms<mode>4): Ditto.
4370 (fnma<mode>4): Ditto.
4371 (fnms<mode>4): Ditto.
4372 (fma4i_fmadd_<mode>): Ditto.
4373 (*fma_fmadd_<mode>): Ditto.
4374 (*fma_fmsub_<mode>): Ditto.
4375 (*fma_fnmadd_<mode>): Ditto.
4376 (*fma_fnmsub_<mode>): Ditto.
4377 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
4378 (*fma_fmaddsub_<mode>): Ditto.
4379 (*fma_fmsubadd_<mode>): Ditto.
4380 (*fmai_fmadd_<mode>): Ditto.
4381 (*fmai_fmsub_<mode>): Ditto.
4382 (*fmai_fnmadd_<mode>): Ditto.
4383 (*fmai_fnmsub_<mode>): Ditto.
4385 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4386 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4387 Sergey Lega <sergey.s.lega@intel.com>
4388 Anna Tikhonova <anna.tikhonova@intel.com>
4389 Ilya Tocar <ilya.tocar@intel.com>
4390 Andrey Turetskiy <andrey.turetskiy@intel.com>
4391 Ilya Verbin <ilya.verbin@intel.com>
4392 Kirill Yukhin <kirill.yukhin@intel.com>
4393 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4395 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
4396 (VI124_256): Changed to ...
4397 (VI124_256_48_512): This.
4398 (ssepackmode): Extended with wider modes.
4399 (<code><mode>3): Changed iterator.
4400 (*avx2_<code><mode>3): Ditto.
4401 (vec_pack_trunc_<mode>): Ditto.
4403 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4404 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4405 Sergey Lega <sergey.s.lega@intel.com>
4406 Anna Tikhonova <anna.tikhonova@intel.com>
4407 Ilya Tocar <ilya.tocar@intel.com>
4408 Andrey Turetskiy <andrey.turetskiy@intel.com>
4409 Ilya Verbin <ilya.verbin@intel.com>
4410 Kirill Yukhin <kirill.yukhin@intel.com>
4411 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4413 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
4414 (VI8F_256_512): Ditto.
4415 (abs<mode>2): Changed iterator.
4416 (avx2_perm<mode>): Changed to ...
4417 (<avx2_avx512f>_perm<mode>): This.
4418 (avx2_perm<mode>_1): Changed to ...
4419 (<avx2_avx512f>_perm<mode>_1): This.
4421 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4422 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4423 Sergey Lega <sergey.s.lega@intel.com>
4424 Anna Tikhonova <anna.tikhonova@intel.com>
4425 Ilya Tocar <ilya.tocar@intel.com>
4426 Andrey Turetskiy <andrey.turetskiy@intel.com>
4427 Ilya Verbin <ilya.verbin@intel.com>
4428 Kirill Yukhin <kirill.yukhin@intel.com>
4429 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4431 * config/i386/sse.md (VI48_AVX512F): New.
4432 (VI48_AVX2): Changed to ...
4433 (VI48_AVX2_48_AVX512F): This.
4434 (avx2_ashrv<mode>): Changed to ...
4435 (<avx2_avx512f>_ashrv<mode>): This.
4436 (avx2_<shift_insn>v<mode>): Changed to ...
4437 (<avx2_avx512f>_<shift_insn>v<mode>): This.
4439 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4440 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4441 Sergey Lega <sergey.s.lega@intel.com>
4442 Anna Tikhonova <anna.tikhonova@intel.com>
4443 Ilya Tocar <ilya.tocar@intel.com>
4444 Andrey Turetskiy <andrey.turetskiy@intel.com>
4445 Ilya Verbin <ilya.verbin@intel.com>
4446 Kirill Yukhin <kirill.yukhin@intel.com>
4447 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4449 * config/i386/sse.md (VI4_AVX512F): New.
4450 (VI8_AVX2_AVX512F): Ditto.
4451 (mul<mode>3): Extended with wider modes.
4452 (*<sse4_1_avx2>_mul<mode>3): Ditto.
4453 (mul<mode>3): Ditto.
4454 (vec_widen_<s>mult_odd_<mode>): Ditto.
4456 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4457 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4458 Sergey Lega <sergey.s.lega@intel.com>
4459 Anna Tikhonova <anna.tikhonova@intel.com>
4460 Ilya Tocar <ilya.tocar@intel.com>
4461 Andrey Turetskiy <andrey.turetskiy@intel.com>
4462 Ilya Verbin <ilya.verbin@intel.com>
4463 Kirill Yukhin <kirill.yukhin@intel.com>
4464 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4466 * config/i386/sse.md (VI2_AVX512F): New.
4467 (VI124_AVX512F): Ditto.
4468 (sseunpackmode): Extended with wider modes.
4469 (sseunpackfltmode): Ditto.
4470 (vec_unpacks_float_hi_<mode>): Ditto.
4471 (vec_unpacks_float_lo_<mode>): Ditto.
4472 (vec_unpacku_float_hi_<mode>): Ditto.
4473 (vec_unpacku_float_lo_<mode>): Ditto.
4474 (vec_unpacks_lo_<mode>): Ditto.
4475 (vec_unpacks_hi_<mode>): Ditto.
4476 (vec_unpacku_lo_<mode>): Ditto.
4477 (vec_unpacku_hi_<mode>): Ditto.
4479 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4480 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4481 Sergey Lega <sergey.s.lega@intel.com>
4482 Anna Tikhonova <anna.tikhonova@intel.com>
4483 Ilya Tocar <ilya.tocar@intel.com>
4484 Andrey Turetskiy <andrey.turetskiy@intel.com>
4485 Ilya Verbin <ilya.verbin@intel.com>
4486 Kirill Yukhin <kirill.yukhin@intel.com>
4487 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4489 * config/i386/i386.md (multdiv): New.
4490 (multdiv_mnemonic): Ditto.
4491 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
4492 (<sse>_vm<multdiv_mnemonic><mode>3): This.
4493 (<sse>_vmdiv<mode>3): Removed.
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>
4505 * config/i386/sse.md (V): Extended with wider modes.
4507 (ssehalfvecmode): Ditto.
4509 (ssepackfltmode): Ditto.
4510 (avx_vec_concat<mode>): Ditto.
4511 (V_256_512): New iterator.
4512 (VF2_512_256): Ditto.
4513 (si2dfmode): New attribute.
4514 (si2dfmodelower): Ditto.
4516 (concat_tg_mode): Ditto.
4517 (floatv4siv4df2): Changed to ...
4518 (float<si2dfmodelower><mode>2): This.
4519 (avx_cvtps2pd256): Changed to ...
4520 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
4521 (vec_pack_trunc_v4df): Changed to ...
4522 (vec_pack_trunc_<mode>): This.
4523 (avx_vpermil<mode>): Changed to ...
4524 (<sse2_avx_avx512f>_vpermil<mode>): This.
4525 (<fixsuffix>fix_truncv8dfv8si2): New.
4526 (vec_pack_sfix_trunc_v8df): Ditto.
4527 (avx512f_rndscale<mode>): Ditto.
4528 (avx512f_roundpd512): Ditto.
4529 (vec_pack_ufix_trunc_<mode>): Updated iterator.
4531 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4532 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4533 Sergey Lega <sergey.s.lega@intel.com>
4534 Anna Tikhonova <anna.tikhonova@intel.com>
4535 Ilya Tocar <ilya.tocar@intel.com>
4536 Andrey Turetskiy <andrey.turetskiy@intel.com>
4537 Ilya Verbin <ilya.verbin@intel.com>
4538 Kirill Yukhin <kirill.yukhin@intel.com>
4539 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4541 * config/i386/i386.md (any_fix): New iterator.
4542 (fixsuffix): New attribute.
4543 * config/i386/sse.md (VF1): Extened with wider modes.
4547 (sseintvecmodelower): Ditto.
4548 (ssescalarmode): Ditto.
4549 (ssescalarnum): Ditto.
4551 (ssexmmmode): Ditto.
4552 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
4553 (<sse>_rcp<mode>2): Change iterator.
4554 (rsqrt<mode>2): Ditto.
4555 (<sse>_rsqrt<mode>2): Ditto.
4556 (avx2_vec_dup<mode>): Ditto.
4557 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
4558 (round<mode>2_sfix): Ditto.
4559 (avx2_pbroadcast<mode>): Ditto.
4560 (*andnot<mode>3): Handle XI mode.
4561 (*<code><mode>3): Ditto.
4562 (AVXTOSSEMODE): Removed.
4563 (avx_vpermil<mode>): Changed to ...
4564 (<sse2_avx_avx512f>_vpermil<mode>): This.
4566 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4567 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4568 Sergey Lega <sergey.s.lega@intel.com>
4569 Anna Tikhonova <anna.tikhonova@intel.com>
4570 Ilya Tocar <ilya.tocar@intel.com>
4571 Andrey Turetskiy <andrey.turetskiy@intel.com>
4572 Ilya Verbin <ilya.verbin@intel.com>
4573 Kirill Yukhin <kirill.yukhin@intel.com>
4574 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4576 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
4577 (<sse>_comi): Ditto.
4578 (<sse>_ucomi): Ditto.
4579 (sse_cvtss2siq_2): Ditto.
4580 (sse2_cvtsd2si): Ditto.
4581 (sse2_cvtsd2siq): Ditto.
4582 (sse2_cvttsd2si): Ditto.
4583 (sse2_cvttsd2siq): Ditto.
4584 (<shift_insn><mode>3): Ditto.
4585 (sse2_cvtsi2sdq): Update constraint and prefix.
4586 (sse_cvtsi2ss): Update prefix.
4587 (sse_cvtsi2ssq): Ditto.
4589 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4591 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
4592 ignore internal calls.
4594 2013-10-11 Richard Biener <rguenther@suse.de>
4596 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
4597 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
4598 TREE_OVERFLOW is set.
4600 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
4602 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
4604 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
4605 * gimple.h: Likewise.
4607 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
4609 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
4612 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
4614 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
4615 * configure: Regenerate.
4617 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4619 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
4620 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
4621 formatting fixes, use pp_colon instead of pp_character (..., ':'),
4622 similarly pp_right_paren.
4623 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
4624 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
4625 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
4626 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
4628 (omp_declare_simd_clauses_equal,
4629 omp_remove_redundant_declare_simd_attrs): New functions.
4630 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
4631 (walk_tree_1): Handle new OpenMP 4.0 clauses.
4632 * tree.h (OMP_LOOP_CHECK): Define.
4633 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
4634 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
4635 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
4636 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
4637 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
4638 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
4639 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
4640 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
4641 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
4642 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
4643 OMP_CLAUSE_SIMDLEN_EXPR): Define.
4644 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
4645 (omp_remove_redundant_declare_simd_attrs): New prototype.
4646 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
4647 GIMPLE_OMP_TEAMS): New codes.
4648 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
4649 * omp-low.c (struct omp_context): Add cancel_label and cancellable
4651 (target_nesting_level): New variable.
4652 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
4653 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
4654 for collapse > 1 static schedule unless ordered.
4655 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
4656 (determine_parallel_type): Adjust get_ws_args_for caller.
4657 (install_var_field): Handle mask & 4 for double indirection.
4658 (scan_sharing_clauses): Ignore shared clause on teams construct.
4659 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
4660 (create_omp_child_function): If inside target or declare target
4661 constructs, set "omp declare target" attribute on the child function.
4662 (find_combined_for): New function.
4663 (scan_omp_parallel): Handle combined loops.
4664 (scan_omp_target, scan_omp_teams): New functions.
4665 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
4666 restrictions and set ctx->cancellable for cancellable constructs.
4667 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
4668 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
4669 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
4670 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
4671 (omp_clause_aligned_alignment): New function.
4672 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
4673 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
4674 on teams constructs. Handle user defined reductions and new
4676 (lower_reduction_clauses): Don't set placeholder to address of ref
4677 if it has already the right type.
4678 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
4679 (expand_parallel_call): Use the new non-_start suffixed builtins,
4680 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
4681 and GOMP_parallel_end after the call.
4682 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
4683 (expand_omp_for_init_counts): Handle combined loops.
4684 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
4686 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
4687 end of cancellable loops.
4688 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
4689 Likewise. Handle collapse > 1 loops.
4690 (expand_omp_simd): Handle combined loops.
4691 (expand_omp_for): Add inner_stmt argument, adjust callers of
4692 expand_omp_for* functions, use expand_omp_for_static*chunk even
4693 for collapse > 1 unless ordered.
4694 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
4695 of cancellable sections.
4696 (expand_omp_single): Remove need_barrier variable, just rely on
4697 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
4698 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
4699 (expand_omp_atomic_load, expand_omp_atomic_store,
4700 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
4701 (expand_omp_target): New function.
4702 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
4703 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
4704 (build_omp_regions_1): Immediately close region for
4705 GF_OMP_TARGET_KIND_UPDATE.
4706 (maybe_add_implicit_barrier_cancel): New function.
4707 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
4709 (lower_omp_single): Likewise. Add clobber after the barrier.
4710 (lower_omp_taskgroup): New function.
4711 (lower_omp_for): Handle combined loops. Adjust
4712 lower_rec_input_clauses caller. Handle cancellation.
4713 (lower_depend_clauses): New function.
4714 (lower_omp_taskreg): Lower depend clauses. Adjust
4715 lower_rec_input_clauses caller. Add clobber after the call. Handle
4717 (lower_omp_target, lower_omp_teams): New functions.
4718 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
4719 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
4720 and GOMP_cancellation_point calls.
4721 (lower_omp): Fold stmts inside of target region.
4722 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
4723 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4724 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
4725 (BT_FN_VOID_OMPFN_PTR_UINT,
4726 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
4727 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
4728 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
4729 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
4730 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
4731 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
4732 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
4733 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
4734 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
4735 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
4736 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
4737 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
4738 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
4739 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
4740 BUILT_IN_GOMP_PARALLEL_END.
4741 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
4742 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4743 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
4744 GF_OMP_FOR_KIND_DISTRIBUTE.
4745 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
4746 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
4747 (dump_gimple_omp_return): Print lhs if it has any.
4748 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
4749 gimple_omp_atomic_seq_cst_p.
4750 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
4751 and GIMPLE_OMP_TEAMS.
4752 * langhooks.c (lhd_omp_mappable_type): New function.
4753 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
4754 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
4756 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
4757 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
4758 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
4759 (struct gimplify_omp_ctx): Add combined_loop field.
4760 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
4761 on stmts inside of target region.
4762 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
4763 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
4764 ORT_TARGET and ORT_TARGET_DATA.
4765 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
4767 (omp_notice_threadprivate_variable): Complain about threadprivate
4768 variables in target region.
4769 (omp_notice_variable): Complain about vars with non-mappable type
4770 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
4771 (omp_check_private): Ignore ORT_TARGET* regions.
4772 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
4773 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
4774 (find_combined_omp_for): New function.
4775 (gimplify_omp_for): Handle gimplification of combined loops.
4776 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
4778 (gimplify_omp_target_update): New function.
4779 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
4780 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
4781 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
4782 (gimplify_body): If fndecl has "omp declare target" attribute, add
4783 implicit ORT_TARGET context around it.
4784 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
4785 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
4786 * tree-nested.c (convert_nonlocal_reference_stmt,
4787 convert_local_reference_stmt, convert_gimple_call): Handle
4788 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4789 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
4790 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
4791 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
4792 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
4793 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
4794 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
4795 BUILT_IN_GOMP_LOOP_END_CANCEL,
4796 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
4797 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
4798 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
4799 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
4800 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
4801 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
4802 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
4803 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
4804 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
4805 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
4806 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
4807 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
4808 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
4809 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
4810 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
4811 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4812 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
4813 gimple_build_omp_teams): New functions.
4814 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
4815 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
4816 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
4817 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4818 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
4819 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
4820 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
4821 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
4822 GF_OMP_ATOMIC_SEQ_CST): New.
4823 (gimple_build_omp_taskgroup, gimple_build_omp_target,
4824 gimple_build_omp_teams): New prototypes.
4825 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
4826 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4827 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
4828 GIMPLE_OMP_SINGLE as end of range.
4829 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
4830 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
4831 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
4832 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
4833 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
4834 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
4835 gimple_omp_target_kind, gimple_omp_target_set_kind,
4836 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
4837 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
4838 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
4839 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
4840 gimple_omp_teams_set_clauses): New inlines.
4841 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
4842 and GIMPLE_OMP_TASKGROUP.
4843 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
4844 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
4845 enum omp_clause_proc_bind_kind): New.
4846 (union omp_clause_subcode): Add depend_kind, map_kind and
4847 proc_bind_kind fields.
4848 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
4849 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4850 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
4851 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
4852 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
4854 2013-10-10 Teresa Johnson <tejohnson@google.com>
4856 * predict.c (tree_estimate_probability): Add new parameter
4857 for estimate_bb_frequencies.
4858 (estimate_bb_frequencies): Add new parameter to force estimation.
4859 (rebuild_frequencies): When max frequency in function is small,
4860 recompute counts from frequencies.
4861 * predict.h (estimate_bb_frequencies): New parameter.
4863 2013-10-10 David Malcolm <dmalcolm@redhat.com>
4865 * ipa-inline.c (ipa_inline): Fix leak of "order" when
4866 optimizations are disabled.
4868 2013-10-10 David Malcolm <dmalcolm@redhat.com>
4870 * coverage.c (coverage_finish): Fix leak of da_file_name.
4872 2013-10-10 Jan Hubicka <jh@suse.cz>
4874 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
4875 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
4878 2013-10-10 Jakub Jelinek <jakub@redhat.com>
4881 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
4882 if any labels are in FALLTHRU_BB, use a special label emitted
4883 immediately after the asm goto insn rather than label_rtx
4885 (expand_asm_stmt): Adjust caller.
4886 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
4887 edge if the last insn in predecessor is a jump with single successor,
4888 but it isn't simplejump_p.
4890 2013-10-10 Richard Biener <rguenther@suse.de>
4892 PR tree-optimization/58656
4893 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
4895 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4897 * gimplify.c: Include expr.h and tm_p.h for targets with special
4898 va-arg padding requirements.
4900 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4902 * tree-flow.h: Move some prototypes to gimple.h.
4903 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
4904 * gimple.h: Relocate some prototypes from tree-flow.h
4905 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
4907 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
4908 (build_va_arg_indirect_ref): Relocate and make static.
4909 (std_gimplify_va_arg_expr): Relocate here.
4910 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
4911 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
4913 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4915 * doc/md.texi: Document the mnemonic attribute.
4917 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4920 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
4921 (set_attr_alternative x ...) when searching for user defined
4924 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4926 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
4928 2013-10-09 Easwaran Raman <eraman@google.com>
4930 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
4931 * cfgexpand.c (defer_stack_allocation): ...use here
4932 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
4934 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
4936 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
4937 (operand_equal_for_value_replacement): New function, extracted from
4938 value_replacement and enhanced to catch more cases.
4939 (value_replacement): Use operand_equal_for_value_replacement.
4941 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4943 * loop-doloop.c (doloop_modify, doloop_optimize): Use
4944 get_max_loop_iterations.
4946 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4948 * config/arm/aarch-common.c (arm_early_load_addr_dep):
4949 Place comment above function.
4951 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4953 * tree-flow.h: Remove all remaining prototypes, enums and structs that
4954 are not related to tree-cfg.c.
4955 * tree-ssa-address.h: New file. Relocate prototypes.
4956 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
4957 (addr_for_mem_ref): New. Combine call to get_address_description and
4958 return addr_for_mem_ref.
4959 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
4960 * tree-ssa-live.h: Adjust prototypes.
4961 * passes.c: Include tree-ssa-live.h.
4962 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
4963 * graphite.c (graphite_transform_loops): Make static.
4964 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
4965 make_pass_graphite, pass_data_graphite_transforms,
4966 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
4967 * ipa-pure-const.c (warn_function_noreturn): Make static.
4968 (execute_warn_function_noreturn, gate_warn_function_noreturn,
4969 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
4970 Relocate from tree-cfg.c
4971 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
4973 (execute_warn_function_noreturn, gate_warn_function_noreturn,
4974 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
4975 Move to ipa-pure-const.c.
4976 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
4977 Relocate from tree-optimize.c.
4978 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
4979 make_pass_fixup_cfg): Move to tree-cfg.c.
4980 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
4981 Relocate some prototypes.
4982 * tree-data-ref.h (tree_check_data_deps) Add prototype.
4983 * tree-dump.c (dump_function_to_file): Remove prototype.
4984 Add tree-flow.h to the include file.
4985 * tree-dump.h: Remove prototype.
4986 * tree-parloops.h: New File. Add prototypes.
4987 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
4988 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
4989 from tree-ssa-loop.c.
4990 * tree-predcom.c (run_tree_predictive_commoning,
4991 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
4992 Relocate here from tree-ssa-loop.c.
4993 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
4994 ssa_name_values.release ().
4995 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
4996 (ssa_name_values): Relocate from tree-flow.h.
4997 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
4998 * tree-ssa-loop.c (run_tree_predictive_commoning,
4999 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
5000 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5001 make_pass_graphite, pass_data_graphite_transforms,
5002 make_pass_graphite_transforms, gate_tree_parallelize_loops,
5003 tree_parallelize_loops, pass_data_parallelize_loops,
5004 make_pass_parallelize_loops): Move to other files.
5005 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
5007 * tree.h: Remove prototypes from tree-address.c.
5009 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5011 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
5012 (struct int_tree_map): Move to tree-hasher.h
5013 (SCALE, LABEL, PERCENT): Move to gimple.h
5014 * tree-flow-inline.h: Delete. Move functions to other files.
5015 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
5016 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
5017 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
5018 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
5019 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
5020 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
5021 * tree-hasher.h: Don't include tree-flow.h.
5022 (struct int_tree_map): Relocate from tree-flow.h.
5023 * tree-sra.c (contains_view_convert_expr_p): Relocate from
5024 tree-flow-inline.h and make static.
5025 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
5027 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
5028 tree-flow-inline.h and make static.
5029 * tree.h (is_global_var, may_be_aliased): Relocate from
5031 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
5032 * value-prof.c: No longer include tree-flow-inline.h.
5033 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
5035 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5037 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
5038 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
5039 (enum move_pos): Move to tree-ssa-loop-im.h
5040 * cfgloop.h: Move some prototypes.
5041 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
5042 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
5043 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
5044 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
5045 (loop_containing_stmt): Relocate from tree-flow-inline.h.
5046 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
5047 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
5048 (enum move_pos): Relocate here.
5049 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
5051 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
5052 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
5053 make_pass_lim): Relocate here from tree-ssa-loop.c.
5054 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
5056 (loop_edge_to_cancel, unloop_loops): Make static.
5057 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5058 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
5059 (tree_complete_unroll, gate_tree_complete_unroll,
5060 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
5061 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5062 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
5063 * tree-ssa-loop-ivopts.c: Remove local prototypes.
5064 (stmt_invariant_in_loop_p): Remove unused function.
5065 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
5066 * tree-ssa-loop-manip.h: New file. Add prototypes.
5067 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
5068 (gcov_type_to_double_int): Move to cfgloop.h.
5069 (double_int_cmp, bound_index,
5070 estimate_numbers_of_iterations_loop): Make static.
5071 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
5072 (max_loop_iterations): Factor out get_max_loop_iterations.
5073 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
5075 * tree-ssa-loop-niter.h: New file. Add prototypes.
5076 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
5077 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
5078 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
5079 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
5080 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
5081 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
5082 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
5083 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
5084 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
5085 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
5086 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5087 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
5088 pass_data_complete_unroll, make_pass_complete_unroll,
5089 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5090 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
5091 tree-ssa-loop-ivcanon.c.
5092 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
5093 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
5094 tree-ssa-loop-prefetch.c.
5095 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
5097 (get_lsm_tmp_name): Relocate and add suffix parameter.
5098 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
5099 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
5100 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
5101 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
5102 (get_estimated_loop_iterations): Factor out accessor from
5103 estimated_loop_iterations in tree-ssa-loop-niter.c.
5104 (get_max_loop_iterations): Factor out accessor from
5105 _max_loop_iterations in tree-ssa-niter.c.
5106 * loop-unroll.c (decide_unroll_constant_iterations,
5107 decide_unroll_runtime_iterations, decide_peel_simple,
5108 decide_unroll_stupid): Use new get_* accessors.
5110 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
5112 PR tree-optimization/20318
5113 * doc/extend.texi (returns_nonnull): New function attribute.
5114 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
5116 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
5117 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
5119 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
5122 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
5123 false if both components are bitfields.
5125 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5127 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
5128 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
5129 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
5130 * config/aarch64/aarch64.h
5131 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
5132 * config/aarch64/aarch64-simd-builtins.def
5133 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
5135 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5137 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
5139 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5141 * config/aarch64/arm_neon.h (vdiv_f64): Added.
5143 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5145 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
5146 (vneg_f64): New intrinsic.
5147 (vneg_s8): Asm replaced with C.
5148 (vneg_s16): Likewise.
5149 (vneg_s32): Likewise.
5150 (vneg_s64): New intrinsic.
5151 (vnegq_f32): Asm replaced with C.
5152 (vnegq_f64): Likewise.
5153 (vnegq_s8): Likewise.
5154 (vnegq_s16): Likewise.
5155 (vnegq_s32): Likewise.
5156 (vnegq_s64): Likewise.
5158 2013-10-09 Renlin Li <Renlin.Li@arm.com>
5160 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
5162 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5164 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
5165 packed stack special handling.
5166 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
5167 back to fixed stack slots for FPRs saved due to stdarg.
5169 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5171 * config/s390/s390.c (s390_frame_info): Restructure function.
5173 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5175 * config/s390/s390.c (struct s390_frame_layout): New field
5177 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
5178 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
5179 regs_ever_clobbered to char*.
5180 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
5181 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
5182 (s390_register_info): Enable FPR save slots. Move/Copy some
5183 functionality into ...
5184 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
5185 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
5187 (s390_frame_info): Do gpr slot allocation here now. stdarg does
5188 not imply a stack frame.
5189 (s390_init_frame_layout): Remove variable clobbered_regs.
5190 (s390_update_register_info): Remove function.
5191 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
5192 cannot be used for register renaming.
5193 (s390_hard_regno_scratch_ok): New function.
5194 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
5195 (s390_initial_elimination_offset): Change offset calculation of
5196 the return address pointer.
5197 (save_gprs): Deal with only r6 being saved from the call-saved regs.
5198 (restore_gprs): Set frame related flag.
5199 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
5200 (s390_emit_prologue): Call s390_register_info instead of
5201 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
5202 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
5203 (s390_optimize_prologue): Call s390_optimize_register_info.
5204 Try to remove also FPR slot save/restore INSNs. Remove frame
5205 related flags from restore INSNs.
5207 2013-10-08 DJ Delorie <dj@redhat.com>
5209 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
5212 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
5213 avoid conflict with the MI use of %c.
5214 * config/rl78/rl78-real.md: change %c to %C throughout.
5215 * config/rl78/rl78-virt.md: Likewise.
5217 2013-10-08 Jan Hubicka <jh@suse.cz>
5219 * config/i386/i386.c (ix86_option_override_internal): Switch
5220 to SSE math for -ffast-math when target ISA supports SSE2.
5222 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
5224 * tree-flow.h: Remove some prototypes.
5225 * tree.h: Remove some protypes, add a couple.
5226 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
5227 using_eh_for_cleanups_p): Add interface routines for front ends.
5228 * tree-eh.h: New file. Add protoptyes.
5229 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
5230 (add_stmt_to_eh_lp_fn): Make static.
5231 (lower_try_finally): Use new using_eh_for_cleanups_p.
5232 * emit-rtl.c: Include tree-eh.h.
5233 * gimple.h: Include tree-eh.h.
5235 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
5237 PR tree-optimization/58480
5238 * tree-vrp.c (infer_nonnull_range): New function.
5239 (infer_value_range): Call infer_nonnull_range.
5241 2013-10-08 Dehao Chen <dehao@google.com>
5243 PR tree-optimization/58619
5244 * tree-inline.c (copy_phis_for_bb): Combine location data
5247 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5250 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
5251 RTX_FRAME_RELATED_P on INSN.
5253 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5255 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
5256 (altivec_expand_vec_perm_const): Call it.
5258 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5260 * config/rs6000/vector.md (mov<mode>): Emit permuted move
5261 sequences for LE VSX loads and stores at expand time.
5262 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
5264 * config/rs6000/rs6000.c (rs6000_const_vec): New.
5265 (rs6000_gen_le_vsx_permute): New.
5266 (rs6000_gen_le_vsx_load): New.
5267 (rs6000_gen_le_vsx_store): New.
5268 (rs6000_gen_le_vsx_move): New.
5269 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
5270 (*vsx_le_perm_load_v4si): New.
5271 (*vsx_le_perm_load_v8hi): New.
5272 (*vsx_le_perm_load_v16qi): New.
5273 (*vsx_le_perm_store_v2di): New.
5274 (*vsx_le_perm_store_v4si): New.
5275 (*vsx_le_perm_store_v8hi): New.
5276 (*vsx_le_perm_store_v16qi): New.
5277 (*vsx_xxpermdi2_le_<mode>): New.
5278 (*vsx_xxpermdi4_le_<mode>): New.
5279 (*vsx_xxpermdi8_le_V8HI): New.
5280 (*vsx_xxpermdi16_le_V16QI): New.
5281 (*vsx_lxvd2x2_le_<mode>): New.
5282 (*vsx_lxvd2x4_le_<mode>): New.
5283 (*vsx_lxvd2x8_le_V8HI): New.
5284 (*vsx_lxvd2x16_le_V16QI): New.
5285 (*vsx_stxvd2x2_le_<mode>): New.
5286 (*vsx_stxvd2x4_le_<mode>): New.
5287 (*vsx_stxvd2x8_le_V8HI): New.
5288 (*vsx_stxvd2x16_le_V16QI): New.
5290 2013-10-07 Renlin Li <Renlin.Li@arm.com>
5292 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
5294 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5296 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
5297 loop to work also for 31bit ABI.
5298 Save the stack pointer for frame_size > 0.
5300 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5302 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
5303 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
5304 constraint letters from expanders.
5305 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
5306 retry count to general_operand.
5307 ("tabort"): Give operand 0 a mode.
5308 ("tabort_1"): Add mode and constraint letter for operand 0.
5309 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
5311 2013-10-04 Jeff Law <law@redhat.com>
5313 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
5315 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
5317 (thread_across_edge): Here. Call it.
5319 2013-10-04 Cary Coutant <ccoutant@google.com>
5321 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
5322 discarding a location list expression (or a piece of one).
5324 2013-10-03 Jan Hubicka <jh@suse.cz>
5326 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
5327 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
5328 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
5330 2013-10-03 Jan Hubicka <jh@suse.cz>
5332 * config/i386/i386.c (ix86_option_override_internal): Do not enable
5333 accumulate-outgoing-args when producing unwind info.
5335 2013-10-03 Wei Mi <wmi@google.com>
5337 * lra-constraints.c (insert_move_for_subreg): New function
5338 extracted from simplify_operand_subreg.
5339 (simplify_operand_subreg): Add reload for paradoxical subreg.
5341 2013-10-03 Rong Xu <xur@google.com>
5343 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
5344 the candidate of builtin_expect such that we should fix the
5345 size/time estimation.
5346 (estimate_function_body_sizes): Do the acutally size/time fix-up
5349 2013-10-03 Rong Xu <xur@google.com>
5351 * predict.c (tree_predict_by_opcode): Get the probability
5352 for builtin_expect from param builtin_expect_probability.
5353 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
5354 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
5355 * doc/invoke.texi: Add documentation for builtin-expect-probability.
5357 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
5360 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
5361 * calls.c (alloca_call_p): Use get_callee_fndecl.
5362 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
5363 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
5365 (vrp_visit_stmt): Remove duplicated code.
5367 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
5369 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
5370 ceildf2, btruncdf2, instead of vsx_* name.
5372 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
5373 iterators to only do V2DF and V4SF here. Move the DF code to
5374 rs6000.md where it is combined with SF mode. Replace <VSv> with
5375 just 'v' since only vector operations are handled with these insns
5376 after moving the DF support to rs6000.md.
5377 (vsx_sub<mode>3): Likewise.
5378 (vsx_mul<mode>3): Likewise.
5379 (vsx_div<mode>3): Likewise.
5380 (vsx_fre<mode>2): Likewise.
5381 (vsx_neg<mode>2): Likewise.
5382 (vsx_abs<mode>2): Likewise.
5383 (vsx_nabs<mode>2): Likewise.
5384 (vsx_smax<mode>3): Likewise.
5385 (vsx_smin<mode>3): Likewise.
5386 (vsx_sqrt<mode>2): Likewise.
5387 (vsx_rsqrte<mode>2): Likewise.
5388 (vsx_fms<mode>4): Likewise.
5389 (vsx_nfma<mode>4): Likewise.
5390 (vsx_copysign<mode>3): Likewise.
5391 (vsx_btrunc<mode>2): Likewise.
5392 (vsx_floor<mode>2): Likewise.
5393 (vsx_ceil<mode>2): Likewise.
5394 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
5395 (vsx_sminsf3): Likewise.
5396 (vsx_fmadf4): Likewise.
5397 (vsx_fmsdf4): Likewise.
5398 (vsx_nfmadf4): Likewise.
5399 (vsx_nfmsdf4): Likewise.
5400 (vsx_cmpdf_internal1): Likewise.
5402 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
5403 simpler to select whether a target has SPE or traditional floating
5404 point support in iterators.
5405 (TARGET_DF_SPE): Likewise.
5406 (TARGET_SF_FPR): Likewise.
5407 (TARGET_DF_FPR): Likewise.
5408 (TARGET_SF_INSN): Macros to say whether floating point support
5409 exists for a given operation for expanders.
5410 (TARGET_DF_INSN): Likewise.
5412 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
5413 combining of SF/DF mode operations, using both traditional and VSX
5421 (abs<mode>2): Combine SF/DF modes using traditional floating point
5422 instructions. Add support for using the upper DF registers with
5423 VSX support, and SF registers with power8-vector support. Update
5424 expanders for operations supported by both the SPE and traditional
5425 floating point units.
5426 (abs<mode>2_fpr): Likewise.
5427 (nabs<mode>2): Likewise.
5428 (nabs<mode>2_fpr): Likewise.
5429 (neg<mode>2): Likewise.
5430 (neg<mode>2_fpr): Likewise.
5431 (add<mode>3): Likewise.
5432 (add<mode>3_fpr): Likewise.
5433 (sub<mode>3): Likewise.
5434 (sub<mode>3_fpr): Likewise.
5435 (mul<mode>3): Likewise.
5436 (mul<mode>3_fpr): Likewise.
5437 (div<mode>3): Likewise.
5438 (div<mode>3_fpr): Likewise.
5439 (sqrt<mode>3): Likewise.
5440 (sqrt<mode>3_fpr): Likewise.
5441 (fre<Fs>): Likewise.
5442 (rsqrt<mode>2): Likewise.
5443 (cmp<mode>_fpr): Likewise.
5444 (smax<mode>3): Likewise.
5445 (smin<mode>3): Likewise.
5446 (smax<mode>3_vsx): Likewise.
5447 (smin<mode>3_vsx): Likewise.
5448 (negsf2): Delete SF operations that are merged with DF.
5455 (fmasf4_fpr): Likewise.
5456 (fmssf4_fpr): Likewise.
5457 (nfmasf4_fpr): Likewise.
5458 (nfmssf4_fpr): Likewise.
5459 (sqrtsf2): Likewise.
5460 (rsqrtsf_internal1): Likewise.
5461 (smaxsf3): Likewise.
5462 (sminsf3): Likewise.
5463 (cmpsf_internal1): Likewise.
5464 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
5465 (negdf2): Delete DF operations that are merged with SF.
5467 (nabsdf2): Likewise.
5473 (rsqrtdf_internal1): Likewise.
5474 (fmadf4_fpr): Likewise.
5475 (fmsdf4_fpr): Likewise.
5476 (nfmadf4_fpr): Likewise.
5477 (nfmsdf4_fpr): Likewise.
5478 (sqrtdf2): Likewise.
5479 (smaxdf3): Likewise.
5480 (smindf3): Likewise.
5481 (cmpdf_internal1): Likewise.
5482 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
5483 (btrunc<mode>2): Delete separate expander, and combine with the
5484 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
5485 (btrunc<mode>2_fpr): Likewise.
5486 (ceil<mode>2): Likewise.
5487 (ceil<mode>2_fpr): Likewise.
5488 (floor<mode>2): Likewise.
5489 (floor<mode>2_fpr): Likewise.
5490 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
5491 Add support for using the upper registers with VSX and
5492 power8-vector. Move insns to be closer to the define_expands. On
5493 VSX systems, prefer the traditional form of FMA over the VSX
5494 version, since the traditional form allows the target not to
5495 overlap with the inputs.
5496 (fms<mode>4_fpr): Likewise.
5497 (nfma<mode>4_fpr): Likewise.
5498 (nfms<mode>4_fpr): Likewise.
5500 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5501 Richard Earnshaw <richard.earnshaw@arm.com>
5503 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
5504 (struct mult_cost_table): Likewise.
5505 (struct mem_cost_table): Likewise.
5506 (struct fp_cost_table): Likewise.
5507 (struct vector_cost_table): Likewise.
5508 (cpu_cost_table): Likewise.
5509 * config/arm/arm.opt (mold-rts-costs): New option.
5510 (mnew-generic-costs): Likewise.
5511 * config/arm/arm.c (generic_extra_costs): New table.
5512 (cortexa15_extra_costs): Likewise.
5513 (arm_slowmul_tune): Use NULL as new costs.
5514 (arm_fastmul_tune): Likewise.
5515 (arm_strongarm_tune): Likewise.
5516 (arm_xscale_tune): Likewise.
5517 (arm_9e_tune): Likewise.
5518 (arm_v6t2_tune): Likewise.
5519 (arm_cortex_a5_tune): Likewise.
5520 (arm_cortex_a9_tune): Likewise.
5521 (arm_v6m_tune): Likewise.
5522 (arm_fa726te_tune): Likewise.
5523 (arm_cortex_a15_tune): Use cortex15_extra_costs.
5524 (arm_cortex_tune): Use generict_extra_costs.
5525 (shifter_op_p): New function.
5526 (arm_unspec_cost): Likewise.
5527 (LIBCALL_COST): Define.
5528 (arm_new_rtx_costs): New function.
5529 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
5530 table is available. Use old costs otherwise unless mnew-generic-costs
5532 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
5533 (cpu_cost_table): Declare.
5535 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
5538 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
5539 (*subs_mul_imm_<mode>)
5540 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
5541 (*sub_<shift>_<mode>)
5542 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
5543 Remove k constraint.
5545 2013-10-03 Ian Bolton <ian.bolton@arm.com>
5547 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
5549 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
5551 2013-10-02 Teresa Johnson <tejohnson@google.com>
5553 * predict.c (probably_never_executed): New function.
5554 (probably_never_executed_bb_p): Invoke probably_never_executed.
5555 (probably_never_executed_edge_p): Ditto.
5556 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
5557 Treat profile insanities conservatively.
5559 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
5561 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
5563 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
5565 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
5566 use it. Use smaller increase for scratch. Don't increase reject
5567 for early clobber scratch.
5568 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
5569 setting eliminated regs except setting fp from hfp.
5570 (lra_eliminate): Check lra_insn_recog_data on NULL.
5572 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5575 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
5576 setting -mvsx-timode by default until the underlying problem is fixed.
5577 (RS6000_CPU, power7 defaults): Likewise.
5579 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5581 * config/x-linux (host-linux.o): Remove header dependencies.
5582 Use $(COMPILE) and $(POSTCOMPILE).
5583 * config/t-linux-android (linux-android.o): Ditto.
5585 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5587 * Makefile.in (expmed.o-warn): Remove.
5589 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5591 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
5592 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
5594 2013-10-02 Teresa Johnson <tejohnson@google.com>
5596 * dojump.c (do_jump_1): Divide probability between
5597 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
5599 2013-10-02 Tom Tromey <tromey@redhat.com>
5601 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
5603 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5605 * tree-flow.h: Remove some prototypes.
5606 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
5607 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
5608 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
5609 mark_virtual_phi_result_for_renaming): Relocate here.
5610 * tree-into-ssa.h: Add prototypes.
5611 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
5612 single_pred_before_succ_order.
5613 (blocks_in_phiopt_order): Rename and move to cfganal.c.
5614 (nonfreeing_call_p) Move to gimple.c.
5615 * cfganal.c (single_pred_before_succ_order): Move and renamed from
5617 * basic-block.h (single_pred_before_succ_order): Add prototype.
5618 * gimple.c (nonfreeing_call_p): Relocate here.
5619 * gimple.h: Add prototype.
5620 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
5621 * tree-ssa-dom.h: New file. Relocate prototypes here.
5622 * tree-ssa.h: Include tree-ssa-dom.h.
5624 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5626 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
5627 Use $(COMPILE) and $(POSTCOMPILE).
5629 * config/alpha/x-alpha (driver-alpha.o): Ditto.
5631 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5633 * tree-flow.h: Remove some prototypes.
5634 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
5635 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5636 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
5637 propagate_tree_value*): Move from here to...
5638 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
5639 propagate_tree_value*): Relocate here.
5640 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5641 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
5642 * gimple-fold.c: Remove gimple-fold.h from include list.
5643 * tree-vrp.c: Remove gimple-fold.h from include list.
5644 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
5645 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
5646 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
5647 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
5648 * sese.c: Add tree-ssa-propagate.h to include list.
5650 2013-10-02 Richard Biener <rguenther@suse.de>
5652 * tree-loop-distribution.c: Include tree-vectorizer.h for
5654 (enum partition_kind): Remove PKIND_REDUCTION.
5655 (struct partition_s): Remove has_writes member, add reduction_p member.
5656 (partition_alloc): Adjust.
5657 (partition_builtin_p): Likewise.
5658 (partition_has_writes): Remove.
5659 (partition_reduction_p): New function.
5660 (partition_merge_into): Likewise.
5661 (generate_code_for_partition): Commonize builtin partition
5663 (rdg_cannot_recompute_vertex_p): Remove.
5664 (already_processed_vertex_p): Likewise.
5665 (rdg_flag_vertex): Do not set has_writes.
5666 (classify_partition): Adjust.
5667 (rdg_build_partitions): Do not set has_writes, treat all
5668 partitions as useful.
5669 (distribute_loop): Record number of library calls generated. Adjust.
5670 (tree_loop_distribution): Report number of loops and library
5671 calls generated as opt-info.
5673 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5675 * tree-flow.h: Include new .h files. Move prototypes.
5676 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
5677 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
5678 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
5679 * tree-pretty-print.h: Add prototypes from tree-flow.h.
5680 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
5681 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
5682 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
5683 * tree.h (get_ref_base_and_extent): Move prototype out.
5684 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
5686 * gimple-low.h: New File. Add prototypes from tree-flow.h.
5687 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
5688 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
5689 * tree-scalar-evolution.c: Include tree.h.
5690 * sese.c: Include tree.h.
5691 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
5692 * dwarf2out.c: Include tree-dfa.h.
5693 * tree-chrec.c: Include tree.h.
5694 * tree-data-ref.c: Include tree.h.
5696 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
5698 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
5701 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5703 * config/t-sol2 (sol2-c.o): Remove header dependencies.
5704 Use $(COMPILE) and $(POSTCOMPILE).
5705 (sol2-cxx.o): Likewise.
5706 (sol2-stubs.o): Likewise.
5708 * config/x-solaris (host-solaris.o): Likewise.
5710 * config/sparc/t-sparc (sparc.o): Remove.
5711 (sparc-c.o): Remove header dependencies.
5712 Use $(COMPILE) and $(POSTCOMPILE).
5713 * config/sparc/x-sparc: Likewise.
5715 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
5717 * config/arc/arc-opts.h: Add 2013 to Copyright years.
5718 * config/arc/arc700.md: Likewise.
5719 * config/arc/arc-modes.def: Likewise.
5720 * config/arc/arc-simd.h: Likewise.
5721 * config/arc/t-arc-uClibc: Likewise.
5722 * config/arc/t-arc-newlib: Likewise.
5724 2013-10-02 Renlin Li <renlin.li@arm.com>
5726 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
5728 (aarch64_expand_epilogue): Likewise.
5730 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5731 Yufeng Zhang <yufeng.zhang@arm.com>
5733 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
5735 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
5736 'base_in' represent a conversion and legal_cast_p_1 holds; set
5737 'base_in' with the returned value from get_unwidened.
5739 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5741 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
5742 plus_constant is not used.
5744 2013-10-01 Wei Mi <wmi@google.com>
5746 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
5747 * config/i386/i386.md: Add define_peephole2 to
5748 break partial reg stall for cvtss2sd/cvtsd2ss.
5750 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5752 * config/arc/arc.c (pass_arc_ifcvt::clone):
5753 Update for ctxt_ -> m_ctxt change.
5755 2013-10-01 Jeff Law <law@redhat.com>
5757 * tree-ssa-threadupdate.c (struct redirection_data): Delete
5758 outgoing_edge and intermediate_edge fields. Instead store the path.
5759 (redirection_data::hash): Hash on the last edge's destination index.
5760 (redirection_data::equal): Check the entire thread path.
5761 (lookup_redirectio_data): Corresponding changes.
5762 (create_edge_and_update_destination_phis): Likewise.
5763 (thread_single_edge): Likewise.
5765 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5766 Diego Novillo <dnovillo@google.com>
5768 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
5769 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
5770 (vld32wh_insn, vld32wl_insn): Delete commented-out old
5771 versions of these patterns.
5773 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
5774 (__builtin_arc_aligned): Likewise.
5776 * config/arc/arc.md: Expand adc_0 comment stating the intended
5777 purpose and why it isn't ready.
5778 Replace commented out call_value_via_label_mixed with a
5779 plain comment about bl_s.
5781 * config/arc/arc.c (stdio.h): Don't include directly.
5782 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
5783 Remove if (1) condition.
5784 (arc_encode_section_info): Fix comment.
5786 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5788 * config/arc/arc.c (arc_conditional_register_usage):
5789 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
5790 Also set reg_alloc_order for DMA config regs.
5792 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5793 Jeremy Bennett <jeremy.bennett@embecosm.com>
5795 * doc/install.texi (--with-cpu): Mention ARC.
5796 (arc-*-elf32): New paragraph.
5797 (arc-linux-uclibc): Likewise.
5798 * doc/md.texi (Machine Constraints): Add ARC part.
5799 * doc/invoke.texi: (menu): Add ARC Options.
5800 (Machine Dependent Options) <ARC Options>: Add synopsis.
5801 (node ARC Options): Add.
5802 * doc/extend.texi (long_call / short_call attribute): Add ARC.
5803 (ARC Built-in Functions): New section defining
5804 generic ARC built-in functions.
5805 (ARC SIMD Built-in Functions): New section defining SIMD specific
5807 (Declaring Attributes of Functions): Extended
5808 description of short_call and long_call attributes for ARC and
5809 added index entries.
5811 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
5812 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
5813 Joern Rennecke <joern.rennecke@embecosm.com>
5814 Muhammad Khurram Riaz <khurram.riaz@arc.com>
5815 Brendan Kehoe <brendan@zen.org>
5816 Michael Eager <eager@eagercon.com>
5817 Simon Cook <simon.cook@embecosm.com>
5818 Jeremy Bennett <jeremy.bennett@embecosm.com>
5820 * config/arc, common/config/arc: New directories.
5822 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5823 Brendan Kehoe <brendan@zen.org>
5824 Simon Cook <simon.cook@embecosm.com>
5826 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
5828 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
5830 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
5831 * tree-ssa-coalesce.h: New. Move prototype to here.
5832 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
5833 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
5834 (gimple_can_coalesce_p): Move to...
5835 * gimple.c (gimple_can_coalesce_p): Here.
5837 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
5839 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
5840 (dump_decl_set): Move to gimple.c.
5841 * gimple.h: Don't include tree-ssa-operands.h.
5842 (dump_decl_set): Add prototype.
5843 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
5844 Move to gimple-ssa.h.
5845 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
5846 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
5847 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
5848 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
5849 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
5851 (dump_decl_set): Relocate here.
5852 * gimple-ssa.h: New file.
5853 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
5854 Relocate from gimple.h.
5855 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
5856 * tree-ssa-operands.c (swap_ssa_operands): Rename from
5857 swap_tree_operands and remove non-ssa path.
5858 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
5859 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
5861 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
5862 vect_is_simple_reduction_1): Use swap_ssa_operands.
5863 * tree-flow.h: Move various prototypes to tree-phinodes.h.
5864 (enum need_phi_state): Move to tree-into-ssa.c.
5865 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
5866 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
5867 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
5868 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
5869 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
5870 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
5871 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
5872 num_imm_uses): Move to ssa-iterators.h.
5873 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
5874 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
5876 (op_iter_done, op_iter_next_def, op_iter_next_tree,
5877 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
5878 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
5879 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
5880 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
5881 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
5882 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
5883 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
5884 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
5885 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
5886 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
5887 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
5888 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
5889 (set_phi_nodes): Move to tree-phinodes.h.
5890 * tree-ssa-operands.h (enum ssa_op_iter_type,
5891 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
5892 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
5893 (dump_decl_set): Remove prototype.
5894 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
5895 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
5896 (set_phi_nodes): Relocate from tree-flow-inline.h.
5897 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
5899 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
5900 include list. Temporarily add gimple.h to include list.
5901 * ssa-iterators.h: New file.
5902 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
5903 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
5904 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
5905 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
5906 Relocate from tree-ssa-operands.h.
5907 (delink_imm_use, link_imm_use_to_list, link_imm_use,
5908 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
5909 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
5910 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
5911 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
5912 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
5913 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
5914 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
5915 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
5916 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
5917 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
5918 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
5919 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
5920 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
5921 Relocate from tree-flow-inline.h.
5922 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
5924 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
5927 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
5928 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
5929 Insert '\t' to output template.
5930 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
5931 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
5932 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
5933 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
5935 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
5937 * doc/install.texi (Host/target specific installation notes for GCC):
5938 Put @anchor before @heading.
5939 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
5940 Use @email for email addresses.
5942 2013-10-01 Jeff Law <law@redhat.com>
5944 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
5945 a vec. Only delete the path if we create one without successfully
5946 registering a jump thread.
5947 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
5949 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
5950 (paths): New vector of jump threading paths.
5951 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
5952 (THREAD_PATH): New accessor macro for the entire thread path.
5953 (lookup_redirection_data): Get intermediate and final outgoing edge
5954 from the thread path.
5955 (create_edge_and_update_destination_phis): Copy the threading path.
5956 (ssa_fix_duplicate_block_edges): Get edges and block types from the
5957 jump threading path.
5958 (ssa_redirect_edges): Get edges and block types from the jump threading
5959 path. Free the path vector.
5960 (thread_block): Get edges from the jump threading path. Look at the
5961 entire path to see if we thread to a loop exit. If we cancel a jump
5962 thread request, then free the path vector.
5963 (thread_single_edge): Get edges and block types from the jump threading
5964 path. Free the path vector.
5965 (thread_through_loop_header): Get edges and block types from the jump
5966 threading path. Free the path vector.
5967 (mark_threaded_blocks): Iterate over the vector of paths and store
5968 the path on the appropriate edge. Get edges and block types from the
5969 jump threading path.
5970 (mark_threaded_blocks): Get edges and block types from the jump
5971 threading path. Free the path vector.
5972 (thread_through_all_blocks): Use the vector of paths rather than
5973 a vector of 3-edge sets.
5974 (register_jump_thread): Accept pointer to a path vector rather
5975 than the path vector itself. Store the path vector for later use.
5978 2013-10-01 Jakub Jelinek <jakub@redhat.com>
5979 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5982 * config/s390/s390.c (s390_split_branches): Modify check for table
5984 (s390_chunkify_start): Rearrange table jump insn check in order to
5985 deal with compare and branch insns correctly.
5987 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
5991 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
5992 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
5993 define_insn_and_split.
5994 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
5995 (shiftsi3_compare): New pattern.
5997 * config/arm/unspecs.md (UNSPEC_RRX): New.
5999 2013-10-01 Alan Modra <amodra@gmail.com>
6001 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
6002 casing inout operands.
6004 2013-10-01 Richard Biener <rguenther@suse.de>
6006 PR tree-optimization/58553
6007 * tree-loop-distribution.c (struct partition_s): Add niter member.
6008 (classify_partition): Populate niter member for the partition
6009 and properly identify whether the relevant store happens before
6010 or after the loop exit.
6011 (generate_memset_builtin): Use niter member from the partition.
6012 (generate_memcpy_builtin): Likewise.
6014 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6016 * vec.h (vec_prefix, vec): Prefix member names with "m_".
6017 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
6019 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6021 * basic-block.h (edge_list): Prefix member names with "m_".
6022 * context.h (context): Likewise.
6023 * domwalk.h (dom_walker): Likewise.
6024 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
6025 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
6026 * hash-table.h (hash_table): Likewise.
6027 * machmode.h (bit_field_mode_iterator): Likewise.
6028 * pass_manager.h (pass_list): Likewise.
6029 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
6030 * tree-pass.h (pass_data): Likewise.
6031 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
6032 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
6033 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
6034 * asan.c (pass_data_asan): Update accordingly.
6035 * cfganal.c (control_dependences::find_control_dependence): Likewise.
6036 (control_dependences::control_dependences): Likewise.
6037 (control_dependences::~control_dependences): Likewise.
6038 (control_dependences::~control_dependences): Likewise.
6039 (control_dependences::get_edges_dependent_on): Likewise.
6040 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
6041 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
6042 * context.c (gcc::context::context): Likewise.
6043 * cprop.c (pass_rtl_cprop::clone): Likewise.
6044 * domwalk.c (dom_walker::walk): Likewise.
6045 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
6046 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
6047 * mode-switching.c (pass_mode_switching::clone): Likewise.
6048 * passes.c (opt_pass::opt_pass): Likewise.
6049 (pass_manager::pass_manager): Likewise.
6050 * predict.c (pass_strip_predict_hints::clone): Likewise.
6051 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
6052 (pass_split_all_insns::clone): Likewise.
6053 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
6055 (bit_field_mode_iterator::next_mode): Likewise.
6056 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
6057 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
6058 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
6059 * tree-complex.c (pass_lower_complex::clone): Likewise.
6060 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
6061 * tree-object-size.c (pass_object_sizes::clone): Likewise.
6062 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
6063 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
6064 (pass_fold_builtins::clone): Likewise.
6065 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
6066 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
6067 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
6068 (pass_cd_dce::clone): Likewise.
6069 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
6070 (pass_phi_only_cprop::clone): Likewise.
6071 * tree-ssa-dse.c (pass_dse::clone): Likewise.
6072 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
6073 * tree-ssa-loop.c (pass_lim::clone): Likewise.
6074 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
6075 * tree-ssa-pre.c (pass_fre::clone): Likewise.
6076 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
6077 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
6078 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
6079 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
6080 * tree-vrp.c (pass_vrp::clone): Likewise.
6081 * tsan.c (pass_tsan::clone): Likewise.
6083 2013-09-30 Jakub Jelinek <jakub@redhat.com>
6086 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
6087 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
6090 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
6091 optimization, punt if sign_bit_p looked through any zero extension.
6093 2013-09-30 Teresa Johnson <tejohnson@google.com>
6095 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
6096 Update redirected out edge count in joiner case.
6097 (ssa_redirect_edges): Common the joiner and non-joiner cases
6098 so that joiner case gets profile updates.
6100 2013-09-30 Richard Biener <rguenther@suse.de>
6102 PR tree-optimization/58554
6103 * tree-loop-distribution.c (classify_partition): Require
6104 unconditionally executed stores for memcpy and memset recognition.
6105 (tree_loop_distribution): Calculate dominance info.
6107 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
6109 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
6110 (NO_PROFILE_COUNTERS): Likewise.
6111 (PROFILE_HOOK): Likewise.
6112 (FUNCTION_PROFILER): Likewise.
6113 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
6115 2013-09-30 Iain Sandoe <iain@codesourcery.com>
6117 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
6118 calls and defines in TARGET_MACHO conditional.
6119 (load_macho_picbase_di): Likewise.
6120 (reload_macho_picbase): Likewise.
6121 (reload_macho_picbase_si): Likewise.
6122 (reload_macho_picbase_di): Likewise.
6123 (nonlocal_goto_receiver): Likewise.
6125 2013-09-30 Nick Clifton <nickc@redhat.com>
6127 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
6128 that use the MSP430X ISA.
6129 (msp430_option_override): Scan -mmcu command line option for any
6130 MCU name that supports the MSP430X ISA.
6131 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
6132 -mmcu options which enable the MSP430X ISA.
6134 2013-09-30 Richard Biener <rguenther@suse.de>
6137 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
6138 before looking for setjmp-like calls.
6140 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6143 * config/darwin-protos.h (machopic_get_function_picbase): New.
6144 * config/darwin.c (machopic_get_function_picbase): New.
6145 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
6146 label for a new func. (load_macho_picbase_di): Likewise.
6147 (reload_macho_picbase): New expand.
6148 (reload_macho_picbase_si): New insn.
6149 (reload_macho_picbase_di): New insn.
6150 (nonlocal_goto_receiver): New define and split.
6151 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
6152 (unspecv enum): Add UNSPECV_NLGR.
6154 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6156 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
6157 that altivec registers are correctly sized on Darwin.
6159 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6161 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
6162 darwin-driver.o): Use COMPILE and POSTCOMPILE.
6163 * config/x-darwin (host-darwin.o): Likewise.
6164 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
6165 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
6166 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
6168 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
6170 * doc/invoke.texi: Fix usage of @tie{} command.
6172 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
6174 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
6176 2013-09-28 Jan Hubicka <jh@suse.cz>
6178 * config/i386/x86-tune.def: Add documentation for each of the options;
6181 2013-09-28 Jan Hubicka <jh@suse.cz>
6183 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
6185 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
6186 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
6187 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
6188 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
6189 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
6191 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
6193 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
6194 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
6195 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
6196 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
6197 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
6198 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
6199 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
6200 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
6201 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
6202 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
6203 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
6204 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
6205 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
6206 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
6207 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
6208 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
6209 gimple.h, godump.c, graphite-clast-to-gimple.c,
6210 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
6211 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
6212 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
6213 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6214 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
6215 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
6216 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
6217 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
6218 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
6219 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
6220 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
6221 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
6222 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
6223 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
6224 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
6225 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
6226 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
6227 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
6228 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
6229 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
6230 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
6231 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
6232 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
6233 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
6234 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6235 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6236 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
6237 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
6238 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
6239 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6240 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6241 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
6242 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
6243 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6244 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6245 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
6246 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
6247 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
6248 whitespace before "(".
6250 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
6252 * expr.h (extract_bit_field): Remove packedp parameter.
6253 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
6254 from forward declaration.
6255 (store_split_bit_field): Remove packedp arg from calls to
6256 extract_fixed_bit_field.
6257 (extract_bit_field_1): Remove packedp parameter and packedp
6258 argument from recursive calls and calls to extract_fixed_bit_field.
6259 (extract_bit_field): Remove packedp parameter and corresponding
6260 arg to extract_bit_field_1.
6261 (extract_fixed_bit_field): Remove packedp parameter. Remove code
6263 (extract_split_bit_field): Remove packedp arg from call to
6264 extract_fixed_bit_field.
6265 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
6266 (copy_blkmode_from_reg): Likewise.
6267 (copy_blkmode_to_reg): Likewise.
6268 (read_complex_part): Likewise.
6269 (store_field): Likewise.
6270 (expand_expr_real_1): Likewise.
6271 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
6272 to extract_bit_field.
6273 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
6274 call to extract_bit_field.
6275 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
6276 call to extract_bit_field.
6277 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
6278 and special packedp behavior from -fstrict-volatile-bitfields
6281 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6283 * lra-eliminations.c (init_elim_table): Guard value_p.
6285 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6287 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
6288 DFmode, DImode, and SFmode in the upper VSX registers based on the
6289 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
6290 if -mpower8-vector. Combine -mvsx-timode handling with the rest
6291 of the VSX register handling.
6293 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
6295 (zero_extendsidi2_lfiwzx): Add support for loading into the
6296 Altivec registers with -mpower8-vector. Use wu/wv constraints to
6297 only do VSX memory options on Altivec registers.
6298 (extendsidi2_lfiwax): Likewise.
6299 (extendsfdf2_fpr): Likewise.
6300 (mov<mode>_hardfloat, SF/SD modes): Likewise.
6301 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
6302 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
6303 (movdi_internal64): Likewise.
6305 2013-09-27 Xinliang David Li <davidxl@google.com>
6307 * opts.c (finish_options): Adjust parameters
6308 according to vect cost model.
6309 (common_handle_option): Set dynamic vect cost
6311 targhooks.c (default_add_stmt_cost): Compute stmt cost
6313 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
6314 Use helper function.
6315 * tree-vectorizer.h (unlimited_cost_model): New function.
6316 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
6317 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
6319 (vect_enhance_data_refs_alignment): Ditto.
6320 * flag-types.h: New enum.
6321 * common/config/i386/i386-common.c (ix86_option_init_struct):
6322 No need to initialize vect_cost_model flag.
6323 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
6326 2013-09-27 Diego Novillo <dnovillo@google.com>
6328 * gimple.h (enum ssa_mode): Remove.
6330 2013-09-27 Paulo Matos <pmatos@broadcom.com>
6332 * cfgloop.h (number_of_loops): Fix typo in check for null.
6334 2013-09-27 Jakub Jelinek <jakub@redhat.com>
6337 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
6338 are children of outermost saved_cfun's loop, and set it up to
6339 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
6340 if loop != loop0 and SESE region contains bbs that belong to loop0.
6342 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6344 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
6345 (binary_scale_code_p, get_base_term, get_index_term): New functions.
6346 (set_address_segment, set_address_base, set_address_index)
6347 (set_address_disp): Accept the argument unconditionally.
6348 (baseness): Remove must_be_base_p and must_be_index_p checks.
6349 (decompose_normal_address): Classify as much as possible in the
6352 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6354 * cse.c (count_reg_usage): Handle INT_LIST.
6355 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
6356 * reginfo.c (reg_scan_mark_refs): Likewise.
6357 * reload1.c (eliminate_regs_1): Likewise.
6359 2013-09-27 Iain Sandoe <iain@codesourcery.com>
6362 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
6365 2013-09-27 Richard Biener <rguenther@suse.de>
6367 PR tree-optimization/58459
6368 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
6369 restriction not propagating into loops.
6371 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
6373 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
6374 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
6375 * doc/tree-ssa.texi (Walking use-def chains): Delete.
6377 2013-09-26 Richard Biener <rguenther@suse.de>
6379 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
6381 2013-09-26 Richard Biener <rguenther@suse.de>
6383 * alias.h (component_uses_parent_alias_set): Rename to ...
6384 (component_uses_parent_alias_set_from): ... this.
6385 * alias.c (component_uses_parent_alias_set): Rename to ...
6386 (component_uses_parent_alias_set_from): ... this and return
6388 (reference_alias_ptr_type_1): Use the result from
6389 component_uses_parent_alias_set_from instead of stripping
6390 components one at a time.
6391 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
6393 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6395 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
6396 Move prototypes to...
6397 * tree-ssa-ter.h: New File. Move prototypes here.
6398 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
6399 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
6400 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
6401 from is_replaceable_p.
6402 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
6403 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
6404 (process_replaceable): Use ter_is_replaceable_p.
6405 (find_replaceable_in_bb): Use ter_is_replaceable_p.
6406 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
6407 newly refactored ssa_is_replaceable_p.
6408 * cfgexpand.c: Include tree-outof-ssa.h.
6409 * ssaexpand.h: Delete.
6411 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6413 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
6414 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
6416 (create_gimple_tmp): Delete.
6417 (get_expr_type, build_assign, build_type_cast): Move to...
6418 * gimple-builder.c: New File.
6419 (get_expr_type): Relocate from gimple.c.
6420 (build_assign, build_type_cast): Change to only create ssanames.
6421 * gimple.h: Move prototypes to...
6422 * gimple-builder.h: New File. Here.
6423 * tree-ssa.h: And here.
6424 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
6425 count_uses_and_derefs): Relocate from gimple.c.
6426 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
6427 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
6428 * tree-ssa-math-opts (execute_cse_reciprocals): Use
6429 gimple_replace_ssa_lhs.
6430 * asan.c: Include gimple-builder.h.
6431 * Makefile.in: Add gimple-builder.o.
6433 2013-09-26 Richard Biener <rguenther@suse.de>
6435 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
6436 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
6437 (loe_visit_block): Use gcc_checking_assert.
6438 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
6439 gimple_assign_ssa_name_copy_p.
6440 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
6443 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
6445 * config/rs6000/t-rs6000 (rs6000.o): Remove.
6446 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
6448 2013-09-26 Richard Biener <rguenther@suse.de>
6450 PR tree-optimization/58539
6451 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
6452 the fact that debug statements are not taking part in loop-closed
6455 2013-09-26 Nick Clifton <nickc@redhat.com>
6457 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
6458 time warning message.
6459 (msp430_print_operand_raw): Delete unused letter parameter.
6460 (TARGET_PRINT_OPERAND_ADDRESS): Define.
6461 (msp430_print_operand_address): New function.
6462 (msp430_print_operand): Move address printing code from here to
6464 * config/msp430/msp430.md (movsipsi2): Add comment in generated
6466 (zero_extendpsisi2): Likewise.
6467 (extendpsisi2): New pattern.
6468 (andneghi3): New pattern.
6470 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
6472 * config/aarch64/aarch64.opt (mlra): New option.
6473 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
6474 (TARGET_LRA_P): Define.
6476 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
6478 * expr.c (expand_assignment): Remove obsolete comment.
6480 2013-09-25 Jeff Law <law@redhat.com>
6482 * tree-flow.h (thread_through_all_blocks): Prototype moved into
6483 tree-ssa-threadupdate.h.
6484 (register_jump_thread): Similarly.
6485 * tree-ssa-threadupdate.h: New header file.
6486 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
6487 * tree-vrp.c: Likewise.
6488 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
6489 (thread_around_empty_blocks): Change type of path vector argument to
6490 an edge,type pair from just an edge. Initialize both elements when
6491 appending to a jump threading path. Tweak references to elements
6493 (thread_across_edge): Similarly. Release memory for the elements
6495 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
6496 (dump_jump_thread_path): New function broken out from
6497 register_jump_thread.
6498 (register_jump_thread): Use dump_jump_thread_path. Change type of
6499 path vector entries. Search the path for NULL edges and dump
6500 the path if one is found. Tweak the conversion of path to 3-edge
6501 form to use the block copy type information embedded in the path.
6503 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6505 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
6507 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6508 Vladimir Makarov <vmakarov@redhat.com>
6510 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
6511 from the least significant bit.
6512 (strip_address_mutations): Add bitfield operations handling.
6513 (must_be_index_p): Add shifting and rotate operations handling.
6514 (set_address_base): Use must_be_base_p predicate.
6515 (set_address_index): Use must_be_index_p predicate.
6517 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
6518 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6519 Sergey Lega <sergey.s.lega@intel.com>
6520 Anna Tikhonova <anna.tikhonova@intel.com>
6521 Ilya Tocar <ilya.tocar@intel.com>
6522 Andrey Turetskiy <andrey.turetskiy@intel.com>
6523 Ilya Verbin <ilya.verbin@intel.com>
6524 Kirill Yukhin <kirill.yukhin@intel.com>
6525 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6527 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
6529 (ix86_expand_vector_move_misalign): Support new unaligned load and
6530 stores and use new names.
6531 (CODE_FOR_sse2_storedqu): Rename to ...
6532 (CODE_FOR_sse2_storedquv16qi): ... this.
6533 (CODE_FOR_sse2_loaddqu): Rename to ...
6534 (CODE_FOR_sse2_loaddquv16qi): ... this.
6535 (CODE_FOR_avx_loaddqu256): Rename to ...
6536 (CODE_FOR_avx_loaddquv32qi): ... this.
6537 (CODE_FOR_avx_storedqu256): Rename to ...
6538 (CODE_FOR_avx_storedquv32qi): ... this.
6539 * config/i386/i386.md (fpint_logic): New.
6540 * config/i386/sse.md (VMOVE): Extend for AVX512.
6544 (VI_UNALIGNED_LOADSTORE): Ditto.
6545 (sse2_avx_avx512f): Ditto.
6546 (sse2_avx2): Extend for AVX512.
6547 (sse4_1_avx2): Ditto.
6548 (avx2_avx512f): New.
6549 (sse): Extend for AVX512.
6552 (avxsizesuffix): Ditto.
6553 (sseintvecmode): Ditto.
6555 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
6556 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
6557 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6558 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
6559 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6560 (<sse2_avx_avx512f>_storedqu<mode): ... this.
6561 (<sse>_movnt<mode>): Replace constraint "x" with "v".
6562 (STORENT_MODE): Extend for AVX512.
6563 (*absneg<mode>2): Replace constraint "x" with "v".
6564 (*mul<mode>3): Ditto.
6565 (*ieee_smin<mode>3): Ditto.
6566 (*ieee_smax<mode>3): Ditto.
6567 (avx_cmp<mode>3): Replace VF with VF_128_256.
6568 (*<sse>_maskcmp<mode>3_comm): Ditto.
6569 (<sse>_maskcmp<mode>3): Ditto.
6570 (<sse>_andnot<mode>3): Extend for AVX512.
6571 (<code><mode>3, anylogic): Replace VF with VF_128_256.
6572 (<code><mode>3, fpint_logic): New.
6573 (*<code><mode>3): Extend for AVX512.
6574 (avx512flogicsuff): New.
6575 (avx512f_<logic><mode>): Ditto.
6576 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
6578 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
6579 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
6580 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
6581 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
6582 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
6583 (xop_vpermil2<mode>3): Ditto.
6584 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
6585 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
6586 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
6587 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
6589 2013-09-25 Tom Tromey <tromey@redhat.com>
6591 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
6592 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
6593 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
6594 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
6595 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
6596 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
6597 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
6598 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
6599 (GRAPHITE_HTAB_H): Remove.
6601 2013-09-25 Tom Tromey <tromey@redhat.com>
6603 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
6605 2013-09-25 Tom Tromey <tromey@redhat.com>
6607 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
6609 2013-09-25 Tom Tromey <tromey@redhat.com>
6611 * config/i386/t-i386 (i386.o): Remove.
6612 (i386-c.o): Use COMPILE and POSTCOMPILE.
6614 2013-09-25 Tom Tromey <tromey@redhat.com>
6616 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
6618 2013-09-25 Tom Tromey <tromey@redhat.com>
6620 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
6621 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
6622 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
6623 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
6624 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
6625 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
6626 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
6627 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
6628 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
6629 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
6630 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
6631 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
6632 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
6633 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
6634 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
6635 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
6636 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
6637 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
6638 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
6639 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
6640 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
6641 (tree-ssa-pre.o, tree-ssa-sccvn.o)
6642 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
6643 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
6644 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
6645 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
6646 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
6647 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
6648 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
6649 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
6650 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
6651 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
6652 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
6653 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
6654 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
6655 (graphite-clast-to-gimple.o, graphite-dependences.o)
6656 (graphite-interchange.o, graphite-poly.o)
6657 (graphite-scop-detection.o, graphite-sese-to-poly.o)
6658 (graphite-optimize-isl.o, tree-vect-loop.o)
6659 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
6660 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
6661 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
6662 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
6663 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
6664 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
6665 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
6666 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
6667 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
6668 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
6669 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
6670 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
6671 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
6672 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
6673 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
6674 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
6675 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
6676 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
6677 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
6678 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
6679 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
6680 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
6681 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
6682 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
6683 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
6684 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
6685 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
6686 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
6687 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
6688 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
6689 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
6690 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
6691 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
6692 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
6693 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
6694 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
6695 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
6696 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
6697 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
6698 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
6699 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
6700 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
6701 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
6702 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
6703 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
6704 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
6705 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
6706 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
6707 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
6708 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
6709 (gcov-dump.o): Remove.
6710 (default-c.o): Use COMPILE and POSTCOMPILE.
6711 (CFLAGS-gcc.o): New variable.
6712 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
6714 2013-09-25 Tom Tromey <tromey@redhat.com>
6716 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
6717 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
6718 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
6719 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
6720 (c-family/c-pragma.o, c-family/c-pretty-print.o)
6721 (c-family/c-semantics.o, c-family/c-ada-spec.o)
6722 (c-family/array-notation-common.o, c-family/stub-objc.o)
6723 (c-family/c-ubsan.o): Remove.
6725 2013-09-25 Tom Tromey <tromey@redhat.com>
6727 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
6729 2013-09-25 Tom Tromey <tromey@redhat.com>
6731 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
6732 to add -DENABLE_SHARED_LIBGCC.
6733 (gcc.o): Don't use subshell.
6735 2013-09-25 Tom Tromey <tromey@redhat.com>
6737 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
6738 * configure.ac: Don't invoke AM_PROG_CC_C_O.
6739 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
6740 * configure, config.in: Rebuild.
6742 2013-09-25 Tom Tromey <tromey@redhat.com>
6744 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
6745 (COMPILE, POSTCOMPILE): New variables.
6746 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
6747 (DEPFILES): New variable.
6748 Include ".Po" files.
6749 * configure.ac: Add checks for dependency checking.
6750 * configure, aclocal.m4: Regenerate.
6752 2013-09-25 Tom Tromey <tromey@redhat.com>
6754 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
6755 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
6757 2013-09-25 Tom Tromey <tromey@redhat.com>
6759 * Makefile.in (generated_files): Add options.h,
6760 target-hooks-def.h, insn-opinit.h,
6761 common/common-target-hooks-def.h, pass-instances.def,
6762 c-family/c-target-hooks-def.h.
6764 2013-09-25 Jeff Law <law@redhat.com>
6766 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
6767 than foo[foo.length () - 1] to access last member in a vec.
6768 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
6770 2013-09-25 Richard Biener <rguenther@suse.de>
6773 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
6775 2013-09-25 Jan Hubicka <jh@suse.cz>
6777 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
6780 2013-09-25 Marek Polacek <polacek@redhat.com>
6783 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
6784 when determining the type name.
6786 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
6788 * config/sh/sh.md: Fix formatting.
6790 2013-09-24 Xinliang David Li <davidxl@google.com>
6792 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
6793 max peel iterations parameter.
6794 * param.def: New parameter.
6795 * doc/invoke.texi: Document New parameter.
6797 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
6799 * gimple-pretty-print.c: Various whitespace tweaks.
6800 * tree-core.h: Likewise.
6801 * tree-pretty-print.c: Likewise.
6802 * tree-ssa-alias.c: Likewise.
6803 * tree-ssa-copy.c: Likewise.
6804 * tree-ssanames.c: Likewise.
6805 * tree-ssanames.h: Likewise.
6806 * tree-vrp.c: Likewise.
6808 2013-09-24 Alan Modra <amodra@gmail.com>
6812 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
6813 for output operands that disallow regs. Don't use EXPAND_WRITE on
6816 2013-09-24 Richard Biener <rguenther@suse.de>
6819 * tree.c (reference_alias_ptr_type): Move ...
6820 * alias.c (reference_alias_ptr_type): ... here and implement
6821 in terms of the new reference_alias_ptr_type_1.
6822 (ref_all_alias_ptr_type_p): New helper.
6823 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
6824 use ref_all_alias_ptr_type_p.
6825 (get_deref_alias_set): Add flag_strict_aliasing check here.
6826 (reference_alias_ptr_type_1): New function, split out from ...
6827 (get_alias_set): ... here.
6828 (alias_ptr_types_compatible_p): New function.
6829 * alias.h (reference_alias_ptr_type): Declare.
6830 (alias_ptr_types_compatible_p): Likewise.
6831 * tree.h (reference_alias_ptr_type): Remove.
6832 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
6833 to compare MEM_REF alias types.
6835 2013-09-24 Richard Biener <rguenther@suse.de>
6837 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
6839 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6841 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
6842 reload helper function arrays into a single array reg_addr.
6843 (reload_fpr_gpr): Likewise.
6844 (reload_gpr_vsx): Likewise.
6845 (reload_vsx_gpr): Likewise.
6846 (struct rs6000_reg_addr): Likewise.
6847 (reg_addr): Likewise.
6848 (rs6000_debug_reg_global): Change rs6000_vector_reload,
6849 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
6850 (rs6000_init_hard_regno_mode_ok): Likewise.
6851 (rs6000_secondary_reload_direct_move): Likewise.
6852 (rs6000_secondary_reload): Likewise.
6854 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
6855 constraints: wu, ww, and wy. Repurpose wv constraint added during
6856 power8 changes. Put wg constraint in alphabetical order.
6858 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
6859 for future work to add ISA 2.07 VSX single precision support.
6860 (-mvsx-scalar-double): Change default from -1 to 1, update
6861 documentation comment.
6862 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
6863 (-mupper-regs-df): New debug switch to control whether DF values
6864 can go in the traditional Altivec registers.
6865 (-mupper-regs-sf): New debug switch to control whether SF values
6866 can go in the traditional Altivec registers.
6868 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
6870 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
6871 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
6872 Add new constraints, wu/ww/wy. Repurpose wv constraint.
6873 (rs6000_debug_legitimate_address_p): Print if we are running
6874 before, during, or after reload.
6875 (rs6000_secondary_reload): Add a comment.
6876 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
6878 * config/rs6000/constraints.md (wa constraint): Sort w<x>
6879 constraints. Update documentation string.
6880 (wd constraint): Likewise.
6881 (wf constraint): Likewise.
6882 (wg constraint): Likewise.
6883 (wn constraint): Likewise.
6884 (ws constraint): Likewise.
6885 (wt constraint): Likewise.
6886 (wx constraint): Likewise.
6887 (wz constraint): Likewise.
6888 (wu constraint): New constraint for ISA 2.07 SFmode scalar
6890 (ww constraint): Likewise.
6891 (wy constraint): Likewise.
6892 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
6893 in the previous submissions.
6894 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
6896 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
6898 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
6899 (REG_BR_PROB): Say that the probability is stored in an int_list.
6900 * reg-notes.def: Update commentary to mention INT_LIST.
6901 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
6902 (INT_LIST): New rtx.
6903 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
6904 * rtlanal.c (int_reg_note_p): New function.
6905 (alloc_reg_note): Assert that the note does not have an int argument.
6906 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
6907 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
6908 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
6909 rather than an INSN_LIST. Handle INT_LIST.
6910 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
6911 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
6912 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
6913 Manipulate them as ints rather than rtxes.
6914 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
6915 * regmove.c (copy_src_to_dest): Likewise.
6916 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
6918 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
6919 into the cases that need it.
6920 * config/arm/arm.c (arm_unwind_emit): Likewise.
6922 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
6923 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
6924 * loop-doloop.c (add_test, doloop_modify): Likewise.
6925 * loop-unswitch.c (compare_and_jump_seq): Likewise.
6926 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
6927 * predict.c (combine_predictions_for_insn): Likewise.
6928 * print-rtl.c (print_rtx): Handle INT_LIST.
6929 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
6930 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
6931 * config/arm/arm.c (emit_unlikely_jump): Likewise.
6932 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
6933 (ix86_split_fp_branch, predict_jump): Likewise.
6934 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
6935 * config/sh/sh.c (expand_cbranchsi4): Likewise.
6936 * config/spu/spu.c (ea_load_store_inline): Likewise.
6938 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
6939 value of a REG_BR_PROB note.
6940 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
6941 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
6942 * emit-rtl.c (try_split): Likewise.
6943 * predict.c (br_prob_note_reliable_p): Likewise.
6944 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
6945 * reorg.c (mostly_true_jump): Likewise.
6946 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
6947 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
6948 * config/i386/i386.c (ix86_print_operand): Likewise.
6949 * config/ia64/ia64.c (ia64_print_operand): Likewise.
6950 * config/mmix/mmix.c (mmix_print_operand): Likewise.
6951 * config/rs6000/rs6000.c (output_cbranch): Likewise.
6952 * config/s390/s390.c (s390_expand_tbegin): Likewise.
6953 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
6954 * config/sparc/sparc.c (output_cbranch): Likewise.
6955 * config/spu/spu.c (get_branch_target): Likewise.
6956 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
6957 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
6959 2013-09-23 Jan Hubicka <jh@suse.cz>
6961 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
6963 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
6964 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
6966 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
6967 (gimple_fold_call): Dump inconsistent devirtualizations; add
6968 sanity check for type based devirtualizations.
6969 * ipa-prop.c: Include ipa-utils.h
6970 (ipa_intraprocedural_devirtualization): Add sanity check.
6971 (try_make_edge_direct_virtual_call): Likewise.
6973 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
6975 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
6976 assignment statements.
6978 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
6980 * gimple-pretty-print.c (dump_ssaname_info): New function.
6981 (dump_gimple_phi): Call it.
6982 (pp_gimple_stmt_1): Likewise.
6983 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
6984 (range_info_def): Declare.
6985 * tree-pretty-print.c (pp_double_int): New function.
6986 (dump_generic_node): Call it.
6987 * tree-pretty-print.h (pp_double_int): Declare.
6988 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
6989 * tree-ssanames.h (range_info_def): New structure.
6990 (value_range_type): Move definition here.
6991 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
6993 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
6995 (set_range_info): New function.
6996 (get_range_info): Likewise.
6997 (duplicate_ssa_name_range_info): Likewise.
6998 (duplicate_ssa_name_fn): Check pointer type and call
6999 duplicate_ssa_name_range_info.
7000 * tree-ssa-copy.c (fini_copy_prop): Likewise.
7001 * tree-vrp.c (value_range_type): Remove definition, now in
7003 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
7004 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
7005 (SSA_NAME_RANGE_INFO): New macro.
7007 2013-09-23 Richard Biener <rguenther@suse.de>
7009 PR tree-optimization/58464
7010 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7011 (phi_trans_add): Change to add conditionally on being not present.
7012 (phi_translate_1): Remove recursion detection here.
7013 (phi_translate): Pre-seed the cache with NULL to catch
7014 recursion here in a more generic way.
7015 (bitmap_find_leader): Adjust comment.
7016 (get_representative_for): Dump value-numbers.
7017 (create_expression_by_pieces): Likewise.
7018 (insert_into_preds_of_block): Likewise.
7020 2013-09-23 Christian Bruel <christian.bruel@st.com>
7023 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
7024 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
7026 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
7029 2013-09-20 Renlin Li <renlin.li@arm.com>
7031 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
7033 (aarch64_expand_epilogue): Likewise.
7034 (aarch64_legitimize_reload_address): Likewise.
7036 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
7038 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
7039 NULL_TREE before pushing them onto the vector. Likewise for labels.
7041 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
7043 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
7044 comparison operators when -fno-trapping-math is in effect.
7045 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
7046 comparison operators in TFmode and assert that unsupported operators
7048 (ia64_print_operand): Likewise.
7050 2013-09-21 Jan Hubicka <jh@suse.cz>
7052 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
7053 (sse_typeless_stores): Enable for core
7054 (sse_load0_by_pxor): Likewise.
7055 (four_jump_limit): Disable for core.
7056 (pad_returns): Likewise.
7057 (avoid_vector_decode): Likewise.
7058 (fuse_cmp_and_branch): Enable for cores.
7059 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
7061 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7064 * config/pa/pa.c (pa_option_override): Disable auto increment and
7065 decrement instructions until reload is completed.
7067 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
7068 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
7069 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
7071 2013-09-20 DJ Delorie <dj@redhat.com>
7072 Nick Clifton <nickc@redhat.com>
7074 * config/rl78/rl78.c: Various whitespace and comment tweaks.
7075 (need_to_save): Save bank 0 on interrupts.
7076 (characterize_address): Strip far address wrappers.
7077 (rl78_as_legitimate_address): Likewise.
7078 (transcode_memory_rtx): Likewise.
7079 (rl78_peep_movhi_p): Disable this peephole after devirt.
7080 (rl78_propogate_register_origins): Forget all origins when a
7082 * config/rl78/rl78-virt.md: Various whitespace tweaks.
7083 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
7085 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
7086 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
7087 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
7089 * config/rl78/constraints.md: Various whitespace and paren tweaks.
7091 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7093 * config/pa/pa.md: In "scc" insn patterns, change output template to
7094 handle const0_rtx in reg_or_0_operand operands.
7096 2013-09-20 Martin Husemann <martin@NetBSD.org>
7099 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
7100 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
7102 2013-09-20 Richard Biener <rguenther@suse.de>
7105 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
7106 remembering SSA name version and block index.
7107 (new_scev_info_str): Adjust.
7108 (hash_scev_info): Likewise. Also hash the block index.
7109 (eq_scev_info): Adjust.
7110 (find_var_scev_info): Likewise.
7111 (struct instantiate_cache_entry): Remove.
7112 (struct instantiate_cache_type): Use a htab to map name, block
7114 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7115 (get_instantiated_value_entry): Likewise.
7116 (hash_idx_scev_info, eq_idx_scev_info): New functions.
7117 (instantiate_scev_name): Adjust.
7119 2013-09-20 Jeff Law <law@redhat.com>
7121 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
7123 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
7125 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
7126 Call aarch64_simd_expand_args to update op[argc].
7128 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7130 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
7133 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7135 * gengtype.c (file_rules): Added rule for *.cc files.
7136 (get_output_file_with_visibility): Give fatal message when no
7139 2013-09-20 Renlin Li <renlin.li@arm.com>
7141 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
7142 (aarch64_expand_epilogue): Likewise.
7143 (aarch64_legitimize_reload_address): Likewise.
7145 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
7148 * expr.c (expand_assignment): Remove misalignp code path.
7150 2013-09-20 Marek Polacek <polacek@redhat.com>
7153 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
7154 TYPE_PRECISION. Add asserts.
7156 2013-09-20 Richard Biener <rguenther@suse.de>
7158 PR tree-optimization/58453
7159 * tree-loop-distribution.c (distribute_loop): Apply the cost
7160 model for -ftree-loop-distribute-patterns, too.
7162 2013-09-20 Richard Biener <rguenther@suse.de>
7165 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
7166 make type comparison less strict.
7168 2013-09-20 Alan Modra <amodra@gmail.com>
7170 * configure: Regenerate.
7171 * aclocal.m4: Regenerate.
7173 2013-09-20 Marek Polacek <polacek@redhat.com>
7176 * doc/extend.texi: Document that attribute used is meant to be used
7177 on variables with static storage duration.
7179 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7181 PR tree-optimization/58472
7182 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
7183 simd_lane_access set inv_p = false.
7184 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
7185 the simduid magic VAR_DECL.
7187 2013-09-19 Jan Hubicka <jh@suse.cz>
7189 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
7191 2013-09-17 Jeff Law <law@redhat.com>
7193 * tree-ssa-dom.c (record_temporary_equivalences): New function
7194 split out of dom_opt_dom_walker::after_dom_children.
7195 (dom_opt_dom_walker::thread_across_edge): Move common code
7196 in here from dom_opt_dom_walker::after_dom_children.
7197 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
7199 2013-09-19 Jan Hubicka <jh@suse.cz>
7201 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
7202 (TARGET_GENERIC): Use PROCESOR_GENERIC
7203 (enum processor_type): Unify generic32 and 64.
7204 * i386.md (cpu): Likewise.
7205 * x86-tune.def (use_leave): Enable for generic32.
7206 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
7207 * athlon.md: Change generic64 to generic in all occurences.
7208 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
7209 (ix86_target_macros_internal): Likewise.
7210 * driver-i386.c (host_detect_local_cpu): Likewise.
7211 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
7213 (generic_memcpy, generic_memset, generic_cost): This one.
7214 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
7215 (m_GENERIC32, m_GENERIC64): Remove.
7216 (m_GENERIC): Turn into one flag.
7217 (processor_target): Unify generic tunnings.
7218 (ix86_option_override_internal): Replace generic32/64 by generic.
7219 (ix86_issue_rate): Likewise.
7220 (ix86_adjust_cost): Likewise.
7222 2013-09-19 Jan Hubicka <jh@suse.cz>
7224 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
7225 of speculative flag.
7227 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7229 * omp-low.c (expand_omp_sections): Always pass len - 1 to
7230 GOMP_sections_start, even if !exit_reachable.
7232 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7234 * lra-constraints.c (need_for_all_save_p): Use macro
7235 HARD_REGNO_CALL_PART_CLOBBERED.
7236 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
7237 set up pseudo conflict hard regs.
7239 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7242 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
7245 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7247 PR rtl-optimization/58438
7248 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
7249 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
7250 from previous subpasses.
7252 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
7254 * arm.c (arm_get_frame_offsets): Validate architecture supports
7255 LDRD/STRD before accepting the tuning preference.
7256 (arm_expand_prologue): Likewise.
7257 (arm_expand_epilogue): Likewise.
7259 2013-09-18 Richard Biener <rguenther@suse.de>
7261 PR tree-optimization/58417
7262 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
7263 have chrecs with symbols defined in the loop as operands.
7264 (chrec_fold_multiply): Likewise.
7265 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
7266 parameters before folding binary operations.
7267 (struct instantiate_cache_entry_hasher): Remove.
7268 (struct instantiate_cache_type): Use a pointer-map.
7269 (instantiate_cache_type::instantiate_cache_type): New function.
7270 (instantiate_cache_type::get): Likewise.
7271 (instantiate_cache_type::set): Likewise.
7272 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7273 (get_instantiated_value_entry): Likewise.
7274 (global_cache): New global.
7275 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
7276 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
7277 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
7278 (instantiate_scev_name): Adjust.
7279 (instantiate_scev): Construct global instead of local cache.
7280 (resolve_mixers): Likewise.
7282 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
7283 Paolo Carlini <paolo.carlini@oracle.com>
7286 * doc/implement-cxx.texi: Fix references to the C++ standards.
7288 2013-09-18 Jakub Jelinek <jakub@redhat.com>
7290 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
7291 * tree-vect-data-refs.c (vect_analyze_data_refs): For
7292 simd_lane_access drs, update also DR_ALIGNED_TO.
7294 2013-09-18 Marek Polacek <polacek@redhat.com>
7297 * doc/extend.texi: Document no_sanitize_undefined attribute.
7298 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
7299 no_sanitize_undefined attribute.
7301 2013-09-18 Nick Clifton <nickc@redhat.com>
7303 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
7304 (ASM_DECLARE_FUNCTION_NAME): Define.
7306 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
7308 * compare-elim.c (find_comparison_dom_walker): New class
7309 (find_comparisons_in_bb): Rename to
7310 find_comparison_dom_walker::before_dom_children
7311 (find_comparisons): Adjust
7312 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
7314 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
7315 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
7316 (init_walk_dominator_tree): Remove declaration.
7317 (fini_walk_dominator_tree): Remove declaration.
7318 * fwprop.c (single_def_use_dom_walker): New class
7319 (single_def_use_enter_block): Convert to
7320 single_def_use_dom_walker::before_dom_children.
7321 (single_def_use_leave_block): Convert to
7322 single_def_use_dom_walker::after_dom_children.
7323 (build_single_def_use_links): Adjust.
7324 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
7326 (find_candidates_in_block): Convert to
7327 find_candidates_dom_walker::before_dom_children.
7328 (execute_strength_reduction): Adjust.
7329 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
7330 (sese_dom_walker): New class.
7331 (sese_dom_walker::sese_dom_walker): New constructor.
7332 (sese_dom_walker::~sese_dom_walker): New destructor.
7333 (build_sese_conditions_before): Convert to
7334 sese_dom_walker::before_dom_children.
7335 (build_sese_conditions_after): Convert to
7336 sese_dom_walker::after_dom_children.
7337 (build_poly_scop): Adjust
7338 * tree-into-ssa.c (rewrite_dom_walker): New class
7339 (rewrite_enter_block): Convert to
7340 rewrite_dom_walker::before_dom_children.
7341 (rewrite_leave_block): Convert to
7342 rewrite_dom_walker::after_dom_children.
7343 (rewrite_update_dom_walker): New class.
7344 (rewrite_update_enter_block): Convert to
7345 rewrite_update_dom_walker::before_dom_children.
7346 (rewrite_update_leave_block): Convert to
7347 rewrite_update_dom_walker::after_dom_children.
7348 (rewrite_blocks, rewrite_into_ssa): Adjust.
7349 (mark_def_dom_walker): New class.
7350 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
7351 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
7352 (mark_def_sites_blocks): Convert to
7353 mark_def_dom_walker::before_dom_children.
7354 (mark_def_site_blocks): Remove.
7355 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
7356 (tree_ssa_dominator_optimize): Adjust.
7357 (dom_thread_across_edge): Convert to method
7358 dom_opt_dom_walker::thread_across_edge.
7359 (dom_opt_enter_block): Convert to member function
7360 dom_opt_dom_walker::before_dom_children.
7361 (dom_opt_leave_block): Convert to member function
7362 dom_opt_dom_walker::after_dom_children.
7363 * tree-ssa-dse.c (dse_dom_walker): New class.
7364 (dse_enter_block): Convert to member function
7365 dse_dom_walker::before_dom_children.
7366 (tree_ssa_dse): Adjust.
7367 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
7368 (determine_invariantness_stmt): Convert to method
7369 invariantness_dom_walker::before_dom_children.
7370 (determine_invariantness): Remove
7371 (move_computations_dom_walker): New class.
7372 (move_computations_stmt): Convert to method
7373 move_computations_dom_walker::before_dom_children.
7374 (move_computations, tree_ssa_lim): Adjust.
7375 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
7376 (nt_init_block): Convert to method
7377 notrappping_dom_walker::before_dom_children.
7378 (nt_fini_block): Convert to method
7379 method nontrapping_dom_walker::after_dom_children.
7380 (get_non_trapping): Adjust.
7381 * tree-ssa-pre.c (eliminate_dom_walker): New class.
7382 (eliminate_bb): Convert to method
7383 eliminate_dom_walker::before_dom_children.
7384 (eliminate_leave_block): Convert to method
7385 eliminate_dom_walker::after_dom_children.
7386 (eliminate): Adjust.
7387 * tree-ssa-strlen.c (strlen_dom_walker): New class.
7388 (strlen_enter_block): Convert to method
7389 strlen_dom_walker::before_dom_children.
7390 (strlen_leave_block): Convert to method
7391 method strlen_dom_walker::after_dom_children.
7392 (tree_ssa_strlen): Adjust.
7393 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
7394 (tree_ssa_uncprop): Adjust.
7395 (uncprop_leave_block): Convert to method
7396 uncprop_dom_walker::after_dom_children.
7397 (uncprop_leave_block): Convert to method
7398 uncprop_dom_walker::before_dom_children.
7400 2013-09-18 Bin Cheng <bin.cheng@arm.com>
7402 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
7403 branch in same basic block. Check both src and dest of the move insn.
7405 2013-09-17 Nick Clifton <nickc@redhat.com>
7407 * config/rl78/rl78-real.md (bf): New pattern.
7409 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
7410 (rl78_print_operand): Do not put a # before a %B.
7411 * config/rl78/rl78.opt: Tweak doc strings.
7413 2013-09-17 DJ Delorie <dj@redhat.com>
7415 * config/rl78/constraints.md (Wcv): Allow up to $r31.
7416 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
7417 (rl78_option_override): Likewise, if -mallregs.
7418 (is_virtual_register): Likewise.
7419 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
7420 (REGNO_OK_FOR_BASE_P): Likewise.
7421 * config/rl78/rl78.opt (-mallregs): New.
7423 2013-09-17 Nick Clifton <nickc@redhat.com>
7425 * config/rl78/rl78.c (need_to_save): Change return type to bool.
7426 For interrupt functions: save all call clobbered registers if the
7427 interrupt handler is not a leaf function.
7428 (rl78_expand_prologue): Always recompute the frame information.
7429 For interrupt functions: only select bank 0 if one of the bank 0
7430 registers is going to be psuhed.
7432 2013-09-17 DJ Delorie <dj@redhat.com>
7434 * config/rl78/constraints.md: For each W* constraint, rename to C*
7435 and create a W* constraint that checks for an optional ES: prefix
7437 * config/rl78/rl78.md (UNS_ES_ADDR): New.
7438 (es_addr): New. Used to wrap far addresses.
7439 * config/rl78/rl78-protos.h (rl78_es_addr): New.
7440 (rl78_es_base): New.
7441 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
7442 wrapped far addresses.
7443 (rl78_print_operand_1): Unwrap far addresses before processing.
7444 (rl78_lo16): Wrap far addresses in unspecs.
7445 (rl78_es_addr): New.
7446 (rl78_es_base): New.
7447 (insn_ok_now): Check for not-yet-wrapped far addresses.
7448 (transcode_memory_rtx): Properly re-wrap far addresses.
7450 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
7452 * config/sparc/t-rtems: Add leon3 multilibs.
7454 2013-09-17 Cong Hou <congh@google.com>
7456 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
7457 when checking the dot production pattern. The type of rhs operand
7458 of multiply is now checked correctly.
7460 2013-09-17 Jeff Law <law@redhat.com>
7462 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
7463 edge implied equivalences into successor phis.
7464 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
7465 here from tree-ssa-threadedge.c.
7466 (mark_threaded_blocks): When threading through a joiner, if both
7467 successors of the joiner's clone reach the same block, verify the
7468 PHI arguments are equal. If not, cancel the jump threading request.
7469 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
7470 tree-ssa-threadupdate.c
7471 (thread_across_edge): Don't check PHI argument equality when
7472 threading through joiner block here.
7474 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
7476 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
7477 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
7478 (warn_uninit, warn_uninitialized_vars,
7479 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7480 Move to tree-ssa-uninit.c.
7481 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
7482 (has_undefined_value_p): New. Pass dependant parts of
7483 ssa_undefined_value_p.
7484 (uninit_undefined_value_p): Use has_undefined_value_p.
7485 (warn_uninit, warn_uninitialized_vars,
7486 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7487 Move from tree-ssa.c.
7488 * tree-ssa.h: Adjust prototypes.
7490 2013-09-17 Jan Hubicka <jh@suse.cz>
7493 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
7494 * ipa-inline.c (can_inline_edge_p): Do not downgrade
7495 FUNCTION_NOT_OPTIMIZED.
7496 * ipa-inline-analysis.c (compute_inline_parameters): Function
7497 not optimized is not inlinable unless it is alwaysinline.
7498 (inline_analyze_function): Force calls in not optimized
7499 function not inlinable.
7501 2013-09-17 Jan Hubicka <jh@suse.cz>
7504 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
7506 * ipa.c (function_and_variable_visibility): Likewise.
7507 * ipa-profile.c (ipa_profile): Likewise.
7509 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
7512 * cgraph.c (cgraph_function_body_availability): Check for ifunc
7513 attribute, and don't inline the resolver in this case.
7515 2013-09-17 Teresa Johnson <tejohnson@google.com>
7517 * coverage.c (get_coverage_counts): Add missing newline.
7519 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7521 PR tree-optimization/58088
7522 * fold-const.c (mask_with_trailing_zeros): New function.
7523 (fold_binary_loc): Make sure we don't recurse infinitely
7524 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
7525 Use mask_with_trailing_zeros where appropriate.
7527 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
7529 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
7530 of 'prev' var to get better distance estimation.
7532 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
7534 * tree-inline.h (struct copy_body_data): Add transform_parameter.
7535 * tree-inline.c (is_parameter_of): New predicate.
7536 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
7537 a parameter has been remapped.
7538 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
7539 (optimize_inline_calls): Initialize transform_parameter.
7540 (copy_gimple_seq_and_replace_locals): Likewise.
7541 (tree_function_versioning): Likewise.
7542 (maybe_inline_call_in_expr): Likewise.
7544 2013-09-17 Nick Clifton <nickc@redhat.com>
7546 * config/msp430/msp430-protos.h: Add prototypes for new functions.
7547 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
7549 (is_attr_func): New function.
7550 (msp430_is_interrupt_func): New function.
7551 (is_naked_func): New function.
7552 (is_reentrant_func): New function.
7553 (is_critical_func): New function.
7554 (msp430_start_function): Add annotations for function attributes.
7555 (msp430_attr): New function.
7556 (msp430_attribute_table): New.
7557 (msp430_function_section): New function.
7558 (TARGET_ASM_FUNCTION_SECTION): Define.
7559 (msp430_builtin): New enum.
7560 (msp430_init_builtins): New function.
7561 (msp430_builtin_devl): New function.
7562 (msp430_expand_builtin): New function.
7563 (TARGET_INIT_BUILTINS): Define.
7564 (TARGET_EXPAND_BUILTINS): Define.
7565 (TARGET_BUILTIN_DECL): Define.
7566 (msp430_expand_prologue): Add support for naked, interrupt,
7567 critical and reentrant functions.
7568 (msp430_expand_epilogue): Likewise.
7569 (msp430_print_operand): Handle 'O' character.
7570 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
7572 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
7573 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
7574 (pushm): Use a 'n' rather than an 'i' constraint.
7575 (msp_return): Add generation of the interrupt return instruction.
7576 (disable_interrupts): New pattern.
7577 (enable_interrupts): New pattern.
7578 (push_intr_state): New pattern.
7579 (pop_intr_state): New pattern.
7580 (bic_SR): New pattern.
7581 (bis_SR): New pattern.
7582 * doc/extend.texi: Document MSP430 function attributes and builtin
7585 2013-09-17 Richard Biener <rguenther@suse.de>
7587 PR tree-optimization/58432
7588 * tree-loop-distribution.c (tree_loop_distribution): Also
7589 scan PHIs for outside loop uses and seed a partition from them.
7591 2013-09-17 Bin Cheng <bin.cheng@arm.com>
7593 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
7594 (restructure_reference): Call backtrace_base_for_ref.
7596 2013-09-17 Alan Modra <amodra@gmail.com>
7599 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
7602 2013-09-16 DJ Delorie <dj@redhat.com>
7604 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
7605 vregs location for RL78/G10.
7606 (rl78_expand_prologue): Avoid SEL on G10.
7607 (rl78_expand_epilogue): Likewise.
7608 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
7609 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
7610 __RL78_G10__ when appropriate.
7611 (ASM_SPEC): Pass -mg10 along to the assembler.
7612 * config/rl78/rl78.md (sel_rb): Disable for G10.
7613 * config/rl78/rl78.opt: Add -mg10 option.
7614 * config/rl78/t-rl78: Add -mg10 multilib.
7616 2013-09-16 Xinliang David Li <davidxl@google.com>
7618 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
7619 * omp-low.c (omp_max_vf): Ditto.
7620 (expand_omp_simd): Ditto.
7621 * tree-vectorizer.c (vectorize_loops): Ditto.
7622 (gate_vect_slp): Ditto.
7623 (gate_increase_alignment): Ditto.
7624 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
7625 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
7626 (gate_tree_vectorize): Name change.
7627 (tree_vectorize): Ditto.
7628 (pass_vectorize::gate): Call new function.
7629 (pass_vectorize::execute): Ditto.
7630 * opts.c: O3 default setting change.
7631 (finish_options): Check new flag.
7632 * doc/invoke.texi: Document new flags.
7633 * common.opt: New flags.
7635 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
7637 * doc/tm.texi.in (Cond Exec Macros): Remove node.
7638 (Condition Code): Don't reference it.
7639 * doc/tm.texi: Regenerate.
7641 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
7644 * lra-constraints.c (undo_optional_reloads): Consider all optional
7645 reload even if it did not get a hard reg.
7647 2013-09-16 Teresa Johnson <tejohnson@google.com>
7649 * dumpfile.c (dump_loc): Remove newline emission.
7650 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
7651 emission to dump_printf_loc calls where missing.
7652 (vect_mark_for_runtime_alias_test): Ditto.
7653 (vect_analyze_data_ref_dependence): Ditto.
7654 (vect_analyze_data_ref_dependences): Ditto.
7655 (vect_slp_analyze_data_ref_dependence): Ditto.
7656 (vect_slp_analyze_data_ref_dependences): Ditto.
7657 (vect_compute_data_ref_alignment): Ditto.
7658 (vect_update_misalignment_for_peel): Ditto.
7659 (vect_verify_datarefs_alignment): Ditto.
7660 (vector_alignment_reachable_p): Ditto.
7661 (vect_get_data_access_cost): Ditto.
7662 (vect_enhance_data_refs_alignment): Ditto.
7663 (vect_find_same_alignment_drs): Ditto.
7664 (vect_analyze_data_refs_alignment): Ditto.
7665 (vect_analyze_group_access): Ditto.
7666 (vect_analyze_data_ref_access): Ditto.
7667 (vect_analyze_data_ref_accesses): Ditto.
7668 (vect_prune_runtime_alias_test_list): Ditto.
7669 (vect_analyze_data_refs): Ditto.
7670 (vect_create_addr_base_for_vector_ref): Ditto.
7671 (vect_create_data_ref_ptr): Ditto.
7672 (vect_grouped_store_supported): Ditto.
7673 (vect_grouped_load_supported): Ditto.
7674 * value-prof.c (check_counter): Ditto.
7675 (check_ic_target): Ditto.
7676 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
7677 (vect_recog_widen_mult_pattern): Ditto.
7678 (vect_recog_widen_sum_pattern): Ditto.
7679 (vect_recog_over_widening_pattern): Ditto.
7680 (vect_recog_widen_shift_pattern): Ditto.
7681 (vect_recog_rotate_pattern): Ditto.
7682 (vect_recog_vector_vector_shift_pattern): Ditto.
7683 (vect_recog_divmod_pattern): Ditto.
7684 (vect_recog_mixed_size_cond_pattern): Ditto.
7685 (vect_recog_bool_pattern): Ditto.
7686 (vect_pattern_recog_1): Ditto.
7687 (vect_pattern_recog): Ditto.
7688 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
7689 (vect_is_simple_iv_evolution): Ditto.
7690 (vect_analyze_scalar_cycles_1): Ditto.
7691 (vect_get_loop_niters): Ditto.
7692 (vect_analyze_loop_1): Ditto.
7693 (vect_analyze_loop_form): Ditto.
7694 (vect_analyze_loop_operations): Ditto.
