ce16f50cf2ff7803a4a0d2fddf9969d4b527486c
[gcc.git] / gcc / ChangeLog
1 2013-11-05 Richard Biener <rguenther@suse.de>
2
3 PR tree-optimization/58955
4 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
5 edge direction.
6
7 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8
9 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
10 little endian.
11 (vec_pack_ufix_trunc_v2df): Likewise.
12
13 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
14
15 PR middle-end/58981
16 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
17 pattern, instead of word_mode.
18
19 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
20 Pmode for size.
21 (set_storage_via_setmem): Likewise.
22
23 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
24
25 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
26 gimple.h and the rest of the condition in eliminate_build.
27 (eliminate_build): Call new routine.
28 * gimple.h (phi_ssa_name_p): Delete.
29
30 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
31
32 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
33 case of no prefix and reserving zero slots, because when that's the
34 case we'll never get here.
35 * vec.h (va_heap::reserve): Don't try and handle
36 vec_prefix::calculate_allocation returning zero because that should
37 never happen.
38
39 2013-11-05 Richard Biener <rguenther@suse.de>
40
41 PR middle-end/58941
42 * tree-dfa.c (get_ref_base_and_extent): Merge common code
43 in MEM_REF and TARGET_MEM_REF handling. Make sure to
44 process trailing array detection before diving into the
45 view-converted object (and possibly apply some extra offset).
46
47 2013-11-05 Joseph Myers <joseph@codesourcery.com>
48
49 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
50 New function.
51 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
52
53 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
54
55 PR tree-optimization/58958
56 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
57 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
58
59 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
60
61 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
62 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
63
64 2013-11-05 Jakub Jelinek <jakub@redhat.com>
65
66 PR tree-optimization/58984
67 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
68 set *SIZE_P if non-NULL on success.
69 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
70 callers.
71 (ipcp_transform_function): Likewise. Punt if size of access
72 is different from TYPE_SIZE on v->value's type.
73
74 2013-11-05 Tobias Burnus <burnus@net-b.de>
75
76 * doc/invoke.texi (-fopenmp-simd): Document new option.
77 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
78 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
79 * tree.c (attribute_value_equal): Ditto.
80
81 2013-11-04 Wei Mi <wmi@google.com>
82
83 * sched-rgn.c (add_branch_dependences): Keep insns in
84 a SCHED_GROUP at the end of BB to remain their location.
85
86 2013-11-04 Wei Mi <wmi@google.com>
87
88 * gcc/config/i386/i386.c (memory_address_length): Extract a part
89 of code to rip_relative_addr_p.
90 (rip_relative_addr_p): New Function.
91 (ix86_macro_fusion_p): Ditto.
92 (ix86_macro_fusion_pair_p): Ditto.
93 * gcc/config/i386/i386.h: Add new tune features about macro-fusion.
94 * gcc/config/i386/x86-tune.def (DEF_TUNE): Ditto.
95 * gcc/doc/tm.texi: Generated.
96 * gcc/doc/tm.texi.in: Ditto.
97 * gcc/haifa-sched.c (try_group_insn): New Function.
98 (group_insns_for_macro_fusion): Ditto.
99 (sched_init): Call group_insns_for_macro_fusion.
100 * gcc/target.def: Add two hooks: macro_fusion_p and
101 macro_fusion_pair_p.
102
103 2013-11-04 Kostya Serebryany <kcc@google.com>
104
105 Update to match the changed asan API.
106 * asan.c (asan_function_start): New function.
107 (asan_emit_stack_protection): Update the string stored in the
108 stack red zone to match new API. Store the PC of the current
109 function in the red zone.
110 (asan_global_struct): Update the __asan_global definition to match
111 the new API.
112 (asan_add_global): Ditto.
113 * asan.h (asan_function_start): New prototype.
114 * final.c (final_start_function): Call asan_function_start.
115 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1 to __asan_init_v3.
116
117 2013-11-04 Wei Mi <wmi@google.com>
118
119 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): Separate
120 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
121 * gcc/config/i386/i386.c (ix86_option_override_internal): Ditto.
122 (ix86_issue_rate): Ditto.
123 (ix86_adjust_cost): Ditto.
124 (ia32_multipass_dfa_lookahead): Ditto.
125 (ix86_sched_init_global): Ditto.
126 (get_builtin_code_for_version): Ditto.
127 * gcc/config/i386/i386.h (enum target_cpu_default): Ditto.
128 (enum processor_type): Ditto.
129 * gcc/config/i386/x86-tune.def (DEF_TUNE): Ditto.
130
131 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
132
133 PR rtl-optimization/58967
134 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
135 !lra_in_progress for mode sizes bigger word.
136
137 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
138
139 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
140 arguments to merge instruction for little endian.
141 (vec_widen_umult_lo_v16qi): Likewise.
142 (vec_widen_smult_hi_v16qi): Likewise.
143 (vec_widen_smult_lo_v16qi): Likewise.
144 (vec_widen_umult_hi_v8hi): Likewise.
145 (vec_widen_umult_lo_v8hi): Likewise.
146 (vec_widen_smult_hi_v8hi): Likewise.
147 (vec_widen_smult_lo_v8hi): Likewise.
148
149 2013-11-04 Ian Lance Taylor <iant@google.com>
150
151 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
152 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
153 builtins that take pointers.
154 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions
155 if set.
156 * lto-wrapper.c (merge_and_complain): Collect
157 OPT_fnon_call_exceptions.
158 (run_gcc): Pass -fnon-call-exceptions.
159
160 2013-11-04 Jakub Jelinek <jakub@redhat.com>
161
162 * optabs.c (expand_vec_perm): Revert one incorrect line from
163 2013-10-31 change.
164
165 PR tree-optimization/58978
166 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
167 use_stmt by single_imm_use directly. Only call single_imm_use
168 on SSA_NAMEs.
169
170 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
171
172 PR rtl-optimization/58968
173 * lra-spills.c (return_regno_p): New function.
174 (lra_final_code_change): Use it.
175
176 2013-11-04 Joseph Myers <joseph@codesourcery.com>
177
178 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document
179 macros.
180 * target.def (float_exceptions_rounding_supported_p): New hook.
181 * targhooks.c (default_float_exceptions_rounding_supported_p): New
182 function.
183 * targhooks.h (default_float_exceptions_rounding_supported_p):
184 Declare.
185 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
186 New @hook.
187 * doc/tm.texi: Regenerate.
188 * config.gcc (powerpc*-*-linux*): Set extra_objs.
189 * config/rs6000/rs6000-linux.c: New file.
190 * config/rs6000/rs6000-protos.h
191 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
192 * config/rs6000/linux.h
193 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
194 * config/rs6000/linux64.h
195 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
196 * config/rs6000/t-linux (rs6000-linux.o): New rule.
197 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
198
199 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
200
201 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
202 Replace the define_insn_and_split with a define_insn and two
203 define_splits, with the split after reload re-permuting the source
204 register to its original value.
205 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
206 (*vsx_le_perm_store_v8hi): Likewise.
207 (*vsx_le_perm_store_v16qi): Likewise.
208
209 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
210
211 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
212 little endian.
213
214 2013-11-04 Jakub Jelinek <jakub@redhat.com>
215
216 PR tree-optimization/58946
217 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
218 bbs with bbinfo[idx].op != NULL before all blocks with
219 bbinfo[idx].op == NULL.
220
221 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
222
223 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
224 (avr_dump_double_int_hex): Likewise.
225 (avr_log_vadump): Remove %D and %X handling.
226 * config/avr/avr.c (avr_double_int_push_digit): Delete.
227 (avr_map_op_t): Change map from double_int to unsigned int.
228 (avr_map_op): Update accordingly.
229 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
230 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
231 unsigned ints rather than double_ints.
232
233 2013-11-03 Marek Polacek <polacek@redhat.com>
234
235 Implement -fsanitize=vla-bound.
236 * opts.c (common_handle_option): Handle vla-bound.