7695 (vect_analyze_loop_2): Ditto.
7696 (vect_analyze_loop): Ditto.
7697 (report_vect_op): Ditto.
7698 (vect_is_slp_reduction): Ditto.
7699 (vect_is_simple_reduction_1): Ditto.
7700 (vect_get_known_peeling_cost): Ditto.
7701 (vect_estimate_min_profitable_iters): Ditto.
7702 (vect_model_reduction_cost): Ditto.
7703 (vect_model_induction_cost): Ditto.
7704 (get_initial_def_for_induction): Ditto.
7705 (vect_create_epilog_for_reduction): Ditto.
7706 (vectorizable_reduction): Ditto.
7707 (vectorizable_induction): Ditto.
7708 (vectorizable_live_operation): Ditto.
7709 (vect_loop_kill_debug_uses): Ditto.
7710 (vect_transform_loop): Ditto.
7711 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
7712 (vect_stmt_relevant_p): Ditto.
7713 (process_use): Ditto.
7714 (vect_mark_stmts_to_be_vectorized): Ditto.
7715 (vect_model_simple_cost): Ditto.
7716 (vect_model_promotion_demotion_cost): Ditto.
7717 (vect_model_store_cost): Ditto.
7718 (vect_get_store_cost): Ditto.
7719 (vect_model_load_cost): Ditto.
7720 (vect_get_load_cost): Ditto.
7721 (vect_init_vector_1): Ditto.
7722 (vect_get_vec_def_for_operand): Ditto.
7723 (vect_finish_stmt_generation): Ditto.
7724 (vectorizable_call): Ditto.
7725 (vectorizable_conversion): Ditto.
7726 (vectorizable_assignment): Ditto.
7727 (vectorizable_shift): Ditto.
7728 (vectorizable_operation): Ditto.
7729 (vectorizable_store): Ditto.
7730 (vectorizable_load): Ditto.
7731 (vectorizable_condition): Ditto.
7732 (vect_analyze_stmt): Ditto.
7733 (vect_transform_stmt): Ditto.
7734 (vect_is_simple_use): Ditto.
7735 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
7736 (vect_can_advance_ivs_p): Ditto.
7737 (vect_update_ivs_after_vectorizer): Ditto.
7738 (vect_do_peeling_for_loop_bound): Ditto.
7739 (vect_gen_niters_for_prolog_loop): Ditto.
7740 (vect_update_inits_of_drs): Ditto.
7741 (vect_create_cond_for_alias_checks): Ditto.
7742 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
7743 (vect_build_slp_tree_1): Ditto.
7744 (vect_supported_load_permutation_p): Ditto.
7745 (vect_analyze_slp_instance): Ditto.
7746 (vect_analyze_slp): Ditto.
7747 (vect_make_slp_decision): Ditto.
7748 (vect_detect_hybrid_slp): Ditto.
7749 (vect_bb_vectorization_profitable_p): Ditto.
7750 (vect_slp_analyze_bb_1): Ditto.
7751 (vect_update_slp_costs_according_to_vf): Ditto.
7752 (vect_get_mask_element): Ditto.
7753 (vect_transform_slp_perm_load): Ditto.
7754 (vect_schedule_slp_instance): Ditto.
7755 (vect_schedule_slp): Ditto.
7756 (vect_slp_transform_bb): Ditto.
7757 * profile.c (read_profile_edge_counts): Ditto.
7758 (compute_branch_probabilities): Ditto.
7759 * coverage.c (get_coverage_counts): Ditto.
7761 2013-09-16 Diego Novillo <dnovillo@google.com>
7763 * tree-core.h: Add missing comment lines from refactoring of tree.h.
7765 2013-09-16 Jan Hubicka <jh@suse.cz>
7767 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
7768 abstract functions; for static functions check the presence of body.
7770 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7772 * config/aarch64/aarch64-simd-builtins.def (fma): New.
7773 * config/aarch64/aarch64-simd.md
7774 (aarch64_mla_elt<mode>): New.
7775 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7776 (aarch64_mls_elt<mode>): Likewise.
7777 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7778 (aarch64_fma4_elt<mode>): Likewise.
7779 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7780 (aarch64_fma4_elt_to_128v2df): Likewise.
7781 (aarch64_fma4_elt_to_64df): Likewise.
7782 (fnma<mode>4): Likewise.
7783 (aarch64_fnma4_elt<mode>): Likewise.
7784 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7785 (aarch64_fnma4_elt_to_128v2df): Likewise.
7786 (aarch64_fnma4_elt_to_64df): Likewise.
7787 * config/aarch64/iterators.md (VDQSF): New.
7788 * config/aarch64/arm_neon.h
7789 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
7790 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
7792 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7794 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
7795 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
7796 (aarch64_mul3_elt_to_128df): Likewise.
7797 (aarch64_mul3_elt_to_64v2df): Likewise.
7798 * config/aarch64/iterators.md (VEL): Also handle DFmode.
7800 (VMUL_CHANGE_NLANES) Likewise.
7803 * config/aarch64/arm_neon.h
7804 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
7806 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7808 * config/aarch64/arm_neon.h
7809 (vcvtx_high_f32_f64): Fix parameters.
7811 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7812 Uros Bizjak <ubizjak@gmail.com>
7814 * config/alpha.c: Include tree-ssanames.h.
7816 2013-09-16 Richard Biener <rguenther@suse.de>
7818 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
7819 (dot_rdg_1): Handle control_dd.
7820 (create_edge_for_control_dependence): New function.
7821 (create_rdg_edges): Add control dependences if asked for.
7822 (build_rdg): Likewise.
7823 (generate_loops_for_partition): If there are not necessary
7824 control stmts remove all their dependencies.
7825 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
7826 (distribute_loop): Pass on control dependences.
7827 (tree_loop_distribution): Compute control dependences and remove
7828 restriction on number of loop nodes.
7830 2013-09-16 Jakub Jelinek <jakub@redhat.com>
7832 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
7835 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
7837 * cse.c (try_const_anchors): Punt on CC modes.
7839 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7841 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
7843 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
7846 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
7849 2013-09-14 Iain Sandoe <iain@codesourcery.com>
7852 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
7854 (darwin_objc1_section): Likewise.
7855 (darwin_file_end): Emit Image Info section when required.
7857 2013-09-14 Jan Hubicka <jh@suse.cz>
7859 * tree-into-ssa.c (gate_into_ssa): New.
7860 (pass_data_build_ssa): Use it.
7861 * cgraph.h (expand_thunk): Update prototype.
7862 * cgraphunit.c (analyze_function): Expand thunks early.
7863 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
7864 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
7865 set lowered flag; do not add new function.
7866 (assemble_thunks_and_aliases): Update.
7867 * tree-ssa.c (gate_init_datastructures): New gate.
7868 (pass_data_init_datastructures): Use it.
7870 2013-09-14 Iain Sandoe <iain@codesourcery.com>
7873 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
7874 xmm register set described in the psABI.
7876 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
7878 * dwarf2out.c (should_emit_struct_debug): Add check
7879 for type_decl variable is not NULL.
7881 2013-09-13 Jacek Caban <jacek@codeweavers.com>
7883 * config.gcc: Use new winnt-c.c target hooks
7884 * config/t-winnt: New file
7885 * config/winnt-c.c: New file
7886 * doc/tm.texi.in: Document new hook
7887 * doc/tm.texi: Regenerated
7889 2013-09-13 Jan Hubicka <jh@suse.cz>
7892 * ipa-inline.c (check_callers): New function.
7893 (check_caller_edge): Remove.
7894 (want_inline_function_to_all_callers_p): Also permit alises that are
7896 (inline_to_all_callers): Terminate the walk when devirtualization
7897 introduce new calls.
7899 2013-09-13 Jan Hubicka <jh@suse.cz>
7901 * ipa-inline-analysis.c (struct growth_data): Add node.
7902 (do_estimate_growth_1): Fix detection of recursion.
7904 2013-09-13 Jakub Jelinek <jakub@redhat.com>
7906 PR tree-optimization/58392
7907 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
7908 to avoid shadowing of outer loop variable. If
7909 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
7910 replace_by_duplicate_decl simduid of loops that have it set and
7911 set dest_cfun->has_simduid_loops and/or
7912 dest_cfun->has_force_vect_loops.
7913 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
7914 instead of maybe_lookup_decl.
7915 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
7916 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
7917 call. Copy over force_vect and copy and remap simduid. Set
7918 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
7919 (copy_cfg_body): Remove blocks_to_copy argument. Use
7920 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
7921 caller. Don't set cfun->has_simduid_loops and/or
7922 cfun->has_force_vect_loops here.
7923 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
7925 (expand_call_inline, tree_function_versioning): Adjust copy_body
7928 2013-09-13 Martin Jambor <mjambor@suse.cz>
7931 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
7932 the assert if the edge was a speculative one.
7934 2013-09-13 Richard Biener <rguenther@suse.de>
7936 * tree-data-ref.h (known_dependences_p): Move here ...
7937 * tree-loop-distribution.c (known_dependences_p): ... from here.
7938 (dump_rdg_component, debug_rdg_component): Remove.
7940 (generate_loops_for_partition): Use gimple_uid instead of
7941 relying on matching stmt visit order.
7942 (rdg_build_partitions): Take starting stmt vector.
7943 (ldist_gen): Merge into ...
7944 (distribute_loop): ... this function. Do not compute starting
7946 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
7948 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7950 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
7951 Set type attribute correctly. Set predicable_short_it attribute.
7952 (cmpsi_shiftsi): Remove %? from output template.
7954 2013-09-13 Richard Biener <rguenther@suse.de>
7956 * tree-loop-distribution.c (struct rdg_component,
7957 rdg_defs_used_in_other_loops_p, free_rdg_components,
7958 rdg_build_components): Remove.
7959 (stmts_from_loop): Do not record virtual PHIs.
7960 (generate_loops_for_partition): Skip virtual PHIs.
7961 (build_rdg_partition_for_component): Rename to ...
7962 (build_rdg_partition_for_vertex): ... this and adjust.
7963 (rdg_build_partitions): Take a vector of starting vertices
7964 instead of components. Remove unnecessary leftover handling.
7965 (ldist_gen): Do not build components or record other stores.
7966 (distribute_loop): Do not distribute loops containing stmts
7969 2013-09-13 Christian Bruel <christian.bruel@st.com>
7972 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
7974 2013-09-13 Kai Tietz <ktietz@redhat.com>
7976 * config.gcc: Separate cases for mingw and cygwin targets,
7977 and add 64-bit cygwin target case.
7979 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
7980 dll-export inline-functions.
7981 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
7984 2013-09-13 Jeff Law <law@redhat.com>
7988 2013-09-06 Jeff Law <law@redhat.com>
7990 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
7991 edge implied equivalences into successor phis.
7993 2013-09-12 DJ Delorie <dj@redhat.com>
7995 * config/rl78/rl78-virt.md: Change from | to \; for asm line
7998 2013-09-12 Brooks Moses <bmoses@google.com>
8001 * Makefile.in: Do not install driver binaries in $(target)/bin.
8003 2013-09-12 DJ Delorie <dj@redhat.com>
8005 * config/rl78/rl78.opt (mrelax): New.
8006 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
8007 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
8009 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
8011 (rl78_expand_epilogue): Likewise.
8013 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
8016 * lra-eliminations.c (remove_reg_equal_offset_note): New.
8017 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
8018 pointer elimination with using remove_reg_equal_offset_note.
8020 2013-09-12 DJ Delorie <dj@redhat.com>
8022 * config/msp430/: New port.
8023 * config.gcc (msp430): Added.
8024 * doc/invoke.texi: Document MSP430 options.
8025 * doc/install.texi: Document msp430-elf
8026 * doc/md.texi: Document msp430-elf
8027 * doc/contrib.texi: Document msp430-elf
8029 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
8032 2013-09-12 Martin Jambor <mjambor@suse.cz>
8035 * ipa-prop.c (remove_described_reference): Give up if the edge in the
8036 reference descriptor is NULL.
8037 (ipa_edge_removal_hook): If owning a reference descriptor, set its
8040 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8042 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
8043 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
8044 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
8045 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
8046 make_temp_ssa_name): move to tree-ssanames.h
8047 * tree-ssa-alias.h: Move prototype.
8048 * tree-ssa.h: Include tree-ssanames.h.
8049 * tree-ssanames.c (FREE_SSANAMES): Move to here.
8050 * tree-ssanames.h: New. Move items from tree-flow*.h
8051 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
8053 2013-09-12 Richard Biener <rguenther@suse.de>
8055 PR tree-optimization/58404
8056 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
8057 propagate non-invariant addresses into dereferences wrapped
8058 in component references.
8060 2013-09-12 Richard Biener <rguenther@suse.de>
8062 PR tree-optimization/58402
8063 * passes.def: Move pass_late_warn_uninitialized later.
8065 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8067 * tree-ssa.h: New. Move content from tree-flow.h and
8069 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
8070 Move prototypes belonging to tree-ssa.c.
8071 * tree-flow-inline.h (redirect_edge_var_map_def,
8072 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
8074 * gimple.h: Adjust prototypes.
8075 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
8077 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
8078 * tree.h: Move prototype to tree-ssa.h.
8079 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
8080 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
8081 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
8082 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
8083 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
8084 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
8085 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
8086 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
8087 graphite-blocking.c, graphite-clast-to-gimple.c,
8088 graphite-dependences.c, graphite-interchange.c,
8089 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
8090 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
8091 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
8092 ipa-reference.c, ipa-split.c, ipa-utils.c,
8093 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
8094 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
8095 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
8096 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
8097 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
8098 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
8099 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
8100 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
8101 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
8102 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
8103 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
8104 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
8105 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
8106 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
8107 value-prof.c, var-tracking.c,
8108 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
8110 2013-09-12 Richard Biener <rguenther@suse.de>
8112 PR tree-optimization/58396
8113 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
8114 (build_rdg): Take a loop-nest parameter, fix memleaks.
8115 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
8117 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8119 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
8122 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
8124 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
8125 The FMA instruction names should have a 'v' prefix.
8127 2013-09-12 Richard Biener <rguenther@suse.de>
8129 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
8130 (dot_rdg): Use popen instead of system in optional code.
8131 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
8132 (already_processed_vertex_p): Adjust.
8133 (has_anti_or_output_dependence, predecessor_has_mem_write,
8134 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
8135 rdg_flag_uses): Remove.
8136 (rdg_flag_vertex): Simplify.
8137 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
8139 (build_rdg_partition_for_component): Process the first vertex
8140 of a component only.
8141 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
8143 2013-09-12 Alan Modra <amodra@gmail.com>
8145 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
8147 2013-09-11 DJ Delorie <dj@redhat.com>
8148 Nick Clifton <nickc@redhat.com>
8150 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
8151 (rl78_stack_based_mem): New.
8152 * config/rl78/constraints.md (Iv08): New.
8161 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
8163 (movsi): Change from expand to insn-and-split.
8164 (ashrsi3): Clobber AX.
8168 * config/rl78/rl78.md (CC_REG): Fix.
8169 (addsi3): Allow memory and immediate operands.
8170 (addsi3_internal): Split into...
8171 (addsi3_internal_virt): ...new, and ...
8172 (addsi3_internal_real): ...new.
8174 (subsi3_internal_virt): New.
8175 (subsi3_internal_real): New.
8176 (mulsi3): Add memory operand.
8177 (mulsi3_rl78): Likewise.
8178 (mulsi3_g13): Likewise.
8179 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
8180 (cbranchqi4_real): Add more constraint options.
8181 (cbranchhi4_real): Expand pattern.
8182 (cbranchhi4_real_signed): New.
8183 (cbranchhi4_real_inverted): New.
8184 (cbranchsi4_real_lt): New.
8185 (cbranchsi4_real_ge): New.
8186 (cbranchsi4_real_signed): New.
8187 (cbranchsi4_real): New.
8189 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
8191 (lshrsi3_virt): Likewise.
8192 (ashlsi3_virt): Likewise.
8193 (cbranchqi4_virt_signed): New.
8194 (cbranchhi4_virt_signed): New.
8195 (cbranchsi4_virt): New.
8196 * config/rl78/rl78.c: Whitespace fixes throughout.
8197 (move_elim_pass): New.
8198 (pass_data_rl78_move_elim): New.
8199 (pass_rl78_move_elim): New.
8200 (make_pass_rl78_move_elim): New.
8201 (rl78_devirt_info): Run devirt earlier.
8202 (rl78_move_elim_info): New.
8203 (rl78_asm_file_start): Register it.
8204 (rl78_split_movsi): New.
8205 (rl78_as_legitimate_address): Allow virtual base registers when
8207 (rl78_addr_space_convert): Remove spurious debug stuff.
8208 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
8209 (rl78_print_operand): More cases for not printing '#'.
8210 (rl78_expand_compare): Remove most of the logic.
8211 (content_memory): New.
8212 (clear_content_memory): New.
8213 (get_content_index): New.
8214 (get_content_name): New.
8215 (display_content_memory): New.
8216 (update_content): New.
8217 (record_content): New.
8218 (already_contains): New.
8219 (insn_ok_now): Re-recog insns with virtual registers.
8220 (add_postponed_content_update): New.
8221 (process_postponed_content_update): New.
8222 (gen_and_emit_move): New.
8223 (transcode_memory_rtx): Record new location content.
8224 Use gen_and_emit_move.
8225 (force_into_acc): New.
8226 (move_to_acc): Use gen_and_emit_move.
8227 (move_from_acc): Likewise.
8228 (move_acc_to_reg): Likewise.
8229 (move_to_x): Likewise.
8230 (move_to_hl): Likewise.
8231 (move_to_de): Likewise.
8232 (rl78_alloc_physical_registers_op1): Record location content.
8233 (has_constraint): New.
8234 (rl78_alloc_physical_registers_op2): Record location content.
8236 (rl78_alloc_physical_registers_ro1): Likewise.
8237 (rl78_alloc_physical_registers_cmp): Likewise.
8238 (rl78_alloc_physical_registers_umul): Likewise.
8239 (rl78_alloc_address_registers_macax): New.
8240 (rl78_alloc_physical_registers): Initialize and set location
8241 content memory as needed.
8242 (rl78_reorg): Make sure split2 is called.
8243 (rl78_rtx_costs): New.
8245 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
8247 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
8248 for (not (neg ...)) and (neg (not ...)) cases.
8250 2013-09-11 Richard Biener <rguenther@suse.de>
8253 * passes.def: Split critical edges before late uninit warning passes.
8254 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
8256 2013-09-11 Jakub Jelinek <jakub@redhat.com>
8258 PR tree-optimization/58385
8259 * fold-const.c (build_range_check): If both low and high are NULL,
8260 use omit_one_operand_loc to preserve exp side-effects.
8262 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8264 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
8266 2013-09-11 Richard Biener <rguenther@suse.de>
8268 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
8269 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
8270 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
8271 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
8272 stmts_from_loop, known_dependences_p, build_empty_rdg,
8273 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
8274 * tree-loop-distribution.c: ... here.
8275 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
8276 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
8277 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
8278 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
8279 * tree-loop-distribution.c: ... here.
8280 * tree-loop-distribution.c: Include gimple-pretty-print.h.
8281 (struct partition_s): Add loops member.
8282 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
8283 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
8284 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
8286 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8287 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8288 Sergey Lega <sergey.s.lega@intel.com>
8289 Anna Tikhonova <anna.tikhonova@intel.com>
8290 Ilya Tocar <ilya.tocar@intel.com>
8291 Andrey Turetskiy <andrey.turetskiy@intel.com>
8292 Ilya Verbin <ilya.verbin@intel.com>
8293 Kirill Yukhin <kirill.yukhin@intel.com>
8294 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8296 * config/i386/constraints.md (k): New.
8298 * config/i386/i386.c (const regclass_map): Add new mask registers.
8299 (dbx_register_map): Ditto.
8300 (dbx64_register_map): Ditto.
8301 (svr4_dbx_register_map): Ditto.
8302 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
8304 (ix86_preferred_reload_class): Disable constants for mask registers.
8305 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
8306 (ix86_hard_regno_mode_ok): Support new mask registers.
8307 (x86_order_regs_for_local_alloc): Ditto.
8308 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
8309 (FIXED_REGISTERS): Add new mask registers.
8310 (CALL_USED_REGISTERS): Ditto.
8311 (REG_ALLOC_ORDER): Ditto.
8312 (VALID_MASK_REG_MODE): New.
8313 (FIRST_MASK_REG): Ditto.
8314 (LAST_MASK_REG): Ditto.
8315 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
8316 (MAYBE_MASK_CLASS_P): New.
8317 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
8318 (REG_CLASS_CONTENTS): Ditto.
8319 (MASK_REGNO_P): New.
8320 (ANY_MASK_REG_P): Ditto.
8321 (HI_REGISTER_NAMES): Add new mask registers.
8322 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
8323 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
8325 (attribute "type"): Add mskmov, msklog.
8326 (attribute "length_immediate"): Support them.
8327 (attribute "memory"): Ditto.
8328 (attribute "prefix_0f"): Ditto.
8329 (*movhi_internal): Support new mask registers.
8330 (*movqi_internal): Ditto.
8331 (define_split): Split out clobber pattern is a logic
8332 insn on mask registers.
8333 (*k<logic><mode>): New.
8334 (*andhi_1): Extend to support mask regs.
8335 (*andqi_1): Extend to support mask regs.
8337 (define_split): Split and-not to and and not if operands
8339 (*<code><mode>_1): Separate HI mode to new pattern...
8340 (*<code>hi_1): This.
8341 (*<code>qi_1): Extend to support mask regs.
8343 (kortestzhi): Ditto.
8344 (kortestchi): Ditto.
8346 (*one_cmpl<mode>2_1): Remove HImode and handle it...
8347 (*one_cmplhi2_1): ...Here, now with mask registers support.
8348 (*one_cmplqi2_1): Support new mask registers.
8349 (HI/QImode arithmetics splitter): Don't split if mask registers
8351 (HI/QImode not splitter): Ditto.
8352 * config/i386/predicated.md (mask_reg_operand): New.
8353 (general_reg_operand): Ditto.
8355 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8357 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
8358 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
8360 2013-09-10 Jeff Law <law@redhat.com>
8362 PR tree-optimization/58380
8363 * tree-ssa-threadupdate.c (thread_block): Recognize another case
8364 of threading through a buried loop header.
8366 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
8367 return value for single successor case.
8369 2013-09-10 Jan Hubicka <jh@suse.cz>
8371 * ipa-devirt.c (ipa_devirt): Enable with LTO.
8373 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
8376 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
8377 support conditional execution.
8378 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
8380 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
8382 * lra.c (lra): Clear lra_optional_reload_pseudos before every
8384 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
8385 Check destination too to check move insn.
8386 (undo_optional_reloads): Add check that the original peudo did not
8387 changed its allocation and the optional reload was inherited on last
8388 inheritance pass. Break loop after deciding to keep optional reload.
8389 (lra_undo_inheritance): Add check that inherited pseudo still in
8392 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
8394 * config/aarch64/aarch64.md (generic_sched): New.
8395 * config/aarch64/aarch64-generic.md (load): Make conditional
8396 on generic_sched attribute.
8397 (nonload): Likewise.
8399 2013-09-10 Jan Hubicka <jh@suse.cz>
8401 * lto-cgraph.c: Include ipa-utils.h.
8402 (compute_ltrans_boundary): Also add possible targets into the boundary.
8404 2013-09-10 Jan Hubicka <jh@suse.cz>
8406 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
8407 VAR_DECL of vtable rather than full expression.
8409 2013-09-10 Jan Hubicka <jh@suse.cz>
8410 Paolo Carlini <paolo.carlini@oracle.com>
8412 * cgraphunit.c (analyze_functions): Save input_location, set it
8413 to UNKNOWN_LOCATION and restore it at the end.
8415 2013-09-10 Martin Jambor <mjambor@suse.cz>
8417 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
8418 represented by a thunk.
8420 2013-09-10 Jeff Law <law@redhat.com>
8422 PR tree-optimization/58343
8423 * tree-ssa-threadupdate.c (thread_block): Identify and disable
8424 jump threading requests through loop headers buried in the middle
8425 of a jump threading path.
8427 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
8428 in return value/type.
8430 2013-09-10 Jakub Jelinek <jakub@redhat.com>
8432 PR rtl-optimization/58365
8433 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
8434 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
8437 2013-09-10 Richard Biener <rguenther@suse.de>
8439 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
8440 * tree-data-ref.c (create_rdg_vertices): Collect all data
8441 references, signal failure to the caller, use data-ref API.
8442 (build_rdg): Compute data references only once. Maintain lifetime
8443 of data references and data dependences from within RDG.
8444 (free_rdg): Free dependence relations.
8445 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
8446 inventing extra dependences.
8447 (distribute_loop): Update for RDG API changes.
8449 2013-09-10 Kai Tietz <ktietz@redhat.com>
8451 * doc/invoke.texi (fms-extensions): Document changed
8452 behavior for ms-abi targets.
8453 * config/i386/i386.c (ix86_option_override_internal):
8454 Set default value of option -fms-extension for ms-abi targets.
8456 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
8458 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
8460 2013-09-10 Alan Modra <amodra@gmail.com>
8463 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
8465 2013-09-10 Alan Modra <amodra@gmail.com>
8467 * config/rs6000/predicates.md (add_cint_operand): New.
8468 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
8469 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
8470 using add_cint_operand.
8471 (largetoc_high_plus_aix): Likewise.
8473 2013-09-09 Jakub Jelinek <jakub@redhat.com>
8475 PR tree-optimization/58364
8476 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
8477 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
8478 the current range can't be an unconditional true or false.
8480 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
8482 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
8484 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
8486 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
8488 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
8491 * doc/invoke.texi (-Wdelete-incomplete): Document it.
8493 2013-09-09 Ian Bolton <ian.bolton@arm.com>
8495 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
8496 NO_REGS for immediate that can't be moved directly into FP_REGS.
8498 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8500 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
8501 comparison with negated operand.
8502 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
8505 2013-09-09 Richard Biener <rguenther@suse.de>
8508 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
8509 of a subloop record all its block as affecting loop-closed SSA form.
8511 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8513 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
8514 of an rtx/bitpos pair.
8515 (store_fixed_bit_field): Update accordingly.
8517 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8519 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
8521 * builtins.c (expand_errno_check): Likewise.
8522 * dwarf2cfi.c (init_return_column_size): Likewise.
8523 * except.c (sjlj_mark_call_sites): Likewise.
8524 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
8525 * lra-constraints.c (emit_inc): Likewise.
8526 * ree.c (combine_set_extension): Likewise.
8527 * regmove.c (fixup_match_2): Likewise.
8528 * reload1.c (inc_for_reload): Likewise.
8530 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8532 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
8533 (force_to_mode, simplify_shift_const_1, simplify_comparison):
8534 Use gen_int_mode with the mode of the associated simplify_* call.
8535 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
8536 * expmed.c (expand_shift_1): Likewise.
8537 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8538 * loop-iv.c (iv_number_of_iterations): Likewise.
8539 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8540 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
8542 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8544 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
8545 of the associated expand_* call.
8546 (asan_emit_stack_protection): Likewise.
8547 * builtins.c (round_trampoline_addr): Likewise.
8548 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
8549 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
8550 (emit_store_flag): Likewise.
8551 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
8552 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
8554 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8555 * ifcvt.c (noce_try_store_flag_constants): Likewise.
8556 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8557 * modulo-sched.c (generate_prolog_epilog): Likewise.
8558 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
8559 (expand_ctz, expand_ffs, expand_unop): Likewise.
8561 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8563 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
8564 of the associated gen_rtx_* call.
8565 * caller-save.c (init_caller_save): Likewise.
8566 * combine.c (find_split_point, make_extraction): Likewise.
8567 (make_compound_operation): Likewise.
8568 * dwarf2out.c (mem_loc_descriptor): Likewise.
8569 * explow.c (plus_constant, probe_stack_range): Likewise.
8570 * expmed.c (expand_mult_const): Likewise.
8571 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
8572 * reload1.c (init_reload): Likewise.
8573 * valtrack.c (cleanup_auto_inc_dec): Likewise.
8574 * var-tracking.c (adjust_mems): Likewise.
8575 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
8576 rather than gen_rtx_fmt_ee.
8578 2013-09-09 Jan Hubicka <jh@suse.cz>
8581 * value-prof.c (gimple_ic): Copy also abnormal edges.
8583 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8585 * asan.c (asan_shadow_cst): Use gen_int_mode.
8587 2013-09-08 Jan Hubicka <jh@suse.cz>
8589 * ipa-profile.c: Add toplevel comment.
8590 (ipa_propagate_frequency_1): Be more conservative when profile is read.
8591 (contains_hot_call_p): New function.
8592 (ipa_propagate_frequency): Set frequencies based on counts when
8594 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
8595 profile; do not tamper with profile after inlining if it is read.
8597 2013-09-08 Jan Hubicka <jh@suse.cz>
8599 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
8602 2013-09-08 Jan Hubicka <jh@suse.cz>
8604 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
8607 2013-09-08 Jeff Law <law@redhat.com>
8610 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
8613 2013-09-08 Andi Kleen <ak@linux.intel.com>
8615 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
8617 2013-09-08 Jan Hubicka <jh@suse.cz>
8619 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
8621 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
8623 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
8624 for non-debug insns.
8625 * lra.c (new_insn_reg): Take the containing insn as a parameter.
8626 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
8627 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
8630 2013-09-08 Jan Hubicka <jh@suse.cz>
8632 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
8633 targets and devirtualize to BUILT_IN_UNREACHABLE.
8634 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
8635 * ipa.c (walk_polymorphic_call_targets): New function.
8636 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
8637 functions; use the new timevar.
8638 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
8639 was removed from the program.
8640 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
8641 not consider it in the walk when its vtable is dead.
8642 (possible_polymorphic_call_targets_1): Pass anonymous flag to
8644 (devirt_variable_node_removal_hook): New function.
8645 (possible_polymorphic_call_targets): Also register
8646 devirt_variable_node_removal_hook.
8647 (ipa_devirt): Do not do non-speculative devirtualization.
8648 (gate_ipa_devirt): One execute if devirtualizing speculatively.
8650 2013-09-08 Jan Hubicka <jh@suse.cz>
8652 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
8653 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
8654 varpool_remove_variable_insertion_hook): Declare.
8655 * varpool.c (varpool_node_hook_list): New structure.
8656 (first_varpool_node_removal_hook,
8657 first_varpool_variable_insertion_hook): New variables.
8658 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
8659 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
8660 varpool_remove_variable_insertion_hook,
8661 varpool_call_variable_insertion_hooks): New functions.
8662 (varpool_remove_node): Use it.
8664 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
8667 * diagnostic.c (diagnostic_build_prefix): When s.file is
8668 "<built-in>" don't output line and column numbers.
8670 2013-09-06 Jan Hubicka <jh@suse.cz>
8672 * cgraphunit.c (expand_thunk): Get body before touching arguments.
8673 * lto-streamer-out.c: Stream thunks, too.
8674 * lto-streamer-in.c (input_function): Pop cfun here
8675 (lto_read_body): Instead of here.
8677 2013-09-06 Caroline Tice <cmtice@google.com>
8679 * doc/install.texi: Add documentation for the --enable-vtable-verify
8680 and the --disable-libvtv configure options.
8682 2013-09-06 Jeff Law <law@redhat.com>
8684 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8685 edge implied equivalences into successor phis.
8687 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
8689 * resource.c (mark_referenced_resources): Handle COND_EXEC.
8691 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
8693 * resource.c (mark_target_live_regs): Compute resources taking
8694 into account if a call is predicated or not.
8696 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
8698 * toplev.c (output_stack_usage): Be prepared for suffixes created by
8699 the compiler in the function names.
8701 2013-09-06 Jan Hubicka <jh@suse.cz>
8704 * ipa-inline.c (has_caller_p): New function.
8705 (want_inline_function_to_all_callers_p): Use it.
8706 (sum_callers, inline_to_all_callers): Break out from ...
8707 (ipa_inline): ... here.
8709 2013-09-06 Jan Hubicka <jh@suse.cz>
8711 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
8712 only when AVX is enabled.
8714 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8716 * config/aarch64/aarch64.md
8717 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
8718 is fpsimd_<load/store>2.
8719 (load_pair<mode>): Likewise.
8720 (store_pair<mode>): Likewise.
8722 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8724 * config/arm/types.md (type): Add "mrs" type.
8725 * config/aarch64/aarch64.md
8726 (aarch64_load_tp_hard): Make type "mrs".
8728 (load_tp_hard): Make type "mrs".
8729 * config/arm/cortex-a15.md: Update with new attributes.
8730 * config/arm/cortex-a5.md: Update with new attributes.
8731 * config/arm/cortex-a53.md: Update with new attributes.
8732 * config/arm/cortex-a7.md: Update with new attributes.
8733 * config/arm/cortex-a8.md: Update with new attributes.
8734 * config/arm/cortex-a9.md: Update with new attributes.
8735 * config/arm/cortex-m4.md: Update with new attributes.
8736 * config/arm/cortex-r4.md: Update with new attributes.
8737 * config/arm/fa526.md: Update with new attributes.
8738 * config/arm/fa606te.md: Update with new attributes.
8739 * config/arm/fa626te.md: Update with new attributes.
8740 * config/arm/fa726te.md: Update with new attributes.
8742 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8744 * config/aarch64/aarch64.md
8745 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
8746 (*movtf_aarch64): Likewise.
8748 (thumb1_movdi_insn): Use "multiple" for type where more than one
8749 instruction is used for a move.
8750 (*arm32_movhf): Likewise.
8751 (*thumb_movdf_insn): Likewise.
8753 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8755 * config/arm/types.md (type): Rename fcpys to fmov.
8757 (*arm_movsi_vfp): Rename type fcpys as fmov.
8758 (*thumb2_movsi_vfp): Likewise
8759 (*movhf_vfp_neon): Likewise
8760 (*movhf_vfp): Likewise
8761 (*movsf_vfp): Likewise
8762 (*thumb2_movsf_vfp): Likewise
8763 (*movsfcc_vfp): Likewise
8764 (*thumb2_movsfcc_vfp): Likewise
8765 * config/aarch64/aarch64-simd.md
8766 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
8767 * config/aarch64/aarch64.md
8768 (*movsi_aarch64): Replace type mov_reg with fmovs.
8769 (*movdi_aarch64): Likewise
8770 (*movsf_aarch64): Likewise
8771 (*movdf_aarch64): Likewise
8773 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
8774 * config/arm/iwmmxt.md
8775 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
8776 * config/arm/arm1020e.md: Update with new attributes.
8777 * config/arm/cortex-a15-neon.md: Update with new attributes.
8778 * config/arm/cortex-a5.md: Update with new attributes.
8779 * config/arm/cortex-a53.md: Update with new attributes.
8780 * config/arm/cortex-a7.md: Update with new attributes.
8781 * config/arm/cortex-a8-neon.md: Update with new attributes.
8782 * config/arm/cortex-a9.md: Update with new attributes.
8783 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8784 * config/arm/cortex-r4f.md: Update with new attributes.
8785 * config/arm/marvell-pj4.md: Update with new attributes.
8786 * config/arm/vfp11.md: Update with new attributes.
8788 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8790 * config/aarch64/aarch64.md
8791 (*madd<mode>): Fix type attribute.
8792 (*maddsi_uxtw): Likewise.
8793 (*msub<mode>): Likewise.
8794 (*msubsi_uxtw): Likewise.
8795 (<su_optab>maddsidi4): Likewise.
8796 (<su_optab>msubsidi4): Likewise.
8798 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8800 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
8801 * config/arm/arm.md (core_cycles): Remove fdiv.
8802 * config/arm/vfp.md:
8803 (*sqrtsf2_vfp): Update for attribute changes.
8804 (*sqrtdf2_vfp): Likewise.
8805 * config/aarch64/aarch64.md:
8806 (sqrt<mode>2): Update for attribute changes.
8807 * config/arm/arm1020e.md: Update with new attributes.
8808 * config/arm/cortex-a15-neon.md: Update with new attributes.
8809 * config/arm/cortex-a5.md: Update with new attributes.
8810 * config/arm/cortex-a53.md: Update with new attributes.
8811 * config/arm/cortex-a7.md: Update with new attributes.
8812 * config/arm/cortex-a8-neon.md: Update with new attributes.
8813 * config/arm/cortex-a9.md: Update with new attributes.
8814 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8815 * config/arm/cortex-r4f.md: Update with new attributes.
8816 * config/arm/marvell-pj4.md: Update with new attributes.
8817 * config/arm/vfp11.md: Update with new attributes.
8819 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8821 * config/arm/types.md
8822 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
8823 * config/aarch64/aarch64.md
8824 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
8826 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
8827 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
8828 (float<GPI:mode><GPF:mode>2): Likewise.
8830 (*truncsisf2_vfp): Update with new attributes.
8831 (*truncsidf2_vfp): Likewise.
8832 (fixuns_truncsfsi2): Likewise.
8833 (fixuns_truncdfsi2): Likewise.
8834 (*floatsisf2_vfp): Likewise.
8835 (*floatsidf2_vfp): Likewise.
8836 (floatunssisf2): Likewise.
8837 (floatunssidf2): Likewise.
8838 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
8839 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
8840 * config/arm/arm1020e.md: Update with new attributes.
8841 * config/arm/cortex-a15-neon.md: Update with new attributes.
8842 * config/arm/cortex-a5.md: Update with new attributes.
8843 * config/arm/cortex-a53.md: Update with new attributes.
8844 * config/arm/cortex-a7.md: Update with new attributes.
8845 * config/arm/cortex-a8-neon.md: Update with new attributes.
8846 * config/arm/cortex-a9.md: Update with new attributes.
8847 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8848 * config/arm/cortex-r4f.md: Update with new attributes.
8849 * config/arm/marvell-pj4.md: Update with new attributes.
8850 * config/arm/vfp11.md: Update with new attributes.
8852 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8854 * config/aarch64/arm_neon.h
8855 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
8856 (vqtbx<1,2,3,4><q>_s8): Likewise.
8858 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8860 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
8861 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
8862 (add<mode>3): Add type attribute.
8863 (add<mode>3): Likewise.
8864 (usadd<mode>3): Likewise.
8865 (ssadd<mode>3): Likewise.
8866 (sub<mode>3): Likewise.
8867 (sub<mode>3): Likewise.
8868 (ussub<mode>3): Likewise.
8869 (sssub<mode>3): Likewise.
8870 (ssmulsa3): Likewise.
8871 (usmulusa3): Likewise.
8872 (arm_usatsihi): Likewise.
8874 (*movdi_vfp): Add types for all instructions.
8875 (*movdi_vfp_cortexa8): Likewise.
8876 (*movhf_vfp_neon): Likewise.
8877 (*movhf_vfp): Likewise.
8878 (*movdf_vfp): Likewise.
8879 (*thumb2_movdf_vfp): Likewise.
8880 (*thumb2_movdfcc_vfp): Likewise.
8881 * config/arm/arm.md: Add type attribute to all insn patterns.
8882 (*thumb1_adddi3): Add type attribute.
8883 (*arm_adddi3): Likewise.
8884 (*adddi_sesidi_di): Likewise.
8885 (*adddi_zesidi_di): Likewise.
8886 (*thumb1_addsi3): Likewise.
8887 (addsi3_compare0): Likewise.
8888 (*addsi3_compare0_scratch): Likewise.
8889 (*compare_negsi_si): Likewise.
8890 (cmpsi2_addneg): Likewise.
8891 (*addsi3_carryin_<optab>): Likewise.
8892 (*addsi3_carryin_alt2_<optab>): Likewise.
8893 (*addsi3_carryin_clobercc_<optab>): Likewise.
8894 (*subsi3_carryin): Likewise.
8895 (*subsi3_carryin_const): Likewise.
8896 (*subsi3_carryin_compare): Likewise.
8897 (*subsi3_carryin_compare_const): Likewise.
8898 (*arm_subdi3): Likewise.
8899 (*thumb_subdi3): Likewise.
8900 (*subdi_di_zesidi): Likewise.
8901 (*subdi_di_sesidi): Likewise.
8902 (*subdi_zesidi_di): Likewise.
8903 (*subdi_sesidi_di): Likewise.
8904 (*subdi_zesidi_ze): Likewise.
8905 (thumb1_subsi3_insn): Likewise.
8906 (*arm_subsi3_insn): Likewise.
8907 (*anddi3_insn): Likewise.
8908 (*anddi_zesidi_di): Likewise.
8909 (*anddi_sesdi_di): Likewise.
8910 (*ne_zeroextracts): Likewise.
8911 (*ne_zeroextracts): Likewise.
8912 (*ite_ne_zeroextr): Likewise.
8913 (*ite_ne_zeroextr): Likewise.
8914 (*anddi_notdi_di): Likewise.
8915 (*anddi_notzesidi): Likewise.
8916 (*anddi_notsesidi): Likewise.
8917 (andsi_notsi_si): Likewise.
8918 (thumb1_bicsi3): Likewise.
8919 (*iordi3_insn): Likewise.
8920 (*iordi_zesidi_di): Likewise.
8921 (*iordi_sesidi_di): Likewise.
8922 (*thumb1_iorsi3_insn): Likewise.
8923 (*xordi3_insn): Likewise.
8924 (*xordi_zesidi_di): Likewise.
8925 (*xordi_sesidi_di): Likewise.
8926 (*arm_xorsi3): Likewise.
8927 (*andsi_iorsi3_no): Likewise.
8928 (*smax_0): Likewise.
8929 (*smax_m1): Likewise.
8930 (*arm_smax_insn): Likewise.
8931 (*smin_0): Likewise.
8932 (*arm_smin_insn): Likewise.
8933 (*arm_umaxsi3): Likewise.
8934 (*arm_uminsi3): Likewise.
8935 (*minmax_arithsi): Likewise.
8936 (*minmax_arithsi_): Likewise.
8937 (*satsi_<SAT:code>): Likewise.
8938 (arm_ashldi3_1bit): Likewise.
8939 (arm_ashrdi3_1bit): Likewise.
8940 (arm_lshrdi3_1bit): Likewise.
8941 (*arm_negdi2): Likewise.
8942 (*thumb1_negdi2): Likewise.
8943 (*arm_negsi2): Likewise.
8944 (*thumb1_negsi2): Likewise.
8945 (*negdi_extendsid): Likewise.
8946 (*negdi_zero_extend): Likewise.
8947 (*arm_abssi2): Likewise.
8948 (*thumb1_abssi2): Likewise.
8949 (*arm_neg_abssi2): Likewise.
8950 (*thumb1_neg_abss): Likewise.
8951 (one_cmpldi2): Likewise.
8952 (extend<mode>di2): Likewise.
8953 (*compareqi_eq0): Likewise.
8954 (*arm_extendhisi2addsi): Likewise.
8955 (*arm_movdi): Likewise.
8956 (*thumb1_movdi_insn): Likewise.
8957 (*arm_movt): Likewise.
8958 (*thumb1_movsi_insn): Likewise.
8959 (pic_add_dot_plus_four): Likewise.
8960 (pic_add_dot_plus_eight): Likewise.
8961 (tls_load_dot_plus_eight): Likewise.
8962 (*thumb1_movhi_insn): Likewise.
8963 (*thumb1_movsf_insn): Likewise.
8964 (*movdf_soft_insn): Likewise.
8965 (*thumb_movdf_insn): Likewise.
8966 (cbranchsi4_insn): Likewise.
8967 (cbranchsi4_scratch): Likewise.
8968 (*negated_cbranchsi4): Likewise.
8969 (*tbit_cbranch): Likewise.
8970 (*tlobits_cbranch): Likewise.
8971 (*tstsi3_cbranch): Likewise.
8972 (*cbranchne_decr1): Likewise.
8973 (*addsi3_cbranch): Likewise.
8974 (*addsi3_cbranch_scratch): Likewise.
8975 (*arm_cmpdi_insn): Likewise.
8976 (*arm_cmpdi_unsig): Likewise.
8977 (*arm_cmpdi_zero): Likewise.
8978 (*thumb_cmpdi_zero): Likewise.
8979 (*deleted_compare): Likewise.
8980 (*mov_scc): Likewise.
8981 (*mov_negscc): Likewise.
8982 (*mov_notscc): Likewise.
8983 (*cstoresi_eq0_thumb1_insn): Likewise.
8984 (cstoresi_nltu_thumb1): Likewise.
8985 (cstoresi_ltu_thu): Likewise.
8986 (thumb1_addsi3_addgeu): Likewise.
8987 (*arm_jump): Likewise.
8988 (*thumb_jump): Likewise.
8989 (*check_arch2): Likewise.
8990 (arm_casesi_internal): Likewise.
8991 (thumb1_casesi_dispatch): Likewise.
8992 (*arm_indirect_jump): Likewise.
8993 (*thumb1_indirect_jump): Likewise.
8995 (*and_scc): Likewise.
8996 (*ior_scc): Likewise.
8997 (*compare_scc): Likewise.
8998 (*cond_move): Likewise.
8999 (*cond_arith): Likewise.
9000 (*cond_sub): Likewise.
9001 (*cmp_ite0): Likewise.
9002 (*cmp_ite1): Likewise.
9003 (*cmp_and): Likewise.
9004 (*cmp_ior): Likewise.
9005 (*ior_scc_scc): Likewise.
9006 (*ior_scc_scc_cmp): Likewise.
9007 (*and_scc_scc): Likewise.
9008 (*and_scc_scc_cmp): Likewise.
9009 (*and_scc_scc_nod): Likewise.
9010 (*negscc): Likewise.
9011 (movcond_addsi): Likewise.
9012 (movcond): Likewise.
9013 (*ifcompare_plus_move): Likewise.
9014 (*if_plus_move): Likewise.
9015 (*ifcompare_move_plus): Likewise.
9016 (*if_move_plus): Likewise.
9017 (*ifcompare_arith_arith): Likewise.
9018 (*if_arith_arith): Likewise.
9019 (*ifcompare_arith_move): Likewise.
9020 (*if_arith_move): Likewise.
9021 (*ifcompare_move_arith): Likewise.
9022 (*if_move_arith): Likewise.
9023 (*ifcompare_move_not): Likewise.
9024 (*if_move_not): Likewise.
9025 (*ifcompare_not_move): Likewise.
9026 (*if_not_move): Likewise.
9027 (*ifcompare_shift_move): Likewise.
9028 (*if_shift_move): Likewise.
9029 (*ifcompare_move_shift): Likewise.
9030 (*if_move_shift): Likewise.
9031 (*ifcompare_shift_shift): Likewise.
9032 (*ifcompare_not_arith): Likewise.
9033 (*ifcompare_arith_not): Likewise.
9034 (*if_arith_not): Likewise.
9035 (*ifcompare_neg_move): Likewise.
9036 (*if_neg_move): Likewise.
9037 (*ifcompare_move_neg): Likewise.
9038 (*if_move_neg): Likewise.
9039 (prologue_thumb1_interwork): Likewise.
9040 (*cond_move_not): Likewise.
9041 (*sign_extract_onebit): Likewise.
9042 (*not_signextract_onebit): Likewise.
9043 (stack_tie): Likewise.
9044 (align_4): Likewise.
9045 (align_8): Likewise.
9046 (consttable_end): Likewise.
9047 (consttable_1): Likewise.
9048 (consttable_2): Likewise.
9049 (consttable_4): Likewise.
9050 (consttable_8): Likewise.
9051 (consttable_16): Likewise.
9052 (*thumb1_tablejump): Likewise.
9053 (prefetch): Likewise.
9054 (force_register_use): Likewise.
9055 (thumb_eh_return): Likewise.
9056 (load_tp_hard): Likewise.
9057 (load_tp_soft): Likewise.
9058 (tlscall): Likewise.
9059 (*arm_movtas_ze): Likewise.
9060 (*arm_rev): Likewise.
9061 (*arm_revsh): Likewise.
9062 (*arm_rev16): Likewise.
9063 * config/arm/thumb2.md
9064 (*thumb2_smaxsi3): Likewise.
9065 (*thumb2_sminsi3): Likewise.
9066 (*thumb32_umaxsi3): Likewise.
9067 (*thumb2_uminsi3): Likewise.
9068 (*thumb2_negdi2): Likewise.
9069 (*thumb2_abssi2): Likewise.
9070 (*thumb2_neg_abss): Likewise.
9071 (*thumb2_movsi_insn): Likewise.
9072 (tls_load_dot_plus_four): Likewise.
9073 (*thumb2_movhi_insn): Likewise.
9074 (*thumb2_mov_scc): Likewise.
9075 (*thumb2_mov_negs): Likewise.
9076 (*thumb2_mov_negs): Likewise.
9077 (*thumb2_mov_nots): Likewise.
9078 (*thumb2_mov_nots): Likewise.
9079 (*thumb2_movsicc_): Likewise.
9080 (*thumb2_movsfcc_soft_insn): Likewise.
9081 (*thumb2_indirect_jump): Likewise.
9082 (*thumb2_and_scc): Likewise.
9083 (*thumb2_ior_scc): Likewise.
9084 (*thumb2_ior_scc_strict_it): Likewise.
9085 (*thumb2_cond_move): Likewise.
9086 (*thumb2_cond_arith): Likewise.
9087 (*thumb2_cond_ari): Likewise.
9088 (*thumb2_cond_sub): Likewise.
9089 (*thumb2_negscc): Likewise.
9090 (*thumb2_movcond): Likewise.
9091 (thumb2_casesi_internal): Likewise.
9092 (thumb2_casesi_internal_pic): Likewise.
9093 (*thumb2_alusi3_short): Likewise.
9094 (*thumb2_mov<mode>_shortim): Likewise.
9095 (*thumb2_addsi_short): Likewise.
9096 (*thumb2_subsi_short): Likewise.
9097 (thumb2_addsi3_compare0): Likewise.
9098 (*thumb2_cbz): Likewise.
9099 (*thumb2_cbnz): Likewise.
9100 (*thumb2_one_cmplsi2_short): Likewise.
9101 (*thumb2_negsi2_short): Likewise.
9102 (*orsi_notsi_si): Likewise.
9103 * config/arm/arm1020e.md: Update with new attributes.
9104 * config/arm/arm1026ejs.md: Update with new attributes.
9105 * config/arm/arm1136jfs.md: Update with new attributes.
9106 * config/arm/arm926ejs.md: Update with new attributes.
9107 * config/arm/cortex-a15.md: Update with new attributes.
9108 * config/arm/cortex-a5.md: Update with new attributes.
9109 * config/arm/cortex-a53.md: Update with new attributes.
9110 * config/arm/cortex-a7.md: Update with new attributes.
9111 * config/arm/cortex-a8.md: Update with new attributes.
9112 * config/arm/cortex-a9.md: Update with new attributes.
9113 * config/arm/cortex-m4.md: Update with new attributes.
9114 * config/arm/cortex-r4.md: Update with new attributes.
9115 * config/arm/fa526.md: Update with new attributes.
9116 * config/arm/fa606te.md: Update with new attributes.
9117 * config/arm/fa626te.md: Update with new attributes.
9118 * config/arm/fa726te.md: Update with new attributes.
9120 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9122 * config/aarch64/aarch64-simd.md
9123 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
9124 <vwx> iterator to ensure correct register choice.
9125 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
9126 (aarch64_sqdmull_n<mode>): Likewise.
9127 (aarch64_sqdmull2_n<mode>_internal): Likewise.
9128 * config/aarch64/arm_neon.h
9129 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
9130 (vml<as><q>_n_<su>16): Likewise.
9131 (vml<as>l_high_lane<q>_<su>16): Likewise.
9132 (vml<as>l_high_n_<su>16): Likewise.
9133 (vml<as>l_lane<q>_<su>16): Likewise.
9134 (vml<as>l_n_<su>16): Likewise.
9135 (vmul<q>_lane<q>_<su>16): Likewise.
9136 (vmul<q>_n_<su>16): Likewise.
9137 (vmull_lane<q>_<su>16): Likewise.
9138 (vmull_n_<su>16): Likewise.
9139 (vmull_high_lane<q>_<su>16): Likewise.
9140 (vmull_high_n_<su>16): Likewise.
9141 (vqrdmulh<q>_n_s16): Likewise.
9143 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
9145 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
9146 have the correct lane parameter.
9148 2013-09-06 Richard Biener <rguenther@suse.de>
9150 * cfganal.c (control_dependences::~control_dependences):
9151 Properly free all of the vector.
9153 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
9156 * config/i386/i386.c (ix86_conditional_register_usage):
9157 Proper initialize extended SSE registers.
9159 2013-09-06 Jan Hubicka <jh@suse.cz>
9161 PR tree-optimization/58311
9162 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
9164 2013-09-06 Jan Hubicka <jh@suse.cz>
9166 * Makefile.in (tree-sra.o): Update dependencies.
9167 * tree-sra.c: Include ipa-utils.h
9168 (scan_function): Use recursive_call_p.
9169 (has_caller_p): New function.
9170 (cgraph_for_node_and_aliases): Count also callers of aliases.
9172 2013-09-06 Jan Hubicka <jh@suse.cz>
9175 * cgraph.h (symtab_semantically_equivalent_p): Declare.
9176 * tree-tailcall.c: Include ipa-utils.h.
9177 (find_tail_calls): Use it.
9178 * ipa-pure-const.c (check_call): Likewise.
9179 * ipa-utils.c (recursive_call_p): New function.
9180 * ipa-utils.h (recursive_call_p): Dclare.
9181 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
9182 (symtab_semantically_equivalent_p): New function.
9183 * Makefile.in (tree-tailcall.o): Update dependencies.
9185 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9187 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
9190 2013-09-06 Richard Biener <rguenther@suse.de>
9192 * lto-streamer.h (lto_global_var_decls): Remove.
9193 * Makefile.in (OBJS): Remove lto-symtab.o.
9194 (lto-symtab.o): Remove.
9195 (GTFILES): Remove lto-symtab.c
9196 * lto-symtab.c: Move to lto/
9198 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9200 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
9201 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
9202 (UNSPEC_FPINT_RINT): New constant definitions.
9203 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
9204 definition with 2 attributes.
9205 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
9206 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
9209 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9211 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
9212 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
9213 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
9214 attribute to "z196_alone".
9215 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
9218 2013-09-06 Richard Biener <rguenther@suse.de>
9220 * basic-block.h (class control_dependences): New.
9221 * tree-ssa-dce.c (control_dependence_map): Remove.
9223 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
9224 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
9225 find_pdom, find_control_dependence, find_all_control_dependences):
9227 (mark_control_dependent_edges_necessary,
9228 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
9229 tree_dce_done, perform_tree_ssa_dce): Adjust.
9230 * cfganal.c (set_control_dependence_map_bit,
9231 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
9232 find_all_control_dependences): Move from tree-ssa-dce.c and
9233 implement as methods of control_dependences class.
9234 (control_dependences::control_dependences): New.
9235 (control_dependences::~control_dependences): Likewise.
9236 (control_dependences::get_edges_dependent_on): Likewise.
9237 (control_dependences::get_edge): Likewise.
9239 2013-09-04 Jan Hubicka <jh@suse.cz>
9241 * tree.c (types_same_for_odr): Drop overactive check.
9242 * ipa-devirt.c (hash_type_name): Likewise.
9244 2013-09-04 Jan Hubicka <jh@suse.cz>
9246 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
9247 (analyze_functions): ... here.
9249 2013-09-04 Jan Hubicka <jh@suse.cz>
9252 * cgraphunit.c (analyze_functions): Clear AUX fields
9253 after processing; initialize assembler name has.
9255 2013-09-05 Jeff Law <law@redhat.com>
9257 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
9258 from thread_around_empty_block. Record threading path into PATH.
9259 Recurse if threading through the initial block is successful.
9260 (thread_across_edge): Corresponding changes to slightly simplify.
9262 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9264 * config/aarch64/aarch64.md
9265 (type): Remove frecpe, frecps, frecpx.
9266 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
9267 fix to be a TARGET_SIMD instruction.
9268 (aarch64_frecps): Remove.
9269 * config/aarch64/aarch64-simd.md
9270 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
9271 (aarch64_frecps<mode>): Handle all float/vector of float modes.
9273 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9274 Sofiane Naci <sofiane.naci@arm.com>
9276 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
9277 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
9278 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
9279 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
9280 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
9281 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
9282 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
9283 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
9284 into "clz, "rbit". Rename "shift" to "shift_imm".
9285 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
9286 changes. Update for attribute changes all occurrences of arlo_* and
9288 * config/arm/arm-fixed.md: Update for attribute changes
9289 all occurrences of arlo_* types.
9290 * config/arm/thumb2.md: Update for attribute changes all occurrences
9292 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
9293 (cortexa7_older_only): Likewise.
9294 (cortexa7_younger): Likewise.
9295 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9296 (1020alu_shift_op): Likewise.
9297 (1020alu_shift_reg_op): Likewise.
9298 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9299 (alu_shift_op): Likewise.
9300 (alu_shift_reg_op): Likewise.
9301 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9302 (11_alu_shift_op): Likewise.
9303 (11_alu_shift_reg_op): Likewise.
9304 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9305 (9_alu_shift_reg_op): Likewise.
9306 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
9308 (cortex_a15_alu_shift): Likewise.
9309 (cortex_a15_alu_shift_reg): Likewise.
9310 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
9312 (cortex_a5_alu_shift): Likewise.
9313 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
9315 (cortex_a53_alu_shift): Likewise.
9316 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
9318 (cortex_a7_alu_reg): Likewise.
9319 (cortex_a7_alu_shift): Likewise.
9320 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
9322 (cortex_a8_alu_shift): Likewise.
9323 (cortex_a8_alu_shift_reg): Likewise.
9324 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9325 (cortex_a9_dp_shift): Likewise.
9326 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
9328 * config/arm/cortex-r4.md
9329 (cortex_r4_alu): Update for attribute changes.
9330 (cortex_r4_mov): Likewise.
9331 (cortex_r4_alu_shift_reg): Likewise.
9332 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9333 (526_alu_shift_op): Likewise.
9334 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9335 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9336 (626te_alu_shift_op): Likewise.
9337 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
9338 (726te_alu_shift_op): Likewise.
9339 (726te_alu_shift_reg_op): Likewise.
9340 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9341 (mp626_alu_shift_op): Likewise.
9342 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
9343 (pj4_alu_conds): Likewise.
9344 (pj4_shift): Likewise.
9345 (pj4_shift_conds): Likewise.
9346 (pj4_alu_shift): Likewise.
9347 (pj4_alu_shift_conds): Likewise.
9348 * config/aarch64/aarch64.md: Update for attribute change
9349 all occurrences of arlo_* and shift* types.
9351 2013-09-05 Mike Stump <mikestump@comcast.net>
9353 * tree.h: Move documentation for tree_function_decl to tree-core.h
9354 with the declaration.
9356 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
9359 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
9360 looking for widest mode.
9362 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
9364 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
9366 2013-09-05 Richard Biener <rguenther@suse.de>
9368 PR tree-optimization/58137
9369 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
9370 Do not create vectors of pointers.
9371 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
9372 types for the components of the vector initializer.
9373 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
9374 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
9376 2013-09-05 Martin Jambor <mjambor@suse.cz>
9378 * ipa-prop.c (remove_described_reference): Accept missing references,
9379 return false if that hppens, otherwise return true.
9380 (cgraph_node_for_jfunc): New function.
9381 (try_decrement_rdesc_refcount): Likewise.
9382 (try_make_edge_direct_simple_call): Use them.
9383 (ipa_edge_removal_hook): Remove references from rdescs.
9384 (ipa_edge_duplication_hook): Clone rdescs and their references
9385 when the new edge has the same caller as the old one.
9386 * cgraph.c (cgraph_resolve_speculation): Remove speculative
9387 reference before removing any edges.
9389 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
9391 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
9393 * thumb2.md (thumb2_storewb_parisi): New pattern.
9395 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
9397 * config/aarch64/aarch64-option-extensions.def: Add
9398 AARCH64_OPT_EXTENSION of 'crc'.
9399 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
9400 (AARCH64_ISA_CRC): Ditto.
9401 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
9402 description of the CRC extension.
9404 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
9406 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
9407 * config/rs6000/linux.h: Ditto.
9408 * alpha/linux.h: Ditto.
9409 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
9410 no_c99_libc_has_function.
9411 * config/c6x/uclinux-elf.h: Ditto.
9412 * config/lm32/uclinux-elf.h: Ditto.
9413 * config/m68k/uclinux.h: Ditto.
9414 * config/moxie/uclinux.h: Ditto.
9415 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
9416 (crisv32-*-linux*, cris-*-linux*): Ditto.
9417 * config/bfin/bfin.c: Include "tm_p.h".
9419 2013-09-05 Richard Biener <rguenther@suse.de>
9421 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
9422 check for a definition without a basic-block.
9424 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9425 Sofiane Naci <sofiane.naci@arm.com>
9427 * config/aarch64/aarch64.md
9428 (*movti_aarch64): Rename r_2_f and f_2_r.
9429 (*movsf_aarch64): Likewise.
9430 (*movdf_aarch64): Likewise.
9431 (*movtf_aarch64): Likewise.
9432 (aarch64_movdi_<mode>low): Likewise.
9433 (aarch64_movdi_<mode>high): Likewise.
9434 (aarch64_mov<mode>high_di): Likewise.
9435 (aarch64_mov<mode>low_di): Likewise.
9436 (aarch64_movtilow_tilow): Likewise.
9437 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
9438 values to config/arm/types.md
9439 (attribute "conds"): Update for attribute change.
9440 (anddi3_insn): Likewise.
9441 (iordi3_insn): Likewise.
9442 (xordi3_insn): Likewise.
9443 (one_cmpldi2): Likewise.
9444 * config/arm/types.md (type): Add Neon types.
9445 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
9446 use "type" attribute.
9447 (movmisalign<mode>_neon_store): Likewise.
9448 (movmisalign<mode>_neon_load): Likewise.
9449 (vec_set<mode>_internal): Likewise.
9450 (vec_setv2di_internal): Likewise.
9451 (vec_extract<mode>): Likewise.
9452 (vec_extractv2di): Likewise.
9453 (add<mode>3_neon): Likewise.
9454 (adddi3_neon): Likewise.
9455 (sub<mode>3_neon): Likewise.
9456 (subdi3_neon): Likewise.
9457 (mul<mode>3_neon): Likewise.
9458 (mul<mode>3add<mode>_neon): Likewise.
9459 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
9460 (fma<VCVTF:mode>4)): Likewise.
9461 (fma<VCVTF:mode>4_intrinsic): Likewise.
9462 (fmsub<VCVTF:mode>4)): Likewise.
9463 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
9464 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
9465 (ior<mode>3): Likewise.
9466 (and<mode>3): Likewise.
9467 (anddi3_neon): Likewise.
9468 (orn<mode>3_neon): Likewise.
9469 (orndi3_neon): Likewise.
9470 (bic<mode>3_neon): Likewise.
9471 (bicdi3_neon): Likewise.
9472 (xor<mode>3): Likewise.
9473 (one_cmpl<mode>2): Likewise.
9474 (abs<mode>2): Likewise.
9475 (neg<mode>2): Likewise.
9476 (umin<mode>3_neon): Likewise.
9477 (umax<mode>3_neon): Likewise.
9478 (smin<mode>3_neon): Likewise.
9479 (smax<mode>3_neon): Likewise.
9480 (vashl<mode>3): Likewise.
9481 (vashr<mode>3_imm): Likewise.
9482 (vlshr<mode>3_imm): Likewise.
9483 (ashl<mode>3_signed): Likewise.
9484 (ashl<mode>3_unsigned): Likewise.
9485 (neon_load_count): Likewise.
9486 (ashldi3_neon_noclobber): Likewise.
9487 (signed_shift_di3_neon): Likewise.
9488 (unsigned_shift_di3_neon): Likewise.
9489 (ashrdi3_neon_imm_noclobber): Likewise.
9490 (lshrdi3_neon_imm_noclobber): Likewise.
9491 (widen_ssum<mode>3): Likewise.
9492 (widen_usum<mode>3): Likewise.
9493 (quad_halves_<code>v4si): Likewise.
9494 (quad_halves_<code>v4sf): Likewise.
9495 (quad_halves_<code>v8hi): Likewise.
9496 (quad_halves_<code>v16qi): Likewise.
9497 (reduc_splus_v2di): Likewise.
9498 (neon_vpadd_internal<mode>): Likewise.
9499 (neon_vpsmin<mode>): Likewise.
9500 (neon_vpsmax<mode>): Likewise.
9501 (neon_vpumin<mode>): Likewise.
9502 (neon_vpumax<mode>): Likewise.
9503 (ss_add<mode>_neon): Likewise.
9504 (us_add<mode>_neon): Likewise.
9505 (ss_sub<mode>_neon): Likewise.
9506 (us_sub<mode>_neon): Likewise.
9507 (neon_vadd<mode>_unspec): Likewise.
9508 (neon_vaddl<mode>): Likewise.
9509 (neon_vaddw<mode>): Likewise.
9510 (neon_vhadd<mode>): Likewise.
9511 (neon_vqadd<mode>): Likewise.
9512 (neon_vaddhn<mode>): Likewise.
9513 (neon_vmul<mode>): Likewise.
9514 (neon_vmla<mode>): Likewise.
9515 (neon_vmlal<mode>): Likewise.
9516 (neon_vmls<mode>): Likewise.
9517 (neon_vmlsl<mode>): Likewise.
9518 (neon_vqdmulh<mode>): Likewise.
9519 (neon_vqdmlal<mode>): Likewise.
9520 (neon_vqdmlsl<mode>): Likewise.
9521 (neon_vmull<mode>): Likewise.
9522 (neon_vqdmull<mode>): Likewise.
9523 (neon_vsub<mode>_unspec): Likewise.
9524 (neon_vsubl<mode>): Likewise.
9525 (neon_vsubw<mode>): Likewise.
9526 (neon_vqsub<mode>): Likewise.
9527 (neon_vhsub<mode>): Likewise.
9528 (neon_vsubhn<mode>): Likewise.
9529 (neon_vceq<mode>): Likewise.
9530 (neon_vcge<mode>): Likewise.
9531 (neon_vcgeu<mode>): Likewise.
9532 (neon_vcgt<mode>): Likewise.
9533 (neon_vcgtu<mode>): Likewise.
9534 (neon_vcle<mode>): Likewise.
9535 (neon_vclt<mode>): Likewise.
9536 (neon_vcage<mode>): Likewise.
9537 (neon_vcagt<mode>): Likewise.
9538 (neon_vtst<mode>): Likewise.
9539 (neon_vabd<mode>): Likewise.
9540 (neon_vabdl<mode>): Likewise.
9541 (neon_vaba<mode>): Likewise.
9542 (neon_vabal<mode>): Likewise.
9543 (neon_vmax<mode>): Likewise.
9544 (neon_vmin<mode>): Likewise.
9545 (neon_vpaddl<mode>): Likewise.
9546 (neon_vpadal<mode>): Likewise.
9547 (neon_vpmax<mode>): Likewise.
9548 (neon_vpmin<mode>): Likewise.
9549 (neon_vrecps<mode>): Likewise.
9550 (neon_vrsqrts<mode>): Likewise.
9551 (neon_vqabs<mode>): Likewise.
9552 (neon_vqneg<mode>): Likewise.
9553 (neon_vcls<mode>): Likewise.
9554 (clz<mode>2): Likewise.
9555 (popcount<mode>2): Likewise.
9556 (neon_vrecpe): Likewise.
9557 (neon_vrsqrte): Likewise.
9558 (neon_vget_lane<mode>_sext_internal): Likewise.
9559 (neon_vget_lane<mode>_zext_internal): Likewise.
9560 (neon_vdup_n<mode>): Likewise.
9561 (neon_vdup_nv2di): Likewise.
9562 (neon_vdpu_lane<mode>_internal): Likewise.
9563 (neon_vswp<mode>): Likewise.
9564 (float<mode><V_cvtto>2): Likewise.
9565 (floatuns<mode><V_cvtto>2): Likewise.
9566 (fix_trunc<mode><V_cvtto>)2): Likewise
9567 (fixuns_trunc<mode><V_cvtto)2): Likewise.
9568 (neon_vcvt<mode>): Likewise.
9569 (neon_vcvtv4sfv4hf): Likewise.
9570 (neon_vcvtv4hfv4sf): Likewise.
9571 (neon_vcvt_n<mode>): Likewise.
9572 (neon_vmovn<mode>): Likewise.
9573 (neon_vqmovn<mode>): Likewise.
9574 (neon_vqmovun<mode>): Likewise.
9575 (neon_vmovl<mode>): Likewise.
9576 (neon_vmul_lane<mode>): Likewise.
9577 (neon_vmull_lane<mode>): Likewise.
9578 (neon_vqdmull_lane<mode>): Likewise.
9579 (neon_vqdmulh_lane<mode>): Likewise.
9580 (neon_vmla_lane<mode>): Likewise.
9581 (neon_vmlal_lane<mode>): Likewise.
9582 (neon_vqdmlal_lane<mode>): Likewise.
9583 (neon_vmls_lane<mode>): Likewise.
9584 (neon_vmlsl_lane<mode>): Likewise.
9585 (neon_vqdmlsl_lane<mode>): Likewise.
9586 (neon_vext<mode>): Likewise.
9587 (neon_vrev64<mode>): Likewise.
9588 (neon_vrev32<mode>): Likewise.
9589 (neon_vrev16<mode>): Likewise.
9590 (neon_vbsl<mode>_internal): Likewise.
9591 (neon_vshl<mode>): Likewise.
9592 (neon_vqshl<mode>): Likewise.
9593 (neon_vshr_n<mode>): Likewise.
9594 (neon_vshrn_n<mode>): Likewise.
9595 (neon_vqshrn_n<mode>): Likewise.
9596 (neon_vqshrun_n<mode>): Likewise.
9597 (neon_vshl_n<mode>): Likewise.
9598 (neon_vqshl_n<mode>): Likewise.
9599 (neon_vqshlu_n<mode>): Likewise.
9600 (neon_vshll_n<mode>): Likewise.
9601 (neon_vsra_n<mode>): Likewise.
9602 (neon_vsri_n<mode>): Likewise.
9603 (neon_vsli_n<mode>): Likewise.
9604 (neon_vtbl1v8qi): Likewise.
9605 (neon_vtbl2v8qi): Likewise.
9606 (neon_vtbl3v8qi): Likewise.
9607 (neon_vtbl4v8qi): Likewise.
9608 (neon_vtbx1v8qi): Likewise.
9609 (neon_vtbx2v8qi): Likewise.
9610 (neon_vtbx3v8qi): Likewise.
9611 (neon_vtbx4v8qi): Likewise.
9612 (neon_vtrn<mode>_internal): Likewise.
9613 (neon_vzip<mode>_internal): Likewise.
9614 (neon_vuzp<mode>_internal): Likewise.
9615 (neon_vld1<mode>): Likewise.
9616 (neon_vld1_lane<mode>): Likewise.
9617 (neon_vld1_dup<mode>): Likewise.
9618 (neon_vld1_dupv2di): Likewise.
9619 (neon_vst1<mode>): Likewise.
9620 (neon_vst1_lane<mode>): Likewise.
9621 (neon_vld2<mode>): Likewise.
9622 (neon_vld2_lane<mode>): Likewise.
9623 (neon_vld2_dup<mode>): Likewise.
9624 (neon_vst2<mode>): Likewise.
9625 (neon_vst2_lane<mode>): Likewise.
9626 (neon_vld3<mode>): Likewise.
9627 (neon_vld3qa<mode>): Likewise.
9628 (neon_vld3qb<mode>): Likewise.
9629 (neon_vld3_lane<mode>): Likewise.
9630 (neon_vld3_dup<mode>): Likewise.
9631 (neon_vst3<mode>): Likewise.
9632 (neon_vst3qa<mode>): Likewise.
9633 (neon_vst3qb<mode>): Likewise.
9634 (neon_vst3_lane<mode>): Likewise.
9635 (neon_vld4<mode>): Likewise.
9636 (neon_vld4qa<mode>): Likewise.
9637 (neon_vld4qb<mode>): Likewise.
9638 (neon_vld4_lane<mode>): Likewise.
9639 (neon_vld4_dup<mode>): Likewise.
9640 (neon_vst4<mode>): Likewise.
9641 (neon_vst4qa<mode>): Likewise.
9642 (neon_vst4qb<mode>): Likewise.
9643 (neon_vst4_lane<mode>): Likewise.
9644 (neon_vec_unpack<US>_lo_<mode>): Likewise.
9645 (neon_vec_unpack<US>_hi_<mode>): Likewise.
9646 (neon_vec_<US>mult_lo_<mode>): Likewise.
9647 (neon_vec_<US>mult_hi_<mode>): Likewise.
9648 (neon_vec_<US>shiftl_<mode>): Likewise.
9649 (neon_unpack<US>_<mode>): Likewise.
9650 (neon_vec_<US>mult_<mode>): Likewise.
9651 (vec_pack_trunc_<mode>): Likewise.
9652 (neon_vec_pack_trunk_<mode>): Likewise.
9653 (neon_vabd<mode>_2): Likewise.
9654 (neon_vabd<mode>_3): Likewise.
9655 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
9656 (thumb2_movsi_vfp): Likewise.
9657 (movdi_vfp): Likewise.
9658 (movdi_vfp_cortexa8): Likewise.
9659 (movhf_vfp_neon): Likewise.
9660 (movhf_vfp): Likewiwse.
9661 (movsf_vfp): Likewiwse.
9662 (thumb2_movsf_vfp): Likewiwse.
9663 (movdf_vfp): Likewise.
9664 (thumb2_movdf_vfp): Likewise.
9665 (movsfcc_vfp): Likewise.
9666 (thumb2_movsfcc_vfp): Likewise.
9667 (movdfcc_vfp): Likewise.
9668 (thumb2_movdfcc_vfp): Likewise.
9669 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
9670 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
9671 (v10_v2c): Likewise.
9672 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
9674 (cortex_a15_neon_int_2): Likewise.
9675 (cortex_a15_neon_int_3): Likewise.
9676 (cortex_a15_neon_int_4): Likewise.
9677 (cortex_a15_neon_int_5): Likewise.
9678 (cortex_a15_neon_vqneg_vqabs): Likewise.
9679 (cortex_a15_neon_vmov): Likewise.
9680 (cortex_a15_neon_vaba): Likewise.
9681 (cortex_a15_neon_vaba_qqq): Likewise.
9682 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9683 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9684 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
9685 scalar_64_32_long_scalar): Likewise.
9686 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9687 (cortex_a15_neon_mla_qqq_8_16): Likewise.
9688 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9689 lotype_qdd_64_32_long): Likewise.
9690 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9691 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9692 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
9693 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9694 (cortex_a15_neon_shift_1): Likewise.
9695 (cortex_a15_neon_shift_2): Likewise.
9696 (cortex_a15_neon_shift_3): Likewise.
9697 (cortex_a15_neon_vshl_ddd): Likewise.
9698 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9699 (cortex_a15_neon_vsra_vrsra): Likewise.
9700 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
9701 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
9702 (cortex_a15_neon_fp_vmul_ddd): Likewise.
9703 (cortex_a15_neon_fp_vmul_qqd): Likewise.
9704 (cortex_a15_neon_fp_vmla_ddd): Likewise.
9705 (cortex_a15_neon_fp_vmla_qqq): Likewise.
9706 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
9707 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
9708 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9709 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9710 (cortex_a15_neon_bp_simple): Likewise.
9711 (cortex_a15_neon_bp_2cycle): Likewise.
9712 (cortex_a15_neon_bp_3cycle): Likewise.
9713 (cortex_a15_neon_vld1_1_2_regs): Likewise.
9714 (cortex_a15_neon_vld1_3_4_regs): Likewise.
9715 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9716 (cortex_a15_neon_vld2_4_regs): Likewise.
9717 (cortex_a15_neon_vld3_vld4): Likewise.
9718 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9719 (cortex_a15_neon_vst1_3_4_regs): Likewise.
9720 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
9721 (cortex_a15_neon_vst3_vst4): Likewise.
9722 (cortex_a15_neon_vld1_vld2_lane): Likewise.
9723 (cortex_a15_neon_vld3_vld4_lane" 10
9724 (cortex_a15_neon_vst1_vst2_lane): Likewise.
9725 (cortex_a15_neon_vst3_vst4_lane): Likewise.
9726 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
9727 (cortex_a15_neon_ldm_2): Likewise.0
9728 (cortex_a15_neon_stm_2): Likewise.
9729 (cortex_a15_neon_mcr): Likewise.
9730 (cortex_a15_neon_mcr_2_mcrr): Likewise.
9731 (cortex_a15_neon_mrc): Likewise.
9732 (cortex_a15_neon_mrrc): Likewise.
9733 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9735 (cortex_a15_alu_shift): Likewise.
9736 (cortex_a15_alu_shift_reg): Likewise.
9737 (cortex_a15_mult32): Likewise.
9738 (cortex_a15_mult64): Likewise.
9739 (cortex_a15_block): Likewise.
9740 (cortex_a15_branch): Likewise.
9741 (cortex_a15_load1): Likewise.
9742 (cortex_a15_load3): Likewise.
9743 (cortex_a15_store1): Likewise.
9744 (cortex_a15_store3): Likewise.
9745 (cortex_a15_call): Likewise.
9746 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
9747 (cortex_a5_f2r): Likewise.
9748 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
9750 (cortex_a53_f2r): Likewise.
9751 * config/arm/cortex-a7.md
9752 (cortex_a7_branch): Update for attribute change.
9753 (cortex_a7_call): Likewise.
9754 (cortex_a7_alu_imm): Likewise.
9755 (cortex_a7_alu_reg): Likewise.
9756 (cortex_a7_alu_shift): Likewise.
9757 (cortex_a7_mul): Likewise.
9758 (cortex_a7_load1): Likewise.
9759 (cortex_a7_store1): Likewise.
9760 (cortex_a7_load2): Likewise.
9761 (cortex_a7_store2): Likewise.
9762 (cortex_a7_load3): Likewise.
9763 (cortex_a7_store3): Likewise.
9764 (cortex_a7_load4): Likewise.
9765 (cortex_a7_store4): Likewise.
9766 (cortex_a7_fpalu): Likewise.
9767 (cortex_a7_fconst): Likewise.
9768 (cortex_a7_fpmuls): Likewise.
9769 (cortex_a7_neon_mul): Likewise.
9770 (cortex_a7_fpmacs): Likewise.
9771 (cortex_a7_neon_mla: Likewise.
9772 (cortex_a7_fpmuld: Likewise.
9773 (cortex_a7_fpmacd: Likewise.
9774 (cortex_a7_fpfmad: Likewise.
9775 (cortex_a7_fdivs: Likewise.
9776 (cortex_a7_fdivd: Likewise.
9777 (cortex_a7_r2f: Likewise.
9778 (cortex_a7_f2r: Likewise.
9779 (cortex_a7_f_flags: Likewise.
9780 (cortex_a7_f_loads: Likewise.
9781 (cortex_a7_f_loadd: Likewise.
9782 (cortex_a7_f_stores: Likewise.
9783 (cortex_a7_f_stored: Likewise.
9784 (cortex_a7_neon): Likewise.
9785 * config/arm/cortex-a8-neon.md
9786 (cortex_a8_neon_mrc): Update for attribute change.
9787 (cortex_a8_neon_mrrc): Likewise.
9788 (cortex_a8_neon_int_1): Likewise.
9789 (cortex_a8_neon_int_2): Likewise.
9790 (cortex_a8_neon_int_3): Likewise.
9791 (cortex_a8_neon_int_4): Likewise.
9792 (cortex_a8_neon_int_5): Likewise.
9793 (cortex_a8_neon_vqneg_vqabs): Likewise.
9794 (cortex_a8_neon_vmov): Likewise.
9795 (cortex_a8_neon_vaba): Likewise.
9796 (cortex_a8_neon_vaba_qqq): Likewise.
9797 (cortex_a8_neon_vsma): Likewise.
9798 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9799 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9800 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
9801 long_scalar): Likewise.
9802 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9803 (cortex_a8_neon_mla_qqq_8_16): Likewise.
9804 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9805 long_scalar_qdd_64_32_long): Likewise.
9806 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9807 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9808 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
9809 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9810 (cortex_a8_neon_shift_1): Likewise.
9811 (cortex_a8_neon_shift_2): Likewise.
9812 (cortex_a8_neon_shift_3): Likewise.
9813 (cortex_a8_neon_vshl_ddd): Likewise.
9814 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9815 (cortex_a8_neon_vsra_vrsra): Likewise.
9816 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
9817 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
9818 (cortex_a8_neon_fp_vsum): Likewise.
9819 (cortex_a8_neon_fp_vmul_ddd): Likewise.
9820 (cortex_a8_neon_fp_vmul_qqd): Likewise.
9821 (cortex_a8_neon_fp_vmla_ddd): Likewise.
9822 (cortex_a8_neon_fp_vmla_qqq): Likewise.
9823 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
9824 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
9825 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9826 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9827 (cortex_a8_neon_bp_simple): Likewise.
9828 (cortex_a8_neon_bp_2cycle): Likewise.
9829 (cortex_a8_neon_bp_3cycle): Likewise.
9830 (cortex_a8_neon_ldr): Likewise.
9831 (cortex_a8_neon_str): Likewise.
9832 (cortex_a8_neon_vld1_1_2_regs): Likewise.
9833 (cortex_a8_neon_vld1_3_4_regs): Likewise.
9834 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9835 (cortex_a8_neon_vld2_4_regs): Likewise.
9836 (cortex_a8_neon_vld3_vld4): Likewise.
9837 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9838 (cortex_a8_neon_vst1_3_4_regs): Likewise.
9839 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
9840 (cortex_a8_neon_vst3_vst4): Likewise.
9841 (cortex_a8_neon_vld1_vld2_lane): Likewise.
9842 (cortex_a8_neon_vld3_vld4_lane): Likewise.
9843 (cortex_a8_neon_vst1_vst2_lane): Likewise.
9844 (cortex_a8_neon_vst3_vst4_lane): Likewise.
9845 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
9846 (cortex_a8_neon_mcr): Likewise.
9847 (cortex_a8_neon_mcr_2_mcrr): Likewise.
9848 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
9849 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
9851 (ca9_neon_mrrc): Likewise.
9852 (cortex_a9_neon_int_1): Likewise.
9853 (cortex_a9_neon_int_2): Likewise.
9854 (cortex_a9_neon_int_3): Likewise.
9855 (cortex_a9_neon_int_4): Likewise.
9856 (cortex_a9_neon_int_5): Likewise.
9857 (cortex_a9_neon_vqneg_vqabs): Likewise.
9858 (cortex_a9_neon_vmov): Likewise.
9859 (cortex_a9_neon_vaba): Likewise.
9860 (cortex_a9_neon_vaba_qqq): Likewise.
9861 (cortex_a9_neon_vsma): Likewise.
9862 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9863 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9864 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
9865 long_scalar): Likewise.
9866 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9867 (cortex_a9_neon_mla_qqq_8_16): Likewise.
9868 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9869 long_scalar_qdd_64_32_long): Likewise.
9870 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9871 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9872 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
9873 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9874 (cortex_a9_neon_shift_1): Likewise.
9875 (cortex_a9_neon_shift_2): Likewise.
9876 (cortex_a9_neon_shift_3): Likewise.
9877 (cortex_a9_neon_vshl_ddd): Likewise.
9878 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9879 (cortex_a9_neon_vsra_vrsra): Likewise.
9880 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
9881 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
9882 (cortex_a9_neon_fp_vsum): Likewise.
9883 (cortex_a9_neon_fp_vmul_ddd): Likewise.
9884 (cortex_a9_neon_fp_vmul_qqd): Likewise.
9885 (cortex_a9_neon_fp_vmla_ddd): Likewise.
9886 (cortex_a9_neon_fp_vmla_qqq): Likewise.
9887 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
9888 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
9889 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9890 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9891 (cortex_a9_neon_bp_simple): Likewise.
9892 (cortex_a9_neon_bp_2cycle): Likewise.
9893 (cortex_a9_neon_bp_3cycle): Likewise.
9894 (cortex_a9_neon_ldr): Likewise.
9895 (cortex_a9_neon_str): Likewise.
9896 (cortex_a9_neon_vld1_1_2_regs): Likewise.
9897 (cortex_a9_neon_vld1_3_4_regs): Likewise.
9898 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9899 (cortex_a9_neon_vld2_4_regs): Likewise.
9900 (cortex_a9_neon_vld3_vld4): Likewise.
9901 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9902 (cortex_a9_neon_vst1_3_4_regs): Likewise.
9903 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
9904 (cortex_a9_neon_vst3_vst4): Likewise.
9905 (cortex_a9_neon_vld1_vld2_lane): Likewise.
9906 (cortex_a9_neon_vld3_vld4_lane): Likewise.
9907 (cortex_a9_neon_vst1_vst2_lane): Likewise.
9908 (cortex_a9_neon_vst3_vst4_lane): Likewise.
9909 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
9910 (cortex_a9_neon_mcr): Likewise.
9911 (cortex_a9_neon_mcr_2_mcrr): Likewise.
9912 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
9913 (cortex_a9_fps): Likewise.
9914 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
9916 (cortex_m4_fmuls): Likewise.
9917 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
9919 (cortex_r4_mrc): Likewise.
9920 * config/arm/iterators.md: Update comment referring to neon_type.
9921 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
9922 (iwmmxt_movsi_insn): Likewise.
9923 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
9925 (pj4_core_to_vfp): Likewise.
9926 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
9928 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
9929 (vfp_fstore): Likewise.
9930 * doc/md.texi: Change references to neon_type to refer to type.
9932 2013-09-04 Dodji Seketeli <dodji@redhat.com>
9934 * tree.h (DECL_BUILT_IN): Fix typo in comment.
9936 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
9938 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
9941 2013-09-04 Easwaran Raman <eraman@google.com>
9944 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
9945 (build_and_add_sum): Use it.
9946 (appears_later_in_bb): Simplify code.
9948 2013-09-04 Teresa Johnson <tejohnson@google.com>
9950 * dumpfile.c (dump_finish): Don't close stderr/stdout.
9952 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
9954 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
9956 2013-09-04 Jan Hubicka <jh@suse.cz>
9958 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
9959 * ipa-devirt.c: Include diganostic.h
9960 (odr_type_d): Add types and types_set.
9961 (hash_type_name): Work for types with vtables during LTO.
9962 (odr_hasher::remove): Fix comment; destroy types_set.
9963 (add_type_duplicate): New function,
9964 (get_odr_type): Use it.
9965 (dump_type_inheritance_graph): Dump type duplicates.
9966 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
9968 * tree.c (types_same_for_odr): Give exact answers on types with
9971 2013-09-04 Dodji Seketeli <dodji@redhat.com>
9973 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
9974 explaining their differences.
9976 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
9978 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
9980 2013-09-03 Jeff Law <law@redhat.com>
9982 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
9983 when not threading through a joiner block. Pass joiner/no joiner
9984 state to register_jump_thread.
9985 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
9986 state from argument rather than implying on path length.
9987 Dump the entire jump thread path into debugging dump.
9988 * tree-flow.h (register_jump_thread): Update prototype.
9990 2013-08-29 Xinliang David Li <davidxl@google.com>
9992 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
9993 Remove a trivial gcc_assert.
9995 2013-08-29 Xinliang David Li <davidxl@google.com>
9997 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
9998 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
9999 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10000 Delay base decl alignment adjustment.
10001 * tree-vectorizer.c (vect_destroy_datarefs): New function.
10002 * tree-vectorizer.h: New data structure.
10003 (set_dr_misalignment): New function.
10004 (dr_misalignment): Ditto.
10005 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
10006 (vectorizable_load): Ditto.
10007 (ensure_base_align): New function.
10008 (vectorize_loops): Add dbg_cnt support.
10009 (execute_vect_slp): Ditto.
10010 * dbgcnt.def: New debug counter.
10011 * Makefile: New dependency.
10013 2013-09-03 Meador Inge <meadori@codesourcery.com>
10017 2013-08-30 Meador Inge <meadori@codesourcery.com>
10019 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10021 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
10023 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
10024 function descriptor.
10026 2013-09-03 Richard Biener <rguenther@suse.de>
10028 * tree-affine.c (add_elt_to_tree): Fix association issue,
10029 avoid useless converts and make sure to always return a
10030 properly typed result.
10032 2013-09-03 Richard Biener <rguenther@suse.de>
10034 PR middle-end/57656
10035 * fold-const.c (negate_expr_p): Fix division case.
10036 (negate_expr): Likewise.
10038 2013-09-03 Richard Biener <rguenther@suse.de>
10041 * tree-streamer-out.c: Include tm.h.
10042 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
10044 2013-09-03 Jan Hubicka <jh@suse.cz>
10046 * tree-profile.c (tree_profiling): Cleanup CFG when done.
10048 2013-09-03 Alan Modra <amodra@gmail.com>
10050 * config.gcc (powerpc*-*-linux*): Add support for little-endian
10051 multilibs to big-endian target and vice versa.
10052 * config/rs6000/t-linux64: Use := assignment on all vars.
10053 (MULTILIB_EXTRA_OPTS): Remove fPIC.
10054 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
10055 * config/rs6000/t-linux64le: New file.
10056 * config/rs6000/t-linux64bele: New file.
10057 * config/rs6000/t-linux64lebe: New file.
10059 2013-09-02 Jan Hubicka <jh@suse.cz>
10061 * ipa-inline-transform.c (inline_transform): Do not
10062 optimize_inline_calls when not optimizing.
10064 2013-09-02 Jan Hubicka <jh@suse.cz>
10066 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
10067 duplicated nodes for assembler names.
10068 * symtab.c (symtab_unregister_node): Do not attempt to unlink
10069 hard registers from assembler name hash.
10071 2013-09-02 Jan Hubicka <jh@suse.cz>
10073 * ipa-split.c (execute_split_functions): Split externally visible
10074 functions called once.
10076 2013-09-02 Martin Jambor <mjambor@suse.cz>
10079 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
10080 linked list. When finding the correct duplicate, also consider also
10081 the caller in additon to its inlined_to node.
10083 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
10085 * config/aarch64/aarch64-simd-builtins.def
10086 (dup_lane_scalar): Remove.
10087 * config/aarch64/aarch64-simd.md
10088 (aarch64_simd_dup): Add 'w->w' alternative.
10089 (aarch64_dup_lane<mode>): Allow for VALL.
10090 (aarch64_dup_lane_scalar<mode>): Remove.
10091 (aarch64_dup_lane_<vswap_width_name><mode>): New.
10092 (aarch64_get_lane_signed<mode>): Add w->w altenative.
10093 (aarch64_get_lane_unsigned<mode>): Likewise.
10094 (aarch64_get_lane<mode>): Likewise.
10095 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
10096 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
10097 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
10098 (VCON): Change container of V2SF.
10099 (vswap_width_name): Likewise.
10100 * config/aarch64/arm_neon.h
10101 (__aarch64_vdup_lane_any): New.
10102 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10103 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
10104 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10106 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
10108 PR middle-end/56382
10109 * expr.c (emit_move_complex): Do not move complex FP values as parts if
10110 the source or the destination is a single hard register.
10112 2013-09-02 Richard Biener <rguenther@suse.de>
10114 PR middle-end/57511
10115 * tree-scalar-evolution.c (instantiate_scev_name): Allow
10118 2013-09-02 Richard Biener <rguenther@suse.de>
10120 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
10121 arithmetic to sizetype.
10123 2013-09-02 Bin Cheng <bin.cheng@arm.com>
10125 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
10126 Find auto-increment use both before and after candidate.
10128 2013-09-02 Marek Polacek <polacek@redhat.com>
10130 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
10132 2013-09-01 Jan Hubicka <jh@suse.cz>
10134 * Makefile.in: Add ipa-profile.o
10135 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
10136 * cgraph.c (struct cgraph_propagate_frequency_data,
10137 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
10138 ipa-profile.c; replace cgraph_ by ipa_ prefix.
10139 * cgraph.h (cgraph_propagate_frequency): Remove.
10140 * ipa-inline-analysis.c: Include ipa-utils.h;
10141 drop duplicated cfgloop.h.
10142 (inline_update_callee_summaries): Update.
10143 * ipa-profile.c: New file.
10144 * ipa-utils.h (ipa_propagate_frequency): Declare.
10145 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
10146 data-streamer.h, value-prof.h.
10147 (symtab_remove_unreachable_nodes): Update profile.
10148 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
10149 account_time_size, cmp_counts, dump_histogram,
10150 ipa_profile_generate_summary, ipa_profile_write_summary,
10151 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
10152 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
10153 Move to ipa-profile.c.
10155 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
10157 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
10159 2013-09-01 Jan Hubicka <jh@suse.cz>
10161 * common.opt (fdevirtualize-speculatively): New function.
10162 * invoke.texi (fdevirtualize-speculatively): Document.
10163 * ipa-devirt.c: Include ipa-inline.h
10164 (likely_target_p): New function.
10165 (ipa_devirt): New function.
10166 (gate_ipa_devirt): New function.
10167 (pass_data_ipa_devirt): New static var.
10168 (pass_ipa_devirt): Likewise.
10169 (make_pass_ipa_devirt): New function.
10170 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
10171 (common_handle_option): Disable devirtualization when
10172 value range profiling is available.
10173 * passes.def (pass_ipa_devirt): Add.
10174 * timever.def (TV_IPA_DEVIRT): New timevar.
10175 * tree-pass.h (make_pass_ipa_devirt):
10177 2013-09-01 Iain Sandoe <iain@codesourcery.com>
10179 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
10180 include sanitize(undefined).
10182 2013-08-31 Diego Novillo <dnovillo@google.com>
10184 * Makefile.in (TREE_CORE_H): Define.
10186 (GTFILES): Add tree-core.h.
10187 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
10189 * tree-core.h: New file.
10190 Move all data structures, enum, typedefs, global
10191 declarations and constants from ...
10192 * tree.h: ... here.
10194 2013-08-31 Jan Hubicka <jh@suse.cz>
10196 * bulitins.c (expand_builtin): Do not early exit for gcov
10197 instrumented functions.
10199 2013-08-31 Marek Polacek <polacek@redhat.com>
10201 * ubsan.c: Include tm_p.h.
10203 2013-08-31 Jan Hubicka <jh@suse.cz>
10205 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
10208 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
10209 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
10210 * tree-cfg.c (verify_gimple_label): ... here.
10211 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
10212 (ipa_merge_profiles): New function.
10213 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
10214 (lto_input_function_body): Likewise.
10215 * ipa-utils.h (ipa_merge_profiles): Declare.
10216 * lto-streamer.h (lto_input_function_body): Update prototype.
10217 (emit_label_in_global_context_p): Remove.
10218 * lto-symtab.c: Include ipa-utils.h
10219 (lto_cgraph_replace_node): Use ipa_merge_profiles.
10221 2013-08-31 Jan Hubicka <jh@suse.cz>
10223 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
10225 2013-08-31 Jan Hubicka <jh@suse.cz>
10227 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
10229 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
10231 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
10232 "cmp" RTX before signed_comparison_operator check to account
10233 for "code" changes.
10235 2013-08-30 Jan Hubicka <jh@suse.cz>
10237 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
10238 (detect_type_change_1): Rename to ...
10239 (detect_type_change): ... this one; early return on non-polymorphic
10241 (detect_type_change_ssa): Add comp_type parameter; update
10242 use of detect_type_change.
10243 (compute_complex_assign_jump_func): Add param_type parameter;
10244 update use of detect_type_change_ssa.
10245 (compute_complex_ancestor_jump_func): Likewise.
10246 (ipa_get_callee_param_type): New function.
10247 (ipa_compute_jump_functions_for_edge): Compute parameter type;
10248 update calls to the jump function computation functions.
10250 2013-08-30 Teresa Johnson <tejohnson@google.com>
10251 Steven Bosscher <steven@gcc.gnu.org>
10253 * cfgrtl.c (fixup_new_cold_bb): New routine.
10254 (commit_edge_insertions): Invoke fixup_partitions.
10255 (find_partition_fixes): New routine.
10256 (fixup_partitions): Ditto.
10257 (verify_hot_cold_block_grouping): Update comments.
10258 (rtl_verify_edges): Invoke find_partition_fixes.
10259 (rtl_verify_bb_pointers): Update comments.
10260 (rtl_verify_bb_layout): Ditto.
10261 * basic-block.h (probably_never_executed_edge_p): Declare.
10262 (fixup_partitions): Ditto.
10263 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
10264 * bb-reorder.c (sanitize_hot_paths): New function.
10265 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
10266 sanitize_hot_paths.
10267 * predict.c (probably_never_executed_edge_p): New routine.
10268 * cfg.c (check_bb_profile): Add partition insanity warnings.
10270 2013-08-30 Meador Inge <meadori@codesourcery.com>
10272 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10274 2013-08-30 Marek Polacek <polacek@redhat.com>
10276 * Makefile.in (ubsan.o): Add.
10277 (c-family/c-ubsan.o): Add.
10278 (builtins.o): Add ubsan.h dependency.
10279 * ubsan.h: New file.
10280 * ubsan.c: New file.
10281 * common.opt: Add -fsanitize=undefined option.
10282 (flag_sanitize): Add variable.
10283 (fsanitize=): Add option. Add Driver.
10284 (fsanitize=thread): Remove option.
10285 (fsanitize=address): Likewise.
10286 (static-libubsan): New option.
10287 * doc/invoke.texi: Document the new flag and -static-libubsan.
10288 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
10289 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
10290 * builtin-attrs.def (ATTR_COLD): Define.
10291 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10292 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
10293 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
10294 * flag-types.h (sanitize_code): New enum.
10295 * opts.c (common_handle_option): Parse command line arguments
10296 of -fsanitize=. Add -fsanitize=unreachable option.
10297 * varasm.c (get_variable_section): Adjust.
10298 (assemble_noswitch_variable): Likewise.
10299 (assemble_variable): Likewise.
10300 (output_constant_def_contents): Likewise.
10301 (categorize_decl_for_section): Likewise.
10302 (place_block_symbol): Likewise.
10303 (output_object_block): Likewise.
10304 * builtins.def: Likewise.
10305 * toplev.c (compile_file): Likewise.
10306 (process_options): Likewise.
10307 * cppbuiltin.c: Likewise.
10308 * tsan.c (tsan_pass): Likewise.
10309 (tsan_gate): Likewise.
10310 (tsan_gate_O0): Likewise.
10311 * cfgexpand.c (partition_stack_vars): Likewise.
10312 (expand_stack_vars): Likewise.
10313 (defer_stack_allocation): Likewise.
10314 (expand_used_vars): Likewise.
10315 * cfgcleanup.c (old_insns_match_p): Likewise.
10316 * asan.c (asan_finish_file): Likewise.
10317 (asan_instrument): Likewise.
10318 (gate_asan): Likewise.
10319 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
10320 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10321 (asan_global_struct): Use pointer_sized_int_node instead
10322 calling build_nonstandard_integer_type.
10323 (initialize_sanitizer_builtins): Likewise.
10324 (asan_finish_file): Likewise.
10325 * gcc.c: Document %{%:function(args):X}.
10326 (static_spec_functions): Add sanitize.
10327 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
10328 store funcval != NULL there.
10329 (do_spec_1): Adjust handle_spec_function caller.
10330 (handle_braces): Allow %:function(args) as condition.
10331 (sanitize_spec_function): New function.
10332 (ADD_STATIC_LIBUBSAN_LIBS): Define.
10333 (LIBUBSAN_SPEC): Likewise.
10334 (LIBUBSAN_EARLY_SPEC): Likewise.
10335 (SANITIZER_SPEC): Handle libubsan.
10336 (SANITIZER_EARLY_SPEC): Likewise.
10337 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
10338 instead of fsanitize=address.
10339 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
10340 instead of fsanitize=address*.
10341 * builtins.c: Include ubsan.h.
10342 (fold_builtin_0): Instrument __builtin_unreachable.
10343 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
10344 instead of flag_asan.
10345 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
10346 (pointer_sized_int_node): Define.
10347 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
10349 2013-08-30 Mike Stump <mikestump@comcast.net>
10351 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
10354 2013-08-29 Jan Hubicka <jh@suse.cz>
10356 * cgraph.c (cgraph_function_body_availability): Handle weakref
10358 * passes.def: Remove pass_fixup_cfg.
10359 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
10360 track when we need to remove functions.
10361 (gate_ipa_inline): Execute inlining always; add comment why.
10362 (pass_data_ipa_inline): Remove TODO_remove_functions.
10363 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
10364 do not produce summaries.
10365 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
10366 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
10367 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
10368 constant pool and vtable.
10370 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
10372 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
10373 New arm_neon.h's internal macros to specify 64-bit constants.
10374 Avoid using stdint.h's macros.
10376 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
10378 * recog.c (verify_changes): Verify that changes[i].old is non-zero
10379 before applying REG_P.
10381 2013-08-30 Jakub Jelinek <jakub@redhat.com>
10383 PR tree-optimization/58277
10384 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
10385 after seeing too many stmts with vdef in between dombb and current
10386 bb, invalidate everything.
10388 2013-08-30 Richard Biener <rguenther@suse.de>
10390 * fold-const.c (fold_single_bit_test): Fix overflow test.
10392 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
10394 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
10395 and which can live in a register, always retrieve the value on entry.
10396 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
10397 passed by invisible reference specially.
10398 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
10399 (vt_add_function_parameter): Correctly deal with a parameter passed by
10400 invisible reference.
10402 2013-08-30 Jan Hubicka <jh@suse.cz>
10404 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
10406 2013-08-30 Richard Biener <rguenther@suse.de>
10408 PR tree-optimization/58228
10409 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
10410 allow invariant loads in nested loop vectorization.
10412 2013-08-30 Richard Biener <rguenther@suse.de>
10414 PR tree-optimization/58223
10415 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
10416 (has_anti_or_output_dependence): ... this and adjust to also
10417 look for output dependences.
10418 (mark_nodes_having_upstream_mem_writes): Adjust.
10419 (rdg_flag_uses): Likewise.
10421 2013-08-30 Richard Biener <rguenther@suse.de>
10423 PR tree-optimization/58010
10424 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
10425 assert that we have a loop-closed PHI.
10427 2013-08-29 Jan Hubicka <jh@suse.cz>
10429 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
10430 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
10431 * lto-section-in.c (lto_free_function_in_decl_state): New function.
10432 (lto_free_function_in_decl_state_for_node): New function.
10434 2013-08-29 Xinliang David Li <davidxl@google.com>
10436 * loop-unroll.c (report_unroll_peel): Minor message change.
10437 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
10438 Emit alignment peeling message with default -fopt-info.
10439 (vect_loop_versioning): Emit loop version info message.
10440 * tree-vectorizer.c (vectorize_loops): Minor message change.
10441 (execute_vect_slp): Ditto.
10443 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
10445 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
10446 of the clone from the DECL_NAME of the original function.
10448 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
10450 * passes.c (register_pass): Add overload.
10451 * tree-pass.h (register_pass): Forward declare it. Add comment.
10453 2013-08-29 Jan Hubicka <jh@suse.cz>
10455 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
10456 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
10457 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
10458 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10459 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
10460 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
10461 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10462 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
10464 2013-08-29 Teresa Johnson <tejohnson@google.com>
10466 * dumpfile.c (dump_loc): Output column number.
10467 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
10468 * doc/invoke.texi: Document optall -fopt-info flag.
10469 * profile.c (read_profile_edge_counts): Use new dump framework.
10470 (compute_branch_probabilities): Ditto.
10471 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
10472 when pass not in any opt group.
10473 * pass_manager.h (pass_manager::get_pass_profile): New method.
10474 * value-prof.c (check_counter): Use new dump framework.
10475 (check_ic_target): Ditto.
10476 * coverage.c (get_coverage_counts): Ditto.
10477 (coverage_init): Setup new dump framework.
10479 2013-08-29 Richard Biener <rguenther@suse.de>
10481 PR tree-optimization/58246
10482 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
10483 handle the dominance check inside a basic-block.
10485 2013-08-29 Richard Biener <rguenther@suse.de>
10487 PR middle-end/57287
10488 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
10489 of default defs that appear in abnormal PHI nodes.
10491 2013-08-29 Richard Biener <rguenther@suse.de>
10493 PR tree-optimization/57685
10494 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
10495 single-use operands to avoid exponential complexity.
10497 2013-08-28 Dehao Chen <dehao@google.com>
10499 * ipa-inline.c (edge_badness): Fix integer underflow.
10501 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
10503 * gtm-builtins.def (_ITM_free): Declare leaf.
10505 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10508 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
10509 (*tls_local_dynamic_base_64_largepic): Likewise.
10510 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
10511 Remove predicate from call operand.
10512 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
10513 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
10515 2013-08-28 Jeff Law <law@redhat.com>
10517 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
10518 checks for the number of predecessors and successors allowed.
10519 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
10520 which require copying a joiner block if there is a request which
10521 is a subpath that requires no joiner block copying.
10523 2013-08-28 Jan Hubicka <jh@suse.cz>
10525 * lto-streamer-out.c (DFS_write_tree_body): Drop
10526 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
10527 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
10528 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
10529 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
10530 Do not read DECL_ERROR_ISSUED.
10531 (unpack_ts_decl_with_vis_value_fields): Do not read
10533 (lto_input_ts_binfo_tree_pointers): Do not read
10534 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
10535 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
10536 write DECL_ERROR_ISSUED..
10537 (pack_ts_decl_with_vis_value_fields): Do not write
10539 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
10540 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
10541 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
10542 * tree.h (tree_decl_common): Update comment.
10543 (DECL_ERROR_ISSUED): Remove.
10545 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10547 PR middle-end/58257
10548 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
10550 2013-08-28 Jan Hubicka <jh@suse.cz>
10552 * builtins.def (free): Declare leaf.
10554 2013-08-27 David Malcolm <dmalcolm@redhat.com>
10556 * gdbhooks.py: New.
10557 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
10558 * configure: Regenerate.
10560 2013-08-27 Martin Jambor <mjambor@suse.cz>
10562 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
10563 (ipa_ancestor_jf_data): Likewise.
10564 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
10565 (ipa_get_jf_pass_through_type_preserved): New function.
10566 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
10567 (ipa_get_jf_ancestor_type_preserved): New function.
10568 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
10569 (ipa_get_jf_ancestor_result): Likewise.
10570 (propagate_vals_accross_pass_through): Use
10571 ipa_get_jf_pass_through_result to do all the value mappings.
10572 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
10573 type_preserved flag.
10574 (ipa_set_jf_cst_copy): New function.
10575 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
10576 (ipa_set_jf_arith_pass_through): Likewise.
10577 (ipa_set_ancestor_jf): Likewise.
10578 (compute_complex_assign_jump_func): Set type_preserved instead of
10580 (ipa_compute_jump_functions_for_edge): Likewise.
10581 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
10582 (update_jump_functions_after_inlining): Update type_preserved.
10583 Explicitely create jump functions when combining one with pass_through.
10584 (ipa_write_jump_function): Stream the type_preserved flags.
10585 (ipa_read_jump_function): Likewise.
10587 2013-08-27 Jakub Jelinek <jakub@redhat.com>
10588 Aldy Hernandez <aldyh@redhat.com>
10590 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
10591 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
10592 * function.h (struct function): Add has_force_vect_loops and
10594 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
10595 * gimple.c (gimple_build_omp_critical): Add KIND argument and
10597 * gimple.def: Update CLAUSES comments.
10598 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
10599 (gimple_build_omp_for): Add argument to prototype.
10600 (gimple_omp_for_kind): New.
10601 (gimple_omp_for_set_kind): New.
10602 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
10603 GOVD_DATA_SHARE_CLASS.
10604 (enum omp_region_type): Add ORT_SIMD.
10605 (gimple_add_tmp_var): Handle ORT_SIMD.
10606 (gimplify_var_or_parm_decl): Same.
10607 (is_gimple_stmt): Same.
10608 (omp_firstprivatize_variable): Same.
10609 (omp_add_variable): Only use splay_tree_insert if lookup failed.
10610 (omp_notice_variable): Handle ORT_SIMD.
10611 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
10612 (omp_check_private): Handle ORT_SIMD.
10613 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10614 OMP_CLAUSE_SAFELEN.
10615 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
10616 Handle OMP_CLAUSE_LASTPRIVATE.
10617 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10618 OMP_CLAUSE_SAFELEN.
10619 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
10620 (gimplify_expr): Handle OMP_SIMD.
10621 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
10622 (expand_GOMP_SIMD_VF): New.
10623 (expand_GOMP_SIMD_LAST_LANE): New.
10624 * internal-fn.def (GOMP_SIMD_LANE): New.
10625 (GOMP_SIMD_VF): New.
10626 (GOMP_SIMD_LAST_LANE): New.
10627 * omp-low.c: Include target.h.
10628 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
10629 OMP_CLAUSE_SAFELEN.
10630 (check_omp_nesting_restrictions): Same.
10632 (lower_rec_simd_input_clauses): New.
10633 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
10635 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
10636 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
10637 (expand_omp_build_assign): New.
10638 (expand_omp_for_init_counts): New.
10639 (expand_omp_for_init_vars): New.
10640 (extract_omp_for_update_vars): New.
10641 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
10642 and rewrite accordingly.
10643 (expand_omp_simd): New.
10644 (expand_omp_for): Use expand_omp_simd.
10645 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
10646 (lower_omp_for): Do not lower the body.
10647 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
10648 in their own loops.
10649 * tree-flow.h (find_omp_clause): Remove prototype.
10650 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
10651 forcing vectorization of the loop, or if flag_tree_vectorize.
10652 (gate_tree_if_conversion): Similarly.
10653 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
10654 gimple_build_omp_for.
10655 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
10656 * tree-parloops (create_parallel_loop): Pass kind argument to
10657 gimple_build_omp_for.
10658 * tree-pretty-print.c (dump_omp_clause): Add cases for
10659 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
10660 OMP_CLAUSE__SIMDUID_.
10661 (dump_generic_node): Handle OMP_SIMD.
10662 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
10663 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
10664 unroll OMP_SIMD loops here.
10665 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
10666 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
10668 (vect_analyze_data_refs): Handle simd loops.
10669 * tree-vect-loop.c (vectorizable_live_operation): Handle
10671 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
10672 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
10673 (vectorizable_load): Same.
10674 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
10675 (struct simduid_to_vf): New.
10676 (simduid_to_vf::hash): New.
10677 (simduid_to-vf::equal): New.
10678 (struct simd_array_to_simduid): New.
10679 (simd_array_to_simduid::hash): New.
10680 (simd_array_to_simduid::equal): New.
10681 (adjust_simduid_builtins): New.
10682 (struct note_simd_array_uses_struct): New.
10683 (note_simd_array_uses_cb): New.
10684 (note_simd_array_uses): New.
10685 (vectorize_loops): Handle simd hints and adjust simd builtins
10687 * tree-vectorizer.h (struct _stmt_vec_info): Add
10688 simd_lane_access_p field.
10689 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
10690 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
10691 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
10692 (omp_clause_code_name): Same.
10693 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
10694 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
10695 * tree.def (OMP_SIMD): New entry.
10696 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
10697 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
10698 (OMP_CLAUSE_DECL): Adjust range for new clauses.
10699 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
10700 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
10701 (OMP_CLAUSE_LINEAR_STEP): New.
10702 (OMP_CLAUSE_SAFELEN_EXPR): New.
10703 (OMP_CLAUSE__SIMDUID__DECL): New.
10704 (find_omp_clause): New prototype.
10706 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
10708 * config/i386/driver-i386.c (host_detect_local_cpu): Update
10709 Haswell processor detection.
10711 2013-08-27 Christian Widmer <shadow@umbrox.de>
10714 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
10715 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
10716 AVX2 capable processors.
10718 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
10720 * config/aarch64/arm_neon.h: Replace all inline asm implementations
10721 of vget_low_* with implementations in terms of other intrinsics.
10723 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
10725 PR middle-end/57219
10726 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
10727 values to -1, 0 and 1.
10729 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
10731 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
10732 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
10733 (<optab><mode>3_insn): Remove.
10734 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
10735 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
10736 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
10737 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
10738 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
10739 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
10740 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
10741 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
10742 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
10743 (ror<mode>3_insn): Likewise.
10744 * config/aarch64/predicates.md (aarch64_simd_register): New.
10746 2013-08-27 Richard Biener <rguenther@suse.de>
10748 PR tree-optimization/57521
10749 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
10750 one edge is non-critical.
10751 (find_phi_replacement_condition): Make sure to use a non-critical
10752 edge. Cleanup and remove old bug workarounds.
10753 (bb_postdominates_preds): Remove.
10754 (if_convertible_loop_p_1): Do not compute post-dominators.
10755 (combine_blocks): Do not free post-dominators.
10756 (main_tree_if_conversion): Likewise.
10757 (pass_data_if_conversion): Add TODO_verify_ssa.
10759 2013-08-27 DJ Delorie <dj@redhat.com>
10761 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
10763 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
10765 * function.c (assign_parm_find_data_types): Set passed_mode and
10766 nominal_mode to the TYPE_MODE of nominal_type for the built
10767 pointer type in case of the struct-pass-by-reference.
10769 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10771 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
10772 (UINT16_TYPE): Change default to "unsigned int".
10774 * config/avr/avr.opt (mfract-convert-truncate): New option.
10775 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
10776 is set, round negative fractional integers according to n1169
10777 when converting to integer types.
10779 2013-08-26 Jan Hubicka <jh@suse.cz>
10781 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
10782 methods can not be called indirectly when their address is not taken.
10784 2013-08-26 Jan Hubicka <jh@suse.cz>
10786 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
10789 2013-08-26 Jan Hubicka <jh@suse.cz>
10791 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
10794 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10796 * reload.c (find_valid_class): Allow classes that do not include
10797 FIRST_PSEUDO_REGISTER - 1.
10799 2013-08-26 Jan Hubicka <jh@suse.cz>
10801 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
10802 fix edge count/frequency when speculation failed; fix type check
10803 for the direct call.
10805 2013-08-26 Jan Hubicka <jh@suse.cz>
10807 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
10809 2013-08-26 Jan Hubicka <jh@suse.cz>
10811 * ipa-inline-transform.c (inline_transform): Be ready for basic block
10812 to be changed by edge redirection.
10814 2013-08-26 Jan Hubicka <jh@suse.cz>
10816 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
10817 formating; add sanity check.
10818 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
10819 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
10820 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
10821 (resolve_noninline_speculation): Update callee keys, too.
10823 2013-08-26 Jan Hubicka <jh@suse.cz>
10825 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
10826 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
10828 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10830 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
10833 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
10835 * config/i386/i386.c (ix86_debug_options): Remove prototype.
10836 (x86_64_elf_select_section): Ditto.
10837 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
10839 (ix86_pass_by_reference): Ditto.
10840 (output_set_got): Ditto.
10841 (ix86_unary_operator_ok): Ditto.
10842 (ix86_expand_builtin): Ditto.
10844 2013-08-23 Jan Hubicka <jh@suse.cz>
10846 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
10848 2013-08-23 Jan Hubicka <jh@suse.cz>
10850 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
10851 (tree_decl_with_vis): Add FINAL field.
10853 2013-08-23 Jeff Law <law@redhat.com>
10855 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
10856 the debugging dump when the expression is fully redundant.
10858 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
10860 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
10861 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
10862 * pretty-print.c (pp_formatted_text_data): Likewise.
10863 (pp_write_text_to_stream): Likewise.
10864 (pp_write_text_as_dot_label_to_stream): Likewise.
10865 (pp_append_r): Likewise.
10866 (pp_format): Likewise.
10867 (pp_flush): Likewise.
10868 (pp_clear_output_area): Likewise.
10869 (pp_append_text): Likewise.
10870 (pp_formatted_text): Likewise.
10871 (pp_remaining_character_count_for_line): Likewise.
10872 (pp_newline): Likewise.
10873 (pp_character): Likewise.
10874 (output_buffer::~output_buffer): Define.
10875 (pretty_printer::~pretty_printer): Destruct output buffer.
10876 * pretty-print.h (output_buffer::~output_buffer): Declare.
10877 (pretty_printer::~pretty_printer): Declare virtual.
10879 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
10882 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
10883 HOST_WIDE_INT_M1U): New macros.
10884 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
10885 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
10886 unsigned -1 for lshift.
10887 * cse.c (cse_insn): Likewise.
10888 * double-int.c (rshift_double, lshift_double): Likewise.
10889 * builtins.c (fold_builtin_bitop): Likewise.
10890 * combine.c (force_to_mode): Likewise.
10891 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
10892 * simplify-rtx.c (simplify_const_unary_operation,
10893 simplify_const_binary_operation): Likewise.
10894 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
10895 check_va_list_escapes): Likewise.
10896 * rtlanal.c (nonzero_bits1): Likewise.
10897 * expmed.c (expand_smod_pow2): Likewise.
10898 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
10900 2013-08-23 Jan Hubicka <jh@suse.cz>
10902 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
10903 as having address taken.
10905 2013-08-23 Jan Hubicka <jh@suse.cz>
10907 * ipa-utils.h (method_class_type): Declare.
10908 * ipa-devirt.c (method_class_type): Export.
10910 * cgraphunit.c (analyze_functions): Do basic devirtualization;
10911 do not walk base classes of anonymous types.
10913 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
10915 PR rtl-optimization/58220
10916 PR regression/58221
10917 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
10918 handle SEQUENCE insns properly.
10920 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
10922 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
10924 (pp_newline_and_indent): Likewise.
10925 (pp_separate_with): Likewise.
10926 * pretty-print.c (pp_newline_and_flush): Define.
10927 (pp_newline_and_indent): Likewise.
10928 (pp_separate_with): Likewise.
10930 2013-08-23 Jakub Jelinek <jakub@redhat.com>
10933 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
10934 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
10936 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
10938 * config/i386/predicates.md (ext_sse_reg_operand): New.
10939 * config/i386/i386.md (*movti_internal): Use
10940 predicate to determine if EVEX is needed.
10941 (*movsi_internal): Ditto.
10942 (*movdf_internal): Ditto.
10943 (*movsf_internal): Ditto.
10944 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
10946 2013-08-23 Jakub Jelinek <jakub@redhat.com>
10948 PR tree-optimization/58209
10949 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
10950 (find_tail_calls): Give up for pointer result types if m is non-NULL.
10951 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
10952 emit POINTER_PLUS_EXPR.
10953 (create_tailcall_accumulator): For pointer result type accumulate in
10956 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
10958 * configure.ac: Add backslashes missing from the last change.
10959 * configure: Regenerate.
10961 2013-08-22 Jan Hubicka <jh@susue.cz>
10963 * ipa.c (function_and_variable_visibility): First remember function
10964 was global and then make it local.
10966 2013-08-22 Julian Brown <julian@codesourcery.com>
10968 * configure.ac: Add aarch64 to list of arches which use "nop" in
10970 * configure: Regenerate.
10972 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10974 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
10975 gnu_libc_has_function.
10976 * config/s390/tpf.h: Likewise.
10978 2013-08-22 Jan Hubicka <jh@susue.cz>
10980 * timevar.c (validate_phases): Add cast.
10982 2013-08-22 Jan Hubicka <jh@susue.cz>
10984 * timevar.c (validate_phases): Use size_t for memory.
10985 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
10987 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
10989 * pretty-print.h (output_buffer::output_buffer): Declare.
10990 (pretty_printer::pretty_printer): Likewise.
10991 (pp_construct): Remove.
10992 * pretty-print.c (output_buffer::output_buffer): Define.
10993 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
10994 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
10995 (print_gimple_expr): Likewise.
10996 (print_gimple_seq): Likewise.
10997 (gimple_dump_bb): Likewise.
10998 * sched-vis.c (dump_value_slim): Likewise.
10999 (dump_insn_slim): Likewise.
11000 (dump_rtl_slim): Likewise.
11001 (str_pattern_slim): Likewise.
11002 * tree-mudflap.c (mf_varname_tree): Likewise.
11003 * graph.c (print_graph_cfg): Likewise.
11004 (start_graph_dump): Likewise.
11005 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
11007 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
11008 pretty printer initialization.
11009 * coretypes.h (diagnostic_context): Remove superflous type alias
11011 (pretty_printer): Likewise. Declare directly as a class.
11012 (pretty_print_info): Remove declaration as class.
11013 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
11014 and pp_clear_output_area.
11015 (asan_add_global): Likewise.
11017 2013-08-22 Jan Hubicka <jh@suse.cz>
11019 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
11020 * ipa-utils.h (update_type_inheritance_graph): Declare.
11021 (possible_polymorphic_call_target_p): Declare.
11022 (possible_polymorphic_call_target_p): New.
11023 * ipa-devirt.c: Update toplevel comments.
11024 (cached_polymorphic_call_targets): Move up.
11025 (odr_type_d): Move ID down.
11026 (polymorphic_type_binfo_p): Update comment.
11027 (odr_hasher::remove): Likewise;
11028 (get_odr_type): Set anonymous_namespace.
11029 (dump_odr_type): Dump it.
11030 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
11031 (maybe_record_node): Record node in cached_polymorphic_call_targets.
11032 (record_binfo): Add comment.
11033 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
11035 (devirt_node_removal_hook): Do not iCE when cache is freed.
11036 (possible_polymorphic_call_target_p): New predicate.
11037 (update_type_inheritance_graph): New function.
11039 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
11040 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
11041 Sergey Lega <sergey.s.lega@intel.com>
11042 Anna Tikhonova <anna.tikhonova@intel.com>
11043 Ilya Tocar <ilya.tocar@intel.com>
11044 Andrey Turetskiy <andrey.turetskiy@intel.com>
11045 Ilya Verbin <ilya.verbin@intel.com>
11046 Kirill Yukhin <kirill.yukhin@intel.com>
11047 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11049 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
11050 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
11051 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
11052 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
11053 (OPTION_MASK_ISA_AVX2_UNSET): Update.
11054 (OPTION_MASK_ISA_AVX512F_UNSET): New.
11055 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
11056 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
11057 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
11058 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
11059 OPT_mavx512pf, OPT_mavx512er cases.
11060 * config/i386/constraints.md (v): New constraint.
11061 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
11062 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
11063 (bit_AVX512CD): New.
11064 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11065 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
11066 * config/i386/i386-c.c (ix86_target_macros_internal):
11067 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
11069 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
11070 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
11071 * config/i386/i386.c (regclass_map, dbx_register_map)
11072 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
11073 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
11074 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
11075 -mavx512pf options.
11076 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
11077 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
11078 -mavx512pf options. Fix formatting.
11079 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
11080 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
11081 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
11082 -mavx512cd, -mavx512pf options.
11083 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
11084 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
11085 (ix86_preferred_output_reload_class): Replace SSE_REGS with
11087 (ix86_hard_regno_mode_ok): Support 512-bit registers.
11088 (ix86_set_reg_reg_cost): Ditto.
11089 (x86_order_regs_for_local_alloc): Ditto.
11090 (MAX_VECT_LEN): Extend to 64-byte.
11091 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
11092 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
11093 (TARGET_AVX512ER, TARGET_AVX512CD): New.
11094 (BIGGEST_ALIGNMENT): Extend to 512-bits.
11095 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
11096 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
11097 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
11098 (SSE_REG_MODE_P): Support new modes.
11099 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
11100 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
11101 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
11102 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
11103 (REG_CLASS_CONTENTS): Add new registers.
11104 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
11105 (EXT_REX_SSE_REGNO_P): New.
11106 (HI_REGISTER_NAMES): Add new registers.
11107 * config/i386/i386.md: Define constants for new registers.
11108 (mode): Add new 512-bit modes.
11109 (prefix): Support evex prefix.
11110 (isa): Support avx512f, noavx512f, fma_avx512f.
11111 (ssemodesuffix): Add new 512-bit modes.
11113 (*movxi_internal_avx512f): Ditto.
11114 (*movdi_internal): Replace constraint "x" with the new constraint "v".
11116 (*movsi_internal): Likewise.
11117 (*movdf_internal): Likewise.
11118 (*movsf_internal): Likewise.
11119 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
11120 (<code><mode>3): Likewise.
11121 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
11123 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
11124 with the new constraint "v".
11125 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
11127 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
11128 with the new constraint "v".
11129 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
11130 (<sse2>_storedqu<avxsizesuffix>): Likewise.
11131 (*<plusminus_insn><mode>3): Likewise.
11132 (<sse>_vm<plusminus_insn><mode>3): Likewise.
11133 (*mul<mode>3): Likewise.
11134 (<sse>_vmmul<mode>3): Likewise.
11135 (<sse>_div<mode>3): Likewise.
11136 (<sse>_vmdiv<mode>3): Likewise.
11137 (<sse>_sqrt<mode>2): Likewise.
11138 (<sse>_vmsqrt<mode>2): Likewise.
11139 (*<code><mode>3_finite): Likewise.
11140 (*<code><mode>3) <smaxmin>: Likewise.
11141 (<sse>_vm<code><mode>3): Likewise.
11142 (*<code><mode>3) <any_logic>: Likewise.
11143 (*fma_fmadd_<mode>): Likewise.
11144 (*fma_fmsub_<mode>): Likewise.
11145 (*fma_fnmadd_<mode>): Likewise.
11146 (*fma_fnmsub_<mode>): Likewise.
11147 (*fma_fmaddsub_<mode>): Likewise.
11148 (*fma_fmsubadd_<mode>): Likewise.
11149 (*fmai_fmadd_<mode>): Likewise.
11150 (*fmai_fmsub_<mode>): Likewise.
11151 (*fmai_fnmadd_<mode>): Likewise.
11152 (*fmai_fnmsub_<mode>): Likewise.
11153 (sse_cvtsi2ss): Likewise.
11154 (sse_cvtsi2ssq): Likewise.
11155 (sse_cvtss2si): Likewise.
11156 (sse_cvtss2si_2): Likewise.
11157 (sse_cvtss2siq): Likewise.
11158 (sse_cvtss2siq_2): Likewise.
11159 (sse_cvttss2si): Likewise.
11160 (sse_cvtss2siq_2): Likewise.
11161 (float<sseintvecmodelower><mode>2): Likewise.
11162 (sse2_cvtsd2si_2): Likewise.
11163 (sse2_cvtsd2siq_2): Likewise.
11164 (*<plusminus_insn><mode>3): Likewise.
11165 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
11166 (*<sse4_1_avx2>_mul<mode>3): Likewise.
11167 (ashr<mode>3): Likewise.
11168 (<shift_insn><mode>3): Likewise.
11169 (avx2_<code><mode>3): Likewise.
11170 (*avx2_<code><mode>3): Likewise.
11171 (*andnot<mode>3): Likewise.
11172 (*<code><mode>3) <any_logic>: Likewise.
11173 (abs<mode>2): Likewise.
11174 (avx2_permvar<mode>): Likewise.
11175 (avx2_perm<mode>_1): Likewise.
11176 (*avx_vpermilp<mode>): Likewise.
11177 (avx_vpermilvar<mode>3): Likewise.
11178 (avx2_ashrv<mode>): Likewise.
11179 (avx2_<shift_insn>v<mode>): Likewise.
11180 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
11182 * doc/rtl.texi: Document XImode.
11184 2013-08-21 Jeff Law <law@redhat.com>
11186 * tree-flow.h (register_jump_thread): Pass vector of edges
11187 instead of each important edge.
11188 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
11189 thread path into a vector and pass that to register_jump_thread.
11190 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
11191 passed in edge vector to the current 3-edge form.
11194 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11196 * dce.c (fini_dce): Call df_analyze again just in case
11197 delete_unmarked_insns removed anything.
11199 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
11201 * reload.h (struct reg_equivs): Rename to ..
11202 (struct reg_equivs_s): .. this.
11204 2013-08-20 Martin Liska <marxin.liska@gmail.com>
11206 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
11208 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11210 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
11212 2013-08-21 Jeff Law <law@redhat.com>
11214 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
11215 simplify assignments too. If the RHS collapses to a singleton
11216 range, then return the value for the range.
11218 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
11220 * config/i386/sse.md (V16): Rename to...
11222 (mov<mode>): Update iterator name.
11223 (*mov<mode>_internal): Ditto.
11224 (push<mode>1): Ditto.
11225 (movmisalign<mode>): Ditto.
11227 2013-08-20 Jan Hubicka <jh@suse.cz>
11230 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
11231 entry for direct edges.
11232 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
11234 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11236 Revert my last two changes, r201865 and r201864:
11239 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11241 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11242 instances can own GC refs.
11244 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11245 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11246 (gcc::context::gt_pch_nx): Likewise.
11247 (gcc::context::gt_pch_nx): Likewise.
11248 * ggc.h (gt_ggc_mx <T>): New.
11249 (gt_pch_nx_with_op <T>): New.
11250 (gt_pch_nx <T>): New.
11251 * passes.c (opt_pass::gt_ggc_mx): New.
11252 (opt_pass::gt_pch_nx): New.
11253 (opt_pass::gt_pch_nx_with_op): New.
11254 (pass_manager::gt_ggc_mx): New.
11255 (pass_manager::gt_pch_nx): New.
11256 (pass_manager::gt_pch_nx_with_op): New.
11257 (pass_manager::operator new): Use
11258 ggc_internal_cleared_alloc_stat rather than xcalloc.
11259 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11260 (pass_manager::gt_ggc_mx): New.
11261 (pass_manager::gt_pch_nx): New.
11262 (pass_manager::gt_pch_nx_with_op): New.
11263 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11264 (opt_pass::operator new): New.
11265 (opt_pass::gt_ggc_mx): New.
11266 (opt_pass::gt_pch_nx): New.
11267 (opt_pass::gt_pch_nx_with_op): New.
11270 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11272 * Makefile.in (GTFILES): Add context.h.
11273 * context.c (gcc::context::operator new): New.
11274 (gcc::context::gt_ggc_mx): New.
11275 (gcc::context::gt_pch_nx): New.
11276 (gcc::context::gt_pch_nx): New.
11277 * context.h (gcc::context): Add GTY((user)) marking.
11278 (gcc::context::operator new): New.
11279 (gcc::context::gt_ggc_mx): New.
11280 (gcc::context::gt_pch_nx): New.
11281 (gcc::context::gt_pch_nx): New.
11282 (g): Add GTY marking.
11283 (gt_ggc_mx (gcc::context *)): New.
11284 (gt_pch_nx (gcc::context *)): New.
11285 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11286 void *cookie)): New.
11287 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11289 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11291 * dce.c (fini_dce): Call df_analyze again just in case
11292 delete_unmarked_insns removed anything.
11294 2013-08-20 Teresa Johnson <tejohnson@google.com>
11296 PR rtl-optimizations/57451
11297 * final.c (reemit_insn_block_notes): Prevent lexical blocks
11298 from crossing split section boundaries.
11300 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11302 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
11303 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
11304 with MULTLIB_DEFAULTS.
11306 2013-08-20 Nick Clifton <nickc@redhat.com>
11308 * target.def (narrow_volatile_bitfield): Note that the default
11309 value is false, not !TARGET_STRICT_ALIGN.
11310 * doc/tm.texi: Regenerate.
11312 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
11314 Fix LIB_SPEC for systems without libpthread.
11316 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
11317 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
11319 * config/i386/linux-common.h: Likewise.
11320 * config/mips/linux-common.h: Likewise.
11322 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11324 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
11327 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11329 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11330 instances can own GC refs.
11332 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11333 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11334 (gcc::context::gt_pch_nx): Likewise.
11335 (gcc::context::gt_pch_nx): Likewise.
11336 * ggc.h (gt_ggc_mx <T>): New.
11337 (gt_pch_nx_with_op <T>): New.
11338 (gt_pch_nx <T>): New.
11339 * passes.c (opt_pass::gt_ggc_mx): New.
11340 (opt_pass::gt_pch_nx): New.
11341 (opt_pass::gt_pch_nx_with_op): New.
11342 (pass_manager::gt_ggc_mx): New.
11343 (pass_manager::gt_pch_nx): New.
11344 (pass_manager::gt_pch_nx_with_op): New.
11345 (pass_manager::operator new): Use
11346 ggc_internal_cleared_alloc_stat rather than xcalloc.
11347 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11348 (pass_manager::gt_ggc_mx): New.
11349 (pass_manager::gt_pch_nx): New.
11350 (pass_manager::gt_pch_nx_with_op): New.
11351 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11352 (opt_pass::operator new): New.
11353 (opt_pass::gt_ggc_mx): New.
11354 (opt_pass::gt_pch_nx): New.
11355 (opt_pass::gt_pch_nx_with_op): New.
11357 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11359 * Makefile.in (GTFILES): Add context.h.
11360 * context.c (gcc::context::operator new): New.
11361 (gcc::context::gt_ggc_mx): New.
11362 (gcc::context::gt_pch_nx): New.
11363 (gcc::context::gt_pch_nx): New.
11364 * context.h (gcc::context): Add GTY((user)) marking.
11365 (gcc::context::operator new): New.
11366 (gcc::context::gt_ggc_mx): New.
11367 (gcc::context::gt_pch_nx): New.
11368 (gcc::context::gt_pch_nx): New.
11369 (g): Add GTY marking.
11370 (gt_ggc_mx (gcc::context *)): New.
11371 (gt_pch_nx (gcc::context *)): New.
11372 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11373 void *cookie)): New.
11374 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11376 2013-08-20 Alan Modra <amodra@gmail.com>
11379 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
11380 (rs6000_emit_epilogue): Likewise.
11382 2013-08-19 Dehao Chen <dehao@google.com>
11384 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
11386 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
11387 Jakub Jelinek <jakub@redhat.com>
11389 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
11390 (BUILT_IN_FABSD64): Likewise.
11391 (BUILT_IN_FABSD128): Likewise.
11392 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
11393 (fold_builtin_1): Likewise.
11394 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
11395 destination and source operands.
11396 (*abstd2_fpr): Likewise.
11397 (*nabstd2_fpr): Likewise.
11399 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
11401 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
11404 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
11406 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
11409 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
11411 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
11412 * builtins.c (default_libc_has_function): New.
11413 (gnu_libc_has_function): Ditto.
11414 (no_c99_libc_has_function): Ditto.
11415 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
11416 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
11417 (fold_builtin_sincos): Likewise.
11418 (fold_builtin_cexp): Likewise.
11419 * builtins.def (DEF_C94_BUILTIN): Likewise.
11420 (DEF_C99_BUILTIN): Likewise.
11421 (DEF_C99_C90RES_BUILTIN): Likewise.
11422 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
11423 definitions to using this define.
11424 * config/darwin-protos.h (darwin_libc_has_function): New.
11425 * config/darwin.c (darwin_libc_has_function): Ditto.
11426 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
11427 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
11428 * config/darwin.h: Ditto.
11429 * config/elfos.h: Ditto.
11430 * config/freebsd.h: Ditto.
11431 * config/i386/cygming.h: Ditto.
11432 * config/i386/djgpp.h: Ditto.
11433 * config/i386/i386-interix.h: Ditto.
11434 * config/microblaze/microblaze.h: Ditto.
11435 * config/mmix/mmix.h: Ditto.
11436 * config/gnu-user.h: Ditto.
11437 * config/ia64/hpux.h: Ditto.
11438 * config/pa/pa-hpux.h: Ditto.
11439 * config/pdp11/pdp11.h: Ditto.
11440 * config/picochip/picochip.h: Ditto.
11441 * config/linux.h: Ditto.
11442 * config/netbsd.h: Ditto.
11443 * config/openbsd.h: Ditto.
11444 * config/rs6000/aix43.h: Ditto.
11445 * config/rs6000/aix51.h: Ditto.
11446 * config/rs6000/aix52.h: Ditto.
11447 * config/rs6000/aix53.h: Ditto.
11448 * config/rs6000/aix61.h: Ditto.
11449 * config/rs6000/darwin.h: Ditto.
11450 * config/rs6000/linux.h: Ditto.
11451 * config/rs6000/linux64.h: Ditto.
11452 * config/s390/tpf.h: Ditto.
11453 * config/sol2-10.h: Ditto.
11454 * config/sol2.h: Ditto.
11455 * config/vms/vms.h: Ditto.
11456 * config/vxworks.h: Ditto.
11457 * config/linux-android.c (linux_android_libc_has_function):
11458 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
11459 * config/linux-protos.h (linux_android_libc_has_function):
11461 * config/i386/i386.c (ix86_libc_has_function): New.
11462 * config/i386/i386-protos.h
11463 (ix86_libc_has_function): New declaration.
11464 * config/i386/i386.md
11465 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
11466 ("isinf<mode>2): Likewise.
11467 * convert.c (convert_to_integer): Using new target hook
11468 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
11469 TARGET_C99_FUNCTIONS.
11470 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
11471 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
11472 * coretypes.h (function_class): New enum for different
11473 classes of functions.
11474 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
11475 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
11476 (TARGET_HAS_SINCOS): Likewise.
11477 (TARGET_LIBC_HAS_FUNCTION): New.
11478 * doc/tm.texi: Regenerated.
11479 * targhooks.h (default_libc_has_function): New declaration.
11480 (no_c99_libc_has_function): Ditto.
11481 (gnu_libc_has_function): Ditto.
11482 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
11483 and TARGET_HAS_SINCOS.
11485 2013-08-18 Jan Hubicka <jh@suse.cz>
11487 * Makeifle-in (ipa-devirt.o): New.
11488 (GTFILES): Add ipa-utils.h and ipa-devirt.c
11489 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
11490 (analyze_functions): Look into possible targets of polymorphic call.
11491 * dumpfile.c (dump_files): Add type-inheritance dump.
11492 * dumpfile.h (TDI_inheritance): New.
11493 * ipa-devirt.c: New file.
11494 * ipa-utils.h (odr_type_d): Forward declare.
11495 (odr_type): New type.
11496 (build_type_inheritance_graph): Declare.
11497 (possible_polymorphic_call_targets): Declare and introduce inline
11498 variant when only edge is pased.
11499 (dump_possible_polymorphic_call_targets): Likewise.
11500 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
11501 * tree.c (type_in_anonymous_namespace_p): Break out from ...
11502 (types_same_for_odr): ... here.
11503 * tree.h (type_in_anonymous_namespace_p): Declare.
11505 2013-08-18 Jakub Jelinek <jakub@redhat.com>
11507 PR tree-optimization/58006
11508 * tree-parloops.c (take_address_of): Don't ICE if get_name
11510 (eliminate_local_variables_stmt): Remove clobber stmts.
11512 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
11514 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
11515 error message is issued for an alias to undefined symbol.
11517 2013-08-18 Jan Hubicka <jh@suse.cz>
11519 * cgraph.c (cgraph_create_indirect_edge): Discover
11520 polymorphic calls and record basic info into indirect_info.
11521 * gimple-fold.c (gimple_fold_call): When doing BINFO based
11522 devirtualization, ignore objc function calls.
11523 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
11524 call with no parm index info.
11525 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
11526 * tree.c (virtual_method_call_p): New function.
11527 * tree.h (virtual_method_call_p): Declare.
11529 2013-08-16 Jan Hubicka <jh@suse.cz>
11531 PR middle-end/58179
11532 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
11534 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
11536 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
11539 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11541 * gengtype.c (type_for_name): Add special-case support for
11542 locating types within the "gcc::" namespace.
11543 (open_base_files): Emit a "using namespace gcc" directive.
11545 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11548 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
11549 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
11551 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
11552 array instead of each individual operand as a separate argument.
11553 (emit_fusion_gpr_load): Likewise.
11554 (expand_fusion_gpr_load): Add new function declaration.
11556 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
11557 signature to have the operands passed as an array, instead of as
11558 separate arguments. Allow ZERO_EXTEND to be in the memory
11559 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
11560 depend on the register live/dead flags when peepholes are run.
11561 (expand_fusion_gpr_load): New function to be called from the
11562 peephole2 pass, to change the register that addis sets to be the
11564 (emit_fusion_gpr_load): Change the calling signature to have the
11565 operands passed as an array, instead of as separate arguments.
11566 Allow ZERO_EXTEND to be in the memory address, and also
11567 SIGN_EXTEND if -mpower8-fusion-sign.
11569 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
11570 unspec enumeration.
11571 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
11572 adjust the register addis loads up in the peephole2 pass. Do not
11573 depend on the register live/dead state when the peephole pass is done.
11575 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11577 * gengtype.c (create_user_defined_type): Ensure that the kind
11578 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
11579 declaration is seen before the GTY((user)) marking.
11581 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
11584 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
11586 2013-08-16 Jan Hubicka <jh@suse.cz>
11588 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
11589 arugment expected_type.
11590 (gimple_fold_call): Use it.
11591 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
11592 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
11593 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
11594 (try_make_edge_direct_virtual_call): Likewise.
11595 * tree.c (obj_type_ref_class): New.
11596 * tree.h (obj_type_ref_class): Use it.
11598 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
11600 * sched-vis.c (rtl_slim_pp_initialized): Remove.
11601 (rtl_slim_pp): Likewise.
11602 (init_rtl_slim_pretty_print): Likewise.
11603 (dump_value_slim): Don't call it. Use local pretty printer.
11604 (dump_insn_slim): Likewise.
11605 (dump_rtl_slim): Likewise.
11606 (str_pattern_slim): Likewise.
11607 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
11610 2013-08-16 Jakub Jelinek <jakub@redhat.com>
11612 PR tree-optimization/58164
11613 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
11614 walk gimple_goto_dest of GIMPLE_GOTO.
11616 PR tree-optimization/58165
11617 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
11618 bi_call must be the last stmt in a bb, don't split_block, instead
11619 use fallthru edge from it and give up if there is none.
11620 Release conds vector when returning early.
11622 2013-08-14 Xinliang David Li <davidxl@google.com>
11624 * config/i386/i386.c (ix86_option_override_internal):
11625 Remove unused variable and field.
11627 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11630 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
11631 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
11632 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
11633 and Linux, correct BLKmode alignment when 128-bit alignment is
11634 required and compatibility flag is not set.
11635 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
11636 for zero-size arguments when compatibility flag is not set.
11638 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11640 PR tree-optimization/58145
11641 * tree-sra.c (build_ref_for_offset): If prev_base has
11642 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
11644 2013-08-14 Xinliang David Li <davidxl@google.com>
11646 * config/i386/i386.c (ix86_option_override_internal):
11647 Fix uninitialized variable error.
11649 2013-08-14 Xinliang David Li <davidxl@google.com>
11651 * config/i386/i386.opt: Define two new options.
11652 * config/i386/x86-tune.def: Add arch selector field in macros.
11653 * config/i386/i386.h: Adjust macro definition.
11654 * config/i386/i386.c (ix86_option_override_internal):
11656 (parse_mtune_ctrl_str): New function.
11657 (set_ix86_tune_features): New function.
11658 (ix86_function_specific_restore): Call the new helper function.
11660 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
11662 PR rtl-optimization/57662
11663 * sel-sched.c (code_motion_process_successors): When the current insn
11664 is removed after the recursive traversal, break from the loop.
11665 Add comments and debug printouts.
11667 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11668 Alexandre Oliva <aoliva@redhat.com>
11671 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
11672 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
11673 there also UNSPEC_PLTOFF.
11675 2013-08-14 Marek Polacek <polacek@redhat.com>
11677 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
11678 NUM_CONDITIONS bit positions.
11680 2013-08-13 Cary Coutant <ccoutant@google.com>
11682 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
11683 (attr_checksum): Hash vector contents instead of pointer.
11684 (attr_checksum_ordered): Likewise.
11686 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
11688 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
11689 when Pmode != word_mode. Add length_address attribute.
11690 (sse3_monitor_<mode>): Merge from sse3_monitor and
11691 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
11692 Pmode != word_mode. Update insn length attribute.
11693 * config/i386/i386.c (ix86_option_override_internal): Update
11694 ix86_gen_monitor selection for merged sse3_monitor insn.
11696 2013-08-13 Julian Brown <julian@codesourcery.com>
11698 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
11699 perform invalid legitimization on greater-than-word-size modes for
11700 TARGET_E500_DOUBLE.
11702 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
11704 * ira.c (setup_class_translate_array): Use aclass instead of cl
11705 for classes not fully covered by allocno classes.
11707 2013-08-13 Jakub Jelinek <jakub@redhat.com>
11709 PR tree-optimization/57661
11710 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
11711 * tree-inline.c (tree_function_versioning): Initialize it.
11712 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
11713 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
11714 that is not being copied.
11717 * asan.c (instrument_strlen_call): Fix typo in comment.
11718 Use char * type even for the lhs of POINTER_PLUS_EXPR.
11720 2013-08-13 Steve Ellcey <sellcey@mips.com>
11722 * config/mips/mips.md (prefetch): Use lw instead of ld on
11723 loongson in 32bit mode.
11725 2013-08-13 Nick Clifton <nickc@redhat.com>
11727 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
11729 2013-08-13 Jan Hubicka <jh@suse.cz>
11731 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
11732 introduced edge; fix typo in sanity check.
11733 (cgraph_resolve_speculation): Export; improve diagnostic.
11734 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
11735 speculation at type mismatch.
11736 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
11737 (cgraph_resolve_speculation): Declare.
11738 (symtab_can_be_discarded): New function.
11739 * value-prof.c (gimple_ic_transform): Remove actual transform code.
11740 * ipa-inline-transform.c (speculation_removed): New global var.
11741 (clone_inlined_nodes): See if speculation can be removed.
11742 (inline_call): If speculations was removed, we growths may not match.
11743 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
11744 (speculation_useful_p): New function.
11745 (resolve_noninline_speculation): New function.
11746 (inline_small_functions): Resolve useless speculations.
11747 * ipa-inline.h (speculation_useful_p): Declare
11748 * ipa.c (can_replace_by_local_alias): Simplify.
11749 (ipa_profile): Produce speculative calls in non-lto, too;
11750 add simple cost model; produce local aliases.
11752 2013-08-13 David Malcolm <dmalcolm@redhat.com>
11754 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
11757 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
11759 * config/i386/i386.c (ix86_function_versions): Use error + inform.
11761 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
11763 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
11764 iterator instead of X87MODEF.
11766 2013-08-12 Perez Read <netfirewall@gmail.com>
11769 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
11770 operand 0 for intel asm alternative.
11771 (*movabs<mode>_2): Ditto for operand 1.
11773 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
11775 * config/aarch64/arm_none.h
11776 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
11778 2013-08-12 Nick Clifton <nickc@redhat.com>
11780 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
11782 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
11784 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
11785 expand for QI/HImode operand to produce more effictive code for
11786 unsigned char(short) --> float(double) conversion.
11788 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
11790 * doc/invoke.texi: Mention that -ftls-model does not force the final
11793 2013-08-12 Marek Polacek <polacek@redhat.com>
11794 Marc Glisse <marc.glisse@inria.fr>
11796 PR tree-optimization/57980
11797 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
11798 when creating -1 constant.
11800 2013-08-10 Jan Hubicka <jh@suse.cz>
11802 Workaround binutils PR14342.
11803 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
11804 (gimple_init_edge_profiler): Likewise.
11805 (gimple_gen_ic_func_profiler): Likewise.
11807 2013-08-09 Jan Hubicka <jh@suse.cz>
11809 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
11811 2013-08-09 Xinliang David Li <davidxl@google.com>
11813 * config/i386/stringop.def: New file.
11814 * config/i386/stringop.opt: New file.
11815 * config/i386/i386-opts.h: Include stringopt.def.
11816 * config/i386/i386.opt: Include stringopt.opt.
11817 * config/i386/i386.c (ix86_option_override_internal):
11818 Override default size based stringop inline strategies with options.
11819 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
11822 2013-08-09 Jan Hubicka <jh@suse.cz>
11824 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
11826 2013-08-09 Jan Hubicka <jh@suse.cz>
11828 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
11830 (dump_cgraph_node): Dump profile-id.
11831 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
11832 and common_target_probability.
11833 * lto-cgraph.c (lto_output_edge): Stream common targets.
11834 (lto_output_node): Stream profile ids.
11835 (input_node): Stream profile ids.
11836 (input_edge): Stream common targets.
11837 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
11838 * ipa.c: Include value-prof.h
11839 (ipa_profile_generate_summary): Turn indirect call statement histograms
11840 into common targets.
11841 (ipa_profile): Turn common targets into speculative edges.
11843 2013-08-09 Jan Hubicka <jh@suse.cz>
11845 * cgraph.h (cgraph_node): Add profile_id.
11846 * value-prof.c (cgraph_node_map): Turn into pointer_map.
11847 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
11848 (del_node_map): Update.
11849 (find_func_by_funcdef_no): Replace by ...
11850 (find_func_by_profile_id): ... this one.
11851 (gimple_ic_transform): Do not remove useful histograms when
11852 speculation is not done; dump info when indirect call removal
11854 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
11855 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
11856 (__gcov_indirect_call_profiler_v2): .. this one.
11857 * profile.h (init_node_map): Update.
11858 * coverage.c (coverage_compute_profile_id): New function.
11859 * coverage.h (coverage_compute_profile_id): Declare.
11860 * tree-profile.c (init_ic_make_global_vars): Make
11861 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
11862 (gimple_init_edge_profiler): Update prototype of
11863 __gcov_indirect_call_profiler.
11864 (gimple_gen_ic_func_profiler): Simplify.
11865 (tree_profiling): Use init_node_map
11867 2013-08-09 Jan Hubicka <jh@suse.cz>
11869 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
11870 non-speculative refs.
11871 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
11872 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
11873 (cgraph_set_call_stmt): Likewise.
11874 (cgraph_create_edge_1): Fix release checking compilatoin;
11875 clear lto_stmt_uid.
11876 (cgraph_free_edge): Free indirect info.
11877 (cgraph_turn_edge_to_speculative): New function.
11878 (cgraph_speculative_call_info): New function.
11879 (cgraph_make_edge_direct): Return direct edge; handle speculation.
11880 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
11881 (dump_cgraph_node): Dump speculation.
11882 (verify_edge_count_and_frequency): Accept speculative edges.
11883 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
11884 (verify_cgraph_node): Handle speculation.
11885 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
11886 (cgraph_set_call_stmt): Update prototype.
11887 (cgraph_make_edge_direct): Update prototype.
11888 (cgraph_speculative_call_info): Declare.
11889 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
11890 to change; update call of ipa_find_references.
11891 * ipa-ref.c (ipa_record_reference): Fix return value; clear
11892 lto_stmt_uid and speculative flags.
11893 (ipa_dump_references): Dump speculation.
11894 (ipa_clone_references): Clone speculative flag.
11895 (ipa_clone_referring): Likewise.
11896 (ipa_clone_ref): New function.
11897 (ipa_find_reference): Look into lto_stmt_uids
11898 (ipa_clear_stmts_in_references): Do not clear speculative calls.
11899 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
11900 (ipa_find_reference): Update declaration.
11901 (ipa_clone_ref): Declare.
11902 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
11903 stream speculative flag.
11904 (lto_output_ref): Stream statements uids and speculation.
11905 (input_ref): Likewise.
11906 (input_edge): Stream speuclation.
11907 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
11908 (cgraph_set_call_stmt_including_clones): Handle speculation.
11909 * ipa-inline.c (heap_edge_removal_hook): New function.
11910 (inline_small_functions): Register it.
11911 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
11912 also initialize refs.
11913 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
11915 (try_make_edge_direct_simple_call): Likewise.
11916 (try_make_edge_direct_simple_call): Likewise.
11917 (update_indirect_edges_after_inlining): Likewise.
11918 (remove_described_reference): Look proper lto_stmt_uid.
11919 (propagate_controlled_uses): Likewise.
11920 (propagate_controlled_uses): Liekwise.
11921 * tree-inline.c (copy_bb): Copy speculative edges.
11922 (redirect_all_calls): New function.
11923 (copy_cfg_body): Do redirection after loop info is updated.
11924 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
11926 2013-08-09 Jan Hubicka <jh@suse.cz>
11928 * lto-streamer-out.c (output_function): Renumber PHIs.
11929 * lto-streamer-in.c (input_function): Likewise.
11931 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
11933 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
11934 (get_lane_unsigned): Likewise.
11935 (dup_lane_scalar): Likewise.
11936 (get_lane): enable for VALL.
11937 * config/aarch64/aarch64-simd.md
11938 (aarch64_dup_lane_scalar<mode>): Remove.
11939 (aarch64_get_lane_signed<mode>): Likewise.
11940 (aarch64_get_lane_unsigned<mode>): Likewise.
11941 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
11942 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
11943 (aarch64_get_lane<mode>): Enable for all vector modes.
11944 (aarch64_get_lanedi): Remove misleading constraints.
11945 * config/aarch64/arm_neon.h
11946 (__aarch64_vget_lane_any): Define.
11947 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
11948 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
11949 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
11950 * config/aarch64/iterators.md (VDQQH): New.
11951 (VDQQHS): Likewise.
11952 (vwcore): Likewise.
11954 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
11956 * configure.ac: Add GAS check for LEON instructions on SPARC.
11957 * configure: Regenerate.
11958 * config.in: Likewise.
11959 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
11961 * config/sparc/sparc.opt (LEON, LEON3): New masks.
11962 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
11964 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
11965 (AS_LEON_FLAG): New macro.
11966 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
11967 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
11968 Deal with LEON and LEON3 for the memory model.
11969 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
11970 (atomic_compare_and_swap<mode>_1): Likewise.
11971 (*atomic_compare_and_swap<mode>_1): Likewise.
11973 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11975 * config/arm/neon.md (vcond): Fix floating-point vector
11976 comparisons against 0.
11978 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
11980 * lra-constraints.c (emit_spill_move): Remove assert.
11981 (process_alt_operands): Add more debugging
11982 output. Increase reject for spilling into memory. Decrease
11983 reject for reloading scratch.
11984 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
11986 2013-08-08 Steve Ellcey <sellcey@mips.com>
11988 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
11989 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
11990 micromips incompatible. Add nan2008.
11991 (MULTILIB_DIRNAMES): Add nan2008.
11992 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
11993 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
11994 and micromips incompatible. Add nan2008.
11995 (MULTILIB_DIRNAMES): Add nan2008.
11996 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
11998 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
12000 PR rtl-optimization/58079
12001 * combine.c (combine_simplify_rtx): Avoid using SUBST if
12002 simplify_comparison has widened a comparison with an integer.
12004 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12006 * config/arm/neon.md (movmisalign<mode>): Disable when we
12007 don't allow unaligned accesses.
12008 (*movmisalign<mode>_neon_store): Likewise.
12009 (*movmisalign<mode>_neon_load): Likewise.
12010 (*movmisalign<mode>_neon_store): Likewise.
12011 (*movmisalign<mode>_neon_load): Likewise.
12013 2013-08-08 Jan Hubicka <jh@suse.cz>
12015 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
12016 (make_pass_rebuild_cgraph_edges): Also clear references.
12017 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
12018 * ipa-inline-transform.c (inline_transform): Remove all references
12020 * cgraphunit.c (expand_function): Remove all references after
12022 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
12023 (ipa_find_reference): Rewrite to iterator.
12024 (remove_stmt_references): Likewise.
12025 (ipa_clear_stmts_in_references): New function.
12026 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
12027 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
12029 * ipa-split.c (split_function): Remove references in split function.
12031 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12034 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
12035 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
12037 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12040 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
12041 suggested mode for the assignment isn't compatible with the
12042 registers required.
12044 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12047 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
12049 2013-08-07 Xinliang David Li <davidxl@google.com>
12051 * config/i386/i386.opt: New option -mtune-ctrl=.
12052 * config/i386/x86-tune.def: New file.
12053 * config/i386/i386.h: include x86-tune.def.
12054 * config/i386/i386.c (ix86_option_override_internal):
12055 Parsing -mtune-ctrl= option and set tune features.
12057 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
12060 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
12061 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
12064 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12066 * rtl.h (update_alignments): Declare.
12067 * final.c (grow_label_align): New function extracted from...
12068 (shorten_branches): ...here. Call it.
12069 (update_alignments): New function.
12070 * reorg.c (sibling_labels): New variable.
12071 (get_label_before): Add SIBLING parameter. If it is non-zero, push
12072 the new label along with it onto the sibling_labels vector.
12073 (fill_simple_delay_slots): Adjust call to get_label_before.
12074 (fill_slots_from_thread): Likewise.
12075 (relax_delay_slots): Likewise.
12076 (make_return_insns): Likewise.
12077 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
12079 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12081 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
12082 document its semantics.
12083 (diagnostic_report_diagnostic): Adjust accordingly.
12085 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12087 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
12088 (sparc_option_override): ...and port to new C++ pass API.
12089 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
12091 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
12093 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
12095 2013-08-06 Caroline Tice <cmtice@google.com>
12097 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
12098 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
12099 * tree-pass.h: Add pass_vtable_verify.
12100 * varasm.c (assemble_variable): Add code to properly set the comdat
12101 section and name for the .vtable_map_vars section.
12102 (assemble_vtyv_preinit_initializer): New function.
12103 (default_sectin_type_flags): Make sure .vtable_map_vars section has
12105 * output.h: Add function decl for assemble_vtv_preinit_initializer.
12106 * vtable-verify.c: New file.
12107 * vtable-verify.h: New file.
12108 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
12109 initialiation levels.
12110 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
12111 * passes.def: Insert pass_vtable_verify.
12112 * aclocal.m4: Reorder includes.
12113 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
12114 -fvtv-counts options.
12115 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
12116 as appropriate, if -fvtable-verify=... is used.
12117 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
12118 -fvtable-verify=... is used.
12119 * Makefile.in (OBJS): Add vtable-verify.o to list.
12120 (vtable-verify.o): Add new build rule.
12121 (GTFILES): Add vtable-verify.c to list.
12122 * common.opt (fvtable-verify=): New flag.
12123 (vtv_priority): Values for fvtable-verify= flag.
12124 (fvtv-counts): New flag.
12125 (fvtv-debug): New flag.
12126 * tree.h (save_vtable_map_decl): New extern function decl.
12128 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12130 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
12131 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
12132 (pass_data_rl78_devirt): ...new pass_data instance and...
12133 (make_pass_rl78_devirt): ...new function.
12134 (rl78_asm_file_start): Port pass registration to new C++ API.
12136 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12138 * coretypes.h (rtl_opt_pass): Add.
12139 (gcc::context): Add.
12140 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
12141 (epiphany_init): Port to new C++ pass API.
12142 (epiphany_optimize_mode_switching): Likewise.
12143 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
12144 (pass_manager::get_pass_mode_switching): New.
12145 (pass_manager::get_pass_peephole2): New.
12146 * mode-switching.c (pass_mode_switching): Add clone method.
12147 * recog.c (pass_peephole2): Add clone method.
12148 (pass_split_all_insns): Add clone method.
12150 2013-08-06 David Malcolm <dmalcolm@redhat.com>
12152 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
12153 (mips_option_override): ...here, porting to new C++ API for passes.
12155 2013-08-06 Jan Hubicka <jh@suse.cz>
12157 * cgraph.c (cgraph_get_body): New function based on lto.c
12159 * cgraph.h (cgraph_get_body): Declare.
12160 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
12162 * cgraphunit.c (expand_function): Get body prior expanding.
12163 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
12164 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
12166 * passes.c (do_per_function_toporder): Get body.
12167 * tree-inline.c (expand_call_inline): Get body prior inlining it.
12168 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
12170 2013-08-06 Martin Jambor <mjambor@suse.cz>
12173 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
12174 is not re-finalized. Rename second parameter to no_collect.
12176 2013-08-06 Martin Jambor <mjambor@suse.cz>
12178 PR middle-end/58041
12179 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
12180 MEM_REF has proper alignment information.
12182 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
12185 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
12187 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
12188 rtx (*) (rtx, ...) with insn_gen_fn.
12189 * genoutput.c (output_insn_data): Cast gen_? function pointers to
12190 insn_gen_fn::stored_funcptr. Add initializer braces.
12192 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12194 Rewrite how instances of passes are cloned to remove assumptions
12195 about their sizes (thus allowing pass subclasses to have
12196 additional data fields, albeit non-GC-managed ones at this point).
12198 * passes.c (make_pass_instance): Now that passes have clone
12199 methods, rewrite this function to eliminate XNEW and memcpy
12200 calls that used hardcoded sizes. Since this function no longer
12201 creates pass instances, rename it to...
12202 (add_pass_instance): ...this. Document the old way that passes were
12203 numbered and flagged, and rework this function to continue using it.
12204 (next_pass_1): Add an initial_pass argument for use by
12206 (position_pass): When adding multiple instances of a pass, use
12207 the pass's clone method, rather than relying on the XNEW/memcpy
12208 within the former make_pass_instance (now add_pass_instance).
12209 (pass_manager::pass_manager): When invoking next_pass_1, also supply
12210 the initial instance of the current pass within the pass manager.
12212 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12214 This is the automated part of the conversion of passes from C
12215 structs to C++ classes.
12217 Patch autogenerated by refactor_passes.py from
12218 https://github.com/davidmalcolm/gcc-refactoring-scripts
12219 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
12221 * asan.c (pass_asan): Convert from a global struct to a subclass of
12222 gimple_opt_pass along with...
12223 (pass_data_asan): ...new pass_data instance and...
12224 (make_pass_asan): ...new function.
12225 (pass_asan_O0): Convert from a global struct to a subclass of
12226 gimple_opt_pass along with...
12227 (pass_data_asan_O0): ...new pass_data instance and...
12228 (make_pass_asan_O0): ...new function.
12229 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
12230 subclass of rtl_opt_pass along with...
12231 (pass_data_inc_dec): ...new pass_data instance and...
12232 (make_pass_inc_dec): ...new function.
12233 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
12234 a subclass of rtl_opt_pass along with...
12235 (pass_data_reorder_blocks): ...new pass_data instance and...
12236 (make_pass_reorder_blocks): ...new function.
12237 (pass_duplicate_computed_gotos): Convert from a global struct to a
12238 subclass of rtl_opt_pass along with...
12239 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
12240 (make_pass_duplicate_computed_gotos): ...new function.
12241 (pass_partition_blocks): Convert from a global struct to a subclass of
12242 rtl_opt_pass along with...
12243 (pass_data_partition_blocks): ...new pass_data instance and...
12244 (make_pass_partition_blocks): ...new function.
12245 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
12246 struct to a subclass of rtl_opt_pass along with...
12247 (pass_data_branch_target_load_optimize1): ...new pass_data instance
12249 (make_pass_branch_target_load_optimize1): ...new function.
12250 (pass_branch_target_load_optimize2): Convert from a global struct to a
12251 subclass of rtl_opt_pass along with...
12252 (pass_data_branch_target_load_optimize2): ...new pass_data instance
12254 (make_pass_branch_target_load_optimize2): ...new function.
12255 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
12256 of rtl_opt_pass along with...
12257 (pass_data_jump): ...new pass_data instance and...
12258 (make_pass_jump): ...new function.
12259 (pass_jump2): Convert from a global struct to a subclass of
12260 rtl_opt_pass along with...
12261 (pass_data_jump2): ...new pass_data instance and...
12262 (make_pass_jump2): ...new function.
12263 * cfgexpand.c (pass_expand): Convert from a global struct to a
12264 subclass of rtl_opt_pass along with...
12265 (pass_data_expand): ...new pass_data instance and...
12266 (make_pass_expand): ...new function.
12267 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
12268 of rtl_opt_pass along with...
12269 (pass_data_free_cfg): ...new pass_data instance and...
12270 (make_pass_free_cfg): ...new function.
12271 (pass_into_cfg_layout_mode): Convert from a global struct to a
12272 subclass of rtl_opt_pass along with...
12273 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
12274 (make_pass_into_cfg_layout_mode): ...new function.
12275 (pass_outof_cfg_layout_mode): Convert from a global struct to a
12276 subclass of rtl_opt_pass along with...
12277 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
12278 (make_pass_outof_cfg_layout_mode): ...new function.
12279 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
12280 struct to a subclass of gimple_opt_pass along with...
12281 (pass_data_build_cgraph_edges): ...new pass_data instance and...
12282 (make_pass_build_cgraph_edges): ...new function.
12283 (pass_rebuild_cgraph_edges): Convert from a global struct to a
12284 subclass of gimple_opt_pass along with...
12285 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
12286 (make_pass_rebuild_cgraph_edges): ...new function.
12287 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
12288 subclass of gimple_opt_pass along with...
12289 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
12291 (make_pass_remove_cgraph_callee_edges): ...new function.
12292 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
12293 struct to a subclass of rtl_opt_pass along with...
12294 (pass_data_stack_adjustments): ...new pass_data instance and...
12295 (make_pass_stack_adjustments): ...new function.
12296 * combine.c (pass_combine): Convert from a global struct to a subclass
12297 of rtl_opt_pass along with...
12298 (pass_data_combine): ...new pass_data instance and...
12299 (make_pass_combine): ...new function.
12300 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
12301 global struct to a subclass of rtl_opt_pass along with...
12302 (pass_data_compare_elim_after_reload): ...new pass_data instance
12304 (make_pass_compare_elim_after_reload): ...new function.
12305 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
12306 of rtl_opt_pass along with...
12307 (pass_data_rtl_cprop): ...new pass_data instance and...
12308 (make_pass_rtl_cprop): ...new function.
12309 * cse.c (pass_cse): Convert from a global struct to a subclass of
12310 rtl_opt_pass along with...
12311 (pass_data_cse): ...new pass_data instance and...
12312 (make_pass_cse): ...new function.
12313 (pass_cse2): Convert from a global struct to a subclass of
12314 rtl_opt_pass along with...
12315 (pass_data_cse2): ...new pass_data instance and...
12316 (make_pass_cse2): ...new function.
12317 (pass_cse_after_global_opts): Convert from a global struct to a
12318 subclass of rtl_opt_pass along with...
12319 (pass_data_cse_after_global_opts): ...new pass_data instance and...
12320 (make_pass_cse_after_global_opts): ...new function.
12321 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
12322 of rtl_opt_pass along with...
12323 (pass_data_ud_rtl_dce): ...new pass_data instance and...
12324 (make_pass_ud_rtl_dce): ...new function.
12325 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
12326 rtl_opt_pass along with...
12327 (pass_data_fast_rtl_dce): ...new pass_data instance and...
12328 (make_pass_fast_rtl_dce): ...new function.
12329 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
12330 a subclass of rtl_opt_pass along with...
12331 (pass_data_df_initialize_opt): ...new pass_data instance and...
12332 (make_pass_df_initialize_opt): ...new function.
12333 (pass_df_initialize_no_opt): Convert from a global struct to a
12334 subclass of rtl_opt_pass along with...
12335 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
12336 (make_pass_df_initialize_no_opt): ...new function.
12337 (pass_df_finish): Convert from a global struct to a subclass of
12338 rtl_opt_pass along with...
12339 (pass_data_df_finish): ...new pass_data instance and...
12340 (make_pass_df_finish): ...new function.
12341 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
12342 rtl_opt_pass along with...
12343 (pass_data_rtl_dse1): ...new pass_data instance and...
12344 (make_pass_rtl_dse1): ...new function.
12345 (pass_rtl_dse2): Convert from a global struct to a subclass of
12346 rtl_opt_pass along with...
12347 (pass_data_rtl_dse2): ...new pass_data instance and...
12348 (make_pass_rtl_dse2): ...new function.
12349 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
12350 subclass of rtl_opt_pass along with...
12351 (pass_data_dwarf2_frame): ...new pass_data instance and...
12352 (make_pass_dwarf2_frame): ...new function.
12353 * except.c (pass_set_nothrow_function_flags): Convert from a global
12354 struct to a subclass of rtl_opt_pass along with...
12355 (pass_data_set_nothrow_function_flags): ...new pass_data instance
12357 (make_pass_set_nothrow_function_flags): ...new function.
12358 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
12359 subclass of rtl_opt_pass along with...
12360 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
12362 (make_pass_convert_to_eh_region_ranges): ...new function.
12363 * final.c (pass_compute_alignments): Convert from a global struct to a
12364 subclass of rtl_opt_pass along with...
12365 (pass_data_compute_alignments): ...new pass_data instance and...
12366 (make_pass_compute_alignments): ...new function.
12367 (pass_final): Convert from a global struct to a subclass of
12368 rtl_opt_pass along with...
12369 (pass_data_final): ...new pass_data instance and...
12370 (make_pass_final): ...new function.
12371 (pass_shorten_branches): Convert from a global struct to a subclass of
12372 rtl_opt_pass along with...
12373 (pass_data_shorten_branches): ...new pass_data instance and...
12374 (make_pass_shorten_branches): ...new function.
12375 (pass_clean_state): Convert from a global struct to a subclass of
12376 rtl_opt_pass along with...
12377 (pass_data_clean_state): ...new pass_data instance and...
12378 (make_pass_clean_state): ...new function.
12379 * function.c (pass_instantiate_virtual_regs): Convert from a global
12380 struct to a subclass of rtl_opt_pass along with...
12381 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
12382 (make_pass_instantiate_virtual_regs): ...new function.
12383 (pass_leaf_regs): Convert from a global struct to a subclass of
12384 rtl_opt_pass along with...
12385 (pass_data_leaf_regs): ...new pass_data instance and...
12386 (make_pass_leaf_regs): ...new function.
12387 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
12388 subclass of rtl_opt_pass along with...
12389 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
12391 (make_pass_thread_prologue_and_epilogue): ...new function.
12392 (pass_match_asm_constraints): Convert from a global struct to a
12393 subclass of rtl_opt_pass along with...
12394 (pass_data_match_asm_constraints): ...new pass_data instance and...
12395 (make_pass_match_asm_constraints): ...new function.
12396 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
12397 subclass of rtl_opt_pass along with...
12398 (pass_data_rtl_fwprop): ...new pass_data instance and...
12399 (make_pass_rtl_fwprop): ...new function.
12400 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
12401 rtl_opt_pass along with...
12402 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
12403 (make_pass_rtl_fwprop_addr): ...new function.
12404 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
12405 rtl_opt_pass along with...
12406 (pass_data_rtl_pre): ...new pass_data instance and...
12407 (make_pass_rtl_pre): ...new function.
12408 (pass_rtl_hoist): Convert from a global struct to a subclass of
12409 rtl_opt_pass along with...
12410 (pass_data_rtl_hoist): ...new pass_data instance and...
12411 (make_pass_rtl_hoist): ...new function.
12412 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
12413 subclass of gimple_opt_pass along with...
12414 (pass_data_lower_cf): ...new pass_data instance and...
12415 (make_pass_lower_cf): ...new function.
12416 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
12417 from a global struct to a subclass of gimple_opt_pass along with...
12418 (pass_data_strength_reduction): ...new pass_data instance and...
12419 (make_pass_strength_reduction): ...new function.
12420 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
12421 of rtl_opt_pass along with...
12422 (pass_data_rtl_ifcvt): ...new pass_data instance and...
12423 (make_pass_rtl_ifcvt): ...new function.
12424 (pass_if_after_combine): Convert from a global struct to a subclass of
12425 rtl_opt_pass along with...
12426 (pass_data_if_after_combine): ...new pass_data instance and...
12427 (make_pass_if_after_combine): ...new function.
12428 (pass_if_after_reload): Convert from a global struct to a subclass of
12429 rtl_opt_pass along with...
12430 (pass_data_if_after_reload): ...new pass_data instance and...
12431 (make_pass_if_after_reload): ...new function.
12432 * init-regs.c (pass_initialize_regs): Convert from a global struct to
12433 a subclass of rtl_opt_pass along with...
12434 (pass_data_initialize_regs): ...new pass_data instance and...
12435 (make_pass_initialize_regs): ...new function.
12436 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
12437 of ipa_opt_pass_d along with...
12438 (pass_data_ipa_cp): ...new pass_data instance and...
12439 (make_pass_ipa_cp): ...new function.
12440 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
12441 global struct to a subclass of gimple_opt_pass along with...
12442 (pass_data_inline_parameters): ...new pass_data instance and...
12443 (make_pass_inline_parameters): ...new function.
12444 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
12445 subclass of gimple_opt_pass along with...
12446 (pass_data_early_inline): ...new pass_data instance and...
12447 (make_pass_early_inline): ...new function.
12448 (pass_ipa_inline): Convert from a global struct to a subclass of
12449 ipa_opt_pass_d along with...
12450 (pass_data_ipa_inline): ...new pass_data instance and...
12451 (make_pass_ipa_inline): ...new function.
12452 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
12453 struct to a subclass of gimple_opt_pass along with...
12454 (pass_data_local_pure_const): ...new pass_data instance and...
12455 (make_pass_local_pure_const): ...new function.
12456 (pass_ipa_pure_const): Convert from a global struct to a subclass of
12457 ipa_opt_pass_d along with...
12458 (pass_data_ipa_pure_const): ...new pass_data instance and...
12459 (make_pass_ipa_pure_const): ...new function.
12460 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
12461 to a subclass of ipa_opt_pass_d along with...
12462 (pass_data_ipa_reference): ...new pass_data instance and...
12463 (make_pass_ipa_reference): ...new function.
12464 * ipa-split.c (pass_split_functions): Convert from a global struct to
12465 a subclass of gimple_opt_pass along with...
12466 (pass_data_split_functions): ...new pass_data instance and...
12467 (make_pass_split_functions): ...new function.
12468 (pass_feedback_split_functions): Convert from a global struct to a
12469 subclass of gimple_opt_pass along with...
12470 (pass_data_feedback_split_functions): ...new pass_data instance and...
12471 (make_pass_feedback_split_functions): ...new function.
12472 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
12473 global struct to a subclass of simple_ipa_opt_pass along with...
12474 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
12476 (make_pass_ipa_function_and_variable_visibility): ...new function.
12477 (pass_ipa_free_inline_summary): Convert from a global struct to a
12478 subclass of simple_ipa_opt_pass along with...
12479 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
12480 (make_pass_ipa_free_inline_summary): ...new function.
12481 (pass_ipa_whole_program_visibility): Convert from a global struct to a
12482 subclass of ipa_opt_pass_d along with...
12483 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
12485 (make_pass_ipa_whole_program_visibility): ...new function.
12486 (pass_ipa_profile): Convert from a global struct to a subclass of
12487 ipa_opt_pass_d along with...
12488 (pass_data_ipa_profile): ...new pass_data instance and...
12489 (make_pass_ipa_profile): ...new function.
12490 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
12491 ipa_opt_pass_d along with...
12492 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
12493 (make_pass_ipa_cdtor_merge): ...new function.
12494 * ira.c (pass_ira): Convert from a global struct to a subclass of
12495 rtl_opt_pass along with...
12496 (pass_data_ira): ...new pass_data instance and...
12497 (make_pass_ira): ...new function.
12498 (pass_reload): Convert from a global struct to a subclass of
12499 rtl_opt_pass along with...
12500 (pass_data_reload): ...new pass_data instance and...
12501 (make_pass_reload): ...new function.
12502 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
12503 subclass of rtl_opt_pass along with...
12504 (pass_data_cleanup_barriers): ...new pass_data instance and...
12505 (make_pass_cleanup_barriers): ...new function.
12506 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
12507 of rtl_opt_pass along with...
12508 (pass_data_loop2): ...new pass_data instance and...
12509 (make_pass_loop2): ...new function.
12510 (pass_rtl_loop_init): Convert from a global struct to a subclass of
12511 rtl_opt_pass along with...
12512 (pass_data_rtl_loop_init): ...new pass_data instance and...
12513 (make_pass_rtl_loop_init): ...new function.
12514 (pass_rtl_loop_done): Convert from a global struct to a subclass of
12515 rtl_opt_pass along with...
12516 (pass_data_rtl_loop_done): ...new pass_data instance and...
12517 (make_pass_rtl_loop_done): ...new function.
12518 (pass_rtl_move_loop_invariants): Convert from a global struct to a
12519 subclass of rtl_opt_pass along with...
12520 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
12521 (make_pass_rtl_move_loop_invariants): ...new function.
12522 (pass_rtl_unswitch): Convert from a global struct to a subclass of
12523 rtl_opt_pass along with...
12524 (pass_data_rtl_unswitch): ...new pass_data instance and...
12525 (make_pass_rtl_unswitch): ...new function.
12526 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
12527 subclass of rtl_opt_pass along with...
12528 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
12530 (make_pass_rtl_unroll_and_peel_loops): ...new function.
12531 (pass_rtl_doloop): Convert from a global struct to a subclass of
12532 rtl_opt_pass along with...
12533 (pass_data_rtl_doloop): ...new pass_data instance and...
12534 (make_pass_rtl_doloop): ...new function.
12535 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
12536 a subclass of rtl_opt_pass along with...
12537 (pass_data_lower_subreg): ...new pass_data instance and...
12538 (make_pass_lower_subreg): ...new function.
12539 (pass_lower_subreg2): Convert from a global struct to a subclass of
12540 rtl_opt_pass along with...
12541 (pass_data_lower_subreg2): ...new pass_data instance and...
12542 (make_pass_lower_subreg2): ...new function.
12543 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
12544 struct to a subclass of ipa_opt_pass_d along with...
12545 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
12546 (make_pass_ipa_lto_gimple_out): ...new function.
12547 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
12548 of ipa_opt_pass_d along with...
12549 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
12550 (make_pass_ipa_lto_finish_out): ...new function.
12551 * mode-switching.c (pass_mode_switching): Convert from a global struct
12552 to a subclass of rtl_opt_pass along with...
12553 (pass_data_mode_switching): ...new pass_data instance and...
12554 (make_pass_mode_switching): ...new function.
12555 * modulo-sched.c (pass_sms): Convert from a global struct to a
12556 subclass of rtl_opt_pass along with...
12557 (pass_data_sms): ...new pass_data instance and...
12558 (make_pass_sms): ...new function.
12559 * omp-low.c (pass_expand_omp): Convert from a global struct to a
12560 subclass of gimple_opt_pass along with...
12561 (pass_data_expand_omp): ...new pass_data instance and...
12562 (make_pass_expand_omp): ...new function.
12563 (pass_lower_omp): Convert from a global struct to a subclass of
12564 gimple_opt_pass along with...
12565 (pass_data_lower_omp): ...new pass_data instance and...
12566 (make_pass_lower_omp): ...new function.
12567 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
12568 of gimple_opt_pass along with...
12569 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
12570 (make_pass_diagnose_omp_blocks): ...new function.
12571 * passes.c (pass_early_local_passes): Convert from a global struct to
12572 a subclass of simple_ipa_opt_pass along with...
12573 (pass_data_early_local_passes): ...new pass_data instance and...
12574 (make_pass_early_local_passes): ...new function.
12575 (pass_all_early_optimizations): Convert from a global struct to a
12576 subclass of gimple_opt_pass along with...
12577 (pass_data_all_early_optimizations): ...new pass_data instance and...
12578 (make_pass_all_early_optimizations): ...new function.
12579 (pass_all_optimizations): Convert from a global struct to a subclass
12580 of gimple_opt_pass along with...
12581 (pass_data_all_optimizations): ...new pass_data instance and...
12582 (make_pass_all_optimizations): ...new function.
12583 (pass_all_optimizations_g): Convert from a global struct to a subclass
12584 of gimple_opt_pass along with...
12585 (pass_data_all_optimizations_g): ...new pass_data instance and...
12586 (make_pass_all_optimizations_g): ...new function.
12587 (pass_rest_of_compilation): Convert from a global struct to a subclass
12588 of rtl_opt_pass along with...
12589 (pass_data_rest_of_compilation): ...new pass_data instance and...
12590 (make_pass_rest_of_compilation): ...new function.
12591 (pass_postreload): Convert from a global struct to a subclass of
12592 rtl_opt_pass along with...
12593 (pass_data_postreload): ...new pass_data instance and...
12594 (make_pass_postreload): ...new function.
12595 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
12596 subclass of rtl_opt_pass along with...
12597 (pass_data_gcse2): ...new pass_data instance and...
12598 (make_pass_gcse2): ...new function.
12599 * postreload.c (pass_postreload_cse): Convert from a global struct to
12600 a subclass of rtl_opt_pass along with...
12601 (pass_data_postreload_cse): ...new pass_data instance and...
12602 (make_pass_postreload_cse): ...new function.
12603 * predict.c (pass_profile): Convert from a global struct to a subclass
12604 of gimple_opt_pass along with...
12605 (pass_data_profile): ...new pass_data instance and...
12606 (make_pass_profile): ...new function.
12607 (pass_strip_predict_hints): Convert from a global struct to a subclass
12608 of gimple_opt_pass along with...
12609 (pass_data_strip_predict_hints): ...new pass_data instance and...
12610 (make_pass_strip_predict_hints): ...new function.
12611 * recog.c (pass_peephole2): Convert from a global struct to a subclass
12612 of rtl_opt_pass along with...
12613 (pass_data_peephole2): ...new pass_data instance and...
12614 (make_pass_peephole2): ...new function.
12615 (pass_split_all_insns): Convert from a global struct to a subclass of
12616 rtl_opt_pass along with...
12617 (pass_data_split_all_insns): ...new pass_data instance and...
12618 (make_pass_split_all_insns): ...new function.
12619 (pass_split_after_reload): Convert from a global struct to a subclass
12620 of rtl_opt_pass along with...
12621 (pass_data_split_after_reload): ...new pass_data instance and...
12622 (make_pass_split_after_reload): ...new function.
12623 (pass_split_before_regstack): Convert from a global struct to a
12624 subclass of rtl_opt_pass along with...
12625 (pass_data_split_before_regstack): ...new pass_data instance and...
12626 (make_pass_split_before_regstack): ...new function.
12627 (pass_split_before_sched2): Convert from a global struct to a subclass
12628 of rtl_opt_pass along with...
12629 (pass_data_split_before_sched2): ...new pass_data instance and...
12630 (make_pass_split_before_sched2): ...new function.
12631 (pass_split_for_shorten_branches): Convert from a global struct to a
12632 subclass of rtl_opt_pass along with...
12633 (pass_data_split_for_shorten_branches): ...new pass_data instance
12635 (make_pass_split_for_shorten_branches): ...new function.
12636 * ree.c (pass_ree): Convert from a global struct to a subclass of
12637 rtl_opt_pass along with...
12638 (pass_data_ree): ...new pass_data instance and...
12639 (make_pass_ree): ...new function.
12640 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
12641 subclass of rtl_opt_pass along with...
12642 (pass_data_stack_regs): ...new pass_data instance and...
12643 (make_pass_stack_regs): ...new function.
12644 (pass_stack_regs_run): Convert from a global struct to a subclass of
12645 rtl_opt_pass along with...
12646 (pass_data_stack_regs_run): ...new pass_data instance and...
12647 (make_pass_stack_regs_run): ...new function.
12648 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
12649 subclass of rtl_opt_pass along with...
12650 (pass_data_cprop_hardreg): ...new pass_data instance and...
12651 (make_pass_cprop_hardreg): ...new function.
12652 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
12653 subclass of rtl_opt_pass along with...
12654 (pass_data_reginfo_init): ...new pass_data instance and...
12655 (make_pass_reginfo_init): ...new function.
12656 * regmove.c (pass_regmove): Convert from a global struct to a subclass
12657 of rtl_opt_pass along with...
12658 (pass_data_regmove): ...new pass_data instance and...
12659 (make_pass_regmove): ...new function.
12660 * regrename.c (pass_regrename): Convert from a global struct to a
12661 subclass of rtl_opt_pass along with...
12662 (pass_data_regrename): ...new pass_data instance and...
12663 (make_pass_regrename): ...new function.
12664 * reorg.c (pass_delay_slots): Convert from a global struct to a
12665 subclass of rtl_opt_pass along with...
12666 (pass_data_delay_slots): ...new pass_data instance and...
12667 (make_pass_delay_slots): ...new function.
12668 (pass_machine_reorg): Convert from a global struct to a subclass of
12669 rtl_opt_pass along with...
12670 (pass_data_machine_reorg): ...new pass_data instance and...
12671 (make_pass_machine_reorg): ...new function.
12672 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
12673 of rtl_opt_pass along with...
12674 (pass_data_sched): ...new pass_data instance and...
12675 (make_pass_sched): ...new function.
12676 (pass_sched2): Convert from a global struct to a subclass of
12677 rtl_opt_pass along with...
12678 (pass_data_sched2): ...new pass_data instance and...
12679 (make_pass_sched2): ...new function.
12680 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
12681 to a subclass of rtl_opt_pass along with...
12682 (pass_data_stack_ptr_mod): ...new pass_data instance and...
12683 (make_pass_stack_ptr_mod): ...new function.
12684 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
12685 to a subclass of rtl_opt_pass along with...
12686 (pass_data_rtl_store_motion): ...new pass_data instance and...
12687 (make_pass_rtl_store_motion): ...new function.
12688 * tracer.c (pass_tracer): Convert from a global struct to a subclass
12689 of gimple_opt_pass along with...
12690 (pass_data_tracer): ...new pass_data instance and...
12691 (make_pass_tracer): ...new function.
12692 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
12693 to a subclass of gimple_opt_pass along with...
12694 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
12695 (make_pass_diagnose_tm_blocks): ...new function.
12696 (pass_lower_tm): Convert from a global struct to a subclass of
12697 gimple_opt_pass along with...
12698 (pass_data_lower_tm): ...new pass_data instance and...
12699 (make_pass_lower_tm): ...new function.
12700 (pass_tm_init): Convert from a global struct to a subclass of
12701 gimple_opt_pass along with...
12702 (pass_data_tm_init): ...new pass_data instance and...
12703 (make_pass_tm_init): ...new function.
12704 (pass_tm_mark): Convert from a global struct to a subclass of
12705 gimple_opt_pass along with...
12706 (pass_data_tm_mark): ...new pass_data instance and...
12707 (make_pass_tm_mark): ...new function.
12708 (pass_tm_edges): Convert from a global struct to a subclass of
12709 gimple_opt_pass along with...
12710 (pass_data_tm_edges): ...new pass_data instance and...
12711 (make_pass_tm_edges): ...new function.
12712 (pass_tm_memopt): Convert from a global struct to a subclass of
12713 gimple_opt_pass along with...
12714 (pass_data_tm_memopt): ...new pass_data instance and...
12715 (make_pass_tm_memopt): ...new function.
12716 (pass_ipa_tm): Convert from a global struct to a subclass of
12717 simple_ipa_opt_pass along with...
12718 (pass_data_ipa_tm): ...new pass_data instance and...
12719 (make_pass_ipa_tm): ...new function.
12720 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
12721 subclass of gimple_opt_pass along with...
12722 (pass_data_call_cdce): ...new pass_data instance and...
12723 (make_pass_call_cdce): ...new function.
12724 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
12725 subclass of gimple_opt_pass along with...
12726 (pass_data_build_cfg): ...new pass_data instance and...
12727 (make_pass_build_cfg): ...new function.
12728 (pass_split_crit_edges): Convert from a global struct to a subclass of
12729 gimple_opt_pass along with...
12730 (pass_data_split_crit_edges): ...new pass_data instance and...
12731 (make_pass_split_crit_edges): ...new function.
12732 (pass_warn_function_return): Convert from a global struct to a
12733 subclass of gimple_opt_pass along with...
12734 (pass_data_warn_function_return): ...new pass_data instance and...
12735 (make_pass_warn_function_return): ...new function.
12736 (pass_warn_function_noreturn): Convert from a global struct to a
12737 subclass of gimple_opt_pass along with...
12738 (pass_data_warn_function_noreturn): ...new pass_data instance and...
12739 (make_pass_warn_function_noreturn): ...new function.
12740 (pass_warn_unused_result): Convert from a global struct to a subclass
12741 of gimple_opt_pass along with...
12742 (pass_data_warn_unused_result): ...new pass_data instance and...
12743 (make_pass_warn_unused_result): ...new function.
12744 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
12745 a subclass of gimple_opt_pass along with...
12746 (pass_data_merge_phi): ...new pass_data instance and...
12747 (make_pass_merge_phi): ...new function.
12748 * tree-complex.c (pass_lower_complex): Convert from a global struct to
12749 a subclass of gimple_opt_pass along with...
12750 (pass_data_lower_complex): ...new pass_data instance and...
12751 (make_pass_lower_complex): ...new function.
12752 (pass_lower_complex_O0): Convert from a global struct to a subclass of
12753 gimple_opt_pass along with...
12754 (pass_data_lower_complex_O0): ...new pass_data instance and...
12755 (make_pass_lower_complex_O0): ...new function.
12756 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
12757 subclass of gimple_opt_pass along with...
12758 (pass_data_lower_eh): ...new pass_data instance and...
12759 (make_pass_lower_eh): ...new function.
12760 (pass_refactor_eh): Convert from a global struct to a subclass of
12761 gimple_opt_pass along with...
12762 (pass_data_refactor_eh): ...new pass_data instance and...
12763 (make_pass_refactor_eh): ...new function.
12764 (pass_lower_resx): Convert from a global struct to a subclass of
12765 gimple_opt_pass along with...
12766 (pass_data_lower_resx): ...new pass_data instance and...
12767 (make_pass_lower_resx): ...new function.
12768 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
12769 of gimple_opt_pass along with...
12770 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
12771 (make_pass_lower_eh_dispatch): ...new function.
12772 (pass_cleanup_eh): Convert from a global struct to a subclass of
12773 gimple_opt_pass along with...
12774 (pass_data_cleanup_eh): ...new pass_data instance and...
12775 (make_pass_cleanup_eh): ...new function.
12776 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
12777 to a subclass of simple_ipa_opt_pass along with...
12778 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
12779 (make_pass_ipa_lower_emutls): ...new function.
12780 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
12781 a subclass of gimple_opt_pass along with...
12782 (pass_data_if_conversion): ...new pass_data instance and...
12783 (make_pass_if_conversion): ...new function.
12784 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
12785 subclass of gimple_opt_pass along with...
12786 (pass_data_build_ssa): ...new pass_data instance and...
12787 (make_pass_build_ssa): ...new function.
12788 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
12789 global struct to a subclass of gimple_opt_pass along with...
12790 (pass_data_loop_distribution): ...new pass_data instance and...
12791 (make_pass_loop_distribution): ...new function.
12792 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
12793 subclass of gimple_opt_pass along with...
12794 (pass_data_mudflap_1): ...new pass_data instance and...
12795 (make_pass_mudflap_1): ...new function.
12796 (pass_mudflap_2): Convert from a global struct to a subclass of
12797 gimple_opt_pass along with...
12798 (pass_data_mudflap_2): ...new pass_data instance and...
12799 (make_pass_mudflap_2): ...new function.
12800 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
12801 subclass of gimple_opt_pass along with...
12802 (pass_data_mudflap_1): ...new pass_data instance and...
12803 (make_pass_mudflap_1): ...new function.
12804 (pass_mudflap_2): Convert from a global struct to a subclass of
12805 gimple_opt_pass along with...
12806 (pass_data_mudflap_2): ...new pass_data instance and...
12807 (make_pass_mudflap_2): ...new function.
12808 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
12809 gimple_opt_pass along with...
12810 (pass_data_nrv): ...new pass_data instance and...
12811 (make_pass_nrv): ...new function.
12812 (pass_return_slot): Convert from a global struct to a subclass of
12813 gimple_opt_pass along with...
12814 (pass_data_return_slot): ...new pass_data instance and...
12815 (make_pass_return_slot): ...new function.
12816 * tree-object-size.c (pass_object_sizes): Convert from a global struct
12817 to a subclass of gimple_opt_pass along with...
12818 (pass_data_object_sizes): ...new pass_data instance and...
12819 (make_pass_object_sizes): ...new function.
12820 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
12821 global struct to a subclass of gimple_opt_pass along with...
12822 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
12824 (make_pass_cleanup_cfg_post_optimizing): ...new function.
12825 (pass_fixup_cfg): Convert from a global struct to a subclass of
12826 gimple_opt_pass along with...
12827 (pass_data_fixup_cfg): ...new pass_data instance and...
12828 (make_pass_fixup_cfg): ...new function.
12829 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
12830 (make_pass_mudflap_1): ...new function.
12831 (pass_mudflap_2): Replace declaration with that of...
12832 (make_pass_mudflap_2): ...new function.
12833 (pass_asan): Replace declaration with that of...
12834 (make_pass_asan): ...new function.
12835 (pass_asan_O0): Replace declaration with that of...
12836 (make_pass_asan_O0): ...new function.
12837 (pass_tsan): Replace declaration with that of...
12838 (make_pass_tsan): ...new function.
12839 (pass_tsan_O0): Replace declaration with that of...
12840 (make_pass_tsan_O0): ...new function.
12841 (pass_lower_cf): Replace declaration with that of...
12842 (make_pass_lower_cf): ...new function.
12843 (pass_refactor_eh): Replace declaration with that of...
12844 (make_pass_refactor_eh): ...new function.
12845 (pass_lower_eh): Replace declaration with that of...
12846 (make_pass_lower_eh): ...new function.
12847 (pass_lower_eh_dispatch): Replace declaration with that of...
12848 (make_pass_lower_eh_dispatch): ...new function.
12849 (pass_lower_resx): Replace declaration with that of...
12850 (make_pass_lower_resx): ...new function.
12851 (pass_build_cfg): Replace declaration with that of...
12852 (make_pass_build_cfg): ...new function.
12853 (pass_early_tree_profile): Replace declaration with that of...
12854 (make_pass_early_tree_profile): ...new function.
12855 (pass_cleanup_eh): Replace declaration with that of...
12856 (make_pass_cleanup_eh): ...new function.
12857 (pass_sra): Replace declaration with that of...
12858 (make_pass_sra): ...new function.
12859 (pass_sra_early): Replace declaration with that of...
12860 (make_pass_sra_early): ...new function.
12861 (pass_early_ipa_sra): Replace declaration with that of...
12862 (make_pass_early_ipa_sra): ...new function.
12863 (pass_tail_recursion): Replace declaration with that of...
12864 (make_pass_tail_recursion): ...new function.
12865 (pass_tail_calls): Replace declaration with that of...
12866 (make_pass_tail_calls): ...new function.
12867 (pass_tree_loop): Replace declaration with that of...
12868 (make_pass_tree_loop): ...new function.
12869 (pass_tree_loop_init): Replace declaration with that of...
12870 (make_pass_tree_loop_init): ...new function.
12871 (pass_lim): Replace declaration with that of...
12872 (make_pass_lim): ...new function.
12873 (pass_tree_unswitch): Replace declaration with that of...
12874 (make_pass_tree_unswitch): ...new function.
12875 (pass_predcom): Replace declaration with that of...
12876 (make_pass_predcom): ...new function.
12877 (pass_iv_canon): Replace declaration with that of...
12878 (make_pass_iv_canon): ...new function.
12879 (pass_scev_cprop): Replace declaration with that of...
12880 (make_pass_scev_cprop): ...new function.
12881 (pass_empty_loop): Replace declaration with that of...
12882 (make_pass_empty_loop): ...new function.
12883 (pass_record_bounds): Replace declaration with that of...
12884 (make_pass_record_bounds): ...new function.
12885 (pass_graphite): Replace declaration with that of...
12886 (make_pass_graphite): ...new function.
12887 (pass_graphite_transforms): Replace declaration with that of...
12888 (make_pass_graphite_transforms): ...new function.
12889 (pass_if_conversion): Replace declaration with that of...
12890 (make_pass_if_conversion): ...new function.
12891 (pass_loop_distribution): Replace declaration with that of...
12892 (make_pass_loop_distribution): ...new function.
12893 (pass_vectorize): Replace declaration with that of...
12894 (make_pass_vectorize): ...new function.
12895 (pass_slp_vectorize): Replace declaration with that of...
12896 (make_pass_slp_vectorize): ...new function.
12897 (pass_complete_unroll): Replace declaration with that of...
12898 (make_pass_complete_unroll): ...new function.
12899 (pass_complete_unrolli): Replace declaration with that of...
12900 (make_pass_complete_unrolli): ...new function.
12901 (pass_parallelize_loops): Replace declaration with that of...
12902 (make_pass_parallelize_loops): ...new function.
12903 (pass_loop_prefetch): Replace declaration with that of...
12904 (make_pass_loop_prefetch): ...new function.
12905 (pass_iv_optimize): Replace declaration with that of...
12906 (make_pass_iv_optimize): ...new function.
12907 (pass_tree_loop_done): Replace declaration with that of...
12908 (make_pass_tree_loop_done): ...new function.
12909 (pass_ch): Replace declaration with that of...
12910 (make_pass_ch): ...new function.
12911 (pass_ccp): Replace declaration with that of...
12912 (make_pass_ccp): ...new function.
12913 (pass_phi_only_cprop): Replace declaration with that of...
12914 (make_pass_phi_only_cprop): ...new function.
12915 (pass_build_ssa): Replace declaration with that of...
12916 (make_pass_build_ssa): ...new function.
12917 (pass_build_alias): Replace declaration with that of...
12918 (make_pass_build_alias): ...new function.
12919 (pass_build_ealias): Replace declaration with that of...
12920 (make_pass_build_ealias): ...new function.
12921 (pass_dominator): Replace declaration with that of...
12922 (make_pass_dominator): ...new function.
12923 (pass_dce): Replace declaration with that of...
12924 (make_pass_dce): ...new function.
12925 (pass_dce_loop): Replace declaration with that of...
12926 (make_pass_dce_loop): ...new function.
12927 (pass_cd_dce): Replace declaration with that of...
12928 (make_pass_cd_dce): ...new function.
12929 (pass_call_cdce): Replace declaration with that of...
12930 (make_pass_call_cdce): ...new function.
12931 (pass_merge_phi): Replace declaration with that of...
12932 (make_pass_merge_phi): ...new function.
12933 (pass_split_crit_edges): Replace declaration with that of...
12934 (make_pass_split_crit_edges): ...new function.
12935 (pass_pre): Replace declaration with that of...
12936 (make_pass_pre): ...new function.
12937 (pass_profile): Replace declaration with that of...
12938 (make_pass_profile): ...new function.
12939 (pass_strip_predict_hints): Replace declaration with that of...
12940 (make_pass_strip_predict_hints): ...new function.
12941 (pass_lower_complex_O0): Replace declaration with that of...
12942 (make_pass_lower_complex_O0): ...new function.
12943 (pass_lower_complex): Replace declaration with that of...
12944 (make_pass_lower_complex): ...new function.
12945 (pass_lower_vector): Replace declaration with that of...
12946 (make_pass_lower_vector): ...new function.
12947 (pass_lower_vector_ssa): Replace declaration with that of...
12948 (make_pass_lower_vector_ssa): ...new function.
12949 (pass_lower_omp): Replace declaration with that of...
12950 (make_pass_lower_omp): ...new function.
12951 (pass_diagnose_omp_blocks): Replace declaration with that of...
12952 (make_pass_diagnose_omp_blocks): ...new function.
12953 (pass_expand_omp): Replace declaration with that of...
12954 (make_pass_expand_omp): ...new function.
12955 (pass_expand_omp_ssa): Replace declaration with that of...
12956 (make_pass_expand_omp_ssa): ...new function.
12957 (pass_object_sizes): Replace declaration with that of...
12958 (make_pass_object_sizes): ...new function.
12959 (pass_strlen): Replace declaration with that of...
12960 (make_pass_strlen): ...new function.
12961 (pass_fold_builtins): Replace declaration with that of...
12962 (make_pass_fold_builtins): ...new function.
12963 (pass_stdarg): Replace declaration with that of...
12964 (make_pass_stdarg): ...new function.
12965 (pass_early_warn_uninitialized): Replace declaration with that of...
12966 (make_pass_early_warn_uninitialized): ...new function.
12967 (pass_late_warn_uninitialized): Replace declaration with that of...
12968 (make_pass_late_warn_uninitialized): ...new function.
12969 (pass_cse_reciprocals): Replace declaration with that of...
12970 (make_pass_cse_reciprocals): ...new function.
12971 (pass_cse_sincos): Replace declaration with that of...
12972 (make_pass_cse_sincos): ...new function.
12973 (pass_optimize_bswap): Replace declaration with that of...
12974 (make_pass_optimize_bswap): ...new function.
12975 (pass_optimize_widening_mul): Replace declaration with that of...
12976 (make_pass_optimize_widening_mul): ...new function.
12977 (pass_warn_function_return): Replace declaration with that of...
12978 (make_pass_warn_function_return): ...new function.
12979 (pass_warn_function_noreturn): Replace declaration with that of...
12980 (make_pass_warn_function_noreturn): ...new function.
12981 (pass_cselim): Replace declaration with that of...
12982 (make_pass_cselim): ...new function.
12983 (pass_phiopt): Replace declaration with that of...
12984 (make_pass_phiopt): ...new function.
12985 (pass_forwprop): Replace declaration with that of...
12986 (make_pass_forwprop): ...new function.
12987 (pass_phiprop): Replace declaration with that of...
12988 (make_pass_phiprop): ...new function.
12989 (pass_tree_ifcombine): Replace declaration with that of...
12990 (make_pass_tree_ifcombine): ...new function.
12991 (pass_dse): Replace declaration with that of...
12992 (make_pass_dse): ...new function.
12993 (pass_nrv): Replace declaration with that of...
12994 (make_pass_nrv): ...new function.
12995 (pass_rename_ssa_copies): Replace declaration with that of...
12996 (make_pass_rename_ssa_copies): ...new function.
12997 (pass_sink_code): Replace declaration with that of...
12998 (make_pass_sink_code): ...new function.
12999 (pass_fre): Replace declaration with that of...
13000 (make_pass_fre): ...new function.
13001 (pass_check_data_deps): Replace declaration with that of...
13002 (make_pass_check_data_deps): ...new function.
13003 (pass_copy_prop): Replace declaration with that of...
13004 (make_pass_copy_prop): ...new function.
13005 (pass_vrp): Replace declaration with that of...
13006 (make_pass_vrp): ...new function.
13007 (pass_uncprop): Replace declaration with that of...
13008 (make_pass_uncprop): ...new function.
13009 (pass_return_slot): Replace declaration with that of...
13010 (make_pass_return_slot): ...new function.
13011 (pass_reassoc): Replace declaration with that of...
13012 (make_pass_reassoc): ...new function.
13013 (pass_rebuild_cgraph_edges): Replace declaration with that of...
13014 (make_pass_rebuild_cgraph_edges): ...new function.
13015 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
13016 (make_pass_remove_cgraph_callee_edges): ...new function.
13017 (pass_build_cgraph_edges): Replace declaration with that of...
13018 (make_pass_build_cgraph_edges): ...new function.
13019 (pass_local_pure_const): Replace declaration with that of...
13020 (make_pass_local_pure_const): ...new function.
13021 (pass_tracer): Replace declaration with that of...
13022 (make_pass_tracer): ...new function.
13023 (pass_warn_unused_result): Replace declaration with that of...
13024 (make_pass_warn_unused_result): ...new function.
13025 (pass_diagnose_tm_blocks): Replace declaration with that of...
13026 (make_pass_diagnose_tm_blocks): ...new function.
13027 (pass_lower_tm): Replace declaration with that of...
13028 (make_pass_lower_tm): ...new function.
13029 (pass_tm_init): Replace declaration with that of...
13030 (make_pass_tm_init): ...new function.
13031 (pass_tm_mark): Replace declaration with that of...
13032 (make_pass_tm_mark): ...new function.
13033 (pass_tm_memopt): Replace declaration with that of...
13034 (make_pass_tm_memopt): ...new function.
13035 (pass_tm_edges): Replace declaration with that of...
13036 (make_pass_tm_edges): ...new function.
13037 (pass_split_functions): Replace declaration with that of...
13038 (make_pass_split_functions): ...new function.
13039 (pass_feedback_split_functions): Replace declaration with that of...
13040 (make_pass_feedback_split_functions): ...new function.
13041 (pass_strength_reduction): Replace declaration with that of...
13042 (make_pass_strength_reduction): ...new function.
13043 (pass_ipa_lower_emutls): Replace declaration with that of...
13044 (make_pass_ipa_lower_emutls): ...new function.
13045 (pass_ipa_function_and_variable_visibility): Replace declaration with
13047 (make_pass_ipa_function_and_variable_visibility): ...new function.
13048 (pass_ipa_tree_profile): Replace declaration with that of...
13049 (make_pass_ipa_tree_profile): ...new function.
13050 (pass_early_local_passes): Replace declaration with that of...
13051 (make_pass_early_local_passes): ...new function.
13052 (pass_ipa_whole_program_visibility): Replace declaration with that
13054 (make_pass_ipa_whole_program_visibility): ...new function.
13055 (pass_ipa_lto_gimple_out): Replace declaration with that of...
13056 (make_pass_ipa_lto_gimple_out): ...new function.
13057 (pass_ipa_increase_alignment): Replace declaration with that of...
13058 (make_pass_ipa_increase_alignment): ...new function.
13059 (pass_ipa_inline): Replace declaration with that of...
13060 (make_pass_ipa_inline): ...new function.
13061 (pass_ipa_free_lang_data): Replace declaration with that of...
13062 (make_pass_ipa_free_lang_data): ...new function.
13063 (pass_ipa_free_inline_summary): Replace declaration with that of...
13064 (make_pass_ipa_free_inline_summary): ...new function.
13065 (pass_ipa_cp): Replace declaration with that of...
13066 (make_pass_ipa_cp): ...new function.
13067 (pass_ipa_reference): Replace declaration with that of...
13068 (make_pass_ipa_reference): ...new function.
13069 (pass_ipa_pure_const): Replace declaration with that of...
13070 (make_pass_ipa_pure_const): ...new function.
13071 (pass_ipa_pta): Replace declaration with that of...
13072 (make_pass_ipa_pta): ...new function.
13073 (pass_ipa_lto_finish_out): Replace declaration with that of...
13074 (make_pass_ipa_lto_finish_out): ...new function.
13075 (pass_ipa_tm): Replace declaration with that of...
13076 (make_pass_ipa_tm): ...new function.
13077 (pass_ipa_profile): Replace declaration with that of...
13078 (make_pass_ipa_profile): ...new function.
13079 (pass_ipa_cdtor_merge): Replace declaration with that of...
13080 (make_pass_ipa_cdtor_merge): ...new function.
13081 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
13083 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13084 (pass_init_datastructures): Replace declaration with that of...
13085 (make_pass_init_datastructures): ...new function.
13086 (pass_fixup_cfg): Replace declaration with that of...
13087 (make_pass_fixup_cfg): ...new function.
13088 (pass_expand): Replace declaration with that of...
13089 (make_pass_expand): ...new function.
13090 (pass_instantiate_virtual_regs): Replace declaration with that of...
13091 (make_pass_instantiate_virtual_regs): ...new function.
13092 (pass_rtl_fwprop): Replace declaration with that of...
13093 (make_pass_rtl_fwprop): ...new function.
13094 (pass_rtl_fwprop_addr): Replace declaration with that of...
13095 (make_pass_rtl_fwprop_addr): ...new function.
13096 (pass_jump): Replace declaration with that of...
13097 (make_pass_jump): ...new function.
13098 (pass_jump2): Replace declaration with that of...
13099 (make_pass_jump2): ...new function.
13100 (pass_lower_subreg): Replace declaration with that of...
13101 (make_pass_lower_subreg): ...new function.
13102 (pass_cse): Replace declaration with that of...
13103 (make_pass_cse): ...new function.
13104 (pass_fast_rtl_dce): Replace declaration with that of...
13105 (make_pass_fast_rtl_dce): ...new function.
13106 (pass_ud_rtl_dce): Replace declaration with that of...
13107 (make_pass_ud_rtl_dce): ...new function.
13108 (pass_rtl_dce): Replace declaration with that of...
13109 (make_pass_rtl_dce): ...new function.
13110 (pass_rtl_dse1): Replace declaration with that of...
13111 (make_pass_rtl_dse1): ...new function.
13112 (pass_rtl_dse2): Replace declaration with that of...
13113 (make_pass_rtl_dse2): ...new function.
13114 (pass_rtl_dse3): Replace declaration with that of...
13115 (make_pass_rtl_dse3): ...new function.
13116 (pass_rtl_cprop): Replace declaration with that of...
13117 (make_pass_rtl_cprop): ...new function.
13118 (pass_rtl_pre): Replace declaration with that of...
13119 (make_pass_rtl_pre): ...new function.
13120 (pass_rtl_hoist): Replace declaration with that of...
13121 (make_pass_rtl_hoist): ...new function.
13122 (pass_rtl_store_motion): Replace declaration with that of...
13123 (make_pass_rtl_store_motion): ...new function.
13124 (pass_cse_after_global_opts): Replace declaration with that of...
13125 (make_pass_cse_after_global_opts): ...new function.
13126 (pass_rtl_ifcvt): Replace declaration with that of...
13127 (make_pass_rtl_ifcvt): ...new function.
13128 (pass_into_cfg_layout_mode): Replace declaration with that of...
13129 (make_pass_into_cfg_layout_mode): ...new function.
13130 (pass_outof_cfg_layout_mode): Replace declaration with that of...
13131 (make_pass_outof_cfg_layout_mode): ...new function.
13132 (pass_loop2): Replace declaration with that of...
13133 (make_pass_loop2): ...new function.
13134 (pass_rtl_loop_init): Replace declaration with that of...
13135 (make_pass_rtl_loop_init): ...new function.
13136 (pass_rtl_move_loop_invariants): Replace declaration with that of...
13137 (make_pass_rtl_move_loop_invariants): ...new function.
13138 (pass_rtl_unswitch): Replace declaration with that of...
13139 (make_pass_rtl_unswitch): ...new function.
13140 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
13141 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13142 (pass_rtl_doloop): Replace declaration with that of...
13143 (make_pass_rtl_doloop): ...new function.
13144 (pass_rtl_loop_done): Replace declaration with that of...
13145 (make_pass_rtl_loop_done): ...new function.
13146 (pass_web): Replace declaration with that of...
13147 (make_pass_web): ...new function.
13148 (pass_cse2): Replace declaration with that of...
13149 (make_pass_cse2): ...new function.
13150 (pass_df_initialize_opt): Replace declaration with that of...
13151 (make_pass_df_initialize_opt): ...new function.
13152 (pass_df_initialize_no_opt): Replace declaration with that of...
13153 (make_pass_df_initialize_no_opt): ...new function.
13154 (pass_reginfo_init): Replace declaration with that of...
13155 (make_pass_reginfo_init): ...new function.
13156 (pass_inc_dec): Replace declaration with that of...
13157 (make_pass_inc_dec): ...new function.
13158 (pass_stack_ptr_mod): Replace declaration with that of...
13159 (make_pass_stack_ptr_mod): ...new function.
13160 (pass_initialize_regs): Replace declaration with that of...
13161 (make_pass_initialize_regs): ...new function.
13162 (pass_combine): Replace declaration with that of...
13163 (make_pass_combine): ...new function.
13164 (pass_if_after_combine): Replace declaration with that of...
13165 (make_pass_if_after_combine): ...new function.
13166 (pass_ree): Replace declaration with that of...
13167 (make_pass_ree): ...new function.
13168 (pass_partition_blocks): Replace declaration with that of...
13169 (make_pass_partition_blocks): ...new function.
13170 (pass_match_asm_constraints): Replace declaration with that of...
13171 (make_pass_match_asm_constraints): ...new function.
13172 (pass_regmove): Replace declaration with that of...
13173 (make_pass_regmove): ...new function.
13174 (pass_split_all_insns): Replace declaration with that of...
13175 (make_pass_split_all_insns): ...new function.
13176 (pass_fast_rtl_byte_dce): Replace declaration with that of...
13177 (make_pass_fast_rtl_byte_dce): ...new function.
13178 (pass_lower_subreg2): Replace declaration with that of...
13179 (make_pass_lower_subreg2): ...new function.
13180 (pass_mode_switching): Replace declaration with that of...
13181 (make_pass_mode_switching): ...new function.
13182 (pass_sms): Replace declaration with that of...
13183 (make_pass_sms): ...new function.
13184 (pass_sched): Replace declaration with that of...
13185 (make_pass_sched): ...new function.
13186 (pass_ira): Replace declaration with that of...
13187 (make_pass_ira): ...new function.
13188 (pass_reload): Replace declaration with that of...
13189 (make_pass_reload): ...new function.