237 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE):
238 Define.
239 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
240 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
241
242 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
243
244 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
245 little endian.
246
247 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
248
249 * config/i386/constraints.md (Ts, Tv): New address constrains.
250 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
251 constraint for address_no_seg_operand.
252 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
253 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
254 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
255 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
256 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
257 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
258 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
259 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
260 constraint for vsib_address_operand.
261
262 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
263
264 * gcse.c (pre_delete): Remove references to regmove from comments.
265 * recog.c: (validate_replace_rtx_1): Likewise.
266 * config/rl78/rl78.c: Likewise.
267 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
268 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
269 * common/config/mmix/mmix-common.c: Likewise.
270
271 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
272
273 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
274 * gimplify.c (gimplify_compound_lval): Likewise.
275 * graphite-clast-to-gimple.c (gloog): Likewise.
276 * graphite-dependences.c (loop_is_parallel_p): Likewise.
277 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
278 (limit_scop); Likewise.
279 (build_scops): Likewise.
280 (dot_scop): Likewise.
281 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
282 (build_scop_drs): Likewise.
283 (insert_stmts): Likewise.
284 (insert_out_of_ssa_copy): Likewise.
285 (remove_phi): Likewise.
286 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
287 * hw-doloop.c (discover_loop): Likewise.
288 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
289 * tree-dfa.c (dump_enumerated_decls): Likewise.
290 * tree-if-conv.c (if_convertable_loop_p): Likewise.
291 * tree-inline.c (tree_function_versioning): Likewise.
292 * tree-loop-distribution.c (build_rdg): Likewise.
293 (rdg_flag_vertex_and_dependent): Likewise.
294 (distribute_loop): Likewise.
295 * tree-parloops.c (loop_parallel_p): Likewise.
296 (eliminate_local_variables): Likewise.
297 (separate_decls_in_region): Likewise.
298 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
299 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
300 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
301 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
302 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
303 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
304 (vectorizable_condition): Likewise.
305
306 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
307
308 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
309 * configure: Regenerate.
310 * config/i386/i386.md (*movdi_internal): Change
311 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
312 (*movdf_internal): Ditto.
313 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
314 * config/i386/sse.md (vec_concatv2di): Output interunit movq
315 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
316
317 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
318
319 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
320 after comment.
321
322 2013-10-31 David Malcolm <dmalcolm@redhat.com>
323
324 Automated part of renaming of symtab_node_base to symtab_node.
325
326 Patch autogenerated by rename_symtab.py from
327 https://github.com/davidmalcolm/gcc-refactoring-scripts
328 revision 58bb219cc090b2f4516a9297d868c245495ee622
329 with ChangeLog entry fixed up by hand.
330
331 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
332 symtab_node.
333 (cgraph_node_for_asm): Likewise.
334 * cgraph.h (symtab_node_base): Likewise.
335 (cgraph_node): Likewise.
336 (varpool_node): Likewise.
337 (is_a_helper <cgraph_node>::test): Likewise.
338 (is_a_helper <varpool_node>::test): Likewise.
339 (symtab_nodes): Likewise.
340 (symtab_register_node): Likewise.
341 (symtab_unregister_node): Likewise.
342 (symtab_remove_node): Likewise.
343 (symtab_get_node): Likewise.
344 (symtab_node_for_asm): Likewise.
345 (symtab_node_asm_name): Likewise.
346 (symtab_node_name): Likewise.
347 (symtab_insert_node_to_hashtable): Likewise.
348 (symtab_add_to_same_comdat_group): Likewise.
349 (symtab_dissolve_same_comdat_group_list): Likewise.
350 (dump_symtab_node): Likewise.
351 (debug_symtab_node): Likewise.
352 (dump_symtab_base): Likewise.
353 (verify_symtab_node): Likewise.
354 (verify_symtab_base): Likewise.
355 (symtab_used_from_object_file_p): Likewise.
356 (symtab_alias_ultimate_target): Likewise.
357 (symtab_resolve_alias): Likewise.
358 (fixup_same_cpp_alias_visibility): Likewise.
359 (symtab_for_node_and_aliases): Likewise.
360 (symtab_nonoverwritable_alias): Likewise.
361 (availability symtab_node_availability): Likewise.
362 (symtab_semantically_equivalent_p): Likewise.
363 (fixup_same_cpp_alias_visibility): Likewise.
364 (symtab_prevail_in_asm_name_hash): Likewise.
365 (cgraph): Likewise.
366 (varpool): Likewise.
367 (varpool_first_variable): Likewise.
368 (varpool_next_variable): Likewise.
369 (varpool_first_static_initializer): Likewise.
370 (varpool_next_static_initializer): Likewise.
371 (varpool_first_defined_variable): Likewise.
372 (varpool_next_defined_variable): Likewise.
373 (cgraph_first_defined_function): Likewise.
374 (cgraph_next_defined_function): Likewise.
375 (cgraph_first_function): Likewise.
376 (cgraph_next_function): Likewise.
377 (cgraph_first_function_with_gimple_body): Likewise.
378 (cgraph_next_function_with_gimple_body): Likewise.
379 (symtab_alias_target): Likewise.
380 (symtab_real_symbol_p): Likewise.
381 (symtab_can_be_discarded): Likewise.
382 * cgraphbuild.c (mark_address): Likewise.
383 (mark_load): Likewise.
384 (mark_store): Likewise.
385 * cgraphunit.c (decide_is_symbol_needed): Likewise.
386 (first): Likewise.
387 (enqueue_node): Likewise.
388 (referred_to_p): Likewise.
389 (cgraph_process_same_body_aliases): Likewise.
390 (analyze_functions): Likewise.
391 (handle_alias_pairs): Likewise.
392 (output_weakrefs): Likewise.
393 (compile): Likewise.
394 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
395 * ipa-inline-analysis.c (inline_write_summary): Likewise.
396 * ipa-prop.c (remove_described_reference): Likewise.
397 (try_decrement_rdesc_refcount): Likewise.
398 (ipa_edge_duplication_hook): Likewise.
399 * ipa-ref.c (ipa_record_reference): Likewise.
400 (ipa_maybe_record_reference): Likewise.
401 (ipa_clone_ref): Likewise.
402 (ipa_clone_references): Likewise.
403 (ipa_clone_referring): Likewise.
404 (ipa_find_reference): Likewise.
405 (ipa_remove_stmt_references): Likewise.
406 (ipa_clear_stmts_in_references): Likewise.
407 * ipa-ref.h (symtab_node_base): Likewise.
408 (ipa_ref): Likewise.
409 (ipa_record_reference): Likewise.
410 (ipa_maybe_record_reference): Likewise.
411 (ipa_clone_references): Likewise.
412 (ipa_clone_referring): Likewise.
413 (ipa_clone_ref): Likewise.
414 (ipa_find_reference): Likewise.
415 (ipa_remove_stmt_references): Likewise.
416 (ipa_clear_stmts_in_references): Likewise.
417 * ipa-reference.c (ipa_reference_write_optimization_summary):
418 Likewise.
419 * ipa.c (enqueue_node): Likewise.
420 (process_references): Likewise.
421 (walk_polymorphic_call_targets): Likewise.
422 (symtab_remove_unreachable_nodes): Likewise.
423 (address_taken_from_non_vtable_p): Likewise.
424 (comdat_can_be_unshared_p_1): Likewise.
425 (comdat_can_be_unshared_p): Likewise.
426 (can_replace_by_local_alias): Likewise.
427 (function_and_variable_visibility): Likewise.
428 * is-a.h: Likewise (within example in comment).
429 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
430 (lto_symtab_encoder_encode): Likewise.
431 (lto_symtab_encoder_delete_node): Likewise.
432 (lto_symtab_encoder_in_partition_p): Likewise.
433 (lto_set_symtab_encoder_in_partition): Likewise.
434 (output_refs): Likewise.
435 (compute_ltrans_boundary): Likewise.
436 (output_symtab): Likewise.
437 (input_node): Likewise.
438 (input_ref): Likewise.
439 (input_edge): Likewise.
440 (input_cgraph_1): Likewise.
441 (input_refs): Likewise.
442 (output_cgraph_opt_summary): Likewise.
443 (input_node_opt_summary): Likewise.
444 (input_cgraph_opt_section): Likewise.
445 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
446 Likewise.
447 * lto-streamer-out.c (lto_output): Likewise.
448 (output_symbol_p): Likewise.
449 (produce_symtab): Likewise.
450 * lto-streamer.h (lto_encoder_entry): Likewise.
451 (lto_free_function_in_decl_state_for_node): Likewise.
452 (lto_symtab_encoder_encode): Likewise.
453 (lto_symtab_encoder_delete_node): Likewise.
454 (lto_symtab_encoder_in_partition_p): Likewise.
455 (lto_set_symtab_encoder_in_partition): Likewise.
456 (lto_symtab_encoder_lookup): Likewise.
457 (lsei_node): Likewise.
458 (lto_symtab_encoder_deref): Likewise.
459 * symtab.c (symtab_hash): Likewise.
460 (assembler_name_hash): Likewise.
461 (symtab_nodes): Likewise.
462 (hash_node): Likewise.
463 (eq_node): Likewise.
464 (hash_node_by_assembler_name): Likewise.
465 (eq_assembler_name): Likewise.
466 (insert_to_assembler_name_hash): Likewise.
467 (unlink_from_assembler_name_hash): Likewise.
468 (symtab_prevail_in_asm_name_hash): Likewise.
469 (symtab_register_node): Likewise.
470 (symtab_insert_node_to_hashtable): Likewise.
471 (symtab_unregister_node): Likewise.
472 (symtab_get_node): Likewise.
473 (symtab_remove_node): Likewise.
474 (symtab_initialize_asm_name_hash): Likewise.
475 (symtab_node_for_asm): Likewise.
476 (symtab_add_to_same_comdat_group): Likewise.
477 (symtab_dissolve_same_comdat_group_list): Likewise.
478 (symtab_node_asm_name): Likewise.
479 (symtab_node_name): Likewise.
480 (dump_symtab_base): Likewise.
481 (dump_symtab_node): Likewise.
482 (dump_symtab): Likewise.
483 (debug_symtab_node): Likewise.
484 (verify_symtab_base): Likewise.
485 (verify_symtab_node): Likewise.
486 (verify_symtab): Likewise.
487 (symtab_used_from_object_file_p): Likewise.
488 (symtab_node_availability): Likewise.
489 (symtab_alias_ultimate_target): Likewise.
490 (fixup_same_cpp_alias_visibility): Likewise.
491 (symtab_resolve_alias): Likewise.
492 (symtab_for_node_and_aliases): Likewise.
493 (symtab_for_node_and_aliases): Likewise.
494 (symtab_nonoverwritable_alias_1): Likewise.
495 (symtab_nonoverwritable_alias): Likewise.
496 (symtab_semantically_equivalent_p): Likewise.
497 * value-prof.c (init_node_map): Likewise.
498 * varasm.c (find_decl): Likewise.
499 * varpool.c (varpool_node_for_asm): Likewise.
500 (varpool_remove_unreferenced_decls): Likewise.
501
502 2013-10-31 David Malcolm <dmalcolm@redhat.com>
503
504 Manual part of renaming of symtab_node_base to symtab_node.
505
506 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
507 clashes with the preferred name for the base class.
508 (const_symtab_node): Remove redundant typedef.
509
510 2013-10-31 Jakub Jelinek <jakub@redhat.com>
511
512 * optabs.c (expand_vec_perm): Avoid vector mode punning
513 SUBREGs in SET_DEST.
514 * expmed.c (store_bit_field_1): Likewise.
515 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
516 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
517 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
518 * config/i386/i386.c (ix86_expand_vector_move_misalign,
519 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
520 ix86_expand_sse_unpack, ix86_expand_args_builtin,
521 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
522 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
523 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
524 expand_vec_perm_vpshufb2_vpermq,
525 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
526 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
527 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
528 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
529 than *d itself.
530
531 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
532
533 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
534 Use gcc_unreachable for unhandled modes. Do not check results of
535 expand_simple_binop. If not expanded to target, move the result.
536
537 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
538 Shiva Chen <shiva0217@gmail.com>
539
540 * config.gcc (nds32*-*-*): Add nds32 target.
541 * config/nds32/nds32.c: New file.
542 * config/nds32/nds32.h: New file.
543 * config/nds32/nds32.md: New file.
544 * config/nds32/constants.md: New file.
545 * config/nds32/constraints.md: New file.
546 * config/nds32/iterators.md: New file.
547 * config/nds32/nds32-doubleword.md: New file.
548 * config/nds32/nds32-intrinsic.md: New file.
549 * config/nds32/nds32_intrinsic.h: New file.
550 * config/nds32/nds32-modes.def: New file.
551 * config/nds32/nds32-multiple.md: New file.
552 * config/nds32/nds32.opt: New file.
553 * config/nds32/nds32-opts.h: New file.
554 * config/nds32/nds32-protos.h: New file.
555 * config/nds32/nds32-peephole2.md: New file.
556 * config/nds32/pipelines.md: New file.
557 * config/nds32/predicates.md: New file.
558 * config/nds32/t-mlibs: New file.
559 * common/config/nds32: New directory and files.
560
561 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
562 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
563 * doc/install.texi (Cross-Compiler-Specific Options): Document
564 --with-nds32-lib for nds32 target.
565 * doc/extend.texi (Function Attributes, Target Builtins): Document
566 nds32 specific attributes.
567
568 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
569
570 * lra-constraints (process_alt_operands): Use the result
571 elimination register for operand when matching constraints.
572
573 2013-10-31 Jakub Jelinek <jakub@redhat.com>
574
575 * tree-vrp.c (maybe_set_nonzero_bits): New function.
576 (remove_range_assertions): Call it.
577
578 * tree.c (tree_ctz): New function.
579 * tree.h (tree_ctz): New prototype.
580 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
581 first argument from tree to const_tree.
582 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
583 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
584 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
585 static.
586 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
587 * tree-vect-loop.c (vect_analyze_loop_operations,
588 vect_transform_loop): Don't force scalar loop for bound just because
589 number of iterations is unknown, only do it if it is not known to be
590 a multiple of vectorization_factor.
591 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
592
593 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
594 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
595 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
596 (set_nonzero_bits, get_nonzero_bits): New prototypes.
597 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
598 a default def isn't partially constant.
599 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
600 is known to be partially zero.
601 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
602 to see if a default def isn't partially constant.
603 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
604 creation of a range, if VR_RANGE, try to improve nonzero_bits from
605 the range.
606 (set_nonzero_bits, get_nonzero_bits): New functions.
607
608 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
609 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
610 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
611 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
612 uses but in the condition and ASSERT_EXPR and the other successor of
613 the predecessor bb is __builtin_unreachable (), set_range_info of the
614 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
615
616 2013-10-31 Martin Jambor <mjambor@suse.cz>
617
618 PR rtl-optimization/58934
619 Revert:
620 2013-10-30 Martin Jambor <mjambor@suse.cz>
621 PR rtl-optimization/10474
622 * ira.c (find_moveable_pseudos): Do not calculate dominance info
623 nor df analysis.
624 (interesting_dest_for_shprep): New function.
625 (split_live_ranges_for_shrink_wrap): Likewise.
626 (ira): Calculate dominance info and df analysis. Call
627 split_live_ranges_for_shrink_wrap.
628
629 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
630 Yury Gribov <y.gribov@samsung.com>
631
632 PR sanitizer/58543
633 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
634 shadow pointer to avoid clobbering the main one.
635
636 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
637
638 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
639
640 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
641
642 PR bootstrap/58933
643 * ira-color.c (update_costs_from_copies): Add new parameter. Use
644 it for calling update_costs_from_allocno.