13190 (pass_clean_state): Replace declaration with that of...
13191 (make_pass_clean_state): ...new function.
13192 (pass_branch_prob): Replace declaration with that of...
13193 (make_pass_branch_prob): ...new function.
13194 (pass_value_profile_transformations): Replace declaration with that
13196 (make_pass_value_profile_transformations): ...new function.
13197 (pass_postreload_cse): Replace declaration with that of...
13198 (make_pass_postreload_cse): ...new function.
13199 (pass_gcse2): Replace declaration with that of...
13200 (make_pass_gcse2): ...new function.
13201 (pass_split_after_reload): Replace declaration with that of...
13202 (make_pass_split_after_reload): ...new function.
13203 (pass_branch_target_load_optimize1): Replace declaration with that
13205 (make_pass_branch_target_load_optimize1): ...new function.
13206 (pass_thread_prologue_and_epilogue): Replace declaration with that
13208 (make_pass_thread_prologue_and_epilogue): ...new function.
13209 (pass_stack_adjustments): Replace declaration with that of...
13210 (make_pass_stack_adjustments): ...new function.
13211 (pass_peephole2): Replace declaration with that of...
13212 (make_pass_peephole2): ...new function.
13213 (pass_if_after_reload): Replace declaration with that of...
13214 (make_pass_if_after_reload): ...new function.
13215 (pass_regrename): Replace declaration with that of...
13216 (make_pass_regrename): ...new function.
13217 (pass_cprop_hardreg): Replace declaration with that of...
13218 (make_pass_cprop_hardreg): ...new function.
13219 (pass_reorder_blocks): Replace declaration with that of...
13220 (make_pass_reorder_blocks): ...new function.
13221 (pass_branch_target_load_optimize2): Replace declaration with that
13223 (make_pass_branch_target_load_optimize2): ...new function.
13224 (pass_leaf_regs): Replace declaration with that of...
13225 (make_pass_leaf_regs): ...new function.
13226 (pass_split_before_sched2): Replace declaration with that of...
13227 (make_pass_split_before_sched2): ...new function.
13228 (pass_compare_elim_after_reload): Replace declaration with that of...
13229 (make_pass_compare_elim_after_reload): ...new function.
13230 (pass_sched2): Replace declaration with that of...
13231 (make_pass_sched2): ...new function.
13232 (pass_stack_regs): Replace declaration with that of...
13233 (make_pass_stack_regs): ...new function.
13234 (pass_stack_regs_run): Replace declaration with that of...
13235 (make_pass_stack_regs_run): ...new function.
13236 (pass_df_finish): Replace declaration with that of...
13237 (make_pass_df_finish): ...new function.
13238 (pass_compute_alignments): Replace declaration with that of...
13239 (make_pass_compute_alignments): ...new function.
13240 (pass_duplicate_computed_gotos): Replace declaration with that of...
13241 (make_pass_duplicate_computed_gotos): ...new function.
13242 (pass_variable_tracking): Replace declaration with that of...
13243 (make_pass_variable_tracking): ...new function.
13244 (pass_free_cfg): Replace declaration with that of...
13245 (make_pass_free_cfg): ...new function.
13246 (pass_machine_reorg): Replace declaration with that of...
13247 (make_pass_machine_reorg): ...new function.
13248 (pass_cleanup_barriers): Replace declaration with that of...
13249 (make_pass_cleanup_barriers): ...new function.
13250 (pass_delay_slots): Replace declaration with that of...
13251 (make_pass_delay_slots): ...new function.
13252 (pass_split_for_shorten_branches): Replace declaration with that of...
13253 (make_pass_split_for_shorten_branches): ...new function.
13254 (pass_split_before_regstack): Replace declaration with that of...
13255 (make_pass_split_before_regstack): ...new function.
13256 (pass_convert_to_eh_region_ranges): Replace declaration with that
13258 (make_pass_convert_to_eh_region_ranges): ...new function.
13259 (pass_shorten_branches): Replace declaration with that of...
13260 (make_pass_shorten_branches): ...new function.
13261 (pass_set_nothrow_function_flags): Replace declaration with that of...
13262 (make_pass_set_nothrow_function_flags): ...new function.
13263 (pass_dwarf2_frame): Replace declaration with that of...
13264 (make_pass_dwarf2_frame): ...new function.
13265 (pass_final): Replace declaration with that of...
13266 (make_pass_final): ...new function.
13267 (pass_rtl_seqabstr): Replace declaration with that of...
13268 (make_pass_rtl_seqabstr): ...new function.
13269 (pass_release_ssa_names): Replace declaration with that of...
13270 (make_pass_release_ssa_names): ...new function.
13271 (pass_early_inline): Replace declaration with that of...
13272 (make_pass_early_inline): ...new function.
13273 (pass_inline_parameters): Replace declaration with that of...
13274 (make_pass_inline_parameters): ...new function.
13275 (pass_update_address_taken): Replace declaration with that of...
13276 (make_pass_update_address_taken): ...new function.
13277 (pass_convert_switch): Replace declaration with that of...
13278 (make_pass_convert_switch): ...new function.
13279 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
13280 to a subclass of simple_ipa_opt_pass along with...
13281 (pass_data_ipa_tree_profile): ...new pass_data instance and...
13282 (make_pass_ipa_tree_profile): ...new function.
13283 * tree-sra.c (pass_sra_early): Convert from a global struct to a
13284 subclass of gimple_opt_pass along with...
13285 (pass_data_sra_early): ...new pass_data instance and...
13286 (make_pass_sra_early): ...new function.
13287 (pass_sra): Convert from a global struct to a subclass of
13288 gimple_opt_pass along with...
13289 (pass_data_sra): ...new pass_data instance and...
13290 (make_pass_sra): ...new function.
13291 (pass_early_ipa_sra): Convert from a global struct to a subclass of
13292 gimple_opt_pass along with...
13293 (pass_data_early_ipa_sra): ...new pass_data instance and...
13294 (make_pass_early_ipa_sra): ...new function.
13295 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
13296 subclass of gimple_opt_pass along with...
13297 (pass_data_ccp): ...new pass_data instance and...
13298 (make_pass_ccp): ...new function.
13299 (pass_fold_builtins): Convert from a global struct to a subclass of
13300 gimple_opt_pass along with...
13301 (pass_data_fold_builtins): ...new pass_data instance and...
13302 (make_pass_fold_builtins): ...new function.
13303 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
13304 subclass of gimple_opt_pass along with...
13305 (pass_data_copy_prop): ...new pass_data instance and...
13306 (make_pass_copy_prop): ...new function.
13307 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
13308 global struct to a subclass of gimple_opt_pass along with...
13309 (pass_data_rename_ssa_copies): ...new pass_data instance and...
13310 (make_pass_rename_ssa_copies): ...new function.
13311 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
13312 subclass of gimple_opt_pass along with...
13313 (pass_data_dce): ...new pass_data instance and...
13314 (make_pass_dce): ...new function.
13315 (pass_dce_loop): Convert from a global struct to a subclass of
13316 gimple_opt_pass along with...
13317 (pass_data_dce_loop): ...new pass_data instance and...
13318 (make_pass_dce_loop): ...new function.
13319 (pass_cd_dce): Convert from a global struct to a subclass of
13320 gimple_opt_pass along with...
13321 (pass_data_cd_dce): ...new pass_data instance and...
13322 (make_pass_cd_dce): ...new function.
13323 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
13324 subclass of gimple_opt_pass along with...
13325 (pass_data_dominator): ...new pass_data instance and...
13326 (make_pass_dominator): ...new function.
13327 (pass_phi_only_cprop): Convert from a global struct to a subclass of
13328 gimple_opt_pass along with...
13329 (pass_data_phi_only_cprop): ...new pass_data instance and...
13330 (make_pass_phi_only_cprop): ...new function.
13331 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
13332 subclass of gimple_opt_pass along with...
13333 (pass_data_dse): ...new pass_data instance and...
13334 (make_pass_dse): ...new function.
13335 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
13336 a subclass of gimple_opt_pass along with...
13337 (pass_data_forwprop): ...new pass_data instance and...
13338 (make_pass_forwprop): ...new function.
13339 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
13340 struct to a subclass of gimple_opt_pass along with...
13341 (pass_data_tree_ifcombine): ...new pass_data instance and...
13342 (make_pass_tree_ifcombine): ...new function.
13343 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
13344 subclass of gimple_opt_pass along with...
13345 (pass_data_ch): ...new pass_data instance and...
13346 (make_pass_ch): ...new function.
13347 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
13348 subclass of gimple_opt_pass along with...
13349 (pass_data_tree_loop): ...new pass_data instance and...
13350 (make_pass_tree_loop): ...new function.
13351 (pass_tree_loop_init): Convert from a global struct to a subclass of
13352 gimple_opt_pass along with...
13353 (pass_data_tree_loop_init): ...new pass_data instance and...
13354 (make_pass_tree_loop_init): ...new function.
13355 (pass_lim): Convert from a global struct to a subclass of
13356 gimple_opt_pass along with...
13357 (pass_data_lim): ...new pass_data instance and...
13358 (make_pass_lim): ...new function.
13359 (pass_tree_unswitch): Convert from a global struct to a subclass of
13360 gimple_opt_pass along with...
13361 (pass_data_tree_unswitch): ...new pass_data instance and...
13362 (make_pass_tree_unswitch): ...new function.
13363 (pass_predcom): Convert from a global struct to a subclass of
13364 gimple_opt_pass along with...
13365 (pass_data_predcom): ...new pass_data instance and...
13366 (make_pass_predcom): ...new function.
13367 (pass_vectorize): Convert from a global struct to a subclass of
13368 gimple_opt_pass along with...
13369 (pass_data_vectorize): ...new pass_data instance and...
13370 (make_pass_vectorize): ...new function.
13371 (pass_graphite): Convert from a global struct to a subclass of
13372 gimple_opt_pass along with...
13373 (pass_data_graphite): ...new pass_data instance and...
13374 (make_pass_graphite): ...new function.
13375 (pass_graphite_transforms): Convert from a global struct to a subclass
13376 of gimple_opt_pass along with...
13377 (pass_data_graphite_transforms): ...new pass_data instance and...
13378 (make_pass_graphite_transforms): ...new function.
13379 (pass_check_data_deps): Convert from a global struct to a subclass of
13380 gimple_opt_pass along with...
13381 (pass_data_check_data_deps): ...new pass_data instance and...
13382 (make_pass_check_data_deps): ...new function.
13383 (pass_iv_canon): Convert from a global struct to a subclass of
13384 gimple_opt_pass along with...
13385 (pass_data_iv_canon): ...new pass_data instance and...
13386 (make_pass_iv_canon): ...new function.
13387 (pass_scev_cprop): Convert from a global struct to a subclass of
13388 gimple_opt_pass along with...
13389 (pass_data_scev_cprop): ...new pass_data instance and...
13390 (make_pass_scev_cprop): ...new function.
13391 (pass_record_bounds): Convert from a global struct to a subclass of
13392 gimple_opt_pass along with...
13393 (pass_data_record_bounds): ...new pass_data instance and...
13394 (make_pass_record_bounds): ...new function.
13395 (pass_complete_unroll): Convert from a global struct to a subclass of
13396 gimple_opt_pass along with...
13397 (pass_data_complete_unroll): ...new pass_data instance and...
13398 (make_pass_complete_unroll): ...new function.
13399 (pass_complete_unrolli): Convert from a global struct to a subclass of
13400 gimple_opt_pass along with...
13401 (pass_data_complete_unrolli): ...new pass_data instance and...
13402 (make_pass_complete_unrolli): ...new function.
13403 (pass_parallelize_loops): Convert from a global struct to a subclass
13404 of gimple_opt_pass along with...
13405 (pass_data_parallelize_loops): ...new pass_data instance and...
13406 (make_pass_parallelize_loops): ...new function.
13407 (pass_loop_prefetch): Convert from a global struct to a subclass of
13408 gimple_opt_pass along with...
13409 (pass_data_loop_prefetch): ...new pass_data instance and...
13410 (make_pass_loop_prefetch): ...new function.
13411 (pass_iv_optimize): Convert from a global struct to a subclass of
13412 gimple_opt_pass along with...
13413 (pass_data_iv_optimize): ...new pass_data instance and...
13414 (make_pass_iv_optimize): ...new function.
13415 (pass_tree_loop_done): Convert from a global struct to a subclass of
13416 gimple_opt_pass along with...
13417 (pass_data_tree_loop_done): ...new pass_data instance and...
13418 (make_pass_tree_loop_done): ...new function.
13419 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
13420 struct to a subclass of gimple_opt_pass along with...
13421 (pass_data_cse_reciprocals): ...new pass_data instance and...
13422 (make_pass_cse_reciprocals): ...new function.
13423 (pass_cse_sincos): Convert from a global struct to a subclass of
13424 gimple_opt_pass along with...
13425 (pass_data_cse_sincos): ...new pass_data instance and...
13426 (make_pass_cse_sincos): ...new function.
13427 (pass_optimize_bswap): Convert from a global struct to a subclass of
13428 gimple_opt_pass along with...
13429 (pass_data_optimize_bswap): ...new pass_data instance and...
13430 (make_pass_optimize_bswap): ...new function.
13431 (pass_optimize_widening_mul): Convert from a global struct to a
13432 subclass of gimple_opt_pass along with...
13433 (pass_data_optimize_widening_mul): ...new pass_data instance and...
13434 (make_pass_optimize_widening_mul): ...new function.
13435 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
13436 subclass of gimple_opt_pass along with...
13437 (pass_data_phiopt): ...new pass_data instance and...
13438 (make_pass_phiopt): ...new function.
13439 (pass_cselim): Convert from a global struct to a subclass of
13440 gimple_opt_pass along with...
13441 (pass_data_cselim): ...new pass_data instance and...
13442 (make_pass_cselim): ...new function.
13443 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
13444 subclass of gimple_opt_pass along with...
13445 (pass_data_phiprop): ...new pass_data instance and...
13446 (make_pass_phiprop): ...new function.
13447 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
13448 subclass of gimple_opt_pass along with...
13449 (pass_data_pre): ...new pass_data instance and...
13450 (make_pass_pre): ...new function.
13451 (pass_fre): Convert from a global struct to a subclass of
13452 gimple_opt_pass along with...
13453 (pass_data_fre): ...new pass_data instance and...
13454 (make_pass_fre): ...new function.
13455 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
13456 subclass of gimple_opt_pass along with...
13457 (pass_data_reassoc): ...new pass_data instance and...
13458 (make_pass_reassoc): ...new function.
13459 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
13460 subclass of gimple_opt_pass along with...
13461 (pass_data_sink_code): ...new pass_data instance and...
13462 (make_pass_sink_code): ...new function.
13463 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
13464 subclass of gimple_opt_pass along with...
13465 (pass_data_strlen): ...new pass_data instance and...
13466 (make_pass_strlen): ...new function.
13467 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
13468 struct to a subclass of gimple_opt_pass along with...
13469 (pass_data_build_alias): ...new pass_data instance and...
13470 (make_pass_build_alias): ...new function.
13471 (pass_build_ealias): Convert from a global struct to a subclass of
13472 gimple_opt_pass along with...
13473 (pass_data_build_ealias): ...new pass_data instance and...
13474 (make_pass_build_ealias): ...new function.
13475 (pass_ipa_pta): Convert from a global struct to a subclass of
13476 simple_ipa_opt_pass along with...
13477 (pass_data_ipa_pta): ...new pass_data instance and...
13478 (make_pass_ipa_pta): ...new function.
13479 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
13480 subclass of gimple_opt_pass along with...
13481 (pass_data_uncprop): ...new pass_data instance and...
13482 (make_pass_uncprop): ...new function.
13483 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
13484 global struct to a subclass of gimple_opt_pass along with...
13485 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
13486 (make_pass_late_warn_uninitialized): ...new function.
13487 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
13488 to a subclass of gimple_opt_pass along with...
13489 (pass_data_init_datastructures): ...new pass_data instance and...
13490 (make_pass_init_datastructures): ...new function.
13491 (pass_early_warn_uninitialized): Convert from a global struct to a
13492 subclass of gimple_opt_pass along with...
13493 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
13494 (make_pass_early_warn_uninitialized): ...new function.
13495 (pass_update_address_taken): Convert from a global struct to a
13496 subclass of gimple_opt_pass along with...
13497 (pass_data_update_address_taken): ...new pass_data instance and...
13498 (make_pass_update_address_taken): ...new function.
13499 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
13500 struct to a subclass of gimple_opt_pass along with...
13501 (pass_data_release_ssa_names): ...new pass_data instance and...
13502 (make_pass_release_ssa_names): ...new function.
13503 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
13504 subclass of gimple_opt_pass along with...
13505 (pass_data_stdarg): ...new pass_data instance and...
13506 (make_pass_stdarg): ...new function.
13507 * tree-switch-conversion.c (pass_convert_switch): Convert from a
13508 global struct to a subclass of gimple_opt_pass along with...
13509 (pass_data_convert_switch): ...new pass_data instance and...
13510 (make_pass_convert_switch): ...new function.
13511 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
13512 to a subclass of gimple_opt_pass along with...
13513 (pass_data_tail_recursion): ...new pass_data instance and...
13514 (make_pass_tail_recursion): ...new function.
13515 (pass_tail_calls): Convert from a global struct to a subclass of
13516 gimple_opt_pass along with...
13517 (pass_data_tail_calls): ...new pass_data instance and...
13518 (make_pass_tail_calls): ...new function.
13519 * tree-vect-generic.c (pass_lower_vector): Convert from a global
13520 struct to a subclass of gimple_opt_pass along with...
13521 (pass_data_lower_vector): ...new pass_data instance and...
13522 (make_pass_lower_vector): ...new function.
13523 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
13524 gimple_opt_pass along with...
13525 (pass_data_lower_vector_ssa): ...new pass_data instance and...
13526 (make_pass_lower_vector_ssa): ...new function.
13527 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
13528 to a subclass of gimple_opt_pass along with...
13529 (pass_data_slp_vectorize): ...new pass_data instance and...
13530 (make_pass_slp_vectorize): ...new function.
13531 (pass_ipa_increase_alignment): Convert from a global struct to a
13532 subclass of simple_ipa_opt_pass along with...
13533 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
13534 (make_pass_ipa_increase_alignment): ...new function.
13535 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
13536 gimple_opt_pass along with...
13537 (pass_data_vrp): ...new pass_data instance and...
13538 (make_pass_vrp): ...new function.
13539 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
13540 subclass of simple_ipa_opt_pass along with...
13541 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
13542 (make_pass_ipa_free_lang_data): ...new function.
13543 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
13544 gimple_opt_pass along with...
13545 (pass_data_tsan): ...new pass_data instance and...
13546 (make_pass_tsan): ...new function.
13547 (pass_tsan_O0): Convert from a global struct to a subclass of
13548 gimple_opt_pass along with...
13549 (pass_data_tsan_O0): ...new pass_data instance and...
13550 (make_pass_tsan_O0): ...new function.
13551 * var-tracking.c (pass_variable_tracking): Convert from a global
13552 struct to a subclass of rtl_opt_pass along with...
13553 (pass_data_variable_tracking): ...new pass_data instance and...
13554 (make_pass_variable_tracking): ...new function.
13555 * web.c (pass_web): Convert from a global struct to a subclass of
13556 rtl_opt_pass along with...
13557 (pass_data_web): ...new pass_data instance and...
13558 (make_pass_web): ...new function.
13559 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
13560 declaration with that of...
13561 (make_pass_mode_switch_use): ...new function.
13562 (pass_resolve_sw_modes): Replace declaration with that of...
13563 (make_pass_resolve_sw_modes): ...new function.
13564 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
13565 from a global struct to a subclass of rtl_opt_pass along with...
13566 (pass_data_mode_switch_use): ...new pass_data instance and...
13567 (make_pass_mode_switch_use): ...new function.
13568 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
13569 from a global struct to a subclass of rtl_opt_pass along with...
13570 (pass_data_resolve_sw_modes): ...new pass_data instance and...
13571 (make_pass_resolve_sw_modes): ...new function.
13572 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
13573 struct to a subclass of rtl_opt_pass along with...
13574 (pass_data_insert_vzeroupper): ...new pass_data instance and...
13575 (make_pass_insert_vzeroupper): ...new function.
13576 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
13577 global struct to a subclass of rtl_opt_pass along with...
13578 (pass_data_work_around_errata): ...new pass_data instance and...
13579 (make_pass_work_around_errata): ...new function.
13580 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
13581 struct to a subclass of rtl_opt_pass along with...
13582 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
13583 (make_pass_mips_machine_reorg2): ...new function.
13585 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13587 * passes.c (pass_manager::operator new): New.
13589 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13591 Handwritten part of conversion of passes to C++ classes.
13593 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
13594 (toplev.o): Add dep on PASS_MANAGER_H.
13595 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
13596 of early local pases to reflect this moving from a global to a
13597 member of gcc::pass_manager.
13598 (cgraph_add_new_function): Likewise.
13599 * lto-cgraph.c (lto_output_node): Update for conversion of
13600 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
13601 * passes.c (opt_pass::clone): New.
13602 (opt_pass::gate): New.
13603 (opt_pass::execute): New.
13604 (opt_pass::opt_pass): New.
13605 (pass_manager::execute_early_local_passes): New.
13606 (pass_manager::execute_pass_mode_switching): new.
13607 (finish_optimization_passes): Convert to...
13608 (pass_manager::finish_optimization_passes): ...this.
13609 (finish_optimization_passes): Update for conversion of passes to
13611 (register_dump_files_1): Use has_gate since we cannot portably
13612 check a vtable entry against NULL.
13613 (dump_one_pass): Likewise.
13614 (ipa_write_summaries_2): Likewise.
13615 (ipa_write_optimization_summaries_1): Likewise.
13616 (ipa_read_summaries_1): Likewise.
13617 (ipa_read_optimization_summaries_1): Likewise.
13618 (execute_ipa_stmt_fixups): Likewise.
13619 (pass_manager::pass_manager): Rewrite pass-creation, invoking
13620 pass-creation functions rather than wiring up globals, and
13621 storing the results in fields of pass_manager generated using
13622 pass-instances.def.
13623 (pass_manager::dump_profile_report): Update for conversion of
13624 passes to C++ classes.
13625 (pass_manager::execute_ipa_summary_passes): Likewise.
13626 (execute_one_ipa_transform_pass): Likewise.
13627 (execute_one_pass): Use has_gate and has_execute since we cannot
13628 portably check a vtable entry against NULL.
13629 * pass_manager.h (pass_manager::finish_optimization_passes): New.
13630 (pass_manager): Use pass-instances.def to add fields for the
13631 various pass instances.
13632 * toplev.c (finalize): Update for move of
13633 finish_optimization_passes to a method of gcc::pass_manager.
13634 * toplev.h (finish_optimization_passes): Move to method of class
13636 * tree-pass.h (struct pass_data): New.
13637 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
13638 (opt_pass::gate): Convert to virtual function.
13639 (opt_pass::~opt_pass): New.
13640 (opt_pass::clone): New.
13641 (opt_pass::execute): Convert to virtual function.
13642 (opt_pass::opt_pass): New.
13643 (opt_pass::ctxt_): new.
13644 (gimple_opt_pass): Convert to subclass of opt_pass.
13645 (gimple_opt_pass::gimple_opt_pass): New.
13646 (rtl_opt_pass): Convert to subclass of opt_pass.
13647 (rtl_opt_pass::rtl_opt_pass): New.
13648 (ipa_opt_pass_d): Convert to subclass of opt_pass.
13649 (ipa_opt_pass_d::ipa_opt_pass_d): New.
13650 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
13651 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
13652 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
13653 invocation of pass_mode_switching to reflect this moving from a
13654 global to a member of gcc::pass_manager.
13655 (ix86_option_override): Rework how pass_insert_vzeroupper is
13656 added to the pass_manager to reflect autogenerated changes.
13657 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
13659 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
13661 PR rtl-optimization/57708
13662 * recog.c (peep2_find_free_register): Validate all regs in a
13665 2013-08-05 Jan Hubicka <jh@suse.cz>
13668 * cgraph.c (verify_cgraph_node): Accept local flags from other
13670 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
13671 (function_and_variable_visibility): Likewise.
13672 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
13674 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13676 * graph.c (init_graph_slim_pretty_print): Remove.
13677 (print_graph_cfg): Do not call it. Use local pretty printer.
13678 (start_graph_dump): Likewise.
13680 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13682 * gimple-pretty-print.c (buffer): Remove.
13683 (initialized): Likewise.
13684 (maybe_init_pretty_print): Likewise.
13685 (print_gimple_stmt): Do not call it. Use non-static local
13686 pretty_printer variable.
13687 (print_gimple_expr): Likewise.
13688 (print_gimple_seq): Likewise.
13689 (gimple_dump_bb): Likewise.
13691 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13693 * asan.c (asan_pp): Remove.
13694 (asan_pp_initialized): Likewise.
13695 (asan_pp_initialize): Likewise.
13696 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
13697 (asan_emit_stack_protection): Tidy. Use local pretty printer.
13698 (asan_add_global): Likewise.
13700 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
13702 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
13703 * diagnostic.h (diagnostic_flush_buffer): Adjust.
13704 * pretty-print.c (pp_formatted_text_data): Likewise.
13705 (pp_indent): Rename from pp_base_indent.
13706 (pp_format): Rename from pp_base_format.
13707 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
13708 (pp_format_verbatim): Rename from pp_base_format_verbatim.
13709 (pp_flush): Rename from pp_base_flush.
13710 (pp_set_line_maximum_length): Rename from
13711 pp_base_set_line_maximum_length.
13712 (pp_clear_output_area): Rename from pp_base_clear_output_area.
13713 (pp_set_prefix): Rename from pp_base_set_prefix.
13714 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
13715 (pp_emit_prefix): Rename from pp_base_emit_prefix.
13716 (pp_append_text): Rename from pp_base_append_text.
13717 (pp_formatted_text): Rename from pp_base_formatted_text.
13718 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
13719 (pp_remaining_character_count_for_line): Rename from
13720 pp_base_remaining_character_count_for_line.
13721 (pp_newline): Rename from pp_base_newline.
13722 (pp_character): Rename from pp_base_character.
13723 (pp_string): Rename from pp_base_string.
13724 (pp_maybe_space): Rename from pp_base_maybe_space.
13725 * asan.c (asan_pp_string): Adjust.
13726 (asan_emit_stack_protection): Likewise.
13727 (asan_add_global): Likewise.
13728 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
13729 * tree-mudflap.c (mf_varname_tree): Likewise.
13730 * tree-pretty-print.c (pp_tree_identifier): Rename from
13731 pp_base_tree_identifier.
13732 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
13733 Declare as function.
13735 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13737 * pretty-print.h (pp_bar_bar): New.
13738 (pp_ampersand_ampersand): Likewise.
13739 (pp_less_equal): Likewise.
13740 (pp_greater_equal): Likewise.
13741 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
13742 printer functions instead of pp_string or operators and punctuators.
13743 (dump_gimple_call): Likewise.
13744 (dump_gimple_omp_for): Likewise.
13745 (dump_gimple_transaction): Likewise.
13746 (dump_gimple_phi): Likewise.
13747 (pp_gimple_stmt_1): Likewise.
13748 * sched-vis.c (print_insn): Likewise.
13749 * tree-mudflap.c (mf_varname_tree): Likewise.
13750 * tree-pretty-print.c (dump_block_node): Likewise.
13751 (dump_generic_node): Likewise.
13753 2013-08-02 Jan Hubicka <jh@suse.cz>
13755 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
13757 * lto-streamer-out.c (tree_is_indexable): Results decls and
13758 parm decls are not indexable.
13759 (DFS_write_tree_body): Do not follow args and results.
13760 (hash_tree): Likewise.
13761 (output_functions): Rearrange so struct function is needed
13762 only when real body is output; be able to also ouptut abstract
13763 functions; output DECL_ARGUMENTS and DECL_RESULT.
13764 (lto_output): When not in WPA, ale store abstract functions.
13765 (write_symbol): Do not care about RESULT_DECL.
13766 (output_symbol_p): Handle correctly sbtract decls.
13767 * lto-streamer-in.c (input_function): Rearrange so struct
13768 function can be NULL at entry; allow streaming of
13769 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
13770 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
13771 sanity check during LTO.
13772 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
13773 RESULT_DECl and DECL_ARGUMENTS.
13774 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
13777 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13779 * pretty-print.h (pp_underscore): New.
13781 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
13782 printer functions instead of pp_character.
13783 (dump_binary_rhs): Likewise.
13784 (dump_ternary_rhs): Likewise.
13785 (dump_gimple_call_args): Likewise.
13786 (pp_points_to_solution): Likewise.
13787 (dump_gimple_call): Likewise.
13788 (dump_gimple_switch): Likewise.
13789 (dump_gimple_cond): Likewise.
13790 (dump_gimple_bind): Likewise.
13791 (dump_gimple_try): Likewise.
13792 (dump_gimple_omp_for): Likewise.
13793 (dump_gimple_omp_continue): Likewise.
13794 (dump_gimple_omp_single): Likewise.
13795 (dump_gimple_omp_sections): Likewise.
13796 (dump_gimple_omp_block): Likewise.
13797 (dump_gimple_omp_critical): Likewise.
13798 (dump_gimple_transaction): Likewise.
13799 (dump_gimple_asm): Likewise.
13800 (dump_gimple_phi): Likewise.
13801 (dump_gimple_omp_parallel): Likewise.
13802 (dump_gimple_omp_task): Likewise.
13803 (dump_gimple_omp_atomic_load): Likewise.
13804 (dump_gimple_omp_atomic_store): Likewise.
13805 (dump_gimple_mem_ops): Likewise.
13806 (pp_gimple_stmt_1): Likewise.
13807 (pp_cfg_jump): Likewise.
13808 (dump_implicit_edges): Likewise.
13809 (gimple_dump_bb_for_graph): Likewise.
13810 * graph.c (draw_cfg_node): Likewise.
13811 * langhooks.c (lhd_print_error_function): Likewise.
13812 * sched-vis.c (print_exp): Likewise.
13813 (print_value): Likewise.
13814 (print_pattern): Likewise.
13815 (print_insn): Likewise.
13816 (rtl_dump_bb_for_graph): Likewise.
13817 * tree-pretty-print.c (dump_function_declaration): Likewise.
13818 (dump_array_domain): Likewise.
13819 (dump_omp_clause): Likewise.
13820 (dump_location): Likewise.
13821 (dump_generic_node): Likewise.
13822 (print_struct_decl): Likewise.
13823 * diagnostic.c (diagnostic_show_locus): Use pp_space.
13825 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
13827 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
13828 candidate table when replacing a candidate statement.
13829 (replace_rhs_if_not_dup): Likewise.
13830 (replace_one_candidate): Likewise.
13832 2013-08-02 Jan Hubicka <jh@suse.cz>
13833 Martin Liska <marxin.liska@gmail.com>
13835 * cgraphunit.c (add_new_function): Fix logic when adding from
13837 (assemble_thunk): Rename to ...
13838 (expand_thunk); .. this one; export; get it working with
13839 general functions; make produced gimple valid.
13840 * cgraph.h (expand_thunk): Declare.
13842 2013-08-02 Jan Hubicka <jh@suse.cz>
13844 * ipa-cp.c (gather_context_independent_values): Use
13845 ipa_get_param_move_cost.
13846 (get_replacement_map): Remove PARAM; move parameter folding
13848 (create_specialized_node): Update.
13849 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
13850 assert that we have gimple body; update move_cost.
13851 (count_formal_params): Assert that we have gimple body.
13852 (ipa_dump_param): New function.
13853 (ipa_alloc_node_params): Break out from ...
13854 (ipa_initialize_node_params): ... here.
13855 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
13856 (ipa_write_node_info): Stream move costs.
13857 (ipa_read_node_info): Read move costs.
13858 (ipa_update_after_lto_read): Do not recompute node params.
13859 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
13860 (ipa_get_param): Check we are not in WPA.
13861 (ipa_get_param_move_cost): New.
13862 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
13863 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
13864 parm numbers to be present.
13866 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
13868 PR rtl-optimization/58048
13869 * lra-constraints.c (process_alt_operands): Don't check asm
13870 operand on register.
13872 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
13874 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
13875 the implied StoreLoad barrier for atomic operations if before.
13877 2013-08-02 Jan Hubicka <jh@suse.cz>
13878 Martin Liska <marxin.liska@gmail.com>
13880 * cgraph.c (cgraph_function_body_availability): Do not check
13882 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
13883 symtab_node_availability): Declare.
13884 * ipa.c (can_replace_by_local_alias): New.
13885 (function_and_variable_visibility): Use it.
13886 * symtab.c (symtab_for_node_and_aliases,
13887 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
13889 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
13891 PR rtl-optimization/57963
13892 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
13893 (lra_constraints): Use them.
13895 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
13897 * config/arm/types.md (define_attr "type"): Add "load_acq"
13899 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
13901 (cortex_a53_store1): Likewise.
13903 2013-08-01 Jan Hubicka <jh@suse.cz>
13905 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
13906 partitions are not needed.
13908 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
13910 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
13911 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
13912 MAYBE_NON_Q_CLASS_P where appropriate.
13914 2013-08-01 Jan Hubicka <jh@suse.cz>
13916 * cgraph.h (release_function_body): Declare.
13917 * tree.c (free_lang_data_in_decl): Free, parameters and return values
13918 of unused delcarations.
13920 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13922 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
13923 RTL form when subtracting a constant.
13925 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13927 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
13928 Generate canonical plus rtx with negated immediate instead of minus
13930 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
13932 2013-08-01 Jan Hubicka <jh@suse.cz>
13934 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
13935 (cgraph_release_function_body): Likewise.
13936 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
13937 * cgraph.h (cgrpah_node): Rename abstract_and_needed
13938 to used_as_abstract_origin.
13939 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
13940 symbols used as abstract origins.
13941 * cgraphunit.c (analyze_functions): Update.
13942 * ipa.c (symtab_remove_unreachable_nodes): Recompute
13943 used_as_abstract_origin.
13944 * tree-inline.c (tree_function_versioning): Update
13945 used_as_abstract_origin; be ready for DECL_RESULT and
13946 DECL_ARGUMENTS to be NULL.
13948 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
13949 for abstract functions.
13950 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
13953 2013-08-01 Jan Hubicka <jh@suse.cz>
13955 * profile.c (compute_value_histograms): Fix thinko.
13957 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
13959 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
13960 aarch-common-protos.h to extra_headers.
13961 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
13962 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
13963 * config/aarch64/t-aarch64 (aarch-common.o): Define.
13965 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
13967 * config/aarch64/aarch64.md (define_attr "type"): Delete.
13968 Include "../arm/types.md". Define "type" attribute for all patterns.
13969 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
13972 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
13974 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
13975 to support power8 load fusion.
13976 (fusion_gpr_mem_load): Likewise.
13978 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
13980 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
13981 declarations for power8 load fusion.
13982 (emit_fusion_gpr_load): Likewise.
13984 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
13985 tuning for power8, turn on fusion mode by default. Turn on sign
13986 extending fusion mode if normal fusion mode is on, and we are at
13988 (fusion_gpr_load_p): New function, return true if we can fuse an
13989 addis instruction with a dependent load to a GPR.
13990 (emit_fusion_gpr_load): Emit the instructions for power8 load
13993 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
13994 (VSX load fusion peepholes): New peepholes to fuse together an
13995 addi instruction with a VSX load instruction.
13997 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
13998 peepholes to fuse an addis instruction with a load to a GPR base
13999 register. If we are supporting sign extending fusions, convert
14000 sign extending loads to zero extending loads and add an explicit
14003 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14005 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
14006 aarch-common-protos.h to extra_headers.
14007 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14008 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
14009 (arm_early_store_addr_dep): Likewise.
14010 (arm_no_early_alu_shift_dep): Likewise.
14011 (arm_no_early_alu_shift_value_dep): Likewise.
14012 (arm_no_early_mul_dep): Likewise.
14013 (arm_no_early_store_addr_dep): Likewise.
14014 (arm_mac_accumulator_is_mul_result): Likewise.
14015 (arm_mac_accumulator_is_result): Likewise.
14016 * config/arm/aarch-common.c: ... here. New file.
14017 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
14019 (arm_early_store_addr_dep): Likewise.
14020 (arm_no_early_alu_shift_dep): Likewise.
14021 (arm_no_early_alu_shift_value_dep): Likewise.
14022 (arm_no_early_mul_dep): Likewise.
14023 (arm_no_early_store_addr_dep): Likewise.
14024 (arm_mac_accumulator_is_mul_result): Likewise.
14025 (arm_mac_accumulator_is_result): Likewise.
14026 * config/arm/aarch-common-protos.h: ... here. New file.
14027 * config/arm/t-arm (aarch-common.o): Define.
14029 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14031 * config/arm/arm.md: Include new file "types.md".
14032 (define_attr "type"): Move from here to ...
14033 (define_attr "mul32"): Likewise.
14034 (define_attr "mul64"): Likewise.
14035 * config/arm/types.md: ... here. New file.
14037 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
14039 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
14040 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
14042 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
14044 * gen-pass-instances.awk: Fix offset of substr().
14046 2013-07-31 David Malcolm <dmalcolm@redhat.com>
14048 * Makefile.in (pass-instances.def): New.
14049 (passes.o): Replace dependency on passes.def with one on
14052 * gen-pass-instances.awk: New.
14054 * passes.c (pass_manager::pass_manager): Use pass-instances.def
14055 rather than passes.def, updating local definition of NEXT_PASS
14056 macro to add an extra NUM parameter (currently unused).
14058 2013-07-30 David Malcolm <dmalcolm@redhat.com>
14060 * Makefile.in (PASS_MANAGER_H): New.
14061 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
14062 (passes.o): Likewise.
14063 (statistics.o): Likewise.
14064 (cgraphunit.o): Likewise.
14065 (context.o): Depend on PASS_MANAGER_H.
14067 * pass_manager.h: New.
14069 * cgraphunit.c (cgraph_add_new_function): Update for moves
14070 of globals to fields of pass_manager.
14071 (analyze_function): Likewise.
14072 (expand_function): Likewise.
14073 (ipa_passes): Likewise.
14074 (compile): Likewise.
14076 * context.c (context::context): New.
14077 * context.h (context::context): New.
14078 (context::get_passes): New.
14079 (context::passes_): New.
14081 * lto-cgraph.c (input_node): Update for moves of globals to
14082 fields of pass_manager.
14084 * passes.c (all_passes): Remove, in favor of a field of the
14085 same name within the new class pass_manager.
14086 (all_small_ipa_passes): Likewise.
14087 (all_lowering_passes): Likewise.
14088 (all_regular_ipa_passes): Likewise.
14089 (all_late_ipa_passes): Likewise.
14090 (all_lto_gen_passes): Likewise.
14091 (passes_by_id): Likewise.
14092 (passes_by_id_size): Likewise.
14093 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14094 the new class pass_manager.
14095 (set_pass_for_id): Convert to...
14096 (pass_manager::set_pass_for_id): ...method.
14097 (get_pass_for_id): Convert to...
14098 (pass_manager::get_pass_for_id): ...method.
14099 (register_one_dump_file): Move body of implementation into...
14100 (pass_manager::register_one_dump_file): ...here.
14101 (register_dump_files_1): Convert to...
14102 (pass_manager::register_dump_files_1): ...method.
14103 (register_dump_files): Convert to...
14104 (pass_manager::register_dump_files): ...method.
14105 (create_pass_tab): Update for moves of globals to fields of
14107 (dump_passes): Move body of implementation into...
14108 (pass_manager::dump_passes): ...here.
14109 (register_pass): Move body of implementation into...
14110 (pass_manager::register_pass): ...here.
14111 (init_optimization_passes): Convert into...
14112 (pass_manager::pass_manager): ...constructor for new
14113 pass_manager class, and initialize the pass_lists array.
14114 (check_profile_consistency): Update for moves of globals to
14115 fields of pass_manager.
14116 (dump_profile_report): Move body of implementation into...
14117 (pass_manager::dump_profile_report): ...here.
14118 (ipa_write_summaries_1): Update for moves of pass lists from
14119 being globals to fields of pass_manager.
14120 (ipa_write_optimization_summaries): Likewise.
14121 (ipa_read_summaries): Likewise.
14122 (ipa_read_optimization_summaries): Likewise.
14123 (execute_all_ipa_stmt_fixups): Likewise.
14125 * statistics.c (statistics_fini): Update for moves of globals to
14126 fields of pass_manager.
14128 * toplev.c (general_init): Replace call to
14129 init_optimization_passes with construction of the pass_manager
14132 * tree-pass.h (all_passes): Remove, in favor of a field of the
14133 same name within the new class pass_manager.
14134 (all_small_ipa_passes): Likewise.
14135 (all_lowering_passes): Likewise.
14136 (all_regular_ipa_passes): Likewise.
14137 (all_lto_gen_passes): Likewise.
14138 (all_late_ipa_passes): Likewise.
14139 (passes_by_id): Likewise.
14140 (passes_by_id_size): Likewise.
14141 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14142 the new class pass_manager.
14143 (get_pass_for_id): Remove.
14145 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14147 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
14150 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14152 * arm.md (mulhi3): New expand pattern.
14154 2013-07-30 Jan Hubicka <jh@suse.cz>
14155 Martin Liska <marxin.liska@gmail.com>
14157 * profile.c (compute_value_histograms): Do not ICE when
14158 there is mismatch only on some counters.
14160 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14162 PR rtl-optimization/57637
14163 * function.c (move_insn_for_shrink_wrap): Also check the
14164 GEN set of the LIVE problem for the liveness analysis
14165 if it exists, otherwise give up.
14167 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14169 PR tree-optimization/57993
14170 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
14171 replaced statement in the candidate table.
14172 (phi_add_costs): Return infinite cost when the hidden basis does
14173 not dominate all phis on which the candidate is dependent.
14174 (replace_one_candidate): Record replaced statement in the
14177 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
14179 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
14180 (ashlv2si3): New expander.
14181 (*ashlv2si3_i): New define_insn_and_split.
14182 * predicates.md (float_operation): Allow patterns with three
14183 basic sub-patterns.
14185 PR rtl-optimization/58021
14186 * mode-switching.c (create_pre_exit): Always split off preceding
14187 insns if we are not at the basic block head.
14189 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
14191 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
14192 (UCLIBC_DYNAMIC_LINKER): New macro.
14193 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
14195 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
14196 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
14198 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
14199 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
14200 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
14201 for SF and DF modes. Use ieee_quad_format for TF mode.
14202 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
14203 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
14204 (mips_option_override): Handle `-mnan=legacy'.
14205 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
14206 `-mabs=2008' and `-mnan=2008'.
14207 (OPTION_DEFAULT_SPECS): Add "nan" default.
14208 (ASM_SPEC): Handle `-mnan='.
14209 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
14210 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
14211 comment accordingly.
14212 (neg<mode>2): Likewise.
14213 * config/mips/mips.opt (mabs, mnan): New options.
14214 * doc/install.texi (Configuration): Document `--with-nan=' option.
14215 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
14217 (MIPS Options): Document them.
14218 * config.gcc <mips*-*-*>: Handle `--with-nan='.
14219 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
14220 * configure: Regenerate.
14221 * config.in: Regenerate.
14223 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14225 * config/i386/i386.md (float post-reload splitters): Do not check
14226 for subregs of SSE registers.
14228 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14229 H.J. Lu <hongjiu.lu@intel.com>
14233 * config/i386/i386.md (post-reload splitter
14234 to avoid partial SSE reg dependency stalls): New pattern.
14236 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
14238 * config/s390/s390.md ("movcc"): Swap load and store instructions.
14240 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
14242 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14243 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
14245 2013-07-26 Cary Coutant <ccoutant@google.com>
14247 * dwarf2out.c (die_checksum_ordered): Don't include template
14248 instantiations in signature.
14249 (is_template_parameter): New function.
14250 (is_template_instantiation): New function.
14251 (generate_skeleton_bottom_up): Don't include template instantiations
14253 (generate_skeleton): Likewise.
14254 (break_out_comdat_types): Move recursive call to break out nested
14256 (prune_unused_types_mark_generic_parms_dies): Call
14257 is_template_parameter.
14259 2013-07-26 Ian Bolton <ian.bolton@arm.com>
14261 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
14262 uses vector registers.
14263 * config/aarch64/iterators.md: Add attributes rtn and vas.
14265 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14266 Richard Earnshaw <richard.earnshaw@arm.com>
14268 * combine.c (simplify_comparison): Re-canonicalize operands
14270 * config/arm/arm.md (movcond_addsi): New splitter.
14272 2013-07-25 Sterling Augustine <saugustine@google.com>
14274 * dwarf2out.c (size_of_pubnames): Move code to...
14275 (include_pubname_in_output): ...here. New.
14276 (want_pubnames): Rearrange.
14277 (output_pubnames): Call include_pubname_in_output. Move assertion.
14279 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14281 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
14283 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14286 * doc/extend.texi: Remove obsolete builtins. Fix
14287 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
14289 2013-07-25 Jan Hubicka <jh@suse.cz>
14291 * cgraph.c (release_function_body): Break out from ...
14292 (cgraph_release_function_body): ... this one; also release DECL_RESULT
14293 and DECL_ARGUMENTS.
14294 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
14295 old_tree in the map.
14296 (create_specialized_node): Update.
14297 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
14299 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
14300 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
14301 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
14302 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
14305 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14307 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
14308 addsi3_carryin_alt2_<optab>): Correct output template.
14310 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14312 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
14313 Adjust for arm_restrict_it.
14314 Remove trailing whitespace.
14316 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
14318 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
14319 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
14321 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
14323 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
14325 PR rtl-optimization/57960
14326 * lra-constraints.c (process_alt_operands): Use the right mode
14327 when checking strict_low.
14329 2013-07-25 Jan Hubicka <jh@suse.cz>
14331 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
14332 * cgraph.c (cgraph_remove_node): Do not release function body
14333 when in cgraph streaming.
14334 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
14335 in other partitions are not considered reachable; fix handling of
14338 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14340 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
14342 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14347 * config/arm/arm.md ("*sibcall_insn): Replace use of
14348 Ss with US. Adjust output for v5 and v4t.
14349 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
14351 * config/arm/constraints.md ("Ss"): Rename to US.
14353 2013-07-25 Terry Guo <terry.guo@arm.com>
14355 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
14356 shift_add/shift_sub0/shift_sub1 RTXs.
14358 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
14359 Anton Blanchard <anton@au1.ibm.com>
14361 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
14362 (altivec_vpks<VI_char>ss): Likewise.
14363 (altivec_vpks<VI_char>us): Likewise.
14364 (altivec_vpku<VI_char>us): Likewise.
14365 (altivec_vpku<VI_char>um): Likewise.
14367 2013-07-24 David Malcolm <dmalcolm@redhat.com>
14369 Introduce context class.
14371 * Makefile.in (CONTEXT_H): New.
14372 (OBJS): Add context.o.
14373 (toplev.o): Add CONTEXT_H to dependencies.
14376 * toplev.c (general_init): Create the singleton gcc::context instance.
14382 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
14384 PR rtl-optimization/57968
14385 * mode-switching.c (create_pre_exit): Allow instructions that
14386 don't set a return register to need a non-exit mode.
14388 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14389 Anton Blanchard <anton@au1.ibm.com>
14391 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
14392 operands to vperm for little endian.
14393 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
14394 of lvsl to create the control mask for a vperm for little endian.
14396 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14397 Anton Blanchard <anton@au1.ibm.com>
14399 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
14400 two operands for little-endian.
14402 2013-07-23 Steve Ellcey <sellcey@mips.com>
14404 * config/mips/mips.c (mips_case_values_threshold): New.
14405 (TARGET_CASE_VALUES_THRESHOLD): Define.
14407 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14408 Anton Blanchard <anton@au1.ibm.com>
14410 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
14411 selection of field for vector splat in little endian mode.
14413 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
14415 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
14416 expanders to rs6000.md.
14417 (ior<mode>3): Likewise.
14418 (and<mode>3): Likewise.
14419 (one_cmpl<mode>2): Likewise.
14420 (nor<mode>3): Likewise.
14421 (andc<mode>3): Likewise.
14422 (eqv<mode>3): Likewise.
14423 (nand<mode>3): Likewise.
14424 (orc<mode>3): Likewise.
14426 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
14429 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
14430 to split multi-word logical operations.
14431 (rs6000_split_logical_di): Likewise.
14432 (rs6000_split_logical): Likewise.
14434 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
14435 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
14436 and allow TImode operations in 32-bit.
14437 (vsx_and<mode>3_64bit): Likewise.
14438 (vsx_ior<mode>3_32bit): Likewise.
14439 (vsx_ior<mode>3_64bit): Likewise.
14440 (vsx_xor<mode>3_32bit): Likewise.
14441 (vsx_xor<mode>3_64bit): Likewise.
14442 (vsx_one_cmpl<mode>2_32bit): Likewise.
14443 (vsx_one_cmpl<mode>2_64bit): Likewise.
14444 (vsx_nor<mode>3_32bit): Likewise.
14445 (vsx_nor<mode>3_64bit): Likewise.
14446 (vsx_andc<mode>3_32bit): Likewise.
14447 (vsx_andc<mode>3_64bit): Likewise.
14448 (vsx_eqv<mode>3_32bit): Likewise.
14449 (vsx_eqv<mode>3_64bit): Likewise.
14450 (vsx_nand<mode>3_32bit): Likewise.
14451 (vsx_nand<mode>3_64bit): Likewise.
14452 (vsx_orc<mode>3_32bit): Likewise.
14453 (vsx_orc<mode>3_64bit): Likewise.
14455 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
14456 logical types in GPRs.
14458 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
14459 logical insns to rs6000.md, and allow TImode operations in
14461 (altivec_ior<mode>3): Likewise.
14462 (altivec_xor<mode>3): Likewise.
14463 (altivec_one_cmpl<mode>2): Likewise.
14464 (altivec_nor<mode>3): Likewise.
14465 (altivec_andc<mode>3): Likewise.
14467 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
14468 attributes for moving the 128-bit logical operations into
14470 (BOOL_REGS_OUTPUT): Likewise.
14471 (BOOL_REGS_OP1): Likewise.
14472 (BOOL_REGS_OP2): Likewise.
14473 (BOOL_REGS_UNARY): Likewise.
14474 (BOOL_REGS_AND_CR0): Likewise.
14475 (one_cmpl<mode>2): Add support for DI logical operations on
14476 32-bit, splitting the operations to 32-bit.
14477 (anddi3): Likewise.
14478 (iordi3): Likewise.
14479 (xordi3): Likewise.
14480 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
14481 changes to combine the 32/64-bit code, allow logical operations on
14482 TI mode in 32-bit, and to use similar match_operator patterns like
14483 scalar mode uses. Combine the Altivec and VSX code for logical
14484 operations, and move it here.
14485 (ior<mode>3, 128-bit types): Likewise.
14486 (xor<mode>3, 128-bit types): Likewise.
14487 (one_cmpl<mode>3, 128-bit types): Likewise.
14488 (nor<mode>3, 128-bit types): Likewise.
14489 (andc<mode>3, 128-bit types): Likewise.
14490 (eqv<mode>3, 128-bit types): Likewise.
14491 (nand<mode>3, 128-bit types): Likewise.
14492 (orc<mode>3, 128-bit types): Likewise.
14493 (and<mode>3_internal): Likewise.
14494 (bool<mode>3_internal): Likewise.
14495 (boolc<mode>3_internal1): Likewise.
14496 (boolc<mode>3_internal2): Likewise.
14497 (boolcc<mode>3_internal1): Likewise.
14498 (boolcc<mode>3_internal2): Likewise.
14499 (eqv<mode>3_internal1): Likewise.
14500 (eqv<mode>3_internal2): Likewise.
14501 (one_cmpl1<mode>3_internal): Likewise.
14503 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14505 * config/microblaze/microblaze.c (microblaze_expand_prologue):
14506 Rename flag_stack_usage to flag_stack_usage_info.
14508 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14510 * config/microblaze/sync.md: New file.
14511 * config/microblaze/microblaze.md: Include sync.md
14512 * config/microblaze/microblaze.c: Add print_operand 'y'.
14513 * config/microblaze/constraints.md: Add memory_contraint
14514 'Q' which is a single register.
14516 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
14518 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
14520 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14522 * reload.c (find_reloads): Exit loop once we find this operand
14523 cannot be reloaded somehow for this alternative.
14525 * reload.c (find_reloads): Exit loop once we find a hard register.
14527 * rtlanal.c (computed_jump_p): Exit loop once we find label
14530 * i386.c (ix86_pad_returns): Exit loop after setting replace.
14532 * cfgloopmanip.c (remove_path): Exit loop after setting
14535 * gensupport.c (subst_dup): Avoid loop if code is not
14536 MATCH_DUP nor MATCH_OP_DUP.
14538 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
14540 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
14542 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14544 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
14545 true for SP_REGNUM if mode == ptr_mode.
14546 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
14547 with value R0_REGNUM + 31.
14549 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14551 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
14552 pad pointer-typed argument downward.
14554 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14556 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
14557 and __ILP32__ when the ILP32 model is in use.
14559 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14561 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
14562 (aarch64_load_symref_appropriately): In the case of
14563 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
14564 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
14565 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
14566 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
14567 if 'mode' doesn't equal to 'ptr_mode'.
14568 (aarch64_output_mi_thunk): Add an assertion on the alignment of
14569 'vcall_offset'; change to call aarch64_emit_move differently depending
14570 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
14571 to calculate the upper bound of 'vcall_offset'.
14572 (aarch64_cannot_force_const_mem): Change to also return true if
14574 (aarch64_legitimize_reload_address): In the case of large
14575 displacements, add new local variable 'xmode' and an assertion
14576 based on it; change to use 'xmode' to generate the new rtx and
14578 (aarch64_asm_trampoline_template): Change to generate the template
14579 differently depending on TARGET_ILP32 or not; change to use
14580 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
14581 (aarch64_trampoline_size): Removed.
14582 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
14583 and replace immediate literals with it. Change to use 'ptr_mode'
14584 instead of 'DImode' and call convert_memory_address if the mode
14585 of 'fnaddr' doesn't equal to 'ptr_mode'.
14586 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
14588 (aarch64_elf_asm_destructor): Likewise.
14589 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
14590 on TARGET_ILP32 instead of aarch64_trampoline_size.
14591 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
14592 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
14593 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
14594 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
14595 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
14596 (storewb_pair<GPI:mode>_<P:mode>): ... this.
14597 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
14598 depending on the value of 'mode'.
14599 (add_losym_<mode>): New.
14600 (ldr_got_small_<mode>): New, based on ldr_got_small.
14601 (ldr_got_small): Remove.
14602 (ldr_got_small_sidi): New.
14603 * config/aarch64/iterators.md (P): New.
14604 (PTR): Change to 'ptr_mode' in the condition.
14606 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14608 * config.gcc (aarch64*-*-*): Support --with-abi.
14609 (aarch64*-*-elf): Support --with-multilib-list.
14610 (aarch64*-*-linux*): Likewise.
14611 (supported_defaults): Add abi to aarch64*-*-*.
14612 * configure.ac: Mention AArch64 for --with-multilib-list.
14613 * configure: Re-generated.
14614 * config/aarch64/biarchilp32.h: New file.
14615 * config/aarch64/biarchlp64.h: New file.
14616 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
14618 (MULTILIB_DEFAULTS): Ditto.
14619 (DRIVER_SELF_SPECS): Ditto.
14620 (ASM_SPEC): Update to also substitute -mabi.
14621 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
14622 file whose name depends on -mabi= and -mbig-endian.
14623 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
14625 (POINTER_SIZE): New define.
14626 (POINTERS_EXTEND_UNSIGNED): Ditto.
14627 (enum aarch64_abi_type): New enumeration tag.
14628 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
14629 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
14630 (TARGET_ILP32): New define.
14631 * config/aarch64/aarch64.opt (mabi): New.
14632 (aarch64_abi): New.
14633 (ilp32, lp64): New values for -mabi.
14634 * config/aarch64/t-aarch64 (comma): New define.
14635 (MULTILIB_OPTIONS): Ditto.
14636 (MULTILIB_DIRNAMES): Ditto.
14637 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
14638 * doc/invoke.texi: Document -mabi for AArch64.
14640 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
14642 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
14644 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14645 Anton Blanchard <anton@au1.ibm.com>
14647 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
14648 endianness when selecting field to splat.
14650 2013-07-22 Eric Christopher <echristo@gmail.com>
14652 * dwarf2out.c (die_odr_checksum): New function to use
14653 CHECKSUM_ macros and ULEB128 for DIE tag.
14654 (generate_type_signature): Use.
14656 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
14658 * config.gcc (sparc*-*-*): Accept leon3 processor.
14659 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
14660 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
14661 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
14662 * config/sparc/sparc.opt (enum processor_type): Add leon3.
14663 (mfix-ut699): Adjust comment.
14664 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
14665 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
14666 (CPP_CPU_SPEC): Likewise.
14667 (ASM_CPU_SPEC): Likewise.
14668 * config/sparc/sparc.c (leon3_cost): New constant.
14669 (sparc_option_override): Add leon3 support.
14670 (mem_ref): New function.
14671 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
14672 (sparc_do_work_around_errata): Look into the instruction in the delay
14673 slot and adjust accordingly. Add fix for the data cache nullify issues
14674 of the UT699. Change insertion position for the NOP.
14675 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
14676 (leon3_load): New reservation.
14677 (leon_store): Bump latency to 2.
14678 (grfpu): New automaton.
14679 (grfpu_alu): New unit.
14680 (grfpu_ds): Likewise.
14681 (leon_fp_alu): Adjust.
14682 (leon_fp_mult): Delete.
14683 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
14684 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
14685 * config/sparc/sparc.md (cpu): Add leon3.
14686 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
14687 (swapsi): Likewise.
14688 (atomic_test_and_set): Likewise.
14689 (ldstub): Likewise.
14691 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
14693 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
14694 default for R5900 targets.
14695 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
14696 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
14697 * config/mips/mips.c (mips_option_override): Report an error for
14698 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
14699 for -march=r5900 -mhard-float.
14701 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14703 * df-problems.c (can_move_insns_across): Exit loop once we
14704 find a non-fixed, non-global register.
14706 * ipa-pure-const.c (propagate_nothrow): Exit loop after
14709 * omega.c (omega_eliminate_red): Break after setting red_found.
14710 (omega_problem_has_red_equations): Similarly after setting found.
14711 (omega_query_variable): Similarly after setting coupled.
14713 2013-07-22 Marek Polacek <polacek@redhat.com>
14715 * gimplify.c: Don't include gimple.h twice.
14717 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14719 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
14720 instead of TARGET_THUMB1.
14721 (Pz): New constraint.
14722 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
14724 (compare_negsi_si): Likewise.
14725 (compare_addsi2_op0): Likewise.
14726 (compare_addsi2_op1): Likewise.
14727 (addsi3_carryin_<optab>): Likewise.
14728 (addsi3_carryin_alt2_<optab>): Likewise.
14729 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
14730 for arm_restrict_it.
14731 (subsi3_carryin): Likewise.
14732 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
14733 (minmax_arithsi): Disable for arm_restrict_it.
14734 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
14735 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
14736 (satsi_<SAT:code>_shift): Likewise.
14737 (arm_shiftsi3): Add alternative for 16-bit encoding.
14738 (arm32_movhf): Disable for arm_restrict_it.
14739 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
14740 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
14742 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
14744 * config/arm/arm.md (attribute "insn"): Delete.
14745 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
14746 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
14747 (not_shiftsi): Update for attribute change.
14748 (not_shiftsi_compare0): Likewise.
14749 (not_shiftsi_compare0_scratch): Likewise.
14750 (arm_one_cmplsi2): Likewise.
14751 (thumb1_one_cmplsi2): Likewise.
14752 (notsi_compare0): Likewise.
14753 (notsi_compare0_scratch): Likewise.
14754 (thumb1_movdi_insn): Likewise.
14755 (arm_movsi_insn): Likewise.
14756 (movhi_insn_arch4): Likewise.
14757 (movhi_bytes): Likewise.
14758 (arm_movqi_insn): Likewise.
14759 (thumb1_movqi_insn): Likewise.
14760 (arm32_movhf): Likewise.
14761 (thumb1_movhf): Likewise.
14762 (arm_movsf_soft_insn): Likewise.
14763 (thumb1_movsf_insn): Likewise.
14764 (thumb_movdf_insn): Likewise.
14765 (movsicc_insn): Likewise.
14766 (movsfcc_soft_insn): Likewise.
14767 (and_scc): Likewise.
14768 (cond_move): Likewise.
14769 (if_move_not): Likewise.
14770 (if_not_move): Likewise.
14771 (if_shift_move): Likewise.
14772 (if_move_shift): Likewise.
14773 (if_shift_shift): Likewise.
14774 (if_not_arith): Likewise.
14775 (if_arith_not): Likewise.
14776 (cond_move_not): Likewise.
14777 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
14778 (neon_mov<mode>): Likewise.
14779 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
14780 (thumb2_movsi_vfp): Likewise.
14781 (movsf_vfp): Likewise.
14782 (thumb2_movsf_vfp): Likewise.
14783 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
14785 (cortexa7_older_only): Likewise.
14786 (cortexa7_younger): Likewise.
14787 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
14788 (1020alu_shift_op): Likewise.
14789 (1020alu_shift_reg_op): Likewise.
14790 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
14791 (alu_shift_op): Likewise.
14792 (alu_shift_reg_op): Likewise.
14793 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
14794 (11_alu_shift_op): Likewise.
14795 (11_alu_shift_reg_op): Likewise.
14796 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
14797 (9_alu_shift_reg_op): Likewise.
14798 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
14800 (cortex_a15_alu_shift): Likewise.
14801 (cortex_a15_alu_shift_reg): Likewise.
14802 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
14803 (cortex_a5_alu_shift): Likewise.
14804 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
14806 (cortex_a53_alu_shift): Likewise.
14807 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
14809 (cortex_a7_alu_reg): Likewise.
14810 (cortex_a7_alu_shift): Likewise.
14811 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
14812 (cortex_a8_alu_shift): Likewise.
14813 (cortex_a8_alu_shift_reg): Likewise.
14814 (cortex_a8_mov): Likewise.
14815 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
14816 (cortex_a9_dp_shift): Likewise.
14817 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
14818 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
14819 (cortex_r4_mov): Likewise.
14820 (cortex_r4_alu_shift): Likewise.
14821 (cortex_r4_alu_shift_reg): Likewise.
14822 * config/arm/fa526.md (526_alu_op): Update for attribute change.
14823 (526_alu_shift_op): Likewise.
14824 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
14825 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
14826 (626te_alu_shift_op): Likewise.
14827 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
14828 (726te_alu_op): Likewise.
14829 (726te_alu_shift_op): Likewise.
14830 (726te_alu_shift_reg_op): Likewise.
14831 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
14832 (mp626_alu_shift_op): Likewise.
14833 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
14834 (pj4_alu_e1_conds): Likewise.
14835 (pj4_alu): Likewise.
14836 (pj4_alu_conds): Likewise.
14837 (pj4_shift): Likewise.
14838 (pj4_shift_conds): Likewise.
14839 (pj4_alu_shift): Likewise.
14840 (pj4_alu_shift_conds): Likewise.
14842 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14844 * config/arm/predicates.md (shiftable_operator_strict_it):
14846 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
14847 Disable cond_exec version for arm_restrict_it.
14848 (thumb2_smaxsi3): Convert to generate cond_exec.
14849 (thumb2_sminsi3): Likewise.
14850 (thumb32_umaxsi3): Likewise.
14851 (thumb2_uminsi3): Likewise.
14852 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
14853 (thumb2_neg_abssi2): Likewise.
14854 (thumb2_mov_scc): Add alternative for 16-bit encoding.
14855 (thumb2_movsicc_insn): Adjust alternatives.
14856 (thumb2_mov_negscc): Disable for arm_restrict_it.
14857 (thumb2_mov_negscc_strict_it): New pattern.
14858 (thumb2_mov_notscc_strict_it): New pattern.
14859 (thumb2_mov_notscc): Disable for arm_restrict_it.
14860 (thumb2_ior_scc): Likewise.
14861 (thumb2_ior_scc_strict_it): New pattern.
14862 (thumb2_cond_move): Adjust for arm_restrict_it.
14863 (thumb2_cond_arith): Disable for arm_restrict_it.
14864 (thumb2_cond_arith_strict_it): New pattern.
14865 (thumb2_cond_sub): Adjust for arm_restrict_it.
14866 (thumb2_movcond): Likewise.
14867 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
14868 (thumb2_zero_extendhisi2_v6): Likewise.
14869 (thumb2_zero_extendqisi2_v6): Likewise.
14870 (orsi_notsi_si): Likewise.
14871 (orsi_not_shiftsi_si): Likewise.
14873 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
14875 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
14876 instruction sequence is 1 byte shorter.
14878 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
14880 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
14881 it is not needed after split.
14883 2013-07-20 Iain Sandoe <iain@codesourcery.com>
14886 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
14887 second label for nonlocal goto receivers. Don't output pic base labels
14888 unless we're producing PIC; mark that action unreachable().
14889 (ix86_save_reg): If the function contains a nonlocal label, save the
14891 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
14892 * config/darwin.c (emitted_pic_label_num): New GTY.
14893 (update_pic_label_number_if_needed): New.
14894 (machopic_output_function_base_name): Adjust for nonlocal receiver
14896 (machopic_should_output_picbase_label): New.
14897 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
14898 (nonlocal_goto_receiver): New insn and split.
14900 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
14902 * config/aarch64/aarch64-builtins.c
14903 (aarch64_fold_builtin): Fold abs in all modes.
14904 * config/aarch64/aarch64-simd-builtins.def
14905 (abs): Enable for all modes.
14906 * config/aarch64/arm_neon.h
14907 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
14908 (vabs_f64): Add missing intrinsic.
14910 2013-07-19 Ian Bolton <ian.bolton@arm.com>
14912 * config/aarch64/arm_neon.h (vabs_s64): New function
14914 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
14917 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
14918 * config/avr/avr.md (adjust_len): Add `round'.
14919 * config/avr/avr-protos.h (avr_out_round): New prototype.
14920 (avr_out_plus): Add `out_label' argument.
14921 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
14922 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
14923 Handle the case where `insn' is just a pattern.
14924 (avr_out_bitop): Handle the case where `insn' is just a pattern.
14925 (avr_out_round): New function.
14926 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
14928 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
14930 * config/microblaze/microblaze.c (microblaze_expand_prologue):
14931 Add check for flag_stack_usage to handle -fstack-usage support
14933 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
14935 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
14936 interaction for new Power8 flags and VSX.
14938 2013-07-18 Sriraman Tallam <tmsriram@google.com>
14940 PR middle-end/57698
14941 * tree-inline.c (expand_call_inline): Emit errors during
14942 early_inlining only if optimization is not turned on.
14944 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14948 * passes.c (init_optimization_passes): Move the construction of
14949 the pass hierarchy into a new passes.def file.
14951 * Makefile.in (passes.o): Add dependency on passes.def.
14953 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14955 * passes.c (init_optimization_passes): Introduce macros for
14956 constructing the tree of passes (INSERT_PASSES_AFTER,
14957 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
14958 TERMINATE_PASS_LIST).
14960 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
14961 Wei Mi <wmi@google.com>
14963 PR rtl-optimization/57878
14964 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
14966 (reload_pseudo_compare_func): Check nregs first for reload
14969 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14971 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
14973 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
14975 * read-rtl.c (validate_const_int): Once an invalid character is
14976 seen, quit the loop.
14978 * gengtype.c (write_roots): Similarly once we find the "deletable"
14979 or "if_marked" option.
14981 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
14983 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
14984 "xtab" and "sat". Move value "clz" from here to ...
14985 (attriubte "type"): ... here.
14986 (satsi_<SAT:code>): Delete "insn" attribute.
14987 (satsi_<SAT:code>_shift): Likewise.
14988 (arm_zero_extendqisi2addsi): Likewise.
14989 (arm_extendqisi2addsi): Likewise.
14990 (clzsi2): Update for attribute changes.
14991 (rbitsi2): Likewise.
14992 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
14994 (arm_usatsihi): Likewise.
14995 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
14997 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
14999 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
15000 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
15001 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
15002 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
15003 in alphabetical order.
15004 (attribute "core_cycles"): Update for attribute changes.
15005 (arm_addsi3): Likewise.
15006 (addsi3_compare0): Likewise.
15007 (addsi3_compare0_scratch): Likewise.
15008 (addsi3_compare_op1): Likewise.
15009 (addsi3_compare_op2): Likewise.
15010 (compare_addsi2_op0): Likewise.
15011 (compare_addsi2_op1): Likewise.
15012 (addsi3_carryin_shift_<optab>): Likewise.
15013 (subsi3_carryin_shift): Likewise.
15014 (rsbsi3_carryin_shift): Likewise.
15015 (arm_subsi3_insn): Likewise.
15016 (subsi3_compare0): Likewise.
15017 (subsi3_compare): Likewise.
15018 (arm_andsi3_insn): Likewise.
15019 (thumb1_andsi3_insn): Likewise.
15020 (andsi3_compare0): Likewise.
15021 (andsi3_compare0_scratch): Likewise.
15022 (zeroextractsi_compare0_scratch
15023 (andsi_not_shiftsi_si): Likewise.
15024 (iorsi3_insn): Likewise.
15025 (iorsi3_compare0): Likewise.
15026 (iorsi3_compare0_scratch): Likewise.
15027 (arm_xorsi3): Likewise.
15028 (thumb1_xorsi3_insn): Likewise.
15029 (xorsi3_compare0): Likewise.
15030 (xorsi3_compare0_scratch): Likewise.
15031 (satsi_<SAT:code>_shift): Likewise.
15033 (arm_shiftsi3): Likewise.
15034 (shiftsi3_compare0): Likewise.
15035 (not_shiftsi): Likewise.
15036 (not_shiftsi_compare0): Likewise.
15037 (not_shiftsi_compare0_scratch): Likewise.
15038 (arm_one_cmplsi2): Likewise.
15039 (thumb_one_complsi2): Likewise.
15040 (notsi_compare0): Likewise.
15041 (notsi_compare0_scratch): Likewise.
15042 (thumb1_zero_extendhisi2): Likewise.
15043 (arm_zero_extendhisi2): Likewise.
15044 (arm_zero_extendhisi2_v6): Likewise.
15045 (arm_zero_extendhisi2addsi): Likewise.
15046 (thumb1_zero_extendqisi2): Likewise.
15047 (thumb1_zero_extendqisi2_v6): Likewise.
15048 (arm_zero_extendqisi2): Likewise.
15049 (arm_zero_extendqisi2_v6): Likewise.
15050 (arm_zero_extendqisi2addsi): Likewise.
15051 (thumb1_extendhisi2): Likewise.
15052 (arm_extendhisi2): Likewise.
15053 (arm_extendhisi2_v6): Likewise.
15054 (arm_extendqisi): Likewise.
15055 (arm_extendqisi_v6): Likewise.
15056 (arm_extendqisi2addsi): Likewise.
15057 (thumb1_extendqisi2): Likewise.
15058 (thumb1_movdi_insn): Likewise.
15059 (arm_movsi_insn): Likewise.
15060 (movsi_compare0): Likewise.
15061 (movhi_insn_arch4): Likewise.
15062 (movhi_bytes): Likewise.
15063 (arm_movqi_insn): Likewise.
15064 (thumb1_movqi_insn): Likewise.
15065 (arm32_movhf): Likewise.
15066 (thumb1_movhf): Likewise.
15067 (arm_movsf_soft_insn): Likewise.
15068 (thumb1_movsf_insn): Likewise.
15069 (movdf_soft_insn): Likewise.
15070 (thumb_movdf_insn): Likewise.
15071 (arm_cmpsi_insn): Likewise.
15072 (cmpsi_shiftsi): Likewise.
15073 (cmpsi_shiftsi_swp): Likewise.
15074 (arm_cmpsi_negshiftsi_si): Likewise.
15075 (movsicc_insn): Likewise.
15076 (movsfcc_soft_insn): Likewise.
15077 (arith_shiftsi): Likewise.
15078 (arith_shiftsi_compare0
15079 (arith_shiftsi_compare0_scratch
15080 (sub_shiftsi): Likewise.
15081 (sub_shiftsi_compare0
15082 (sub_shiftsi_compare0_scratch
15083 (and_scc): Likewise.
15084 (cond_move): Likewise.
15085 (if_plus_move): Likewise.
15086 (if_move_plus): Likewise.
15087 (if_move_not): Likewise.
15088 (if_not_move): Likewise.
15089 (if_shift_move): Likewise.
15090 (if_move_shift): Likewise.
15091 (if_shift_shift): Likewise.
15092 (if_not_arith): Likewise.
15093 (if_arith_not): Likewise.
15094 (cond_move_not): Likewise.
15095 (thumb1_ashlsi3): Set type attribute.
15096 (thumb1_ashrsi3): Likewise.
15097 (thumb1_lshrsi3): Likewise.
15098 (thumb1_rotrsi3): Likewise.
15099 (shiftsi3_compare0_scratch): Likewise.
15100 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
15101 (neon_mov<mode>): Likewise.
15102 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
15104 (thumb2_movsi_insn): Likewise.
15105 (thumb2_cmpsi_neg_shiftsi): Likewise.
15106 (thumb2_extendqisi_v6): Likewise.
15107 (thumb2_zero_extendhisi2_v6): Likewise.
15108 (thumb2_zero_extendqisi2_v6): Likewise.
15109 (thumb2_shiftsi3_short): Likewise.
15110 (thumb2_addsi3_compare0_scratch): Likewise.
15111 (orsi_not_shiftsi_si): Likewise.
15112 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
15113 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
15115 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
15116 (1020alu_shift_op): Likewise.
15117 (1020alu_shift_reg_op): Likewise.
15118 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
15119 (alu_shift_op): Likewise.
15120 (alu_shift_reg_op): Likewise.
15121 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
15122 (11_alu_shift_op): Likewise.
15123 (11_alu_shift_reg_op): Likewise.
15124 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
15125 (9_alu_shift_reg_op): Likewise.
15126 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15128 (cortex_a15_alu_shift): Likewise.
15129 (cortex_a15_alu_shift_reg): Likewise.
15130 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
15132 (cortex_a5_alu_shift): Likewise.
15133 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
15135 (cortex_a53_alu_shift): Likewise.
15136 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15138 (cortex_a7_alu_reg): Likewise.
15139 (cortex_a7_alu_shift): Likewise.
15140 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
15142 (cortex_a8_alu_shift): Likewise.
15143 (cortex_a8_alu_shift_reg): Likewise.
15144 (cortex_a8_mov): Likewise.
15145 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
15146 (cortex_a9_dp_shift): Likewise.
15147 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
15149 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
15151 (cortex_r4_mov): Likewise.
15152 (cortex_r4_alu_shift): Likewise.
15153 (cortex_r4_alu_shift_reg): Likewise.
15154 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
15155 (526_alu_shift_op): Likewise.
15156 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
15157 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
15158 (626te_alu_shift_op): Likewise.
15159 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
15160 (726te_alu_op): Likewise.
15161 (726te_alu_shift_op): Likewise.
15162 (726te_alu_shift_reg_op): Likewise.
15163 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
15164 (mp626_alu_shift_op): Likewise.
15165 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
15166 (pj4_alu_e1_conds): Likewise.
15167 (pj4_alu): Likewise.
15168 (pj4_alu_conds): Likewise.
15169 (pj4_shift): Likewise.
15170 (pj4_shift_conds): Likewise.
15171 (pj4_alu_shift): Likewise.
15172 (pj4_alu_shift_conds): Likewise.
15173 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15175 (cortexa7_older_only): Likewise.
15176 (cortexa7_younger): Likewise.
15178 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15180 * ipa-pure-const.c (generate_summary): Rename to...
15181 (pure_const_generate_summary): ... this.
15183 2013-07-17 Iain Sandoe <iain@codesourcery.com>
15185 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
15187 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
15190 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
15193 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15195 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
15196 enabled without -march=zEC12.
15197 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
15200 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15202 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
15203 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
15204 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
15205 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
15206 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
15207 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
15208 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
15210 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
15211 and ISA_HAS_NMADD3_NMSUB3.
15212 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
15213 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
15214 (nmsub4<mode>, nmsub3<mode>): Likewise.
15215 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
15217 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15219 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
15220 TARGET_MIPS5400 checking.
15222 2013-07-16 Jakub Jelinek <jakub@redhat.com>
15223 Peter Bergner <bergner@vnet.ibm.com>
15225 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
15226 registers in the comment.
15227 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
15228 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
15229 rather than FIRST_PSEUDO_REGISTERS.
15231 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
15233 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
15234 enable extra ISA flags with TARGET_HTM.
15236 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15238 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
15241 2013-07-15 Cong Hou <congh@google.com>
15243 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
15244 in compare function for sorting.
15246 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
15248 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
15249 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
15250 * config/rs6000/rs6000.opt: Add -mhtm option.
15251 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
15252 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
15253 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
15254 __HTM__ if the HTM instructions are available.
15255 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
15256 htm_spr_reg_operand): New define_predicates.
15257 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
15258 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
15260 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
15261 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
15262 HTM builtin functions.
15263 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
15264 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
15265 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
15266 (rs6000_builtin_mask_calculate): Likewise.
15267 (rs6000_option_override_internal): Likewise.
15268 (bdesc_htm): Add new HTM builtin support.
15269 (htm_spr_num): New function.
15270 (htm_spr_regno): Likewise.
15271 (rs6000_htm_spr_icode): Likewise.
15272 (htm_expand_builtin): Likewise.
15273 (htm_init_builtins): Likewise.
15274 (rs6000_expand_builtin): Add support for HTM builtin functions.
15275 (rs6000_init_builtins): Likewise.
15276 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
15278 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
15279 (TARGET_HTM, MASK_HTM): Define macros.
15280 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
15281 (FIXED_REGISTERS): Likewise.
15282 (CALL_USED_REGISTERS): Likewise.
15283 (CALL_REALLY_USED_REGISTERS): Likewise.
15284 (REG_ALLOC_ORDER): Likewise.
15285 (enum reg_class): Likewise.
15286 (REG_CLASS_NAMES): Likewise.
15287 (REG_CLASS_CONTENTS): Likewise.
15288 (REGISTER_NAMES): Likewise.
15289 (ADDITIONAL_REGISTER_NAMES): Likewise.
15290 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
15291 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
15292 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
15293 * config/rs6000/htm.md: New file.
15294 * config/rs6000/htmintrin.h: New file.
15295 * config/rs6000/htmxlintrin.h: New file.
15297 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
15299 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15300 Define SYMBOL_TINY_GOT, update comment.
15301 * config/aarch64/aarch64.c
15302 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
15303 (aarch64_expand_mov_immediate): Likewise.
15304 (aarch64_print_operand): Likewise.
15305 (aarch64_classify_symbol): Likewise.
15306 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
15307 (ldr_got_tiny): Define.
15309 2013-07-13 Tobias Grosser <tobias@grosser.es>
15311 PR tree-optimization/54094
15312 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
15313 scheduling dimension for the parallelism check from the polyhedral
15314 information in the AST.
15315 * graphite-dependences.c (carries_deps): Do not assume the schedule is
15318 2013-07-13 Jason Merrill <jason@redhat.com>
15320 * print-tree.c (debug_vec_tree): Use debug_raw.
15321 (debug_raw (vec<tree, va_gc> &)): New.
15322 (debug_raw (vec<tree, va_gc> *)): New.
15323 * tree.h: Declare them.
15325 2013-07-13 Bin Cheng <bin.cheng@arm.com>
15327 * ifcvt.c (ifcvt_after_combine): New static variable.
15328 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
15330 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
15331 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
15332 rest_of_handle_if_after_reload): Pass new argument for if_convert.
15334 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
15336 * config/mips/mips.c (mips_expand_call): Remove empty statement.
15338 2013-07-12 Michael Matz <matz@suse.de>
15340 PR middle-end/55771
15341 * convert.c (convert_to_real): Reject non-float inner types.
15343 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
15345 * config/aarch64/aarch64-protos.h
15346 (aarch64_simd_immediate_valid_for_move): Remove.
15347 * config/aarch64/aarch64.c (simd_immediate_info): New member.
15348 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
15350 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
15352 2013-07-11 Steve Ellcey <sellcey@mips.com>
15354 * config/mips/mips.c (mips_conditional_register_usage): Do not
15355 use t[0-7] registers in MIPS16 mode when optimizing for size.
15357 2013-07-11 Sriraman Tallam <tmsriram@google.com>
15359 * config/i386/i386.c (dispatch_function_versions): Fix array
15360 indexing of function_version_info to match actual_versions.
15362 2013-07-11 Teresa Johnson <tejohnson@google.com>
15364 * vec.h (struct va_gc): Move release out-of-line.
15365 (va_gc::release): Call ggc_free on released vec.
15367 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15369 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
15370 Require GOT register as additional operand in UNSPEC.
15371 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
15372 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15373 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15374 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
15375 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
15376 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
15377 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
15379 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
15382 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
15383 name seen by assembler/linker rather if available.
15385 2013-07-11 Andreas Schwab <schwab@suse.de>
15387 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
15389 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
15391 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
15393 2013-07-10 Joseph Myers <joseph@codesourcery.com>
15395 * doc/tm.texi.in: Move hook documentation to ....
15396 * target.def: ... here.
15398 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
15399 text on @hook line.
15400 * doc/tm.texi: Regenerate.
15402 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
15405 * doc/invoke.texi: Document -Wconditionally-supported.
15407 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15410 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
15413 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15416 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
15417 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
15418 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
15419 Remove duplicate devices.
15420 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
15421 * config/avr/t-multilib: Regenerate.
15422 * config/avr/avr-tables.opt: Regenerate.
15423 * doc/avr-mmcu.texi: Regenerate.
15425 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15428 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
15430 2013-07-10 Graham Stott <graham.stott@btinternet.com>
15432 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
15433 the cost of MULT when optimizing for size.
15435 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
15437 * config/cr16/cr16-protos.h: Don't include target.h.
15439 2013-07-09 Joseph Myers <joseph@codesourcery.com>
15441 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
15442 adjust register size for TDmode and TFmode for VSX registers.
15444 2013-07-08 Kai Tietz <ktietz@redhat.com>
15447 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
15448 hook_bool_const_tree_true.
15450 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15452 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
15453 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
15454 * config/s390/s390.md: Define FPR*_REGNUM constants.
15455 Fix FPR2_REGNUM constant (18 -> 17).
15456 ("*trunc<BFP:mode><DFP_ALL:mode>2")
15457 ("*trunc<DFP_ALL:mode><BFP:mode>2")
15458 ("trunc<BFP:mode><DFP_ALL:mode>2")
15459 ("trunc<DFP_ALL:mode><BFP:mode>2")
15460 ("*extend<BFP:mode><DFP_ALL:mode>2")
15461 ("*extend<DFP_ALL:mode><BFP:mode>2")
15462 ("extend<BFP:mode><DFP_ALL:mode>2")
15463 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
15466 2013-07-08 Graham Stott <graham.stott@btinternet.com>
15468 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
15470 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15472 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
15473 and cfun_fpr_bit_p to cfun_fpr_save_p.
15474 (s390_frame_area, s390_register_info, s390_frame_info)
15475 (s390_emit_prologue, s390_emit_epilogue)
15476 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
15478 * config/s390/s390.h: Define *_REGNUM macros for floating point
15481 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
15483 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
15485 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
15487 PR rtl-optimization/57786
15488 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
15489 and break out of the loop when it is set to false.
15491 2013-07-08 Jakub Jelinek <jakub@redhat.com>
15494 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
15495 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
15496 (const_int 63)) 0)).
15497 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
15498 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
15499 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
15501 PR rtl-optimization/57829
15502 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
15503 mask bits outside of mode are just sign-extension from mode to HWI.
15505 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
15507 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
15508 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
15509 adjust_address instead of change_address to keep info about alignment.
15510 (emit_strmov): Remove.
15511 (emit_memmov): New function.
15512 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
15513 (expand_movmem_epilogue): Likewise and return updated rtx for
15515 (expand_constant_movmem_prologue): Likewise and return updated rtx for
15516 destination and source.
15517 (decide_alignment): Refactor, handle vector_loop.
15518 (ix86_expand_movmem): Likewise.
15519 (ix86_expand_setmem): Likewise.
15520 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
15522 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
15524 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
15525 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
15527 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15529 * config/i386/sse.md (sse_movlhps): Change alternative 3
15530 of operand 2 to "m".
15532 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15535 * config/i386/sse.md (iptr): New mode attribute.
15536 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
15537 (<sse>_vm<plusminus_insn><mode>3): Ditto.
15538 (<sse>_vmmul<mode>3): Ditto.
15539 (<sse>_vmdiv<mode>3): Ditto.
15540 (sse_vmrcpv4sf2): Ditto.
15541 (<sse>_vmsqrt<mode>2): Ditto.
15542 (sse_vmrsqrtv4sf2): Ditto.
15543 (<sse>_vm<code><mode>3): Ditto.
15544 (avx_vmcmp<mode>3): Ditto.
15545 (<sse>_vmmaskcmp<mode>3): Ditto.
15546 (<sse>_comi): Ditto.
15547 (<sse>_ucomi): Ditto.
15548 (*xop_vmfrcz_<mode>): Ditto.
15549 (*fmai_fmadd_<mode>): Ditto.
15550 (*fmai_fmsub_<mode>): Ditto.
15551 (*fmai_fnmadd_<mode>): Ditto.
15552 (*fmai_fnmsub_<mode>): Ditto.
15553 (*fma4i_vmfmadd_<mode>): Ditto.
15554 (*fma4i_vmfmsub_<mode>): Ditto.
15555 (*fma4i_vmfnmadd_<mode>): Ditto.
15556 (*fma4i_vmfnmsub_<mode>): Ditto.
15557 (*xop_vmfrcz_<mode>): Ditto.
15558 (sse_cvtps2pi): Ditto.
15559 (sse_cvttps2pi): Ditto.
15560 (sse_cvtss2si): Ditto.
15561 (sse_cvtss2si_2): Ditto.
15562 (sse_cvtss2siq_2): Ditto.
15563 (sse_cvttss2si): Ditto.
15564 (sse_cvttss2siq): Ditto.
15565 (sse_cvtsd2si): Ditto.
15566 (sse_cvtsd2si_2): Ditto.
15567 (sse_cvtsd2siq_2): Ditto.
15568 (sse_cvttsd2si): Ditto.
15569 (sse_cvttsd2siq): Ditto.
15570 (sse_cvtsd2ss): Ditto.
15571 (sse_cvtss2sd): Ditto.
15572 (avx2_pbroadcast<mode>): Ditto.
15573 (avx2_pbroadcast<mode>_1): Ditto.
15574 (*avx_vperm_broadcast_v4sf): Ditto.
15576 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
15577 (sse_movlhps): Ditto.
15578 (sse_storehps): Ditto.
15579 (sse_loadhps): Ditto.
15580 (sse_storelps): Ditto.
15581 (sse_loadlps): Ditto.
15582 (*vec_concatv4sf): Ditto.
15583 (*vec_interleave_highv2df): Ditto.
15584 (*vec_interleave_lowv2df): Ditto.
15585 (*vec_extractv2df_1_sse): Ditto.
15586 (*vec_extractv2df_0_sse): Ditto.
15587 (sse2_storelpd): Ditto.
15588 (sse2_loadlpd): Ditto.
15589 (sse2_movsd): Ditto.
15590 (*vec_concatv4si): Ditto.
15591 (vec_concatv2di): Ditto.
15593 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
15594 for Intel asm dialect.
15595 (mmx_punpcklwd): Ditto.
15596 (mmx_punpckldq): Ditto.
15598 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
15599 for intel assembler dialect.
15601 2013-07-06 Jakub Jelinek <jakub@redhat.com>
15604 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
15605 for BUILT_IN_C{LZ,LRSB}*.
15606 * tree.h (CASE_INT_FN): Add FN##IMAX case.
15607 * tree-vrp.c (extract_range_basic): Handle
15608 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
15609 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
15610 fall thru to code calling set_value*.
15611 * builtins.c (expand_builtin): Remove *IMAX cases.
15612 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
15613 if width is bigger than 2*HWI.
15615 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
15617 PR rtl-optimization/55342
15618 * lra-int.h (lra_subreg_reload_pseudos): New.
15619 * lra.c: Add undoing optional reloads to the block diagram.
15620 (lra_subreg_reload_pseudos): New.
15621 (lra_optional_reload_pseudos): Change comments.
15622 (lra): Init and clear lra_subreg_reload_pseudos. Clear
15623 lra_optional_reload_pseudos after undo transformations.
15624 * lra-assigns.c (pseudo_prefix_title): New.
15625 (lra_setup_reg_renumber): Use it.
15626 (spill_for): Ditto. Check subreg reload pseudos too.
15627 (assign_by_spills): Consider subreg reload pseudos too.
15628 * lra-constraints.c (simplify_operand_subreg): Use
15629 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
15630 (curr_insn_transform): Recognize and do optional reloads.
15631 (undo_optional_reloads): New.
15632 (lra_undo_inheritance): Call undo_optional_reloads.
15634 2013-07-05 Thomas Quinot <quinot@adacore.com>
15636 * tree-complex.c (expand_complex_operations_1): Fix typo.
15638 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
15640 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
15641 (tune_params): New member 'const vec_costs'.
15642 * config/aarch64/aarch64.c (generic_vector_cost): New.
15643 (generic_tunings): New member 'generic_vector_cost'.
15644 (aarch64_builtin_vectorization_cost): New.
15645 (aarch64_add_stmt_cost): New.
15646 (TARGET_VECTORIZE_ADD_STMT_COST): New.
15647 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
15649 2013-07-03 Jakub Jelinek <jakub@redhat.com>
15652 * config/i386/predicates.md (vsib_address_operand): Disallow
15653 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
15655 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
15657 PR middle-end/55030
15658 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
15659 expand_builtin_setjmp_receiver.
15660 (expand_label): Adjust, call expand_builtin_setjmp_receiver
15661 with NULL for the label parameter.
15662 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
15663 the frame-pointer. Adjust comments.
15664 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
15665 only if LABEL is non-NULL.
15667 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
15669 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
15670 (ARM_ABI_AAPCS64): Ditto.
15672 (ARM_DEFAULT_ABI): Ditto.
15674 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
15676 * config/aarch64/aarch64-builtins.c
15677 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
15678 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
15680 * config/aarch64/aarch64-simd.md
15681 (aarch64_ld1<VALL:mode>): New.
15682 (aarch64_st1<VALL:mode>): Likewise.
15683 * config/aarch64/arm_neon.h
15684 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
15686 2013-07-02 Sriraman Tallam <tmsriram@google.com>
15688 * config/i386/i386.c (gate_insert_vzeroupper): Check if
15690 (ix86_option_override_internal):Turn on all -mavx target flags by
15691 default as they are dependent on AVX anyway.
15693 2013-07-02 Cary Coutant <ccoutant@google.com>
15695 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
15696 deterministic hash.
15697 (loc_checksum_ordered): Likewise.
15698 (hash_loc_operands): Remove inline keyword.
15700 2013-07-02 Jakub Jelinek <jakub@redhat.com>
15702 PR tree-optimization/57741
15703 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
15704 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
15705 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
15706 Allow REAL_CST step_exprs if flag_associative_math.
15707 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
15709 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15711 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
15713 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15715 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
15717 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15719 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
15721 (iorsi3_insn): Likewise.
15722 (arm_xorsi3): Likewise.
15724 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
15726 * arm.md (attribute "wtype"): Delete. Move attribute values from here
15728 (attribute "type"): ... here, and prefix with "wmmx_".
15729 (attribute "core_cycles"): Update for attribute changes.
15730 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
15731 (tbcstv4hi): Likewise.
15732 (tbcstv2si): Likewise.
15733 (iwmmxt_iordi3): Likewise.
15734 (iwmmxt_xordi3): Likewise.
15735 (iwmmxt_anddi3): Likewise.
15736 (iwmmxt_nanddi3): Likewise.
15737 (iwmmxt_arm_movdi): Likewise.
15738 (iwmmxt_movsi_insn): Likewise.
15739 (mov<mode>_internal): Likewise.
15740 (and<mode>3_iwmmxt): Likewise.
15741 (ior<mode>3_iwmmxt): Likewise.
15742 (xor<mode>3_iwmmxt): Likewise.
15743 (add<mode>3_iwmmxt): Likewise.
15744 (ssaddv8qi3): Likewise.
15745 (ssaddv4hi3): Likewise.
15746 (ssaddv2si3): Likewise.
15747 (usaddv8qi3): Likewise.
15748 (usaddv4hi3): Likewise.
15749 (usaddv2si3): Likewise.
15750 (sub<mode>3_iwmmxt): Likewise.
15751 (sssubv8qi3): Likewise.
15752 (sssubv4hi3): Likewise.
15753 (sssubv2si3): Likewise.
15754 (ussubv8qi3): Likewise.
15755 (ussubv4hi3): Likewise.
15756 (ussubv2si3): Likewise.
15757 (mulv4hi3_iwmmxt): Likewise.
15758 (smulv4hi3_highpart): Likewise.
15759 (umulv4hi3_highpart): Likewise.
15760 (iwmmxt_wmacs): Likewise.
15761 (iwmmxt_wmacsz): Likewise.
15762 (iwmmxt_wmacu): Likewise.
15763 (iwmmxt_wmacuz): Likewise.
15764 (iwmmxt_clrdi): Likewise.
15765 (iwmmxt_clrv8qi): Likewise.
15766 (iwmmxt_clr4hi): Likewise.
15767 (iwmmxt_clr2si): Likewise.
15768 (iwmmxt_uavgrndv8qi3): Likewise.
15769 (iwmmxt_uavgrndv4hi3): Likewise.
15770 (iwmmxt_uavgv8qi3): Likewise.
15771 (iwmmxt_uavgv4hi3): Likewise.
15772 (iwmmxt_tinsrb): Likewise.
15773 (iwmmxt_tinsrh): Likewise.
15774 (iwmmxt_tinsrw): Likewise.
15775 (iwmmxt_textrmub): Likewise.
15776 (iwmmxt_textrmsb): Likewise.
15777 (iwmmxt_textrmuh): Likewise.
15778 (iwmmxt_textrmsh): Likewise.
15779 (iwmmxt_textrmw): Likewise.
15780 (iwmxxt_wshufh): Likewise.
15781 (eqv8qi3): Likewise.
15782 (eqv4hi3): Likewise.
15783 (eqv2si3): Likewise.
15784 (gtuv8qi3): Likewise.
15785 (gtuv4hi3): Likewise.
15786 (gtuv2si3): Likewise.
15787 (gtv8qi3): Likewise.
15788 (gtv4hi3): Likewise.
15789 (gtv2si3): Likewise.
15790 (smax<mode>3_iwmmxt): Likewise.
15791 (umax<mode>3_iwmmxt): Likewise.
15792 (smin<mode>3_iwmmxt): Likewise.
15793 (umin<mode>3_iwmmxt): Likewise.
15794 (iwmmxt_wpackhss): Likewise.
15795 (iwmmxt_wpackwss): Likewise.
15796 (iwmmxt_wpackdss): Likewise.
15797 (iwmmxt_wpackhus): Likewise.
15798 (iwmmxt_wpackwus): Likewise.
15799 (iwmmxt_wpackdus): Likewise.
15800 (iwmmxt_wunpckihb): Likewise.
15801 (iwmmxt_wunpckihh): Likewise.
15802 (iwmmxt_wunpckihw): Likewise.
15803 (iwmmxt_wunpckilb): Likewise.
15804 (iwmmxt_wunpckilh): Likewise.
15805 (iwmmxt_wunpckilw): Likewise.
15806 (iwmmxt_wunpckehub): Likewise.
15807 (iwmmxt_wunpckehuh): Likewise.
15808 (iwmmxt_wunpckehuw): Likewise.
15809 (iwmmxt_wunpckehsb): Likewise.
15810 (iwmmxt_wunpckehsh): Likewise.
15811 (iwmmxt_wunpckehsw): Likewise.
15812 (iwmmxt_wunpckelub): Likewise.
15813 (iwmmxt_wunpckeluh): Likewise.
15814 (iwmmxt_wunpckeluw): Likewise.
15815 (iwmmxt_wunpckelsb): Likewise.
15816 (iwmmxt_wunpckelsh): Likewise.
15817 (iwmmxt_wunpckelsw): Likewise.
15818 (ror<mode>3): Likewise.
15819 (ashr<mode>3_iwmmxt): Likewise.
15820 (lshr<mode>3_iwmmxt): Likewise.
15821 (ashl<mode>3_iwmmxt): Likewise.
15822 (ror<mode>3_di): Likewise.
15823 (ashr<mode>3_di): Likewise.
15824 (lshr<mode>3_di): Likewise.
15825 (ashl<mode>3_di): Likewise.
15826 (iwmmxt_wmadds): Likewise.
15827 (iwmmxt_wmaddu): Likewise.
15828 (iwmmxt_tmia): Likewise.
15829 (iwmmxt_tmiaph): Likewise.
15830 (iwmmxt_tmiabb): Likewise.
15831 (iwmmxt_tmiatb): Likewise.
15832 (iwmmxt_tmiabt): Likewise.
15833 (iwmmxt_tmiatt): Likewise.
15834 (iwmmxt_tmovmskb): Likewise.
15835 (iwmmxt_tmovmskh): Likewise.
15836 (iwmmxt_tmovmskw): Likewise.
15837 (iwmmxt_waccb): Likewise.
15838 (iwmmxt_wacch): Likewise.
15839 (iwmmxt_waccw): Likewise.
15840 (iwmmxt_waligni): Likewise.
15841 (iwmmxt_walignr): Likewise.
15842 (iwmmxt_walignr0): Likewise.
15843 (iwmmxt_walignr1): Likewise.
15844 (iwmmxt_walignr2): Likewise.
15845 (iwmmxt_walignr3): Likewise.
15846 (iwmmxt_wsadb): Likewise.
15847 (iwmmxt_wsadh): Likewise.
15848 (iwmmxt_wsadbz): Likewise.
15849 (iwmmxt_wsadhz): Likewise.
15850 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
15851 (iwmmxt_wabsdiffb): Likewise.
15852 (iwmmxt_wabsdiffh): Likewise.
15853 (iwmmxt_wabsdiffw): Likewise.
15854 (iwmmxt_waddsubhx): Likewise
15855 (iwmmxt_wsubaddhx): Likewise.
15856 (addc<mode>3): Likewise.
15857 (iwmmxt_avg4): Likewise.
15858 (iwmmxt_avg4r): Likewise.
15859 (iwmmxt_wmaddsx): Likewise.
15860 (iwmmxt_wmaddux): Likewise.
15861 (iwmmxt_wmaddsn): Likewise.
15862 (iwmmxt_wmaddun): Likewise.
15863 (iwmmxt_wmulwsm): Likewise.
15864 (iwmmxt_wmulwum): Likewise.
15865 (iwmmxt_wmulsmr): Likewise.
15866 (iwmmxt_wmulumr): Likewise.
15867 (iwmmxt_wmulwsmr): Likewise.
15868 (iwmmxt_wmulwumr): Likewise.
15869 (iwmmxt_wmulwl): Likewise.
15870 (iwmmxt_wqmulm): Likewise.
15871 (iwmmxt_wqmulwm): Likewise.
15872 (iwmmxt_wqmulmr): Likewise.
15873 (iwmmxt_wqmulwmr): Likewise.
15874 (iwmmxt_waddbhusm): Likewise.
15875 (iwmmxt_waddbhusl): Likewise.
15876 (iwmmxt_wqmiabb): Likewise.
15877 (iwmmxt_wqmiabt): Likewise.
15878 (iwmmxt_wqmiatb): Likewise.
15879 (iwmmxt_wqmiatt): Likewise.
15880 (iwmmxt_wqmiabbn): Likewise.
15881 (iwmmxt_wqmiabtn): Likewise.
15882 (iwmmxt_wqmiatbn): Likewise.
15883 (iwmmxt_wqmiattn): Likewise.
15884 (iwmmxt_wmiabb): Likewise.
15885 (iwmmxt_wmiabt): Likewise.
15886 (iwmmxt_wmiatb): Likewise.
15887 (iwmmxt_wmiatt): Likewise.
15888 (iwmmxt_wmiabbn): Likewise.
15889 (iwmmxt_wmiabtn): Likewise.
15890 (iwmmxt_wmiatbn): Likewise.
15891 (iwmmxt_wmiattn): Likewise.
15892 (iwmmxt_wmiawbb): Likewise.
15893 (iwmmxt_wmiawbt): Likewise.
15894 (iwmmxt_wmiawtb): Likewise.
15895 (iwmmxt_wmiawtt): Likewise.
15896 (iwmmxt_wmiawbbn): Likewise.
15897 (iwmmxt_wmiawbtn): Likewise.
15898 (iwmmxt_wmiawtbn): Likewise.
15899 (iwmmxt_wmiawttn): Likewise.
15900 (iwmmxt_wmerge): Likewise.
15901 (iwmmxt_tandc<mode>3): Likewise.
15902 (iwmmxt_torc<mode>3): Likewise.
15903 (iwmmxt_torvsc<mode>3): Likewise.
15904 (iwmmxt_textrc<mode>3): Likewise.
15905 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
15906 (wmmxt_pack): Likewise.
15907 (wmmxt_mult_c1): Likewise.
15908 (wmmxt_mult_c2): Likewise.
15909 (wmmxt_alu_c1): Likewise.
15910 (wmmxt_alu_c2): Likewise.
15911 (wmmxt_alu_c3): Likewise.
15912 (wmmxt_transfer_c1): Likewise.
15913 (wmmxt_transfer_c2): Likewise.
15914 (wmmxt_transfer_c3): Likewise.
15915 (marvell_f_iwmmxt_wstr): Likewise.
15916 (marvell_f_iwmmxt_wldr): Likewise.
15918 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
15920 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
15922 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15925 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15926 * lra-constraints.c (need_for_split_p): Check call used hard regs
15927 living through calls.
15929 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
15930 call used regs for call insn.
15932 2013-06-28 Jakub Jelinek <jakub@redhat.com>
15935 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
15936 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
15938 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15940 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
15942 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15944 * lra-constraints.c (need_for_split_p): Check call used hard regs
15945 living through calls.
15947 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
15950 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
15951 to tie with any other modes. Eliminate Altivec vector mode tests,
15952 since these are a subset of ALTIVEC or VSX vector modes. Simplify
15953 code, to return 0 if testing MODE2 for a condition, if we've
15954 already tested MODE1 for the same condition.
15956 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
15958 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
15961 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
15963 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15964 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
15966 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
15968 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
15970 (aarch64_symbolic_constant_p): Remove.
15971 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
15972 static. Fix line length and white space.
15973 (aarch64_symbolic_constant_p): Remove.
15974 * config/aarch64/predicates.md (aarch64_valid_symref):
15975 Use aarch64_classify_symbol_expression.
15977 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15979 * config/arm/constraints.md (Ts): New constraint.
15980 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
15982 (compare_scc): Use "Ts" constraint for operand 0.
15983 (ior_scc_scc): Likewise.
15984 (and_scc_scc): Likewise.
15985 (and_scc_scc_nodom): Likewise.
15986 (ior_scc_scc_cmp): Likewise for operand 7.
15987 (and_scc_scc_cmp): Likewise.
15988 * config/arm/thumb2.md (thumb2_movsi_insn):
15989 Add alternatives for 16-bit encodings.
15990 (thumb2_movhi_insn): Likewise.
15991 (thumb2_movsicc_insn): Likewise.
15992 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
15993 (thumb2_negscc): Use "Ts" constraint.
15994 Move mvn instruction outside cond_exec block.
15995 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
15996 for 16-bit encodings.
15998 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16000 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
16002 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
16003 (mulsi3subsi): Likewise.
16004 (mulsidi3adddi): Likewise.
16005 (mulsidi3_v6): Likewise.
16006 (umulsidi3_v6): Likewise.
16007 (umulsidi3adddi_v6): Likewise.
16008 (smulsi3_highpart_v6): Likewise.
16009 (umulsi3_highpart_v6): Likewise.
16010 (mulhisi3tb): Likewise.
16011 (mulhisi3bt): Likewise.
16012 (mulhisi3tt): Likewise.
16013 (maddhisi4): Likewise.
16014 (maddhisi4tb): Likewise.
16015 (maddhisi4tt): Likewise.
16016 (maddhidi4): Likewise.
16017 (maddhidi4tb): Likewise.
16018 (maddhidi4tt): Likewise.
16019 (zeroextractsi_compare0_scratch): Likewise.
16020 (insv_zero): Likewise.
16021 (insv_t2): Likewise.
16022 (anddi_notzesidi_di): Likewise.
16023 (anddi_notsesidi_di): Likewise.
16024 (andsi_notsi_si): Likewise.
16025 (iordi_zesidi_di): Likewise.
16026 (xordi_zesidi_di): Likewise.
16027 (andsi_iorsi3_notsi): Likewise.
16028 (smax_0): Likewise.
16029 (smax_m1): Likewise.
16030 (smin_0): Likewise.
16031 (not_shiftsi): Likewise.
16032 (unaligned_loadsi): Likewise.
16033 (unaligned_loadhis): Likewise.
16034 (unaligned_loadhiu): Likewise.
16035 (unaligned_storesi): Likewise.
16036 (unaligned_storehi): Likewise.
16037 (extv_reg): Likewise.
16038 (extzv_t2): Likewise.
16039 (divsi3): Likewise.
16040 (udivsi3): Likewise.
16041 (arm_zero_extendhisi2addsi): Likewise.
16042 (arm_zero_extendqisi2addsi): Likewise.
16043 (compareqi_eq0): Likewise.
16044 (arm_extendhisi2_v6): Likewise.
16045 (arm_extendqisi2addsi): Likewise.