645 (assign_hard_reg): Call restore_costs_from_copies only for
646 !retry_p. Pass new argument to update_costs_from_copies.
647 (color_pass): Pass new argument to update_costs_from_copies.
648 (ira_mark_allocation_change): Ditto.
649
650 2013-10-30 Sharad Singhai <singhai@google.com>
651
652 PR middle-end/58134
653 * opts.c (common_handle_option): Remove deprecated option
654 -ftree-vectorizer-verbose.
655 * doc/invoke.texi (Debugging Options): Ditto.
656 * opts-global.c (handle_common_deferred_options): Ditto.
657 (dump_remap_tree_vectorizer_verbose): Delete.
658 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
659
660 2013-10-30 DJ Delorie <dj@redhat.com>
661
662 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
663 that take no arguments.
664
665 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
666
667 PR other/58545
668 * reload1.c (update_eliminables_and_spill): New function, broken
669 out of reload.
670 (reload): Use it. Check for frame size change after frame size
671 alignment, and call update_eliminables_and_spill first if continue-ing.
672
673 2013-10-30 Cong Hou <congh@google.com>
674
675 PR target/58762
676 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
677 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
678 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
679
680 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
681
682 PR rtl-optimization/58369
683 * reload1.c (compute_reload_subreg_offset): New function.
684 (choose_reload_regs): Use it to pass endian-correct
685 offset to subreg_regno_offset.
686
687 2013-10-30 Tobias Burnus <burnus@net-b.de>
688
689 PR other/33426
690 * gcc/tree-cfg.c (replace_loop_annotate): Replace warning by
691 warning_at.
692
693 2013-10-30 Jason Merrill <jason@redhat.com>
694
695 * configure.ac (loose_warn): Add -Wno-format if
696 --disable-build-format-warnings.
697
698 2013-10-30 David Malcolm <dmalcolm@redhat.com>
699
700 * cgraphunit.c (analyze_functions): Split symtab_node declarations
701 onto multiple lines to make things easier for rename_symtab.py.
702
703 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
704 (symtab_semantically_equivalent_p): Likewise.
705
706 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
707
708 PR target/58784
709 * lra.c (check_rtl): Remove address check before LRA work.
710
711 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
712
713 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
714 POINTER_PLUS_EXPR in the defining statement.
715
716 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
717
718 * regmove.c: Remove.
719 * tree-pass.h (make_pass_regmove): Remove.
720 * timevar.def (TV_REGMOVE): Remove.
721 * passes.def (pass_regmove): Remove.
722 * opts.c (default_options_table): Remove entry for regmove.
723 * doc/passes.texi: Remove regmove pass description.
724 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
725 options.
726 (-fdump-rtl-regmove): Ditto.
727 * common.opt (foptimize-register-move, fregmove): Ignore.
728 * Makefile.in (OBJS): Remove regmove.o.
729 * regmove.c: Remove.
730 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
731 and type.
732 (struct ira_allocno) New member allocno_prefs.
733 (ALLOCNO_PREFS): New macro.
734 (ira_prefs, ira_prefs_num): New external vars.
735 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
736 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
737 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
738 Ditto.
739 (ira_add_allocno_copy_to_list): Remove prototype.
740 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
741 (ira_pref_iterator): New type.
742 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
743 (FOR_EACH_PREF): New macro.
744 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
745 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
746 code.
747 (ira_setup_alts): New function.
748 (decrease_live_ranges_number): New function.
749 (ira): Call the above function.
750 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
751 (ira_create_allocno): Initialize allocno prefs.
752 (pref_pool, pref_vec): New static vars.
753 (initiate_prefs, find_allocno_pref, ira_create_pref): New
754 functions.
755 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
756 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
757 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
758 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
759 (ira_add_allocno_copy_to_list): Make static. Rename to
760 add_allocno_copy_to_list.
761 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
762 swap_allocno_copy_ends_if_necessary.
763 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
764 ira_remove_allocno_prefs.
765 (ira_flattening): Ditto.
766 (ira_build): Call initiate_prefs, print_prefs.
767 (ira_destroy): Call finish_prefs.
768 * ira-color.c (struct update_cost_record): New.
769 (struct allocno_color_data): Add new member update_cost_records.
770 (update_cost_record_pool): New static var.
771 (init_update_cost_records, get_update_cost_record): New functions.
772 (free_update_cost_record_list, finish_update_cost_records): Ditto.
773 (struct update_cost_queue_elem): Add member from.
774 (initiate_cost_update): Call init_update_cost_records.
775 (finish_cost_update): Call finish_update_cost_records.
776 (queue_update_cost, get_next_update_cost): Add new param from.
777 (Update_allocno_cost, update_costs_from_allocno): New functions.
778 (update_costs_from_prefs): Ditto.
779 (update_copy_costs): Rename to update_costs_from_copies.
780 (restore_costs_from_copies): New function.
781 (update_conflict_hard_regno_costs): Don't go back.
782 (assign_hard_reg): Call restore_costs_from_copies. Add printing
783 more debug info.
784 (pop_allocnos): Add priniting more debug info.
785 (color_allocnos): Remove prefs for conflicting hard regs.
786 Call update_costs_from_prefs.
787 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
788 (get_dup_num): Rename, modify, and move to ira.c
789 (process_regs_for_copy): Add prefs.
790 (add_insn_allocno_copies): Put src as first arg of
791 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
792 * ira-costs.c (record_reg_classes): Modify and move code into
793 record_operands_costs.
794 (find_costs_and_classes): Create prefs for the hard reg of small
795 reg class.
796 (process_bb_node_for_hard_reg_moves): Add prefs.
797
798 2013-10-30 Richard Biener <rguenther@suse.de>
799
800 PR middle-end/57100
801 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
802 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
803 as worker for ...
804 (pre_and_rev_post_order_compute): ... which now wraps it.
805 * graph.c (draw_cfg_nodes_no_loops): Use
806 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
807 on cfun.
808
809 2013-10-30 Christian Bruel <christian.bruel@st.com>
810
811 * gcc/config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
812 (sh_expand_cmpstr): Handle known align and schedule improvements.
813 * gcc/config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
814 * gcc/config/sh/sh.md (cmpstrnsi): New pattern.
815
816 2013-10-30 Martin Jambor <mjambor@suse.cz>
817
818 PR rtl-optimization/10474
819 * ira.c (find_moveable_pseudos): Do not calculate dominance info
820 nor df analysis.
821 (interesting_dest_for_shprep): New function.
822 (split_live_ranges_for_shrink_wrap): Likewise.
823 (ira): Calculate dominance info and df analysis. Call
824 split_live_ranges_for_shrink_wrap.
825
826 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
827
828 PR target/58854
829 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
830
831 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
832
833 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
834 * tree.h (POINTER_BOUNDS_P): New.
835 (BOUNDED_TYPE_P): New.
836 (BOUNDED_P): New.
837 (pointer_bounds_type_node): New.
838 * tree.c (build_common_tree_nodes): Initialize
839 pointer_bounds_type_node.
840 * gimple.h (gimple_call_get_nobnd_arg_index): New.
841 (gimple_call_num_nobnd_args): New.
842 (gimple_call_nobnd_arg): New.
843 (gimple_return_retbnd): New.
844 (gimple_return_set_retbnd): New
845 * gimple.c (gimple_build_return): Increase number of ops
846 for return statement.
847 (gimple_call_get_nobnd_arg_index): New.
848 * gimple-pretty-print.c (dump_gimple_return): Print second op.
849
850 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
851
852 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
853 with "chkp ctor" and "bnd_legacy" attributes.
854 * gimplify.c (gimplify_init_constructor): Avoid infinite
855 loop during gimplification of bounds initializer.
856
857 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
858
859 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
860 (handle_bnd_legacy): New.
861 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
862 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
863 attributes.
864
865 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
866
867 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
868 (BT_FN_PTR_CONST_PTR): New.
869 (BT_FN_CONST_PTR_CONST_PTR): New.
870 (BT_FN_PTR_CONST_PTR_SIZE): New.
871 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
872 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
873 (BT_FN_VOID_CONST_PTR_SIZE): New.
874 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
875 * chkp-builtins.def: New.
876 * builtins.def: include chkp-builtins.def.
877 (DEF_CHKP_BUILTIN): New.
878 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
879 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
880 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
881 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
882 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
883 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
884 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
885 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
886 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
887 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
888 * common.opt (fcheck-pointer-bounds): New.
889 * toplev.c (process_options): Check Pointer Bounds Checker is
890 supported.
891 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
892
893 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
894
895 * target.def (builtin_chkp_function): New.
896 (chkp_bound_type): New.
897 (chkp_bound_mode): New.
898 (fn_abi_va_list_bounds_size): New.
899 (load_bounds_for_arg): New.
900 (store_bounds_for_arg): New.
901 * targhooks.h (default_load_bounds_for_arg): New.
902 (default_store_bounds_for_arg): New.
903 (default_fn_abi_va_list_bounds_size): New.
904 (default_chkp_bound_type): New.
905 (default_chkp_bound_mode): New.
906 (default_builtin_chkp_function): New.
907 * targhooks.c (default_load_bounds_for_arg): New.
908 (default_store_bounds_for_arg): New.
909 (default_fn_abi_va_list_bounds_size): New.
910 (default_chkp_bound_type): New.
911 (default_chkp_bound_mode); New.
912 (default_builtin_chkp_function): New.
913 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
914 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
915 (TARGET_STORE_BOUNDS_FOR_ARG): New.
916 (TARGET_BUILTIN_CHKP_FUNCTION): New.
917 (TARGET_CHKP_BOUND_TYPE): New.
918 (TARGET_CHKP_BOUND_MODE): New.
919 * doc/tm.texi: Regenerated.
920 * langhooks.h (lang_hooks): Add chkp_supported field.
921 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
922 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
923
924 2013-10-29 Andrew Pinski <apinski@cavium.com>
925
926 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
927 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
928 fails, combining the branches anyways.
929 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
930 increases the number of combinings.
931 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
932
933 2013-10-29 Mike Stump <mikestump@comcast.net>
934
935 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
936 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
937 (make_vector_mode): Increase namebuf to 16.
938 (emit_insn_modes_h): When processing BImode, don't
939 also match partial int modes.
940 (emit_class_narrowest_mode): Likewise.
941
942 * config/bfin/bfin-modes.def: Add precision to PDI.
943 * config/m32c/m32c-modes.def: Add precision to PSI.
944 * config/msp430/msp430-modes.def: Add precision to PSI.
945 * config/rs6000/rs6000-modes.def: Add precision to PTI.
946 * config/sh/sh-modes.def: Add precision to PSI and PDI.
947
948 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
949
950 PR target/54236
951 * config/sh/sh.md (*addc): Rename existing variations to ...
952 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
953 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
954 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
955 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
956
957 2013-10-29 Teresa Johnson <tejohnson@google.com>
958
959 PR ipa/58862
960 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
961 insanities when updating probabilities.
962
963 2013-10-29 David Malcolm <dmalcolm@redhat.com>
964
965 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
966 prettyprinter for cgraph_node to reflect the conversion of the
967 symtable types to a C++ class hierarchy: it now *is* a
968 symtab_node_base, rather than having one (named "symbol").
969
970 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
971
972 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
973 __cilkrts_pop_frame. If matched, then return true for built-in
974 function name.
975 (expand_builtin): Added BUILT_IN_CILK_DETACH and
976 BUILT_IN_CILK_POP_FRAME case.
977 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
978 (lhs_cilk_detect_spawn): Likewise.
979 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
980 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
981 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
982 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
983 (LANG_HOOKS_CILKPLUS): Likewise.
984 * tree.h (CILK_SPAWN_FN): Likewise.
985 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
986 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
987 (OBJS): Added cilk-common.o.
988 (BUILTINS_DEF): Added cilk-builtins.def.
989 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
990 (lhd_cilk_detect_spawn): Likewise.
991 * langhooks.h (lang_hooks_for_cilkplus): New struct.
992 (struct lang_hooks): Added new field called "cilkplus."
993 * cilk-common.c: New file.
994 * cilk.h: Likewise.
995 * cilk-builtins.def: Likewise.
996 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
997 "__cilk" macro and set it to 200.
998 * function.h (struct function::cilk_frame_decl): New field.
999 (struct function::is_cilk_function): Likewise.
1000 (struct function::calls_cilk_spawn): Likewise.
1001 * gimplify.c (gimplify_call_expr): Added a check if the function call
1002 being gimplified is a spawn detach point. If so, then add pop_frame
1003 and detach function calls.
1004 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
1005 for gimplifying _Cilk_spawn and _Cilk_sync statements.
1006 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
1007 function. If so, added a _Cilk_sync and gimplified it.
1008 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
1009 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
1010 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
1011 spawner function.
1012 (can_inline_edge_p): Prevent inling of spawnee function.
1013 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
1014 for functions that use Cilk keywords.
1015 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
1016 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
1017 CILK_SYNC_STMT cases.
1018 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
1019 trees.
1020 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
1021 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
1022 * passes.texi (Cilk Keywords): New section that describes the compiler
1023 code changes for handling Cilk Keywords.
1024
1025 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1026
1027 Patch autogenerated by refactor_symtab.py from
1028 https://github.com/davidmalcolm/gcc-refactoring-scripts
1029 revision 58bb219cc090b2f4516a9297d868c245495ee622
1030
1031 * asan.c (asan_finish_file): Update for conversion of symtab types to
1032 a true class hierarchy.
1033 * cfgexpand.c (estimated_stack_frame_size): Likewise.
1034 * cgraph.c (cgraph_get_body): Likewise.
1035 (cgraph_get_create_real_symbol_node): Likewise.
1036 (verify_cgraph_node): Likewise.
1037 (verify_edge_corresponds_to_fndecl): Likewise.
1038 (verify_edge_count_and_frequency): Likewise.
1039 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
1040 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1041 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
1042 (cgraph_node_cannot_return): Likewise.
1043 (cgraph_set_pure_flag_1): Likewise.
1044 (cgraph_set_const_flag_1): Likewise.
1045 (cgraph_set_nothrow_flag_1): Likewise.
1046 (cgraph_make_node_local_1): Likewise.
1047 (cgraph_for_node_and_aliases): Likewise.
1048 (cgraph_for_node_thunks_and_aliases): Likewise.
1049 (cgraph_node_can_be_local_p): Likewise.
1050 (cgraph_node_cannot_be_local_p_1): Likewise.
1051 (cgraph_function_body_availability): Likewise.
1052 (dump_cgraph_node): Likewise.
1053 (cgraph_rtl_info): Likewise.
1054 (cgraph_mark_address_taken_node): Likewise.
1055 (cgraph_remove_node): Likewise.
1056 (cgraph_release_function_body): Likewise.
1057 (cgraph_update_edges_for_call_stmt_node): Likewise.
1058 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
1059 (cgraph_make_edge_direct): Likewise.
1060 (cgraph_resolve_speculation): Likewise.
1061 (cgraph_speculative_call_info): Likewise.
1062 (cgraph_turn_edge_to_speculative): Likewise.
1063 (cgraph_create_edge_1): Likewise.
1064 (cgraph_set_call_stmt): Likewise.
1065 (cgraph_node_for_asm): Likewise.
1066 (cgraph_add_thunk): Likewise.
1067 (cgraph_same_body_alias): Likewise.
1068 (cgraph_create_function_alias): Likewise.
1069 (cgraph_create_node): Likewise.
1070 (cgraph_create_empty_node): Likewise.
1071 (record_function_versions): Likewise.
1072 (used_from_object_file_p): Likewise.
1073 * cgraph.h (symtab_can_be_discarded): Likewise.
1074 (symtab_real_symbol_p): Likewise.
1075 (cgraph_mark_force_output_node): Likewise.
1076 (cgraph_edge_recursive_p): Likewise.