16046 (arm_movt): Likewise.
16047 (thumb2_ldrd): Likewise.
16048 (thumb2_ldrd_base): Likewise.
16049 (thumb2_ldrd_base_neg): Likewise.
16050 (thumb2_strd): Likewise.
16051 (thumb2_strd_base): Likewise.
16052 (thumb2_strd_base_neg): Likewise.
16053 (arm_negsi2): Add alternative for 16-bit encoding.
16054 (arm_one_cmplsi2): Likewise.
16056 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16058 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
16059 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
16060 (movdfcc): Likewise.
16061 * config/arm/vfp.md (*thumb2_movsf_vfp):
16062 Disable predication for arm_restrict_it.
16063 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
16064 (*thumb2_movdfcc_vfp): Likewise.
16065 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
16066 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
16067 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
16068 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
16069 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
16070 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
16071 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
16072 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
16073 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
16074 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
16075 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
16076 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
16077 Disable predication for arm_restrict_it.
16079 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
16081 * config/i386/bmiintrin.h (_bextr_u32): New.
16082 (_bextr_u64): Ditto.
16084 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
16086 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
16087 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
16088 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
16089 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
16090 * config/mips/n32-elf.h: ...this new file.
16092 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
16095 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
16096 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
16098 2013-06-27 Catherine Moore <clm@codesourcery.com>
16100 * config/mips/mips-tables.opt: Regenerate.
16101 * config/mips/mips-cpus.def: Add m14ke and m14kec.
16102 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
16103 * doc/invoke.texi: Add -m14kc.
16105 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16108 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
16109 constraints of operand 1 and 2.
16112 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
16113 to match RTL canonicalization. Swap predicates and
16114 constraints of operand 1 and 2.
16116 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
16118 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
16119 Process OP_INOUT regs for splitting too.
16121 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16123 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
16124 decl before the loop, initialize to NULL.
16125 (vectorizable_load): Initialize ptr_incr to NULL.
16127 2013-06-27 Martin Jambor <mjambor@suse.cz>
16130 * ipa-ref.h (ipa_maybe_record_reference): Declare.
16131 * ipa-ref.c (ipa_maybe_record_reference): New function.
16132 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
16133 * ipa-cp.c (create_specialized_node): Record potential references from
16135 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
16137 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16139 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
16140 parameter 'mode' of type 'enum machine_mode mode'; change to pass
16141 'mode' to force_reg.
16142 (aarch64_add_offset): Update calls to aarch64_force_temporary.
16143 (aarch64_expand_mov_immediate): Likewise.
16145 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16147 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
16148 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
16150 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16152 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
16153 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16154 (struct machine_function): Add tbegin_p.
16155 (s390_canonicalize_comparison): Fold CC mode compares to
16156 conditional jump if possible.
16157 (s390_emit_jump): Return the emitted jump.
16158 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
16159 Handle CCRAWmode compares.
16160 (s390_option_override): Default to -mhtm if available.
16161 (s390_reg_clobbered_rtx): Handle floating point regs as well.
16162 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
16163 FPRs instead of df_regs_ever_live_p.
16164 (s390_optimize_nonescaping_tx): New function.
16165 (s390_init_frame_layout): Extend clobbered_regs array to cover
16167 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
16168 (s390_expand_tbegin): New function.
16169 (enum s390_builtin): New enum definition.
16170 (code_for_builtin): New array definition.
16171 (s390_init_builtins): New function.
16172 (s390_expand_builtin): New function.
16173 (TARGET_INIT_BUILTINS): Define.
16174 (TARGET_EXPAND_BUILTIN): Define.
16175 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
16176 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
16177 (s390_alc_comparison): Likewise.
16178 * config/s390/s390-modes.def: Add CCRAWmode.
16179 * config/s390/s390.h (processor_flags): Add PF_TX.
16180 (TARGET_CPU_HTM): Define macro.
16181 (TARGET_HTM): Define macro.
16182 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
16183 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
16184 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16185 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
16186 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
16187 (TBEGIN_MASK, TBEGINC_MASK): New constants.
16188 ("*cc_to_int"): Move up.
16189 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
16190 constants other than 0.
16191 ("*ccraw_to_int"): New insn and splitter definition.
16192 ("tbegin", "tbegin_nofloat", "tbegin_retry")
16193 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
16194 ("tx_assist"): New expander.
16195 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
16196 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
16197 * config/s390/s390.opt: Add -mhtm option.
16198 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
16199 * config/s390/htmxlintrin.h: New file.
16200 * config/s390/htmintrin.h: New file.
16201 * config/s390/s390intrin.h: New file.
16202 * doc/extend.texi: Document htm builtins.
16203 * config.gcc: Add the new header files to extra_headers.
16205 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
16207 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
16208 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
16210 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
16211 Pat Haugen <pthaugen@us.ibm.com>
16212 Peter Bergner <bergner@vnet.ibm.com>
16214 * config/rs6000/power8.md: New.
16215 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
16216 setting for power8 entry.
16217 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
16218 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
16219 test for Power4/Power5 only.
16220 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
16222 (force_new_group): Adjust comment.
16223 * config/rs6000/rs6000.md: Include power8.md.
16225 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
16227 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
16228 * config/arm/arm-protos.h (arm_max_conditional_execute): New
16230 (tune_params): Update comment.
16231 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
16232 (arm_max_conditional_execute): New function.
16233 (thumb2_final_prescan_insn): Use max_insn_skipped and
16234 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
16236 2013-06-25 Jakub Jelinek <jakub@redhat.com>
16238 PR tree-optimization/57705
16239 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
16240 SSA_NAME step, provided that it is not defined inside the loop.
16241 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
16242 (get_initial_def_for_induction): Handle SSA_NAME IV step.
16244 2013-06-25 Martin Jambor <mjambor@suse.cz>
16246 PR middle-end/57670
16247 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
16248 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
16250 (ipa_note_param_call): Initialize member_ptr flag.
16251 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
16252 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
16253 (ipa_write_indirect_edge_info): Stream member_ptr flag.
16254 (ipa_read_indirect_edge_info): Likewise.
16256 2013-06-25 Richard Biener <rguenther@suse.de>
16258 PR middle-end/56977
16259 * passes.c (init_optimization_passes): Move pass_fold_builtins
16260 and pass_dce earlier with -Og.
16262 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
16264 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
16265 <BIT_FIELD_REF>: Remove trailing TAB.
16266 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
16269 2013-06-24 Martin Jambor <mjambor@suse.cz>
16271 PR tree-optimization/57358
16272 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
16273 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
16274 (ipa_analyze_params_uses): Generate pessimistic info when true.
16276 2013-06-24 Martin Jambor <mjambor@suse.cz>
16278 PR tree-optimization/57539
16279 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
16280 global.inlined_to of the new node to it. All callers changed.
16281 * ipa-inline-transform.c (clone_inlined_nodes): New variable
16282 inlining_into, pass it to cgraph_clone_node.
16283 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
16284 ipa_free_edge_args_substructures.
16285 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
16286 rdesc linked list. Do not assert rdesc edges have inlined caller.
16287 Assert we have found an rdesc in the rdesc list.
16289 2013-06-24 Richard Biener <rguenther@suse.de>
16291 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
16292 (pointer_set_lookup): Declare.
16293 (class pointer_map): New template class implementing a
16294 generic pointer to T map.
16295 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
16296 pointer_map<T>::contains, pointer_map<T>::insert,
16297 pointer_map<T>::traverse): New functions.
16298 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
16299 (pointer_set_lookup): New function.
16300 (pointer_set_contains): Use pointer_set_lookup.
16301 (pointer_set_insert): Likewise.
16302 (insert_aux): Remove.
16303 (struct pointer_map_t): Embed a pointer_set_t.
16304 (pointer_map_create): Adjust.
16305 (pointer_map_destroy): Likewise.
16306 (pointer_map_contains): Likewise.
16307 (pointer_map_insert): Likewise.
16308 (pointer_map_traverse): Likewise.
16309 * tree-streamer.h (struct streamer_tree_cache_d): Use a
16310 pointer_map<unsigned> instead of a pointer_map_t.
16311 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
16312 (streamer_tree_cache_lookup): Likewise.
16313 (streamer_tree_cache_create): Likewise.
16314 (streamer_tree_cache_delete): Likewise.
16315 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
16316 pointer_map<unsigned> instead of a pointer_map_t.
16317 (lto_init_tree_ref_encoder): Adjust.
16318 (lto_destroy_tree_ref_encoder): Likewise.
16319 * lto-section-out.c (lto_output_decl_index): Likewise.
16320 (lto_record_function_out_decl_state): Likewise.
16321 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
16323 2013-06-24 Richard Biener <rguenther@suse.de>
16325 PR tree-optimization/57488
16326 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
16328 2013-06-24 Alan Modra <amodra@gmail.com>
16330 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
16331 (gen_easy_altivec_constant): Likewise.
16332 * config/rs6000/predicates.md (easy_vector_constant_add_self,
16333 easy_vector_constant_msb): Likewise.
16335 2013-06-23 Jakub Jelinek <jakub@redhat.com>
16338 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
16339 add missing return true.
16341 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
16344 * config/sh/predicates.md (general_extend_operand): Invoke
16345 general_movsrc_operand for memory operands.
16346 (general_movsrc_operand): Allow reg+reg addressing, do not use
16347 general_operand for memory operands.
16349 2013-06-23 Sriraman Tallam <tmsriram@google.com>
16351 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
16352 when current target options does not apply.
16353 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
16354 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
16355 * config/i386/bmiintrin.h: Pass appropriate target
16356 attributes to header.
16357 * config/i386/mmintrin.h: Ditto.
16358 * config/i386/nmmintrin.h: Ditto.
16359 * config/i386/avx2intrin.h: Ditto.
16360 * config/i386/fxsrintrin.h: Ditto.
16361 * config/i386/tbmintrin.h: Ditto.
16362 * config/i386/xsaveintrin.h: Ditto.
16363 * config/i386/f16cintrin.h: Ditto.
16364 * config/i386/xtestintrin.h: Ditto.
16365 * config/i386/xsaveoptintrin.h: Ditto.
16366 * config/i386/bmi2intrin.h: Ditto.
16367 * config/i386/lzcntintrin.h: Ditto.
16368 * config/i386/smmintrin.h: Ditto.
16369 * config/i386/wmmintrin.h: Ditto.
16370 * config/i386/x86intrin.h: Remove all header include guards.
16371 * config/i386/prfchwintrin.h: Ditto.
16372 * config/i386/pmmintrin.h: Ditto.
16373 * config/i386/tmmintrin.h: Ditto.
16374 * config/i386/xmmintrin.h: Ditto.
16375 * config/i386/popcntintrin.h: Ditto.
16376 * config/i386/rdseedintrin.h: Ditto.
16377 * config/i386/ammintrin.h: Ditto.
16378 * config/i386/emmintrin.h: Ditto.
16379 * config/i386/immintrin.h: Remove all header include guards.
16380 * config/i386/fma4intrin.h: Ditto.
16381 * config/i386/lwpintrin.h: Ditto.
16382 * config/i386/xopintrin.h: Ditto.
16383 * config/i386/ia32intrin.h: Ditto.
16384 * config/i386/avxintrin.h: Ditto.
16385 * config/i386/rtmintrin.h: Ditto.
16386 * config/i386/fmaintrin.h: Ditto.
16387 * config/i386/mm3dnow.h: Ditto.
16389 2013-06-22 Sriraman Tallam <tmsriram@google.com>
16391 * common/config/i386/i386-common.c: Handle LZCNT.
16393 2013-06-22 Andi Kleen <ak@linux.intel.com>
16395 * doc/extend.texi: Use __atomic_store_n instead of
16396 __atomic_store in HLE example.
16398 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
16400 * config/sh/sh.c: Remove <cstdlib> workaround.
16402 2013-06-21 Andi Kleen <ak@linux.intel.com>
16404 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
16406 2013-06-21 Andi Kleen <ak@linux.intel.com>
16408 * doc/extend.texi: Document that __atomic_clear and
16409 __atomic_test_and_set should only be used with bool.
16411 2013-06-20 Jan Hubicka <jh@suse.cz>
16413 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
16414 types_same_for_odr.
16415 * tree.c (decls_same_for_odr): New function.
16416 (same_for_edr): New function.
16417 (types_same_for_odr): New function.
16418 (get_binfo_at_offset): Use it.
16419 * tree.h (types_same_for_odr): Declare.
16421 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
16422 Jason Merrill <jason@redhat.com>
16424 * system.h: Include <cstdlib> as well as <stdlib.h>.
16426 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
16429 * config/i386/i386.c (construct_container): Report error if
16430 long double is used with disabled x87 float returns.
16432 2013-06-20 Jan Hubicka <jh@suse.cz>
16434 * lto-cgraph.c (input_symtab): Do not set cgraph state.
16436 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
16438 PR rtl-optimization/57425
16439 PR rtl-optimization/57569
16440 * alias.c (write_dependence_p): Remove parameters mem_mode and
16441 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
16442 Changed all callers.
16443 (canon_anti_dependence): Get comments and semantics in sync.
16444 Add parameter mem_canonicalized. Changed all callers.
16445 * rtl.h (canon_anti_dependence): Update prototype.
16447 2013-06-20 Richard Biener <rguenther@suse.de>
16449 * data-streamer-in.c (streamer_read_uhwi): Optimize single
16450 byte case, inline streamer_read_uchar and defer section
16453 2013-06-20 Richard Biener <rguenther@suse.de>
16455 PR tree-optimization/57584
16456 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
16457 SSA names into the expanded expression that take part in
16458 abnormal coalescing.
16460 2013-06-19 Sharad Singhai <singhai@google.com>
16462 * gcov.c (print_usage): Handle new option.
16463 (process_args): Ditto.
16464 (get_gcov_intermediate_filename): New function.
16465 (output_intermediate_file): New function.
16466 (output_gcov_file): New function
16467 (generate_results): Handle new option.
16468 (release_function): Relase demangled name.
16469 (read_graph_file): Handle demangled name.
16470 (output_lines): Ditto.
16471 * doc/gcov.texi: Document gcov intermediate format.
16473 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
16476 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
16477 (lra_emit_add): Use the functions. Add comment about Y as an
16480 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
16483 * collect2.c (collect_atexit): New.
16484 (collect_exit): Delete.
16485 (main): Register collect_atexit with atexit.
16486 (collect_wait): Change collect_exit to exit.
16488 * collect2.h (collect_exit): Delete.
16489 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
16491 2013-06-19 Wei Mi <wmi@google.com>
16493 PR rtl-optimization/57518
16494 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
16495 if regno is used in paradoxical subreg.
16496 (update_equiv_regs): Check pdx_subregs[regno] before
16497 set a reg to be equivalent with a mem.
16499 2013-06-19 Matthias Klose <doko@ubuntu.com>
16502 * file-find.h (find_a_file): Add a mode parameter.
16503 * file-find.c (find_a_file): Likewise.
16504 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
16505 with X_OK for the executables.
16506 * collect2.c (main): Call find_a_file with X_OK.
16508 2013-06-19 Steve Ellcey <sellcey@mips.com>
16511 * config/mips/mips.md (casesi_internal_mips16_<mode>):
16512 Use NEXT_INSN instead of next_real_insn.
16514 2013-06-19 Jan Hubicka <jh@suse.cz>
16516 * cgraph.h (const_value_known_p): Replace by ...
16517 (ctor_for_folding): .. this one.
16518 * cgraphunit.c (process_function_and_variable_attributes): Use it.
16519 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
16520 * expr.c (expand_expr_real_1): Likewise.
16521 (string_constant): Likewise.
16522 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
16523 * ipa.c (process_references): Likewise.
16524 (symtab_remove_unreachable_nodes): Likewise.
16525 * ipa-inline-analysis.c (param_change_prob): Likewise.
16526 * gimple-fold.c (canonicalize_constructor_val): Likewise.
16527 (get_base_constructor): Likwise.
16528 * varpool.c (varpool_remove_node): Likewise.
16529 (varpool_remove_initializer): LIkewise.
16530 (dump_varpool_node): LIkwise.
16531 (const_value_known_p): Rewrite to ...
16532 (ctor_for_folding): ... this one.
16534 2013-06-19 Jakub Jelinek <jakub@redhat.com>
16537 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
16538 PERSONALITY in $PATH derived prefixes.
16540 2013-06-19 Jeff Law <law@redhat.com>
16542 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
16545 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
16546 (simplify_bitwise_binary): Use it to simpify certain binary ops on
16549 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
16551 * config/arm/vfp.md: Move VFP instruction classification documentation
16553 * config/arm/arm.md: ... here. Update instruction classification
16556 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
16558 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
16559 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
16560 pattern. Use more efficient sequences on ARMv5 and Thumb2.
16562 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
16565 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16566 with NEXT_INSN. Use tablejump_p to check for jump table data
16569 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
16572 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
16573 that now in C++ the value is correct per the C++ standards.
16575 2013-06-19 Richard Biener <rguenther@suse.de>
16577 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
16578 for global context.
16580 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16583 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16586 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16587 with next_active_insn.
16589 2013-06-18 Sriraman Tallam <tmsriram@google.com>
16591 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
16592 functions are inlined during failures to flag an error.
16593 * tree-inline.c (expand_call_inline): Allow the error to be flagged
16594 in early inline pass.
16596 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
16598 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
16601 2013-06-18 Julian Brown <julian@codesourcery.com>
16603 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
16604 Permit virtual register pre-reload if !strict.
16605 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
16607 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
16609 * config/arm/neon.md (movmisalign<mode>): Use
16610 neon_perm_struct_or_reg_operand instead of
16611 neon_struct_or_register_operand.
16612 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
16613 neon_permissive_struct_operand instead of neon_struct_operand.
16614 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
16615 neon_vector_mem_operand.
16616 * config/arm/predicates.md (neon_struct_operand): Adjust call to
16617 neon_vector_mem_operand.
16618 (neon_permissive_struct_operand): New.
16619 (neon_struct_or_register_operand): Rename to...
16620 (neon_perm_struct_or_reg_operand): This. Adjust call to
16621 neon_vector_mem_operand.
16623 2013-06-18 Richard Biener <rguenther@suse.de>
16625 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
16626 * lto-streamer.h: Include pointer-set.h.
16627 (struct lto_decl_slot): Remove.
16628 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
16629 Remove next_index entry.
16630 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16631 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16632 (lto_init_tree_ref_encoder): Adjust.
16633 (lto_destroy_tree_ref_encoder): Likewise.
16634 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16635 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16636 (lto_output_decl_index): Adjust.
16637 (lto_new_out_decl_state): Likewise.
16638 (lto_record_function_out_decl_state): Likewise.
16639 * lto-streamer-out.c (copy_function): Likewise.
16641 2013-06-18 Richard Biener <rguenther@suse.de>
16643 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
16644 * cgraphunit.c: Include cfgloop.h.
16645 (init_lowered_empty_function): Initialize the loop tree.
16646 (assemble_thunk): Insert new BBs into loops.
16648 2013-06-18 Richard Biener <rguenther@suse.de>
16650 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
16651 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
16652 the map from cache entry to cache index optional.
16653 (streamer_tree_cache_replace_tree): Adjust accordingly.
16654 (streamer_tree_cache_append): Likewise.
16655 (streamer_tree_cache_delete): Likewise.
16656 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
16657 streamer cache map from cache entry to cache index.
16658 * lto-streamer-out.c (create_output_block): Adjust.
16660 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
16662 * config/arm/arm.md (attribute "insn"): Move multiplication and
16663 division attributes to...
16664 (attribute "type"): ... here. Remove mult.
16665 (attribute "mul32"): New attribute.
16666 (attribute "mul64"): Add umaal.
16667 (*arm_mulsi3): Update attributes.
16668 (*arm_mulsi3_v6): Likewise.
16669 (*thumb_mulsi3): Likewise.
16670 (*thumb_mulsi3_v6): Likewise.
16671 (*mulsi3_compare0): Likewise.
16672 (*mulsi3_compare0_v6): Likewise.
16673 (*mulsi_compare0_scratch): Likewise.
16674 (*mulsi_compare0_scratch_v6): Likewise.
16675 (*mulsi3addsi): Likewise.
16676 (*mulsi3addsi_v6): Likewise.
16677 (*mulsi3addsi_compare0): Likewise.
16678 (*mulsi3addsi_compare0_v6): Likewise.
16679 (*mulsi3addsi_compare0_scratch): Likewise.
16680 (*mulsi3addsi_compare0_scratch_v6): Likewise.
16681 (*mulsi3subsi): Likewise.
16682 (*mulsidi3adddi): Likewise.
16683 (*mulsi3addsi_v6): Likewise.
16684 (*mulsidi3adddi_v6): Likewise.
16685 (*mulsidi3_nov6): Likewise.
16686 (*mulsidi3_v6): Likewise.
16687 (*umulsidi3_nov6): Likewise.
16688 (*umulsidi3_v6): Likewise.
16689 (*umulsidi3adddi): Likewise.
16690 (*umulsidi3adddi_v6): Likewise.
16691 (*smulsi3_highpart_nov6): Likewise.
16692 (*smulsi3_highpart_v6): Likewise.
16693 (*umulsi3_highpart_nov6): Likewise.
16694 (*umulsi3_highpart_v6): Likewise.
16695 (mulhisi3): Likewise.
16696 (*mulhisi3tb): Likewise.
16697 (*mulhisi3bt): Likewise.
16698 (*mulhisi3tt): Likewise.
16699 (maddhisi4): Likewise.
16700 (*maddhisi4tb): Likewise.
16701 (*maddhisi4tt): Likewise.
16702 (maddhidi4): Likewise.
16703 (*maddhidi4tb): Likewise.
16704 (*maddhidi4tt): Likewise.
16705 (divsi3): Likewise.
16706 (udivsi3): Likewise.
16707 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
16708 (thumb2_mulsi_short_compare0): Likewise.
16709 (thumb2_mulsi_short_compare0_scratch): Likewise.
16710 * config/arm/arm1020e.md (1020mult1): Update attribute change.
16711 (1020mult2): Likewise.
16712 (1020mult3): Likewise.
16713 (1020mult4): Likewise.
16714 (1020mult5): Likewise.
16715 (1020mult6): Likewise.
16716 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
16718 (cortex_a15_mult64): Likewise.
16719 (cortex_a15_sdiv): Likewise.
16720 (cortex_a15_udiv): Likewise.
16721 * config/arm/arm1026ejs.md (mult1): Update attribute change.
16727 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
16728 (pj4_ir_div): Likewise.
16729 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
16730 (11_mult2): Likewise.
16731 (11_mult3): Likewise.
16732 (11_mult4): Likewise.
16733 (11_mult5): Likewise.
16734 (11_mult6): Likewise.
16735 (11_mult7): Likewise.
16736 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
16737 (cortex_a8_mla): Likewise.
16738 (cortex_a8_mull): Likewise.
16739 (cortex_a8_smulwy): Likewise.
16740 (cortex_a8_smlald): Likewise.
16741 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
16742 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
16743 (cortex_r4_mul_3): Likewise.
16744 (cortex_r4_mla_4): Likewise.
16745 (cortex_r4_mla_3): Likewise.
16746 (cortex_r4_smlald): Likewise.
16747 (cortex_r4_mull): Likewise.
16748 (cortex_r4_sdiv): Likewise.
16749 (cortex_r4_udiv): Likewise.
16750 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
16751 (cortex_a7_idiv): Likewise.
16752 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
16753 (9_mult2): Likewise.
16754 (9_mult3): Likewise.
16755 (9_mult4): Likewise.
16756 (9_mult5): Likewise.
16757 (9_mult6): Likewise.
16758 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
16759 (cortex_a53_sdiv): Likewise.
16760 (cortex_a53_udiv): Likewise.
16761 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
16762 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
16763 (mp626_mult2): Likewise.
16764 (mp626_mult3): Likewise.
16765 (mp626_mult4): Likewise.
16766 * config/arm/fa526.md (526_mult1): Update attribute change.
16767 (526_mult2): Likewise.
16768 * config/arm/arm-generic.md (mult): Update attribute change.
16769 (mult_ldsched_strongarm): Likewise.
16770 (mult_ldsched): Likewise.
16771 (multi_cycle): Likewise.
16772 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
16773 * config/arm/fa606te.md (606te_mult1): Update attribute change.
16774 (606te_mult2): Likewise.
16775 (606te_mult3): Likewise.
16776 (606te_mult4): Likewise.
16777 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
16778 (cortex_a9_mac16): Likewise.
16779 (cortex_a9_multiply): Likewise.
16780 (cortex_a9_mac): Likewise.
16781 (cortex_a9_multiply_long): Likewise.
16782 * config/arm/fa626te.md (626te_mult1): Update attribute change.
16783 (626te_mult2): Likewise.
16784 (626te_mult3): Likewise.
16785 (626te_mult4): Likewise.
16787 2013-06-18 Richard Biener <rguenther@suse.de>
16790 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
16792 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16795 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16796 with next_active_insn.
16798 2013-06-18 Alan Modra <amodra@gmail.com>
16800 * config/rs6000/rs6000.h (enum data_align): New.
16801 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
16802 (DATA_ABI_ALIGNMENT): Define.
16803 (CONSTANT_ALIGNMENT): Correct comment.
16804 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
16805 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
16807 2013-06-17 David Malcolm <dmalcolm@redhat.com>
16809 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
16810 ATTRIBUTE_UNUSED marking.
16812 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
16814 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
16815 alternative and update.
16816 (aarch64_dup_lanedi): Delete.
16817 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
16818 * config/aarch64/aarch64-simd-builtins.def: Update.
16820 2013-06-17 Richard Biener <rguenther@suse.de>
16822 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
16823 (lto_input_scc): Declare.
16824 (lto_input_tree_1): Likewise.
16825 (struct lto_stats_d): Add num_tree_bodies_output and
16826 num_pickle_refs_output.
16827 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
16828 (lto_read_tree_1): Split out from ...
16829 (lto_read_tree): ... this.
16830 (lto_input_scc): New function.
16831 (lto_input_tree_1): Split out from ...
16832 (lto_input_tree): ... this. Handle LTO_tree_scc.
16833 (lto_data_in_create): Create the streamer cache without hashes.
16834 * lto-streamer-out.c (create_output_block): Create the streamer
16835 cache with hashes when not doing WPA.
16836 (lto_write_tree_1): Split out from ...
16837 (lto_write_tree): ... this.
16838 (get_symbol_initial_value): New function.
16839 (lto_output_tree_1): Split out from ...
16840 (lto_output_tree): ... this. Write trees as series of SCCs
16841 using a DFS walk via DFS_write_tree.
16842 (struct sccs, struct scc_entry): New types.
16843 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
16844 (DFS_write_tree_body): New function.
16845 (DFS_write_tree): Likewise.
16846 (hash_tree): Likewise.
16847 (scc_entry_compare): Likewise.
16848 (hash_scc): Likewise.
16849 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
16850 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
16851 TREE_CHAIN as regular reference.
16852 (streamer_read_integer_cst): Remove.
16853 (streamer_get_pickled_tree): Adjust.
16854 * tree-streamer-out.c (streamer_write_chain): Disable streaming
16855 of DECL_EXTERNALs in BLOCK_VARS for now.
16856 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
16858 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
16859 Add hash value argument and record that if hashes are recorded
16861 (streamer_tree_cache_insert_1): Adjust.
16862 (streamer_tree_cache_insert): Likewise.
16863 (streamer_tree_cache_insert_at): Rename to ...
16864 (streamer_tree_cache_replace_tree): ... this and adjust.
16865 (streamer_tree_cache_append): Adjust.
16866 (record_common_node): Likewise.
16867 (streamer_tree_cache_create): Add argument whether to
16868 record hash values together with trees.
16869 (streamer_tree_cache_delete): Adjust.
16870 * tree-streamer.h (struct streamer_tree_cache_d): Add
16872 (streamer_read_integer_cst): Remove.
16873 (streamer_tree_cache_insert): Adjust.
16874 (streamer_tree_cache_append): Likewise.
16875 (streamer_tree_cache_insert_at): Rename to ...
16876 (streamer_tree_cache_replace_tree): ... this and adjust.
16877 (streamer_tree_cache_create): Add argument whether to record hashes.
16878 (streamer_tree_cache_get): Rename to ...
16879 (streamer_tree_cache_get_tree): ... this.
16880 (streamer_tree_cache_get_hash): New function.
16881 * tree.c (cache_integer_cst): New function.
16882 * tree.h (cache_integer_cst): Declare.
16883 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
16884 * lto-symtab.c (lto_varpool_replace_node): Only release
16885 DECL_INITIAL of non-prevailing decls.
16886 * varpool.c (varpool_remove_initializer): Do not release
16887 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
16889 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
16891 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
16892 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
16893 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
16894 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
16895 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
16896 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
16897 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
16898 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
16899 instead of TARGET_64BIT.
16900 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
16901 Require ISA_HAS_<D>DIV.
16903 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
16905 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
16906 (mips*-*-linux*): Move default with_llsc setting to where other
16908 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
16910 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
16911 Likewise. Remove default with_tune setting. Move default float
16912 setting to its own block. Handle with_llsc in the same block as above.
16914 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
16916 PR rtl-optimization/57425
16917 PR rtl-optimization/57569
16918 * alias.c (write_dependence_p): Add new parameters mem_mode,
16919 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
16920 Changed all callers.
16921 (canon_anti_dependence): New function.
16922 * cse.c (check_dependence): Use canon_anti_dependence.
16923 * cselib.c (cselib_invalidate_mem): Likewise.
16924 * rtl.h (canon_anti_dependence): Declare.
16926 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
16928 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
16929 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
16930 ".set mips3" for 64-bit targets.
16932 2013-06-15 Dehao Chen <dehao@google.com>
16934 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
16935 * gimple-low.c (gimple_check_call_matching_types): Likewise.
16936 (gimple_check_call_args): Likewise.
16937 * value-prof.c (check_ic_target): Likewise.
16938 * ipa-inline.c (early_inliner): Likewise.
16939 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
16940 * cgraph.c (cgraph_create_edge_1): Likewise.
16941 (cgraph_make_edge_direct): Likewise.
16943 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
16946 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
16947 rs6000_output_move_128bit to handle emitting quad memory
16948 operations. Set attribute length to 8 bytes.
16950 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
16952 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
16954 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
16955 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
16956 (aarch64_<su>mlsl<mode>): Likewise.
16958 2013-06-14 Mike Stump <mikestump@comcast.net>
16960 * Makefile.in (TARGET_H): Add insn-codes.h.
16962 2013-06-14 Alan Modra <amodra@gmail.com>
16964 PR middle-end/57134
16965 PR middle-end/57586
16966 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
16967 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
16968 bitfield expansion when EXPAND_MEMORY.
16969 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
16971 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
16973 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
16974 test for clearing quad memory on 32-bit later.
16976 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
16978 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
16979 (fold_negate_expr): Likewise.
16980 (fold_real_zero_addition_p): Handle vectors.
16981 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
16983 2013-06-14 Alan Modra <amodra@gmail.com>
16985 * varasm.c (force_const_mem): Revert 2013-06-07 change.
16987 2013-06-13 Jan Hubicka <jh@suse.cz>
16989 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
16990 Local comdats are not externally visible.
16991 * symtab.c (dump_symtab_base): Dump externally visible.
16992 (verify_symtab_base): Verify back links in the symtab hash.
16994 2013-06-13 Bin Cheng <bin.cheng@arm.com>
16996 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
16997 CONVERT_EXPR as equal nodes.
16999 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17001 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
17003 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17005 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
17006 Generalize to complex and vector.
17007 * tree.c (build_all_ones_cst): New function.
17008 * tree.h (build_all_ones_cst): Declare it.
17010 2013-06-13 Alan Modra <amodra@gmail.com>
17012 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
17013 * config/rs6000/rs6000.md (signbittf2): New insn.
17014 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
17015 (abstf2_internal, cmptf_internal2): Likewise.
17016 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
17018 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
17019 Pat Haugen <pthaugen@us.ibm.com>
17020 Peter Bergner <bergner@vnet.ibm.com>
17022 * config/rs6000/rs6000.c (emit_load_locked): Add support for
17023 power8 byte, half-word, and quad-word atomic instructions.
17024 (emit_store_conditional): Likewise.
17025 (rs6000_expand_atomic_compare_and_swap): Likewise.
17026 (rs6000_expand_atomic_op): Likewise.
17028 * config/rs6000/sync.md (larx): Add new modes for power8.
17030 (AINT): New mode iterator to include TImode as well as normal
17031 integer modes on power8.
17032 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
17033 that VSX registers are not considered. Use AINT mode iterator
17034 instead of INT1 to allow inclusion of quad word atomic operations
17036 (load_locked<mode>): Likewise.
17037 (store_conditional<mode>): Likewise.
17038 (atomic_compare_and_swap<mode>): Likewise.
17039 (atomic_exchange<mode>): Likewise.
17040 (atomic_nand<mode>): Likewise.
17041 (atomic_fetch_<fetchop_name><mode>): Likewise.
17042 (atomic_nand_fetch<mode>): Likewise.
17043 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
17045 (ATOMIC): On power8, add QImode, HImode modes.
17046 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
17047 modes that promote to SImode.
17048 (load_lockedti): Convert TImode arguments to PTImode, so that we
17049 get a guaranteed even/odd register pair.
17050 (load_lockedpti): Likewise.
17051 (store_conditionalti): Likewise.
17052 (store_conditionalpti): Likewise.
17054 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
17055 atomic load/store instructions.
17058 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
17060 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
17062 (insn_count): New attribute, with most cases extracted from...
17063 (length): ...here. Redefine most cases in terms of insn_count.
17064 (single_insn): Delete.
17065 (can_delay): Use insn_count to check for single instructions.
17066 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
17067 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
17068 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
17069 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
17070 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
17071 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
17072 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
17073 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
17074 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
17075 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
17076 rather than "length".
17077 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
17078 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
17079 Use "insn_count" rather than "length".
17080 * config/mips/mips-dsp.md
17081 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
17082 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
17085 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
17087 PR tree-optimization/57361
17088 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
17090 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
17092 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
17094 (aarch64_simd_combine<mode>): New instruction expansion.
17095 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
17096 function prototype.
17097 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
17098 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
17100 2013-06-12 Jan Hubicka <jh@suse.cz>
17102 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
17103 decl has when in streaming stage.
17104 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
17105 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
17107 2013-06-12 Roland Stigge <stigge@antcom.de>
17110 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
17112 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17114 PR tree-optimization/57537
17115 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
17116 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
17118 2013-06-12 Richard Biener <rguenther@suse.de>
17120 * data-streamer.h (streamer_write_char_stream): CSE
17121 obs->current_pointer.
17122 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
17123 streamer_write_char_stream manually and optimize the resulting loop.
17124 (streamer_write_hwi_stream): Likewise.
17126 2013-06-12 Jan Hubicka <jh@suse.cz>
17128 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
17129 * cgraph.h (varpool_create_empty_node): Declare.
17130 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
17132 * symtab.c (symtab_unregister_node): Be lax about missin entries
17134 (symtab_get_node): Update comment.
17135 * varpool.c (varpool_create_empty_node): Break out from ...
17136 (varpool_node_for_decl): ... here.
17137 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
17139 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
17141 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
17142 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
17143 part. Use straight-line flow at the end.
17144 <COMPONENT_REF>: Remove superfluous else.
17145 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
17147 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17150 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
17151 target hook even for !TREE_PUBLIC decls. If no resolution info
17152 is available, return false for common and external decls.
17154 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17156 * config/rl78/constraints.md (U): New constraint.
17157 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
17160 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17163 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
17164 to allow returning address to AT_PLATFORM name.
17166 2013-06-11 Jan Hubicka <jh@suse.cz>
17168 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
17169 * cgraph.h (symtab_node_base): Add weakref flag.
17170 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
17171 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
17172 (output_weakrefs): Use weakref flag.
17173 * fold-const.c (simple_operand_p): Handle WEAK.
17174 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
17175 * ipa.c (varpool_externally_visible_p): Drop weakref.
17176 (function_and_variable_visibility): Update comment; fix weakref
17177 sanity checks; do not clear DECL_WEAK on them.
17178 * lto-cgraph.c (lto_output_node): update.
17179 (lto_output_varpool_node): Update.
17180 (input_overwrite_node): Update.
17181 (input_node): Update.
17182 (input_varpool_node): Update.
17183 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
17184 (lto_symtab_merge_symbols): Add sanity check.
17185 (lto_symtab_prevailing_decl): Do not special case weakrefs.
17186 * passes.c (rest_of_decl_compilation): Set static flag, too.
17187 * symtab.c (dump_symtab_base): Dump weakref.
17188 (verify_symtab_base): Sanity check weakrefs.
17189 (symtab_make_decl_local): Remove duplicated code.
17190 (symtab_alias_ultimate_target): Simplify.
17191 * varpool.c (varpool_create_variable_alias): Set weakref flag.
17193 2013-06-11 Tom de Vries <tom@codesourcery.com>
17195 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
17196 sequence_vect == NULL.
17198 2013-06-11 DJ Delorie <dj@redhat.com>
17200 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
17201 (rl78_unwind_word_mode): New.
17203 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17205 * final.c (debug_prefix_maps): Make static.
17207 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17209 * function.c (initial_trampoline): Remove stray copy.
17211 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
17213 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
17215 2013-06-11 Martin Jambor <mjambor@suse.cz>
17217 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
17218 within bounds at the beginning of the function.
17220 2013-06-11 Alan Modra <amodra@gmail.com>
17222 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
17224 (get_named_section): Don't NULL !DECL_P decl.
17226 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
17228 * doc/invoke.texi (core-avx2): Document.
17230 (atom): Updated with MOVBE.
17232 2013-06-11 Richard Biener <rguenther@suse.de>
17234 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
17236 2013-06-11 Anton Blanchard <anton@samba.org>
17238 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
17239 correct shift value in little-endian mode.
17241 2013-06-11 Jakub Jelinek <jakub@redhat.com>
17244 * varasm.c (get_variable_align): Move #endif to the right place.
17246 2013-06-10 Cary Coutant <ccoutant@google.com>
17248 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
17249 for hash so that hash table traversal order is deterministic.
17251 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
17252 Pat Haugen <pthaugen@us.ibm.com>
17253 Peter Bergner <bergner@vnet.ibm.com>
17255 * config/rs6000/vector.md (GPR move splitter): Do not split moves
17256 of vectors in GPRS if they are direct moves or quad word load or
17259 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
17261 (direct_move_p): Likewise.
17262 (quad_load_store_p): Likewise.
17264 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
17265 classes into bins based on the physical register type.
17266 (reg_class_to_reg_type): Likewise.
17267 (IS_STD_REG_TYPE): Likewise.
17268 (IS_FP_VECT_REG_TYPE): Likewise.
17269 (reload_fpr_gpr): Arrays to determine what insn to use if we can
17270 use direct move instructions.
17271 (reload_gpr_vsx): Likewise.
17272 (reload_vsx_gpr): Likewise.
17273 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
17274 information that is a simplification of register classes. Also
17275 precalculate direct move reload helpers.
17276 (direct_move_p): New function to return true if the operation can
17277 be done as a direct move instruciton.
17278 (quad_load_store_p): New function to return true if the operation
17279 is a quad memory operation.
17280 (rs6000_legitimize_address): If quad memory, only allow register
17281 indirect for TImode addresses.
17282 (rs6000_legitimate_address_p): Likewise.
17283 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
17284 (rs6000_reload_register_type): Likewise.
17285 (register_to_reg_type): Return register type.
17286 (rs6000_secondary_reload_simple_move): New helper function for
17287 secondary reload and secondary memory needed to identify anything
17288 that is a simple move, and does not need reloading.
17289 (rs6000_secondary_reload_direct_move): New helper function for
17290 secondary reload to identify cases that can be done with several
17291 instructions via the direct move instructions.
17292 (rs6000_secondary_reload_move): New helper function for secondary
17293 reload to identify moves between register types that can be done.
17294 (rs6000_secondary_reload): Add support for quad memory operations
17295 and for direct move.
17296 (rs6000_secondary_memory_needed): Likewise.
17297 (rs6000_debug_secondary_memory_needed): Change argument names.
17298 (rs6000_output_move_128bit): New function to return the move to
17299 use for 128-bit moves, including knowing about the various
17300 limitations of quad memory operations.
17302 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
17303 memory operations. call rs6000_output_move_128bit for the actual
17304 instruciton(s) to generate.
17305 (vsx_movti_64bit): Likewise.
17307 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
17308 (UNSPEC_P8V_MTVSRWZ): Likewise.
17309 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
17310 (UNSPEC_P8V_MTVSRD): Likewise.
17311 (UNSPEC_P8V_XXPERMDI): Likewise.
17312 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
17313 (UNSPEC_FUSION_GPR): Likewise.
17314 (FMOVE128_GPR): New iterator for direct move.
17315 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
17316 (f32_sv): Likewise.
17317 (f32_dm): Likewise.
17318 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
17319 loads and direct move instructions.
17320 (zero_extendsidi2_lfiwzx): Likewise.
17321 (extendsidi2_lfiwax): Likewise.
17322 (extendsidi2_nocell): Likewise.
17323 (floatsi<mode>2_lfiwax): Likewise.
17324 (lfiwax): Likewise.
17325 (floatunssi<mode>2_lfiwzx): Likewise.
17326 (lfiwzx): Likewise.
17327 (fix_trunc<mode>_stfiwx): Likewise.
17328 (fixuns_trunc<mode>_stfiwx): Likewise.
17329 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
17330 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
17331 (parity<mode>2_cmpb): Set length/type attr.
17332 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
17333 for 'mr.' to fast_compare.
17334 (bpermd_<mode>): Change type attr to popcnt.
17335 (p8_fmrgow_<mode>): New insns for power8 direct move support.
17336 (p8_mtvsrwz_1): Likewise.
17337 (p8_mtvsrwz_2): Likewise.
17338 (reload_fpr_from_gpr<mode>): Likewise.
17339 (p8_mtvsrd_1): Likewise.
17340 (p8_mtvsrd_2): Likewise.
17341 (p8_xxpermdi_<mode>): Likewise.
17342 (reload_vsx_from_gpr<mode>): Likewise.
17343 (reload_vsx_from_gprsf): Likewise.
17344 (p8_mfvsrd_3_<mode>): LIkewise.
17345 (reload_gpr_from_vsx<mode>): Likewise.
17346 (reload_gpr_from_vsxsf): Likewise.
17347 (p8_mfvsrd_4_disf): Likewise.
17348 (multi-word GPR splits): Do not split direct moves or quad memory
17351 2013-06-10 David Malcolm <dmalcolm@redhat.com>
17353 * tree-into-ssa.c (interesting_blocks): Make static.
17355 2013-06-10 Jakub Jelinek <jakub@redhat.com>
17358 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
17359 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
17360 Use DATA_ABI_ALIGNMENT for that case instead if defined.
17361 (get_variable_align): New function.
17362 (get_variable_section, emit_bss, emit_common,
17363 assemble_variable_contents, place_block_symbol): Use
17364 get_variable_align instead of DECL_ALIGN.
17365 (assemble_noswitch_variable): Add align argument, use it
17366 instead of DECL_ALIGN.
17367 (assemble_variable): Adjust caller. Use get_variable_align
17368 instead of DECL_ALIGN.
17369 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
17371 (DATA_ABI_ALIGNMENT): Define.
17372 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
17373 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
17374 opt is false, only return the psABI mandated alignment increase.
17375 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
17376 (DATA_ABI_ALIGNMENT): ... this.
17377 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
17378 (DATA_ABI_ALIGNMENT): ... this.
17379 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
17380 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
17381 (DATA_ABI_ALIGNMENT): ... this.
17382 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
17383 * doc/tm.texi: Regenerated.
17385 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
17387 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
17388 cmp_code to construct REG_EQUAL note.
17390 2013-06-09 Jakub Jelinek <jakub@redhat.com>
17393 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
17394 that operands[2] doesn't overlap with operands[0].
17396 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
17397 Jan Hubicka <jh@suse.cz>
17399 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
17400 hack to mark symbols as used.
17402 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
17404 PR rtl-optimization/57559
17405 * lra-constraints.c (process_alt_operands): Don't discourage
17406 memory with known offset for offsetable memory constraint.
17407 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
17409 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
17411 * varasm.c (struct oc_local_state): Reorder fields.
17412 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
17413 and adjust accordingly.
17414 (output_constructor): Reorder initialization code and adjust call to
17415 output_constructor_bitfield.
17417 2013-06-07 Jan Hubicka <jh@suse.cz>
17419 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
17421 2013-06-07 David Malcolm <dmalcolm@redhat.com>
17423 * tree-object-size.c (unknown): Make const.
17425 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17427 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
17428 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
17429 for last alternative in the cpu_facility attribute.
17431 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17434 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
17435 (xordi3): Change operand 2 constraint to arm_xordi_operand.
17436 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
17437 * config/arm/constraints.md (Dg): New constraint.
17438 * config/arm/neon.md (xordi3_neon): Remove.
17439 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
17440 * config/arm/predicates.md (arm_xordi_operand): New predicate.
17442 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17444 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
17445 Clean up alternatives.
17447 2013-06-07 Alan Modra <amodra@gmail.com>
17449 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
17452 2013-06-07 Alan Modra <amodra@gmail.com>
17454 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
17456 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17458 * config/arm/constraints.md (Df): New constraint.
17459 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
17460 Correct length attribute for last two alternatives.
17462 2013-06-07 Alan Modra <amodra@gmail.com>
17464 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
17465 override user -mfp-in-toc.
17466 (offsettable_ok_by_alignment): Consider just the current access
17467 rather than the whole object, unless BLKmode. Handle
17468 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
17469 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
17470 for -mcmodel=medium.
17471 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
17472 override user -mfp-in-toc or -msum-in-toc. Default to
17473 -mno-fp-in-toc for -mcmodel=medium.
17475 2013-06-06 DJ Delorie <dj@redhat.com>
17477 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
17478 TARGET_VALID_POINTER_MODE.
17480 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
17481 Pat Haugen <pthaugen@us.ibm.com>
17482 Peter Bergner <bergner@vnet.ibm.com>
17484 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
17485 Document new power8 builtins.
17487 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
17488 condition code register, to allow 128-bit logical operations to be
17489 done in the VSX or GPR registers.
17490 (nor<mode>3): Use the canonical form for nor.
17491 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
17492 vclz*, and vpopcnt* vector instructions.
17493 (nand<mode>3): Likewise.
17494 (orc<mode>3): Likewise.
17495 (clz<mode>2): LIkewise.
17496 (popcount<mode>2): Likewise.
17498 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
17499 that only the GPRs are recognized.
17501 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
17502 support for new power8 builtins.
17504 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
17506 (xscvdpspn): Likewise.
17512 (vpopcnt): Likewise.
17513 (vpopcntb): Likewise.
17514 (vpopcnth): Likewise.
17515 (vpopcntw): Likewise.
17516 (vpopcntd): Likewise.
17518 (vmrgew): Likewise.
17519 (vmrgow): Likewise.
17521 (eqv_v16qi3): Likewise.
17522 (eqv_v8hi3): Likewise.
17523 (eqv_v4si3): Likewise.
17524 (eqv_v2di3): Likewise.
17525 (eqv_v4sf3): Likewise.
17526 (eqv_v2df3): Likewise.
17528 (nand_v16qi3): Likewise.
17529 (nand_v8hi3): Likewise.
17530 (nand_v4si3): Likewise.
17531 (nand_v2di3): Likewise.
17532 (nand_v4sf3): Likewise.
17533 (nand_v2df3): Likewise.
17535 (orc_v16qi3): Likewise.
17536 (orc_v8hi3): Likewise.
17537 (orc_v4si3): Likewise.
17538 (orc_v2di3): Likewise.
17539 (orc_v4sf3): Likewise.
17540 (orc_v2df3): Likewise.
17542 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
17543 allow power8 quad mode in 64-bit.
17544 (rs6000_builtin_vectorized_function): Add support to vectorize
17545 ISA 2.07 count leading zeros, population count builtins.
17546 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
17547 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
17548 (builtin_function_type): Add vgbbd builtin function which takes an
17550 (altivec_expand_vec_perm_const): Add support for new power8 merge
17553 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
17554 that does not include TImdoe for use with 32-bit.
17555 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
17557 (UNSPEC_VSX_CVDPSPN): Likewise.
17558 (vsx_xscvdpspn): Likewise.
17559 (vsx_xscvspdpn): Likewise.
17560 (vsx_xscvdpspn_scalar): Likewise.
17561 (vsx_xscvspdpn_directmove): Likewise.
17562 (vsx_and<mode>3): Split logical operations into 32-bit and
17563 64-bit. Add support to do logical operations on TImode as well as
17564 VSX vector types. Allow logical operations to be done in either
17565 VSX registers or in general purpose registers in 64-bit mode. Add
17566 splitters if GPRs were used. For AND, add clobber of CCmode to
17567 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
17569 (vsx_and<mode>3_32bit): Likewise.
17570 (vsx_and<mode>3_64bit): Likewise.
17571 (vsx_ior<mode>3): Likewise.
17572 (vsx_ior<mode>3_32bit): Likewise.
17573 (vsx_ior<mode>3_64bit): Likewise.
17574 (vsx_xor<mode>3): Likewise.
17575 (vsx_xor<mode>3_32bit): Likewise.
17576 (vsx_xor<mode>3_64bit): Likewise.
17577 (vsx_one_cmpl<mode>2): Likewise.
17578 (vsx_one_cmpl<mode>2_32bit): Likewise.
17579 (vsx_one_cmpl<mode>2_64bit): Likewise.
17580 (vsx_nor<mode>3): Likewise.
17581 (vsx_nor<mode>3_32bit): Likewise.
17582 (vsx_nor<mode>3_64bit): Likewise.
17583 (vsx_andc<mode>3): Likewise.
17584 (vsx_andc<mode>3_32bit): Likewise.
17585 (vsx_andc<mode>3_64bit): Likewise.
17586 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
17587 and xxlorc instructions.
17588 (vsx_eqv<mode>3_64bit): Likewise.
17589 (vsx_nand<mode>3_32bit): Likewise.
17590 (vsx_nand<mode>3_64bit): Likewise.
17591 (vsx_orc<mode>3_32bit): Likewise.
17592 (vsx_orc<mode>3_64bit): Likewise.
17594 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
17596 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
17598 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
17599 (p8_vmrgow): Likewise.
17600 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
17601 GPRs to be split under VSX.
17602 (p8v_clz<mode>2): Add power8 count leading zero support.
17603 (p8v_popcount<mode>2): Add power8 population count support.
17604 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
17607 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
17610 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
17612 (vec_nand): Likewise.
17613 (vec_vclz): Likewise.
17614 (vec_vclzb): Likewise.
17615 (vec_vclzd): Likewise.
17616 (vec_vclzh): Likewise.
17617 (vec_vclzw): Likewise.
17618 (vec_vgbbd): Likewise.
17619 (vec_vmrgew): Likewise.
17620 (vec_vmrgow): Likewise.
17621 (vec_vpopcnt): Likewise.
17622 (vec_vpopcntb): Likewise.
17623 (vec_vpopcntd): Likewise.
17624 (vec_vpopcnth): Likewise.
17625 (vec_vpopcntw): Likewise.
17627 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17629 PR rtl-optimization/57468
17630 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
17633 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17635 PR rtl-optimization/57459
17636 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
17637 type when setting live regs.
17639 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17641 * config/s390/s390.opt (mlra): New option.
17642 * config/s390/s390.c (s390_decompose_address): Check displacement
17643 for all registers for LRA.
17644 (s390_secondary_reload): Don't used secondary reloads for LRA.
17645 (s390_lra_p): New function.
17646 (TARGET_LRA_P): Define.
17647 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
17648 of attribute cpu_facility to zarch for the last alternative.
17649 (*cmpmem_short): Ditto.
17651 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
17653 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
17654 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
17655 (arm_expand_prologue): Likewise.
17657 2013-06-06 Teresa Johnson <tejohnson@google.com>
17660 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
17661 as this is now done by redirect_edge_and_branch_force.
17662 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
17663 barriers, and fix interaction with splitting.
17664 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
17665 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
17666 reflect changes made in the routine.
17667 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
17668 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
17669 since this is called in cfglayout mode, and replace partition fixup
17670 with assert as that is now done by force_nonfallthru_and_redirect.
17671 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
17672 already be marked with region crossing note.
17673 (insert_section_boundary_note): Make non-static, gate on flag
17674 has_bb_partition, rewrite to also check for multiple partitions.
17675 (rest_of_handle_reorder_blocks): Remove call to
17676 insert_section_boundary_note, now done later during free_cfg.
17677 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
17678 * bb-reorder.h (insert_section_boundary_note): Declare.
17679 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
17680 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
17681 invoke insert_section_boundary_note.
17682 (try_redirect_by_replacing_jump): Remove unnecessary
17683 check for region crossing note.
17684 (fixup_partition_crossing): New function.
17685 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
17686 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
17687 in non-cfglayout mode.
17688 (force_nonfallthru_and_redirect): Fixup partition boundaries,
17689 remove old code that tried to do this. Emit barrier correctly
17690 when we are in cfglayout mode.
17691 (last_bb_in_partition): New function.
17692 (rtl_split_edge): Correctly fixup partition boundaries.
17693 (commit_one_edge_insertion): Remove old code that tried to
17694 fixup region crossing edge since this is now handled in
17695 split_block, and set up insertion point correctly since
17696 block may now end in a jump.
17697 (verify_hot_cold_block_grouping): Guard against checking when not in
17698 linearized RTL mode.
17699 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
17701 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
17702 rtl_verify_flow_info, so not called in cfglayout mode.
17703 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
17704 (fixup_reorder_chain): Remove old code that attempted to fixup region
17705 crossing note as this is now handled in force_nonfallthru_and_redirect.
17706 (duplicate_insn_chain): Don't duplicate switch section notes.
17707 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
17709 * basic-block.h (emit_barrier_after_bb): Declare.
17711 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17713 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
17714 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
17715 arm_usatsihi): Adjust alternatives for arm_restrict_it.
17717 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17719 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
17721 * config/arm/ldmstm.md: Regenerate.
17723 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17725 * config/arm/sync.md (atomic_loaddi_1):
17726 Disable predication for arm_restrict_it.
17727 (arm_load_exclusive<mode>): Likewise.
17728 (arm_load_exclusivesi): Likewise.
17729 (arm_load_exclusivedi): Likewise.
17730 (arm_load_acquire_exclusive<mode>): Likewise.
17731 (arm_load_acquire_exclusivesi): Likewise.
17732 (arm_load_acquire_exclusivedi): Likewise.
17733 (arm_store_exclusive<mode>): Likewise.
17734 (arm_store_exclusive<mode>): Likewise.
17735 (arm_store_release_exclusivedi): Likewise.
17736 (arm_store_release_exclusive<mode>): Likewise.
17738 2013-06-06 Richard Biener <rguenther@suse.de>
17740 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
17742 (lto_tag_is_tree_code_p): Adjust.
17743 (lto_tag_is_gimple_code_p): Likewise.
17744 (lto_gimple_code_to_tag): Likewise.
17745 (lto_tag_to_gimple_code): Likewise.
17746 (lto_tree_code_to_tag): Likewise.
17747 (lto_tag_to_tree_code): Likewise.
17748 * data-streamer.h (streamer_write_hwi_in_range): Use
17749 uhwi streaming to stream the normalized range.
17750 (streamer_read_hwi_in_range): Likewise.
17752 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17754 * config/arm/arm.md (enabled_for_depr_it): New attribute.
17755 (predicable_short_it): Likewise.
17756 (predicated): Likewise.
17757 (enabled): Handle above.
17758 (define_cond_exec): Set predicated attribute to yes.
17760 2013-06-05 Mike Stump <mikestump@comcast.net>
17762 * gdbinit.in (__FUNCTION__): Add.
17764 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
17766 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
17767 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
17769 2013-06-05 Jan Hubicka <jh@suse.cz>
17771 * varasm.c (mark_decl_referenced): Revert the removal until targets
17774 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
17776 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
17777 instead of mark_decl_referenced.
17779 2013-06-05 Jan Hubicka <jh@suse.cz>
17781 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
17782 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
17783 and symtab_used_from_object_file_p.
17784 (cgraph_make_node_local_1): Clear forced_by_abi.
17785 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
17786 * cgraph.h (symtab_node_base): Add forced_by_abi.
17787 (decide_is_variable_needed): Remove.
17788 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
17789 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
17790 (decide_is_symbol_needed): ... this one; handle symbols in general;
17791 always analyze virtuals; honnor forced_by_abi.
17792 (cgraph_finalize_function): Update.
17793 (varpool_finalize_decl): Update.
17794 (symbol_defined_and_needed): Remove.
17795 (analyze_functions): Update.
17796 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17797 output_refs, input_overwrite_node): Handle forced_by_abi.
17798 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
17799 (address_taken_from_non_vtable_p): ... this one.
17800 (comdat_can_be_unshared_p_1): New function.
17801 (cgraph_comdat_can_be_unshared_p): Rename to ...
17802 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
17803 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
17804 (function_and_variable_visibility): Clear forced_by_abi as needed.
17805 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
17806 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
17807 * symtab.c (dump_symtab_base): Dump forced_by_abi.
17808 * varpool.c (decide_is_variable_needed): Remove.
17810 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17812 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
17813 (arm_option_override): Override arm_restrict_it where appropriate.
17814 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
17815 * config/arm/arm.opt (mrestrict-it): New command-line option.
17816 * doc/invoke.texi: Document -mrestrict-it.
17818 2013-06-05 David Malcolm <dmalcolm@redhat.com>
17820 * tsan.c (tsan_atomic_table): Make const.
17822 2013-06-05 Richard Biener <rguenther@suse.de>
17824 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
17825 index associated with the tree we are supposed to replace.
17826 * tree-streamer-out.c (pack_ts_base_value_fields): Output
17827 TREE_ASM_WRITTEN as zero for everything but SSA names.
17829 2013-06-05 David Malcolm <dmalcolm@redhat.com>
17831 * tree-ssa-structalias.c (call_stmt_vars): Make static.
17833 2013-06-04 Jan Hubicka <jh@suse.cz>
17835 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
17836 (input_node, input_varpool_node): Handle correctly external same
17838 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
17839 nodes at ltrans stage.
17841 2013-06-04 Jan Hubicka <jh@suse.cz>
17843 * ipa-inline.c (update_caller_keys): Fix availability test.
17844 (update_callee_keys): Likewise.
17845 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
17846 to follow ELF standard.
17848 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
17850 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
17851 (mips64r5900el-*-elf*): New configurations.
17852 * config/mips/mips-cpus.def (r5900): New processor.
17853 * config/mips/mips-tables.opt: Regenerate.
17854 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
17855 (mips_issue_rate): Handle PROCESSOR_R5900.
17856 (mips_reorg_process_insns): Force reorder mode for the R5900.
17857 * config/mips/mips.h (TARGET_MIPS5900): Define.
17858 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
17860 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
17862 * config/mips/mips.md (processor): Add r5900.
17863 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
17865 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17867 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
17868 into function to generate MOVI instruction.
17869 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
17870 (aarch64_preferred_simd_mode): Turn into wrapper.
17871 (aarch64_output_scalar_simd_mov_immediate): New function.
17872 * config/aarch64/aarch64-protos.h: Add prototype for above.
17874 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17876 * config/aarch64/aarch64.c (simd_immediate_info): Remove
17877 element_char member.
17878 (sizetochar): Return signed char.
17879 (aarch64_simd_valid_immediate): Remove elchar and other
17880 unnecessary variables.
17881 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
17882 Calculate element_char as required.
17883 * config/aarch64/aarch64-protos.h: Update and move prototype
17884 for aarch64_output_simd_mov_immediate.
17885 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
17888 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17890 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
17891 information completed by aarch64_simd_valid_immediate.
17892 (aarch64_legitimate_constant_p): Update arguments.
17893 (aarch64_simd_valid_immediate): Work with struct rather than many
17895 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
17896 (aarch64_simd_make_constant): Update arguments.
17897 (aarch64_output_simd_mov_immediate): Work with struct rather than
17898 many pointers. Output immediate directly rather than as operand.
17899 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
17901 * config/aarch64/constraints.md (Dn): Update arguments.
17903 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17905 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
17907 (aarch64_simd_immediate_valid_for_move): Remove.
17908 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
17909 (aarch64_simd_make_constant): Update call.
17910 (aarch64_output_simd_mov_immediate): Update call.
17911 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
17913 * config/aarch64/constraints.md (Dn): Update call.
17915 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17917 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
17918 return type to bool for prototype.
17919 (aarch64_legitimate_constant_p): Check for true instead of not -1.
17920 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
17921 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
17923 2013-06-04 Catherine Moore <clm@codesourcery.com>
17925 * config/mips/mips.opt (meva): New.
17926 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
17927 (ASM_SPEC): Handle -meva.
17928 * doc/invoke.texi (meva): Document.
17930 2013-06-04 Alan Modra <amodra@gmail.com>
17932 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
17935 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17937 * rtl.def: Add extra fourth optional field to define_cond_exec.
17938 * gensupport.c (process_one_cond_exec): Process attributes from
17940 * doc/md.texi: Document fourth field in define_cond_exec.
17942 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
17944 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
17945 out the processing order as in store_bit_field_1.
17947 2013-06-04 Jan Hubicka <jh@suse.cz>
17949 PR middle-end/57500
17950 * cgraphunit.c (cgraph_process_same_body_aliases): Create
17951 non-VAR_DECL node if it does not exist yet.
17953 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
17955 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
17956 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
17957 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
17958 target_cpu_default setting.
17960 2013-06-03 Teresa Johnson <tejohnson@google.com>
17962 * dumpfile.c (opt_info_switch_p): Change -fopt-info
17963 default to -fopt-info=optimized instead of all.
17964 * doc/invoke.texi: Ditto.
17965 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
17966 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
17967 (execute_vect_slp): Emit BB vectorization success under
17968 MSG_OPTIMIZED_LOCATIONS.
17969 * tree-vect-slp.c (vect_slp_transform_bb): Change
17970 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
17971 * tree-vect-loop.c (vect_transform_loop): Ditto.
17973 2013-06-03 Jason Merrill <jason@redhat.com>
17976 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
17977 Use TARGET_EXPR for C++.
17979 2013-06-03 Jakub Jelinek <jakub@redhat.com>
17981 PR rtl-optimization/57268
17982 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
17983 if DEBUG_INSN_P (insn).
17986 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
17988 PR rtl-optimization/57268
17989 * sched-deps.c (sched_analyze_2): Flush dependence lists if
17990 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
17992 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
17994 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
17995 (ix86_avoid_lea_for_addr): Likewise.
17996 (exact_dependency_1): Likewise.
17997 (ix86_adjust_cost): Likewise.
17998 (swap_top_of_ready_list): Fix formatting and !reload_completed check
18000 (do_reorder_for_imul): Fix typo, formatting and
18001 !reload_completed check removed.
18002 (ix86_sched_reorder): Fix typo and formatting.
18003 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
18006 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
18008 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
18010 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
18012 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
18013 <STRING_CST>: Likewise.
18014 <VECTOR_CST>: Likewise.
18016 2013-06-01 Janus Weil <janus@gcc.gnu.org>
18017 Mikael Morin <mikael@gcc.gnu.org>
18019 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
18020 * config.in: Regenerated.
18021 * configure: Regenerated.
18023 2013-06-01 Jan Hubicka <jh@suse.cz>
18025 PR middle-end/57366
18026 * cgraphunit.c (compile): When weakref is not supported,
18027 set up transparent aliases before final output pass.
18028 * varasm.c (assemble_alias): Do not try to do it here.
18030 2013-06-01 Jan Hubicka <jh@suse.cz>
18032 PR middle-end/57467
18033 * passes.c (for_per_function): Skip unanalyzed functions.
18035 2013-06-01 Jan Hubicka <jh@suse.cz>
18037 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
18038 (lto_symtab_merge_symbols_1): ... this one.
18039 (lto_symtab_merge_cgraph_nodes): Rename to ...
18040 (lto_symtab_merge_symbols): ... this one; simplify.
18041 * cgraph.c (same_body_aliases_done): Rename to ...
18042 (cpp_implicit_aliases_done): ... this one.
18043 (cgraph_create_function_alias): Update.
18044 (cgraph_same_body_alias): Update.
18045 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
18046 (verify_edge_corresponds_to_fndecl): Simplify.
18047 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
18048 (cgraph_node): Remove same_body_alias.
18049 (varpool_node): Remove alias_of and extra_name_alias.
18050 (same_body_aliases_done): Rename to ..
18051 (cpp_implicit_aliases_done): ... this one.
18052 (symtab_alias_ultimate_target): Add default parameter.
18053 (symtab_resolve_alias): New function.
18054 (fixup_same_cpp_alias_visibility): Declare.
18055 (cgraph_function_node): Add default parameter.
18056 (cgraph_node_asm_name): Likewise.
18057 (cgraph_function_or_thunk_node): Add default parameter; do
18058 not ICE when it is NULL.
18059 (varpool_variable_node): Likewise.
18060 * tree-emutls.c (create_emultls_var): Update.
18061 (ipa_lower_emutls): Update.
18062 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
18063 (cgraph_reset_node): Reset alias info.
18064 (cgraph_finalize_function): Update.
18065 (fixup_same_cpp_alias_visibility): Move to symtab.c.
18066 (analyze_function): Simplify.
18067 (cgraph_process_same_body_aliases): Simplify.
18068 (analyze_functions): Fixup same body aliases.
18069 (handle_alias_pairs): Simplify.
18070 (assemble_thunk): Update.
18071 (assemble_thunks_and_aliases): Update.
18072 (output_weakrefs): Rewrite.
18073 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
18074 (lto_output_varpool_node): Likewise.
18075 (compute_ltrans_boundary): Remve assert.
18076 (get_alias_symbol): New functoin.
18077 (input_node): Rewrite alias handling.
18078 (input_varpool_node): Likewise.
18079 * ipa-pure-const.c (propagate_pure_const): Fix formating.
18080 * ipa.c (process_references): Handle weakrefs correctly.
18081 (symtab_remove_unreachable_nodes): Likewise.
18082 * trans-mem.c (get_cg_data): Update.
18083 (ipa_tm_create_version_alias): Update.
18084 (ipa_tm_execute): Update.
18085 * symtab.c (dump_symtab_base): Dump aliases.
18086 (verify_symtab_base): Verify aliases.
18087 (symtab_node_availability): New function.
18088 (symtab_alias_ultimate_target): Simplify.
18089 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
18090 handle all the fixup cases.
18091 (symtab_resolve_alias): New function.
18092 * passes.c (ipa_write_summaries): Handle weakrefs.
18093 * varpool.c (varpool_analyze_node): Simplify.
18094 (assemble_aliases): Update.
18095 (varpool_create_variable_alias): Simplify.
18096 (varpool_extra_name_alias): Simplify.
18097 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
18098 (lto_symtab_merge_symbols): ... this one.
18100 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
18103 PR rtl-optimization/57268
18104 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18105 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18107 2013-06-01 Tobias Burnus <burnus@net-b.de>
18109 Partially reverted:
18110 2013-05-31 Tobias Burnus <burnus@net-b.de>
18112 PR middle-end/57073
18113 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
18116 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18118 PR rtl-optimization/57268
18119 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18120 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18122 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18124 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
18125 unordered comparison operators when -fno-trapping-math is in effect
18127 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
18128 and implement unordered comparison operators properly on the e500.
18130 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18132 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
18133 for constant scalar integers.
18134 (simplify_relational_operation_1): Likewise.
18136 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
18138 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
18139 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
18142 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18143 Igor Zamyatin <igor.zamyatin@intel.com>
18145 Silvermont (SLM) architecture performance tuning.
18146 * config/i386/i386.h (enum ix86_tune_indices): Add
18147 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
18148 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
18150 * config/i386/i386.c (initial_ix86_tune_features)
18151 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
18152 (ix86_lea_outperforms): Handle Silvermont tuning.
18153 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
18155 (ix86_use_lea_for_mov): Likewise.
18156 (ix86_avoid_lea_for_addr): Likewise.
18157 (ix86_lea_for_add_ok): Likewise.
18158 (exact_dependency_1): New function.
18159 (exact_store_load_dependency): Likewise.
18160 (ix86_adjust_cost): Handle Silvermont tuning.
18161 (do_reoder_for_imul): Likewise.
18162 (swap_top_of_ready_list): New function.
18163 (ix86_sched_reorder): Changed to handle Silvermont tuning.
18165 * config/i386/i386.md (peepholes that split memory operand in fp
18168 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18170 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
18171 Remove un-necessary braces.
18173 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18175 * config/aarch64/aarch64.c (aarch64_classify_symbol):
18176 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
18178 2013-05-31 Tobias Burnus <burnus@net-b.de>
18180 PR middle-end/57073
18181 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
18183 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18186 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
18187 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
18188 * config/arm/neon.md (iordi3_neon): Remove.
18189 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
18190 * config/arm/predicates.md (imm_for_neon_logic_operand):
18191 Move to earlier in the file.
18192 (neon_logic_op2): Likewise.
18193 (arm_iordi_operand_neon): New predicate.
18195 2013-05-31 Richard Biener <rguenther@suse.de>
18197 PR tree-optimization/57478
18198 PR tree-optimization/57453
18199 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
18202 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18204 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
18205 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
18207 2013-05-30 Tobias Burnus <burnus@net-b.de>
18208 Thomas Koenig <tkoenig@gcc.gnu.org>
18210 PR middle-end/57073
18211 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
18212 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
18214 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
18216 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
18218 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
18220 * target.def (register_usage_leveling_p): New hook.
18221 * targhooks.c (default_register_usage_leveling_p): New.
18222 * targhooks.h (default_register_usage_leveling_p): New prototype.
18223 * lra-assigns.c (register_usage_leveling_p): Use the hook.
18224 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
18225 * doc/tm.texi: Update.
18226 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
18228 2013-05-30 Ian Bolton <ian.bolton@arm.com>
18230 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
18231 (*insv_reg<mode>): New define_insn.
18233 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
18235 PR rtl-optimization/57439
18236 * postreload.c (move2add_valid_value_p): Check that we have
18237 a zero subreg_regno_offset when accessing the register in
18238 the requested mode.
18240 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18241 Igor Zamyatin <igor.zamyatin@intel.com>
18243 Silvermont (SLM) architecture pipeline model, tuning and
18245 * config.gcc: Add slm config options and target.
18247 * config/i386/slm.md: New.
18249 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
18251 * config/i386/i386-c.c (ix86_target_macros_internal): New case
18253 (ix86_target_macros_internal): Likewise.
18255 * config/i386/i386.c (slm_cost): New cost.
18256 (m_SLM): New macro flag.
18257 (initial_ix86_tune_features): Set m_SLM.
18258 (x86_accumulate_outgoing_args): Likewise.
18259 (x86_arch_always_fancy_math_387): Likewise.
18260 (processor_target_table): Add slm cost.
18261 (cpu_names): Add slm cpu name.
18262 (x86_option_override_internal): Set SLM ISA.
18263 (ix86_issue_rate): New case PROCESSOR_SLM.
18264 (ia32_multipass_dfa_lookahead): Likewise.
18265 (fold_builtin_cpu): Add slm.
18267 * config/i386/i386.h (TARGET_SLM): New target macro.
18268 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
18269 (processor_type): Add PROCESSOR_SLM.
18271 * config/i386/i386.md (cpu): Add new value "slm".
18272 (slm.md): Include slm.md.
18274 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
18275 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18277 * config/arm/arm-protos.h: Add and update function protos.
18278 * config/arm/arm.c (use_simple_return_p): New added.
18279 (thumb2_expand_return): Check simple_return flag.
18280 * config/arm/arm.md: Add simple_return and conditional simple_return.
18281 * config/arm/iterators.md: Add iterator for return and simple_return.
18283 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18285 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
18286 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
18287 (arm_emit_vfp_multi_reg_pop): Likewise.
18288 (thumb2_emit_ldrd_pop): Likewise.
18289 (arm_expand_epilogue): Add misc REG_CFA notes.
18290 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
18292 2013-05-29 Lawrence Crowl <crowl@google.com>
18294 * config/arm/t-arm: Update for below.
18296 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
18297 Change type to hash_table. Update dependent calls and types.
18299 * config/i386/t-cygming: Update for below.
18301 * config/i386/t-interix: Update for below.
18303 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
18304 Change type to hash_table. Update dependent calls and types.
18305 (i386_find_on_wrapper_list::wrappers): Likewise.
18307 * config/ia64/t-ia64: Update for below.
18309 * config/ia64/ia64.c (bundle_state_table):
18310 Change type to hash_table. Update dependent calls and types.
18312 * config/mips/mips.c (mips_reorg_process_insns::htab):
18313 Change type to hash_table. Update dependent calls and types.
18315 * config/sol2.c (solaris_comdat_htab):
18316 Change type to hash_table. Update dependent calls and types.
18318 * config/t-sol2: Update for above.
18320 2013-05-29 Teresa Johnson <tejohnson@google.com>
18322 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
18323 functions are not yet marked as defined.
18325 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
18326 Pat Haugen <pthaugen@us.ibm.com>
18327 Peter Bergner <bergner@vnet.ibm.com>
18329 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
18333 (vrotl<mode>3): Likewise.
18334 (vashl<mode>3): Likewise.
18335 (vlshr<mode>3): Likewise.
18336 (vashr<mode>3): Likewise.
18338 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18339 support for power8 V2DI builtins.
18341 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
18342 power8 V2DI builtins.
18343 (vupkhsw): Likewise.
18344 (vupklsw): Likewise.
18345 (vaddudm): Likewise.
18346 (vminsd): Likewise.
18347 (vmaxsd): Likewise.
18348 (vminud): Likewise.
18349 (vmaxud): Likewise.
18350 (vpkudum): Likewise.
18351 (vpksdss): Likewise.
18352 (vpkudus): Likewise.
18353 (vpksdus): Likewise.
18358 (vsubudm): Likewise.
18359 (vcmpequd): Likewise.
18360 (vcmpgtsd): Likewise.
18361 (vcmpgtud): Likewise.
18362 (vcmpequd_p): Likewise.
18363 (vcmpgtsd_p): Likewise.
18364 (vcmpgtud_p): Likewise.
18365 (vupkhsw): Likewise.
18366 (vupklsw): Likewise.
18367 (vaddudm): Likewise.
18368 (vmaxsd): Likewise.
18369 (vmaxud): Likewise.
18370 (vminsd): Likewise.
18371 (vminud): Likewise.
18372 (vpksdss): Likewise.
18373 (vpksdus): Likewise.
18374 (vpkudum): Likewise.
18375 (vpkudus): Likewise.
18380 (vsubudm): Likewise.
18382 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
18383 support for power8 V2DI instructions.
18385 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
18386 power8 V2DI instructions. Combine pack and unpack insns to use an
18387 iterator for each mode. Check whether a particular mode supports
18388 Altivec instructions instead of just checking TARGET_ALTIVEC.
18389 (UNSPEC_VPKUWUM): Likewise.
18390 (UNSPEC_VPKSHSS): Likewise.
18391 (UNSPEC_VPKSWSS): Likewise.
18392 (UNSPEC_VPKUHUS): Likewise.
18393 (UNSPEC_VPKSHUS): Likewise.
18394 (UNSPEC_VPKUWUS): Likewise.
18395 (UNSPEC_VPKSWUS): Likewise.
18396 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
18397 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
18398 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
18399 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
18400 (UNSPEC_VUPKHSB): Likewise.
18401 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
18402 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
18403 (UNSPEC_VUPKHSH): Likewise.
18404 (UNSPEC_VUPKLSB): Likewise.
18405 (UNSPEC_VUPKLSH): Likewise.
18407 (VI_char): Likewise.
18408 (VI_scalar): Likewise.
18409 (VI_unit): Likewise.
18411 (VP_small): Likewise.
18412 (VP_small_lc): Likewise.
18413 (VU_char): Likewise.
18414 (add<mode>3): Likewise.
18415 (altivec_vaddcuw): Likewise.
18416 (altivec_vaddu<VI_char>s): Likewise.
18417 (altivec_vadds<VI_char>s): Likewise.
18418 (sub<mode>3): Likewise.
18419 (altivec_vsubcuw): Likewise.
18420 (altivec_vsubu<VI_char>s): Likewise.
18421 (altivec_vsubs<VI_char>s): Likewise.
18422 (altivec_vavgs<VI_char>): Likewise.
18423 (altivec_vcmpbfp): Likewise.
18424 (altivec_eq<mode>): Likewise.
18425 (altivec_gt<mode>): Likewise.
18426 (altivec_gtu<mode>): Likewise.
18427 (umax<mode>3): Likewise.
18428 (smax<mode>3): Likewise.
18429 (umin<mode>3): Likewise.
18430 (smin<mode>3): Likewise.
18431 (altivec_vpkuhum): Likewise.
18432 (altivec_vpkuwum): Likewise.
18433 (altivec_vpkshss): Likewise.
18434 (altivec_vpkswss): Likewise.
18435 (altivec_vpkuhus): Likewise.
18436 (altivec_vpkshus): Likewise.
18437 (altivec_vpkuwus): Likewise.
18438 (altivec_vpkswus): Likewise.
18439 (altivec_vpks<VI_char>ss): Likewise.
18440 (altivec_vpks<VI_char>us): Likewise.
18441 (altivec_vpku<VI_char>us): Likewise.
18442 (altivec_vpku<VI_char>um): Likewise.
18443 (altivec_vrl<VI_char>): Likewise.
18444 (altivec_vsl<VI_char>): Likewise.
18445 (altivec_vsr<VI_char>): Likewise.
18446 (altivec_vsra<VI_char>): Likewise.
18447 (altivec_vsldoi_<mode>): Likewise.
18448 (altivec_vupkhsb): Likewise.
18449 (altivec_vupkhs<VU_char>): Likewise.
18450 (altivec_vupkls<VU_char>): Likewise.
18451 (altivec_vupkhsh): Likewise.
18452 (altivec_vupklsb): Likewise.
18453 (altivec_vupklsh): Likewise.
18454 (altivec_vcmpequ<VI_char>_p): Likewise.
18455 (altivec_vcmpgts<VI_char>_p): Likewise.
18456 (altivec_vcmpgtu<VI_char>_p): Likewise.
18457 (abs<mode>2): Likewise.
18458 (vec_unpacks_hi_v16qi): Likewise.
18459 (vec_unpacks_hi_v8hi): Likewise.
18460 (vec_unpacks_lo_v16qi): Likewise.
18461 (vec_unpacks_hi_<VP_small_lc>): Likewise.
18462 (vec_unpacks_lo_v8hi): Likewise.
18463 (vec_unpacks_lo_<VP_small_lc>): Likewise.
18464 (vec_pack_trunc_v8h): Likewise.
18465 (vec_pack_trunc_v4si): Likewise.
18466 (vec_pack_trunc_<mode>): Likewise.
18468 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
18470 (vec_vmaxsd): Likewise.
18471 (vec_vmaxud): Likewise.
18472 (vec_vminsd): Likewise.
18473 (vec_vminud): Likewise.
18474 (vec_vpksdss): Likewise.
18475 (vec_vpksdus): Likewise.
18476 (vec_vpkudum): Likewise.
18477 (vec_vpkudus): Likewise.
18478 (vec_vrld): Likewise.
18479 (vec_vsld): Likewise.
18480 (vec_vsrad): Likewise.
18481 (vec_vsrd): Likewise.
18482 (vec_vsubudm): Likewise.
18483 (vec_vupkhsw): Likewise.
18484 (vec_vupklsw): Likewise.
18486 2013-05-29 Jan Hubicka <jh@suse.cz>
18488 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
18489 flags; reorder rest of fields in more consistent way.
18490 (varpool_node): Remove analyzed, finalized and alias.
18491 (cgraph_ndoe): Likewise.
18492 (symtab_alias_ultimate_target): New function.
18493 (cgraph_function_node): Move offline.
18494 (cgraph_reset_node): Declare.
18495 (cgraph_comdat_can_be_unshared_p): Remove.
18496 (varpool_remove_initializer): Declare.
18497 (varpool_first_defined_variable, varpool_next_defined_variable
18498 cgraph_first_defined_function, cgraph_next_defined_function): Update.
18499 (cgraph_function_with_gimple_body_p): Update.
18500 (varpool_all_refs_explicit_p): Update.
18501 (symtab_alias_target): New function.
18502 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
18503 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
18504 (cgraph_function_or_thunk_node): Simplify using
18505 symtab_alias_ultimate_target.
18506 (varpool_variable_node): Likewise.
18507 * cgraph.c (cgraph_create_function_alias): Update.
18508 (cgraph_add_thunk): Update.
18509 (cgraph_remove_node): Update.
18510 (dump_cgraph_node): Do not dump removed flags.
18511 (cgraph_function_body_availability): Update.
18512 (cgraph_propagate_frequency): Update.
18513 (verify_cgraph_node): Check sanity of local flag.
18514 (cgraph_function_node): Move here from cgraph.h; revamp for
18515 cgraph_function_or_thunk_node.
18516 * lto-symtab.c (lto_varpool_replace_node): Update.
18517 (lto_symtab_resolve_can_prevail_p): Update.
18518 (lto_symtab_merge_cgraph_nodes): Update.
18519 * ipa-cp.c (determine_versionability, initialize_node_lattices,
18520 propagate_constants_accross_call, devirtualization_time_bonus,
18521 ipcp_propagate_stage): Update.
18522 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
18523 * ipa-inline-transform.c (clone_inlined_nodes,
18524 preserve_function_body_p): Update.
18525 * ipa-reference.c (propagate): Update.
18526 (write_node_summary_p): Update.
18527 * toplev.c (wrapup_global_declaration_2): Update.
18528 * cgraphunit.c (cgraph_analyze_function): Rename to ...
18529 (analyze_function) ... this one.
18530 (cgraph_process_new_functions): Update.
18531 (cgraph_reset_node): Export.
18532 (cgraph_finalize_function): Update.
18533 (cgraph_add_new_function): Update.
18534 (process_function_and_variable_attributes): Update.
18535 (varpool_finalize_decl): Update.
18536 (symbol_finalized): Remove.
18537 (symbol_finalized_and_needed): Rename to ...
18538 (symbol_defined_and_needed): ... update.
18539 (cgraph_analyze_functions): Update.
18540 (handle_alias_pairs): Update.
18541 (mark_functions_to_output): Update.
18542 (assemble_thunk): Update.
18543 (output_in_order): Update.
18544 (output_weakrefs): Update.
18545 (finalize_compilation_unit): Update.
18546 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
18547 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
18548 input_node, input_varpool_node): Update.
18549 * dbxout.c (dbxout_expand_expr): Update.
18550 * cgraphclones.c (cgraph_clone_node): Update.
18551 (cgraph_copy_node_for_versioning): Update.
18552 (cgraph_materialize_clone): Update.
18553 (cgraph_materialize_all_clones): Update.
18554 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
18555 propagate_pure_const, propagate_nothrow): Update.
18556 * lto-streamer-out.c (lto_output, write_symbol): Update.
18557 * ipa-utils.c (ipa_reverse_postorder): Update.
18558 * ipa-inline.c (can_inline_edge_p): Update.
18559 (update_caller_keys, ipa_inline): Update.
18560 * dwarf2out.c (reference_to_unused,
18561 premark_types_used_by_global_vars_helper): Update.
18562 * tree-eh.c (tree_could_trap_p): Update.
18563 * ipa-split.c (consider_split, execute_split_functions): Update.
18564 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
18565 has_addr_references_p): Update; move ahead in file for better
18567 (process_references): Simplify.
18568 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
18569 bodies are removed.
18570 (cgraph_comdat_can_be_unshared_p): Make static.
18571 (cgraph_externally_visible_p): Update.
18572 (varpool_externally_visible_p): Update.
18573 (function_and_variable_visibility): Update.
18574 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
18575 ipa_tm_mark_force_output_node): Update.
18576 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
18577 estimate_edge_devirt_benefit, inline_generate_summary,
18578 inline_write_summary): Update.
18579 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
18580 * ipa-prop.c (ipa_compute_jump_functions): Update.
18581 (ipa_print_node_params, ipa_prop_read_section,
18582 ipa_update_after_lto_read, read_replacements_section): Update.
18583 * varasm.c (mark_decl_referenced): Update.
18584 (assemble_alias, dump_tm_clone_pairs): Update.
18585 * tree-inline.c (copy_bb): Update.
18586 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
18588 * symtab.c (dump_symtab_base): Print new flags.
18589 (verify_symtab_base): Verify new flags.
18590 (symtab_alias_ultimate_target): New function.
18591 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
18592 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
18594 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
18596 * i386.c (ix86_get_function_versions_dispatcher,
18597 ix86_generate_version_dispatcher_body): Update.
18598 (fold_builtin_cpu): Use varpool_add_new_variable.
18599 * varpool.c (varpool_remove_initializer): Break out from ...
18600 (varpool_remove_node): ... this one.
18601 (dump_varpool_node, varpool_node_for_asm,
18602 cgraph_variable_initializer_availability, varpool_analyze_node,
18603 varpool_assemble_decl, varpool_remove_unreferenced_decls,
18604 varpool_finalize_named_section_flags, varpool_create_variable_alias):
18607 2013-05-29 Jan Hubicka <jh@suse.cz>
18609 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
18611 2013-05-29 Easwaran Raman <eraman@google.com>
18613 PR tree-optimization/57442
18614 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
18615 when control exits the main loop.
18617 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
18619 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
18621 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
18622 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
18623 * rx/t-rx: Add rx100 under multi library matches option for nofpu
18626 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18628 PR tree-optimization/57441
18629 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
18630 Don't limit size of incr_vec to number of candidates.
18632 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
18634 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
18635 and mips16 directories.
18636 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
18637 (MULTILIB_DIRNAMES): Ditto.
18638 (MULTILIB_EXCEPTIONS): Add new exceptions.
18639 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
18640 (MULTILIB_DIRNAMES): Ditto.
18641 (MULTILIB_EXCEPTIONS): Add new exceptions.
18643 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18644 Marcus Shawcroft <marcus.shawcroft@arm.com>
18646 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
18647 SYMBOL_TINY_ABSOLUTE.
18648 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
18649 SYMBOL_TINY_ABSOLUTE.
18650 (aarch64_expand_mov_immediate): Likewise.
18651 (aarch64_classify_symbol): Likewise.
18652 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
18653 Permit SYMBOL_TINY_ABSOLUTE.
18654 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
18656 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18657 Marcus Shawcroft <marcus.shawcroft@arm.com>
18659 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
18660 Refactor if/switch. Replace gcc_assert with if.
18662 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
18664 * config/i386/i386.c (initial_ix86_tune_features): Enable
18665 FP Reassociation for AMD bdver1 and bdver2.
18667 2013-05-29 Martin Jambor <mjambor@suse.cz>
18669 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
18670 and IMAGPART_EXPR do not occur within other handled_components.
18672 2013-05-29 Richard Biener <rguenther@suse.de>
18674 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
18675 access on whether the use is in the BB we currently try to
18677 (vect_bb_vectorization_profitable_p): Pass the BB we currently
18678 vectorize to vect_bb_slp_scalar_cost.
18680 2013-05-29 Richard Biener <rguenther@suse.de>
18682 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
18683 computing scalar cost offsetted by stmts that are kept live
18685 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
18686 for computation of scalar cost.
18688 2013-05-28 Steve Ellcey <sellcey@mips.com>
18690 * config/mips/mips-cpus.def (mips32r2): Change processor type.
18692 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
18694 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
18695 array notation built-in reduction functions.
18696 * doc/passes.texi (Passes): Added documentation about changes done
18698 * doc/invoke.texi (C Dialect Options): Added documentation about
18699 the -fcilkplus flag.
18700 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
18701 (BUILTINS_DEF): Depend on cilkplus.def.
18702 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
18703 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
18704 * cilkplus.def: New file.
18706 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
18708 PR rtl-optimization/57439
18709 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
18711 2013-05-28 Easwaran Raman <eraman@google.com>
18713 PR tree-optimization/57337
18714 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
18715 (find_insert_point): Correctly identify the insertion point
18716 when two statements with the same UID is compared.
18718 2013-05-28 Richard Biener <rguenther@suse.de>
18720 PR tree-optimization/56787
18721 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
18722 from the list of data references.
18723 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
18725 (vect_analyze_loop_operations): Likewise.
18726 (vect_transform_loop): Remove clobbers.
18728 2013-05-28 Martin Jambor <mjambor@suse.cz>
18730 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
18731 and REALPART_EXPRs have scalar type.
18733 2013-05-28 Richard Biener <rguenther@suse.de>
18735 PR tree-optimization/57411
18736 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
18738 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
18739 virtual operand propagation.
18741 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18743 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
18744 destination register for bmasksi_vis.
18745 (vector_init_bshuffle): Likewise.
18746 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
18748 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18750 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
18751 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
18752 mode if the instruction isn't available in the original mode.
18753 * config/sparc/sparc.opt (mfix-ut699): New option.
18754 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
18755 (divdf3): Turn into expander.
18756 (divdf3_nofix): New insn.
18757 (divdf3_fix): Likewise.
18758 (divsf3): Disable if -mfix-ut699.
18759 (sqrtdf2): Turn into expander.
18760 (sqrtdf2_nofix): New insn.
18761 (sqrtdf2_fix): Likewise.
18762 (sqrtsf2): Disable if -mfix-ut699.
18764 2013-05-27 Richard Biener <rguenther@suse.de>
18766 PR middle-end/57412
18767 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
18768 block for the new loop.
18770 2013-05-27 Richard Biener <rguenther@suse.de>
18772 PR tree-optimization/57343
18773 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
18774 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
18775 (number_of_iterations_cond): Do not build the folded tree.
18777 2013-05-27 Richard Biener <rguenther@suse.de>
18780 PR middle-end/57381
18781 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
18782 OEP_CONSTANT_ADDRESS_OF retained.
18784 PR tree-optimization/57417
18785 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
18786 for unchanged base.
18787 (set_ssa_val_to): Compare addresses using
18788 get_addr_base_and_unit_offset.
18790 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
18792 PR rtl-optimization/56833
18793 * postreload.c (move2add_record_mode): New function.
18794 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
18795 (move2add_use_add2_insn): Use move2add_record_sym_value.
18796 (move2add_use_add3_insn): Likewise.
18797 (reload_cse_move2add): Use move2add_valid_value_p and
18798 move2add_record_mode. Invalidate call-clobbered and REG_INC
18799 affected regs by setting reg_mode to VOIDmode.
18800 (move2add_note_store): Don't pretend the inside of a SUBREG is
18801 the actual destination. Invalidate single/leading registers by
18802 setting reg_mode to VOIDmode.
18803 Use move2add_record_sym_value, move2add_valid_value_p and
18804 move2add_record_mode.
18806 2013-05-27 Richard Biener <rguenther@suse.de>
18808 PR tree-optimization/57396
18809 * tree-affine.c (double_int_constant_multiple_p): Properly
18810 return false for val == 0 and div != 0.
18812 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18814 * config/mips/mips.h: Use #elif in preprocessor conditions.
18816 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18819 * config/mips/constraints.md (kl): New constraint.
18820 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
18821 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
18822 constraint for operand 0. Split after CSE for MIPS16. Emit a move
18823 from LO for MIPS16.
18824 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
18826 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18829 * config/mips/mips.c (mips_can_inline_p): New function.
18830 (TARGET_CAN_INLINE_P): Define.
18832 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
18834 * sched-int.h (ds_t, dw_t): Make unsigned int.
18835 Fix documentation that describes how all the ds_t bits are used.
18836 Reserve the last bit for delayed-branch scheduling.
18837 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
18838 (BITS_PER_DEP_WEAK): Fix definition and documentation.
18839 (gen_dep_weak_1): Remove prototype.
18840 * sched-deps.c (get_dep_weak_1): Make static.
18841 * target.def (speculate_insn, needs_block_p, gen_spec_check,
18842 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
18843 * doc/tm.texi: Regenerate.
18844 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
18846 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
18849 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
18851 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
18852 Sandra Loosemore <sandra@codesourcery.com>
18854 * config.gcc (powerpc-*): Allow native for with-cpu.
18856 2013-05-24 Jeff Law <law@redhat.com>
18858 PR tree-optimization/57124
18859 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
18860 conversion feeding a condition if the range has an overflow
18861 if -fstrict-overflow. Add warnings for when we do make the
18864 2013-05-24 Dehao Chen <dehao@google.com>
18866 * tree-cfg.c (locus_discrim_map): Fix the typo.
18867 (locus_discrim_hasher): Likewise.
18868 (locus_discrim_hasher::hash): Likewise.
18869 (locus_discrim_hasher::equal): Likewise.
18871 2013-05-24 Martin Jambor <mjambor@suse.cz>
18873 PR tree-optimization/57294
18874 * cgraph.h (ipa_record_stmt_references): Declare.
18875 * cgraphbuild.c (ipa_record_stmt_references): New function.
18876 (build_cgraph_edges): Use ipa_record_stmt_references.
18877 (rebuild_cgraph_edges): Likewise.
18878 (cgraph_rebuild_references): Likewise.
18879 * ipa-prop.c (ipa_modify_call_arguments): Discard references
18880 associated with the old statement and build references from the
18881 newly built statements.
18882 * ipa-ref.c (ipa_remove_stmt_references): New function.
18883 * ipa-ref.h (ipa_remove_stmt_references): Declare.
18885 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
18887 * lra-constraints.c (emit_spill_move): Use smaller mode for
18889 (check_and_process_move): Consider mem-reg moves for secondary
18891 (curr_insn_transform): Don't lose insns emitted before for
18892 secondary memory moves.
18893 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
18894 reg set up in the current insn.
18896 2013-05-24 Dehao Chen <dehao@google.com>
18898 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
18900 (locus_descrim_hasher::equal): Likewise.
18901 (build_gimple_cfg): New discriminator assignment algorithm.
18902 (make_edges): Likewise.
18903 (next_discriminator_for_locus): Likewise.
18904 (same_line_p): Likewise.
18905 (assign_discriminators): Likewise.
18906 (make_cond_expr_edges): Likewise.
18907 (make_gimple_switch_edges): Likewise.
18908 (make_goto_expr_edges): Likewise.
18909 (make_gimple_asm_edges): Likewise.
18911 2013-05-24 Ian Bolton <ian.bolton@arm.com>
18913 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
18914 X format specifier to only display bottom 16 bits.
18915 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
18916 immediate to match for operand 2, since it will be masked.
18918 2013-05-24 Richard Biener <rguenther@suse.de>
18920 PR tree-optimization/57287
18921 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
18922 all SSA names that occur in abnormal PHIs.
18924 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
18927 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
18928 that index is not negative.
18930 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
18932 PR rtl-optimization/55177
18933 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
18934 (simplify_byte_swapping_operation): New.
18935 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
18936 (simplify_relational_operation_1): Deal with BSWAP.
18938 2013-05-23 Richard Henderson <rth@redhat.com>
18941 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
18942 (ix86_reorg): Call it.
18944 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
18947 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
18948 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
18949 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
18951 2013-05-23 Christian Bruel <christian.bruel@st.com>
18954 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
18956 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18957 Marcus Shawcroft <marcus.shawcroft@arm.com>
18959 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
18960 * config/aarch64/constraints.md (Usa): Remove.
18961 * doc/md.texi (AArch64 Usa): Remove.
18963 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18964 Marcus Shawcroft <marcus.shawcroft@arm.com>
18966 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
18967 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
18968 * config/aarch64/predicates.md (aarch64_const_address): Remove.
18969 (aarch64_mov_operand): Use aarch64_mov_operand_p.
18971 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
18973 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
18974 instruction (AdvSIMD).
18975 * config/aarch64/aarch64-builtins.c
18976 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
18977 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
18979 2013-05-23 Martin Jambor <mjambor@suse.cz>
18981 PR middle-end/57347
18982 * tree.h (contains_bitfld_component_ref_p): Declare.
18983 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
18984 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
18986 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
18987 not access a bit-field. Assert all final offsets are byte-aligned.
18989 2013-05-23 Richard Biener <rguenther@suse.de>
18991 PR tree-optimization/57380
18992 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
18993 least one invariant or re-used load.
18994 * passes.c (init_optimization_passes): Move pass_phiprop before
18997 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
18999 * config/aarch64/aarch64-simd.md
19000 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
19002 2013-05-23 Richard Biener <rguenther@suse.de>
19004 PR middle-end/57381
19005 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19006 OEP_CONSTANT_ADDRESS_OF retained.
19008 2013-05-23 Jakub Jelinek <jakub@redhat.com>
19010 PR middle-end/57344
19011 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
19012 don't lower unit. Handle unit not being always BITS_PER_WORD.
19014 2013-05-23 Richard Biener <rguenther@suse.de>
19016 PR rtl-optimization/57341
19017 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
19018 instead of true_dependence.
19020 2013-05-22 David Malcolm <dmalcolm@redhat.com>
19022 * bb-reorder.c (branch_threshold): Make const.
19023 (exec_threshold): Ditto.
19025 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19026 Pat Haugen <pthaugen@us.ibm.com>
19027 Peter Bergner <bergner@vnet.ibm.com>
19029 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
19030 documentation for the power8 crypto builtins.
19032 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
19034 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
19035 macros for defining power8 builtin functions.
19036 (BU_P8V_AV_2): Likewise.
19037 (BU_P8V_AV_P): Likewise.
19038 (BU_P8V_VSX_1): Likewise.
19039 (BU_P8V_OVERLOAD_1): Likewise.
19040 (BU_P8V_OVERLOAD_2): Likewise.
19041 (BU_CRYPTO_1): Likewise.
19042 (BU_CRYPTO_2): Likewise.
19043 (BU_CRYPTO_3): Likewise.
19044 (BU_CRYPTO_OVERLOAD_1): Likewise.
19045 (BU_CRYPTO_OVERLOAD_2): Likewise.
19046 (XSCVSPDP): Fix typo, point to the correct instruction.
19047 (VCIPHER): Add power8 crypto builtins.
19048 (VCIPHERLAST): Likewise.
19049 (VNCIPHER): Likewise.
19050 (VNCIPHERLAST): Likewise.
19051 (VPMSUMB): Likewise.
19052 (VPMSUMH): Likewise.
19053 (VPMSUMW): Likewise.
19054 (VPERMXOR_V2DI): Likewise.
19055 (VPERMXOR_V4SI: Likewise.
19056 (VPERMXOR_V8HI: Likewise.
19057 (VPERMXOR_V16QI: Likewise.
19058 (VSHASIGMAW): Likewise.
19059 (VSHASIGMAD): Likewise.
19060 (VPMSUM): Likewise.
19061 (VPERMXOR): Likewise.
19062 (VSHASIGMA): Likewise.
19064 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
19065 __CRYPTO__ if the crypto instructions are available.
19066 (altivec_overloaded_builtins): Add support for overloaded power8
19069 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
19070 support for power8 crypto builtins.
19071 (builtin_function_type): Likewise.
19072 (altivec_init_builtins): Add support for builtins that take vector
19073 long long (V2DI) arguments.
19075 * config/rs6000/crypto.md: New file, define power8 crypto
19078 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19079 Pat Haugen <pthaugen@us.ibm.com>
19080 Peter Bergner <bergner@vnet.ibm.com>
19082 * doc/invoke.texi (Option Summary): Add power8 options.
19083 (RS/6000 and PowerPC Options): Likewise.
19085 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
19086 constraints.md instead of rs6000.h. Reorder w* constraints. Add
19087 wm, wn, wr documentation.
19089 * config/rs6000/constraints.md (wm): New constraint for VSX
19090 registers if direct move instructions are enabled.
19091 (wn): New constraint for no registers.
19092 (wq): New constraint for quad word even GPR registers.
19093 (wr): New constraint if 64-bit instructions are enabled.
19094 (wv): New constraint if power8 vector instructions are enabled.
19095 (wQ): New constraint for quad word memory locations.
19097 * config/rs6000/predicates.md (const_0_to_15_operand): New
19098 constraint for 0..15 for crypto instructions.
19099 (gpc_reg_operand): If VSX allow registers in VSX registers as well
19100 as GPR and floating point registers.
19101 (int_reg_operand): New predicate to match only GPR registers.
19102 (base_reg_operand): New predicate to match base registers.
19103 (quad_int_reg_operand): New predicate to match even GPR registers
19104 for quad memory operations.
19105 (vsx_reg_or_cint_operand): New predicate to allow vector logical
19106 operations in both GPR and VSX registers.
19107 (quad_memory_operand): New predicate for quad memory operations.
19108 (reg_or_indexed_operand): New predicate for direct move support.
19110 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
19111 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
19112 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
19113 (POWERPC_MASKS): Add power8 options.
19114 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
19117 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
19118 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
19120 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
19121 (-mpower8-fusion): New power8 options.
19122 (-mpower8-fusion-sign): Likewise.
19123 (-mpower8-vector): Likewise.
19124 (-mcrypto): Likewise.
19125 (-mdirect-move): Likewise.
19126 (-mquad-memory): Likewise.
19128 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
19129 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
19131 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
19132 (rs6000_debug_vector_unit): Add p8_vector.
19133 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
19134 definitions. Also print fusion state.
19135 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
19136 (rs6000_builtin_mask_calculate): Add power8 builtin support.
19137 (rs6000_option_override_internal): Add support for power8.
19138 (rs6000_common_init_builtins): Add debugging for skipped builtins
19139 if -mdebug=builtin.
19140 (rs6000_adjust_cost): Add power8 support.
19141 (rs6000_issue_rate): Likewise.
19142 (insn_must_be_first_in_group): Likewise.
19143 (insn_must_be_last_in_group): Likewise.
19144 (force_new_group): Likewise.
19145 (rs6000_register_move_cost): Likewise.
19146 (rs6000_opt_masks): Likewise.
19148 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
19149 power8 capable assembler, default to power7 options.
19150 (TARGET_DIRECT_MOVE): Likewise.
19151 (TARGET_CRYPTO): Likewise.
19152 (TARGET_P8_VECTOR): Likewise.
19153 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
19154 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
19155 (VECTOR_MEM_P8_VECTOR_P): Likewise.
19156 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
19157 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
19158 (TARGET_XSCVDPSPN): Likewise.
19159 (TARGET_XSCVSPDPN): Likewsie.
19160 (TARGET_SYNC_HI_QI): Likewise.
19161 (TARGET_SYNC_TI): Likewise.
19162 (MASK_CRYPTO): Likewise.
19163 (MASK_DIRECT_MOVE): Likewise.
19164 (MASK_P8_FUSION): Likewise.
19165 (MASK_P8_VECTOR): Likewise.
19166 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
19167 temporary used by some of the direct move instructions to get two FP
19168 temporary registers does not force creation of a stack frame.
19169 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
19170 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
19171 that any VSX registers are tieable, even if they are also an
19172 Altivec vector mode.
19173 (r6000_reg_class_enum): Add wm, wr, wv constraints.
19174 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
19175 (RS6000_BTM_CRYPTO): Likewise.
19176 (RS6000_BTM_COMMON): Likewise.
19178 * config/rs6000/rs6000.md (cpu attribute): Add power8.
19179 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
19180 (enum rs6000_vector): Add power8 vector support.
19182 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19186 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
19187 (any_sibcall_could_use_r3): this and handle indirect calls.
19188 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
19190 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19192 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
19194 2013-05-22 Richard Biener <rguenther@suse.de>
19196 PR middle-end/57349
19197 * profile.c (branch_prob): Do not split blocks that are
19198 abnormally receiving from ECF_RETURNS_TWICE functions.
19200 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19202 * recog.c (offsettable_address_addr_space_p): Fix calculation of
19203 address mode. Move pointer mode initialization to the same place.
19205 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
19207 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
19208 while it has any effect.
19210 2013-05-21 Easwaran Raman <eraman@google.com>
19212 PR tree-optimization/57322
19213 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
19214 UID of the statement added to the BB to be 1.
19216 2013-05-21 Jakub Jelinek <jakub@redhat.com>
19218 PR tree-optimization/57331
19219 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
19220 of conversion from pointer type to integral type with integer.
19222 2013-05-21 Martin Jambor <mjambor@suse.cz>
19225 * ipa-prop.c (ipa_read_node_info): Process param_used and
19226 controlled_uses in the same order as when writing.
19228 2013-05-21 Magnus Granberg <baldrick@free.fr>
19231 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
19233 2013-05-21 Richard Biener <rguenther@suse.de>
19235 PR tree-optimization/57318
19236 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
19237 estimate stmts with side-effects as likely eliminated.
19239 2013-05-21 Richard Biener <rguenther@suse.de>
19241 PR tree-optimization/57330
19242 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
19243 preserve the call stmts fntype.
19245 2013-05-21 Richard Biener <rguenther@suse.de>
19247 PR tree-optimization/57303
19248 * tree-ssa-sink.c (statement_sink_location): Improve killing
19249 stmt detection and properly handle self-assignments.
19251 2013-05-21 Christian Bruel <christian.bruel@st.com>
19253 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
19254 spanning registers. LEAF_REG_REMAP is supported only for contiguous
19255 registers. Set register size out of the PARALLEL loop.
19257 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
19260 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
19261 (*fmasf4, *fmasf4_media): New insns.
19263 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19265 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
19266 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
19267 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
19268 (mips_idiv_insns): Update the comments to say that the returned
19269 instruction counts are in units of BASE_INSN_LENGTH.
19270 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
19271 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
19272 using 2 rather than 4 as the length of indirect MIPS16 and
19273 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
19274 length of a NOP. Don't divide MIPS16 lengths by 2.
19275 (mips16_split_long_branches): Assume a branch is long if the
19276 length is greater than 4 rather than 8.
19277 * config/mips/mips.md (length): Give MIPS16 lengths directly,
19278 rather than multiplying them by 2. Multiply instruction counts
19279 by BASE_INSN_LENGTH rather than 4.
19280 (*jump_mips16, tls_get_tp_mips16_<mode>)
19281 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
19283 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19285 * config/mips/mips.md (extended_mips16): Remove branch case.