1077 (symtab_alias_target): Likewise.
1078 (varpool_all_refs_explicit_p): Likewise.
1079 (varpool_can_remove_if_no_refs): Likewise.
1080 (cgraph_only_called_directly_or_aliased_p): Likewise.
1081 (cgraph_next_function_with_gimple_body): Likewise.
1082 (cgraph_first_function_with_gimple_body): Likewise.
1083 (cgraph_function_with_gimple_body_p): Likewise.
1084 (cgraph_next_function): Likewise.
1085 (cgraph_first_function): Likewise.
1086 (cgraph_next_defined_function): Likewise.
1087 (cgraph_first_defined_function): Likewise.
1088 (varpool_next_defined_variable): Likewise.
1089 (varpool_first_defined_variable): Likewise.
1090 (varpool_next_static_initializer): Likewise.
1091 (varpool_first_static_initializer): Likewise.
1092 (varpool_next_variable): Likewise.
1093 (varpool_first_variable): Likewise.
1094 (varpool_node_name): Likewise.
1095 (varpool): Likewise.
1096 (cgraph): Likewise.
1097 (is_a_helper <varpool_node>::test): Likewise.
1098 (is_a_helper <cgraph_node>::test): Likewise.
1099 (varpool_variable_node): Likewise.
1100 (cgraph_function_or_thunk_node): Likewise.
1101 (varpool_alias_target): Likewise.
1102 (cgraph_alias_target): Likewise.
1103 (cgraph_node_name): Likewise.
1104 (varpool_node_asm_name): Likewise.
1105 (cgraph_node_asm_name): Likewise.
1106 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
1107 (cgraph_rebuild_references): Likewise.
1108 (rebuild_cgraph_edges): Likewise.
1109 (record_eh_tables): Likewise.
1110 (build_cgraph_edges): Likewise.
1111 (mark_store): Likewise.
1112 (mark_load): Likewise.
1113 (mark_address): Likewise.
1114 (record_type_list): Likewise.
1115 (record_reference): Likewise.
1116 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
1117 (cgraph_materialize_clone): Likewise.
1118 (cgraph_function_versioning): Likewise.
1119 (cgraph_copy_node_for_versioning): Likewise.
1120 (update_call_expr): Likewise.
1121 (cgraph_find_replacement_node): Likewise.
1122 (cgraph_create_virtual_clone): Likewise.
1123 (cgraph_clone_node): Likewise.
1124 * cgraphunit.c (compile): Likewise.
1125 (output_weakrefs): Likewise.
1126 (output_in_order): Likewise.
1127 (expand_function): Likewise.
1128 (assemble_thunks_and_aliases): Likewise.
1129 (expand_thunk): Likewise.
1130 (mark_functions_to_output): Likewise.
1131 (handle_alias_pairs): Likewise.
1132 (analyze_functions): Likewise.
1133 (walk_polymorphic_call_targets): Likewise.
1134 (varpool_finalize_decl): Likewise.
1135 (process_function_and_variable_attributes): Likewise.
1136 (cgraph_process_same_body_aliases): Likewise.
1137 (analyze_function): Likewise.
1138 (cgraph_add_new_function): Likewise.
1139 (cgraph_finalize_function): Likewise.
1140 (referred_to_p): Likewise.
1141 (cgraph_reset_node): Likewise.
1142 (cgraph_process_new_functions): Likewise.
1143 (enqueue_node): Likewise.
1144 (decide_is_symbol_needed): Likewise.
1145 * coverage.c (coverage_compute_profile_id): Likewise.
1146 * dbxout.c (dbxout_expand_expr): Likewise.
1147 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
1148 (reference_to_unused): Likewise.
1149 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1150 * gimplify.c (unvisit_body): Likewise.
1151 (unshare_body): Likewise.
1152 * ipa-cp.c (ipcp_generate_summary): Likewise.
1153 (ipcp_decision_stage): Likewise.
1154 (identify_dead_nodes): Likewise.
1155 (decide_whether_version_node): Likewise.
1156 (decide_about_value): Likewise.
1157 (perhaps_add_new_callers): Likewise.
1158 (create_specialized_node): Likewise.
1159 (update_profiling_info): Likewise.
1160 (ipcp_propagate_stage): Likewise.
1161 (estimate_local_effects): Likewise.
1162 (good_cloning_opportunity_p): Likewise.
1163 (devirtualization_time_bonus): Likewise.
1164 (propagate_constants_accross_call): Likewise.
1165 (initialize_node_lattices): Likewise.
1166 (ipcp_cloning_candidate_p): Likewise.
1167 (determine_versionability): Likewise.
1168 (print_all_lattices): Likewise.
1169 (print_lattice): Likewise.
1170 (ipcp_discover_new_direct_edges): Likewise.
1171 * ipa-devirt.c (ipa_devirt): Likewise.
1172 (likely_target_p): Likewise.
1173 (update_type_inheritance_graph): Likewise.
1174 (possible_polymorphic_call_target_p): Likewise.
1175 (dump_possible_polymorphic_call_targets): Likewise.
1176 (devirt_variable_node_removal_hook): Likewise.
1177 (record_binfo): Likewise.
1178 (maybe_record_node): Likewise.
1179 (build_type_inheritance_graph): Likewise.
1180 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1181 (inline_generate_summary): Likewise.
1182 (inline_analyze_function): Likewise.
1183 (do_estimate_growth): Likewise.
1184 (simple_edge_hints): Likewise.
1185 (estimate_node_size_and_time): Likewise.
1186 (estimate_edge_devirt_benefit): Likewise.
1187 (compute_inline_parameters): Likewise.
1188 (estimate_function_body_sizes): Likewise.
1189 (compute_bb_predicates): Likewise.
1190 (initialize_inline_failed): Likewise.
1191 (dump_inline_summary): Likewise.
1192 (dump_inline_edge_summary): Likewise.
1193 * ipa-inline-transform.c (inline_transform): Likewise.
1194 (preserve_function_body_p): Likewise.
1195 (save_inline_function_body): Likewise.
1196 (inline_call): Likewise.
1197 (clone_inlined_nodes): Likewise.
1198 (can_remove_node_now_p): Likewise.
1199 (can_remove_node_now_p_1): Likewise.
1200 * ipa-inline.c (early_inliner): Likewise.
1201 (early_inline_small_functions): Likewise.
1202 (inline_always_inline_functions): Likewise.
1203 (ipa_inline): Likewise.
1204 (flatten_function): Likewise.
1205 (inline_small_functions): Likewise.
1206 (speculation_useful_p): Likewise.
1207 (recursive_inlining): Likewise.
1208 (update_caller_keys): Likewise.
1209 (reset_edge_caches): Likewise.
1210 (update_edge_key): Likewise.
1211 (edge_badness): Likewise.
1212 (relative_time_benefit): Likewise.
1213 (want_inline_self_recursive_call_p): Likewise.
1214 (want_inline_small_function_p): Likewise.
1215 (want_early_inline_function_p): Likewise.
1216 (num_calls): Likewise.
1217 (can_early_inline_edge_p): Likewise.
1218 (can_inline_edge_p): Likewise.
1219 (report_inline_failed_reason): Likewise.
1220 * ipa-profile.c (ipa_profile): Likewise.
1221 (ipa_propagate_frequency): Likewise.
1222 (ipa_propagate_frequency_1): Likewise.
1223 (ipa_profile_generate_summary): Likewise.
1224 * ipa-prop.c (ipcp_transform_function): Likewise.
1225 (read_replacements_section): Likewise.
1226 (ipa_prop_read_section): Likewise.
1227 (ipa_modify_call_arguments): Likewise.
1228 (ipa_print_node_params): Likewise.
1229 (propagate_controlled_uses): Likewise.
1230 (update_indirect_edges_after_inlining): Likewise.
1231 (remove_described_reference): Likewise.
1232 (ipa_make_edge_direct_to_target): Likewise.
1233 (ipa_analyze_node): Likewise.