19286 (length): Remove duplicated extended_mips16 test.
19288 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19290 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
19292 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
19294 * recog.h (Recog_data): Rename to...
19295 (recog_data_d): ...this.
19296 (recog_data): Update accordingly.
19297 * recog.c (recog_data): Likewise.
19298 * reload.c (save_recog_data): Likewise.
19299 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
19300 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
19302 2013-05-17 Julian Brown <julian@codesourcery.com>
19304 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
19305 found in a REG_EQUAL note, invalidate it.
19307 2013-05-17 Easwaran Raman <eraman@google.com>
19309 * tree-ssa-reassoc.c (find_insert_point): New function.
19310 (insert_stmt_after): Likewise.
19311 (get_def_stmt): Likewise.
19312 (ensure_ops_are_available): Likewise.
19313 (not_dominated_by): Likewise.
19314 (rewrite_expr_tree): Do not move statements beyond what is
19315 necessary. Remove call to swap_ops_for_binary_stmt...
19316 (reassociate_bb): ... and move it here.
19317 (build_and_add_sum): Assign UIDs for new statements.
19318 (linearize_expr): Likewise.
19319 (do_reassoc): Renumber gimple statement UIDs.
19321 2013-05-17 Jan Hubicka <jh@suse.cz>
19323 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
19325 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
19326 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
19328 (output_weakrefs): Update.
19330 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
19331 Martin Jambor <mjambor@suse.cz>
19333 PR middle-end/57276
19334 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
19335 value that corresponds to the given aggval is found in values vector.
19337 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
19339 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
19340 sse, sse2, sse3, ssse3 and sse4a flags to options.
19342 2013-05-17 David Malcolm <dmalcolm@redhat.com>
19344 * gengtype-state.c: (s_expr_writer): New class, to handle
19345 prettifying of output layout of s-expressions.
19346 (state_writer): New class, to write out gtype.state.
19347 (state_written_type_count): Move this variable into member data of
19349 (s_expr_writer::s_expr_writer): New code: constructor for new class
19350 (state_writer::state_writer(): ditto
19351 (s_expr_writer::write_new_line): New function
19352 (s_expr_writer::write_any_indent): ditto
19353 (s_expr_writer::begin_s_expr): ditto
19354 (s_expr_writer::end_s_expr): ditto
19355 (write_state_fileloc): convert to method of state_writer...
19356 (state_writer:: write_state_fileloc): ...and use methods of
19357 s_expr_writer to write indentation into the gtype.state output file
19358 to visually represent the hierarchical structure of the list
19360 (write_state_fields): ditto, renaming to...
19361 (state_writer::write_state_fields)
19362 (write_state_a_string): ditto, renaming to...
19363 (state_writer::write_state_a_string)
19364 (write_state_string_option): ditto, renaming to...
19365 (state_writer::write_state_string_option)
19366 (write_state_type_option): ditto, renaming to...
19367 (state_writer::write_state_type_option)
19368 (write_state_nested_option): ditto, renaming to...
19369 (state_writer::write_state_nested_option)
19370 (write_state_option): ditto, renaming to...
19371 (state_writer::write_state_option)
19372 (write_state_options): ditto, renaming to...
19373 (state_writer::write_state_options)
19374 (write_state_lang_bitmap): ditto, renaming to...
19375 (state_writer::write_state_lang_bitmap)
19376 (write_state_version): ditto, renaming to...
19377 (state_writer::write_state_version)
19378 (write_state_scalar_type): ditto, renaming to...
19379 (state_writer::write_state_scalar_type)
19380 (write_state_string_type): ditto, renaming to...
19381 (state_writer::write_state_string_type)
19382 (write_state_undefined_type): ditto, renaming to...
19383 (state_writer::write_state_undefined_type)
19384 (write_state_struct_union_type): ditto, renaming to...
19385 (state_writer::write_state_struct_union_type)
19386 (write_state_struct_type): ditto, renaming to...
19387 (state_writer::write_state_struct_type)
19388 (write_state_user_struct_type): ditto, renaming to...
19389 (state_writer::write_state_user_struct_type)
19390 (write_state_lang_struct_type): ditto, renaming to...
19391 (state_writer::write_state_lang_struct_type)
19392 (write_state_param_struct_type): ditto, renaming to...
19393 (state_writer::write_state_param_struct_type)
19394 (write_state_pointer_type): ditto, renaming to...
19395 (state_writer::write_state_pointer_type)
19396 (write_state_array_type): ditto, renaming to...
19397 (state_writer::write_state_array_type)
19398 (write_state_gc_used): ditto, renaming to...
19399 (state_writer::write_state_gc_used)
19400 (write_state_common_type_content): ditto, renaming to...
19401 (state_writer::write_state_common_type_content)
19402 (write_state_type): ditto, renaming to...
19403 (state_writer::write_state_type)
19404 (write_state_pair_list): ditto, renaming to...
19405 (state_writer::write_state_pair_list)
19406 (write_state_pair): ditto, renaming to...
19407 (state_writer::write_state_pair)
19408 (write_state_typedefs): ditto, renaming to...
19409 (state_writer::write_state_typedefs)
19410 (write_state_structures): ditto, renaming to...
19411 (state_writer::write_state_structures)
19412 (write_state_param_structs): ditto, renaming to...
19413 (state_writer::write_state_param_structs)
19414 (write_state_variables): ditto, renaming to...
19415 (state_writer::write_state_variables)
19416 (write_state_srcdir): ditto, renaming to...
19417 (state_writer::write_state_srcdir)
19418 (write_state_files_list): ditto, renaming to...
19419 (state_writer::write_state_files_list)
19420 (write_state_languages): ditto, renaming to...
19421 (state_writer::write_state_languages)
19422 (write_state): create a state_writer instance and use it when
19423 writing out the state file
19425 2013-05-17 Mike Stump <mikestump@comcast.net>
19427 PR rtl-optimization/57304
19428 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
19429 accessing DF_REF_REAL_LOC.
19431 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19433 PR rtl-optimization/57281
19434 PR rtl-optimization/57300
19435 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
19436 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
19437 what the other splitter did if the registers are dead.
19439 2013-05-17 Richard Biener <rguenther@suse.de>
19441 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
19444 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19446 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
19449 2013-05-17 Marek Polacek <polacek@redhat.com>
19451 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
19452 length when doing non-zero store of storing '\0' to '\0'.
19454 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19456 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
19457 vect_external_def oprnd1 with loop_vinfo, try to emit
19458 optional cast, negation and and stmts on the loop preheader
19459 edge instead of into the pattern def seq.
19461 PR tree-optimization/57051
19462 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
19463 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
19465 2013-05-16 Nick Clifton <nickc@redhat.com>
19467 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
19468 (rl78_is_naked_func): New function.
19469 (rl78_expand_prologue): Skip prologue generation for naked functions.
19470 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
19471 * doc/extend.texi (naked): Add RL78 to the list of processors
19472 that supports this attribute.
19474 2013-05-16 Jeff Law <law@redhat.com>
19476 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
19478 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19480 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
19481 cache parameters using detect_caches_amd also for CYRIX,
19482 NSC and TM2 signatures.
19484 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19485 Dzianis Kahanovich <mahatma@eu.by>
19489 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
19490 VIA/Centaur processors and determine their cache parameters
19491 using detect_caches_amd.
19493 2013-05-16 Teresa Johnson <tejohnson@google.com>
19495 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
19496 (rtl_verify_edges): New function.
19497 (rtl_verify_bb_insns): Ditto.
19498 (rtl_verify_bb_pointers): Ditto.
19499 (rtl_verify_bb_insn_chain): Ditto.
19500 (rtl_verify_fallthru): Ditto.
19501 (rtl_verify_bb_layout): Ditto.
19502 (rtl_verify_flow_info_1): Outline checks into new functions.
19503 (rtl_verify_flow_info): Ditto.
19505 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
19507 * cfghooks.c (copy_bbs): Add update_dominance argument.
19508 * cfghooks.h (copy_bbs): Update prototype.
19509 * tree-cfg.c (gimple_duplicate_sese_region):
19510 Add update_dominance argument.
19511 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
19512 * tree-ssa-loop-ch.c (copy_loop_headers): Update
19513 gimple_duplicate_sese_region call.
19514 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
19515 Update copy_bbs call.
19516 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
19517 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
19519 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19521 * tree-vectorizer.h (NUM_PATTERNS): Increment.
19522 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
19523 vect_recog_rotate_pattern.
19524 (vect_recog_rotate_pattern): New function.
19526 2013-05-16 Jason Merrill <jason@redhat.com>
19528 * Makefile.in (LLINKER): New variable.
19529 (mostlyclean): Remove link mutex.
19530 * configure.ac: Handle --enable-link-mutex.
19531 * lock-and-run.sh: New script.
19533 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19536 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
19539 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19541 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
19543 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
19545 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
19546 * config/arm/arm.c (next_consecutive_mem): New function.
19547 (gen_movmem_ldrd_strd): Likewise.
19548 * config/arm/arm.md (movmemqi): Update condition and code.
19549 (unaligned_loaddi, unaligned_storedi): New patterns.
19551 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19553 * config.gcc: Obsolete *-*-solaris2.9*.
19554 * doc/install.texi (Specific, *-*-solaris2*): Document it.
19556 2013-05-16 Richard Biener <rguenther@suse.de>
19558 * passes.c (init_optimization_passes): Move pass_parallelize_loops
19559 earlier, after GRAPHITE transforms and IV canonicalization.
19561 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19563 * omp-low.c (extract_omp_for_data): For collapsed loops,
19564 if at least one of the loops is known at compile time to
19565 iterate zero times, set count to 0.
19566 (expand_omp_regimplify_p): New function.
19567 (expand_omp_for_generic): For collapsed loops, if at least
19568 one of the loops isn't known to iterate at least once,
19569 add runtime check with setting count to 0.
19570 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
19571 For unsigned types if it isn't known at compile time that
19572 the loop will iterate at least once, add runtime check to bypass
19573 the whole loop if initial condition isn't true.
19575 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
19577 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
19579 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
19581 PR middle-end/57286
19582 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
19583 transformations to avoid an infinite loop.
19585 2013-05-16 Marek Polacek <polacek@redhat.com>
19587 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
19589 2013-05-15 Leif Ekblad <leif@rdos.net>
19591 * config/i386/i386.c (ix86_decompose_address): Use
19592 DEFAULT_TLS_SEG_REG to access TLS segment register.
19593 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
19594 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
19595 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
19597 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
19600 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
19601 sibling calls to functions that would normally be lazily bound,
19602 unless $gp is call-clobbered.
19604 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
19606 * config/i386/i386.c (ix86_option_override_internal): Update
19607 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
19608 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
19609 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
19610 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
19612 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
19614 2013-05-15 Andreas Schwab <schwab@suse.de>
19616 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
19617 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
19620 2013-05-15 Teresa Johnson <tejohnson@google.com>
19622 * loop-unroll.c (report_unroll_peel): Check decision before
19623 emitting unroll/peel message.
19625 2013-05-15 Teresa Johnson <tejohnson@google.com>
19627 * function.h (has_bb_partition): New rtl_data flag.
19628 (bb_reorder_complete): Ditto.
19629 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
19630 instead of flag_reorder_blocks_and_partition.
19631 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
19632 with some enhancements.
19633 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
19634 * bb-reorder.c (connect_traces): Check for has_bb_partition
19635 instead of flag_reorder_blocks_and_partition.
19636 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
19637 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
19638 verify_hot_cold_block_grouping.
19639 (partition_hot_cold_basic_blocks): Set has_bb_partition.
19641 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19644 * config/arm/predicates.md (call_insn_operand): New predicate.
19645 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
19646 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
19647 if insn is not a tail call.
19648 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
19650 * config/arm/arm.h (enum reg_class): New caller save register class.
19651 (REG_CLASS_NAMES): Likewise.
19652 (REG_CLASS_CONTENTS): Likewise.
19653 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
19656 2013-05-15 Richard Biener <rguenther@suse.de>
19658 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
19659 of MSG_OPTIMIZED_LOCATIONS.
19660 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
19661 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
19663 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
19664 of MSG_OPTIMIZED_LOCATIONS.
19665 (execute_vect_slp): Likewise.
19666 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
19667 (vect_create_cond_for_alias_checks): Likewise.
19668 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
19669 (vect_recog_widen_mult_pattern): Likewise.
19670 (vect_recog_widen_sum_pattern): Likewise.
19671 (vect_recog_over_widening_pattern): Likewise.
19672 (vect_recog_widen_shift_pattern): Likewise.
19673 (vect_recog_vector_vector_shift_pattern): Likewise.
19674 (vect_recog_divmod_pattern): Likewise.
19675 (vect_recog_mixed_size_cond_pattern): Likewise.
19676 (vect_recog_bool_pattern): Likewise.
19677 (vect_pattern_recog_1): Likewise.
19679 2013-05-15 Martin Jambor <mjambor@suse.cz>
19681 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
19682 non-functions to builtin_unreachable.
19683 * ipa-inline-transform.c (inline_call): Do not assert estimates were
19684 correct when new direct edges were discovered.
19686 2013-05-15 Martin Jambor <mjambor@suse.cz>
19688 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
19689 header, print symbol order instead of node uid, print more information
19690 about indirect edge targets.
19691 (ipa_make_edge_direct_to_target): Print symbol order instead of node
19693 (ipa_make_edge_direct_to_target): Likewise.
19694 (remove_described_reference): Likewise.
19695 (propagate_controlled_uses): Likewise.
19696 (ipa_print_node_params): Also print symbol order.
19697 (ipcp_transform_function): Print symbol order instead of node uids.
19698 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
19699 (cgraph_get_create_real_symbol_node): Likewise.
19700 * ipa-cp.c (print_lattice): Likewise.
19701 (print_all_lattices): Likewise.
19702 (determine_versionability): Likewise.
19703 (initialize_node_lattices): Likewise.
19704 (estimate_local_effects): Likewise.
19705 (update_profiling_info): Likewise.
19706 (create_specialized_node): Likewise.
19707 (perhaps_add_new_callers): Likewise.
19708 (decide_about_value): Likewise.
19709 (decide_whether_version_node): Likewise.
19710 (identify_dead_nodes): Likewise.
19711 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
19712 (dump_inline_summary): Likewise.
19713 (estimate_node_size_and_time): Likewise.
19714 (inline_analyze_function): Likewise.
19715 * ipa-inline.c (report_inline_failed_reason): Likewise.
19716 (want_early_inline_function_p): Likewise.
19717 (edge_badness): Likewise.
19718 (update_edge_key): Likewise.
19719 (inline_small_functions): Likewise. Add dumping of order to two other
19721 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
19722 instead of node uids.
19723 (propagate_pure_const): Likewise.
19724 (propagate_pure_const): Likewise.
19725 * ipa-utils.c (dump_cgraph_node_set): Likewise.
19726 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
19727 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
19729 * tree-pretty-print.c (dump_function_header): Likewise.
19730 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
19731 Print symbol order instead of node uids.
19733 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19735 * config/s390/s390.c (s390_register_move_cost): Don't impose the
19736 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
19738 2013-05-15 Richard Biener <rguenther@suse.de>
19740 PR tree-optimization/57275
19741 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
19742 return value for fail to do runtime alias checks for gather loads.
19744 2013-05-15 Jan Hubicka <jh@suse.cz>
19748 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
19749 weakrefs are not external.
19750 (lto_symtab_merge_decls): Fix thinko when dealing with
19751 non-lto_symtab decls.
19752 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
19753 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
19754 * varpool.c (dump_varpool_node): Dump more flags.
19756 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19758 * config/i386/i386.c (processor_alias_table): Add instruction
19759 FSGSBASE for AMD bdver3 architecture.
19761 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19763 * tree.c (warn_deprecated_use): Print file:line using locus color.
19764 * diagnostic.c (diagnostic_report_current_module): Print file:line
19765 and file:line:column using locus color.
19767 2013-05-14 Mike Stump <mikestump@comcast.net>
19769 * gdbinit.in: Add __null.
19771 2013-05-14 Mike Stump <mikestump@comcast.net>
19773 * recog.h: Rename struct recog_data to Recog_data.
19774 * recog.c: Likewise.
19775 * reload.c (can_reload_into): Likewise.
19776 * config/picochip/picochip.c: Likewise.
19778 2013-05-14 Mike Stump <mikestump@comcast.net>
19780 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
19782 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
19784 * resource.h (struct resources): Remove unch_memory member.
19785 (CLEAR_RESOURCE): Don't clear unch_memory.
19786 * resource.c (mark_referenced_resources): Don't set it.
19787 (mark_set_resources): Likewise.
19788 (mark_target_live_regs): Don't clear it.
19789 (init_resource_info): Likewise.
19790 * reorg.c (resource_conflicts_p): Don't compare it.
19791 (redundant_insn): Don't set it.
19793 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
19795 * emit-rtl.c (next_label): Remove unused function.
19796 (skip_consecutive_labels, link_cc0_insns): Move to ...
19797 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
19798 only place where these functions are used, and make them static.
19800 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
19802 * fold-const.c (fold_negate_expr): Handle vectors.
19803 (fold_truth_not_expr): Make it static.
19804 (fold_invert_truthvalue): New static function.
19805 (invert_truthvalue_loc): Handle vectors. Do not call
19806 fold_truth_not_expr directly.
19807 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
19808 <TRUTH_NOT_EXPR>: Do not cast to boolean.
19809 (fold_comparison): Handle vector constants.
19810 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
19811 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
19812 * tree.h (fold_truth_not_expr): Remove declaration.
19814 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
19816 * config/aarch64/aarch64-simd.md
19817 (aarch64_vcond_internal<mode>): Rename to...
19818 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
19819 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
19820 float modes. Clarify all iterator modes.
19821 (vcond<mode><mode>): Use new name for vcond expanders.
19822 (vcond<v_cmp_result><mode>): Likewise.
19823 (vcondu<mode><mode>: Likewise.
19824 * config/aarch64/iterators.md (VDQF_COND): New.
19826 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
19829 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
19830 variable for the shift amount. Check that we shift by non-negative
19833 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
19836 * config/arm/arm.h (EPILOGUE_USES): Only return true
19837 for LR_REGNUM after epilogue_completed.
19839 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
19841 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
19842 is error_mark_node.
19844 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19847 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
19848 and Solaris 11+/x86 with gld.
19849 * configure: Regenerate.
19851 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19853 * expmed.c (expand_shift_1): Canonicalize rotates by
19854 constant bitsize / 2 to bitsize - 1.
19855 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
19856 case ROTATERT>: Likewise.
19859 2013-05-10 Jakub Jelinek <jakub@redhat.com>
19861 * config/i386/i386.md (rotateinv): New code attr.
19862 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
19863 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
19864 roll $31, %eax, etc.
19866 2013-05-14 Richard Biener <rguenther@suse.de>
19868 PR middle-end/57235
19869 * tree-eh.c (sink_clobbers): Give up for successors with
19870 multiple predecessors and no virtual uses.
19872 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
19874 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
19875 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
19877 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19879 PR middle-end/57251
19880 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
19881 the case when both op0 and op1 have VOIDmode.
19883 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
19885 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
19886 in multiply-accumulate mode.
19888 2013-05-13 Guozhi Wei <carrot@google.com>
19890 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
19892 2013-05-13 Kai Tietz <ktietz@redhat.com>
19895 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
19896 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
19897 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
19898 * config/i386/i386.c (ix86_option_override_internal): Likewise.
19899 (ix86_expand_prologue): Likewise.
19900 (ix86_expand_split_stack_prologue): Likewise.
19901 (legitimate_pic_address_disp_p): Likewise.
19902 (legitimize_pic_address): Likewise.
19903 (legitimize_tls_address): Likewise.
19904 (legitimize_pe_coff_symbol): Likewise.
19905 (output_pic_addr_const): Likewise.
19906 (construct_plt_address): Likewise.
19907 (ix86_expand_call): Likewise.
19908 (x86_output_mi_thunk): Likewise.
19909 (x86_function_profiler): Likewise.
19911 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
19913 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
19914 similar switch cases.
19915 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
19916 (aarch64_simd_mov_to_<mode>low): Delete.
19917 (aarch64_simd_mov_to_<mode>high): Delete.
19918 (move_lo_quad_<mode>): Add w<-r alternative.
19919 (aarch64_simd_move_hi_quad_<mode>): Likewise.
19920 (aarch64_simd_mov_from_*): Update type attribute.
19921 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
19924 2013-05-13 Jan Hubicka <jh@suse.cz>
19926 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
19927 * config/i386/i386.c (ix86_compute_frame_layout,
19928 ix86_expand_epilogue, emit_i387_cw_initialization,
19929 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
19930 ix86_local_alignment): Fix use of size/speed predicates.
19932 2013-05-13 Jakub Jelinek <jakub@redhat.com>
19934 PR tree-optimization/45216
19935 PR tree-optimization/57157
19936 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
19937 the (-Y) & (B - 1) variant if OP is |.
19938 * expmed.c (expand_shift_1): For rotations by const0_rtx just
19939 return shifted. Use (-op1) & (prec - 1) as other_amount
19940 instead of prec - op1.
19942 2013-05-13 Martin Jambor <mjambor@suse.cz>
19944 PR middle-end/42371
19945 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
19946 (ipa_constant_data): New type.
19947 (ipa_jump_func): Use ipa_constant_data to hold information about
19948 constant jump functions.
19949 (ipa_get_jf_constant): Adjust to jump function type changes.
19950 (ipa_get_jf_constant_rdesc): New function.
19951 (ipa_param_descriptor): New field controlled_uses.
19952 (ipa_get_controlled_uses): New function.
19953 (ipa_set_controlled_uses): Likewise.
19954 * ipa-ref.h (ipa_find_reference): Declare.
19955 * ipa-prop.c (ipa_cst_ref_desc): New type.
19956 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
19958 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
19959 New parameter cs. Adjust all callers.
19960 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
19961 (remove_described_reference): New function.
19962 (jfunc_rdesc_usable): Likewise.
19963 (try_make_edge_direct_simple_call): Decrement controlled use count,
19964 attempt to remove reference if it hits zero.
19965 (combine_controlled_uses_counters): New function.
19966 (propagate_controlled_uses): Likewise.
19967 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
19968 (ipa_edge_duplication_hook): Duplicate reference descriptions.
19969 (ipa_print_node_params): Print described use counter.
19970 (ipa_write_jump_function): Adjust to jump function type changes.
19971 (ipa_read_jump_function): New parameter CS, pass it to
19972 ipa_set_jf_constant. Adjust caller.
19973 (ipa_write_node_info): Stream controlled use count
19974 (ipa_read_node_info): Likewise.
19975 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
19977 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
19978 count. Remove cloning-added reference if it reaches zero.
19979 * ipa-ref.c (ipa_find_reference): New function.
19981 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19983 * config/i386/i386.c (processor_target_table): Modified default
19984 alignment values for AMD BD and BT architectures.
19986 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
19988 * tree-vect-generic.c (uniform_vector_p): Move ...
19989 * tree.c (uniform_vector_p): ... here.
19990 * tree.h (uniform_vector_p): Declare it.
19991 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
19994 2013-05-13 Jakub Jelinek <jakub@redhat.com>
19996 PR tree-optimization/57230
19997 * tree-ssa-strlen.c (handle_char_store): Record length for
19998 array store from STRING_CST.
20000 PR tree-optimization/57230
20001 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
20004 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
20006 * config/epiphany/epiphany.c (epiphany_init): Check size of
20007 NUM_MODES_FOR_MODE_SWITCHING.
20008 (epiphany_expand_prologue):
20009 Remove CONFIG_REGNUM initial value handling code.
20010 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
20011 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
20012 (emit_set_fp_mode, epiphany_mode_after): Likewise.
20013 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
20014 Don't return 1 for FP_MODE_NONE.
20015 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
20016 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
20017 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
20018 * config/epiphany/epiphany.md (save_config): New pattern.
20020 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
20022 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
20024 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
20026 * config/i386/i386.md (memory): Handle sseishft1.
20027 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
20028 (*vec_extractv2di_1): Ditto.
20030 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
20032 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
20035 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
20037 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
20038 Add mthumb/march=armv7-a multilib.
20039 Add mthumb/march=armv7-r multilib.
20040 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
20042 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
20044 * config/v850/t-rtems: Add more multilibs.
20046 2013-05-10 Richard Biener <rguenther@suse.de>
20048 PR tree-optimization/57214
20049 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
20050 not propagate from SSA names that occur in abnormal PHI nodes.
20052 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
20054 * stor-layout.c (element_precision): New function.
20055 * machmode.h (element_precision): Declare it.
20056 * tree.c (build_minus_one_cst): New function.
20057 (element_precision): Likewise.
20058 * tree.h (build_minus_one_cst): Declare new function.
20059 (element_precision): Likewise.
20060 * fold-const.c (operand_equal_p): Use element_precision.
20061 (fold_binary_loc): Handle vector types.
20062 * convert.c (convert_to_integer): Use element_precision.
20063 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
20066 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
20068 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
20069 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
20070 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
20071 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20072 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
20073 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
20074 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
20075 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20076 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
20077 (Uuw8): New constraints.
20078 (Usb4): Move into alphabetical order.
20079 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
20080 (sd8_operand, ub8_operand, uw8_operand): New predicates.
20081 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
20082 previously unnamed patterns.
20083 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
20084 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
20085 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
20086 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
20087 of set_attr_alternative/if_then_else. Use extended_mips16 instead
20088 of specific lengths.
20090 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20092 * config/i386/i386.md (rotateinv): New code attr.
20093 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20094 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20095 roll $31, %eax, etc.
20097 PR tree-optimization/45216
20098 PR tree-optimization/57157
20099 * tree-ssa-forwprop.c (simplify_rotate): New function.
20100 (ssa_forward_propagate_and_combine): Call it.
20102 2013-05-10 Richard Biener <rguenther@suse.de>
20104 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
20105 disable peeling when we version for aliasing.
20106 (vector_alignment_reachable_p): Honor explicit user alignment.
20107 (vect_supportable_dr_alignment): Likewise.
20108 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
20109 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
20110 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
20111 then peeling to arrange for the cost-model check to come first.
20113 2013-05-10 Alan Modra <amodra@gmail.com>
20115 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
20116 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
20117 * configure: Regenerate.
20119 2013-05-10 Alan Modra <amodra@gmail.com>
20122 * varasm.c (default_elf_select_section): Move !DECL_P check..
20123 (get_named_section): ..to here before calling get_section_name.
20125 (default_section_type_flags): Add DECL_P check.
20126 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
20127 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
20129 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
20131 * config/epiphany/epiphany.c (epiphany_expand_prologue):
20132 When using gen_stack_adjust_str with a register offset, add a
20133 REG_FRAME_RELATED_EXPR note.
20135 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
20137 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
20138 (*vec_extractv4si_zext_mem): Ditto.
20139 (*vec_extractv2di): Add 0->x and x->x alternatives.
20140 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
20141 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
20143 2013-05-09 Jason Merrill <jason@redhat.com>
20145 N3639 C++1y VLA support
20146 * gimplify.c (gimplify_vla_decl): Don't touch an existing
20149 * tree.c (build_constructor_va): New.
20150 * tree.h: Declare it.
20152 2013-05-09 Martin Jambor <mjambor@suse.cz>
20155 * gimple-fold.c (canonicalize_constructor_val): Call
20156 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
20158 2013-05-09 Jan Hubicka <jh@suse.cz>
20159 Richard Biener <rguenther@suse.de>
20162 * symtab.c (symtab_make_decl_local): Do not add private names.
20164 2013-05-09 Jan Hubicka <jh@suse.cz>
20167 * symtab.c (insert_to_assembler_name_hash): Handle clones.
20168 (unlink_from_assembler_name_hash): Likewise.
20169 (symtab_prevail_in_asm_name_hash, symtab_register_node,
20170 symtab_unregister_node, symtab_initialize_asm_name_hash,
20171 change_decl_assembler_name): Update.
20173 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
20175 * config/aarch64/aarch64.md: New movtf split.
20176 (*movtf_aarch64): Update.
20177 (aarch64_movdi_tilow): Handle TF modes and rename to
20178 aarch64_movdi_<mode>low.
20179 (aarch64_movdi_tihigh): Handle TF modes and rename to
20180 aarch64_movdi_<mode>high
20181 (aarch64_movtihigh_di): Handle TF modes and rename to
20182 aarch64_mov<mode>high_di
20183 (aarch64_movtilow_di): Handle TF modes and rename to
20184 aarch64_mov<mode>low_di
20185 (aarch64_movtilow_tilow): Remove spurious whitespace.
20186 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
20188 (aarch64_print_operand): Update.
20190 2013-05-09 Alan Modra <amodra@gmail.com>
20192 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
20194 * configure: Regenerate.
20196 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20198 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
20199 splitter preparation statements.
20200 * config/i386/sse.md (*vec_extract* splitters): Ditto.
20201 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
20204 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20206 * gimple-ssa-strength-reduction.c (count_candidates): Change
20207 return value to int.
20208 (analyze_candidates_and_replace): Change type of length to int.
20210 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20212 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
20213 (*vec_extract<mode>): Use VI12_128 mode iterator.
20214 (*vec_extract<mode>_mem): Ditto.
20215 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
20218 2013-05-08 Diego Novillo <dnovillo@google.com>
20223 2012-08-17 Diego Novillo <dnovillo@google.com>
20226 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
20227 * config.in: Regenerate.
20228 * configure: Regenerate.
20229 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
20231 2013-05-08 Jan Hubicka <jh@suse.cz>
20234 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
20235 * cgraph.h (symtab_node_base): Add unique_name.
20236 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20237 input_overwrite_node, input_varpool_node): Stream unique_name.
20238 * cgraphclones.c (cgraph_create_virtual_clone,
20239 cgraph_function_versioning): Set unique_name.
20240 * ipa.c (function_and_variable_visibility): Set unique_name.
20242 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20244 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
20245 (alloc_cand_and_find_basis): Restrict conditional candidate
20246 processing to CAND_MULTs.
20248 2013-05-08 Jan Hubicka <jh@suse.cz>
20251 lto-symtab.c (lto_symtab_symbol_p): New function.
20252 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
20253 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
20254 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
20255 Skip static symbols.
20257 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
20259 PR tree-optimization/57200
20260 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
20261 Only call inform if the preceding warning_at returns true.
20263 2013-05-07 Han Shen <shenhan@google.com>
20265 * cfgexpand.c (record_or_union_type_has_array_p): New function.
20266 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
20267 * common.opt (fstack-protector-strong): New option.
20268 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
20269 * doc/invoke.texi (Optimization Options): Document
20270 "-fstack-protector-strong".
20271 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
20273 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20275 * config/mips/mips.c (mips_machine_reorg2): Return 0.
20277 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
20279 * ira.c (update_equiv_regs): Add insn having equiv memory even if
20280 it is not lhs of the insn.
20281 (setup_reg_equiv): Remove insn having equiv memory which it is not
20283 * lra-constraints.c (process_address): Try to improve generation
20284 code for address base + disp.
20285 (lra_constraints): Make correct the code for checking insn setting
20286 up backward equivalence. Remove insn only if it is in the init
20288 * lra-eliminations.c (update_reg_eliminate): Change return value.
20289 (lra_eliminate): Use the result.
20291 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
20293 * config/i386/sse.md (ssescalarnummask): New mode attribute.
20294 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
20295 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
20296 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
20297 register target operands.
20298 (*vec_extractv8hi_sse2): New pattern.
20299 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
20300 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
20301 (*vec_extract<mode>_mem): New insn and split pattern.
20303 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
20305 * config/arm/arm.c (arm_asan_shadow_offset): New function.
20306 (TARGET_ASAN_SHADOW_OFFSET): Define.
20307 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
20308 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
20310 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20312 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
20313 (incr_vec_index): Return -1 if increment not found.
20314 (create_add_on_incoming_edge): Assert if increment not found.
20315 (record_increment): Limit number of increments recorded.
20316 (all_phi_incrs_profitable): Return false if an increment not found.
20317 (replace_profitable_candidates): Don't process increments that were
20319 (analyze_candidates_and_replace): Limit size of incr_vec.
20321 2013-05-07 Richard Biener <rguenther@suse.de>
20323 * calls.c (special_function_p): setjmp-like functions are leaf.
20324 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
20325 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
20327 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
20329 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
20330 (aarch64_simd_mov<mode>): New expander.
20331 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
20332 (aarch64_simd_mov_to_<mode>high): Likewise.
20333 (aarch64_simd_mov_from_<mode>low): Likewise.
20334 (aarch64_simd_mov_from_<mode>high): Likewise.
20335 (aarch64_dup_lane<mode>): Update.
20336 (aarch64_dup_lanedi): New instruction pattern.
20337 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
20338 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
20340 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20342 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
20343 (replace_mult_candidate): Remove unnecessary argument; remove
20344 unnecessary parameter from call to introduce_cast_before_cand.
20345 (replace_unconditional_candidate): Remove unnecessary parameter
20346 from call to replace_mult_candidate.
20347 (replace_conditional_candidate): Likewise.
20348 (insert_initializers): Use make_temp_ssa_name.
20349 (introduce_cast_before_cand): Remove unnecessary argument; use
20350 make_temp_ssa_name.
20351 (replace_one_candidate): Remove unnecessary argument; remove
20352 unnecessary parameter from calls to introduce_cast_before_cand.
20353 (replace_profitable_candidates): Remove unnecessary parameters
20354 from calls to replace_one_candidate.
20356 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20358 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
20359 phi def as possibly hiding a basis for a CAND_ADD whose operands
20360 have been commuted in the analysis.
20361 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
20363 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20365 * config/aarch64/aarch64.md
20366 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
20367 shift value between 0-4.
20369 2013-05-07 Richard Biener <rguenther@suse.de>
20371 * double-int.h (rshift): New overload.
20372 * double-int.c (rshift): New function.
20373 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
20374 (create_reference_ops_from_ref): Remove.
20375 (vn_reference_insert): Use shared ops for constructing the
20376 reference and copy it.
20378 2013-05-07 Richard Biener <rguenther@suse.de>
20380 PR middle-end/57190
20381 * tree-eh.c (sink_clobbers): Properly propagate
20382 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
20384 2013-05-07 Jakub Jelinek <jakub@redhat.com>
20386 PR tree-optimization/57149
20387 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
20388 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
20389 collect_phi_def_edges, execute_late_warn_uninitialized): Use
20390 uninit_undefined_value_p instead of ssa_undefined_value_p.
20393 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
20394 for modifier == EXPAND_INITIALIZER.
20396 2013-05-07 Anton Blanchard <anton@samba.org>
20398 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
20399 for powerpc64 little endian.
20400 * configure: Regenerate.
20402 2013-05-06 Graham Stott <grahams@btinternet.com>
20404 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
20405 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
20406 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
20407 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
20409 2013-05-06 Graham Stott <grahams@btinternet.com>
20411 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
20412 codes which allow non-lvalues.
20414 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
20416 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
20417 components are all 1s.
20418 (integer_minus_onep): New function.
20419 * tree.h (integer_minus_onep): Declare it.
20420 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
20421 integer_minus_onep instead of integer_all_onesp.
20423 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20426 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
20427 variations of these patterns.
20429 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20431 * config/i386/i386.md (isa): Add x64_sse4 member.
20432 (enabled): Handle x64_sse4.
20433 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
20434 instruction for 64bit SSE4_1 targets. Update insn attributes.
20435 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
20436 instruction for SSE4_1 targets. Update insn attributes.
20437 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20438 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
20439 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
20441 (*vec_extractv4si): Rename from *sse4_1_pextrd.
20442 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
20443 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
20445 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20448 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
20450 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
20452 * final.c (do_assembler_dialects): Don't handle curly braces and
20453 vertical bar escaped by % as dialect delimiters.
20454 (output_asm_insn): Print curly braces and vertical bar if escaped
20455 by % and ASSEMBLER_DIALECT defined.
20456 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
20457 * doc/tm.texi: Regenerated.
20459 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20461 * config/mips/mips.c: Include tree-pass.h.
20462 (mips_reorg): Split in pre- and post-dbr_schedule parts.
20463 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
20464 (pass_mips_machine_reorg2): New machine specific pass.
20465 (insert_pass_mips_machine_reorg2): New pass plugin definition.
20466 (mips_option_override): Register the new pass.
20467 * rtl.h (cleanup_barriers): Remove prototype.
20468 (dbr_schedule): Likewise.
20469 * jump.c (cleanup_barriers): Make static.
20470 * reorg.c (dbr_schedule): Likewise.
20472 2013-05-06 Richard Biener <rguenther@suse.de>
20474 PR tree-optimization/57185
20475 * tree-parloops.c (add_field_for_reduction): Handle anonymous
20476 SSA names properly.
20478 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20481 * config/i386/i386.c (add_parameter_dependencies): Add dependence
20482 between "first_arg" and "insn", not "last" and "insn".
20484 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20486 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
20487 (find_candidates_in_block): Re-enable slsr_process_phi.
20488 (create_phi_basis): Fix double counting of candidate adjustment.
20490 2013-05-06 Richard Biener <rguenther@suse.de>
20492 PR middle-end/57147
20493 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
20494 the edge is also fallthru, preserve it and just clear the
20496 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
20497 also complex, preserve that and just clear the fallthru flag.
20498 * tree-inline.c (update_ssa_across_abnormal_edges): Also
20499 update virtual operands.
20501 2013-05-06 Alan Modra <amodra@gmail.com>
20503 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
20504 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
20505 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
20506 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
20507 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20508 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
20510 2013-05-06 Alan Modra <amodra@gmail.com>
20512 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
20513 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
20514 (DEFAULT_ASM_ENDIAN): Define.
20515 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20516 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
20517 Update -K PIC clause from sysv4.h.
20518 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
20519 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
20521 2013-05-06 Alan Modra <amodra@gmail.com>
20523 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
20524 twice for little-endian.
20525 (ashrdi3_no_power, ashrdi3): Support little-endian.
20527 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20530 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
20531 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
20533 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
20534 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
20535 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
20538 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
20539 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20541 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
20542 * config/pa/pa.opt: Make mbig-switch a no-op.
20543 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
20544 (CASE_VECTOR_MODE): Always return SImode.
20545 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
20546 for the !TARGET_BIG_SWITCH case.
20547 * config/pa/pa-linux.h: Likewise.
20548 * config/pa/pa-openbsd.h: Likewise.
20549 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
20550 * config/pa/pa.md (short_jump): Remove define_insn.
20551 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
20552 (casesi0): Remove define_insn.
20553 (type): Remove btable_branch.
20554 (pa_combine_type): Likewise.
20555 (in_nullified_branch_delay): Likewise.
20556 (in_call_delay): Likewise.
20557 (define_delay): Likewise.
20558 (define_insn_reservation "Z3"): Likewise.
20559 (define_insn_reservation "Z4"): Likewise.
20560 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
20561 (pa_adjust_insn_length): Remove adjustment for btable branches.
20562 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
20565 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
20567 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20568 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
20569 Add m->r,x alternatives.
20570 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
20571 splitters using SWI48x mode iterator.
20572 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
20573 TARGET_64BIT. Add m->x alternative.
20574 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
20575 Add o->x alternative. Enable for TARGET_SSE.
20576 (sse_storeq): Remove expander.
20577 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
20578 with memory input operand.
20579 (*vec_extractv2di_1 splitter): New.
20580 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
20581 * config/i386/i386.md (ssevecmodelower): New mode attribute.
20583 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20585 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
20586 (INT_LOWPART): Delete.
20587 (extract_MB): Adjust.
20588 (extract_ME): Adjust.
20589 (print_operand): Adjust.
20591 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20593 * config/rs6000/predicates.md (reg_or_add_cint_operand,
20594 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20595 (reg_or_logical_cint_operand, easy_fp_constant,
20596 logical_const_operand): Delete "CONST_DOUBLE" case.
20597 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
20598 "HOST_BITS_PER_WIDE_INT == 64" test.
20599 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
20600 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
20601 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
20602 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
20604 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
20605 CONST_DOUBLE DImode/VOIDmode case.
20606 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
20607 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20608 CONST_DOUBLE VOIDmode case.
20609 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20610 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
20611 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
20612 Delete CONST_DOUBLE case.
20613 (splitters for mov FMOVE64 const_double): Delete
20614 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20615 "HOST_BITS_PER_WIDE_INT >= 64" test.
20616 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
20618 (mov DI const_double): Delete.
20620 2013-05-04 Jakub Jelinek <jakub@redhat.com>
20622 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
20623 on op shows all bits zero in mode of a lowpart subreg, return zero.
20625 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
20628 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
20629 to save TFmode registers and DImode to save TImode registers for
20630 caller save operations.
20631 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
20632 mark being partially clobbered since they only use the first
20635 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
20636 and TDmode only use the upper 64-bits of each VSX register.
20638 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20640 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
20641 (find_candidates_in_block): Disable slsr_process_phi.
20643 2013-05-03 Guozhi Wei <carrot@google.com>
20645 * coverage.c (coverage_obj_init): Move the construction of gcov
20647 (build_init_ctor): ... here.
20649 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20651 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
20652 (slsr_cand_d): Redefine def_phi.
20653 (stride_status, phi_adjust_status, count_phis_status): New enums.
20654 (find_phi_def): New.
20655 (find_basis_for_base_expr): New.
20656 (find_basis_for_candidate): Handle hidden bases.
20657 (alloc_cand_and_find_basis): Handle phi candidates.
20658 (slsr_process_phi): New.
20659 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
20660 (create_mul_imm_cand): Likewise.
20661 (create_add_ssa_cand): Exclude phi base candidates.
20662 (create_add_imm_cand): Likewise.
20663 (slsr_process_cast): Likewise.
20664 (slsr_process_copy): Likewise.
20665 (find_candidates_in_block): Handle phi candidates.
20666 (dump_candidate): Likewise.
20667 (unconditional_cands): Delete.
20668 (unconditional_cands_with_known_stride_p): Delete.
20669 (phi_dependent_cand_p): New.
20670 (cand_increment): Handle phi-dependent candidates.
20671 (replace_dependent): Delete.
20672 (replace_mult_candidate): New.
20673 (replace_unconditional_candidate): New.
20674 (incr_vec_index): Move to avoid forward reference.
20675 (create_add_on_incoming_edge): New.
20676 (create_phi_basis): New.
20677 (replace_dependents): Delete.
20678 (replace_conditional_candidate): New.
20679 (phi_add_costs): New.
20680 (replace_uncond_cands_and_profitable_phis): New.
20681 (record_increment): Handle phi adjustments.
20682 (record_phi_increments): New.
20683 (record_increments): Handle phi adjustments.
20684 (phi_incr_cost): New.
20685 (lowest_cost_path): Handle phis.
20686 (total_savings): Likewise.
20687 (analyze_increments): Likewise.
20688 (ncd_with_phi): New.
20689 (ncd_of_cand_and_phis): New.
20690 (nearest_common_dominator_for_cands): Handle phi increments.
20691 (all_phi_incrs_profitable): New.
20692 (replace_profitable_candidates): Handle phi-dependent candidates.
20693 (analyze_candidates_and_replace): Likewise.
20695 2013-05-03 Teresa Johnson <tejohnson@google.com>
20698 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
20699 do not exceed REG_BR_PROB_BASE.
20701 2013-05-03 Jeff Law <law@redhat.com>
20703 PR tree-optimization/57144
20704 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
20705 operand of the condition will bit into the new type when eliminating
20706 a cast feeding a condition.
20708 2013-05-03 Jakub Jelinek <jakub@redhat.com>
20710 PR rtl-optimization/57130
20711 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
20712 of COMPARE as in_code to the recursive call if needed.
20714 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
20716 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
20717 (enabled): Handle new members.
20718 * config/i386/sse.md (*vec_concatv2si): Merge from
20719 *vec_concatv2si_sse2 and vec_concatv2si_sse.
20720 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
20722 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
20724 PR tree-optimization/57027
20725 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
20726 for fnms opportunity, check we got the prerequisite kind
20727 of tree / gimple before using accessor functions.
20729 2013-05-03 Richard Biener <rguenther@suse.de>
20731 * double-int.h (lshift): New overload without precision
20732 and arith argument.
20733 (operator *=, operator +=, operator -=): Move ...
20734 * double-int.c (operator *=, operator +=, operator -=): ... here
20735 and implement more efficiently.
20736 (mul_double_with_sign): Remove.
20737 (lshift_double): Adjust to take unsinged shift argument, push
20738 dispatching code to callers.
20739 (mul_double_wide_with_sign): Add early out for callers that
20740 are not interested in high parts or overflow.
20741 (lshift): New function.
20742 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
20743 dispatch code here.
20744 (lrotate, rrotate): Use logical shifts.
20745 * expr.c (get_inner_reference): Use lshift.
20746 * fixed-value.c (do_fixed_divide): Likewise.
20747 * tree-dfa.c (get_ref_base_and_extent): Likewise.
20748 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
20749 (indirect_refs_may_alias_p): Likewise.
20750 (stmt_kills_ref_p_1): Likewise.
20752 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20754 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
20756 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20758 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
20759 scalar form of FABD instruction.
20761 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
20763 * lra-constraints.c (process_alt_operands): Add checking alt
20764 number to choose the best alternative.
20766 2013-05-02 Richard Biener <rguenther@suse.de>
20768 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
20769 bitmap and its handling.
20770 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
20772 2013-05-02 Richard Biener <rguenther@suse.de>
20774 PR middle-end/57140
20775 * tree-inline.c (copy_loops): Properly handle removed loops.
20776 (copy_cfg_body): Mark destination loops for fixup if source
20777 loops needed fixup.
20779 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
20782 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
20783 generating simple_return for naked functions.
20785 2013-05-02 Martin Jambor <mjambor@suse.cz>
20787 PR middle-end/56988
20788 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
20789 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
20791 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
20792 ipa_agg_replacement_value structures.
20793 (known_aggs_to_agg_replacement_list): Likewise.
20794 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
20795 (read_agg_replacement_chain): Likewise.
20796 (ipcp_transform_function): Also check that by_ref flags match.
20798 2013-05-02 Richard Biener <rguenther@suse.de>
20800 * graphds.h (struct graph): Add obstack member.
20801 * graphds.c (new_graph): Initialize obstack and allocate
20803 (add_edge): Allocate edge from the obstack.
20804 (free_graph): Free the obstack instead of all edges and vertices.
20806 2013-05-02 Teresa Johnson <tejohnson@google.com>
20808 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
20810 * cfg.c (update_bb_profile_for_threading): Ditto.
20811 * tree-inline.c (copy_bb): Ditto.
20812 (copy_edges_for_bb): Ditto.
20813 (initialize_cfun): Ditto.
20814 (copy_cfg_body): Ditto.
20815 (expand_call_inline): Ditto.
20816 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
20817 (estimate_node_size_and_time): Ditto.
20818 (inline_merge_summary): Ditto.
20819 * cgraphclones.c (cgraph_clone_edge): Ditto.
20820 (cgraph_clone_node): Ditto.
20821 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20822 (compute_trg_info): Ditto.
20824 2013-05-02 Ian Bolton <ian.bolton@arm.com>
20826 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
20827 S reg when fp attribute set.
20828 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
20830 2013-05-02 Ian Bolton <ian.bolton@arm.com>
20832 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
20834 (*and_one_cmplsi3_compare0_uxtw): Likewise.
20835 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
20836 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
20838 2013-05-02 Richard Biener <rguenther@suse.de>
20840 * tree-scalar-evolution.c (scev_info_hasher): Remove.
20841 (struct instantiate_cache_entry): New type.
20842 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
20843 (struct instantiate_cache_type): New type.
20844 (set_instantiated_value, get_instantiated_value): Remove.
20845 (get_instantiated_value_entry): New function.
20846 (instantiate_scev_name): Use the new cache and adjust.
20847 (instantiate_scev_poly): Adjust.
20848 (instantiate_scev_binary): Likewise.
20849 (instantiate_array_ref): Likewise.
20850 (instantiate_scev_convert): Likewise.
20851 (instantiate_scev_not): Likewise.
20852 (instantiate_scev_3): Likewise.
20853 (instantiate_scev_2): Likewise.
20854 (instantiate_scev_r): Likewise.
20855 (instantiate_scev): Likewise.
20856 (resolve_mixers): Likewise.
20858 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
20861 * lra-constraints.c (best_small_class_operands_num): Remove.
20862 (process_alt_operands): Remove small_class_operands_num. Take
20863 small classes operands into losers and only if the operand is not
20864 matched. Modify debugging output.
20865 (curr_insn_transform): Remove best_small_class_operands_num.
20868 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20870 * config/aarch64/aarch64-builtins.c
20871 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
20872 * config/aarch64/aarch64-simd-builtins.def
20873 (reduc_splus_): Add new modes.
20874 (reduc_uplus_): New.
20875 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
20876 (reduc_uplus_v4sf): Likewise.
20877 (reduc_splus_v4sf): Likewise.
20878 (aarch64_addv<mode>): Likewise.
20879 (reduc_uplus_<mode>): Likewise.
20880 (reduc_splus_<mode>): Likewise.
20881 (aarch64_addvv2di): Likewise.
20882 (reduc_uplus_v2di): Likewise.
20883 (reduc_splus_v2di): Likewise.
20884 (aarch64_addvv2si): Likewise.
20885 (reduc_uplus_v2si): Likewise.
20886 (reduc_splus_v2si): Likewise.
20887 (reduc_<sur>plus_<mode>): New.
20888 (reduc_<sur>plus_v2di): Likewise.
20889 (reduc_<sur>plus_v2si): Likewise.
20890 (reduc_<sur>plus_v4sf): Likewise.
20891 (aarch64_addpv4sf): Likewise.
20892 * config/aarch64/arm_neon.h
20893 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
20894 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
20895 add UNSPEC_SADDV, UNSPEC_UADDV.
20897 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
20899 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20901 * config/aarch64/arm_neon.h
20902 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
20904 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20906 * config/aarch64/aarch64-builtins
20907 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
20909 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20911 * config/aarch64/aarch64-simd-builtins.def
20912 (reduc_smax_): New.
20913 (reduc_smin_): Likewise.
20914 (reduc_umax_): Likewise.
20915 (reduc_umin_): Likewise.
20916 (reduc_smax_nan_): Likewise.
20917 (reduc_smin_nan_): Likewise.
20920 (smax): Update for V2SF, V4SF and V2DF modes.
20923 (smin_nan): Likewise.
20924 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
20925 (<su><maxmin><mode>3): ...This, refactor.
20926 (s<maxmin><mode>3): New.
20927 (<maxmin_uns><mode>3): Likewise.
20928 (reduc_<maxmin_uns>_<mode>): Refactor.
20929 (reduc_<maxmin_uns>_v4sf): Likewise.
20930 (reduc_<maxmin_uns>_v2si): Likewise.
20931 (aarch64_<fmaxmin><mode>: Remove.
20932 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
20934 (vmin<q>_f<32,64>): Likewise.
20935 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
20937 (su): Add mappings for smax, smin, umax, umin.
20939 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
20940 (FMAXMIN): Rename as...
20941 (FMAXMIN_UNS): ...This.
20943 (fmaxminv): Likewise.
20944 (fmaxmin): Likewise.
20946 (maxmin_uns_op): Likewise.
20948 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20950 * config/aarch64/arm_neon.h
20951 (vac<ge, gt><sd>_f<32, 64>): Rename to...
20952 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
20953 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
20955 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20957 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
20958 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
20960 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20962 * config/aarch64/aarch64-simd.md
20963 (vcond<mode>_internal): Handle special cases for constant masks.
20964 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
20965 (vcondu<mode><mode>): Likewise.
20966 (vcond<v_cmp_result><mode>): New.
20968 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20970 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
20971 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
20972 * config/aarch64/aarch64-simd-builtins.def
20973 (cmeq): Update to BUILTIN_VALLDI.
20978 * config/aarch64/arm_neon.h
20979 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
20980 to builtins or C as appropriate.
20982 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20984 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
20986 (cmhi): Rename to...
20988 * config/aarch64/aarch64-simd.md
20989 (simd_mode): Add SF.
20990 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
20991 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
20992 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
20993 (cstore<mode>_neg): ...This.
20994 * config/aarch64/iterators.md
20996 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
20997 (COMPARISONS): New.
20998 (UCOMPARISONS): Likewise.
20999 (optab): Add missing comparisons.
21005 (VCMP_S): Likewise.
21006 (VCMP_U): Likewise.
21007 (V_cmp_result): Add DF, SF modes.
21008 (v_cmp_result): Likewise.
21010 (vmtype): Likewise.
21011 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
21013 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
21015 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
21016 define_insn to define_insn_and_split.
21017 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
21018 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
21019 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
21020 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
21021 (thumb2_negscc): Likewise.
21023 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21025 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
21027 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21029 * config/arm/thumb2.md: Remove trailing whitespaces.
21031 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21033 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
21034 Use gen_int_mode rather than GEN_INT.
21036 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
21038 * value-prof.c (stream_in_histogram_value): Remove the strayed
21041 2013-04-30 Richard Biener <rguenther@suse.de>
21043 PR middle-end/57122
21044 * cfghooks.c (split_edge): Properly check for the loop latch edge.
21046 2013-04-30 Richard Biener <rguenther@suse.de>
21048 PR middle-end/57107
21049 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
21051 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
21053 PR rtl-optimization/56957
21054 PR rtl-optimization/57105
21055 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
21056 variable. Use just INSN_UID for determining whether an insn
21057 should be only disconnected from the insn stream.
21058 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
21060 2013-04-30 Jakub Jelinek <jakub@redhat.com>
21062 PR tree-optimization/57104
21063 * tsan.c (instrument_expr): Don't instrument accesses to
21064 DECL_HARD_REGISTER VAR_DECLs.
21066 2013-04-30 Richard Biener <rguenther@suse.de>
21068 * function.h (loops_for_fn): New inline function.
21069 (set_loops_for_fn): Likewise.
21070 * cfgloop.h (place_new_loop): Add struct function parameter.
21071 (get_loop): Likewise.
21072 (get_loops): Likewise.
21073 (number_of_loops): Likewise.
21074 (fel_next): Adjust.
21075 (fel_init): Likewise.
21076 * cfg.c (get_loop_copy): Adjust.
21077 * cfgloop.c (flow_loops_dump): Likewise.
21078 (record_loop_exits): Likewise.
21079 (verify_loop_structure): Likewise.
21080 * cfgloopanal.c (mark_irreducible_loops): Likewise.
21081 (estimate_reg_pressure_cost): Likewise.
21082 (mark_loop_exit_edges): Likewise.
21083 * cfgloopmanip.c (place_new_loop): Likewise.
21084 (add_loop): Likewise.
21085 (duplicate_loop): Likewise.
21086 * graph.c (draw_cfg_nodes): Likewise.
21087 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
21088 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
21089 (extract_affine_chrec): Likewise.
21090 (build_scop_iteration_domain): Likewise.
21091 * graphite.c (graphite_initialize): Likewise.
21092 * ira-build.c (create_loop_tree_nodes): Likewise.
21093 (more_one_region_p): Likewise.
21094 (rebuild_regno_allocno_maps): Likewise.
21095 (mark_loops_for_removal): Likewise.
21096 (mark_all_loops_for_removal): Likewise.
21097 (remove_unnecessary_regions): Likewise.
21098 (ira_build): Likewise.
21099 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21100 * loop-init.c (fix_loop_structure): Likewise.
21101 (gate_rtl_move_loop_invariants): Likewise.
21102 (gate_rtl_unswitch): Likewise.
21103 (gate_rtl_unroll_and_peel_loops): Likewise.
21104 (rtl_doloop): Likewise.
21105 * lto-streamer-in.c (input_cfg): Likewise.
21106 * lto-streamer-out.c (output_cfg): Likewise.
21107 * modulo-sched.c (sms_schedule): Likewise.
21108 * predict.c (tree_estimate_probability): Likewise.
21109 (tree_estimate_probability_driver): Likewise.
21110 (estimate_loops): Likewise.
21111 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
21112 (move_sese_region_to_fn): Likewise.
21113 (debug_loop_num): Likewise.
21114 * tree-chrec.c (chrec_evaluate): Likewise.
21115 (hide_evolution_in_other_loops_than_loop): Likewise.
21116 (chrec_component_in_loop_num): Likewise.
21117 (reset_evolution_in_loop): Likewise.
21118 (evolution_function_is_invariant_rec_p): Likewise.
21119 * tree-if-conv.c (main_tree_if_conversion): Likewise.
21120 * tree-inline.c (copy_loops): Likewise.
21121 (copy_cfg_body): Likewise.
21122 (tree_function_versioning): Likewise.
21123 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
21124 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
21126 (add_to_evolution_1): Likewise.
21127 (scev_const_prop): Likewise.
21128 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
21129 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
21130 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
21131 (tree_ssa_lim_initialize): Likewise.
21132 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
21133 (verify_loop_closed_ssa): Likewise.
21134 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
21135 (tree_ssa_loop_im): Likewise.
21136 (tree_ssa_loop_unswitch): Likewise.
21137 (tree_vectorize): Likewise.
21138 (check_data_deps): Likewise.
21139 (tree_ssa_loop_ivcanon): Likewise.
21140 (tree_ssa_loop_bounds): Likewise.
21141 (tree_complete_unroll): Likewise.
21142 (tree_complete_unroll_inner): Likewise.
21143 (tree_parallelize_loops): Likewise.
21144 (tree_ssa_loop_prefetch): Likewise.
21145 (tree_ssa_loop_ivopts): Likewise.
21146 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
21147 * tree-vectorizer.c (vectorize_loops): Likewise.
21149 2013-04-29 Mike Frysinger <vapier@gentoo.org>
21151 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
21152 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
21153 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
21154 with EABI_LINK_SPEC.
21156 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21159 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
21162 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
21165 * lra-constraints.c (process_alt_operands): Discourage a bit more
21166 using memory for pseudos. Print cost dump for alternatives.
21167 Modify cost values for conflicts with early clobbers.
21168 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
21170 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21173 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
21175 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21177 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
21178 from/to S register.
21179 (movdi_aarch64): Support LDR/STR from/to D register.
21181 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21183 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
21184 or higher by default.
21186 2013-04-29 Richard Biener <rguenther@suse.de>
21188 PR middle-end/57075
21189 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
21190 even if not adding abnormal edges for calls that can make
21193 2013-04-29 Richard Biener <rguenther@suse.de>
21195 PR middle-end/57103
21196 * tree-cfg.c (move_stmt_op): Fix condition under which to update
21198 (move_stmt_r): Remove redundant checking.
21200 2013-04-29 Teresa Johnson <tejohnson@google.com>
21203 * basic-block.h (apply_scale): New function.
21204 (apply_probability): Use apply_scale.
21205 * gimple-streamer-in.c (input_bb): Ditto.
21206 * lto-streamer-in.c (input_cfg): Ditto.
21207 * lto-cgraph.c (merge_profile_summaries): Ditto.
21208 * tree-optimize.c (execute_fixup_cfg): Ditto.
21209 * tree-inline.c (copy_bb): Update comment to use apply_scale.
21210 (copy_edges_for_bb): Ditto.
21211 (copy_cfg_body): Ditto.
21213 2013-04-29 Tom de Vries <tom@codesourcery.com>
21215 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
21216 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
21217 (tail_merge_optimize): Handle current_loops == NULL.
21219 2013-04-26 Jeff Law <law@redhat.com>
21221 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
21222 (simplify_cond_using_ranges): Generalize code to simplify
21223 COND_EXPRs where one argument is a constant and the other
21224 is an SSA_NAME created by an integral type conversion.
21226 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21228 * config/arm/arm.md (store_minmaxsi): Use only when
21229 optimize_insn_for_size_p.
21231 2013-04-29 Christian Bruel <christian.bruel@st.com>
21234 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
21236 2013-04-29 Richard Biener <rguenther@suse.de>
21238 PR middle-end/57089
21239 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
21240 loop tree make sure to schedule a fixup for the child as well.
21241 (expand_omp_for_generic): Properly add loops.
21242 (expand_omp_for_static_nochunk): Likewise.
21243 (expand_omp_for_static_chunk): Likewise.
21244 (expand_omp_for): For the degenerate case fixup loops.
21245 (expand_omp_sections): Fix default bb placement in loops.
21246 (expand_omp_atomic_pipeline): Properly add loops.
21248 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21250 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
21252 2013-04-29 Tom de Vries <tom@codesourcery.com>
21254 * tree-ssa-tail-merge.c: Update header comment.
21256 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21258 * config/aarch64/arm_neon.h
21259 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
21260 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
21261 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
21262 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
21263 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
21264 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
21265 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
21266 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
21268 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21270 * config/aarch64/aarch64-simd.md
21271 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
21272 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
21273 fix_trunc, fixuns_trunc.
21274 (ftrunc<VDQF:mode>2): New.
21275 * config/aarch64/iterators.md (optab): Add fix, fixuns.
21276 (fix_trunc_optab): New.
21278 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21280 * config/aarch64/aarch64-builtins.c
21281 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
21282 iceilf, lround, iroundf.
21284 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21287 * config/i386/i386.h (enum ix86_tune_indices)
21288 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
21289 New, split from X86_TUNE_INTER_UNIT_MOVES.
21290 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
21291 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
21292 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
21293 (TARGET_INTER_UNIT_MOVES): Remove.
21294 * config/i386/i386.c (initial_ix86_tune_features): Update.
21295 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
21296 (ix86_expand_convert_uns_didf_sse): Use
21297 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21298 (ix86_expand_vector_init_one_nonzero): Ditto.
21299 (ix86_expand_vector_init_interleave): Ditto.
21300 (inline_secondary_memory_needed): Return true for moves from SSE class
21301 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
21302 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
21303 * config/i386/constraints.md (Yi, Ym): Depend on
21304 TARGET_INTER_UNIT_MOVES_TO_VEC.
21305 (Yj, Yn): New constraints.
21306 * config/i386/i386.md (*movdi_internal): Change constraints of
21307 operand 1 from Yi to Yj and from Ym to Yn.
21308 (*movsi_internal): Ditto.
21309 (*movdf_internal): Ditto.
21310 (*movsf_internal): Ditto.
21311 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
21312 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21313 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
21314 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
21315 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
21316 * config/i386/sse.md (movdi_to_sse): Ditto.
21317 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
21318 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
21319 TARGET_INTER_UNIT_MOVES.
21320 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
21321 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
21322 instead of TARGET_INTER_UNIT_MOVES.
21323 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
21324 operand 1 from Yi to Yj and from Ym to Yn.
21326 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21328 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
21329 (float_truncate_hi_): Likewise.
21330 (float_extend_lo_): Likewise.
21331 (float_truncate_lo_): Likewise.
21332 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
21333 (aarch64_float_extend_lo_v2df): Likewise.
21334 (vec_unpacks_hi_v4sf): Likewise.
21335 (aarch64_float_truncate_lo_v2sf): Likewise.
21336 (aarch64_float_truncate_hi_v4sf): Likewise.
21337 (vec_pack_trunc_v2df): Likewise.
21338 (vec_pack_trunc_df): Likewise.
21340 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21342 * config/aarch64/aarch64-builtins.c
21343 (aarch64_fold_builtin): Fold float conversions.
21344 * config/aarch64/aarch64-simd-builtins.def
21345 (floatv2si, floatv4si, floatv2di): New.
21346 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
21347 * config/aarch64/aarch64-simd.md
21348 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
21349 * config/aarch64/iterators.md (FLOATUORS): New.
21350 (optab): Add float, floatuns.
21351 (su_optab): Likewise.
21353 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21355 * config/aarch64/aarch64-builtins.c
21356 (aarch64_builtin_vectorized_function): Use new names for
21358 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
21359 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
21360 (fcvtzu): Split as...
21361 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
21362 (fcvtas): Split as...
21363 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
21364 (fcvtau): Split as...
21365 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
21366 (fcvtps): Split as...
21367 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
21368 (fcvtpu): Split as...
21369 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
21370 (fcvtms): Split as...
21371 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
21372 (fcvtmu): Split as...
21373 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
21374 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
21375 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
21376 (lfrintnusf, lfrintnudf): Likewise.
21377 * config/aarch64/aarch64-simd.md
21378 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
21380 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
21381 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
21382 (fcvt_pattern): Likewise.
21384 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21386 * config/aarch64/aarch64-simd.md
21387 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
21388 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
21390 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21392 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
21393 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
21394 (vrnd<a,m,n,p>_f32): Implement using builtins.
21395 (vrnd<i,x><q>_f<32, 64>): New.
21397 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21399 * config/aarch64/aarch64-builtins.c
21400 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
21401 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
21402 (frintz): Rename to...
21404 (frintp): Rename to...
21406 (frintm): Rename to...
21408 (frinti): Rename to...
21409 (nearbyint): ...this.
21410 (frintx): Rename to...
21412 (frinta): Rename to...
21414 * config/aarch64/aarch64-simd.md
21415 (aarch64_frint<frint_suffix><mode>): Delete.
21416 (<frint_pattern><mode>2): Convert to insn.
21417 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
21418 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
21419 (frint_pattern): Likewise.
21420 (frint_suffix): Likewise.
21422 2013-04-29 Richard Biener <rguenther@suse.de>
21424 PR tree-optimization/57081
21425 * loop-init.c: Include tree-flow.h.
21426 (loop_optimizer_finalize): Free number of iteration estimates.
21427 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
21429 2013-04-29 Jakub Jelinek <jakub@redhat.com>
21431 PR tree-optimization/57083
21432 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
21433 non-singleton shift count range, zero extend low_bound for uns case.
21435 * config/i386/predicates.md (general_vector_operand): New predicate.
21436 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
21437 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
21438 if they aren't nonimmediate operands. If their original values
21439 satisfy const_vector_equal_evenodd_p, don't shift them.
21440 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
21441 predicates. For the SSE4.1 case force operands[{1,2}] into registers
21442 if not nonimmediate_operand.
21443 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
21444 instead of register_operand.
21445 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
21447 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
21449 * stor-layout.c (finalize_size_functions): Allocate a structure and
21450 reset cfun before dumping the functions.
21452 2013-04-27 Jakub Jelinek <jakub@redhat.com>
21454 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
21457 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
21458 use xop_pmacsdqh if uns_p.
21459 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
21460 the immediate rotate count.
21462 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
21464 * rtl.h (struct rtx_def): Add comment for field jump.
21465 (LRA_SUBREG_P): New macro.
21466 * recog.c (register_operand): Check LRA_SUBREG_P.
21467 * lra.c (lra): Add note at the end of RTL code. Align non-empty
21469 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
21470 (lra_final_code_change): Skip subreg change for operators.
21471 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
21472 if there are no operand changes.
21473 * lra-constraints.c (curr_insn_set): New.
21474 (match_reload): Set LRA_SUBREG_P.
21475 (emit_spill_move): Ditto.
21476 (check_and_process_move): Use curr_insn_set. Process only single
21477 set insns. Don't initialize sec_mem_p and change_p.
21478 (simplify_operand_subreg): Use LRA_SUBREG_P.
21479 (reg_in_class_p): New function.
21480 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
21481 of #ifdef. Add code to remove cycling.
21482 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
21483 non-null disp. Reload inner instead of disp when base and index
21484 are null. Try to put lo_sum into register.
21485 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
21486 (check_and_process_move): Move code for move cost check to
21487 simple_move_p. Remove equiv_substitution.
21488 (simple_move_p): New function.
21489 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
21490 curr_insn_set. Call check_and_process_move only for single set
21491 insns. Use the new function. Move call of check_and_process_move
21492 after operand equiv substitution and address process.
21494 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21497 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
21498 with nonlocal goto receivers or returns twice calls, ignore
21499 unininitialized values from abnormal edges to nl goto receiver
21500 or returns twice call.
21502 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21504 PR tree-optimization/57051
21505 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
21506 and VEC_RSHIFT_EXPR if shift count is a multiple of element
21509 2013-04-26 Richard Biener <rguenther@suse.de>
21511 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
21512 (expand_omp_taskreg): Likewise. Mark loops for fixup.
21513 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
21514 (fixup_loop_arrays_after_move): New function.
21515 (move_sese_region_to_fn): Properly outline the loop tree parts
21516 of the SESE region.
21518 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
21520 * config/i386/i386.md (type, unit): Fix long lines.
21522 2013-04-26 Richard Biener <rguenther@suse.de>
21524 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
21525 (lto-streamer-out.o): Likewise.
21526 * cfgloop.c (init_loops_structure): Export, add struct function
21527 argument and adjust.
21528 (flow_loops_find): Adjust.
21529 * cfgloop.h (enum loop_estimation): Add EST_LAST.
21530 (init_loops_structure): Declare.
21531 * lto-streamer-in.c: Include cfgloop.h.
21532 (input_cfg): Input the loop tree.
21533 * lto-streamer-out.c: Include cfgloop.h.
21534 (output_cfg): Output the loop tree.
21535 (output_struct_function_base): Do not drop PROP_loops.
21537 2013-03-26 Richard Biener <rguenther@suse.de>
21539 * tree-cfg.c (execute_build_cfg): Build the loop tree.
21540 (pass_build_cfg): Provide PROP_loops.
21541 (move_sese_region_to_fn): Remove loops that are outlined into fn
21543 * tree-inline.c: Include cfgloop.h.
21544 (initialize_cfun): Do not drop PROP_loops.
21545 (copy_loops): New function.
21546 (copy_cfg_body): Copy loop structure.
21547 (tree_function_versioning): Initialize destination loop tree.
21548 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
21549 (pass_parallelize_loops): Do IL verification.
21550 * loop-init.c (loop_optimizer_init): Fixup loops if required.
21551 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
21552 the CFG make sure we fixup loops as well.
21553 * tree-ssa-tail-merge.c: Include cfgloop.h.
21554 (replace_block_by): When merging loop latches mark loops for fixup.
21555 * lto-streamer-out.c (output_struct_function_base): Drop
21556 PROP_loops for now.
21557 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
21558 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
21559 * ipa-split.c: Include cfgloop.h.
21560 (split_function): Add the new return block to the loop tree root.
21561 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
21562 whether we have removed the forwarder block.
21563 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
21564 * cfgloop.h (place_new_loop): Declare.
21565 * cfgloopmanip.c (place_new_loop): Export.
21566 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
21567 (tree-switch-conversion.o): Likewise.
21568 (tree-complex.o): Likewise.
21569 (tree-inline.o): Likewise.
21570 (tree-ssa-tailmerge.o): Likewise.
21571 (ipa-split.o): Likewise.
21572 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
21573 (tree-ssa-copy.o): Likewise.
21574 * tree-switch-conversion.c: Include cfgloop.h
21575 (process_switch): If we emit a bit-test cascade, schedule loops
21577 * tree-complex.c: Include cfgloop.h.
21578 (expand_complex_div_wide): Properly add new basic-blocks to loops.
21579 * asan.c: Include cfgloop.h.
21580 (create_cond_insert_point): Properly add new basic-blocks to
21581 loops, schedule loop fixup.
21582 * cfgloop.c (verify_loop_structure): Check that looks are not
21584 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
21586 (expand_omp_for_generic): Likewise.
21587 (expand_omp_sections): Likewise.
21588 (expand_omp_atomic_pipeline): Schedule loops for fixup.
21589 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
21590 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
21591 is initialized, not when loops are present.
21592 * tree-parloops.c (parallelize_loops): Remove checking here.
21593 * passes.c (init_optimization_passes): Schedule a copy-propagation
21594 pass before complete unrolling of inner loops.
21596 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21598 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
21599 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
21600 (colorize_init): Add argument to _WIN32 version.
21601 * toplev.c: Include diagnostic-color.h.
21602 (process_options): Default to -fdiagnostics-color=auto if
21603 GCC_COLORS env var is in the environment.
21604 * common.opt (fdiagnostics-color=): Add Var and Init.
21605 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
21606 env var is in the environment, the default is auto rather than never.
21608 * diagnostic.h (file_name_as_prefix): Add context argument.
21609 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
21610 the string as locus.
21611 * langhooks.c (lhd_print_error_function): Adjust caller.
21613 2013-04-25 Lawrence Crowl <crowl@google.com>
21615 * var-tracking.c (shared_hash_def::htab):
21616 Change type to hash_table. Update dependent calls and types.
21618 2013-04-25 Lawrence Crowl <crowl@google.com>
21620 * Makefile.in: Update as needed below.
21622 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
21623 Move declaration to after the type's method definitons.
21625 * attribs.c (htab_t scoped_attributes::attribute_hash):
21626 Change type to hash_table. Update dependent calls and types.
21628 * bitmap.c (htab_t bitmap_desc_hash):
21629 Change type to hash_table. Update dependent calls and types.
21631 * cselib.c (htab_t cselib_hash_table):
21632 Change type to hash_table. Update dependent calls and types.
21634 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
21635 (hash_string_slot_node): Move implementation into lto-streamer.h
21636 struct string_slot_hasher.
21637 (eq_string_slot_node): Likewise.
21639 * data-streamer-out.c: Update output_block::string_hash_table
21640 dependent calls and types.
21642 * dwarf2cfi.c (htab_t trace_index):
21643 Change type to hash_table. Update dependent calls and types.
21645 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
21646 Change type to hash_table. Update dependent calls and types.
21647 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
21648 (htab_t optimize_external_refs::map): Likewise.
21649 (htab_t output_comp_unit::extern_map): Likewise.
21650 (htab_t output_comdat_type_unit::extern_map): Likewise.
21651 (htab_t output_macinfo::macinfo_htab): Likewise.
21652 (htab_t optimize_location_lists::htab): Likewise.
21653 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
21655 * except.c (htab_t ehspec_hash_type):
21656 Change type to hash_table. Update dependent calls and types.
21657 (assign_filter_values::ttypes): Likewise.
21658 (assign_filter_values::ehspec): Likewise.
21659 (sjlj_assign_call_site_values::ar_hash): Likewise.
21660 (convert_to_eh_region_ranges::ar_hash): Likewise.
21662 * gcse.c (htab_t pre_ldst_table):
21663 Change type to hash_table. Update dependent calls and types.
21665 * ggc-common.c (htab_t saving_htab):
21666 Change type to hash_table. Update dependent calls and types.
21667 (htab_t loc_hash): Likewise.
21668 (htab_t ptr_hash): Likewise.
21669 (call_count): Rename ggc_call_count.
21670 (call_alloc): Rename ggc_call_alloc.
21671 (loc_descriptor): Rename make_loc_descriptor.
21672 (add_statistics): Rename ggc_add_statistics.
21674 * ggc-common.c (saving_htab):
21675 Change type to hash_table. Update dependent calls and types.
21677 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
21678 (push_gimplify_context): Likewise.
21679 (pop_gimplify_context): Likewise.
21680 (struct gimple_temp_hash_elt): Added.
21681 (struct gimplify_hasher): Likewise.
21682 (struct gimplify_ctx.temp_htab):
21683 Change type to hash_table. Update dependent calls and types.
21685 * gimple-fold.c: Include gimplify-ctx.h.
21687 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
21688 Change type to hash_table. Update dependent calls and types.
21689 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
21690 avoid potential global name collision.
21692 * gimplify.c: Include gimplify-ctx.h.
21693 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
21694 (htab_t gimplify_ctx::temp_htab):
21695 Update dependent calls and types for new type hash_table.
21696 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
21697 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
21699 * gimplify-ctx.h: New.
21700 (struct gimple_temp_hash_elt): Move from gimplify.c.
21701 (class gimplify_hasher): New.
21702 (struct gimplify_ctx): Move from gimple.h.
21703 (htab_t gimplify_ctx::temp_htab):
21704 Change type to hash_table. Update dependent calls and types.
21706 * graphite-clast-to-gimple.c: Include graphite-htab.h.
21707 (htab_t ivs_params::newivs_index):
21708 Change type to hash_table. Update dependent calls and types.
21709 (htab_t ivs_params::params_index): Likewise.
21710 (htab_t print_generated_program::params_index): Likewise.
21711 (htab_t gloog::newivs_index): Likewise.
21712 (htab_t gloog::params_index): Likewise.
21714 * graphite.c: Include graphite-htab.h.
21715 4htab_t graphite_transform_loops::bb_pbb_mapping):
21716 Change type to hash_table. Update dependent calls and types.
21718 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
21719 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
21720 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
21722 * graphite-dependences.c: Include graphite-htab.h.
21723 (loop_is_parallel_p): Change hash table type of parameter.
21725 * graphite-htab.h: New.
21726 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
21727 (extern find_pbb_via_hash): Move from graphite-poly.h.
21728 (extern loop_is_parallel_p): Move from graphite-poly.h.
21729 (extern get_loop_body_pbbs): Move from graphite-poly.h.
21731 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
21732 (extern loop_is_parallel_p): Move to graphite-htab.h.
21733 (extern get_loop_body_pbbs): Move to graphite-htab.h.
21735 * haifa-sched.c (htab_t delay_htab):
21736 Change type to hash_table. Update dependent calls and types.
21737 (htab_t delay_htab_i2): Likewise.
21739 * ira-color.c (htab_t allocno_hard_regs_htab):
21740 Change type to hash_table. Update dependent calls and types.
21742 * ira-costs.c (htab_t cost_classes_htab):
21743 Change type to hash_table. Update dependent calls and types.
21745 * loop-invariant.c (htab_t merge_identical_invariants::eq):
21746 Change type to hash_table. Update dependent calls and types.
21748 * loop-iv.c (htab_t bivs):
21749 Change type to hash_table. Update dependent calls and types.
21751 * loop-unroll.c (htab_t opt_info::insns_to_split):
21752 Change type to hash_table. Update dependent calls and types.
21753 (htab_t opt_info::insns_with_var_to_expand): Likewise.
21755 * lto-streamer.h (struct string_slot): Move from data-streamer.h
21756 (struct string_slot_hasher): New.
21757 (htab_t output_block::string_hash_table):
21758 Change type to hash_table. Update dependent calls and types.
21760 * lto-streamer-in.c (freeing_string_slot_hasher): New.
21761 (htab_t file_name_hash_table):
21762 Change type to hash_table. Update dependent calls and types.
21764 * lto-streamer-out.c: Update output_block::string_hash_table dependent
21767 * lto-streamer.c (htab_t tree_htab):
21768 Change type to hash_table. Update dependent calls and types.
21770 * omp-low.c: Include gimplify-ctx.h.
21772 * passes.c (htab_t name_to_pass_map):
21773 Change type to hash_table. Update dependent calls and types.
21774 (pass_traverse): Rename to passes_pass_traverse.
21776 * plugin.c (htab_t event_tab):
21777 Change type to hash_table. Update dependent calls and types.
21779 * postreload-gcse.c (htab_t expr_table):
21780 Change type to hash_table. Update dependent calls and types.
21781 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
21783 * sese.c (debug_rename_map_1): Make extern.
21784 (htab_t copy_bb_and_scalar_dependences::rename_map):
21785 Change type to hash_table. Update dependent calls and types.
21787 * sese.h (extern debug_rename_map): Move to .c file.
21789 * store-motion.c (htab_t store_motion_mems_table):
21790 Change type to hash_table. Update dependent calls and types.
21792 * trans-mem.c (htab_t tm_new_mem_hash):
21793 Change type to hash_table. Update dependent calls and types.
21795 * tree-browser.c (htab_t TB_up_ht):
21796 Change type to hash_table. Update dependent calls and types.
21798 * tree-cfg.c (htab_t discriminator_per_locus):
21799 Change type to hash_table. Update dependent calls and types.
21801 * tree-complex.c: Include tree-hasher.h
21802 (htab_t complex_variable_components):
21803 Change type to hash_table. Update dependent calls and types.
21805 * tree-eh.c (htab_t finally_tree):
21806 Change type to hash_table. Update dependent calls and types.
21808 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
21809 struct int_tree_hasher.
21810 (extern int_tree_map_eq): Likewise.
21811 (uid_decl_map_hash): Removed.
21812 (extern decl_tree_map_eq): Likewise.
21814 * tree-hasher.h: New.
21815 (struct int_tree_hasher): New.
21816 (typedef int_tree_htab_type): New.
21818 * tree-inline.c: Include gimplify-ctx.h.
21820 * tree-mudflap.c: Include gimplify-ctx.h.
21822 * tree-parloops.c: Include tree-hasher.h.
21823 (htab_t eliminate_local_variables_stmt::decl_address):
21824 Change type to hash_table. Update dependent calls and types.
21825 (htab_t separate_decls_in_region::decl_copies): Likewise.
21827 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
21828 Change type to hash_table. Update dependent calls and types.
21830 * tree-sra.c (candidates):
21831 Change type to hash_table. Update dependent calls and types.
21833 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
21835 (int_tree_map_hash): Likewise.
21837 * tree-ssa-dom.c (htab_t avail_exprs):
21838 Change type to hash_table. Update dependent calls and types.
21840 * tree-ssa-live.c (var_map_base_init::tree_to_index):
21841 Change type to hash_table. Update dependent calls and types.
21843 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
21844 Change type to hash_table. Update dependent calls and types.
21846 * tree-ssa-phiopt.c (seen_ssa_names):
21847 Change type to hash_table. Update dependent calls and types.
21849 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
21850 Change type to hash_table. Update dependent calls and types.
21852 * tree-ssa-uncprop.c (equiv):
21853 Change type to hash_table. Update dependent calls and types.
21855 2013-04-25 Jakub Jelinek <jakub@redhat.com>
21857 PR rtl-optimization/57003
21858 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
21859 call note_stores with kill_clobbered_value callback again after
21860 killing regs_invalidated_by_call.
21862 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21864 * config/aarch64/aarch64-simd.md
21865 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
21866 (aarch64_simd_bsl<mode>): Likewise.
21867 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
21869 2013-04-25 Marek Polacek <polacek@redhat.com>
21871 PR tree-optimization/57066
21872 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
21874 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21876 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
21878 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21880 * config/aarch64/aarch64-builtins.c
21881 (aarch64_fold_builtin): New.
21882 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
21883 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
21884 * config/aarch64/aarch64-simd-builtins.def (abs): New.
21885 * config/aarch64/arm_neon.h
21886 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
21888 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21889 Tejas Belagod <tejas.belagod@arm.com>
21891 * config/aarch64/aarch64-builtins.c
21892 (aarch64_gimple_fold_builtin): New.
21893 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
21894 * config/aarch64/aarch64-simd-builtins.def (addv): New.
21895 * config/aarch64/aarch64-simd.md (addpv4sf): New.
21896 (addvv4sf): Update.
21897 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
21899 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21901 * config/aarch64/aarch64.md
21902 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
21904 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21906 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
21907 (*ngcsi_uxtw): New pattern.
21909 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21910 Julian Brown <julian@codesourcery.com>
21912 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
21913 (TB_DREG): Add T_V4HF.
21914 (v4hf_UP): New macro.
21915 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21916 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21917 Handle initialisation of V4HF. Adjust initialisation of reinterpret
21919 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21920 (arm_vector_mode_supported_p): Handle V4HF.
21921 (arm_mangle_map): Handle V4HFmode.
21922 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
21923 * config/arm/arm_neon_builtins.def: Add entries for
21924 vcvtv4hfv4sf, vcvtv4sfv4hf.
21925 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
21926 (neon_vcvtv4hfv4sf): Likewise.
21927 * config/arm/neon-gen.ml: Handle half-precision floating point
21929 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
21930 * config/arm/arm_neon.h: Regenerate.
21931 * config/arm/neon.ml (type elts): Add F16.
21932 (type vectype): Add T_float16x4, T_floatHF.
21933 (type vecmode): Add V4HF.
21934 (type features): Add Requires_FP_bit feature.
21935 (elt_width): Handle F16.
21936 (elt_class): Likewise.
21937 (elt_of_class_width): Likewise.
21938 (mode_of_elt): Refactor.
21939 (type_for_elt): Handle F16, fix error messages.
21940 (vectype_size): Handle T_float16x4.
21941 (vcvt_sh): New function.
21942 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
21943 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
21944 (string_of_mode): Handle V4HF.
21945 * doc/arm-neon-intrinsics.texi: Regenerate.
21947 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21949 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
21950 format specifier in 'X' case.
21952 2013-04-25 Alan Modra <amodra@gmail.com>
21955 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
21956 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
21957 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
21958 Repeat for many other rotate/shift and mask patterns using subregs.
21959 Name lshiftrt insns.
21960 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
21961 on WORDS_BIG_ENDIAN.
21963 2013-04-25 Alan Modra <amodra@gmail.com>
21965 * config.gcc: Support little-endian powerpc-linux targets.
21966 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
21967 (LINK_OS_LINUX_SPEC): Define.
21968 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
21969 Preserve MASK_LITTLE_ENDIAN.
21970 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
21971 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
21972 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
21973 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
21974 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
21975 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
21976 Correct fp word order for little-endian. Don't shift toc entries
21977 smaller than a word for little-endian.
21978 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
21979 (bswapdi2 splits): Correct low-part subreg for little-endian.
21980 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
21981 low/high where such is correct only for be.
21982 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
21983 little-endian for -mcall-aixdesc.
21985 2013-04-25 Alan Modra <amodra@gmail.com>
21987 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
21988 replace_equiv_address_nv.
21990 2013-04-25 Alan Modra <amodra@gmail.com>
21992 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
21994 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
21997 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
21998 * rtl.h (struct rtx_def): ...
22000 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22002 PR rtl-optimizations/57046
22003 * lra-constraints (split_reg): Set up lra_risky_transformations_p
22004 for multi-reg splits.
22006 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
22008 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
22010 2013-04-24 Sterling Augustine <saugustine@google.com>
22012 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
22013 (comp_dir_string, debug_str_dwo_section): New.
22014 (DEBUG_STR_DWO_SECTION): Rename to ...
22015 (DEBUG_DWO_STR_SECTION): ... this.
22016 (DEBUG_NORM_STR_SECTION): Delete.
22017 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
22018 (DEBUG_STR_DWO_SECTION_FLAGS): New.
22019 (find_AT_string): Move most logic to ...
22020 (find_AT_string_in_table): ... here. New.
22021 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
22022 add_skeleton_AT_string. Delete logic.
22023 (output_skeleton_debug_sections): Remove call to
22024 add_top_level_skeleton_die_attrs.
22025 (add_comp_dir_attribute): Move logic to comp_dir_string.
22026 (dwarf2out_init): Initialize debug_str_dwo_section.
22027 (output_indirect_string): Call find_string_form.
22028 (output_indirect_strings): Rewrite.
22029 (prune_unused_types): Empty skeleton_debug_str_hash.
22030 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
22031 (dwarf2out_finish): Call output_indirect_strings.
22033 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22035 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
22037 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22039 * rtl.h (struct rtx_def): Add comment for field jump.
22040 (LRA_SUBREG_P): New macro.
22041 * recog.c (register_operand): Check LRA_SUBREG_P.
22042 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22044 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22045 (lra_final_code_change): Skip subreg change for operators.
22046 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22047 if there are no operand changes.
22048 * lra-constraints.c (curr_insn_set): New.
22049 (match_reload): Set LRA_SUBREG_P.
22050 (emit_spill_move): Ditto.
22051 (check_and_process_move): Use curr_insn_set. Process only single
22052 set insns. Don't initialize sec_mem_p and change_p.
22053 (simplify_operand_subreg): Use LRA_SUBREG_P.
22054 (reg_in_class_p): New function.
22055 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22056 of #ifdef. Add code to remove cycling.
22057 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22058 non-null disp. Reload inner instead of disp when base and index
22059 are null. Try to put lo_sum into register.
22060 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22061 (check_and_process_move): Move code for move cost check to
22062 simple_move_p. Remove equiv_substitution.
22063 (simple_move_p): New function.
22064 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22065 curr_insn_set. Call check_and_process_move only for single set
22066 insns. Use the new function. Move call of check_and_process_move
22067 after operand equiv substitution and address process.
22069 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
22071 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
22072 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
22073 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
22075 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22077 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
22079 2013-04-24 Marek Polacek <polacek@redhat.com>
22081 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
22082 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
22083 (select_loops_exit_conditions): Likewise.
22084 (number_of_iterations_for_all_loops): Likewise.
22085 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
22086 (scev_analysis): Likewise.
22088 2013-04-02 Catherine Moore <clm@codesourcery.com>
22089 Chao-ying Fu <fu@mips.com>
22091 * config/mips/micromips.md (jraddiusp): New pattern.
22092 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
22093 instruction if possible.
22095 2013-04-24 Alan Modra <amodra@gmail.com>
22097 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
22099 2013-04-24 Julian Brown <julian@codesourcery.com>
22100 Chung-Lin Tang <cltang@codesourcery.com>
22102 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
22103 dependency behavior in enumeration type DIE generation. Add TODO note
22104 to comments about future DW_FORM_sdata/udata re-work of related code.
22106 2013-04-23 Lawrence Crowl <crowl@google.com>
22108 * Makefile.in: Update as needed below.
22110 * hash-table.h (class hash_table):
22111 Correct many methods with parameter types compare_type to the correct
22112 value_type. (Correct code was unlikely to notice the change.)
22113 (hash_table::elements_with_deleted) New.
22114 (class hashtable::iterator): New.
22115 (hashtable::begin()): New.
22116 (hashtable::end()): New.
22117 (FOR_EACH_HASH_TABLE_ELEMENT): New.
22119 * statistics.c (statistics_hashes):
22120 Change type to hash_table. Update dependent calls and types.
22122 * tree-into-ssa.c (var_infos):
22123 Change type to hash_table. Update dependent calls and types.
22125 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
22126 Change type to hash_table. Update dependent calls and types.
22128 * tree-ssa-loop-im.c (struct mem_ref.refs):
22129 Change type to hash_table. Update dependent calls and types.
22131 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
22132 Change type to hash_table. Update dependent calls and types.
22134 * tree-ssa-sccvn.c (vn_tables_s::nary):
22135 Change type to hash_table. Update dependent calls and types.
22136 (vn_tables_s::phis): Likewise.
22137 (vn_tables_s::references): Likewise.
22139 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
22140 (vn_reference_eq): Update parameter and return types.
22142 * tree-ssa-structalias.c (pointer_equiv_class_table):
22143 Change type to hash_table. Update dependent calls and types.
22144 (location_equiv_class_table): Likewise.
22146 * tree-vect-data-refs.c: Consequential changes for making
22147 peeling a hash_table.
22149 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
22150 (destroy_loop_vec_info): Dependent hash_table update.
22152 * tree-vectorizer.h (peeling_htab):
22153 Change type to hash_table. Update dependent calls and types.
22155 2013-04-23 Shiva Chen <shiva0217@gmail.com>
22157 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
22158 to check the register content is equal or not.
22159 * lra-constraints.c (match_reload): Use lra_assign_reg_val
22160 to assign register content record.
22161 * lra-eliminations.c (update_reg_eliminate): Use
22162 lra_update_reg_val_offset to update register content offset.
22163 * lra-int.h (struct lra_reg): Add offset member.
22164 (lra_reg_val_equal_p): New static inline function.
22165 (lra_update_reg_val_offset): New static inline function.
22166 (lra_assign_reg_val): New static inline function.
22167 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
22168 to assign register content record.
22169 (initialize_lra_reg_info_element): Initial offset to zero.
22171 2013-04-23 Catherine Moore <clm@codesourcery.com>
22173 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
22174 operands. Record compression.
22176 2013-04-23 Xinliang David Li <davidxl@google.com>
22178 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
22180 2013-04-23 Richard Biener <rguenther@suse.de>
22182 PR middle-end/57036
22183 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
22184 parameter, only add abnormal goto edges from the copied body
22185 if the call could perform abnormal gotos.
22186 (copy_cfg_body): Adjust.
22188 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
22190 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
22192 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
22194 * coretypes.h (gimple_stmt_iterator): Add struct to make
22197 2013-04-23 Richard Biener <rguenther@suse.de>
22199 PR tree-optimization/57026
22200 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
22201 from SSA names occuring in abnormal PHI nodes.
22203 2013-04-22 Andi Kleen <ak@linux.intel.com>
22205 * lto/lto.c (print_lto_report_1): Fix LTO report names.
22207 2013-04-22 Andi Kleen <ak@linux.intel.com>
22209 * lto/lto.c (print_lto_report_1): Declare early.
22210 (read_cgraph_and_symbols): Call print_lto_report_1 early.
22212 2013-04-22 Andi Kleen <ak@linux.intel.com>
22214 * common.opt (-flto-report-wpa): Add.
22215 * doc/invoke.texi (-flto-report-wpa): Add.
22216 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
22219 2013-04-22 Xinliang David Li <davidxl@google.com>
22221 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
22222 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
22223 * Makefile.in: New dependency
22225 David Daney <ddaney.cavm@gmail.com>
22227 * configure.ac (gcc_cv_as_micromips_support): Use the
22228 --fatal-warnings option.
22229 * configure: Regenerate.
22231 2013-04-22 Marek Polacek <polacek@redhat.com>
22234 * tsan.c (instrument_expr): Don't instrument expression
22235 in case its size is zero.
22237 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
22241 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22243 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22245 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22247 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
22248 (gimple_stmt_iterator): New typedef.
22249 * gimple.h (gimple_stmt_iterator): Rename to...
22250 (gimple_stmt_iterator_d): ... This.
22251 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
22252 trees be valid for GIMPLE and GENERIC.
22253 (TARGET_GIMPLE_FOLD_BUILTIN): New.
22254 * gimple-fold.c (gimple_fold_call): Call target hook
22255 gimple_fold_builtin.
22256 * hooks.c (hook_bool_gsiptr_false): New.
22257 * hooks.h (hook_bool_gsiptr_false): New.
22258 * target.def (fold_stmt): New.
22259 * doc/tm.texi: Regenerate.
22261 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
22264 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
22265 a set sp if no stack realignment.
22267 2013-04-22 Nick Clifton <nickc@redhat.com>
22269 * config.gcc (tilegx-linux): Extend extra_objs rather than
22271 (tilepro-linux): Likewise.
22273 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22275 * config/aarch64/aarch64-builtins.c
22277 (CF0, CF1, CF2, CF3, CF4, CF10): New.
22278 (VAR<1-12>): Add MAP parameter.
22279 (BUILTIN_*): Likewise.
22280 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
22281 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
22282 (aarch64_ushl_n<mode>): Likewise.
22283 (aarch64_sshr_n<mode>): Likewise.
22284 (aarch64_ushr_n<mode>): Likewise.
22285 (aarch64_<maxmin><mode>): Likewise.
22286 (aarch64_sqrt<mode>): Likewise.
22287 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
22288 (vshr<q>_n_*): Likewise.
22290 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22292 * config/aarch64/aarch64-builtins.c
22293 (aarch64_simd_builtin_type_mode): Handle SF types.
22295 (BUILTIN_GPF): Define.
22296 (aarch64_init_simd_builtins): Handle SF types.
22297 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
22298 (frecps): Likewise.
22299 (frecpx): Likewise.
22300 * config/aarch64/aarch64-simd.md
22301 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
22302 (aarch64_frecpe<mode>): New.
22303 (aarch64_frecps<mode>): Likewise.
22304 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
22305 (v8type): Add frecp<esx>.
22306 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
22307 (aarch64_frecps<mode>): Likewise.
22308 * config/aarch64/iterators.md (FRECP): New.
22309 (frecp_suffix): Likewise.
22310 * config/aarch64/arm_neon.h
22311 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
22313 2013-04-22 Christian Bruel <christian.bruel@st.com>
22316 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
22317 (REG_CLASS_NAMES): Idem.
22318 (REG_CLASS_CONTENTS): Idem.
22319 (REGCLASS_HAS_FP_REG): Idem.
22320 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
22321 (sh_conditional_register_usage): Idem.
22323 2013-04-21 Jeff Law <law@redhat.com>
22325 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
22326 (ssa_forward_propagate_and_combine): Use it.
22328 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22330 * lra.c: Update the flow chart diagram.
22332 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22334 PR rtl-optimization/56847
22335 * lra-constraints.c (process_alt_operands): Discourage alternative
22336 with non-matche doffsettable memory constraint fro memory with
22339 2013-04-19 Richard Biener <rguenther@suse.de>
22341 PR tree-optimization/56982
22342 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
22344 * gimplify.c (gimplify_call_expr): Notice special calls.
22345 (gimplify_modify_expr): Likewise.
22346 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
22347 abnormal control flow receivers.
22348 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
22349 in the same way as cfun->has_nonlocal_labels.
22350 (gimple_purge_dead_abnormal_call_edges): Likewise.
22351 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
22352 receivers start a basic-block.
22354 2013-04-19 Richard Biener <rguenther@suse.de>
22356 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
22358 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
22359 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
22360 (SLP_TREE_LOAD_PERMUTATION): Add.
22361 (vect_transform_slp_perm_load): Adjust prototype.
22362 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
22363 (vect_free_slp_instance): Likewise.
22364 (vect_create_new_slp_node): Likewise.
22365 (vect_supported_slp_permutation_p): Remove.
22366 (vect_slp_rearrange_stmts): Adjust.
22367 (vect_supported_load_permutation_p): Likewise. Inline
22368 vect_supported_slp_permutation_p here.
22369 (vect_analyze_slp_instance): Compute load permutations per
22370 slp node instead of per instance.
22371 (vect_get_slp_defs): Adjust.
22372 (vect_transform_slp_perm_load): Likewise.
22373 (vect_schedule_slp_instance): Remove redundant code.
22374 (vect_schedule_slp): Remove hack for PR56270, add it ...
22375 * tree-vect-stmts.c (vectorizable_load): ... here, do not
22376 CSE loads for SLP. Adjust.
22378 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22380 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
22381 spelling in two comments.
22383 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22386 * config/arm/arm.c (load_multiple_sequence): Require SP
22387 as base register for loads if SP is in the register list.
22389 2013-04-19 Martin Jambor <mjambor@suse.cz>
22391 PR tree-optimization/56718
22392 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
22393 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
22394 and made public. Adjusted all callers.
22395 (ipa_intraprocedural_devirtualization): New function.
22396 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
22397 (ipa_intraprocedural_devirtualization): Likewise.
22398 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
22400 2013-04-19 Richard Biener <rguenther@suse.de>
22402 PR tree-optimization/57000
22403 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
22405 2013-04-19 Terry Guo <terry.guo@arm.com>
22407 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
22409 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
22410 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
22411 (cortex_m4_fmacs): Use new reservations.
22412 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
22414 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
22416 PR rtl-optimization/56999
22417 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
22419 (lra_coalesce): Remove split_origin_bitmap and related code.
22420 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
22421 ranges if necessary.
22423 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
22425 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
22427 (ix86_expand_call): Remove clobbered_registers array and use
22428 x86_64_ms_sysv_extra_clobbered_registers instead.
22429 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
22431 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
22433 * config/i386/i386.md (*call_rex64_ms_sysv): Use
22434 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
22435 (*call_value_rex64_ms_sysv): Ditto.
22437 2013-04-18 Cary Coutant <ccoutant@google.com>
22439 * dwarf2out.c (output_pubnames): Check die_perennial_p of
22440 parent instead of die_mark.
22442 2013-04-18 Diego Novillo <dnovillo@google.com>
22444 * gimple.c (create_gimple_tmp): New.
22445 (get_expr_type): New.
22446 (build_assign): New.
22447 (build_type_cast): New.
22448 * gimple.h (enum ssa_mode): Define.
22449 (gimple_seq_set_location): New.
22450 * asan.c (build_check_stmt): Change some gimple_build_* calls
22451 to use build_assign and build_type_cast.
22453 2013-04-18 Richard Biener <rguenther@suse.de>
22455 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
22456 handle negative step. Remove redundant checks.
22457 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
22458 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
22459 for negative step and grouped loads fail to vectorize.
22461 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22463 * emit-rtl.c (reset_insn_used_flags): New function.
22464 (reset_all_used_flags): Use it.
22465 (verify_insn_sharing): New function.
22466 (verify_rtl_sharing): Fix verification for SEQUENCEs.
22468 2013-04-18 Jakub Jelinek <jakub@redhat.com>
22470 PR tree-optimization/56984
22471 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
22472 and (x >> M) >= N don't register any assertion if N << M is the
22475 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22477 * lower-subreg.c (resolve_simple_move): If called self-recursive,
22478 do not delete_insn insns that have not yet been emitted, only
22479 unlink them with remove_insn.
22480 * df-scan.c (df_insn_delete): Revert r197492.
22482 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
22484 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
22485 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
22487 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22489 * config/arm/arm.md (movsicc_insn): Convert define_insn into
22490 define_insn_and_split.
22491 (and_scc,ior_scc,negscc): Likewise.
22492 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
22494 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22496 * config/arm/arm.c (use_return_insn): Return 0 for targets that
22497 can benefit from using a sequence of LDRD instructions in epilogue
22498 instead of a single LDM instruction.
22500 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
22503 * doc/extend.texi: Fix typo.
22505 2013-04-17 Richard Biener <rguenther@suse.de>
22507 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
22508 (vect_build_slp_tree): ... here.
22509 (vect_build_slp_tree_1): Compute which stmts of the SLP group
22510 match. Remove special-casing of mismatched complex loads.
22511 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
22512 re-try the match with swapped commutative operands.
22513 (vect_supported_load_permutation_p): Remove special-casing of
22514 mismatched complex loads.
22515 (vect_analyze_slp_instance): Adjust.
22517 2013-04-17 Richard Biener <rguenther@suse.de>
22519 PR rtl-optimization/56921
22520 * cfgloop.h (struct loop): Add simple_loop_desc member.
22521 (struct niter_desc): Mark with GTY(()).
22522 (simple_loop_desc): Do not use aux field but simple_loop_desc.
22523 * loop-iv.c (get_simple_loop_desc): Likewise.
22524 (free_simple_loop_desc): Likewise.
22527 2013-04-16 Richard Biener <rguenther@suse.de>
22529 PR rtl-optimization/56921
22530 * loop-init.c (pass_rtl_move_loop_invariants): Add
22531 TODO_do_not_ggc_collect to todo_flags_finish.
22532 (pass_rtl_unswitch): Same.
22533 (pass_rtl_unroll_and_peel_loops): Same.
22534 (pass_rtl_doloop): Same.
22536 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
22538 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
22539 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
22540 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
22542 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
22543 * tree-streamer.c (record_common_node): Adjust reference in comment.
22545 2013-04-17 Terry Guo <terry.guo@arm.com>
22547 * config/arm/cortex-m4.md: Add a new bypass.
22549 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22551 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
22553 (*subs_<optab><mode>_multp2): New pattern.
22554 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22555 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22557 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22559 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
22560 (*subs_mul_imm_<mode>): New pattern.
22562 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
22565 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
22566 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
22567 (vsx_movti_32bit): Change j->wa to O->wa.
22569 2013-04-16 Richard Biener <rguenther@suse.de>
22571 PR rtl-optimization/56921
22572 * loop-init.c (pass_rtl_move_loop_invariants): Add
22573 TODO_do_not_ggc_collect to todo_flags_finish.
22574 (pass_rtl_unswitch): Same.
22575 (pass_rtl_unroll_and_peel_loops): Same.
22576 (pass_rtl_doloop): Same.
22578 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
22580 * config/arm/arm.c (emit_multi_reg_push): New declaration
22581 for an existing function.
22582 (arm_emit_strd_push): New function.
22583 (arm_expand_prologue): Used here.
22584 (arm_emit_ldrd_pop): New function.
22585 (arm_expand_epilogue): Used here.
22586 (arm_get_frame_offsets): Update condition.
22587 (arm_emit_multi_reg_pop): Add a special case for load of a single
22588 register with writeback.
22590 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
22592 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
22595 2013-04-16 Richard Biener <rguenther@suse.de>
22597 PR tree-optimization/56756
22598 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
22599 (first_mem_ref_loc): New.
22600 (execute_sm): Place the load temporarily before a previous
22601 access instead of in the latch edge to ensure its SSA dependencies
22602 are defined at points dominating the load.
22604 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
22606 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
22607 correct fix by moving header and footer insn to the footer of
22608 the merged basic block. Clear BB_END of the merged-away block.
22610 PR middle-end/43631
22611 * emit-rtl.c (make_note_raw): New function.
22612 (link_insn_into_chain): New static inline function.
22613 (add_insn): Use it.
22614 (add_insn_before, add_insn_after): Factor insn chain linking code...
22615 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
22616 using link_insn_into_chain.
22617 (note_outside_basic_block_p): New helper function for emit_note_after
22618 and emit_note_before.
22619 (emit_note_after): Use nobb variant of add_insn_after if the note
22620 should not be contained in a basic block.
22621 (emit_note_before): Use nobb variant of add_insn_before if the note
22622 should not be contained in a basic block.
22623 (emit_note_copy): Use make_note_raw.
22624 (emit_note): Likewise.
22625 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
22626 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
22627 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
22628 the moved barrier the tail of the basic block it follows.
22629 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
22631 2013-04-15 Jakub Jelinek <jakub@redhat.com>
22633 PR tree-optimization/56962
22634 * gimple-ssa-strength-reduction.c (record_increment): Only set
22635 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
22636 either rhs1 or rhs2 is equal to c->base_expr.
22638 2013-04-15 Richard Biener <rguenther@suse.de>
22640 PR tree-optimization/56933
22641 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
22643 (GROUP_READ_WRITE_DEPENDENCE): Remove.
22644 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
22645 * tree-vect-data-refs.c (vect_analyze_group_access): Move
22646 dependence check ...
22647 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
22649 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
22650 GROUP_READ_WRITE_DEPENDENCE.
22652 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22654 * emit-rtl.c (reset_all_used_flags): New function.
22655 (verify_rtl_sharing): Call reset_all_used_flags before and after
22656 performing the checks.
22658 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22660 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
22661 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
22662 * config/arm/constraints.md (De): New constraint.
22663 * config/arm/neon.md (anddi3_neon): Delete.
22664 (neon_vand<mode>): Expand to standard anddi3 pattern.
22665 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
22666 Move earlier in the file.
22667 (neon_inv_logic_op2): Likewise.
22668 (arm_anddi_operand_neon): New predicate.
22670 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22672 * configure.ac (gcc_cv_ld_as_needed): Set
22673 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
22674 Use -z ignore, -z record on *-*-solaris2*.
22675 (HAVE_LD_AS_NEEDED): Update comment.
22676 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
22677 * configure: Regenerate.
22678 * config.in: Regenerate.
22679 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
22680 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
22681 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
22682 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
22683 equivalents. Fix markup.
22684 * doc/tm.texi: Regenerate.
22686 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
22688 * config/i386/i386.opt: New option mstack-protector-guard=.
22689 * config/i386/i386-opts.h: Add enum stack_protector_guard.
22690 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
22691 TARGET_SSP_TLS_GUARD.
22692 * config/i386/i386.c (ix86_option_override_internal): Set
22693 ix86_stack_protector_guard.
22694 * config/i386/i386.md (stack_protect_set): Enable for
22695 TARGET_SSP_TLS_GUARD only.
22696 (stack_protect_set_<mode>): Ditto.
22697 (stack_protect_test): Ditto.
22698 (stack_protect_test_<mode>): Ditto.
22699 * doc/invoke.texi (i386 Option): Document.
22701 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
22704 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
22705 (S_MODES): Set H_MODE bit.
22706 (SF_MODES): Set only S_MODE and SF_MODE bits.
22707 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
22708 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
22709 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
22710 <MODE_FLOAT>: Likewise.
22712 2013-04-15 Joey Ye <joey.ye@arm.com>
22714 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
22716 2013-04-15 Joey Ye <joey.ye@arm.com>
22718 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
22720 (thumb_far_jump_used_p): Count instruction size and set
22723 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
22725 * reorg.c (fill_simple_delay_slots): Reindent block of code.
22726 * resource.c (mark_target_live_regs): Reformat conditional block.
22728 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22730 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
22731 notes, they are emitted only just before final.
22732 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
22734 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22736 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
22737 * cfgrtl.c (delete_insn): Call it here instead.
22738 * lra-spills.c (lra_final_code_change): Use delete_insn.
22739 * haifa-sched.c (sched_remove_insn): Likewise.
22740 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
22741 returning to the nop pool.
22742 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
22743 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
22745 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22747 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
22748 * doc/tm.texi: Regenerated.
22750 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
22752 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
22755 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22757 * df-core.c (df_find_def): Compare register numbers.
22758 (df_find_use): Likewise.
22760 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
22763 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
22764 lra_in_progress for return.
22766 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22768 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
22769 define_insn into define_insn_and_split and emit movsicc patterns.
22771 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22773 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
22775 2013-04-12 Richard Biener <rguenther@suse.de>
22777 * tree-pass.h (TODO_do_not_ggc_collect): New.
22778 * passes.c (execute_one_ipa_transform_pass): Honor
22779 TODO_do_not_ggc_collect.
22780 (execute_one_pass): Likewise.
22783 2013-04-10 Richard Biener <rguenther@suse.de>
22785 * passes.c (init_optimization_passes): Remove reload pass.
22786 * ira.c (do_reload): Merge into ...
22788 (rest_of_handle_reload): Remove.
22789 (pass_reload): Likewise.
22790 * config/i386/i386.c (ix86_option_override): Refer to ira instead
22791 of reload for vzeroupper pass placement.
22793 2013-04-12 Jakub Jelinek <jakub@redhat.com>
22795 PR tree-optimization/56918
22796 PR tree-optimization/56920
22797 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
22798 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
22799 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
22800 use wide_mul_with_sign method.
22802 2013-04-12 Richard Biener <rguenther@suse.de>
22804 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
22805 not be considered a gimple constant.
22807 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
22809 * fold-const.c (const_binop): Handle vector shifts by a scalar.
22810 (fold_binary_loc): Call const_binop also for mixed vector-scalar
22813 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
22814 Jakub Jelinek <jakub@redhat.com>
22816 * opts.c: Include diagnostic-color.h.
22817 (common_handle_option): Handle OPT_fdiagnostics_color_.
22818 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
22819 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
22820 (diagnostic-color.o): New.
22821 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
22822 (diagnostic_color_rule): New enum.
22823 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
22824 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
22825 the location string.
22826 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
22827 either NULL, or color kind.
22828 * diagnostic-color.c: New file.
22829 * diagnostic-color.h: New file.
22830 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
22832 * doc/invoke.texi (-fdiagnostics-color): Document.
22833 * pretty-print.h (pp_show_color): Define.
22834 (struct pretty_print_info): Add show_color field.
22835 * diagnostic.c: Include diagnostic-color.h.
22836 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
22837 macros. Colorize error:, warning: etc. strings and also the location
22839 (diagnostic_show_locus): Colorize the caret line.
22840 * pretty-print.c: Include diagnostic-color.h.
22841 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
22842 inside of %< %> quotes or quoted through q format modifier.
22844 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22846 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
22848 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22850 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
22851 code in CC_NZ mode.
22852 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
22855 2013-04-11 Marek Polacek <polacek@redhat.com>
22857 PR tree-optimization/48184
22858 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
22860 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
22862 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
22863 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
22864 (skip_simple_arithmetic): Tidy up.
22865 * tree.h (skip_simple_constant_arithmetic): Declare.
22867 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22869 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
22871 2013-04-11 Richard Biener <rguenther@suse.de>
22873 * tree-vect-loop.c (get_initial_def_for_induction): Properly
22874 generate vector constants.
22876 2013-04-11 Richard Biener <rguenther@suse.de>
22878 PR tree-optimization/56878
22879 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
22880 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
22882 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
22883 Prefer to align the DR with the most invariant base address.
22885 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
22887 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
22890 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
22892 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
22893 floating-point vector comparisons against 0.
22895 2013-04-11 Jakub Jelinek <jakub@redhat.com>
22897 PR tree-optimization/56899
22898 * fold-const.c (extract_muldiv_1): Apply distributive law
22899 only if TYPE_OVERFLOW_WRAPS (ctype).
22901 2013-04-11 Bin Cheng <bin.cheng@arm.com>
22904 * ira-costs.c (scan_one_insn): Check whether the source rtx of
22905 loading has side effect.
22907 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
22909 * config/sparc/sparc.c: Include tree-pass.h.
22910 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
22911 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
22912 head of file. Change return type. Split off gate function.
22913 (sparc_gate_work_around_errata): New function.
22914 (pass_work_around_errata): New pass definition.
22915 (insert_pass_work_around_errata) New pass insert definition to
22916 insert pass_work_around_errata just after delayed-branch scheduling.
22917 (sparc_option_override): Insert the pass.
22918 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
22920 2013-04-10 David S. Miller <davem@davemloft.net>
22922 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
22923 or -mcpu=hypersparc.
22925 * target.def (cstore_mode): New hook.
22926 * target.h: Include insn-codes.h
22927 * targhooks.c: Likewise.
22928 (default_cstore_mode): New function.
22929 * targhooks.h: Declare it.
22930 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
22931 * doc/tm.texi: Rebuild.
22932 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
22933 target hook, rather than inspecting the insn_data.
22934 * config/sparc/sparc.c (sparc_cstore_mode): New function.
22935 (TARGET_CSTORE_MODE): Redefine.
22936 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
22938 * config/sparc/predicates.md (cstore_result_operand): New special
22940 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
22941 Use it for operand 0.
22942 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
22943 (*snesi_special): Likewise.
22944 (*snesi_zero): Likewise.
22945 (*seqsi_zero): Likewise.
22946 (*sltu_insn): Likewise.
22947 (*sgeu_insn): Likewise.
22948 (*seqdi_special): Make operand 0 and comparison operation be of
22950 (*snedi_special): Likewise.
22951 (*snedi_special_vis3): Likewise.
22952 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
22953 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
22954 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
22955 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
22956 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
22957 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
22958 (*sltu_extend_sp64): Likewise.
22959 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
22960 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
22961 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
22962 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
22963 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
22965 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
22967 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
22968 (aarch64_start_file): Use the new function.
22970 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
22971 Jason Merrill <jason@redhat.com>
22973 * common.opt: Add -gdwarf.
22974 * opts.c (common_handle_option): Handle it.
22975 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
22977 2013-04-10 Richard Biener <rguenther@suse.de>
22979 * passes.c (execute_todo): Do not call ggc_collect conditional here.
22980 (execute_one_ipa_transform_pass): But unconditionally here.
22981 (execute_one_pass): And here.
22982 (init_optimization_passes): Remove reload pass.
22983 * tree-pass.h (TODO_ggc_collect): Remove.
22984 (pass_reload): Likewise.
22985 * ira.c (do_reload): Merge into ...
22987 (rest_of_handle_reload): Remove.
22988 (pass_reload): Likewise.
22989 * config/i386/i386.c (ix86_option_override): Refer to ira instead
22990 of reload for vzeroupper pass placement.
22991 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
22992 and todo_flags_finish of all passes.
22994 2013-04-10 Richard Biener <rguenther@suse.de>
22996 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
22997 first_const_oprnd field, rename first_def_type to first_op_type.
22998 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
22999 (vect_get_and_check_slp_defs): Always use the type of the
23000 operand. Allow mixed vect_external_def, vect_constant_def types.
23001 (vect_get_constant_vectors): Handle mixed vect_external_def,
23002 vect_constant_def types.
23004 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
23006 PR tree-optimization/55524
23007 * tree-ssa-math-opts.c
23008 (convert_mult_to_fma): Don't use an fms construct
23009 when we don't have an fms operation, but fnma, and it looks
23010 likely that we'll be able to use the latter.
23012 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
23014 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
23016 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
23017 inline fail caused by overwritable functions.
23019 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
23021 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
23022 unnecessary bits in the constant power of two case.
23024 2013-04-10 Richard Biener <rguenther@suse.de>
23026 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
23027 broken code swapping operands.
23028 (vect_build_slp_tree): Do not compute load permutations here.
23029 (vect_analyze_slp_instance): Compute load permutations here,
23030 after building the SLP tree.
23032 2013-04-09 Christian Bruel <christian.bruel@st.com>
23034 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
23035 of next/prev_real_insn.
23037 2013-04-09 Jan Hubicka <jh@suse.cz>
23039 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
23040 Drop aliased parameter.
23041 (function_and_variable_visibility): Do not handle alias pairs.
23042 * cgraph.c (varpool_externally_visible_p): Update prototype.
23043 * varpool.c (varpool_add_new_variable): Update.
23045 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23047 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
23049 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
23051 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
23053 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
23055 2013-04-09 Marek Polacek <polacek@redhat.com>
23057 PR tree-optimization/48762
23058 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
23060 2013-04-09 Richard Biener <rguenther@suse.de>
23062 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
23064 (vect_build_slp_tree): Likewise.
23065 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
23066 calculating the cost of a SLP instance.
23067 (vect_analyze_slp_instance): Use it from here, after building
23070 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23072 PR middle-end/56883
23073 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
23074 expand_omp_for_static_chunk): Use simple_p = true in
23075 force_gimple_operand_gsi calls when assigning to addressable decls.
23077 2013-04-09 Jeff Law <law@redhat.com>
23079 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
23080 when the boolean was created by converting a wider object which
23081 had a boolean range.
23083 2013-04-09 Richard Biener <rguenther@suse.de>
23085 * tree-vectorizer.h (slp_void_p): Remove.
23086 (slp_tree): Typedef before _slp_tree declaration.
23087 (struct _slp_tree): Use a vector of slp_tree as children.
23088 (vect_get_place_in_interleaving_chain): Remove.
23089 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
23091 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
23093 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
23094 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
23095 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
23096 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
23097 Use slp_node instead of slp_void_p and adjust.
23099 2013-04-09 Richard Biener <rguenther@suse.de>
23101 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
23102 work that is not necessary.
23104 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23106 PR tree-optimization/56854
23107 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
23108 forward into clobber stmts if it would change MEM_REF lhs into
23111 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
23113 * tree.c (type_hash_lookup, type_hash_add): Make static.
23114 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
23116 2013-04-09 Richard Biener <rguenther@suse.de>
23118 * tree.h (unsave_expr_now): Remove.
23119 * tree-inline.c (mark_local_for_remap_r): Remove.
23120 (unsave_expr_1): Likewise.
23121 (unsave_r): Likewise.
23122 (unsave_expr_now): Likewise.
23123 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
23124 (propagate_tree_value): Likewise.
23126 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
23128 * doc/rtl.texi (sequence): Rewrite documentation to match the
23129 current use of SEQUENCE rtl objects.
23130 * rtl.def (SEQUENCE): Likewise.
23132 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
23133 Update documentation.
23134 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
23135 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
23137 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
23139 2013-04-08 Teresa Johnson <tejohnson@google.com>
23141 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
23142 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
23144 (estimate_edge_size_and_time): Add comment to suggest using rounding
23146 (estimate_node_size_and_time): Ditto.
23147 (remap_edge_change_prob): Use helper rounding divide methods.
23148 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
23149 (gimple_mod_pow2_value_transform): Ditto.
23150 (gimple_mod_subtract_transform): Ditto.
23151 (gimple_ic_transform): Ditto.
23152 (gimple_stringops_transform): Ditto.
23153 * stmt.c (conditional_probability): Ditto.
23154 (emit_case_dispatch_table): Ditto.
23155 * lto-cgraph.c (merge_profile_summaries): Ditto.
23156 * tree-optimize.c (execute_fixup_cfg): Ditto.
23157 * cfgcleanup.c (try_forward_edges): Ditto.
23158 * cfgloopmanip.c (scale_loop_profile): Ditto.
23160 (duplicate_loop_to_header_edge): Ditto.
23161 (lv_adjust_loop_entry_edge): Ditto.
23162 * tree-vect-loop.c (vect_transform_loop): Ditto.
23163 * profile.c (compute_branch_probabilities): Ditto.
23164 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
23165 * lto-streamer-in.c (input_cfg): Ditto.
23166 * gimple-streamer-in.c (input_bb): Ditto.
23167 * ipa-cp.c (update_profiling_info): Ditto.
23168 (update_specialized_profile): Ditto.
23169 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
23170 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
23172 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23173 (compute_trg_info): Ditto.
23174 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
23175 (purge_dead_edges): Ditto.
23176 * loop-unswitch.c (unswitch_loop): Ditto.
23177 * cgraphclones.c (cgraph_clone_edge): Ditto.
23178 (cgraph_clone_node): Ditto.
23179 * tree-inline.c (copy_bb): Ditto.
23180 (copy_edges_for_bb): Ditto.
23181 (initialize_cfun): Ditto.
23182 (copy_cfg_body): Ditto.
23183 (expand_call_inline): Ditto.
23185 2013-04-08 Kai Tietz <ktietz@redhat.com>
23187 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
23188 TARGET_CYGWIN64 by TARGET_64BIT.
23190 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23192 * config/epiphany/epiphany.md (GPR_1): New constant.
23193 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
23194 * config/epiphany/epiphany.c (gen_compare_reg):
23195 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
23196 is already in place.
23197 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
23198 Don't require being called during rtl expansion; If y operlaps r0,
23200 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
23201 (epiphany_expand_epilogue): Likewise.
23203 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
23204 Don't use CC_FPmode for ORDERED / UNORDERED.
23205 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
23207 * config/epiphany/constraints.md (CnL): New constraint.
23208 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
23209 * config/epiphany/predicates.md (add_operand): Allow 1024.
23211 * config/epiphany/epiphany.md (logical_op): New code iterator.
23212 (op_mnc): New code attribute.
23213 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
23214 (mov_f+1, mov_f+2): New peephole2 patterns.
23216 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
23217 (cstoresi4): Also allow re-use of zero result when doing a NE
23218 comparison to a non-zero operand.
23219 Use (clobber (scratch)) for first insn if the gpr output is not needed.
23221 * config/epiphany/epiphany.md (<insn_opname>v2si3):
23222 Use gen_addsi3_i / gen_subsi3_i.
23224 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23228 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
23229 contain anything but clobbers, at most one __builtin_stack_restore,
23230 optionally debug stmts and final resx, and if it has at least one
23231 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
23232 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
23233 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
23234 which isn't defaut definition, remove them.
23235 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
23236 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
23237 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
23238 with MEM_REF LHS with SSA_NAME address.
23240 2013-04-08 Jeff Law <law@redhat.com>
23242 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
23244 2013-04-08 Richard Biener <rguenther@suse.de>
23246 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
23248 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
23249 determined vector type.
23250 (vect_analyze_data_refs): Likewise.
23251 (vect_get_new_vect_var): Adjust.
23252 (vect_create_destination_var): Preserve SSA name versions.
23253 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
23254 not dump anything here.
23256 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23258 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
23259 Add member lr_slot_known.
23260 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
23262 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
23263 Remove code that sets lr_slot_offset according to what a previous
23264 version of epiphany_emit_save_restore used to do.
23265 (epiphany_emit_save_restore): When doing an lr save or restore,
23266 set/verify lr_slot_known and lr_slot_offset.
23268 2013-04-08 Xinyu Qi <xyqi@marvell.com>
23271 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
23274 2013-04-08 Richard Biener <rguenther@suse.de>
23276 * alias.c (find_base_term): Fix thinko in previous change.
23278 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23280 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
23281 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
23282 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
23283 if possible to compute val.
23284 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
23285 For QImode integers don't require anything about precision. Use
23286 const_with_all_bytes_same to find out if the constant doesn't have
23287 repeated bytes in it.
23289 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23291 * config/s390/s390.c (s390_expand_insv): Only accept insertions
23294 2013-04-08 Marek Polacek <polacek@redhat.com>
23296 PR rtl-optimization/48182
23297 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
23300 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23303 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
23304 nuses, make sure we have a label.
23306 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23309 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
23310 (rs6000_emit_swdiv_low_precision): Remove.
23311 (rs6000_emit_swdiv): Rewrite to handle between one and four
23312 iterations of Newton-Raphson generally; modify required number of
23313 iterations for some cases.
23314 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
23316 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
23318 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
23319 set-but-unused variable.
23321 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
23322 basic blocks of released function bodies garbage-collectable.
23324 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
23325 (struct rtl_opt_pass): Add TODO_df_finish.
23327 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
23329 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23331 * config/arm/constraints.md (q): New constraint.
23332 * config/arm/ldrdstrd.md: New file.
23333 * config/arm/arm.md (ldrdstrd.md) New include.
23334 (arm_movdi): Use "q" instead of "r" constraint
23335 for double-word memory access.
23336 (movdf_soft_insn): Likewise.
23337 * config/arm/vfp.md (movdi_vfp): Likewise.
23338 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
23339 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
23340 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
23341 (mem_ok_for_ldrd_strd): Likewise.
23342 (output_move_double): Update assertion.
23344 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23346 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
23348 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23350 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
23351 define_insn_and_split.
23352 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
23354 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23356 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
23357 define_insn_and_split.
23358 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
23359 (shiftsi3_compare): New pattern.
23360 (rrx): New pattern.
23361 * config/arm/unspecs.md (UNSPEC_RRX): New.
23363 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23365 * config/arm/arm.md (negdi_extendsidi): New pattern.
23366 (negdi_zero_extendsidi): Likewise.
23368 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23370 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
23371 define_insn_and_split.
23372 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
23373 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
23375 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23377 * config/arm/arm.md (arm_subdi3): Convert define_insn into
23378 define_insn_and_split.
23379 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
23380 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
23382 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23384 * config/arm/arm.md (subsi3_carryin): New pattern.
23385 (subsi3_carryin_const): Likewise.
23386 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
23387 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
23389 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23391 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
23393 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23395 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
23396 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
23398 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23400 * config/arm/arm.c (arm_expand_builtin): Change fcode
23401 type to unsigned int.
23403 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23405 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
23407 2013-04-04 Ian Lance Taylor <iant@google.com>
23409 * doc/standards.texi (Standards): The Go frontend supports the Go 1
23412 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
23414 PR middle-end/56729
23415 * df-scan.c (df_insn_delete): Disable failing assert.
23417 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23419 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
23420 New function prototype.
23421 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
23422 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
23423 (arm_builtin_vectorized_function): New function.
23425 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23427 * config/arm/arm_neon_builtins.def: New file.
23428 * config/arm/arm.c (neon_builtin_data): Move contents to
23429 arm_neon_builtins.def.
23430 (enum arm_builtins): Include neon builtin definitions.
23431 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
23432 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
23434 2013-04-04 Marek Polacek <polacek@redhat.com>
23436 PR tree-optimization/48186
23437 * predict.c (maybe_hot_frequency_p): Return false if
23438 HOT_BB_FREQUENCY_FRACTION is 0.
23439 (cgraph_maybe_hot_edge_p): Likewise.
23441 2013-04-04 Richard Biener <rguenther@suse.de>
23443 PR tree-optimization/56826
23444 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
23447 2013-04-04 Richard Biener <rguenther@suse.de>
23449 PR tree-optimization/56213
23450 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
23451 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
23453 2013-04-04 Richard Biener <rguenther@suse.de>
23455 PR tree-optimization/56837
23456 * tree-loop-distribution.c (classify_partition): For non-zero
23457 values require that the value has the same precision as its
23458 mode to be useful as memset value.
23460 2013-04-03 Nick Clifton <nickc@redhat.com>
23462 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
23463 (fmssf4): Use fmsf.s on E3V5 architectures.
23464 (fnmasf4): Use fnmaf.s on E3V5 architectures.
23465 (fnmssf4): Use fnmsf.s on E3V5 architectures.
23467 2013-04-03 Jeff Law <law@redhat.com>
23469 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
23470 (lra-eliminations.o): Likewise.
23472 2013-04-03 Teresa Johnson <tejohnson@google.com>
23474 * gcov-io.c (compute_working_sets): Moved most of body of old
23475 compute_working_sets here from profile.c.
23476 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
23477 (gcov_working_set_t): Moved typedef here from basic-block.h
23478 (compute_working_set): Declare.
23479 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
23480 (get_working_sets): Renamed from compute_working_set,
23481 replace most of body with call to new compute_working_sets.
23482 (get_exec_counts): Replace call to compute_working_sets
23483 to get_working_sets.
23484 * profile.h (get_working_sets): Renamed from compute_working_set.
23485 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
23486 to get_working_sets.
23487 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
23488 * gcov-dump.c (dump_working_sets): New function.
23490 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
23492 * hwint.c (sext_hwi, zext_hwi): New functions.
23493 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
23494 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
23495 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
23496 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
23497 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
23498 (sext_hwi, zext_hwi): New functions.
23500 2013-04-03 Jeff Law <law@redhat.com>
23502 PR tree-optimization/56799
23503 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
23504 back test for widening conversion erroneously dropped in prior change.
23506 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23509 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
23510 instead of next_real_insn.
23512 2013-04-03 Marek Polacek <polacek@redhat.com>
23515 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
23517 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23520 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
23522 (thumb1_output_casesi): Likewise.
23523 (thumb2_output_casesi): Likewise.
23525 2013-04-03 Richard Biener <rguenther@suse.de>
23527 PR tree-optimization/56817
23528 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
23530 (tree_unroll_loops_completely_1): ... new function to manually
23531 walk the loop tree, properly defering outer loops of unrolled
23532 loops to later iterations.
23534 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23536 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
23537 (vectorizable_load): Likewise.
23538 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
23539 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
23541 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23543 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
23546 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
23548 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
23550 2013-04-03 Bin Cheng <bin.cheng@arm.com>
23552 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
23554 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23556 PR tree-optimization/56790
23557 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
23560 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23562 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
23564 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
23565 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
23568 2013-04-03 Jakub Jelinek <jakub@redhat.com>
23571 * tree.h (force_folding_builtin_constant_p): New decl.
23572 * builtins.c (force_folding_builtin_constant_p): New variable.
23573 (fold_builtin_constant_p): Fold immediately also if
23574 force_folding_builtin_constant_p.
23576 2013-04-03 Richard Biener <rguenther@suse.de>
23578 PR tree-optimization/56812
23579 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
23580 DRs of the same interleaving chain are independent.
23582 2013-04-02 Jason Merrill <jason@redhat.com>
23584 * gdbinit.in (pbb): Use debug fn.
23586 2013-04-02 Lawrence Crowl <crowl@google.com>
23588 * sese.h (struct ivtype_map_elt_s): Remove unused.
23589 (extern debug_ivtype_map): Remove unused.
23590 (extern eq_ivtype_map_elts): Remove unused.
23591 * sese.c (debug_ivtype_map): Removed unused.
23592 (debug_ivtype_map_1): Removed unused.
23593 (debug_ivtype_elt): Remove unused.
23594 (eq_ivtype_map_elts): Remove unused.
23597 2013-04-02 Kai Tietz <ktietz@redhat.com>
23600 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
23601 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
23602 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
23604 (legitimize_pe_coff_symbol): Likewise.
23605 (is_imported_p): New helper-function.
23606 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
23607 for Windows x64 targets.
23608 (ix86_expand_prologue): Optimize for pe-coff targets.
23609 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
23610 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
23611 medium/large code-model.
23612 (legitimize_pic_address): Likewise.
23613 (legitimize_tls_address): Likewise.
23614 (ix86_expand_call): Likewise.
23615 (x86_output_mi_thunk): Likewise.
23616 (get_dllimport_decl): Add new beimport argument.
23617 (construct_plt_address): Don't assert for x64 pe-coff targets.
23618 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
23620 (SYMBOL_FLAG_STUBVAR): New macro.
23621 (SYMBOL_REF_STUBVAR_P): Likewise.
23622 * config/i386/winnt.c (stub_list): New structure.
23623 (stub_head): New local variable.
23624 (i386_pe_record_stub): New function.
23625 (i386_pe_file_end): Emit refptr-stubs.
23627 2013-04-02 Jakub Jelinek <jakub@redhat.com>
23629 PR rtl-optimization/56745
23630 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
23631 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
23634 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
23635 and both of them are MEM_REFs, just compare first argument for
23636 equality and attempt to deal even with differing offsets.
23639 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
23640 of gimple_clobber_p to be MEM_REF.
23641 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
23642 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
23643 after gimplification.
23644 * asan.c (get_mem_ref_of_assignment): Don't instrument
23645 gimple_clobber_p stmts.
23646 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
23647 gimple_clobber_p stmt if they have MEM_REF lhs and
23648 are dead because of another gimple_clobber_p stmt.
23649 * tree-ssa-live.c (clear_unused_block_pointer): Treat
23650 gimple_clobber_p stmts like debug stmts.
23651 (remove_unused_locals): Remove clobbers with MEM_REF lhs
23652 that refer to unused VAR_DECLs or uninitialized values.
23653 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
23654 gimple_clobber_p stmts if they refer to removed parameters.
23655 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
23658 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
23660 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
23661 using SWI48 mode attribute.
23663 2013-04-02 Wei Mi <wmi@google.com>
23665 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
23666 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
23667 *<rotate_insn><mode>3_mask in i386.md.
23669 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
23671 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
23673 2013-04-02 Richard Biener <rguenther@suse.de>
23675 PR tree-optimization/56778
23676 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
23677 Runtime alias tests are not supported for gather loads.
23678 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
23679 stmts referenced from SSA operands before updating SSA form.
23681 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
23682 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23684 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
23685 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
23686 * config/arm/cortex-a53.md: New file.
23687 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
23688 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
23689 * config/arm/arm.c (arm_issue_rate): Likewise.
23690 * config/arm/arm-tune.md: Regenerate
23691 * config/arm/arm-tables.opt: Regenerate.
23692 * config/arm/arm-cores.def: Add cortex-a53.
23694 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
23696 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
23699 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23701 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
23702 scalar load/store operations using B/H registers.
23703 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
23705 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23707 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
23709 * config/aarch64/aarch64.c
23710 (aarch64_simd_scalar_immediate_valid_for_move): New.
23711 * config/aarch64/aarch64-protos.h
23712 (aarch64_simd_scalar_immediate_valid_for_move): New.
23713 * config/aarch64/constraints.md (Dh, Dq): New.
23714 * config/aarch64/iterators.md (hq): New.
23716 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
23718 * reorg.c (get_branch_condition): Deal with conditional returns.
23719 (fill_simple_delay_slots): Remove dead code dealing with jumps.
23721 2013-04-01 Wei Mi <wmi@google.com>
23723 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
23724 Truncate operand 2 using %b asm operand modifier.
23725 (*<shift_insn><mode>3_mask): Ditto.
23726 (*<rotate_insn><mode>3_mask): Ditto.
23728 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
23730 PR middle-end/56798
23731 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
23733 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
23735 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
23737 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
23739 2013-03-30 Lawrence Crowl <crowl@google.com>
23741 * dse.c (clear_alias_sets): Remove never set.
23742 (disqualified_clear_alias_sets): Remove never set.
23743 (clear_alias_mode_pool): Remove never set.
23744 (dse_step0): Remove condition that is never true.
23745 (canon_address): Remove condition that is never true.
23746 (dse_step7): Remove condition that is never true.
23747 (rest_of_handle_dse): Remove condition that is never true.
23748 (rest_of_handle_dse::did_global): Remove never read from above.
23749 (dse_step2_spill): Remove never called from above.
23750 (dse_step5_spill): Remove never called from above.
23752 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
23754 * doc/md.texi (Standard Names) <casesi>: Update documentation for
23755 JUMP_TABLE_DATA changes.
23756 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
23757 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
23758 (Insns) <jump_table_data>: New entry.
23759 * doc/tm.texi: Regenerate.
23761 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
23763 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
23764 for table jump at the end of a basic block using tablejump_p.
23765 * targhooks.c (default_invalid_within_doloop): Likewise.
23766 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
23767 target hook implementation that is identical to the default hook.
23768 (rs6000_invalid_within_doloop): Remove.
23770 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
23771 unused variable from tablejump_p call.
23773 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
23774 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
23775 (INSN_DELETED_P): Likewise.
23776 (emit_jump_table_data): New prototype.
23777 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
23778 after 4th as unused.
23779 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
23780 * sched-vis.c (print_insn): Likewise.
23781 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
23782 insn for compatibility with back ends that use next_active_insn to
23783 identify jump table data.
23784 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
23785 (remove_insn): Likewise.
23786 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
23788 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
23789 (emit_jump_table_data): New function.
23791 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
23792 basic block, a JUMP_TABLE_DATA never is.
23793 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
23794 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
23795 off from code handling real insns.
23796 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
23797 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
23798 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
23799 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
23800 is not a NONDEBUG_INSN_P.
23801 * ira-costs.c (scan_one_insn): Likewise.
23802 * jump.c (mark_all_labels): Likewise.
23803 (mark_jump_label_1): Likewise.
23804 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
23805 * lra.c (get_insn_freq): Expect all insns reaching here to be in
23807 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
23808 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
23809 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
23810 JUMP_TABLE_DATA_P insns.
23811 (calculate_elim_costs_all_insns): Likewise.
23812 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
23813 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
23814 (delete_output_reload): Code style fixups.
23815 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
23816 insn flags on this non-insn.
23817 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
23818 as scheduling barriers, for pre-change compatibility.
23819 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
23820 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
23822 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
23823 redundant JUMP_TABLE_DATA_P test.
23824 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
23825 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
23826 (frv_for_each_packet): Likewise.
23827 * config/i386/i386.c (min_insn_size): Likewise.
23828 (ix86_avoid_jump_mispredicts): Likewise.
23829 * config/m32r/m32r.c (m32r_is_insn): Likewise.
23830 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
23831 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
23832 (mips16_insn_length): Robustify.
23833 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
23834 (mips16_split_long_branches): Likewise.
23835 * config/pa/pa.c (pa_combine_instructions): Likewise.
23836 * config/rs6000/rs6000.c (get_next_active_insn): Treat
23837 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
23838 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
23839 as contributing to pool range lengths.
23840 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
23841 Remove redundant JUMP_TABLE_DATA_P test.
23842 (sh_loop_align): Likewise.
23843 (split_branches): Likewise.
23844 (sh_insn_length_adjustment): Likewise.
23845 * config/spu/spu.c (get_branch_target): Likewise.
23847 2013-03-29 Jan Hubicka <jh@suse.cz>
23849 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
23850 gcov streaming; stream hot bb threshold to ltrans.
23851 * predict.c (get_hot_bb_threshold): Break out from ....
23852 (maybe_hot_count_p): ... here.
23853 (set_hot_bb_threshold): New function.
23854 * lto-section-in.c (lto_section_name): Add profile.
23855 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
23856 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
23857 and data-streamer.h
23858 (histogram_entry): New structure.
23859 (histogram, histogram_pool): New global vars.
23860 (histogram_hash): New structure.
23861 (histogram_hash::hash): New method.
23862 (histogram_hash::equal): Likewise.
23863 (account_time_size): New function.
23864 (cmp_counts): New function.
23865 (dump_histogram): New function.
23866 (ipa_profile_generate_summary): New function.
23867 (ipa_profile_write_summary): New function.
23868 (ipa_profile_read_summary): New function.
23869 (ipa_profile): Decide on threshold.
23870 (pass_ipa_profile): Add ipa_profile_write_summary and
23871 ipa_profile_read_summary.
23872 * Makefile.in (ipa.o): Update dependencies.
23873 * lto-streamer.h (LTO_section_ipa_profile): New section.
23875 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
23877 * tree.h (VAR_P): New.
23879 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
23882 * doc/invoke.texi ([-fwhole-program]): Fix typo.
23884 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
23886 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
23887 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
23888 (control_flow_insn_p): Likewise.
23889 * cfgrtl.c (duplicate_insn_chain): Likewise.
23890 * final.c (get_attr_length_1): Likewise.
23891 (shorten_branches): Likewise.
23892 (final_scan_insn): Likewise.
23893 * function.c (instantiate_virtual_regs): Likewise.
23894 * gcse.c (insert_insn_end_basic_block): Likewise.
23895 * ira-costs.c (scan_one_insn): Likewise.
23896 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
23897 * lra.c (check_rtl): Likewise.
23898 * reload1.c (elimination_costs_in_insn): Likewise.
23899 * reorg.c (follow_jumps): Likewise.
23901 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
23902 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
23903 (thumb_far_jump_used_p): Likewise.
23904 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
23905 (workaround_speculation): Likewise.
23906 (add_sched_insns_for_speculation): Likewise.
23907 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
23908 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
23909 (frv_for_each_packet): Likewise.
23910 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
23911 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
23912 (final_emit_insn_group_barriers): Likewise.
23913 * config/m32r/m32r.c (m32r_is_insn): Likewise.
23914 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
23915 (mips16_insn_length): Likewise.
23916 * config/pa/pa.c (pa_reorg): Likewise.
23917 (pa_combine_instructions): Likewise.
23918 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
23919 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
23920 (sh_reorg): Likewise.
23921 (split_branches): Likewise.
23922 * config/spu/spu.c (get_branch_target): Likewise.
23924 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
23927 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
23929 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
23930 Fix declaration name.
23932 2013-03-28 Lawrence Crowl <crowl@google.com>
23934 * graphds.h (struct graph.indicies): Remove unused.
23935 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
23936 (SCOP_ORIGINAL_PDDRS): Remove unused.
23937 * sese.h (extern insert_loop_close_phis): Removed unused.
23938 (extern insert_guard_phis): Removed unused.
23939 (extern ivtype_map_elt_info): Removed unused.
23940 (new_ivtype_map_elt): Removed unused.
23941 * sese.c (ivtype_map_elt_info): Removed unused.
23943 2013-03-28 Lawrence Crowl <crowl@google.com>
23945 * Makefile.in: Add several missing include dependences.
23947 (test-dump.o): New. This object is not added to any executable,
23948 but is present for ad-hoc testing.
23950 (debug (const bitmap_head_def &)): New.
23951 (debug (const bitmap_head_def *)): New.
23953 (extern debug (const bitmap_head_def &)): New.
23954 (extern debug (const bitmap_head_def *)): New.
23956 (debug (edge_def &)): New.
23957 (debug (edge_def *)): New.
23959 (debug (basic_block_def &)): New.
23960 (debug (basic_block_def *)): New.
23962 (dump_node (const_tree, int, FILE *)): Correct source file.
23964 (debug (die_struct &)): New.
23965 (debug (die_struct *)): New.
23967 (extern debug (die_struct &)): New.
23968 (extern debug (die_struct *)): New.
23969 * gimple-pretty-print.c
23970 (debug (gimple_statement_d &)): New.
23971 (debug (gimple_statement_d *)): New.
23972 * gimple-pretty-print.h
23973 (extern debug (gimple_statement_d &)): New.
23974 (extern debug (gimple_statement_d *)): New.
23976 (debug (ira_allocno_copy &)): New.
23977 (debug (ira_allocno_copy *)): New.
23978 (debug (ira_allocno &)): New.
23979 (debug (ira_allocno *)): New.
23981 (extern debug (ira_allocno_copy &)): New.
23982 (extern debug (ira_allocno_copy *)): New.
23983 (extern debug (ira_allocno &)): New.
23984 (extern debug (ira_allocno *)): New.
23986 (debug (live_range &)): New.
23987 (debug (live_range *)): New.
23989 (debug (lra_live_range &)): New.
23990 (debug (lra_live_range *)): New.
23992 (debug (lra_live_range &)): New.
23993 (debug (lra_live_range *)): New.
23995 (debug (omega_pb_d &)): New.
23996 (debug (omega_pb_d *)): New.
23998 (extern debug (omega_pb_d &)): New.
23999 (extern debug (omega_pb_d *)): New.
24001 (debug (const rtx_def &)): New.
24002 (debug (const rtx_def *)): New.
24004 (debug_tree (tree): Move within file.
24005 (debug_raw (const tree_node &)): New.
24006 (debug_raw (const tree_node *)): New.
24007 (dump_tree_via_hooks (const tree_node *, int)): New.
24008 (debug (const tree_node &)): New.
24009 (debug (const tree_node *)): New.
24010 (debug_verbose (const tree_node &)): New.
24011 (debug_verbose (const tree_node *)): New.
24012 (debug_head (const tree_node &)): New.
24013 (debug_head (const tree_node *)): New.
24014 (debug_body (const tree_node &)): New.
24015 (debug_body (const tree_node *)): New.
24016 (debug_vec_tree (tree): Move and reimplement in terms of dump.
24017 (debug (vec<tree, va_gc> &)): New.
24018 (debug (vec<tree, va_gc> *)): New.
24020 (extern debug (const rtx_def &)): New.
24021 (extern debug (const rtx_def *)): New.
24023 (debug_raw (simple_bitmap_def &)): New.
24024 (debug_raw (simple_bitmap_def *)): New.
24025 (debug (simple_bitmap_def &)): New.
24026 (debug (simple_bitmap_def *)): New.
24028 (extern debug (simple_bitmap_def &)): New.
24029 (extern debug (simple_bitmap_def *)): New.
24030 (extern debug_raw (simple_bitmap_def &)): New.
24031 (extern debug_raw (simple_bitmap_def *)): New.
24033 (debug (vinsn_def &)): New.
24034 (debug (vinsn_def *)): New.
24035 (debug_verbose (vinsn_def &)): New.
24036 (debug_verbose (vinsn_def *)): New.
24037 (debug (expr_def &)): New.
24038 (debug (expr_def *)): New.
24039 (debug_verbose (expr_def &)): New.
24040 (debug_verbose (expr_def *)): New.
24041 (debug (vec<rtx> &)): New.
24042 (debug (vec<rtx> *)): New.
24044 (extern debug (vinsn_def &)): New.
24045 (extern debug (vinsn_def *)): New.
24046 (extern debug_verbose (vinsn_def &)): New.
24047 (extern debug_verbose (vinsn_def *)): New.
24048 (extern debug (expr_def &)): New.
24049 (extern debug (expr_def *)): New.
24050 (extern debug_verbose (expr_def &)): New.
24051 (extern debug_verbose (expr_def *)): New.
24052 (extern debug (vec<rtx> &)): New.
24053 (extern debug (vec<rtx> *)): New.
24055 (_list_iter_cond_expr): Make inline instead of static.
24057 (debug (sreal &)): New.
24058 (debug (sreal *)): New.
24060 (extern debug (sreal &)): New.
24061 (extern debug (sreal *)): New.
24063 (extern debug_raw (const tree_node &)): New.
24064 (extern debug_raw (const tree_node *)): New.
24065 (extern debug (const tree_node &)): New.
24066 (extern debug (const tree_node *)): New.
24067 (extern debug_verbose (const tree_node &)): New.
24068 (extern debug_verbose (const tree_node *)): New.
24069 (extern debug_head (const tree_node &)): New.
24070 (extern debug_head (const tree_node *)): New.
24071 (extern debug_body (const tree_node &)): New.
24072 (extern debug_body (const tree_node *)): New.
24073 (extern debug (vec<tree, va_gc> &)): New.
24074 (extern debug (vec<tree, va_gc> *)): New.
24076 (debug (struct loop &)): New.
24077 (debug (struct loop *)): New.
24078 (debug_verbose (struct loop &)): New.
24079 (debug_verbose (struct loop *)): New.
24080 * tree-dump.c: Add header dependence.
24082 (extern debug (struct loop &)): New.
24083 (extern debug (struct loop *)): New.
24084 (extern debug_verbose (struct loop &)): New.
24085 (extern debug_verbose (struct loop *)): New.
24087 (debug (data_reference &)): New.
24088 (debug (data_reference *)): New.
24089 (debug (vec<data_reference_p> &)): New.
24090 (debug (vec<data_reference_p> *)): New.
24091 (debug (vec<ddr_p> &)): New.
24092 (debug (vec<ddr_p> *)): New.
24094 (extern debug (data_reference &)): New.
24095 (extern debug (data_reference *)): New.
24096 (extern debug (vec<data_reference_p> &)): New.
24097 (extern debug (vec<data_reference_p> *)): New.
24098 (extern debug (vec<ddr_p> &)): New.
24099 (extern debug (vec<ddr_p> *)): New.
24101 (debug (pt_solution &)): New.
24102 (debug (pt_solution *)): New.
24104 (extern debug (pt_solution &)): New.
24105 (extern debug (pt_solution *)): New.
24107 (debug (_var_map &)): New.
24108 (debug (_var_map *)): New.
24109 (debug (tree_live_info_d &)): New.
24110 (debug (tree_live_info_d *)): New.
24112 (extern debug (_var_map &)): New.
24113 (extern debug (_var_map *)): New.
24114 (extern debug (tree_live_info_d &)): New.
24115 (extern debug (tree_live_info_d *)): New.
24117 2013-03-28 Jan Hubicka <jh@suse.cz>
24119 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
24121 2013-03-28 Ian Bolton <ian.bolton@arm.com>
24123 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
24124 record only when desired or required.
24126 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
24128 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
24129 *vec_extractv2di_1_rex64. Use x64 isa attribute.
24131 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24133 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
24134 (*andsi3_compare0_uxtw): New pattern.
24135 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
24136 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
24138 2013-03-28 Jan Hubicka <jh@suse.cz>
24140 * data-streamer-in.c (streamer_read_gcov_count): New function.
24141 * gimple-streamer-out.c: Include value-prof.h.
24142 (output_gimple_stmt): Output histogram.
24143 (output_bb): Use streamer_write_gcov_count.
24144 * value-prof.c: Include data-streamer.h
24145 (dump_histogram_value): Add HIST_TYPE_MAX.
24146 (stream_out_histogram_value): New function.
24147 (stream_in_histogram_value): New function.
24148 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
24149 (stream_out_histogram_value, stream_in_histogram_value): Declare.
24150 * data-streamer-out.c (streamer_write_gcov_count): New function.
24151 (streamer_write_gcov_count_stream): New function.
24152 * lto-cgraph.c (lto_output_edge): Update counter streaming.
24153 (lto_output_node): Likewise.
24154 (input_node, input_edge): Likewise.
24155 * lto-streamer-out.c (output_cfg): Update streaming.
24156 * lto-streamer-in.c (input_cfg): Likewise.
24157 * data-streamer.h (streamer_write_gcov_count,
24158 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
24159 * gimple-streamer-in.c: Include value-prof.h
24160 (input_gimple_stmt): Input histograms.
24161 (input_bb): Update profile streaming.
24163 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24165 * genmodes.c (emit_max_int): New function.
24166 (emit_insn_modes_h): Added call to emit_max_function.
24167 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
24169 * machmode.def: Fixed comment.
24171 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24173 * combine.c (try_combine): Removed useless assert.
24174 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
24176 2013-03-28 Marek Polacek <polacek@redhat.com>
24177 Richard Biener <rguenther@suse.de>
24179 PR tree-optimization/56695
24180 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
24181 build signed result of a vector comparison.
24182 * tree-cfg.c (verify_gimple_comparison): Check that a result
24183 of a vector comparison has signed type.
24185 2013-03-28 Richard Biener <rguenther@suse.de>
24187 PR tree-optimization/37021
24188 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
24189 do not restrict gaps between groups.
24190 * tree-vect-stmts.c (vectorizable_load): Properly account for
24191 a gap between groups.
24193 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
24195 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
24196 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
24199 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
24201 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
24202 * doc/extend.texi (Named Address Spaces): Ditto.
24203 (Variable Attributes): Ditto.
24205 2013-03-27 Kai Tietz <ktietz@redhat.com>
24207 * config.build: Add support for cygwin x64 target.
24208 * config.gcc: Likewise.
24209 * config.host: Likewise.
24210 * configure.ac: Likewise
24211 * configure: Regenerated.
24213 2013-03-27 Kai Tietz <ktietz@redhat.com>
24215 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
24216 * config/i386/t-cygwin-w64: New file.
24217 * config/i386/cygwin-w64.h: New file.
24218 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
24219 and add support for x64-cygwin target.
24220 (CPP_SPEC): Likewise.
24221 (CXX_WRAP_SPEC_LIST): Undefine before define.
24222 (LIBGCJ_SONAME): Use 15 as version.
24224 2013-03-27 Richard Biener <rguenther@suse.de>
24226 PR tree-optimization/56716
24227 * tree-ssa-structalias.c (perform_var_substitution): Adjust
24228 dumping for ref nodes.
24230 2013-03-27 Martin Jambor <mjambor@suse.cz>
24232 PR tree-optimization/55334
24233 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
24234 restricted pointers to arrays.
24236 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
24238 * Makefile.in (.SUFFIXES): Add .cc.
24239 (.c.o): Apply same recipe for implicit rule .cc.o.
24241 2013-03-27 Richard Biener <rguenther@suse.de>
24243 PR tree-optimization/37021
24244 * tree-vect-data-refs.c (vect_check_strided_load): Allow
24245 REALPART/IMAGPART_EXPRs around the supported refs.
24246 * tree-ssa-structalias.c (find_func_aliases): Assume that
24247 floating-point values are not used to transfer pointers.
24249 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
24251 * target.def (TARGET_HAS_IFUNC_P): New target hook.
24252 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
24253 * doc/tm.texi: Regenerate.
24254 * targhooks.h (default_has_ifunc_p): New.
24255 * targhooks.c (default_has_ifunc_p): Ditto.
24256 * config/linux-protos.h: New file.
24257 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
24258 hook for linux which disables support of indirect functions in android.
24259 * config/linux-android.c: New file.
24260 * config/t-linux-android.c: Ditto.
24261 * config.gcc: Added new object file linux-android.o.
24262 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
24263 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
24264 * varasm.c (do_assemble_alias): Likewise.
24265 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
24266 doesn't support indirect functions.
24267 * configure: Regenerate.
24269 2013-03-27 Bin Cheng <bin.cheng@arm.com>
24272 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
24273 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
24276 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24278 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
24280 2013-03-27 Terry Guo <terry.guo@arm.com>
24282 * config/arm/arm-cores.def: Added core cortex-r7.
24283 * config/arm/arm-tune.md: Regenerated.
24284 * config/arm/arm-tables.opt: Regenerated.
24285 * doc/invoke.texi: Added entry for core cortex-r7.
24287 2013-03-27 Walter Lee <walt@tilera.com>
24289 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
24290 double-decrement of next_scratch_regno.
24292 2013-03-27 Walter Lee <walt@tilera.com>
24294 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
24296 (insn_v1mulus): Ditto.
24297 (insn_v2muls): Ditto.
24299 2013-03-27 Walter Lee <walt@tilera.com>
24301 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
24302 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
24304 2013-03-27 Walter Lee <walt@tilera.com>
24306 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
24307 (*sibcall_value): Ditto.
24309 2013-03-27 Walter Lee <walt@tilera.com>
24311 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
24312 (insn_mnz_v8qi): ... this ...
24313 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
24315 (insn_v<n>mnz): Replaced by ...
24316 (insn_v1mnz): ... this ...
24317 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
24319 (insn_mz_<mode>): Replaced by ...
24320 (insn_mz_v8qi): ... this ...
24321 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
24323 (insn_v<n>mz): Replaced by ...
24324 (insn_v1mz): ... this ...
24325 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
24328 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
24330 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
24332 2013-03-26 Roland McGrath <mcgrathr@google.com>
24334 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
24335 than fprintf with a non-constant, non-format string.
24337 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24339 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
24340 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
24341 operand 0 predicate.
24342 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
24343 attribute. Use general_x64nomem_operand as operand 1 predicate.
24344 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
24345 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24346 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
24347 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24348 (mov<mode>_insv_1): Remove expander. Merge insn with
24349 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
24350 Use general_x64nomem_operand as operand 1 predicate.
24351 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
24352 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
24353 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24354 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
24355 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
24356 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
24357 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24358 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
24359 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
24360 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
24361 (general_x64nomem_operand): Ditto.
24363 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
24365 * config/rtems.opt: Add -pthread option.
24367 2013-03-26 Richard Biener <rguenther@suse.de>
24369 * alias.c (find_base_term): Avoid redundant and not used recursion.
24370 (base_alias_check): Get the initial base term from the caller.
24371 (true_dependence_1): Compute and pass base terms to base_alias_check.
24372 (write_dependence_p): Likewise.
24373 (may_alias_p): Likewise.
24375 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
24377 * config/aarch64/aarch64.c (aarch64_classify_address): Support
24378 PC-relative load in SI modes and above only.
24380 2013-03-26 Xinyu Qi <xyqi@marvell.com>
24382 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
24383 * config/arm/iwmmxt.md (WCGR0): Update.
24384 (WCGR1, WCGR2, WCGR3): Likewise.
24386 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24388 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
24389 Use x64 and nox64 isa attributes.
24391 2013-03-26 Richard Biener <rguenther@suse.de>
24393 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
24394 alignment computations and rely on get_object_alignment_1
24395 for the !TYPE_P case.
24396 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
24398 2013-03-26 Walter Lee <walt@tilera.com>
24400 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
24401 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
24403 2013-03-25 Jeff Law <law@redhat.com>
24405 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
24406 check for INTEGRAL_TYPE_P that was missing due to checking in
24407 wrong version of prior patch.
24409 2013-03-25 Walter Lee <walt@tilera.com>
24411 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
24412 TILEGX_INSN_SHUFFLEBYTES1.
24413 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
24415 (tilegx_builtins): Ditto.
24416 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
24418 2013-03-25 Walter Lee <walt@tilera.com>
24420 * config/tilegx/tilegx.md (floatsisf2): New pattern.
24421 (floatunssisf2): New pattern.
24422 (floatsidf2): New pattern.
24423 (floatunssidf2): New pattern.
24425 2013-03-25 Walter Lee <walt@tilera.com>
24427 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
24428 tests for constraint J, K, N, P.
24430 2013-03-25 Walter Lee <walt@tilera.com>
24432 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
24433 Use indirect/pcrel encoding.
24434 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
24437 2013-03-25 Steve Ellcey <sellcey@mips.com>
24439 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
24440 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
24441 * config/mips/mips.c (mips_option_override): Set IMADD default.
24442 * config/mips/mips.h (PTF_AVOID_IMADD): New.
24443 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
24444 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
24445 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
24446 * doc/invoke.texi (-mimadd/-mno-imadd): New.
24448 2013-03-25 Jeff Law <law@redhat.com>
24450 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
24451 slightly to avoid creating and folding useless trees. Simplify
24452 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
24454 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
24456 * config/i386/i386.md (*zero_extendsidi2): Merge with
24457 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
24458 * config/i386/predicates.md (x86_64_zext_operand): Rename from
24459 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
24460 targets. Clarify comment.
24462 2013-03-25 Martin Jambor <mjambor@suse.cz>
24464 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
24465 pass-through jump functions differently.
24466 (ipa_read_jump_function): Likewise. Also use setter functions to set
24469 2013-03-25 Martin Jambor <mjambor@suse.cz>
24471 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
24472 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
24474 (ipa_get_indirect_edge_target): New function.
24475 (devirtualization_time_bonus): New parameter known_aggs, pass it to
24476 ipa_get_indirect_edge_target. Update all callers.
24477 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
24478 ipa_get_indirect_edge_target_1 instead of calling
24479 ipa_get_indirect_edge_target.
24480 (create_specialized_node): Pass aggvlas to
24481 ipcp_discover_new_direct_edges.
24483 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24485 * config/arm/arm.md (f_sels, f_seld): New types.
24486 (*cmov<mode>): New pattern.
24487 * config/arm/predicates.md (arm_vsel_comparison_operator): New
24490 2013-03-25 Kai Tietz <ktietz@redhat.com>
24492 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
24493 POSIX-printf for mingw-hosted builds.
24495 2013-03-25 Richard Biener <rguenther@suse.de>
24497 PR middle-end/56694
24498 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
24499 must-not-throw stmt location.
24501 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24503 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
24504 Emit load-acquire versions when acq is true.
24505 (arm_emit_store_exclusive): Add rel parameter.
24506 Emit store-release versions when rel is true.
24507 (arm_split_compare_and_swap): Use acquire-release instructions
24509 of barriers when appropriate.
24510 (arm_split_atomic_op): Likewise.
24511 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
24512 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
24513 (VUNSPEC_SLX): Likewise.
24514 (VUNSPEC_LDA): Likewise.
24515 (VUNSPEC_STL): Likewise.
24516 * config/arm/sync.md (atomic_load<mode>): New pattern.
24517 (atomic_store<mode>): Likewise.
24518 (arm_load_acquire_exclusive<mode>): Likewise.
24519 (arm_load_acquire_exclusivesi): Likewise.
24520 (arm_load_acquire_exclusivedi): Likewise.
24521 (arm_store_release_exclusive<mode>): Likewise.
24523 2013-03-25 Catherine Moore <clm@codesourcery.com>
24525 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
24526 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
24527 * config/mip/predicates.md (lwsp_swsp_operand,
24528 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
24529 sb16_operand, db4_operand, db7_operand, ib3_operand,
24530 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
24531 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
24532 andi16_operand): New predicates.
24533 * config/mips/mips.md (compression): New attribute.
24534 (enabled): New attribute.
24535 (length): Consider compression in computing length.
24536 (shift_compression): New code attribute.
24537 (*add<mode>3): New operands. Record compression.
24538 (sub<mode>3): Likewise.
24539 (one_cmpl<mode>2): Likewise.
24540 (*and<mode>3): Likewise.
24541 (*ior<mode>3): Likewise.
24542 (unnamed pattern for xor): Likewise.
24543 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
24544 (*<optab><mode>3): Likewise.
24545 (*mov<mode>_internal: Likewise.
24546 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
24547 (mips_unsigned_immediate_p): New.
24548 (umips_lwsp_swsp_address_p): New.
24549 (m16_based_address_p): New.
24550 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
24551 (mips_unsigned_immediate_p): New prototype.
24552 (lwsp_swsp_address_p): New prototype.
24553 (m16_based_address_p): New prototype.
24554 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
24555 (mips_signed_immediate_p): New function.
24556 (m16_based_address_p): New function.
24557 (lwsp_swsp_address_p): New function.
24558 (mips_print_operand_punctuation): Recognize short delay slot insns
24559 for microMIPS.add<mode>3"
24561 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24564 * config/arm/iterators.md (v_cmp_result): New mode attribute.
24565 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
24567 2013-03-25 Richard Biener <rguenther@suse.de>
24569 PR tree-optimization/56689
24570 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
24573 2013-03-25 Richard Biener <rguenther@suse.de>
24575 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
24577 (memory_references): Likewise.
24578 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
24579 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
24580 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
24581 (gather_mem_refs_in_loops): Fold into ...
24582 (analyze_memory_references): ... this. Move initialization
24583 to tree_ssa_lim_initialize.
24584 (fill_always_executed_in): Rename to ...
24585 (fill_always_executed_in_1): ... this.
24586 (fill_always_executed_in): Move contains_call computation to
24587 this new function from ...
24588 (tree_ssa_lim_initialize): ... here.
24589 (tree_ssa_lim): Call fill_always_executed_in.
24591 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24593 * postreload.c (reload_combine): Fix code detecting returns.
24595 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24597 * function.c (emit_use_return_register_into_block): On cc0 targets,
24598 do not emit the sequence between cc0 setter and user.
24600 2013-03-25 Kai Tietz <ktietz@redhat.com>
24602 * config/i386/predicates.md (local_symbolic_operand): Interpret
24603 dll-imported symbols as none-local.
24605 2013-03-25 Richard Biener <rguenther@suse.de>
24607 * tree-ssa-loop-im.c (struct depend): Remove.
24608 (struct lim_aux_data): Make depends a vec of gimples.
24609 (free_lim_aux_data): Adjust.
24610 (add_dependency): Likewise.
24611 (set_level): Likewise.
24613 2013-03-25 Richard Biener <rguenther@suse.de>
24615 PR middle-end/56434
24616 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
24617 the pointer returned by calls with ECF_MALLOC set.
24619 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24621 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
24623 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24625 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
24626 using MMXMODE mode iterator.
24627 (*move<mode>_internal): Merge with *movv2sf_internal and
24628 *movv2sf_internal_rex64 using MMXMODE mode iterator.
24630 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24632 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
24633 (record_last_mem_set_info): Likewise.
24635 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
24636 of XNEWVEC followed by memset.
24637 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
24639 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24641 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
24642 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
24643 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
24644 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
24645 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
24646 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
24647 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
24648 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
24649 BARRIER_P instead of GET_CODE.
24651 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24653 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
24654 inaccuracy in the probing code.
24656 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
24657 (ctrapdi4): Likewise.
24659 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24661 * calls.c (expand_call): Add missing guard to code handling return
24662 of non-BLKmode structures in MSB.
24663 * function.c (expand_function_end): Likewise.
24665 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24667 * combine.c (try_combine): Adjust comment. Do not add the set of
24668 insn #0 if the destination indirectly is set or dies in insn #2.
24669 Tidy up code to distribute a new note.
24671 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24673 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
24674 also for alternatives 16 and 17.
24676 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24678 * config/i386/sse.md (*mov<mode>_internal): Merge with
24679 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
24680 Emit insn template depending on type attribute. Use
24681 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24682 movd instead of movq mnemonic for interunit moves. Rewrite mode
24683 attribute calculation. Remove unit attribute calculation.
24684 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
24685 Set prefix_data16 attribute for DImode ssemov types.
24686 Use Ym instead of y for SSE-MMX conversion alternatives.
24687 Reorder operand constraints.
24689 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
24691 * df.h (df_insn_delete): Adjust prototype.
24692 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
24693 and let it decide whether mark the basic block dirty.
24694 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
24695 * df-scan.c (df_insn_info_delete): New helper function, split
24696 off from df_insn_delete.
24697 (df_scan_free_bb_info): Use it.
24698 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
24700 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
24701 that the insn is actually an insn and it has a non-NULL basic block.
24702 Do not mark basic block dirty if only deleting a DEBUG_INSN.
24704 2013-03-22 Richard Biener <rguenther@suse.de>
24706 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
24708 (mem_ref_alloc): Do not allocate them.
24709 (refs_independent_p): Do not query or maintain a cache.
24711 2013-03-22 Richard Biener <rguenther@suse.de>
24713 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
24714 (gather_mem_refs_in_loops): Do not compute it.
24715 (analyze_memory_references): Do not allocate it.
24716 (tree_ssa_lim_finalize): Do not free it.
24717 (for_all_locs_in_loop): Do not query all_refs_in_loop.
24719 2013-03-22 Richard Biener <rguenther@suse.de>
24721 * is-a.h (as_a): Use gcc_checking_assert.
24723 2013-03-22 Ian Bolton <ian.bolton@arm.com>
24725 * config/aarch64/aarch64.c (aarch64_print_operand): New
24726 format specifier for printing a constant in hex.
24727 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
24728 format specifier for printing second operand.
24730 2013-03-22 Richard Biener <rguenther@suse.de>
24732 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
24734 (gather_mem_refs_in_loops): Perform store accumulation here.
24735 (create_vop_ref_mapping_loop): Remove.
24736 (create_vop_ref_mapping): Likewise.
24737 (analyze_memory_references): Initialize refs_stored_in_loop.
24738 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
24739 (record_indep_loop): Remove.
24740 (record_dep_loop): New function.
24741 (ref_indep_loop_p_1): Adjust to only walk over references
24742 in the loop, not its subloops.
24743 (ref_indep_loop_p): Rename to ...
24744 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
24745 maintaining a more fine-grained cache.
24746 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
24747 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
24749 2013-03-22 Richard Biener <rguenther@suse.de>
24751 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
24752 (struct mem_ref): Make accesses_in_loop a vec of a vec of
24753 aggregate mem_ref_loc.
24754 (free_mem_ref_locs): Inline into ...
24755 (memref_free): ... this and adjust.
24756 (mem_ref_alloc): Adjust.
24757 (mem_ref_locs_alloc): Remove.
24758 (record_mem_ref_loc): Adjust.
24759 (get_all_locs_in_loop): Rewrite into ...
24760 (for_all_locs_in_loop): ... this iterator.
24761 (rewrite_mem_ref_loc): New functor.
24762 (rewrite_mem_refs): Use for_all_locs_in_loop.
24763 (sm_set_flag_if_changed): New functor.
24764 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
24765 (ref_always_accessed): New functor.
24766 (ref_always_accessed_p): Use for_all_locs_in_loop.
24768 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24770 * tree-pass.h (PROP_gimple_lvec): New.
24771 * passes.c (dump_properties): Handle PROP_gimple_lvec.
24772 (init_optimization_passes): Move pass_lower_vector.
24773 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
24775 (pass_lower_vector): Provide PROP_gimple_lvec.
24776 (pass_lower_vector_ssa): Likewise.
24777 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
24779 2013-03-21 Mark Wielaard <mjw@redhat.com>
24781 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
24783 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
24785 * config/i386/i386.md (*movdi_internal): Disparage slightly
24786 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
24787 conversion alternatives.
24789 2013-03-21 Jakub Jelinek <jakub@redhat.com>
24791 PR middle-end/48087
24792 * diagnostic.def (DK_WERROR): New kind.
24793 * diagnostic.h (werrorcount): Define.
24794 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
24795 promoted to DK_ERROR, increment DK_WERROR counter instead of
24797 * toplev.c (toplev_main): Call print_ignored_options even if
24798 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
24799 even if just werrorcount is non-zero.
24802 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
24804 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
24805 (string_cst_pool_decl): New function.
24806 (optimize_one_addr_into_implicit_ptr): New function.
24807 (resolve_addr_in_expr): Optimize DWARF location expression
24808 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
24809 which doesn't live in memory, but has DW_AT_location or
24810 DW_AT_const_value, or refers to a string literal, into
24811 DW_OP_GNU_implicit_pointer.
24812 (optimize_location_into_implicit_ptr): New function.
24813 (resolve_addr): If removing DW_AT_location of a variable because
24814 it was DW_OP_addr of address of the variable, but the variable doesn't
24815 live in memory, try to emit const value attribute for the initializer.
24817 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24819 * tree.h (VECTOR_TYPE_P): New macro.
24820 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
24821 TYPE_MODE): Use it.
24822 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
24823 VEC_COND_EXPR cannot be lvalues.
24824 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
24826 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24828 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
24829 Restrict the transformation to equal modes.
24831 2013-03-21 Richard Biener <rguenther@suse.de>
24833 PR tree-optimization/39326
24834 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
24835 (MEM_ANALYZABLE): Adjust.
24836 (record_mem_ref_loc): Move bitmap ops ...
24837 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
24838 unanalyzable refs, do not record locations for it.
24839 (analyze_memory_references): Allocate ref zero as shared
24841 (refs_independent_p): Do not test for unanalyzed mems here.
24842 (ref_indep_loop_p_1): Special-case disambiguation against
24843 the unanalyzed ref.
24844 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
24846 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
24848 * config/arm/arm-protos.h (tune_params): Add
24849 prefer_neon_for_64bits field.
24850 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
24851 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
24852 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
24853 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
24854 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
24855 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
24856 (arm_option_override): Handle -mneon-for-64bits new option.
24857 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
24858 (prefer_neon_for_64bits): Declare new variable.
24859 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
24860 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
24861 (arch_enabled): Handle new arch types. Remove support for onlya8
24863 (one_cmpldi2): Use new arch names.
24864 (zero_extend<mode>di2, extend<mode>di2): Ditto.
24865 * config/arm/arm.opt (mneon-for-64bits): Add option.
24866 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
24867 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
24868 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
24870 * doc/invoke.texi (-mneon-for-64bits): Document.
24872 2013-03-21 Richard Biener <rguenther@suse.de>
24874 PR tree-optimization/39326
24875 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
24876 (sort_bbs_in_loop_postorder_cmp): New function.
24877 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
24879 2013-03-21 Richard Biener <rguenther@suse.de>
24881 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
24882 (vect_insert_into_interleaving_chain): Likewise.
24883 (vect_drs_dependent_in_basic_block): Inline ...
24884 (vect_slp_analyze_data_ref_dependence): ... here. New function,
24886 (vect_analyze_data_ref_dependence): ... here. Simplify.
24887 (vect_check_interleaving): Simplify.
24888 (vect_analyze_data_ref_dependences): Likewise. Split out ...
24889 (vect_slp_analyze_data_ref_dependences): ... this new function.
24890 (dr_group_sort_cmp): New function.
24891 (vect_analyze_data_ref_accesses): Compute data-reference groups
24892 here instead of in vect_analyze_data_ref_dependence. Use
24893 a more efficient algorithm.
24894 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
24895 vect_slp_analyze_data_ref_dependences. Call
24896 vect_analyze_data_ref_accesses earlier.
24897 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
24898 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
24899 (vect_slp_analyze_data_ref_dependences): New prototype.
24901 2013-03-21 Richard Biener <rguenther@suse.de>
24903 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
24904 ref is stored in the loop.
24905 (find_refs_for_sm): Walk only over all stores.
24906 (store_motion_loop): Allocate from lim_bitmap_obstack.
24907 (store_motion): Likewise.
24909 2013-03-21 Richard Biener <rguenther@suse.de>
24911 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
24912 Update virtual SSA form.
24914 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
24916 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
24917 * configure: Regenerate.
24918 * config.in: Regenerate.
24919 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
24920 if !HAVE_LD_EH_FRAME_CIEV3.
24922 2013-03-21 Richard Biener <rguenther@suse.de>
24924 * tree-cfg.c (verify_expr_no_block): New function.
24925 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
24926 nor DECL_VALUE_EXPR have locations with associated blocks.
24927 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
24928 (clear_unused_block_pointer): Remove code dealing with
24929 blocks in DECL_DEBUG_EXPR locations.
24931 2013-03-21 Richard Biener <rguenther@suse.de>
24933 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
24934 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
24935 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
24936 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
24937 instead of DECL_DEBUG_EXPR_IS_FROM.
24938 * gimplify.c (gimplify_modify_expr): Likewise.
24939 * tree-cfg.c (verify_expr_location_1): Likewise.
24940 * tree-complex.c (create_one_component_var): Likewise.
24941 * tree-sra.c (create_access_replacement): Likewise.
24942 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
24943 (clear_unused_block_pointer): Likewise.
24944 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
24945 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
24946 * var-tracking.c (var_debug_decl): Likewise.
24947 (track_expr_p): Likewise.
24948 * tree-inline.c (add_local_variables): Likewise. Set
24949 DECL_HAS_DEBUG_EXPR_P after copying it.
24950 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
24951 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
24953 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
24956 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
24957 * configure: Regenerate.
24958 * config.in: Regenerate.
24959 * config/i386/i386.md (*movdf_internal): Use
24960 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24961 movd instead of movq mnemonic for interunit moves.
24962 (*movdi_internal): Ditto.
24964 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24966 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
24967 (abd<mode>_3): New pattern.
24968 (aba<mode>_3): New pattern.
24969 (fabd<mode>_3): New pattern.
24971 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24973 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
24974 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
24975 occurrence of REGISTER_PREFIX as its empty string.
24977 2013-03-20 Jeff Law <law@redhat.com>
24979 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
24980 addititional equivalences for equality comparisons between an SSA_NAME
24981 and a constant where the SSA_NAME was set from a widening conversion.
24983 2013-03-20 Walter Lee <walt@tilera.com>
24985 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
24987 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
24989 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
24990 depending on type attribute.
24991 (*movti_internal): Ditto.
24992 (*movtf_internal): Ditto.
24993 (*movxf_internal): Ditto.
24994 (*movdf_internal): Ditto.
24995 (*movsf_internal): Ditto.
24997 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
24999 * config/i386/i386.md (*movti_internal): Set prefix attribute to
25000 maybe_vex for sselog1 and ssemov types.
25001 (*movdi_internal): Reorder operand constraints.
25002 (*movsi_internal): Ditto. Set prefix attribute to
25003 maybe_vex for sselog1 and ssemov types.
25004 (*movtf_internal): Set prefix attribute to maybe_vex
25005 for sselog1 and ssemov types.
25006 (*movdf_internal): Ditto. Set prefix_data16 attribute for
25007 DImode ssemov types. Reorder operand constraints.
25008 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
25009 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
25010 attribute for SImode ssemov types. Reorder operand constraints.
25012 2013-03-20 Martin Jambor <mjambor@suse.cz>
25014 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
25015 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
25017 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
25019 * config/rs6000/predicates.md (indexed_address, update_address_mem
25020 update_indexed_address_mem): New predicates.
25021 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
25022 attribute for load/store instructions.
25023 * config/rs6000/dfp.md (movsd_store): Likewise.
25024 (movsd_load): Likewise.
25025 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
25026 (unnamed HI->DI extend define_insn): Likewise.
25027 (unnamed SI->DI extend define_insn): Likewise.
25028 (unnamed QI->SI extend define_insn): Likewise.
25029 (unnamed QI->HI extend define_insn): Likewise.
25030 (unnamed HI->SI extend define_insn): Likewise.
25031 (unnamed HI->SI extend define_insn): Likewise.
25032 (extendsfdf2_fpr): Likewise.
25033 (movsi_internal1): Likewise.
25034 (movsi_internal1_single): Likewise.
25035 (movhi_internal): Likewise.
25036 (movqi_internal): Likewise.
25037 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
25038 attribute for load/store instructions.
25039 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
25041 (mov<mode>_softfloat): Likewise.
25042 (mov<mode>_hardfloat32): Likewise.
25043 (mov<mode>_hardfloat64): Likewise.
25044 (mov<mode>_softfloat64): Likewise.
25045 (movdi_internal32): Likewise.
25046 (movdi_internal64): Likewise.
25047 (probe_stack_<mode>): Likewise.
25049 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
25051 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
25052 floating point, and decimal floating point to reload iterator.
25054 * config/rs6000/constraints.md (wl constraint): New constraints to
25055 return FLOAT_REGS if certain options are used to reduce the number
25056 of separate patterns that exist in the file.
25057 (wx constraint): Likewise.
25058 (wz constraint): Likewise.
25060 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25061 -mdebug=reg, print wg, wl, wx, and wz constraints.
25062 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25063 Initialize the reload functions for 64-bit binary/decimal floating
25065 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25066 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25067 create the buffer on the stack to overcome not having a 32-bit
25069 (rs6000_emit_move): Likewise.
25070 (rs6000_secondary_memory_needed_rtx): Likewise.
25071 (rs6000_alloc_sdmode_stack_slot): Likewise.
25072 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25073 via xxlxor, just like DFmode 0.0.
25075 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
25076 define as 1 if we are running on a power7 or newer.
25077 (enum r6000_reg_class_enum): Add new constraints.
25079 * config/rs6000/dfp.md (movsd): Delete, combine with binary
25080 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25081 with other moves by using conditional constraits (wg). Use LFIWZX
25082 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
25083 (movsd splitter): Likewise.
25084 (movsd_hardfloat): Likewise.
25085 (movsd_softfloat): Likewise.
25087 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
25088 binary and decimal floating point moves.
25089 (fmove_ok): New attributes to combine binary and decimal floating
25090 point moves, and to combine power6x (mfpgpr) moves along normal
25092 (real_value_to_target): Likewise.
25093 (f32_lr): Likewise.
25094 (f32_lm): Likewise.
25095 (f32_li): Likewise.
25096 (f32_sr): Likewise.
25097 (f32_sm): Likewise.
25098 (f32_si): Likewise.
25099 (movsf): Combine binary and decimal floating point moves. Combine
25100 power6x (mfpgpr) moves with other moves by using conditional
25101 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
25102 (mov<mode> for SFmode/SDmode); Likewise.
25103 (SFmode/SDmode splitters): Likewise.
25104 (movsf_hardfloat): Likewise.
25105 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
25106 (movsf_softfloat): Likewise.
25107 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
25109 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
25110 wx and wz constraints.
25112 * config/rs6000/constraints.md (wg constraint): New constraint to
25113 return FLOAT_REGS if -mmfpgpr (power6x) was used.
25115 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
25118 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25119 -mdebug=reg, print wg, wl, wx, and wz constraints.
25120 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25121 Initialize the reload functions for 64-bit binary/decimal floating
25123 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25124 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25125 create the buffer on the stack to overcome not having a 32-bit
25127 (rs6000_emit_move): Likewise.
25128 (rs6000_secondary_memory_needed_rtx): Likewise.
25129 (rs6000_alloc_sdmode_stack_slot): Likewise.
25130 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25131 via xxlxor, just like DFmode 0.0.
25133 * config/rs6000/dfp.md (movdd): Delete, combine with binary
25134 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25135 with other moves by using conditional constraits (wg). Use LFIWZX
25136 and STFIWX for loading SDmode on power7.
25137 (movdd splitters): Likewise.
25138 (movdd_hardfloat32): Likewise.
25139 (movdd_softfloat32): Likewise.
25140 (movdd_hardfloat64_mfpgpr): Likewise.
25141 (movdd_hardfloat64): Likewise.
25142 (movdd_softfloat64): Likewise.
25144 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
25145 64-bit binary and decimal floating point moves.
25146 (FMOVE64X): Likewise.
25147 (movdf): Combine 64-bit binary and decimal floating point moves.
25148 Combine power6x (mfpgpr) moves with other moves by using
25149 conditional constraits (wg).
25150 (mov<mode> for DFmode/DDmode): Likewise.
25151 (DFmode/DDmode splitters): Likewise.
25152 (movdf_hardfloat32): Likewise.
25153 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
25154 (movdf_softfloat32): Likewise.
25155 (movdf_hardfloat64_mfpgpr): Likewise.
25156 (movdf_hardfloat64): Likewise.
25157 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
25158 (movdf_softfloat64): Likewise.
25159 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
25160 (reload_<mode>_load): Move to later in the file so they aren't in
25161 the middle of the floating point move insns.
25162 (reload_<mode>_store): Likewise.
25164 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
25167 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
25168 constraint if -mdebug=reg.
25169 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
25170 Enable using dd reload support if needed.
25172 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
25173 binary and decimal floating point moves in rs6000.md.
25174 (movtd_internal): Likewise.
25176 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
25177 decimal floating point moves.
25179 (movtf_internal): Likewise.
25180 (mov<mode>_internal, TDmode/TFmode): Likewise.
25181 (movtf_softfloat): Likewise.
25182 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
25184 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
25185 movdi_internal64, using wg constraint for move direct operations.
25186 (movdi_internal64): Likewise.
25188 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
25189 MODES_TIEABLE_P for selected modes. Print the numerical value of
25190 the various virtual registers. Use GPR/FPR first/last values,
25191 instead of hard coding the register numbers. Print which modes
25192 have reload functions registered.
25193 (rs6000_option_override_internal): If -mdebug=reg, trace the options
25194 settings before/after setting cpu, target and subtarget settings.
25195 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
25196 and for secondary reload failures in rs6000_secondary_reload_inner.
25197 (rs6000_secondary_reload_fail): Likewise.
25198 (rs6000_secondary_reload_inner): Likewise.
25200 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
25201 macros for first/last GPR and FPR registers.
25202 (LAST_GPR_REGNO): Likewise.
25203 (FIRST_FPR_REGNO): Likewise.
25204 (LAST_FPR_REGNO): Likewise.
25206 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
25207 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
25208 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
25209 (vcond<mode><mode>): Likewise.
25210 (vcondu<mode><mode>): Likewise.
25211 (vector_gtu<mode>): Likewise.
25212 (vector_gte<mode>): Likewise.
25213 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
25214 to prevent the compiler from converting DImode operations to TImode.
25215 (ior<mode>3): Likewise.
25216 (and<mode>3): Likewise.
25217 (one_cmpl<mode>2): Likewise.
25218 (nor<mode>3): Likewise.
25219 (andc<mode>3): Likewise.
25221 * config/rs6000/constraints.md (wt constraint): New constraint
25222 that returns VSX_REGS if TImode is allowed in VSX registers.
25224 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
25225 constant under VSX.
25227 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
25228 similar to TImode, but it is restricted to being in the GPRs.
25230 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
25231 TImode to occupy a single VSX register.
25233 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
25234 -mvsx-timode for power7/power8.
25235 (power7 cpu): Likewise.
25236 (power8 cpu): Likewise.
25238 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
25239 sure that TFmode/TDmode take up two registers if they are ever
25240 allowed in the upper VSX registers.
25241 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
25243 (rs6000_init_hard_regno_mode_ok): Likewise.
25244 (rs6000_debug_reg_global): Add debugging for PTImode and wt
25245 constraint. Print if LRA is turned on.
25246 (rs6000_option_override_internal): Give an error if -mvsx-timode
25247 and VSX is not enabled.
25248 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
25249 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
25250 to reg+offset addressing. Use PTImode when checking offset
25251 addresses for validity.
25252 (reg_offset_addressing_ok_p): Likewise.
25253 (rs6000_legitimate_offset_address_p): Likewise.
25254 (rs6000_legitimize_address): Likewise.
25255 (rs6000_legitimize_reload_address): Likewise.
25256 (rs6000_legitimate_address_p): Likewise.
25257 (rs6000_eliminate_indexed_memrefs): Likewise.
25258 (rs6000_emit_move): Likewise.
25259 (rs6000_secondary_reload): Likewise.
25260 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
25261 reloads to fpr registers to continue to use reg+offset addressing,
25262 but 64-bit reloads to altivec registers need reg+reg addressing.
25263 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
25264 it. Treat LO_SUM like a PLUS operation.
25265 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
25266 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
25267 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
25268 registers to share a register with a smaller sized type, since VSX
25269 puts scalars in the upper 64-bits.
25270 (print_operand): Add support for PTImode.
25271 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
25272 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
25273 registers, but don't have arithmetic support.
25274 (rs6000_memory_move_cost): Add test for VSX.
25275 (rs6000_opt_masks): Add -mvsx-timode.
25277 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
25280 (VSr): Use wt constraint for TImode.
25281 (VSv): Drop TImode support.
25282 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
25283 (vsx_movti_64bit): Likewise.
25284 (vsx_movti_32bit): Likewise.
25285 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
25286 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
25287 one '?' on the appropriate output constraint. Do not allow TImode
25288 logical operations on 32-bit systems.
25289 (vsx_ior<mode>3): Likewise.
25290 (vsx_xor<mode>3): Likewise.
25291 (vsx_one_cmpl<mode>2): Likewise.
25292 (vsx_nor<mode>3): Likewise.
25293 (vsx_andc<mode>3): Likewise.
25294 (vsx_concat_<mode>): Likewise.
25295 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
25297 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
25298 OPTION_MASK_VSX_TIMODE.
25299 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
25300 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
25302 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
25303 (TI2 iterator): New iterator for TImode, PTImode.
25304 (wd mode attribute): Add values for vector types.
25305 (movti_string): Replace TI move operations with operations for TImode
25306 and PTImode. Add support for TImode being allowed in VSX registers.
25307 (mov<mode>_string, TImode/PTImode): Likewise.
25308 (movti_ppc64): Likewise.
25309 (mov<mode>_ppc64, TImode/PTImode): Likewise.
25310 (TI mode splitters): Likewise.
25312 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
25315 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
25317 PR tree-optimization/56355
25318 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
25319 Also handle integers with undefined overflow.
25321 2013-03-20 Catherine Moore <clm@codesourcery.com>
25322 Maciej W. Rozycki <macro@codesourcery.com>
25323 Tom de Vries <tom@codesourcery.com>
25324 Nathan Sidwell <nathan@codesourcery.com>
25325 Iain Sandoe <iain@codesourcery.com>
25326 Nathan Froyd <froydnj@codesourcery.com>
25327 Chao-ying Fu <fu@mips.com>
25329 * doc/extend.texi: (micromips, nomicromips, nocompression):
25330 Document new function attributes.
25331 * doc/invoke.texi (minterlink-compressed, mmicromips,
25332 m14k, m14ke, m14kec): Document new options.
25333 (minterlink-mips16): Update documentation.
25334 * doc/md.texi (ZC, ZD): Document new constraints.
25335 * configure.ac (gcc_cv_as_micromips): Check if linker
25336 supports the .set micromips directive.
25337 * configure: Regenerate.
25338 * config.in: Regenerate.
25339 * config/mips/mips-tables.opt: Regenerate.
25340 * config/mips/micromips.md: New file.
25341 * constraints.md (ZC, ZD): New constraints.
25342 * config/mips/predicates.md (movep_src_register): New predicate.
25343 (movep_src_operand): New predicate.
25344 (non_volatile_mem_operand): New predicate.
25345 * config/mips/mips.md (multimem): New type.
25346 (length): Differentiate between 17-bit and 18-bit branch offsets.
25347 (MOVEP1, MOVEP2): New mode iterator.
25348 (mov_<load>l): Use ZC constraint.
25349 (mov_<load>r): Likewise.
25350 (mov_<store>l): Likewise.
25351 (mov_<store>r): Likewise.
25352 (*branch_equality<mode>_inverted): Add microMIPS support.
25353 (*branch_equality<mode>): Likewise.
25354 (*jump_absolute): Likewise.
25355 (indirect_jump_<mode>): Likewise.
25356 (tablejump_<mode>): Likewise.
25357 (<optab>_internal): Likewise.
25358 (sibcall_internal): Likewise.
25359 (sibcall_value_internal): Likewise.
25360 (prefetch): Use constraint ZD.
25361 * config/mips/mips.opt (minterlink-compressed): New option.
25362 (minterlink-mips16): Now an alias for minterlink-compressed.
25363 (mmicromips): New option.
25364 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
25365 (compare_and_swap_12): Likewise.
25366 (sync_add<mode>): Likewise.
25367 (sync_<optab>_12): Likewise.
25368 (sync_old_<optab>_12): Likewise.
25369 (sync_new_<optab>_12): Likewise.
25370 (sync_nand_12): Likewise.
25371 (sync_old_nand_12): Likewise.
25372 (sync_new_nand_12): Likewise.
25373 (sync_sub<mode>): Likewise.
25374 (sync_old_add<mode>): Likewise.
25375 (sync_old_sub<mode>): Likewise.
25376 (sync_new_add<mode>): Likewise.
25377 (sync_new_sub<mode>): Likewise.
25378 (sync_<optab><mode>): Likewise.
25379 (sync_old_<optab><mode>): Likewise.
25380 (sync_new_<optab><mode>): Likewise.
25381 (sync_nand<mode>): Likewise.
25382 (sync_old_nand<mode>): Likewise.
25383 (sync_new_nand<mode>): Likewise.
25384 (sync_lock_test_and_set<mode>): Likewise.
25385 (test_and_set_12): Likewise.
25386 (atomic_compare_and_swap<mode>): Likewise.
25387 (atomic_exchange<mode>_llsc): Likewise.
25388 (atomic_fetch_add<mode>_llsc): Likewise.
25389 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
25390 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
25391 (umips_save_restore_pattern_p): Likewise.
25392 (umips_load_store_pair_p): Likewise.
25393 (umips_output_load_store_pair): Likewise.
25394 (umips_movep_target_p): Likewise.
25395 (umips_12bit_offset_address_p): Likewise.
25396 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
25397 (mips_base_mips16): Rename this...
25398 (mips_base_compression_flags): ...to this. Update all uses.
25399 (mips_attribute_table): Add micromips, nomicromips and nocompression.
25400 (mips_mips16_decl_p): Delete.
25401 (mips_nomips16_decl_p): Delete.
25402 (mips_get_compress_on_flags): New function.
25403 (mips_get_compress_off_flags): New function.
25404 (mips_get_compress_mode): New function.
25405 (mips_get_compress_on_name): New function.
25406 (mips_get_compress_off_name): New function.
25407 (mips_insert_attributes): Support multiple compression types.
25408 (mips_merge_decl_attributes): Likewise.
25409 (umips_12bit_offset_address_p): New function.
25410 (mips_start_function_definition): Emit .set micromips directive.
25411 (mips_call_may_need_jalx_p): New function.
25412 (mips_function_ok_for_sibcall): Add microMIPS support.
25413 (mips_print_operand_punctuation): Support short delay slots and
25415 (umips_swm_mask, umips_swm_encoding): New.
25416 (umips_build_save_restore): New function.
25417 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
25418 (was_mips16_p): Remove.
25419 (old_compression_mode): New.
25420 (mips_set_compression_mode): New function.
25421 (mips_set_current_function): Add microMIPS support.
25422 (mips_option_override): Likewise.
25423 (umips_save_restore_pattern_p): New function.
25424 (umips_output_save_restore): New function.
25425 (umips_load_store_pair_p_1): New function.
25426 (umips_load_store_pair_p): New function.
25427 (umips_output_load_store_pair_1): New function.
25428 (umips_output_load_store_pair): New function.
25429 (umips_movep_target_p) New function.
25430 (mips_prepare_pch_save): Add microMIPS support.
25431 * config/mips/mips.h (TARGET_COMPRESSION): New.
25432 (TARGET_CPU_CPP_BUILTINS): Update macro
25433 to use new compression flags and to support microMIPS.
25434 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
25435 (MIPS_ARCH_FLOAT_SPEC): Likewise.
25436 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
25437 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
25438 (ASM_SPEC): Support mmicromips and mno-micromips.
25439 (M16STORE_REG_P): New macro.
25440 (MIPS_CALL): Support TARGET_MICROMIPS.
25441 (MICROMIPS_J): New macro.
25442 (mips_base_mips16): Rename this...
25443 (mips_base_compression_flags): ...to this.
25444 (UMIPS_12BIT_OFFSET_P): New macro.
25445 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
25446 (MULTILIB_DIRNAMES): Likewise.
25447 2013-03-20 Richard Biener <rguenther@suse.de>
25449 PR tree-optimization/56661
25450 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
25451 the result does not have to be distinct.
25453 2013-03-20 Richard Biener <rguenther@suse.de>
25455 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
25458 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
25459 Steven Bosscher <steven@gcc.gnu.org>
25461 PR rtl-optimization/56605
25462 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
25464 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25467 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
25468 that require movd instead of movq.
25470 2013-03-20 Richard Biener <rguenther@suse.de>
25472 * tree-ssa-structalias.c (struct variable_info): Add pointer
25473 to the first field of an aggregate with sub-vars. Make
25474 this and the pointer to the next subfield its ID.
25475 (vi_next): New function.
25476 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
25477 storedanything_id, integer_id): Increment by one.
25478 (new_var_info, get_call_vi, lookup_call_clobber_vi,
25479 get_call_clobber_vi): Adjust.
25480 (solution_set_expand): Simplify and speedup.
25481 (solution_set_add): Inline into ...
25482 (set_union_with_increment): ... this. Adjust accordingly.
25483 (do_sd_constraint): Likewise.
25484 (do_ds_constraint): Likewise.
25485 (do_complex_constraint): Simplify.
25486 (build_pred_graph): Adjust.
25487 (solve_graph): Likewise. Simplify and speedup.
25488 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
25489 get_constraint_for_component_ref, get_constraint_for_1,
25490 first_vi_for_offset, first_or_preceding_vi_for_offset,
25491 create_function_info_for, create_variable_info_for_1,
25492 create_variable_info_for, intra_create_variable_infos): Adjust.
25493 (init_base_vars): Push NULL for ID zero.
25494 (compute_points_to_sets): Adjust.
25496 2013-03-20 Richard Biener <rguenther@suse.de>
25498 * cfgloop.c (verify_loop_structure): Streamline and avoid
25499 ICEing on corrupt loop tree.
25500 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
25503 2013-03-20 Richard Biener <rguenther@suse.de>
25505 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
25506 check whether an SSA update is needed.
25508 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
25510 * config/mips/constraints.md (T): Rename to...
25514 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
25515 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
25517 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25519 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
25520 (*subsi3_carryin_uxtw): Likewise.
25522 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25524 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
25525 (*rorsi3_insn_uxtw): Likewise.
25527 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25529 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
25530 (*extrsi5_insn_uxtw): Likewise.
25532 2013-03-19 Richard Biener <rguenther@suse.de>
25534 PR tree-optimization/56273
25535 * passes.c (init_optimization_passes): Move second VRP after DOM.
25537 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
25539 * config/i386/i386.md (*movti_internal): Merge from
25540 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
25541 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
25542 nox64 isa attributes.
25544 2013-03-18 Richard Biener <rguenther@suse.de>
25546 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
25548 (merge_node_constraints): Likewise.
25549 (build_succ_graph): Likewise.
25550 (valid_graph_edge): Inline into single caller.
25551 (unify_nodes): Likewise. Use bitmap_set_bit return value
25553 (scc_visit): Fix formatting and variable use.
25554 (do_sd_constraint): Use gcc_checking_assert.
25555 (do_ds_constraint): Likewise.
25556 (do_complex_constraint): Likewise.
25557 (condense_visit): Likewise. Cleanup.
25558 (dump_pred_graph): New function.
25559 (perform_var_substitution): Dump the pred-graph before
25560 variable substitution.
25561 (find_equivalent_node): Use gcc_checking_assert.
25562 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
25564 2013-03-18 Richard Biener <rguenther@suse.de>
25566 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
25567 Remove cond_expr_stmt_list argument and do not gimplify the
25569 (vect_loop_versioning): Adjust.
25570 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
25571 Cleanup to use less temporaries.
25572 (vect_create_data_ref_ptr): Cleanup.
25574 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25576 PR tree-optimization/56635
25577 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
25578 require types_compatible_p types.
25580 2013-03-18 Nick Clifton <nickc@redhat.com>
25582 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
25583 spurious backslash.
25585 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
25586 Add missing line to comment describing function.
25588 2013-03-18 Richard Biener <rguenther@suse.de>
25590 PR tree-optimization/56210
25591 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
25592 Handle string / character search functions.
25593 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
25595 2013-03-18 Richard Biener <rguenther@suse.de>
25597 PR middle-end/56483
25598 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
25599 and implement properly.
25600 * gimple.h (gimple_cond_single_var_p): Remove.
25602 2013-03-18 Richard Biener <rguenther@suse.de>
25604 * tree-data-ref.h (find_data_references_in_loop): Declare.
25605 * tree-data-ref.c (get_references_in_stmt): Use a stack
25606 vector pre-allocated in the callers.
25607 (find_data_references_in_stmt): Adjust.
25608 (graphite_find_data_references_in_stmt): Likewise.
25609 (create_rdg_vertices): Likewise.
25610 (find_data_references_in_loop): Export.
25611 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
25612 Compute dependences here...
25613 (vect_analyze_data_refs): ...not here. When we encounter
25614 a non-vectorizable data reference in basic-block vectorization
25615 truncate the data reference vector. Do not bother to
25616 fixup data-dependence information for gather loads.
25617 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
25618 of data references, as reported.
25620 2013-03-18 Richard Biener <rguenther@suse.de>
25622 PR tree-optimization/3713
25623 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
25624 has_constants and expr.
25625 (stmt_has_constants): Properly valueize SSA names when deciding
25626 whether the stmt has constants.
25628 2013-03-18 Richard Biener <rguenther@suse.de>
25630 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
25631 whole function when there is nothing to do.
25632 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
25633 * tree-vectorizer.c (vectorize_loops): Update virtual and
25634 loop-closed SSA once.
25635 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
25637 2013-03-18 Richard Biener <rguenther@suse.de>
25639 PR middle-end/56113
25640 * domwalk.c (bb_postorder): New global static.
25641 (cmp_bb_postorder): New function.
25642 (walk_dominator_tree): Replace scheme imposing an order for
25643 visiting dominator sons by one sorting them at the time they
25644 are pushed on the stack.
25646 2013-03-18 Richard Biener <rguenther@suse.de>
25648 PR tree-optimization/39326
25649 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
25650 (struct mem_ref): Replace mem member with ao_ref typed member.
25651 (MEM_ANALYZABLE): Adjust.
25652 (memref_eq): Likewise.
25653 (mem_ref_alloc): Likewise.
25654 (gather_mem_refs_stmt): Likewise.
25655 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
25656 (execute_sm_if_changed_flag_set): Adjust.
25657 (execute_sm): Likewise.
25658 (ref_always_accessed_p): Likewise.
25659 (refs_independent_p): Likewise.
25660 (can_sm_ref_p): Likewise.
25662 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25665 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
25666 return 1 even for !unsignedp.
25668 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25670 * config/i386/i386.md (isa): Add x64 and nox64.
25671 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
25672 (*pushtf): Enable *roF alternative for x64 isa only.
25673 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
25674 mode attribute of integer alternatives to DImode for TARGET_64BIT.
25675 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
25676 (*movtf_internal): Merge from *movtf_internal_rex64 and
25677 *movtf_internal_sse. Use x64 and nox64 isa attributes.
25678 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
25679 nox64 isa attributes.
25680 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
25681 nox64 isa attributes.
25682 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
25684 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25686 * config/alpha/alpha.c (TARGET_LRA_P): New define.
25688 2013-03-17 Jakub Jelinek <jakub@redhat.com>
25691 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
25692 class names. Remove trailing comma after "ALL_REGS".
25694 2013-03-16 Jan Hubicka <jh@suse.cz>
25696 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
25697 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
25698 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
25699 of cgraph_get_create_node.
25700 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
25702 2013-03-16 Jason Merrill <jason@redhat.com>
25705 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
25706 with DW_AT_default_value.
25708 2013-03-16 Jakub Jelinek <jakub@redhat.com>
25710 * BASE-VER: Set to 4.9.0.
25712 2013-03-14 Andi Kleen <ak@linux.intel.com>
25715 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
25716 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
25717 Document _x* TSX intrinsics.
25719 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
25720 David Holsgrove <david.holsgrove@xilinx.com>
25722 * configure.ac: Add MicroBlaze TLS support detection.
25723 * configure: Regenerate.
25724 * config/microblaze/microblaze-protos.h
25725 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
25726 symbol_mentioned_p, label_mentioned_p): Add prototypes.
25727 * config/microblaze/microblaze.c (microblaze_address_type): Add
25728 ADDRESS_TLS and tls_reloc address types.
25729 (microblaze_address_info): Add tls_reloc.
25730 (TARGET_HAVE_TLS): Define.
25731 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
25732 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
25733 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
25734 load_tls_operand, microblaze_call_tls_get_addr,
25735 microblaze_legitimize_tls_address): New functions.
25736 (microblaze_classify_unspec): Handle UNSPEC_TLS.
25737 (get_base_reg): Use microblaze_tls_symbol_p.
25738 (microblaze_classify_address): Handle TLS.
25739 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
25740 label_mentioned_p and microblaze_tls_referenced_p.
25741 (microblaze_legitimize_address): Handle TLS.
25742 (microblaze_address_insns): Handle ADDRESS_TLS.
25743 (pic_address_needs_scratch): Handle TLS.
25744 (print_operand_address): Handle TLS.
25745 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
25746 (microblaze_expand_move): Handle TLS.
25747 (microblaze_legitimate_constant_p): Check
25748 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
25749 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
25750 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
25751 (PIC_OFFSET_TABLE_REGNUM): Set.
25752 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
25753 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
25754 (addsi3, movsi_internal2, movdf_internal): Update constraints
25755 * config/microblaze/predicates.md (arith_plus_operand): Define
25756 (move_operand): Redefine as move_src_operand,
25757 check microblaze_tls_referenced_p.
25759 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25761 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
25762 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
25764 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25766 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
25769 2013-03-14 Jakub Jelinek <jakub@redhat.com>
25771 PR tree-optimization/53265
25772 * common.opt (Waggressive-loop-optimizations): New option.
25773 * tree-ssa-loop-niter.c: Include tree-pass.h.
25774 (do_warn_aggressive_loop_optimizations): New function.
25775 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
25776 if number_of_latch_executions returned constant.
25777 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
25778 early. If number_of_latch_executions returned constant, set
25779 nb_iterations_upper_bound back to it.
25780 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
25782 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
25783 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
25785 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
25786 (MULTILIB_OSDIRNAMES): Set.
25787 * genmultilib: If defaultosdirname doesn't start with :: , set
25788 defaultosdirname2 instead, clear it and emit two . multilib_raw
25789 entries instead of just one.
25791 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
25793 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
25794 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
25795 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
25796 (SUBTARGET_OVERRIDE_OPTIONS): New.
25798 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
25801 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
25802 (musermode): Convert to Var(TARGET_USERMODE).
25803 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
25804 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
25805 * config/sh/sh.c (sh_option_override): Use
25806 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
25807 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
25809 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
25811 (udivsi3_i4_single, divsi3_i4_single): Use
25812 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
25814 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
25816 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
25819 2013-03-13 Richard Biener <rguenther@suse.de>
25821 PR tree-optimization/56608
25822 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
25823 calls when vectorizing basic-blocks.
25825 2013-03-13 Jakub Jelinek <jakub@redhat.com>
25828 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
25831 2013-03-12 Jakub Jelinek <jakub@redhat.com>
25833 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
25835 2013-03-11 Jan Hubicka <jh@suse.cz>
25838 * lto-streamer-out.c (output_symbol_p): Skip references from
25839 constructors of external variables.
25841 2013-03-11 Jan Hubicka <jh@suse.cz>
25843 PR middle-end/56571
25844 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
25846 * emit-rtl.c (verify_rtx_sharing): Likewise.
25847 (copy_insn_1): Likewise.
25848 * rtl.c (copy_rtx): Likewise.
25850 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
25853 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
25854 output_operand_lossage message.
25856 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
25859 * arm.c (shift_op): Validate RTL pattern on the fly.
25860 (arm_print_operand, case 'S'): Don't use shift_operator to validate
25863 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25866 * config/pa/pa.md (call_value): Check for calls to powf and direct to
25867 new call patterns that clobber %fr12.
25868 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
25869 split and postreload patterns.
25870 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
25871 registers %fr12 and %fr12R as call used.
25873 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
25875 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
25876 (canon_address, record_store, replace_read, check_mem_read_rtx,
25877 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
25878 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
25879 rest_of_handle_dse): Likewise.
25881 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
25883 PR middle-end/56524
25884 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
25886 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
25887 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
25888 (save_optabs_if_changed): Replace with...
25889 (init_tree_optimization_optabs): ...this.
25890 * optabs.c (save_optabs_if_changed): Rename to...
25891 (init_tree_optimization_optabs): ...this. Take the optimization node
25892 as argument. Do nothing if the base optabs are already correct.
25893 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
25894 to recompute optabs.
25895 * function.h (function): Remove optabs field.
25896 * function.c (invoke_set_current_function_hook): Call
25897 init_tree_optimization_optabs. Use the result to initialize
25900 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
25902 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
25903 if GTMA_HAS_NO_INSTRUMENTATION.
25904 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
25905 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
25906 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
25907 * gimple-pretty-print.c (dump_gimple_transaction): Handle
25908 GTMA_HAS_NO_INSTRUMENTATION.
25910 2013-03-08 Jakub Jelinek <jakub@redhat.com>
25912 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
25915 2013-03-08 Marek Polacek <polacek@redhat.com>
25916 Jakub Jelinek <jakub@redhat.com>
25918 PR tree-optimization/56478
25919 * predict.c (is_comparison_with_loop_invariant_p): Change the
25920 type of loop_step to tree.
25921 (predict_loops): Adjust.
25922 (predict_iv_comparison): Perform the computations on double_ints.
25924 2013-03-08 Richard Biener <rguenther@suse.de>
25926 PR tree-optimization/56570
25927 * tree-cfg.c (verify_expr_location_1): Verify locations for
25929 * tree-sra.c (create_access_replacement): Strip locations
25930 from DECL_DEBUG_EXPRs.
25932 2013-03-08 Richard Biener <rguenther@suse.de>
25934 * tree-inline.c (expand_call_inline): Do not associate
25935 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
25936 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
25938 2013-03-08 Richard Biener <rguenther@suse.de>
25940 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
25941 or block changes with -Og. Fix for location / block encoding
25942 changes and PHI arguments with locations.
25944 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
25946 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
25948 (struct output_info): Likewise.
25949 (register_overhead): Remove bad gcc_assert.
25950 (bitmap_find_bit): If there is only a single bitmap element, do not
25951 count a miss as a search.
25952 (print_statistics): Update for counter type changes.
25953 (dump_bitmap_statistics): Likewise. Print headers such that they
25954 are properly lined up with the printed counters.
25956 2013-03-07 Jakub Jelinek <jakub@redhat.com>
25958 PR tree-optimization/56559
25959 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
25960 check that it has only a single use.
25962 2013-03-07 Richard Biener <rguenther@suse.de>
25964 * doc/invoke.texi (fwhole-program): Discourage use in combination
25967 2013-03-06 Jakub Jelinek <jakub@redhat.com>
25969 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
25971 PR tree-optimization/56539
25972 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
25973 instead of GSI_CONTINUE_LINKING as last argument to
25974 force_gimple_operand_gsi. Adjust function comment.
25976 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
25979 PR middle-end/56548
25980 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
25981 promoted mode, convert the result back to the original mode.
25983 2013-03-06 Richard Biener <rguenther@suse.de>
25985 PR middle-end/56294
25986 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
25987 (insert_updated_phi_nodes_compare_uids): New function.
25988 (update_ssa): Sort symbols_to_rename after UID before
25989 traversing it to insert PHI nodes.
25991 2013-03-06 Richard Biener <rguenther@suse.de>
25993 PR middle-end/50494
25994 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
25995 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
25998 2013-02-13 Richard Biener <rguenther@suse.de>
26001 * varasm.c (output_constant_def_1): Get the decl representing
26002 the constant as argument.
26003 (output_constant_def): Wrap output_constant_def_1.
26004 (make_decl_rtl): Use output_constant_def_1 with the decl
26005 representing the constant.
26006 (build_constant_desc): Optionally re-use a decl already
26007 representing the constant.
26008 (tree_output_constant_def): Adjust.
26010 2013-03-06 Joey Ye <joey.ye@arm.com>
26013 * gcc.c (convert_white_space): New function.
26014 (main): Handles white space in function name.
26016 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
26019 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
26020 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
26021 to SH_DIV_CALL_TABLE for TARGET_SH2.
26022 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
26024 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
26025 call-table options.
26027 2013-03-05 Sterling Augustine <saugustine@google.com>
26028 Cary Coutant <ccoutant@google.com>
26031 * dwarf2out.c (resolve_addr): Don't call
26032 remove_loc_list_addr_table_entries a second time for the same
26035 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26038 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
26039 (avoid_complex_debug_insns): New function.
26040 (expand_debug_locations): Call it.
26042 PR rtl-optimization/56484
26043 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
26044 lifetimes of hard registers on small register class machines.
26046 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26048 * config/microblaze/microblaze-protos.h: Rename
26049 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
26050 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
26052 (microblaze_fast_interrupt_function_p): New function.
26053 (microblaze_is_interrupt_handler): Rename to
26054 microblaze_is_interrupt_variant and add fast_interrupt check.
26055 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
26056 (save_restore_insns): Likewise.
26057 (compute_frame_size): Likewise.
26058 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
26059 (microblaze_globalize_label): Likewise.
26060 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
26061 * config/microblaze/microblaze.md: Use wrapper
26062 microblaze_is_interrupt_variant.
26064 2013-03-05 Kai Tietz <ktietz@redhat.com>
26066 * sdbout.c (sdbout_one_type): Switch to current function's section
26067 supporting cold/hot.
26069 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26071 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
26074 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26076 PR middle-end/56461
26077 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
26078 if VALGRIND_GET_VBITS is defined, temporarily make object
26079 memory all defined, and restore previous valgrind addressability
26080 and definability afterwards. Free this_object at the end.
26082 PR middle-end/56461
26083 * lra.c (lra): Call lra_clear_live_ranges if live_p,
26084 right before calling lra_create_live_ranges, also call it
26085 when clearing live_p. Only call lra_clear_live_ranges
26086 at the end if live_p.
26088 PR middle-end/56461
26089 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
26091 2013-03-05 Richard Biener <rguenther@suse.de>
26093 PR tree-optimization/56521
26094 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
26097 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
26100 * except.h (remove_unreachable_eh_regions): New prototype.
26101 * except.c (remove_eh_handler_splicer): New function, split out
26102 of remove_eh_handler.
26103 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
26104 warning about running it on many EH regions one at a time.
26105 (remove_unreachable_eh_regions_worker): New function, walk the
26106 EH tree in depth-first order and remove non-marked regions.
26107 (remove_unreachable_eh_regions): New function.
26108 * tree-eh.c (mark_reachable_handlers): New function, split out
26109 from remove_unreachable_handlers.
26110 (remove_unreachable_handlers): Use mark_reachable_handlers and
26111 remove_unreachable_eh_regions.
26112 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
26113 and remove_unreachable_eh_regions.
26115 2013-03-05 Richard Biener <rguenther@suse.de>
26117 PR middle-end/56525
26118 * loop-init.c (fix_loop_structure): Remove loops in two stages,
26119 not freeing them until the end.
26121 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26123 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
26125 2013-03-05 Richard Biener <rguenther@suse.de>
26127 PR tree-optimization/56270
26128 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
26129 of loads after scheduling an SLP instance.