1234 (ipa_analyze_params_uses): Likewise.
1235 (ipa_compute_jump_functions): Likewise.
1236 (ipa_get_callee_param_type): Likewise.
1237 (ipa_print_node_jump_functions): Likewise.
1238 (ipa_initialize_node_params): Likewise.
1239 (ipa_populate_param_decls): Likewise.
1240 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
1241 (write_agg_replacement_chain): Likewise.
1242 (ipa_write_node_info): Likewise.
1243 (ipa_edge_duplication_hook): Likewise.
1244 (try_decrement_rdesc_refcount): Likewise.
1245 * ipa-pure-const.c (propagate_nothrow): Likewise.
1246 (propagate_pure_const): Likewise.
1247 (pure_const_read_summary): Likewise.
1248 (pure_const_write_summary): Likewise.
1249 (analyze_function): Likewise.
1250 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
1251 (ipa_ref_referring_ref_list): Likewise.
1252 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
1253 (ipa_remove_stmt_references): Likewise.
1254 (ipa_find_reference): Likewise.
1255 (ipa_dump_referring): Likewise.
1256 (ipa_dump_references): Likewise.
1257 (ipa_record_reference): Likewise.
1258 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
1259 (ipa_reference_write_optimization_summary): Likewise.
1260 (write_node_summary_p): Likewise.
1261 (propagate): Likewise.
1262 (read_write_all_from_decl): Likewise.
1263 (generate_summary): Likewise.
1264 (analyze_function): Likewise.
1265 (propagate_bits): Likewise.
1266 (ipa_reference_get_not_written_global): Likewise.
1267 (ipa_reference_get_not_read_global): Likewise.
1268 * ipa-split.c (execute_split_functions): Likewise.
1269 (split_function): Likewise.
1270 * ipa-utils.c (ipa_merge_profiles): Likewise.
1271 (dump_cgraph_node_set): Likewise.
1272 (ipa_reverse_postorder): Likewise.
1273 (ipa_edge_within_scc): Likewise.
1274 (ipa_get_nodes_in_cycle): Likewise.
1275 (ipa_free_postorder_info): Likewise.
1276 (ipa_reduced_postorder): Likewise.
1277 (searchc): Likewise.
1278 (recursive_call_p): Likewise.
1279 * ipa.c (ipa_cdtor_merge): Likewise.
1280 (record_cdtor_fn): Likewise.
1281 (function_and_variable_visibility): Likewise.
1282 (varpool_externally_visible_p): Likewise.
1283 (cgraph_externally_visible_p): Likewise.
1284 (comdat_can_be_unshared_p): Likewise.
1285 (comdat_can_be_unshared_p_1): Likewise.
1286 (address_taken_from_non_vtable_p): Likewise.
1287 (ipa_discover_readonly_nonaddressable_vars): Likewise.
1288 (symtab_remove_unreachable_nodes): Likewise.
1289 (walk_polymorphic_call_targets): Likewise.
1290 (process_references): Likewise.
1291 (enqueue_node): Likewise.
1292 (has_addr_references_p): Likewise.
1293 (cgraph_non_local_node_p_1): Likewise.
1294 * is-a.h (varpool_analyze_node): Likewise.
1295 * lto-cgraph.c (input_symtab): Likewise.
1296 (merge_profile_summaries): Likewise.
1297 (input_cgraph_1): Likewise.
1298 (input_edge): Likewise.
1299 (input_varpool_node): Likewise.
1300 (input_node): Likewise.
1301 (input_overwrite_node): Likewise.
1302 (compute_ltrans_boundary): Likewise.
1303 (output_refs): Likewise.
1304 (lto_output_varpool_node): Likewise.
1305 (lto_output_node): Likewise.
1306 (reachable_from_other_partition_p): Likewise.
1307 (referenced_from_other_partition_p): Likewise.
1308 (lto_output_edge): Likewise.
1309 (output_node_opt_summary): Likewise.
1310 (add_node_to): Likewise.
1311 (reachable_from_this_partition_p): Likewise.
1312 (lto_set_symtab_encoder_in_partition): Likewise.
1313 (lto_symtab_encoder_in_partition_p): Likewise.
1314 (lto_set_symtab_encoder_encode_initializer): Likewise.
1315 (lto_symtab_encoder_encode_initializer_p): Likewise.
1316 (lto_set_symtab_encoder_encode_body): Likewise.
1317 (lto_symtab_encoder_encode_body_p): Likewise.
1318 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1319 Likewise.
1320 * lto-streamer-in.c (lto_read_body): Likewise.
1321 (fixup_call_stmt_edges): Likewise.
1322 (fixup_call_stmt_edges_1): Likewise.
1323 * lto-streamer-out.c (produce_symtab): Likewise.
1324 (output_symbol_p): Likewise.
1325 (write_symbol): Likewise.
1326 (lto_output): Likewise.
1327 (copy_function): Likewise.
1328 (output_function): Likewise.
1329 * passes.c (function_called_by_processed_nodes_p): Likewise.
1330 (ipa_write_optimization_summaries): Likewise.
1331 (ipa_write_summaries): Likewise.
1332 (do_per_function_toporder): Likewise.
1333 (do_per_function): Likewise.
1334 (dump_passes): Likewise.
1335 * symtab.c (symtab_semantically_equivalent_p): Likewise.
1336 (symtab_nonoverwritable_alias): Likewise.
1337 (symtab_nonoverwritable_alias_1): Likewise.
1338 (symtab_for_node_and_aliases): Likewise.
1339 (symtab_resolve_alias): Likewise.
1340 (fixup_same_cpp_alias_visibility): Likewise.
1341 (symtab_alias_ultimate_target): Likewise.
1342 (symtab_used_from_object_file_p): Likewise.
1343 (verify_symtab_base): Likewise.
1344 (dump_symtab_base): Likewise.
1345 (symtab_node_name): Likewise.
1346 (symtab_node_asm_name): Likewise.
1347 (symtab_dissolve_same_comdat_group_list): Likewise.
1348 (symtab_add_to_same_comdat_group): Likewise.
1349 (symtab_unregister_node): Likewise.
1350 (symtab_insert_node_to_hashtable): Likewise.
1351 (symtab_register_node): Likewise.
1352 (unlink_from_assembler_name_hash): Likewise.
1353 (insert_to_assembler_name_hash): Likewise.
1354 (eq_assembler_name): Likewise.
1355 (hash_node_by_assembler_name): Likewise.
1356 (eq_node): Likewise.
1357 (hash_node): Likewise.
1358 * toplev.c (wrapup_global_declaration_2): Likewise.
1359 * trans-mem.c (ipa_tm_execute): Likewise.
1360 (ipa_tm_transform_clone): Likewise.
1361 (ipa_tm_transform_transaction): Likewise.
1362 (ipa_tm_transform_calls_redirect): Likewise.
1363 (ipa_tm_insert_gettmclone_call): Likewise.
1364 (ipa_tm_insert_irr_call): Likewise.
1365 (ipa_tm_create_version): Likewise.
1366 (ipa_tm_create_version_alias): Likewise.
1367 (ipa_tm_mark_forced_by_abi_node): Likewise.
1368 (ipa_tm_mark_force_output_node): Likewise.
1369 (ipa_tm_diagnose_tm_safe): Likewise.
1370 (ipa_tm_mayenterirr_function): Likewise.
1371 (ipa_tm_scan_irr_function): Likewise.
1372 (ipa_tm_note_irrevocable): Likewise.
1373 (ipa_tm_scan_calls_clone): Likewise.
1374 (get_cg_data): Likewise.
1375 * tree-eh.c (tree_could_trap_p): Likewise.
1376 * tree-emutls.c (ipa_lower_emutls): Likewise.
1377 (create_emultls_var): Likewise.
1378 (lower_emutls_function_body): Likewise.
1379 (gen_emutls_addr): Likewise.
1380 (emutls_decl): Likewise.
1381 (new_emutls_decl): Likewise.