26131 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26133 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
26135 (check_gcc_parallelize): Run guality.exp as a separate job from
26136 vect.exp with unsorted.exp and $(dg_target_exps) separately from
26137 struct-layout-1.exp with stackalign.exp.
26139 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
26141 PR middle-end/56461
26142 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
26143 load_index sbitmap even if some bit in it isn't set.
26145 PR middle-end/56461
26146 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
26147 (discover_iteration_bound_by_body_walk): Change queues to
26148 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
26149 spelling in comment. Call safe_push on queues[bound_index] directly.
26150 Release queues[queue_index] in every iteration unconditionally.
26151 Release bounds vector.
26153 PR middle-end/56461
26154 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
26155 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
26156 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
26159 2013-03-05 Richard Biener <rguenther@suse.de>
26162 * tree-inline.c (remap_blocks_to_null): New function.
26163 (expand_call_inline): When expanding a call stmt without
26164 an associated block inline remap all callee blocks to NULL.
26166 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26168 PR rtl-optimization/56494
26169 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
26170 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
26171 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
26173 PR middle-end/56461
26174 * sel-sched-ir.c (free_sched_pools): Release
26175 succs_info_pool.stack[succs_info_pool.max_top] vectors too
26176 if succs_info_pool.max_top isn't -1.
26179 * opts.c (opts_obstack, opts_concat): Moved to...
26180 * opts-common.c (opts_obstack, opts_concat): ... here.
26182 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26184 PR middle-end/56461
26185 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
26187 2013-03-04 Martin Jambor <mjambor@suse.cz>
26189 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
26190 all appropriate places.
26192 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
26194 PR tree-optimization/56424
26195 * ipa-split.c (split_function): Do not set the RSO flag if result is
26196 not by reference and its type is a register type.
26198 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
26200 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
26201 (microblaze_legitimate_pic_operand): Likewise
26202 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
26203 new function microblaze_legitimate_pic_operand
26204 * config/microblaze/microblaze-protos.h
26205 (microblaze_legitimate_pic_operand): Declare.
26207 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26209 * config/microblaze/predicates.md (call_insn_simple_operand):
26210 New predicate for supported rtx code types.
26211 * config/microblaze/microblaze.md (call_internal1): Use
26212 call_insn_simple_operand predicate.
26214 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26216 PR middle-end/56461
26217 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
26218 partitions.ordered_remove.
26220 PR middle-end/56461
26221 * tree-vect-stmts.c (vectorizable_conversion): Don't call
26222 vec_oprnds0.create (1) for modifier == NONE.
26224 PR middle-end/56461
26225 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
26226 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
26227 vec_oprnds1 right before pushing anything to it for
26230 PR middle-end/56461
26231 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
26232 set nbbs to 0 instead of having separate code path.
26233 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
26234 instead of false as last argument if returning NULL.
26236 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
26238 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
26239 the attribute is now called "target" instead of "option".
26240 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
26241 * doc/tm.texi.in (Target Attributes): Likewise document the correct
26242 attribute/pragma name for TARGET_OPTION_VALID_P and
26243 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
26244 * doc/tm.texi: Regenerated.
26246 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
26248 * config/microblaze/microblaze.c:
26249 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
26250 * config/microblaze/microblaze.h: Add -mxl-reorder to
26252 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
26253 instructions emitted if TARGET_REORDER.
26254 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
26255 or 0 for -m/-mno case, but initialises as 2 to detect default use case
26258 2013-03-01 Xinliang David Li <davidxl@google.com>
26260 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
26263 2013-03-01 Jakub Jelinek <jakub@redhat.com>
26265 PR middle-end/56461
26266 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
26267 vector even when returning true. Fix up function comment formatting.
26269 PR middle-end/56461
26270 * ira-build.c (ira_loop_nodes_count): New variable.
26271 (create_loop_tree_nodes): Initialize it.
26272 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
26274 PR middle-end/56461
26275 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
26276 method on dr_chain and result_chain.
26277 * tree-vect-stmts.c (vectorizable_store): Only call
26278 result_chain.create if j == 0.
26280 PR middle-end/56461
26281 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
26282 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
26283 before overwriting it.
26285 2013-03-01 Tobias Burnus <burnus@net-b.de>
26287 * doc/extended.texi (C Extensions): Change order in @menu
26289 (Other MIPS Built-in Functions): Move last MIPS entry before
26290 "picoChip Built-in Functions".
26291 (SH Built-in Functions): Move after RX Built-in Functions.
26292 * doc/gcc.texi (Introduction): Change order in @menu
26294 * doc/md.texi (Constraints): Ditto.
26295 * gty.texi (Type Information): Ditto.
26296 (User-provided marking routines for template types): Make
26298 * doc/invoke.texi (AArch64 Options): Move before
26299 "Adapteva Epiphany Options".
26301 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
26302 Jakub Jelinek <jakub@redhat.com>
26305 * asan.c (gate_asan): Lookup no_sanitize_address instead of
26306 no_address_safety_analysis attribute.
26307 * doc/extend.texi (no_address_safety_attribute): Rename to
26308 no_sanitize_address attribute, mention no_address_safety_analysis
26309 attribute as deprecated alias.
26311 2013-02-28 Jakub Jelinek <jakub@redhat.com>
26313 PR middle-end/56461
26314 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
26315 type to vec<vec<tree> > *.
26316 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
26317 to be vec<tree> instead of vec<tree> *, set vec_defs
26318 to vNULL and call vec_defs.create (number_of_vects), adjust other
26320 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
26321 vectorizable_condition): Adjust vect_get_slp_defs callers.
26323 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26325 * config/aarch64/aarch64.c
26326 (aarch64_float_const_representable): Remove unused variable.
26328 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26330 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
26332 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26334 * config/aarch64/aarch64-builtins.c
26335 (aarch64_init_simd_builtins): Make static.
26337 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26339 * config/aarch64/aarch64.c
26340 (aarch64_simd_make_constant): Make static.
26342 2013-02-28 Martin Jambor <mjambor@suse.cz>
26344 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
26345 with no initialization to the RHS of debug statements.
26347 2013-02-28 Martin Jambor <mjambor@suse.cz>
26349 PR tree-optimization/56294
26350 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
26352 (get_access_replacement): Do not call create_access_replacement.
26353 Assert a replacement exists.
26354 (get_repl_default_def_ssa_name): Create the replacement declaration
26357 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26359 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
26360 final_end_function.
26362 2013-02-28 Marek Polacek <polacek@redhat.com>
26364 PR rtl-optimization/56466
26365 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
26366 if we're changing a loop.
26367 (peel_loops_completely): Likewise.
26369 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
26372 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
26374 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26377 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
26378 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
26379 INTX_FTYPE_FX, FX_FTYPE_INTX.
26380 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
26382 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26384 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
26385 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
26386 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
26387 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
26388 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
26389 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
26390 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
26391 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
26392 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
26393 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
26394 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
26395 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
26396 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
26397 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
26398 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
26399 (avrxmega6): Increase max flash segments from 5 to 6.
26400 * config/avr/t-multilib: Regenerate.
26401 * config/avr/avr-tables.opt: Regenerate.
26402 * doc/avr-mmcu.texi: Regenerate.
26404 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26406 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
26407 (avr_device_to_arch): Rename to avr_device_to_ld.
26408 (avr_device_to_as): New prototype.
26409 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
26410 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
26411 * config/avr/driver-avr.c (avr_device_to_as): New.
26412 (avr_device_to_arch): Rename to avr_device_to_ld.
26414 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26416 PR middle-end/56461
26417 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
26418 method on dr_chain and result_chain.
26420 PR middle-end/56461
26421 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
26422 pointer_set_destroy on not_executed_last_iteration.
26424 PR middle-end/56461
26425 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
26427 PR middle-end/56461
26428 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
26429 FOR_EACH_DEFINED_FUNCTION when freeing state.
26431 PR middle-end/56461
26432 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
26434 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
26437 PR middle-end/56461
26438 * ipa-cp.c (decide_whether_version_node): Call vec_free on
26439 known_aggs[i].items and release known_aggs vector.
26441 PR middle-end/56461
26442 * ipa-reference.c (propagate): Free node_info even for alias nodes.
26444 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26446 * config/microblaze/microblaze.c (microblaze_emit_compare):
26447 Use xor for EQ/NE comparisions.
26448 * config/microblaze/microblaze.md (cstoresf4): Add constraints
26449 (cbranchsf4): Adjust operator to comparison_operator.
26451 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26453 PR middle-end/56461
26454 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
26456 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
26457 vec_safe_push, always update *slot.
26458 (redirect_edge_var_map_clear): Use vec_free.
26459 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
26460 (free_var_map_entry): Use vec_free.
26461 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
26462 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
26464 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
26466 PR middle-end/45472
26467 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
26468 when the may_trap_p bit of the exprs being merged differs.
26469 Reorder tests for speculativeness in the logical and operator.
26471 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26473 * incpath.c (add_standard_paths): Use reconcat instead of concat
26474 where appropriate and avoid leaking memory.
26476 * opts.h: Include obstack.h.
26477 (opts_concat): New prototype.
26478 (opts_obstack): New declaration.
26479 * opts.c (opts_concat): New function.
26480 (opts_obstack): New variable.
26481 (init_options_struct): Call gcc_init_obstack on opts_obstack.
26482 (finish_options): Use opts_concat instead of concat
26483 and XOBNEWVEC instead of XNEWVEC.
26484 * opts-common.c (generate_canonical_option, decode_cmdline_option,
26485 generate_option): Likewise.
26486 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
26487 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
26490 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
26491 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
26493 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26495 PR middle-end/56461
26496 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
26498 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26500 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
26501 (arm_block_move_unaligned_straight): Likewise.
26502 (arm_adjust_block_mem): Likewise.
26504 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26507 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
26508 temp, cond and label.
26509 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
26512 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
26513 * config/c6x/c6x.h (dbx_register_map): Update declaration.
26516 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
26517 of prologue/epilogue functions.
26520 * config/tilegx/tilegx.c (tilegx_expand_prologue):
26521 Remove unused variable cfa_offset.
26522 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
26525 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
26526 type promotion to unsigned.
26529 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
26530 for HOST_WIDE_INT of 32 bit / same size as int.
26531 (arm_block_move_unaligned_straight): Likewise.
26532 (arm_adjust_block_mem): Likewise.
26535 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
26538 2013-02-26 Marek Polacek <polacek@redhat.com>
26540 PR tree-optimization/56426
26541 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
26543 2013-02-26 Richard Biener <rguenther@suse.de>
26546 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
26547 unused variable loops.
26549 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26551 PR tree-optimization/56448
26552 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
26553 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
26554 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
26555 later operands of the references, or even first operand for
26556 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
26558 PR tree-optimization/56443
26559 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
26560 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
26561 to type_for_mode langhook.
26563 2013-02-25 Matt Turner <mattst88@gmail.com>
26565 * doc/invoke.texi: Document r4700.
26567 2013-02-25 Richard Biener <rguenther@suse.de>
26569 PR tree-optimization/56175
26570 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
26572 (simplify_bitwise_binary): ... here. Also guard the conversion
26573 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
26575 2013-02-25 Catherine Moore <clm@codesourcery.com>
26578 2013-02-24 Catherine Moore <clm@codesourcery.com>
26579 Maciej W. Rozycki <macro@codesourcery.com>
26580 Tom de Vries <tom@codesourcery.com>
26581 Nathan Sidwell <nathan@codesourcery.com>
26582 Iain Sandoe <iain@codesourcery.com>
26583 Nathan Froyd <froydnj@codesourcery.com>
26584 Chao-ying Fu <fu@mips.com>
26586 * doc/extend.texi: (micromips, nomicromips, nocompression):
26587 Document new function attributes.
26588 * doc/invoke.texi (minterlink-compressed, mmicromips,
26589 m14k, m14ke, m14kec): Document new options.
26590 (minterlink-mips16): Update documentation.
26591 * doc/md.texi (ZC, ZD): Document new constraints.
26592 * configure.ac (gcc_cv_as_micromips): Check if linker
26593 supports the .set micromips directive.
26594 * configure: Regenerate.
26595 * config.in: Regenerate.
26596 * config/mips/mips-tables.opt: Regenerate.
26597 * config/mips/micromips.md: New file.
26598 * constraints.md (ZC, AD): New constraints.
26599 * config/mips/predicates.md (movep_src_register): New predicate.
26600 (movep_src_operand): New predicate.
26601 (non_volatile_mem_operand): New predicate.
26602 * config/mips/mips.md (multimem): New type.
26603 (length): Differentiate between 17-bit and 18-bit branch offsets.
26604 (MOVEP1, MOVEP2): New mode iterator.
26605 (mov_<load>l): Use ZC constraint.
26606 (mov_<load>r): Likewise.
26607 (mov_<store>l): Likewise.
26608 (mov_<store>r): Likewise.
26609 (*branch_equality<mode>_inverted): Add microMIPS support.
26610 (*branch_equality<mode>): Likewise.
26611 (*jump_absolute): Likewise.
26612 (indirect_jump_<mode>): Likewise.
26613 (tablejump_<mode>): Likewise.
26614 (<optab>_internal): Likewise.
26615 (sibcall_internal): Likewise.
26616 (sibcall_value_internal): Likewise.
26617 (prefetch): Use constraint ZD.
26618 * config/mips/mips.opt (minterlink-compressed): New option.
26619 (minterlink-mips16): Now an alias for minterlink-compressed.
26620 (mmicromips): New option.
26621 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26622 (compare_and_swap_12): Likewise.
26623 (sync_add<mode>): Likewise.
26624 (sync_<optab>_12): Likewise.
26625 (sync_old_<optab>_12): Likewise.
26626 (sync_new_<optab>_12): Likewise.
26627 (sync_nand_12): Likewise.
26628 (sync_old_nand_12): Likewise.
26629 (sync_new_nand_12): Likewise.
26630 (sync_sub<mode>): Likewise.
26631 (sync_old_add<mode>): Likewise.
26632 (sync_old_sub<mode>): Likewise.
26633 (sync_new_add<mode>): Likewise.
26634 (sync_new_sub<mode>): Likewise.
26635 (sync_<optab><mode>): Likewise.
26636 (sync_old_<optab><mode>): Likewise.
26637 (sync_new_<optab><mode>): Likewise.
26638 (sync_nand<mode>): Likewise.
26639 (sync_old_nand<mode>): Likewise.
26640 (sync_new_nand<mode>): Likewise.
26641 (sync_lock_test_and_set<mode>): Likewise.
26642 (test_and_set_12): Likewise.
26643 (atomic_compare_and_swap<mode>): Likewise.
26644 (atomic_exchange<mode>_llsc): Likewise.
26645 (atomic_fetch_add<mode>_llsc): Likewise.
26646 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26647 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26648 (umips_save_restore_pattern_p): Likewise.
26649 (umips_load_store_pair_p): Likewise.
26650 (umips_output_load_store_pair): Likewise.
26651 (umips_movep_target_p): Likewise.
26652 (umips_12bit_offset_address_p): Likewise.
26653 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26654 (mips_base_mips16): Rename this...
26655 (mips_base_compression_flags): ...to this. Update all uses.
26656 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26657 (mips_mips16_decl_p): Delete.
26658 (mips_nomips16_decl_p): Delete.
26659 (mips_get_compress_on_flags): New function.
26660 (mips_get_compress_off_flags): New function.
26661 (mips_get_compress_mode): New function.
26662 (mips_get_compress_on_name): New function.
26663 (mips_get_compress_off_name): New function.
26664 (mips_insert_attributes): Support multiple compression types.
26665 (mips_merge_decl_attributes): Likewise.
26666 (umips_12bit_offset_address_p): New function.
26667 (mips_start_function_definition): Emit .set micromips directive.
26668 (mips_call_may_need_jalx_p): New function.
26669 (mips_function_ok_for_sibcall): Add microMIPS support.
26670 (mips_print_operand_punctuation): Support short delay slots and
26672 (umips_swm_mask, umips_swm_encoding): New.
26673 (umips_build_save_restore): New function.
26674 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26675 (was_mips16_p): Remove.
26676 (old_compression_mode): New.
26677 (mips_set_compression_mode): New function.
26678 (mips_set_current_function): Add microMIPS support.
26679 (mips_option_override): Likewise.
26680 (umips_save_restore_pattern_p): New function.
26681 (umips_output_save_restore): New function.
26682 (umips_load_store_pair_p_1): New function.
26683 (umips_load_store_pair_p): New function.
26684 (umips_output_load_store_pair_1): New function.
26685 (umips_output_load_store_pair): New function.
26686 (umips_movep_target_p) New function.
26687 (mips_prepare_pch_save): Add microMIPS support.
26688 * config/mips/mips.h (TARGET_COMPRESSION): New.
26689 (TARGET_CPU_CPP_BUILTINS): Update macro
26690 to use new compression flags and to support microMIPS.
26691 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26692 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26693 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26694 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26695 (ASM_SPEC): Support mmicromips and mno-micromips.
26696 (M16STORE_REG_P): New macro.
26697 (MIPS_CALL): Support TARGET_MICROMIPS.
26698 (MICROMIPS_J): New macro.
26699 (mips_base_mips16): Rename this...
26700 (mips_base_compression_flags): ...to this.
26701 (UMIPS_12BIT_OFFSET_P): New macro.
26702 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26703 (MULTILIB_DIRNAMES): Likewise.
26705 2013-02-25 Tom de Vries <tom@codesourcery.com>
26707 PR rtl-optimization/56131
26708 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
26709 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
26710 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
26712 2013-02-25 Tobias Burnus <burnus@net-b.de>
26714 * doc/invoke.texi (-fsanitize=): Move from optimization
26715 to debugging options.
26717 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26719 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
26721 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26722 Alexander Monakov <amonakov@ispras.ru>
26724 PR middle-end/56077
26725 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
26726 flush pending lists also on non-jumps. Adjust comment.
26728 2013-02-24 Catherine Moore <clm@codesourcery.com>
26729 Maciej W. Rozycki <macro@codesourcery.com>
26730 Tom de Vries <tom@codesourcery.com>
26731 Nathan Sidwell <nathan@codesourcery.com>
26732 Iain Sandoe <iain@codesourcery.com>
26733 Nathan Froyd <froydnj@codesourcery.com>
26734 Chao-ying Fu <fu@mips.com>
26736 * doc/extend.texi: (micromips, nomicromips, nocompression):
26737 Document new function attributes.
26738 * doc/invoke.texi (minterlink-compressed, mmicromips,
26739 m14k, m14ke, m14kec): Document new options.
26740 (minterlink-mips16): Update documentation.
26741 * doc/md.texi (ZC, ZD): Document new constraints.
26742 * configure.ac (gcc_cv_as_micromips): Check if linker
26743 supports the .set micromips directive.
26744 * configure: Regenerate.
26745 * config.in: Regenerate.
26746 * config/mips/mips-tables.opt: Regenerate.
26747 * config/mips/micromips.md: New file.
26748 * constraints.md (ZC, AD): New constraints.
26749 * config/mips/predicates.md (movep_src_register): New predicate.
26750 (movep_src_operand): New predicate.
26751 (non_volatile_mem_operand): New predicate.
26752 * config/mips/mips.md (multimem): New type.
26753 (length): Differentiate between 17-bit and 18-bit branch offsets.
26754 (MOVEP1, MOVEP2): New mode iterator.
26755 (mov_<load>l): Use ZC constraint.
26756 (mov_<load>r): Likewise.
26757 (mov_<store>l): Likewise.
26758 (mov_<store>r): Likewise.
26759 (*branch_equality<mode>_inverted): Add microMIPS support.
26760 (*branch_equality<mode>): Likewise.
26761 (*jump_absolute): Likewise.
26762 (indirect_jump_<mode>): Likewise.
26763 (tablejump_<mode>): Likewise.
26764 (<optab>_internal): Likewise.
26765 (sibcall_internal): Likewise.
26766 (sibcall_value_internal): Likewise.
26767 (prefetch): Use constraint ZD.
26768 * config/mips/mips.opt (minterlink-compressed): New option.
26769 (minterlink-mips16): Now an alias for minterlink-compressed.
26770 (mmicromips): New option.
26771 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26772 (compare_and_swap_12): Likewise.
26773 (sync_add<mode>): Likewise.
26774 (sync_<optab>_12): Likewise.
26775 (sync_old_<optab>_12): Likewise.
26776 (sync_new_<optab>_12): Likewise.
26777 (sync_nand_12): Likewise.
26778 (sync_old_nand_12): Likewise.
26779 (sync_new_nand_12): Likewise.
26780 (sync_sub<mode>): Likewise.
26781 (sync_old_add<mode>): Likewise.
26782 (sync_old_sub<mode>): Likewise.
26783 (sync_new_add<mode>): Likewise.
26784 (sync_new_sub<mode>): Likewise.
26785 (sync_<optab><mode>): Likewise.
26786 (sync_old_<optab><mode>): Likewise.
26787 (sync_new_<optab><mode>): Likewise.
26788 (sync_nand<mode>): Likewise.
26789 (sync_old_nand<mode>): Likewise.
26790 (sync_new_nand<mode>): Likewise.
26791 (sync_lock_test_and_set<mode>): Likewise.
26792 (test_and_set_12): Likewise.
26793 (atomic_compare_and_swap<mode>): Likewise.
26794 (atomic_exchange<mode>_llsc): Likewise.
26795 (atomic_fetch_add<mode>_llsc): Likewise.
26796 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26797 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26798 (umips_save_restore_pattern_p): Likewise.
26799 (umips_load_store_pair_p): Likewise.
26800 (umips_output_load_store_pair): Likewise.
26801 (umips_movep_target_p): Likewise.
26802 (umips_12bit_offset_address_p): Likewise.
26803 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26804 (mips_base_mips16): Rename this...
26805 (mips_base_compression_flags): ...to this. Update all uses.
26806 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26807 (mips_mips16_decl_p): Delete.
26808 (mips_nomips16_decl_p): Delete.
26809 (mips_get_compress_on_flags): New function.
26810 (mips_get_compress_off_flags): New function.
26811 (mips_get_compress_mode): New function.
26812 (mips_get_compress_on_name): New function.
26813 (mips_get_compress_off_name): New function.
26814 (mips_insert_attributes): Support multiple compression types.
26815 (mips_merge_decl_attributes): Likewise.
26816 (umips_12bit_offset_address_p): New function.
26817 (mips_start_function_definition): Emit .set micromips directive.
26818 (mips_call_may_need_jalx_p): New function.
26819 (mips_function_ok_for_sibcall): Add microMIPS support.
26820 (mips_print_operand_punctuation): Support short delay slots and
26822 (umips_swm_mask, umips_swm_encoding): New.
26823 (umips_build_save_restore): New function.
26824 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26825 (was_mips16_p): Remove.
26826 (old_compression_mode): New.
26827 (mips_set_compression_mode): New function.
26828 (mips_set_current_function): Add microMIPS support.
26829 (mips_option_override): Likewise.
26830 (umips_save_restore_pattern_p): New function.
26831 (umips_output_save_restore): New function.
26832 (umips_load_store_pair_p_1): New function.
26833 (umips_load_store_pair_p): New function.
26834 (umips_output_load_store_pair_1): New function.
26835 (umips_output_load_store_pair): New function.
26836 (umips_movep_target_p) New function.
26837 (mips_prepare_pch_save): Add microMIPS support.
26838 * config/mips/mips.h (TARGET_COMPRESSION): New.
26839 (TARGET_CPU_CPP_BUILTINS): Update macro
26840 to use new compression flags and to support microMIPS.
26841 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26842 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26843 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26844 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26845 (ASM_SPEC): Support mmicromips and mno-micromips.
26846 (M16STORE_REG_P): New macro.
26847 (MIPS_CALL): Support TARGET_MICROMIPS.
26848 (MICROMIPS_J): New macro.
26849 (mips_base_mips16): Rename this...
26850 (mips_base_compression_flags): ...to this.
26851 (UMIPS_12BIT_OFFSET_P): New macro.
26852 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26853 (MULTILIB_DIRNAMES): Likewise.
26855 2013-02-24 Jakub Jelinek <jakub@redhat.com>
26858 * target-globals.c (save_target_globals): For init_reg_sets and
26859 target_reinit remporarily set this_fn_optabs to this_target_optabs.
26861 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
26863 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
26864 * config/aarch64/t-aarch64
26865 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
26867 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
26869 PR inline-asm/56148
26870 * lra-constraints.c (process_alt_operands): Reload operand
26871 conflicting with earlier clobber only if no more other conflicting
26874 2013-02-22 Jakub Jelinek <jakub@redhat.com>
26877 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
26878 if not linking a shared library.
26880 2013-02-22 Seth LaForge <sethml@google.com>
26882 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
26884 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
26886 * config/arm/arm.md (split for extendsidi): Update condition.
26887 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
26888 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
26889 (qhs_zextenddi_cstr): Likewise.
26891 2013-02-21 Jakub Jelinek <jakub@redhat.com>
26893 PR middle-end/56420
26894 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
26895 avoid signed wrapping.
26896 (expand_mult): Handle properly multiplication by
26897 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
26898 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
26899 in the compiler if coeff is HOST_WIDE_INT_MIN.
26900 (expand_divmod): Don't make ext_op1 static, change it's type to
26901 uhwi. Avoid undefined behavior in -INTVAL (op1).
26903 PR rtl-optimization/50339
26904 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
26906 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
26907 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
26908 into splitting_ashiftrt field.
26909 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
26911 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
26914 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
26916 PR middle-end/56108
26917 * trans-mem.c (execute_tm_mark): Do not expand transactions that
26918 are sure to go irrevocable.
26920 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
26922 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
26923 scalars are valid operands.
26925 2013-02-21 Martin Jambor <mjambor@suse.cz>
26927 PR tree-optimization/56310
26928 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
26929 only matching indices and non-negative final offsets.
26930 (intersect_aggregates_with_edge): Pass src_idx to
26931 agg_replacements_to_vector. Pass src_idx insstead of index to
26932 intersect_with_agg_replacements.
26934 2013-02-21 Martin Jambor <mjambor@suse.cz>
26936 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
26937 instead of hard-wired defaults.
26939 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
26941 * doc/invoke.texi (MIPS Options): Update documentation of the
26942 floating-point multiply-accumulate instruction restrictions.
26944 2013-02-21 Kostya Serebryany <kcc@google.com>
26946 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
26947 asan_shadow_offset on x86_64 linux.
26949 2013-02-21 Richard Biener <rguenther@suse.de>
26951 PR tree-optimization/56415
26953 2013-02-11 Richard Biener <rguenther@suse.de>
26955 PR tree-optimization/56273
26956 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
26959 2013-02-21 Jakub Jelinek <jakub@redhat.com>
26962 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
26965 PR inline-asm/56405
26966 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
26967 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
26969 2013-02-20 Jan Hubicka <jh@suse.cz>
26971 PR tree-optimization/56265
26972 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
26973 when target is referenced for first time.
26975 2013-02-20 Richard Biener <rguenther@suse.de>
26977 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
26978 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
26979 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
26980 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
26981 not return anything.
26982 (rename_ssa_copies): Do not remove unused locals.
26983 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
26984 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
26985 * passes.c (execute_function_todo): Do not schedule unused locals
26986 removal if cleanup_tree_cfg did something.
26987 * tree-ssa-live.c (remove_unused_locals): Dump statistics
26988 about the number of removed locals.
26990 2013-02-20 Richard Biener <rguenther@suse.de>
26992 PR tree-optimization/56398
26993 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
26995 2013-02-20 Martin Jambor <mjambor@suse.cz>
26997 PR tree-optimization/55334
26998 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
26999 restricted pointers to arrays.
27001 2013-02-20 Richard Biener <rguenther@suse.de>
27002 Jakub Jelinek <jakub@redhat.com>
27004 PR tree-optimization/56396
27005 * tree-ssa-ccp.c (n_const_val): New static variable.
27006 (get_value): Return NULL for SSA names we don't have a lattice
27008 (ccp_initialize): Initialize n_const_val.
27009 * tree-ssa-copy.c (n_copy_of): New static variable.
27010 (init_copy_prop): Initialize n_copy_of.
27011 (get_value): Return NULL_TREE for SSA names we don't have a
27014 2013-02-20 Martin Jambor <mjambor@suse.cz>
27016 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
27018 2013-02-20 Richard Biener <rguenther@suse.de>
27020 * genpreds.c (write_lookup_constraint): Do not compare first
27021 letter of the constraint again.
27023 2013-02-20 Richard Biener <rguenther@suse.de>
27025 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
27027 (get_use_iv_cost): Terminate hashtable walk when coming across
27030 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
27032 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
27033 reassociation for avx2 targets.
27035 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27037 * config/microblaze/microblaze.c: microblaze_has_clz = 0
27038 Add version check for v8.10.a to enable microblaze_has_clz
27039 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
27040 version and TARGET_PATTERN_COMPARE check
27041 * config/microblaze/microblaze.md: New clzsi2 instruction
27043 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27045 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
27046 function before branching.
27048 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27050 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
27052 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
27054 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27056 PR middle-end/55889
27057 * sel-sched.c: Include ira.h.
27058 (implicit_clobber_conflict_p): New function.
27059 (moveup_expr): Use it.
27060 * Makefile.in (sel-sched.o): Depend on ira.h.
27062 2013-02-19 Richard Biener <rguenther@suse.de>
27064 PR tree-optimization/56384
27065 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
27066 (vn_hash_type): Split out from ...
27067 (vn_hash_constant_with_type): ... here.
27068 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
27069 (vn_phi_eq): Compare types from vn_phi_s structure.
27070 (vn_phi_lookup): Populate vn_phi_s type.
27071 (vn_phi_insert): Likewise.
27073 2013-02-19 Jakub Jelinek <jakub@redhat.com>
27075 PR tree-optimization/56350
27076 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
27077 if haven't found reduction or nested cycle operand, rather than
27078 asserting we must find it.
27080 PR tree-optimization/56381
27081 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
27084 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
27085 Jakub Jelinek <jakub@redhat.com>
27088 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
27089 (swap_optab_enable): Same.
27090 (init_all_optabs): Use argument instead of global.
27091 * tree.h (struct tree_optimization_option): New field target_optabs.
27092 * expr.h (init_all_optabs): Add argument to prototype.
27093 (TREE_OPTIMIZATION_OPTABS): New.
27094 (save_optabs_if_changed): Protoize.
27095 * optabs.h: Declare this_fn_optabs.
27096 * optabs.c (save_optabs_if_changed): New.
27097 Declare this_fn_optabs.
27098 (init_optabs): Add argument to init_all_optabs() call.
27099 * function.c (invoke_set_current_function_hook): Handle per
27101 * function.h (struct function): New field optabs.
27102 * config/mips/mips.c (mips_set_mips16_mode): Handle when
27103 optimization_current_node has changed.
27104 * target-globals.h (save_target_globals_default_opts): Protoize.
27105 * target-globals.c (save_target_globals_default_opts): New.
27107 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27110 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
27111 registers %fr12 and %fr12R as call used.
27114 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
27115 and HImode, require all displacements to be an integer multiple of
27117 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
27118 only allow QImode and HImode when reload is in progress and strict is
27119 true. Likewise for symbolic addresses. Use base14_operand to check
27120 displacements in REG+BASE addresses.
27122 2013-02-18 Richard Biener <rguenther@suse.de>
27124 PR tree-optimization/56366
27125 * tree-vect-loop.c (get_initial_def_for_induction): Properly
27126 handle sign-conversion of outer-loop initial induction value.
27128 2013-02-18 Richard Biener <rguenther@suse.de>
27130 PR middle-end/56349
27131 * cfghooks.c (merge_blocks): If we merge a latch into another
27132 block adjust references to it.
27133 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
27134 (verify_loop_structure): Verify that a recorded latch is in fact
27137 2013-02-18 Richard Biener <rguenther@suse.de>
27139 PR tree-optimization/56321
27140 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
27141 order SSA name release and virtual operand unlinking.
27143 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27145 * config/microblaze/microblaze.md (save_stack_block): Define.
27146 (restore_stack_block): Likewise.
27148 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27150 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
27151 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
27152 * config/microblaze/microblaze.c (microblaze_option_override):
27153 Bail out early for PIC modes when target does not support PIC.
27155 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27157 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
27158 Replace with a microblaze version.
27159 (microblaze_trampoline_init): Adapt for microblaze.
27160 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
27163 2013-02-16 Jakub Jelinek <jakub@redhat.com>
27164 Dodji Seketeli <dodji@redhat.com>
27167 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
27168 (instrument_mem_region_access): Do not forget to always put
27169 instrumentation of the of 'base' and 'base + len' in a "if (len !=
27170 0) statement, even for cases where either 'base' or 'base + len'
27171 are not instrumented -- because they have been previously
27172 instrumented. Simplify the logic by putting all the statements
27173 instrument 'base + len' inside a sequence, and then insert that
27174 sequence right before the current insertion point. Then, to
27175 instrument 'base + len', just get an iterator on that statement.
27176 And do not forget to update the pointer to iterator the function
27177 received as argument.
27179 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
27181 PR rtl-optimization/56348
27182 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
27184 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
27186 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
27187 (clean_graph_dump_file): Pass base to start_graph_dump.
27189 2013-02-14 Richard Henderson <rth@redhat.com>
27192 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
27194 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
27196 * collect2-aix.h: Define F_LOADONLY.
27198 2013-02-14 Richard Biener <rguenther@suse.de>
27201 * varasm.c (output_constant_def_1): Get the decl representing
27202 the constant as argument.
27203 (output_constant_def): Wrap output_constant_def_1.
27204 (make_decl_rtl): Use output_constant_def_1 with the decl
27205 representing the constant.
27206 (build_constant_desc): Optionally re-use a decl already
27207 representing the constant.
27208 (tree_output_constant_def): Adjust.
27210 2013-02-14 Dodji Seketeli <dodji@redhat.com>
27213 * asan.c (instrument_builtin_call): Really put the length of the
27214 second source argument into src1_len.
27216 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27218 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
27219 argument. If it is false, don't create edge from then_bb to
27221 (insert_if_then_before_iter): Pass true to it.
27222 (build_check_stmt): Pass false to it.
27223 (transform_statements): Flush hash table only on extended basic
27224 block boundaries, rather than at the beginning of every bb.
27225 Don't flush hash table on nonfreeing_call_p calls.
27226 * tree-flow.h (nonfreeing_call_p): New prototype.
27227 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
27229 2013-02-13 David S. Miller <davem@davemloft.net>
27231 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
27233 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
27236 * ira.c (max_regno_before_ira): Move from ...
27238 (fix_reg_equiv_init): Use max_regno_before_ira instead of
27241 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27243 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
27245 2013-02-13 Richard Biener <rguenther@suse.de>
27248 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
27249 globals in MEM_REFs.
27251 2013-02-13 Richard Biener <rguenther@suse.de>
27253 * loop-init.c (loop_optimizer_init): Clear loop state when
27254 re-initializing preserved loops.
27255 * loop-unswitch.c (unswitch_single_loop): Return whether
27256 we unswitched the loop. Do not verify loop state here.
27257 (unswitch_loops): When we unswitched a loop discover new loops.
27259 2013-02-13 Kostya Serebryany <kcc@google.com>
27261 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
27263 * sanitizer.def: Rename __asan_init to __asan_init_v1.
27265 2013-02-12 Dodji Seketeli <dodji@redhat.com>
27267 Avoid instrumenting duplicated memory access in the same basic block
27268 * Makefile.in (asan.o): Add new dependency on hash-table.h
27269 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
27270 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
27271 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
27272 (free_mem_ref_resources, has_mem_ref_been_instrumented)
27273 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
27274 (get_mem_ref_of_assignment): New functions.
27275 (get_mem_refs_of_builtin_call): Extract from
27276 instrument_builtin_call and tweak a little bit to make it fit with
27278 (instrument_builtin_call): Use the new
27279 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
27280 of is_gimple_builtin_call.
27281 (instrument_derefs, instrument_mem_region_access): Insert the
27282 instrumented memory reference into the hash table.
27283 (maybe_instrument_assignment): Renamed instrument_assignment into
27284 this, and change it to advance the iterator when instrumentation
27285 actually happened and return true in that case. This makes it
27286 homogeneous with maybe_instrument_assignment, and thus give a
27287 chance to callers to be more 'regular'.
27288 (transform_statements): Clear the memory reference hash table
27289 whenever we enter a new BB, when we cross a function call, or when
27290 we are done transforming statements. Use
27291 maybe_instrument_assignment instead of instrumentation. No more
27292 need to special case maybe_instrument_assignment and advance the
27293 iterator after calling it; it's now handled just like
27294 maybe_instrument_call. Update comment.
27296 2013-02-13 Richard Biener <rguenther@suse.de>
27298 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
27299 Fix loop discovery code.
27301 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
27303 PR inline-asm/56148
27304 * lra-constraints.c (process_alt_operands): Match early clobber
27305 operand with itself. Check conflicts with earlyclobber only if
27306 the operand is not reloaded. Prefer to reload conflicting operand
27307 if earlyclobber and matching operands are the same.
27309 2013-02-12 Richard Biener <rguenther@suse.de>
27312 * lto-streamer-out.c (write_symbol): Do not output symbols
27313 for hard register variables.
27315 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
27318 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
27319 (umulsidi3_insn, mulsidi3_insn): New insns.
27321 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
27323 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
27324 (struct tune_params): Add vec_costs field.
27325 * config/arm/arm.c (arm_builtin_vectorization_cost)
27326 (arm_add_stmt_cost): New functions.
27327 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
27328 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
27329 (arm_default_vec_cost): New struct of type cpu_vec_costs.
27330 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
27331 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
27332 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
27333 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
27335 2013-02-12 Richard Biener <rguenther@suse.de>
27338 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
27339 decls again if possible.
27341 2013-02-12 Richard Biener <rguenther@suse.de>
27343 PR middle-end/56288
27344 * tree-ssa.c (verify_ssa_name): Fix check, move
27345 SSA_NAME_IN_FREE_LIST check up.
27347 2013-02-12 Jakub Jelinek <jakub@redhat.com>
27348 Steven Bosscher <steven@gcc.gnu.org>
27350 PR rtl-optimization/56151
27351 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
27352 equal to op0 or op1, and last_insn pattern is CODE operation
27353 with MEM dest and one of the operands matches that MEM.
27355 2013-02-11 Sriraman Tallam <tmsriram@google.com>
27357 * doc/extend.texi: Document Function Multiversioning and "default"
27358 parameter string to target attribute.
27359 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
27360 target attribute parameter is "default".
27361 (ix86_compare_version_priority): Remove checks for target attribute.
27362 (ix86_mangle_function_version_assembler_name): Change error to sorry.
27363 Remove check for target attribute equal to NULL. Add assert.
27364 (ix86_generate_version_dispatcher_body): Change error to sorry.
27366 2013-02-11 Iain Sandoe <iain@codesourcery.com>
27367 Jack Howarth <howarth@bromo.med.uc.edu>
27368 Patrick Marlier <patrick.marlier@gmail.com>
27371 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
27372 define ENDFILE_SPEC as TM_DESTRUCTOR.
27373 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
27375 2013-02-11 Alexander Potapenko <glider@google.com>
27376 Jack Howarth <howarth@bromo.med.uc.edu>
27377 Jakub Jelinek <jakub@redhat.com>
27380 * config/darwin.c (cdtor_record): Rename ctor_record.
27381 (sort_cdtor_records): Rename sort_ctor_records.
27382 (finalize_dtors): New routine to sort destructors by
27383 priority before use in assemble_integer.
27384 (machopic_asm_out_destructor): Use finalize_dtors if needed.
27386 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
27388 PR rtl-optimization/56275
27389 * simplify-rtx.c (avoid_constant_pool_reference): Check that
27390 offset is non-negative and less than cmode size before
27391 calling simplify_subreg.
27393 2013-02-11 Richard Biener <rguenther@suse.de>
27395 PR tree-optimization/56264
27396 * cfgloop.h (fix_loop_structure): Adjust prototype.
27397 * loop-init.c (fix_loop_structure): Return the number of
27398 newly discovered loops.
27399 * tree-cfgcleanup.c (repair_loop_structures): When new loops
27400 are discovered, do a full loop-closed SSA rewrite.
27402 2013-02-11 Richard Biener <rguenther@suse.de>
27404 PR tree-optimization/56273
27405 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27407 (check_array_ref): Fix missing newline in dumps.
27408 (search_for_addr_array): Likewise.
27410 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
27412 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
27414 2013-02-09 Jakub Jelinek <jakub@redhat.com>
27417 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
27419 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
27421 PR rtl-optimization/56246
27422 * lra-constraints.c (simplify_operand_subreg): Try to reuse
27424 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
27425 constraints are satisfied.
27427 2013-02-08 Jeff Law <law@redhat.com>
27430 * emit-rtl.c (reg_is_parm_p): New function.
27431 * regs.h (reg_is_parm_p): New prototype.
27432 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
27433 callee-clobbered registers.
27435 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
27438 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
27439 If there is no implicit builtin declaration, just return NULL.
27441 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27443 * config/i386/sse.md (FMAMODEM): New mode iterator.
27444 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
27445 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
27447 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27449 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
27450 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
27451 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
27453 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27455 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
27456 (microblaze*-*-elf): Likewise.
27457 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
27459 * config/microblaze/microblaze-c.c: Add builtin defines for
27460 _LITTLE_ENDIAN and _BIG_ENDIAN.
27461 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
27462 add to TARGET_DEFAULT flags.
27463 Expand ASM_SPEC and LINK_SPEC.
27464 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
27465 * config/microblaze/microblaze.md: Update extendsidi2 and
27466 movdi_internal instructions to use low-order / high-order reg
27468 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
27469 options and inversemask / mask of LITTLE_ENDIAN.
27470 * config/microblaze/t-microblaze: Expand multilib options to
27471 include mlittle-endian (le) and update exceptions patterns.
27473 2013-02-08 Jakub Jelinek <jakub@redhat.com>
27475 PR rtl-optimization/56195
27476 * lra-constraints.c (get_reload_reg): Don't reuse regs
27477 if they have smaller mode than requested, if they have
27478 wider mode than requested, try to return a SUBREG.
27480 PR tree-optimization/56250
27481 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
27482 if type is unsigned and code isn't MULT_EXPR.
27484 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27486 PR tree-optimization/56064
27487 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
27488 bits according to mode.
27489 * fixed-value.h (fixed_from_double_int)
27490 (const_fixed_from_double_int): Adjust comments.
27492 2013-02-08 Richard Biener <rguenther@suse.de>
27495 * lto-streamer.h (struct data_in): Remove current_file, current_line
27496 and current_col members.
27497 * lto-streamer-out.c (lto_output_location): Stream changed bits
27498 en-block for efficiency.
27499 * lto-streamer-in.c (clear_line_info): Remove.
27500 (lto_input_location): Cache current file, line and column
27501 globally via local statics. Read changed bits en-block.
27502 (input_function): Do not call clear_line_info.
27503 (lto_read_body): Likewise.
27504 (lto_input_toplevel_asms): Likewise.
27506 2013-02-08 Michael Matz <matz@suse.de>
27508 PR tree-optimization/52448
27509 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
27510 (nt_call_phase): New static.
27511 (add_or_mark_expr): Only mark accesses with newer phase than any
27513 (nonfreeing_call_p): New.
27514 (nt_init_block): Update nt_call_phase, mark blocks as visited.
27515 (nt_fini_block): Keep blocks marked as visited.
27516 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
27518 2013-02-08 Richard Biener <rguenther@suse.de>
27520 * ira.c (ira): Free broken dominator information.
27522 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27524 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
27526 2013-02-08 Marek Polacek <polacek@redhat.com>
27528 * cfgloop.c (verify_loop_structure): Add more checking of headers.
27530 2013-02-08 Richard Biener <rguenther@suse.de>
27532 PR middle-end/56181
27533 * cfgloop.h (flow_loops_find): Adjust.
27534 (bb_loop_header_p): Declare.
27535 * cfgloop.c (bb_loop_header_p): New function split out from ...
27536 (flow_loops_find): ... here. Adjust function signature,
27537 support incremental loop structure update.
27538 (verify_loop_structure): Cleanup. Verify a loop is a loop.
27539 * cfgloopmanip.c (fix_loop_structure): Move ...
27540 * loop-init.c (fix_loop_structure): ... here.
27541 (apply_loop_flags): Split out from ...
27542 (loop_optimizer_init): ... here.
27543 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
27544 in incremental mode, only remove dead loops here.
27546 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27549 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
27550 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
27551 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
27552 (*round<mode>3.libgcc): New insns for fixed-modes.
27553 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
27554 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
27555 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
27556 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
27557 implementations. Define to __builtin_avr_absFX,
27558 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
27559 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
27560 __builtin_avr_countlsFX, respectively.
27561 * config/avr/avr-c.c (target.h): Include it.
27562 (enum avr_builtin_id): New enum.
27563 (avr_resolve_overloaded_builtin): New static function.
27564 (avr_register_target_pragmas): Use it to set
27565 targetm.resolve_overloaded_builtin.
27566 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
27567 tree nodes used by DEF_BUILTIN.
27568 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
27569 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
27570 <AVR_BUILTIN_xxBITS>: Same.
27572 2013-02-08 Richard Biener <rguenther@suse.de>
27574 * cfgloop.c (verify_loop_structure): Properly handle
27575 a loop exiting to another loop header.
27576 * ira-int.h (ira_loops): Remove.
27577 * ira.c (ira_loops): Remove.
27578 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
27579 (do_reload): Use loop_optimizer_finalize.
27580 * ira-build.c (create_loop_tree_nodes): Use get_loops and
27581 number_of_loops to access the loop tree.
27582 (more_one_region_p): Likewise.
27583 (finish_loop_tree_nodes): Likewise.
27584 (rebuild_regno_allocno_maps): Likewise.
27585 (mark_loops_for_removal): Likewise.
27586 (mark_all_loops_for_removal): Likewise.
27587 (remove_unnecessary_regions): Likewise.
27588 (ira_build): Likewise.
27589 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
27591 2013-02-08 Richard Biener <rguenther@suse.de>
27593 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
27594 * ipa-pure-const.c (analyze_function): Avoid calling
27595 mark_irreducible_loops twice.
27596 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
27598 2013-02-07 David S. Miller <davem@davemloft.net>
27600 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
27602 * var-tracking.c (vt_add_function_parameter): Test the presence of
27603 HAVE_window_save properly and do not remap argument registers when
27604 we have a leaf function.
27606 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
27609 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
27611 * config/i386/i386.c (ix86_print_operand): Ditto.
27613 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27615 * lra-constraints.c (process_alt_operands): Fix recently added comment.
27617 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27619 PR rtl-optimization/56225
27620 * lra-constraints.c (process_alt_operands): Check that reload hard
27621 reg can hold value for strict_low_part.
27623 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27626 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
27627 dwarf2out_end_function.
27628 (in_first_function_p, maybe_at_text_label_p,
27629 first_loclabel_num_not_at_text_label): New variables.
27630 (dwarf2out_var_location): In the first function find out
27631 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
27632 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
27635 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
27637 PR rtl-optimization/56178
27638 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
27639 SUBREG of a register. Tidy up related block of code.
27640 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
27641 note if the source is a register or a SUBREG of a register.
27643 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27646 * config/rs6000/rs6000.md (ptrm): New mode attr.
27647 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
27648 call_value_indirect_aix<pttrsize>,
27649 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
27652 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
27654 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
27655 if -bnortl. Convert to strcmp and strncmp.
27657 2013-02-07 Alan Modra <amodra@gmail.com>
27660 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
27661 addresses won't wrap when offsetting.
27662 (rs6000_secondary_reload): Provide secondary reloads needed for
27663 wrapping LO_SUM addresses.
27665 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
27667 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
27668 MACH, just __MACH__.
27670 2013-02-06 Richard Biener <rguenther@suse.de>
27672 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
27673 instead of calling fix_loop_structure.
27675 2013-02-06 Jakub Jelinek <jakub@redhat.com>
27677 PR middle-end/56217
27678 * omp-low.c (use_pointer_for_field): Return false if
27679 lower_send_shared_vars doesn't generate any copy-out code.
27681 2013-02-06 Tom de Vries <tom@codesourcery.com>
27683 PR rtl-optimization/56131
27684 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
27685 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
27686 of the label is NULL. Add comment.
27688 2013-02-05 Jakub Jelinek <jakub@redhat.com>
27690 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
27693 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
27694 (STATIC_LIBTSAN_LIBS): Likewise.
27695 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
27696 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
27697 is defined, don't add anything else beyond that.
27698 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
27699 (LINK_COMMAND_SPEC): Use them.
27701 PR tree-optimization/56205
27702 * tree-stdarg.c (check_all_va_list_escapes): Return true if
27703 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
27704 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
27706 2013-02-05 Richard Biener <rguenther@suse.de>
27708 PR tree-optimization/53342
27709 PR tree-optimization/53185
27710 * tree-vectorizer.h (vect_check_strided_load): Remove.
27711 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
27712 not disallow peeling for vectorized strided loads.
27713 (vect_check_strided_load): Make static and simplify.
27714 (vect_analyze_data_refs): Adjust.
27715 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
27716 correctly when vectorizing strided loads.
27718 2013-02-05 Richard Biener <rguenther@suse.de>
27720 * doc/install.texi: Refer to ISL, not PPL.
27722 2013-02-05 Jan Hubicka <jh@suse.cz>
27724 PR tree-optimization/55789
27725 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
27727 2013-02-05 Jan Hubicka <jh@suse.cz>
27729 PR tree-optimization/55789
27730 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
27731 the dead call anyway.
27733 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
27736 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
27738 2013-02-04 Alexander Potapenko <glider@google.com>
27739 Jack Howarth <howarth@bromo.med.uc.edu>
27740 Jakub Jelinek <jakub@redhat.com>
27743 * config/darwin.c (sort_ctor_records): Stabilized qsort
27744 on constructor priority by using original position.
27745 (finalize_ctors): New routine to sort constructors by
27746 priority before use in assemble_integer.
27747 (machopic_asm_out_constructor): Use finalize_ctors if needed.
27749 2013-02-04 Jakub Jelinek <jakub@redhat.com>
27752 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
27753 about visibility on artificial decls.
27754 * config/sol2.c (solaris_assemble_visibility): Likewise.
27756 2013-02-04 Kai Tietz <ktietz@redhat.com>
27759 * config/i386/i386.c (function_value_ms_64): Add additional valtype
27760 argument and improve checking of return-argument types for 16-byte
27762 (ix86_function_value_1): Add additional valtype argument on call
27763 of function_value_64.
27764 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
27765 handling infunction_value_64 function.
27767 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
27769 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
27771 2013-02-04 Richard Biener <rguenther@suse.de>
27773 PR tree-optimization/56188
27774 * tree-ssa-structalias.c (label_visit): Consider case with
27775 initially non-empty points-to set.
27776 (perform_var_substitution): Dump node mapping and clean up.
27778 2013-02-04 Richard Guenther <rguenther@suse.de>
27781 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
27782 node prevail as last resort.
27783 (lto_symtab_merge_decls): Remove guard on LTRANS here.
27784 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
27786 2013-02-04 Richard Biener <rguenther@suse.de>
27788 PR tree-optimization/56113
27789 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
27791 (equiv_class_lookup_or_add): ... this.
27792 (label_visit): Adjust and fix error in previous patch.
27793 (perform_var_substitution): Adjust.
27795 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
27797 * config/sh/divtab.c: Fix formatting and comments throughout the file.
27798 * config/sh/sh4-300.md: Likewise.
27799 * config/sh/sh4a.md: Likewise.
27800 * config/sh/constraints.md: Likewise.
27801 * config/sh/sh.md: Likewise.
27802 * config/sh/netbsd-elf.h: Likewise.
27803 * config/sh/predicates.md: Likewise.
27804 * config/sh/sh-protos.h: Likewise.
27805 * config/sh/ushmedia.h: Likewise.
27806 * config/sh/linux.h: Likewise.
27807 * config/sh/sh.c: Likewise.
27808 * config/sh/superh.h: Likewise.
27809 * config/sh/elf.h: Likewise.
27810 * config/sh/sh4.md: Likewise.
27811 * config/sh/sh.h: Likewise.
27813 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27815 * config/pa/constraints.md: Adjust unused letters. Change "T"
27816 constraint to match_test floating_point_store_memory_operand().
27817 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
27818 (base14_operand): New.
27819 (floating_point_store_memory_operand): New.
27820 (integer_store_memory_operand): Revise to use base14_operand and
27821 reg_plus_base_memory_operand.
27822 (move_dest_operand): Allow symbolic_memory_operands.
27823 (symbolic_memory_operand): Check for LO_SOM.
27824 (symbolic_operand): Change default case to break.
27825 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
27826 CONST_DOUBLE values to be reloaded by putting them into memory when
27827 the destination is a floating point register.
27828 (movdf): Remove code to handle CONST_DOUBLE.
27830 (reload_indf_r1): New.
27831 (reload_insf_r1): New.
27832 Consistently use "Q" and "T" constraints with integer and floating
27833 point move instructions, respectively.
27834 (movdi): Remove FAIL.
27835 Change predicate for source operand unamed DImode move from
27836 general_operand to move_src_operand.
27837 (umulsidi3): Change predicate for destination operand to
27839 Likewise for similar unamed patterns.
27840 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
27841 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
27842 (hppa_legitimize_address): Simplify mask calculation.
27843 (pa_emit_move_sequence): Revised handling of secondary reloads from
27844 REG+D addresses for floating point loads and stores. Directly handle
27845 loading CONST0_RTX (mode) to a floating point register.
27846 (pa_secondary_reload): Handle reloading DF and SFmode constant values
27847 to floating point registers. Don't restrict secondary reloads to
27848 floating point registers to integer modes. Revise some comments and
27850 (TARGET_LEGITIMATE_ADDRESS_P): Define.
27851 (pa_legitimate_address_p): New.
27852 (pa_legitimize_reload_address): New.
27853 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
27854 (STRICT_REG_OK_FOR_BASE_P): New.
27855 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
27856 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
27858 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
27859 Andrew Dixie <andrewd@gentrack.com>
27861 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
27864 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
27866 * expmed.c (extract_bit_field_1): Pass the full width of the
27867 structure to get_best_reg_extraction_insn.
27869 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
27872 * configure.ac (use_cxa_atexit): Add AIX.
27873 * configure: Regenerate.
27875 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
27877 2013-02-01 Jakub Jelinek <jakub@redhat.com>
27880 * final.c (need_profile_function): New variable.
27881 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
27882 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
27883 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
27884 notes, targetm.asm_out.function_prologue doesn't emit anything,
27885 HAVE_prologue and profiler should be emitted before prologue,
27886 set need_profile_function instead of emitting it.
27887 (final_scan_insn): If need_profile_function, emit
27888 profile_function on the first NOTE_INSN_BASIC_BLOCK or
27889 NOTE_INSN_FUNCTION_BEG note.
27891 2013-02-01 Richard Henderson <rth@redhat.com>
27893 * config/rs6000/rs6000.md (smulditi3): New.
27896 * config/alpha/alpha.md (umulditi3): New.
27898 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
27900 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
27901 (ASM_OUTPUT_ALIGNED_LOCAL): New.
27903 2013-02-01 Richard Biener <rguenther@suse.de>
27905 PR tree-optimization/56113
27906 * tree-ssa-structalias.c (label_visit): Reduce work for
27907 single-predecessor nodes.
27909 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
27911 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
27912 range isn't testing for zero.
27914 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
27916 PR middle-end/56113
27917 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
27919 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
27920 Nick Clifton <nickc@redhat.com>
27922 * config/v850/constraints.md (Q): Define as a memory constraint.
27923 * config/v850/predicates.md (label_ref_operand): New predicate.
27924 (e3v5_shift_operand): New predicate.
27925 (ior_operator): New predicate.
27926 * config/v850/t-v850: Add e3v5 multilib.
27927 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
27928 (v850_gen_movdi): Prototype.
27929 * config/v850/v850.c: Add support for e3v5 architecture.
27930 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
27932 (construct_save_jarl): Add e3v5 long JARL support.
27933 (v850_adjust_insn_length): New function. Adjust length of call
27934 insns when using e3v5 instructions.
27935 (v850_gen_movdi): New function: Generate instructions to move a
27937 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
27938 (CPP_SPEC): Define __v850e3v5__ as appropriate.
27939 (TARGET_USE_FPU): Enable for e3v5.
27940 (CONST_OK_FOR_W): New macro.
27941 (ADJUST_INSN_LENGTH): Define.
27942 * config/v850/v850.md (UNSPEC_LOOP): Define.
27943 (attr cpu): Add v850e3v5.
27944 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
27945 (movdi): New pattern.
27946 (movdi_internal): New pattern.
27947 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
27948 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
27949 (cstoresf4): Likewise.
27950 (cstoredf4): Likewise.
27951 (insv): New pattern.
27952 (rotlso3_a): New pattern.
27953 (rotlsi3_b): New pattern
27954 (rotlsi3_v850e3v5): New pattern.
27955 (doloop_begin): New pattern.
27956 (fix_loop_counter): New pattern.
27957 (doloop_end): New pattern.
27958 (branch_normal): Add e3v5 long branch support.
27959 (branch_invert): Likewise.
27960 (branch_z_normal): Likewise.
27961 (branch_z_invert): Likewise.
27962 (branch_nz_normal): Likewise.
27963 (branch_nz_invert): Likewise.
27964 (call_internal_short): Add e3v5 register-indirect JARL support.
27965 (call_internal_long): Likewise.
27966 (call_value_internal_short): Likewise.
27967 (call_value_internal_long): Likewise.
27968 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
27969 (mloop): New option.
27970 * config.gcc: Add support for configuring v840e3v5 target.
27971 * doc/invoke.texi: Document new v850 specific command line options.
27973 2013-01-31 Paul Koning <ni1d@arrl.net>
27977 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
27978 children if parent is a class.
27979 (prune_unused_types_prune): Don't add DW_AT_declaration.
27981 2013-01-31 Richard Biener <rguenther@suse.de>
27983 PR tree-optimization/56157
27984 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
27985 match up operand with SLP child.
27987 2013-01-31 Jason Merrill <jason@redhat.com>
27990 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
27991 parameters the first time.
27992 (gen_scheduled_generic_parms_dies): Check completeness here.
27994 2013-01-31 Richard Biener <rguenther@suse.de>
27996 PR middle-end/53073
27997 * common.opt (faggressive-loop-optimizations): New flag,
27998 enabled by default.
27999 * doc/invoke.texi (faggressive-loop-optimizations): Document.
28000 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
28001 infer_loop_bounds_from_undefined by it.
28003 2013-01-31 Richard Biener <rguenther@suse.de>
28005 PR tree-optimization/56150
28006 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
28007 visit virtual operands.
28008 (find_uses_to_rename_bb): Likewise.
28010 2013-01-31 Richard Biener <rguenther@suse.de>
28012 PR tree-optimization/56150
28013 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
28014 mixed store non-store stmts.
28016 2013-01-30 Jakub Jelinek <jakub@redhat.com>
28019 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
28020 LIBASAN_EARLY_SPEC is defined.
28021 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
28022 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
28024 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
28027 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
28028 invalid args instead of ICEing on it.
28029 (ix86_valid_target_attribute_tree): Return error_mark_node if
28030 ix86_valid_target_attribute_inner_p failed.
28031 (ix86_valid_target_attribute_p): Return false only if
28032 ix86_valid_target_attribute_tree returned error_mark_node. Allow
28033 target("default") attribute.
28034 (sorted_attr_string): Change argument from const char * to tree,
28035 merge in all target attribute arguments rather than just one.
28036 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
28037 instead of free. Avoid using strcat.
28038 (ix86_mangle_function_version_assembler_name): Mangle
28039 target("default") as if no target attribute is present. Adjust
28040 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
28041 instead of xmalloc and XDELETEVEC instead of free.
28042 (ix86_function_versions): Don't return true if one of the decls
28043 doesn't have target attribute. If they don't and one of the decls
28044 is DECL_FUNCTION_VERSIONED, report an error. Adjust
28045 sorted_attr_string caller. Use XDELETEVEC instead of free.
28046 (ix86_supports_function_versions): Remove.
28047 (make_name): Fix up formatting.
28048 (make_dispatcher_decl): Remove resolver_name and its initialization.
28049 Avoid leaking memory.
28050 (is_function_default_version): Return true if there is
28051 target("default") attribute rather than no target attribute at all.
28052 (make_resolver_func): Avoid leaking memory.
28053 (ix86_generate_version_dispatcher_body): Likewise.
28054 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
28055 * target.def (supports_function_versions): Remove.
28056 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
28057 * doc/tm.texi: Regenerated.
28059 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
28061 PR rtl-optimization/56144
28062 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
28063 for values with side effects.
28065 2013-01-30 Richard Biener <rguenther@suse.de>
28067 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
28068 (sparseset_pop): Likewise.
28069 * cfganal.c (compute_idf): Likewise. Increase work-stack size
28070 to be able to use quick_push in the worker loop.
28072 2013-01-30 Marek Polacek <polacek@redhat.com>
28074 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
28076 2013-01-30 Richard Biener <rguenther@suse.de>
28079 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
28081 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
28083 PR tree-optimization/56064
28084 * fixed-value.c (fixed_from_double_int): New function.
28085 * fixed-value.h (fixed_from_double_int): New prototype.
28086 (const_fixed_from_double_int): New static inline function.
28087 * fold-const.c (native_interpret_fixed): New static function.
28088 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
28089 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
28090 (native_encode_fixed): New static function.
28091 (native_encode_expr) <FIXED_CST>: Use it.
28092 (native_interpret_int): Move double_int worker code to...
28093 * double-int.c (double_int::from_buffer): ...this new static method.
28094 * double-int.h (double_int::from_buffer): Prototype it.
28096 2013-01-30 Richard Biener <rguenther@suse.de>
28098 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
28099 New pointer-map and obstack.
28100 (init_alias_vars): Allocate pointer-map and obstack.
28101 (delete_points_to_sets): Free them.
28102 (find_what_var_points_to): Cache result.
28103 (find_what_p_points_to): Adjust for changed interface of
28104 find_what_var_points_to.
28105 (compute_points_to_sets): Likewise.
28106 (ipa_pta_execute): Likewise.
28108 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
28110 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
28111 * configure: Regenerate.
28112 * config.in: Regenerate.
28113 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
28114 #nobits/#progbits if supported.
28116 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
28119 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
28120 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
28121 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
28123 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28125 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
28126 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
28128 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28130 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
28131 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
28133 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28135 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
28137 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
28138 * config/arm/cortex-a7.md: New bypasses using
28139 arm_mac_accumulator_is_result.
28141 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28143 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
28144 (cortex_a7_neon_mla): Likewise.
28145 (cortex_a7_fpfmad): New reservation.
28146 (cortex_a7_fpmacs): Use ffmas and update required units.
28147 (cortex_a7_fpmuld): Update required units and latency.
28148 (cortex_a7_fpmacd): Likewise.
28149 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
28150 (cortex_a7_neon). Likewise.
28151 (bypass) Update participating units.
28153 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28155 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
28156 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
28158 * config/arm/vfp11.md (vfp_farith): Use ffmas.
28159 (vfp_fmul): Use ffmad.
28160 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
28161 (cortex_r4_fmacd): Use ffmad.
28162 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
28163 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
28164 (cortex_a9_fmacd): Use ffmad.
28165 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
28166 (cortex_a8_vfp_macd): Use ffmad.
28167 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
28168 (cortex_a5_fpmacd): Use ffmad.
28169 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
28170 (cortex_a15_vfp_macd): Use ffmad.
28171 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
28173 2013-01-29 Jason Merrill <jason@redhat.com>
28176 * varasm.c (default_assemble_visibility): Don't warn about
28177 visibility on artificial decls.
28179 2013-01-29 Richard Biener <rguenther@suse.de>
28181 PR tree-optimization/56113
28182 * tree-ssa-structalias.c (equiv_class_lookup): Also return
28184 (label_visit): Free duplicate bitmaps and record the leader instead.
28185 (perform_var_substitution): Adjust.
28187 2013-01-29 Richard Biener <rguenther@suse.de>
28189 PR tree-optimization/55270
28190 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
28191 the CFG, schedule loops for fixup.
28193 2013-01-29 Nick Clifton <nickc@redhat.com>
28195 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
28198 2013-01-28 Leif Ekblad <leif@rdos.net>
28200 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
28201 * config/i386/i386.h (TARGET_RDOS): New macro.
28202 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
28203 * config/i386/i386.c (ix86_option_override_internal): For 64bit
28204 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
28205 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
28206 DEFAULT_LARGE_SECTION_THRESHOLD.
28207 * config/i386/i386.md (R14_REG, R15_REG): New constants.
28208 * config/i386/rdos.h: New file.
28209 * config/i386/rdos64.h: New file.
28211 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
28214 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
28217 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28219 PR rtl-optimization/56117
28220 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
28221 call cselib_lookup_from_insn on the MEM before calling
28222 add_insn_mem_dependence.
28224 2013-01-28 Richard Biener <rguenther@suse.de>
28226 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
28227 to a stmt that didn't have one.
28228 (copy_phis_for_bb): Likewise for PHI arguments.
28229 (copy_debug_stmt): Likewise for debug stmts.
28231 2013-01-28 Richard Biener <rguenther@suse.de>
28233 PR tree-optimization/56034
28234 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
28235 (partition_builtin_p): Adjust.
28236 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
28237 it is the last partition.
28238 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
28239 up the vertex for the definition.
28240 (classify_partition): Classify whether a partition is a
28241 PKIND_REDUCTION, thus has uses outside of the loop.
28242 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
28243 Merge all PKIND_REDUCTION partitions into the last partition.
28244 (tree_loop_distribution): Seed partitions from reductions as well.
28246 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28248 PR tree-optimization/56125
28249 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
28250 pow(x,c) into sqrt(x) * powi(x, n/2) or
28251 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
28252 optimizing for size.
28253 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
28254 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
28257 PR tree-optimization/56094
28258 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
28259 to UNKNOWN_LOCATION while gimplifying expr.
28261 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
28264 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
28265 operand 0 in movabs insn template for -masm=intel asm alternative.
28266 (*movabs<mode>_2): Ditto for operand 1.
28268 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
28271 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
28274 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28276 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
28279 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
28281 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
28282 Undef to avoid warning.
28284 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28286 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
28287 * configure: Regenerate.
28289 2013-01-25 Jakub Jelinek <jakub@redhat.com>
28291 PR tree-optimization/56098
28292 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
28293 for stmts with volatile ops.
28294 (cond_store_replacement): Don't optimize if assign has volatile ops.
28295 (cond_if_else_store_replacement_1): Don't optimize if either
28296 then_assign or else_assign have volatile ops.
28297 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
28300 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28302 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
28304 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28306 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
28307 missing ':' in asm example.
28309 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
28311 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
28312 entries into lane and laneq entries.
28313 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
28314 Remove AdvSIMD scalar modes.
28315 (aarch64_sq<r>dmulh_laneq<mode>): New.
28316 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
28318 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
28319 builtin implementations to relfect changes in RTL in aarch64-simd.md.
28320 * config/aarch64/iterators.md (VCOND): New.
28323 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28326 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
28327 Add NULL LIBNAME argument to existing definitions.
28328 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
28329 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
28330 * config/avr/avr.c (DEF_BUILTIN): Same.
28331 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
28332 (avr_expand_builtin): Expand to a vanilla call if a libgcc
28333 implementation is available (DECL_ASSEMBLER_NAME is set).
28334 (avr_fold_absfx): New static function.
28335 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
28336 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
28337 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
28338 AVR_BUILTIN_ABSLLK.
28339 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
28340 (abshk, absk, abslk, absllk): Provide as static inline functions.
28342 2013-01-25 Marek Polacek <polacek@redhat.com>
28344 PR tree-optimization/56035
28345 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
28347 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
28349 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
28350 (*movtf_internal_rex64): Add (!o,C) alternative
28351 (*movxf_internal_rex64): Ditto.
28352 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
28354 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
28356 * doc/invoke.texi: fix typo.
28357 * doc/objc.texi: fix typo.
28359 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
28361 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
28362 for the first two alternatives.
28364 2013-01-24 Diego Novillo <dnovillo@google.com>
28366 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
28367 (ggc-zone.o): Remove.
28368 * configure.ac: Remove option --with-gc.
28369 * configure: Re-generate.
28370 * doc/install.texi: Remove documentation for --with-gc.
28371 * gengtype.c (write_enum_defn): Remove. Update all users.
28372 (write_Types_process_field): Remove generation of gt_e_* argument.
28373 (output_type_enum): Remove. Update all users.
28374 (write_enum_defn): Remove. Update all users.
28375 (enum alloc_zone): Remove. Update all users.
28376 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
28377 * ggc-common.c (ggc_splay_alloc): Remove first argument.
28378 Update all callers.
28379 (struct ptr_data): Remove field TYPE. Update all users.
28380 (gt_pch_note_object): Remove argument TYPE. Update all users.
28381 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
28383 * ggc-none.c (ggc_alloc_typed_stat): Remove.
28384 (struct alloc_zone): Remove.
28385 (ggc_internal_alloc_zone_stat): Remove.
28386 (ggc_internal_cleared_alloc_zone_stat): Remove.
28387 * ggc-page.c (ggc_alloc_typed_stat): Remove.
28388 (ggc_pch_count_object): Remove last argument. Update all users.
28389 (ggc_pch_alloc_object): Remove last argument. Update all users.
28390 (struct alloc_zone): Remove.
28391 * ggc-zone.c: Remove.
28392 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
28393 (struct alloc_zone): Remove.
28394 (ggc_alloc_typed_stat): Remove.
28395 (ggc_alloc_typed): Remove.
28396 (ggc_splay_alloc): Remove first argument.
28397 (rtl_zone): Remove. Update all users.
28398 (tree_zone): Remove. Update all users.
28399 (tree_id_zone): Remove. Update all users.
28400 (ggc_internal_zone_alloc_stat): Remove. Update all users.
28401 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
28402 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
28403 * tree-ssanames.c: Remove references to zone allocator in comments.
28405 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
28407 * config/avr/avr.c (avr_out_fract): Make register numbers that
28408 might be outside of source operand signed.
28410 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
28412 * config/i386/constraints.md (Yf): New constraint.
28413 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
28414 of f constraint to conditionaly disable x87 register preferences.
28415 (*movdf_internal): Ditto.
28416 (*movsf_internal): Ditto.
28418 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
28420 PR inline-asm/55934
28421 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
28422 that have operands with impossible constraints.
28423 Add a FIXME for a speed-up opportunity.
28424 * lra-constraints.c (process_alt_operands): Verify that a class
28425 selected from constraints on asms is valid for the operand mode.
28426 (curr_insn_transform): Remove incorrect comment.
28428 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
28430 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
28431 TOC operand is a valid symbol ref in the constant pool.
28433 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28435 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
28437 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
28440 * config/avr/stdfix.h: New file.
28441 * t-avr (stdfix-gcc.h): New rule to build it.
28442 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
28444 2013-01-23 Kostya Serebryany <kcc@google.com>
28446 * config/darwin.h: remove dependency on
28447 CoreFoundation (asan on Mac OS).
28449 2013-01-23 Jakub Jelinek <jakub@redhat.com>
28452 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
28453 instead of cmpdi_operand for first comparison operand.
28454 Don't assert that comparison operands aren't both constants.
28456 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28458 * doc/install.texi (Downloading the Source): Update references to
28459 downloading separate components.
28461 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28463 * doc/extend.texi (__int128): Improve grammar.
28465 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
28468 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
28469 alternative to (o,r).
28470 (*movdi_internal_rex64): Remove (!o,n) alternative.
28471 (DImode immediate->memory splitter): Remove.
28472 (DImode immediate->memory peephole2): Remove.
28473 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
28474 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
28475 alternative to (!o,*r).
28476 (*movtf_internal_sse): New pattern.
28477 (*movxf_internal_rex64): New pattern.
28478 (*movxf_internal): Disable for TARGET_64BIT.
28479 (*movdf_internal_rex64): Remove (!o,F) alternative.
28481 2013-01-22 Jakub Jelinek <jakub@redhat.com>
28483 PR middle-end/56074
28484 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
28485 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
28486 * tree-vect-loop-manip.c (find_loop_location): Also ignore
28487 stmt locations where LOCATION_LOCUS of the stmt location is
28488 UNKNOWN_LOCATION or BUILTINS_LOCATION.
28491 * config/i386/i386.md (UNSPEC_STOS): New.
28492 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
28493 *strsetqi_1): Add UNSPEC_STOS.
28495 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
28498 * doc/invoke.texi: Remove left over -Wsynth example.
28500 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28502 PR tree-optimization/56051
28503 * fold-const.c (fold_binary_loc): Don't fold
28504 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
28505 a narrowing conversion, or widening conversion from signed
28508 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28510 PR rtl-optimization/56023
28511 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
28512 dependent on debug instruction.
28514 2013-01-21 Martin Jambor <mjambor@suse.cz>
28516 PR middle-end/56022
28517 * function.c (allocate_struct_function): Call
28518 invoke_set_current_function_hook earlier.
28520 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28522 * reload1.c (init_reload): Only initialize reload_obstack
28523 during the first call.
28525 2013-01-21 Marek Polacek <polacek@redhat.com>
28527 * cfgloop.c (verify_loop_structure): Fix up grammar.
28529 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
28531 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
28532 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
28534 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28537 * config/arm/marvell-pj4.md: Update copyright year.
28538 Fix up use of alu to alu_reg and simple_alu_imm.
28540 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28542 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
28544 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
28547 * lra-constraints.c (curr_insn_transform): Don't reuse original
28548 insn for secondary memory move when memory mode should be different.
28550 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28552 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
28553 atomic_storedi_1): New patterns.
28555 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
28557 btver2 pipeline descriptions.
28558 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
28560 * config/i386/i386.md (btver2_decode): New type attributes.
28561 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
28563 * config/i386/btver2.md: New file describing btver2 pipelines.
28565 2013-01-19 Andrew Pinski <apinski@cavium.com>
28567 PR tree-optimization/52631
28568 * tree-ssa-sccvn (visit_use): Before looking up the original
28569 statement, try looking up the simplified expression.
28571 2013-01-19 Anthony Green <green@moxielogic.com>
28573 * config/moxie/moxie.c (moxie_expand_prologue): Set
28574 current_function_static_stack_size.
28576 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28578 PR tree-optimization/56029
28579 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
28580 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
28582 2013-01-18 Sharad Singhai <singhai@google.com>
28584 PR tree-optimization/55995
28585 * dumpfile.c (dump_loc): Print location only if available.
28586 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
28588 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
28591 * lra-constraints.c (curr_insn_transform): Reuse original insn for
28592 secondary memory move.
28593 (inherit_reload_reg): Use rclass instead of cl for
28594 check_secondary_memory_needed_p.
28596 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28598 PR middle-end/56015
28599 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
28600 the case where writing real complex part of target modifies op1.
28602 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
28604 * config/aarch64/aarch64-simd.md
28605 (aarch64_vcond_internal<mode>): Handle unordered cases.
28606 * config/aarch64/iterators.md (v_cmp_result): New.
28608 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
28609 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28611 * config/arm/marvell-pj4.md: New file.
28612 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
28613 * config/arm/arm.md (generic_sched): Add marvell_pj4.
28614 (generic_vfp): Likewise.
28615 * config/arm/arm-cores.def: Add marvell-pj4.
28616 * config/arm/arm-tune.md: Regenerate.
28617 * config/arm/arm-tables.opt: Regenerate.
28618 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
28619 * doc/invoke.texi: Document marvell-pj4.
28621 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
28623 * config/aarch64/arm_neon.h: Map scalar types to standard types.
28625 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28630 * var-tracking.c (negative_power_of_two_p): New.
28631 (global_get_addr_cache, local_get_addr_cache): New.
28632 (get_addr_from_global_cache, get_addr_from_local_cache): New.
28633 (vt_canonicalize_addr): Rewrite using the above. Adjust the
28635 (vt_stack_offset_p): Remove.
28636 (vt_canon_true_dep): Always canonicalize loc's address.
28637 (clobber_overlapping_mems): Make sure we have a MEM.
28638 (local_get_addr_clear_given_value): New.
28639 (val_reset): Clear local cached entries.
28640 (compute_bb_dataflow): Create and release the local cache.
28641 Disable duplicate MEMs clobbering.
28642 (emit_notes_in_bb): Clobber MEMs likewise.
28643 (vt_emit_notes): Create and release the local cache.
28644 (vt_initialize, vt_finalize): Create and release the global
28645 cache, respectively.
28646 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
28648 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28650 PR libmudflap/53359
28651 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
28652 not found in the symtab.
28654 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28657 PR rtl-optimization/55547
28658 PR rtl-optimization/53827
28661 * alias.c (offset_overlap_p): New, factored out of...
28662 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
28663 the conservative special case for symbolic constants. Don't
28664 adjust zero sizes on alignment.
28666 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
28668 PR rtl-optimization/52573
28669 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
28670 REG_UNUSED for the same register.
28672 2013-01-17 Richard Biener <rguenther@suse.de>
28673 Marek Polacek <polacek@redhat.com>
28675 PR rtl-optimization/55833
28676 * loop-unswitch.c (unswitch_loops): Move loop verification...
28677 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
28678 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
28679 Set it to true when we're removing a loop from hierarchy tree in
28680 an irreducible region.
28681 (fix_bb_placements): Adjust caller.
28682 (fix_loop_placements): Likewise.
28684 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
28686 * config/avr/builtins.def (DEF_BUILTIN): Factor out
28687 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
28688 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
28689 Remove ID. Adjust comments.
28690 * config/avr/avr-c.c (avr_builtin_name): Remove.
28691 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
28692 * config/avr/avr.c (avr_tolower): New static function.
28693 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
28694 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
28695 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
28698 2013-01-17 Jan Hubicka <jh@suse.cz>
28700 PR tree-optimization/55273
28701 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
28703 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
28706 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
28707 store through atomic_store<mode>_1.
28708 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
28710 2013-01-17 Martin Jambor <mjambor@suse.cz>
28712 PR tree-optimizations/55264
28713 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
28714 for virtual methods.
28715 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
28716 virtual methods before inlining is over.
28717 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
28719 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
28722 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
28724 PR rtl-optimization/56005
28725 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
28726 pending reads for prefetch.
28728 2013-01-16 Ian Bolton <ian.bolton@arm.com>
28730 * config/aarch64/aarch64.md
28731 (*cstoresi_neg_uxtw): New pattern.
28732 (*cmovsi_insn_uxtw): New pattern.
28733 (*<optab>si3_uxtw): New pattern.
28734 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
28735 (*<optab>si3_insn_uxtw): New pattern.
28736 (*bswapsi2_uxtw): New pattern.
28738 2013-01-16 Richard Biener <rguenther@suse.de>
28740 * tree-inline.c (tree_function_versioning): Remove set but
28741 never used variable.
28743 2013-01-16 Richard Biener <rguenther@suse.de>
28745 PR tree-optimization/55964
28746 * tree-flow.h (rename_variables_in_loop): Remove.
28747 (rename_variables_in_bb): Likewise.
28748 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
28749 (copy_loop_before): Adjust and delete update-ssa status.
28750 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
28751 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
28752 (rename_variables_in_loop): Remove.
28753 (slpeel_update_phis_for_duplicate_loop): Likewise.
28754 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
28755 use available cfg machinery instead of duplicating it.
28756 Update PHI nodes and perform poor-mans SSA update here.
28757 (slpeel_tree_peel_loop_to_edge): Adjust.
28759 2013-01-16 Richard Biener <rguenther@suse.de>
28761 PR tree-optimization/54767
28762 PR tree-optimization/53465
28763 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
28764 (vrp_visit_phi_node): For PHI arguments coming via backedges
28765 drop all symbolical range information.
28766 (execute_vrp): Compute backedges.
28768 2013-01-16 Richard Biener <rguenther@suse.de>
28770 * doc/install.texi: Update CLooG and ISL requirements to
28773 2013-01-16 Christian Bruel <christian.bruel@st.com>
28776 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
28777 (broken_move): Handle UNSPECV_SP_SWITCH_B.
28778 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
28780 2013-01-16 DJ Delorie <dj@redhat.com>
28782 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
28783 (UNSPECV_SP_SWITCH_E): New.
28784 (sp_switch_1): Change to an unspec.
28785 (sp_switch_2): Change to an unspec. Don't use post-inc when we
28788 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
28790 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
28791 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
28792 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
28793 (expand_mem_thread_fence): Ditto.
28794 (expand_mem_signal_fence): Ditto.
28795 (expand_atomic_load): Ditto.
28796 (expand_atomic_store): Ditto.
28798 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
28800 PR rtl-optimization/55547
28801 PR rtl-optimization/53827
28804 * alias.c (memrefs_conflict_p): Set sizes to negative after
28807 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28810 * function.c (thread_prologue_and_epilogue_insns): Always
28811 add crtl->drap_reg to set_up_by_prologue.set, even if
28812 stack_realign_drap is false.
28814 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
28816 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
28817 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
28818 *call): Fix indention.
28820 2013-01-15 Tom de Vries <tom@codesourcery.com>
28823 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
28826 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
28828 PR rtl-optimization/55153
28829 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
28831 2013-01-15 Martin Jambor <mjambor@suse.cz>
28833 PR tree-optimization/55920
28834 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
28835 accesses as grp_to_be_debug_replaced.
28837 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28839 PR tree-optimization/55920
28840 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
28841 there is non-useless type conversion needed from debug rhs to lhs,
28842 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
28844 2013-01-15 Joseph Myers <joseph@codesourcery.com>
28845 Mikael Pettersson <mikpe@it.uu.se>
28848 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
28850 (ASM_OUTPUT_CASE_LABEL): Remove.
28851 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
28852 * final.c (shorten_branches): Update alignment of labels before
28853 jump tables if CASE_VECTOR_SHORTEN_MODE.
28855 2013-01-15 Richard Biener <rguenther@suse.de>
28858 * system.h: Do not include gmp.h for building host tools.
28860 2013-01-15 Richard Biener <rguenther@suse.de>
28862 PR middle-end/55882
28863 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
28864 account for bitpos when computing alignment.
28866 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
28868 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
28869 (ix86_target_macros_internal): Likewise.
28871 * config/i386/i386.c (m_CORE2I7): Removed.
28872 (m_CORE_HASWELL): New macro.
28873 (m_CORE_ALL): Likewise.
28874 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
28875 (initial_ix86_arch_features): Likewise.
28876 (processor_target_table): Initializations for Core avx2.
28877 (cpu_names): New names "core-avx2".
28878 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
28879 PROCESSOR_CORE_HASWELL.
28880 (ix86_issue_rate): New case.
28881 (ia32_multipass_dfa_lookahead): Likewise.
28882 (ix86_sched_init_global): Likewise.
28884 * config/i386/i386.h (TARGET_HASWELL): New macro.
28885 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
28886 (processor_type): New PROCESSOR_HASWELL.
28888 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28890 PR tree-optimization/55955
28891 * tree-vect-loop.c (vectorizable_reduction): Give up early on
28892 *SHIFT_EXPR and *ROTATE_EXPR codes.
28894 PR tree-optimization/48766
28895 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
28896 -ftrapv disable -fwrapv.
28898 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28901 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
28902 etc. to 1 and not to __flash.
28903 Use LL suffix for __INT24_MAX__ with -mint8.
28904 Use ULL suffix for __UINT24_MAX__ with -mint8.
28906 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28908 * config/avr/avr-arch.h
28909 (struct base_arch_s): Use typedef avr_arch_t instead.
28910 (struct arch_info_s): Use typedef avr_arch_info_t instead.
28911 (struct mcu_type_s): Use typedef avr_mcu_t instead.
28912 * config/avr/avr.c: Same.
28913 * config/avr/avr-devices.c: Same.
28914 * config/avr/driver-avr.c: Same.
28915 * config/avr/gen-avr-mmcu-texi.c: Same.
28916 * config/avr/avr-mcus.def: Adjust comment.
28918 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
28920 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
28921 * config/aarch64/iterators.md (VALLDI): New.
28923 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
28924 Andi Kleen <ak@linux.intel.com>
28927 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
28928 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
28931 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28933 * config/avr/avr-stdint.h: Remove trailing blanks.
28934 * config/avr/avr-log.h: Same.
28935 * config/avr/avr-arch.h: Same.
28936 * config/avr/avr-devices.c: Same.
28937 * config/avr/avr-dimode.md: Same.
28938 * config/avr/predicates.md: Same.
28939 * config/avr/avr-c.c: Same. And fix typo.
28941 * config/avr/avr-protos.h: Same. And:
28942 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
28943 (init_cumulative_args): Rename to avr_init_cumulative_args.
28944 (expand_prologue): Rename to avr_expand_prologue.
28945 (expand_epilogue): Rename to avr_expand_epilogue.
28946 (adjust_insn_length): Rename to avr_adjust_insn_length.
28947 (notice_update_cc): Rename to avr_notice_update_cc.
28948 (final_prescan_insn): Rename to avr_final_prescan_insn.
28949 * config/avr/avr.c: Same.
28950 * config/avr/avr.h: Same.
28951 * config/avr/avr.md: Remove trailing blanks.
28952 (prologue): Use avr_expand_prologue.
28953 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
28955 2013-01-14 Richard Biener <rguenther@suse.de>
28957 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
28958 verify_location, collect_subblocks): New functions.
28959 (verify_gimple_in_cfg): Verify that locations only reference
28960 BLOCKs in the functions BLOCK tree.
28962 2013-01-14 Richard Biener <rguenther@suse.de>
28964 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
28966 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
28968 (insert_out_of_ssa_copy_on_edge): Likewise.
28969 (rewrite_close_phi_out_of_ssa): Likewise.
28970 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
28972 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
28973 propagated constants.
28974 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
28977 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28979 * config/avr/avr-modes.def: Add GPL copyright notice.
28981 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
28983 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
28984 MEMMODEL_MASK to determine memory model.
28985 (atomic_store<mode>): Ditto from operands[2].
28986 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
28988 2013-01-13 Jakub Jelinek <jakub@redhat.com>
28991 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
28992 (fold_gimple_assign): Don't call unshare_expr here.
28993 (fold_ctor_reference): Call unshare_expr.
28995 2013-01-13 Terry Guo <terry.guo@arm.com>
28997 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
28998 * doc/fragments.texi: Document MULTILIB_REUSE.
28999 * gcc.c (multilib_reuse): New internal spec.
29000 (set_multilib_dir): Also search multilib from multilib_reuse.
29001 * genmultilib (tmpmultilib3): Refactor code.
29002 (tmpmultilib4): Ditto.
29003 (multilib_reuse): New multilib argument.
29005 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
29007 * Makefile.in: Update copyright.
29009 2013-01-12 Tom de Vries <tom@codesourcery.com>
29011 PR middle-end/55890
29012 * calls.c (expand_call): Check if arg_nr is valid.
29014 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
29016 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
29017 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
29018 documentation. Add missing '__' in front of
29019 __builtin_ia32_packssdw256.
29021 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29024 * config/s390/s390.c (s390_preferred_reload_class): Do not return
29025 NO_REGS for larl operands.
29026 (s390_reload_larl_operand): Use s390_load_address instead of
29029 2013-01-11 Richard Biener <rguenther@suse.de>
29031 * tree-cfg.c (verify_node_sharing_1): Split out from ...
29032 (verify_node_sharing): ... here.
29033 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
29035 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
29037 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
29038 Substitute TREECHECKING.
29039 * configure: Regenerate.
29040 * Makefile.in (TREECHECKING): New.
29042 2013-01-11 Richard Guenther <rguenther@suse.de>
29044 PR tree-optimization/44061
29045 * tree-vrp.c (extract_range_basic): Compute zero as
29046 value-range for __builtin_constant_p of function parameters.
29048 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
29050 Update copyright years.
29052 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
29054 PR rtl-optimization/55672
29055 * lra-eliminations.c (mark_not_eliminable): Permit addition with
29056 const to be eliminable.
29058 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
29060 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
29061 * configure: Regenerate.
29063 2013-01-10 Richard Biener <rguenther@suse.de>
29065 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
29067 2013-01-10 Richard Biener <rguenther@suse.de>
29070 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
29071 locations for virtual PHI arguments.
29072 (rewrite_update_phi_arguments): Likewise.
29074 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
29076 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
29079 2013-01-10 Jakub Jelinek <jakub@redhat.com>
29081 PR tree-optimization/55921
29082 * tree-complex.c (expand_complex_asm): New function.
29083 (expand_complex_operations_1): Call it for GIMPLE_ASM.
29085 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29088 * config/s390/s390.c (s390_symref_operand_p)
29089 (s390_loadrelative_operand_p): Merge the two functions.
29090 (s390_check_qrst_address, print_operand_address): Add parameters
29091 to s390_loadrelative_operand_p invokation.
29092 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
29093 (s390_reload_larl_operand, s390_secondary_reload): Use
29094 s390_loadrelative_operand_p instead of s390_symref_operand_p.
29095 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
29097 2013-01-09 Mike Stump <mikestump@comcast.net>
29099 * dse.c (record_store): Remove unnecessary assert.
29101 2013-01-09 Jan Hubicka <jh@suse.cz>
29103 PR tree-optimization/55569
29104 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
29105 * cfgloop.h (scale_loop_profile): Likewise.
29107 2013-01-09 Jan Hubicka <jh@suse.cz>
29110 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
29112 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
29114 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
29116 PR middle-end/55114
29117 * expr.h (maybe_emit_group_store): Declare.
29118 * expr.c (maybe_emit_group_store): New function.
29119 * builtins.c (expand_builtin_int_roundingfn): Call it.
29120 (expand_builtin_int_roundingfn_2): Likewise.
29122 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
29124 PR rtl-optimization/55829
29125 * lra-constraints.c (match_reload): Add code for absent output.
29126 (curr_insn_transform): Add code for reloads of matched inputs
29129 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
29131 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
29132 attribute of movddup insn to DF.
29133 (*vec_interleave_lowv2df): Ditto.
29134 (vec_dupv2df): Ditto.
29136 2013-01-09 Jan Hubicka <jh@suse.cz>
29138 PR tree-optimiation/55875
29139 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
29140 EVERY_ITERATION parameter.
29141 (number_of_iterations_exit): Check if exit is executed every iteration.
29142 (idx_infer_loop_bounds): Similarly here.
29143 (n_of_executions_at_most): Simplify
29144 to only test for cases where statement is dominated by the
29145 particular bound; handle correctly the "postdominance" test.
29146 (scev_probably_wraps_p): Use max loop iterations info
29147 as a global bound first.
29149 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
29150 Nick Clifton <nickc@redhat.com>
29152 * config/v850/v850.md (cbranchsf4): New pattern.
29153 (cstoresf4): New pattern.
29154 (cbranchdf4): New pattern.
29155 (cstoredf4): New pattern.
29156 (movsicc): Disallow floating point comparisons.
29157 (cmpsf_le_insn): Fix order of operators.
29158 (cmpsf_lt_insn): Likewise.
29159 (cmpsf_eq_insn): Likewise.
29160 (cmpdf_le_insn): Likewise.
29161 (cmpdf_lt_insn): Likewise.
29162 (cmpdf_eq_insn): Likewise.
29163 (cmpsf_ge_insn): Use LE comparison.
29164 (cmpdf_ge_insn): Likewise.
29165 (cmpsf_gt_insn): Use LT comparison.
29166 (cmpdf_gt_insn): Likewise.
29167 (cmpsf_ne_insn): Delete pattern.
29168 (cmpdf_ne_insn): Delete pattern.
29169 * config/v850/v850.c (v850_gen_float_compare): Use
29170 gen_cmpdf_eq_insn for NE comparison.
29171 (v850_float_z_comparison_operator)
29172 (v850_float_nz_comparison_operator): Move from here ...
29173 * config/v850/predicates.md: ... to here. Move GT and GE
29174 comparisons into v850_float_z_comparison_operator.
29175 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
29177 (v850_float_nz_comparison_operator): Likewise.
29179 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29181 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
29182 with calls to gen_insvsi/gen_insvdi.
29184 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29186 * config/i386/i386.c (initial_ix86_tune_features): Set up
29187 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
29189 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
29190 Jakub Jelinek <jakub@redhat.com>
29192 PR tree-optimization/48189
29193 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
29194 If nitercst is 0, don't predict the exit edge.
29196 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29198 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
29199 in asm_fprintf with reg_names.
29200 (aarch64_print_operand_address): Likewise.
29201 (aarch64_return_addr): Likewise.
29202 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
29204 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29206 * config/pa/pa.h (VAL_U6_BITS_P): Define.
29207 (INT_U6_BITS): Likewise.
29208 * config/pa/predicates.md (uint6_operand): New predicate.
29209 (shift5_operand, shift6_operand): Likewise.
29210 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
29212 (lshrdi3): Use shift6_operand.
29213 (shrpsi4, shrpdi4): New insn patterns.
29214 (extzv): Delete expander.
29215 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
29216 predicates in unamed zero extract patterns. Tighten common constraint.
29217 (extv): Delete expander.
29218 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
29219 predicates in unamed sign extract patterns. Tighten common constraint.
29220 (insv): Delete expander.
29221 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
29222 predicates in unamed insert patterns. Tighten common constraint.
29223 Change uint32_operand predicate to uint6_operand predicate in unamed
29224 DImode pattern to insert constant values of type 1...1xxxx.
29226 2013-01-04 Jan Hubicka <jh@suse.cz>
29228 PR tree-optimization/55823
29229 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
29232 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29233 Uros Bizjak <ubizjak@gmail.com>
29235 PR rtl-optimization/55845
29236 * df-problems.c (can_move_insns_across): Stop scanning at
29237 volatile_insn_p source instruction or give up if
29238 across_from .. across_to range contains any volatile_insn_p
29241 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29243 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
29244 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
29246 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
29247 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
29249 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29252 * asan.c (asan_clear_shadow): New function.
29253 (asan_emit_stack_protection): Use it.
29255 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29257 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
29258 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
29259 with tab instead of space.
29261 2013-01-08 Nick Clifton <nickc@redhat.com>
29263 * config/rl78/rl78.c (rl78_expand_prologue): Always select
29264 register bank 0 at the start of an interrupt handler.
29265 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
29268 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29270 * config/aarch64/aarch64-simd.md
29271 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
29272 (aarch64_simd_bsl): Likewise.
29273 (aarch64_vcond_internal<mode>): Likewise.
29274 (vcond<mode><mode>): Likewise.
29275 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
29276 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
29278 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29280 * config/aarch64/aarch64-builtins.c
29281 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
29283 2013-01-08 Martin Jambor <mjambor@suse.cz>
29286 * tree-sra.c (analyze_access_subtree): Return true also after
29287 potentially creating a debug-only replacement.
29289 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29291 PR middle-end/55890
29292 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
29294 PR tree-optimization/54120
29295 * tree-vrp.c (range_fits_type_p): Don't allow
29296 src_precision < precision from signed vr to unsigned_p
29297 if vr->min or vr->max is negative.
29298 (simplify_float_conversion_using_ranges): Test can_float_p
29299 against CODE_FOR_nothing.
29301 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29302 Richard Biener <rguenther@suse.de>
29304 PR middle-end/55851
29305 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
29306 types instead of just INTEGER_TYPE types.
29308 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
29310 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
29313 2013-01-07 Steve Ellcey <sellcey@mips.com>
29316 * config/mips/mips.opt: Change mad to mmad to match documentation.
29318 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29321 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
29322 .progmemx.data now.
29324 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29327 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
29328 (avr_addrspace_t): Add .section_name field.
29329 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
29331 (avr_addrspace): Same. Initialize .section_name. Remove last
29332 NULL entry. Put __memx into .progmemx.data.
29333 (progmem_section_prefix): Remove.
29334 (avr_asm_init_sections): No need to initialize progmem_section.
29335 (avr_asm_named_section): Use avr_addrspace[].section_name to get
29336 section name prefix.
29337 (avr_asm_select_section): Ditto. And use get_unnamed_section to
29338 retrieve the progmem section.
29339 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
29340 boundary to run over avr_addrspace[].
29341 (avr_register_target_pragmas): Ditto.
29343 2013-01-06 Jakub Jelinek <jakub@redhat.com>
29345 * varasm.c (output_constant_def_contents): For asan_protect_global
29346 protected strings, adjust DECL_ALIGN if needed, before testing for
29348 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
29349 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
29351 (output_object_block): For asan protected decls, emit asan padding
29352 after their contents.
29353 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
29354 (asan_finish_file): Test it here instead.
29356 2013-01-07 Nick Clifton <nickc@redhat.com>
29357 Matthias Klose <doko@debian.org>
29358 Doug Kwan <dougkwan@google.com>
29359 H.J. Lu <hongjiu.lu@intel.com>
29362 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
29364 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
29366 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
29368 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
29370 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
29372 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29375 * doc/install.texi (Cross-Compiler-Specific Options): Document
29378 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
29380 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
29381 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
29382 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
29383 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
29384 vqmovun_high_s64): Fix source operand number and update copyright.
29386 2013-01-07 Richard Biener <rguenther@suse.de>
29388 PR middle-end/55890
29389 * gimple.h (gimple_call_builtin_p): New overload.
29390 * gimple.c (validate_call): New function.
29391 (gimple_call_builtin_p): Likewise.
29392 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
29393 Use gimple_call_builtin_p.
29394 (find_func_clobbers): Likewise.
29395 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
29396 (strlen_optimize_stmt): Likewise.
29398 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29400 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
29401 (vld1q_dup_*): Likewise.
29402 (vld1_*): Likewise.
29403 (vld1q_*): Likewise.
29404 (vld1_lane_*): Likewise.
29405 (vld1q_lane_*): Likewise.
29407 2013-01-07 Richard Biener <rguenther@suse.de>
29409 * lto-streamer.h (LTO_minor_version): Bump to 2.
29411 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29413 * config/aarch64/aarch64-protos.h
29414 (aarch64_const_double_zero_rtx_p): Rename to...
29415 (aarch64_float_const_zero_rtx_p): ...this.
29416 (aarch64_float_const_representable_p): New.
29417 (aarch64_output_simd_mov_immediate): Likewise.
29418 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
29419 move immediate case.
29420 * config/aarch64/aarch64.c
29421 (aarch64_const_double_zero_rtx_p): Rename to...
29422 (aarch64_float_const_zero_rtx_p): ...this.
29423 (aarch64_print_operand): Allow printing of new constants.
29424 (aarch64_valid_floating_const): New.
29425 (aarch64_legitimate_constant_p): Check for valid floating-point
29427 (aarch64_simd_valid_immediate): Likewise.
29428 (aarch64_vect_float_const_representable_p): New.
29429 (aarch64_float_const_representable_p): Likewise.
29430 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
29431 (aarch64_output_simd_mov_immediate): New.
29432 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
29433 (*movdf_aarch64): Likewise.
29434 * config/aarch64/constraints.md (Ufc): New.
29435 (Y): call aarch64_float_const_zero_rtx.
29436 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
29438 2013-01-07 Richard Biener <rguenther@suse.de>
29440 PR tree-optimization/55888
29441 PR tree-optimization/55862
29442 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
29443 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
29444 not if it is contained therein.
29446 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29448 * config/avr/t-avr: Typo.
29450 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29453 * config/avr/t-avr: Don't automatically rebuild
29454 $(srcdir)/config/avr/t-multilib
29455 $(srcdir)/config/avr/avr-tables.opt
29456 $(srcdir)/doc/avr-mmcu.texi
29457 (avr-mcus): New phony target to build them on request.
29458 (s-avr-mlib, s-avr-mmcu-texi): Remove.
29459 * avr/avr-mcus.def: Adjust comments.
29461 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
29463 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
29465 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29467 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
29469 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29471 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
29473 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
29475 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
29476 to generate profiling.
29477 * config/rs6000/aix64.h (LIB_SPEC): Same.
29479 2013-01-04 Andrew Pinski <apinski@cavium.com>
29481 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
29483 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
29485 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
29487 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
29489 (ix86_expand_move): Ditto.
29490 (ix86_zero_extend_to_Pmode): Ditto.
29491 (ix86_expand_call): Ditto.
29492 (ix86_expand_special_args_builtin): Ditto.
29493 (ix86_expand_builtin): Ditto.
29495 2013-01-04 Richard Biener <rguenther@suse.de>
29497 PR tree-optimization/55862
29498 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
29499 translating them through PHI nodes.
29501 2013-01-04 Martin Jambor <mjambor@suse.cz>
29503 PR tree-optimization/55755
29504 * tree-sra.c (sra_modify_assign): Do not check that an access has no
29505 children when trying to avoid producing a VIEW_CONVERT_EXPR.
29507 2013-01-04 Marek Polacek <polacek@redhat.com>
29509 PR middle-end/55859
29510 * opts.c (default_options_optimization): Clarify error message.
29512 2013-01-04 Richard Biener <rguenther@suse.de>
29514 PR middle-end/55863
29515 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
29518 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29521 * config/pa/pa.md (movsi): Revert previous change.
29522 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
29525 2013-01-03 Richard Henderson <rth@redhat.com>
29527 * config/i386/i386.c (ix86_expand_move): Always assign to op1
29528 after eliminating TLS symbols.
29530 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
29533 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
29534 * graphite-poly.c (debug_gmp_value): Likewise.
29536 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
29539 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
29540 selected code model, define __code_mode_small__, __code_model_medium__,
29541 __code_model_large__, __code_model_32__ or __code_model_kernel__.
29542 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
29543 xchg temporary register with %k. Declare temporary register as
29545 [__x86_64__]: For medium and large code models, preserve %rbx register.
29547 2013-01-03 Richard Biener <rguenther@suse.de>
29549 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
29550 (dump_subscript): Adjust.
29551 (finalize_ddr_dependent): Do not dump redundant info.
29552 (analyze_siv_subscript): Adjust.
29553 (subscript_dependence_tester): Likewise.
29554 (compute_affine_dependence): Likewise.
29556 2013-01-03 Richard Biener <rguenther@suse.de>
29559 2013-01-03 Richard Biener <rguenther@suse.de>
29561 PR tree-optimization/55857
29562 * tree-vect-stmts.c (vectorizable_load): Do not setup
29563 re-alignment for invariant loads.
29565 2013-01-02 Richard Biener <rguenther@suse.de>
29567 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29568 invariant load do not generate a vector load from the scalar location.
29570 2013-01-03 Richard Biener <rguenther@suse.de>
29572 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
29573 for not vectorizing.
29574 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
29575 not build INDIRECT_REFs, call get_name once only.
29576 (vect_create_data_ref_ptr): Likewise. Dump base object kind
29577 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
29579 2013-01-03 Richard Biener <rguenther@suse.de>
29581 PR tree-optimization/55857
29582 * tree-vect-stmts.c (vectorizable_load): Do not setup
29583 re-alignment for invariant loads.
29585 2013-01-03 Richard Biener <rguenther@suse.de>
29588 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
29589 prefer a built-in decl.
29591 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29593 * gcc.c (process_command): Update copyright notice dates.
29594 * gcov.c (print_version): Likewise.
29595 * gcov-dump.c (print_version): Likewise.
29597 PR rtl-optimization/55838
29598 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
29599 iv0.step, iv1.step and step.
29601 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29602 Marc Glisse <marc.glisse@inria.fr>
29604 PR tree-optimization/55832
29605 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
29606 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
29607 integer_{one,zero}_node.
29609 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29612 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
29613 * var-tracking.c (reverse_op): Don't add reverse ops to
29614 VALUEs that have already
29615 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
29617 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
29619 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
29621 2013-01-02 Teresa Johnson <tejohnson@google.com>
29623 * dumpfile.c (dump_loc): Print filename with location.
29624 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
29625 new location_t parameter to emit complete unroll message with
29626 new dump framework.
29627 (canonicalize_loop_induction_variables): Compute loops location
29628 and pass to try_unroll_loop_completely.
29629 * loop-unroll.c (report_unroll_peel): New function.
29630 (peel_loops_completely): Use new dump format with location
29631 for main dumpfile message, and invoke report_unroll_peel on success.
29632 (decide_unrolling_and_peeling): Ditto.
29633 (decide_peel_once_rolling): Remove old dumpfile message subsumed
29634 by report_unroll_peel.
29635 (decide_peel_completely): Ditto.
29636 (decide_unroll_constant_iterations): Ditto.
29637 (decide_unroll_runtime_iterations): Ditto.
29638 (decide_peel_simple): Ditto.
29639 (decide_unroll_stupid): Ditto.
29640 * cfgloop.c (get_loop_location): New function.
29641 * cfgloop.h (get_loop_location): Declare.
29643 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29645 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
29648 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29650 PR middle-end/55198
29651 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
29652 BLKmode objects when EXPAND_MEMORY is specified.
29654 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29656 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
29658 (fold_builtin_cpu): Do not share cpu model decls across statements.
29660 2013-01-02 Jason Merrill <jason@redhat.com>
29663 * tree.c (build_array_type_1): Revert earlier change.
29665 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
29667 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
29669 * config/aarch64/aarch64-tune.md: Re-generate.
29671 2013-01-02 Richard Biener <rguenther@suse.de>
29673 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29674 invariant load do not generate a vector load from the scalar location.
29676 2013-01-02 Richard Biener <rguenther@suse.de>
29679 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
29680 * configure: Regenerate.
29682 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29684 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
29685 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
29686 (expand_builtin_int_roundingfn_2): Keep the original target around
29687 for the fallback case.
29689 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29691 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
29692 to be clear for sign changes.
29694 2013-01-01 Jan Hubicka <jh@suse.cz>
29696 * ipa-inline-analysis.c: Fix formatting.
29698 2013-01-01 Jakub Jelinek <jakub@redhat.com>
29700 PR tree-optimization/55831
29701 * tree-vect-loop.c (get_initial_def_for_induction): Use
29702 gsi_after_labels instead of gsi_start_bb.
29704 Copyright (C) 2013 Free Software Foundation, Inc.
29706 Copying and distribution of this file, with or without modification,
29707 are permitted in any medium without royalty provided the copyright
29708 notice and this notice are preserved.