1382 * tree-inline.c (tree_function_versioning): Likewise.
1383 (optimize_inline_calls): Likewise.
1384 (expand_call_inline): Likewise.
1385 (estimate_num_insns): Likewise.
1386 (copy_bb): Likewise.
1387 (delete_unreachable_blocks_update_callgraph): Likewise.
1388 * tree-nested.c (gimplify_all_functions): Likewise.
1389 (create_nesting_tree): Likewise.
1390 (check_for_nested_with_variably_modified): Likewise.
1391 * tree-pretty-print.c (dump_function_header): Likewise.
1392 * tree-profile.c (tree_profiling): Likewise.
1393 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
1394 (modify_function): Likewise.
1395 (convert_callers): Likewise.
1396 (convert_callers_for_node): Likewise.
1397 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
1398 (associate_varinfo_to_alias): Likewise.
1399 (create_variable_info_for): Likewise.
1400 (get_constraint_for_ssa_var): Likewise.
1401 * tree-vectorizer.c (increase_alignment): Likewise.
1402 * tree.c (find_decls_types_in_var): Likewise.
1403 (find_decls_types_in_node): Likewise.
1404 (free_lang_data_in_decl): Likewise.
1405 * value-prof.c (gimple_ic_transform): Likewise.
1406 (gimple_ic): Likewise.
1407 (check_ic_target): Likewise.
1408 (init_node_map): Likewise.
1409 * varasm.c (decl_binds_to_current_def_p): Likewise.
1410 (default_binds_local_p_1): Likewise.
1411 (dump_tm_clone_pairs): Likewise.
1412 (assemble_alias): Likewise.
1413 (find_decl): Likewise.
1414 (mark_decl_referenced): Likewise.
1415 * varpool.c (varpool_for_node_and_aliases): Likewise.
1416 (varpool_extra_name_alias): Likewise.
1417 (varpool_create_variable_alias): Likewise.
1418 (add_new_static_var): Likewise.
1419 (varpool_finalize_named_section_flags): Likewise.
1420 (varpool_remove_unreferenced_decls): Likewise.
1421 (enqueue_node): Likewise.
1422 (varpool_assemble_decl): Likewise.
1423 (assemble_aliases): Likewise.
1424 (varpool_analyze_node): Likewise.
1425 (cgraph_variable_initializer_availability): Likewise.
1426 (varpool_add_new_variable): Likewise.
1427 (ctor_for_folding): Likewise.
1428 (dump_varpool_node): Likewise.
1429 (varpool_remove_initializer): Likewise.
1430 (varpool_remove_node): Likewise.
1431 (varpool_node_for_decl): Likewise.
1432 (varpool_create_empty_node): Likewise.
1433 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
1434 (ix86_get_function_versions_dispatcher): Likewise.
1435
1436 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1437
1438 * cgraph.h (symtab_node_base): Convert to a class;
1439 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
1440 chain_next/prev from symtab_node_def.
1441 (cgraph_node): Inherit from symtab_node; add GTY option
1442 tag ("SYMTAB_FUNCTION").
1443 (varpool_node): Inherit from symtab_node; add GTY option
1444 tag ("SYMTAB_VARIABLE").
1445 (symtab_node_def): Remove.
1446 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
1447 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
1448 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
1449 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
1450
1451 * ipa-ref.h (symtab_node_def): Drop.
1452 (symtab_node): Change underlying type from symtab_node_def to
1453 symtab_node_base.
1454 (const_symtab_node): Likwise.
1455
1456 * is-a.h: Update examples in comment.
1457
1458 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
1459 (assembler_name_hash): Likewise.
1460
1461 2013-10-29 Martin Liska <marxin.liska@gmail.com>
1462
1463 * doc/tree-ssa.texi (gimple_phi_result): Document.
1464 (gimple_phi_num_args, gimple_phi_arg): Likewise.
1465 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
1466 (PHI_RESULT, PHI_NUM_ARGS): Remove.
1467 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
1468
1469 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1470
1471 * expr.h: Revert change and include tree-core.h.
1472 * rtl.h: Revert change and don't include tree-core.h.
1473
1474 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1475
1476 * config/darwin.c: Include gimple.h.
1477 * config/i386/winnt.c: Likewise.
1478
1479 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
1480
1481 PR tree-optimization/19831
1482 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
1483
1484 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1485
1486 * tree-outof-ssa.h: Remove include files.
1487 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
1488 * expr.c: Likewise.
1489 * tree-ssa-coalesce.c: Likewise.
1490 * cfgexpand.c: Likewise.
1491 * tree-ssa-ter.c: Likewise.
1492 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
1493 * lto-streamer.h: Likewise.
1494 * cgraphbuild.c: Add gimple.h to include list.
1495 * data-streamer-in.c: Likewise.
1496 * ipa-cp.c: Likewise.
1497 * tree-streamer.c: Likewise.
1498 * lto-compress.c: Likewise.
1499 * ipa-reference.c: Likewise.
1500 * data-streamer-out.c: Likewise.
1501 * lto-cgraph.c: Likewise.
1502 * cgraphclones.c: Likewise.
1503 * ipa-utils.c: Likewise.
1504 * data-streamer.c: Likewise.
1505 * ipa-split.c: Likewise.
1506 * lto-section-in.c: Likewise.
1507 * tree-streamer-out.c: Likewise.
1508 * ipa-prop.c: Likewise.
1509 * tree-streamer-in.c: Likewise.
1510 * symtab.c: Likewise.
1511 * opts-global.c: Likewise.
1512 * lto-opts.c: Likewise.
1513 * lto-section-out.c: Likewise.
1514 * lto-streamer.c: Likewise.
1515 * rtl.h: Add tree-core.h to include list.
1516 * expr.h: Remove tree-core.h from include list.
1517 * gimple.h: Likewise.
1518 * ipa-utils.h: Likewise.
1519 * streamer-hooks.h: Likewise.
1520 * streamer-hooks.c: Include input.h.
1521
1522 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1523
1524 * config/arm/arm.c (cortexa7_extra_costs): New table.
1525 (arm_cortex_a7_tune): New.
1526 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
1527
1528 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
1529
1530 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
1531
1532 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1533
1534 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
1535 to use autogenerated markers for a class-hierarchy, every class
1536 must have a GTY marker.
1537 * gengtype.h (struct type): Add linked list of subclasses to
1538 the "s" member of the union.
1539 (add_subclass): New decl.
1540 * gengtype-state.c (read_state_struct_type): Set up subclass
1541 linked list.
1542 * gengtype.c (get_ultimate_base_class): New.
1543 (add_subclass): New.
1544 (new_structure): Set up subclass linked list.
1545 (set_gc_used_type): Propagate usage information to subclasses.
1546 (output_mangled_typename): Use get_ultimate_base_class.
1547 (walk_subclasses): Use the subclass linked list, avoiding an
1548 O(N^2) when writing out all types.
1549 (walk_type): Issue an error if the base class is missing a tag,
1550 rather than generating bogus C code. Add a gcc_unreachable
1551 default case, in case people omit tags from concrete subclasses,
1552 or get the values wrong.
1553 (write_func_for_structure): Issue an error for subclasses for
1554 which the base doesn't have a "desc", since otherwise the
1555 autogenerated routines for the base would silently fail to visit
1556 any subclass fields.
1557 (write_root): Use get_ultimate_base_class, tweaking constness of
1558 tp to match that function's signature.
1559
1560 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1561
1562 * doc/gty.texi (GTY Options): Add note about inheritance to
1563 description of desc and tag.
1564 (Inheritance and GTY): New.
1565
1566 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1567
1568 * gengtype-parse.c (opts_have): Drop "static" so that
1569 we can use this from gengtype.c.
1570 * gengtype.c (set_gc_used_type): Mark any base class as used;
1571 update field traversal to visit inherited fields.
1572 (output_mangled_typename): Convert references to classes within
1573 an inheritance hierarchy to reference the ultimate base class,
1574 since only it will have gt_ functions.
1575 (get_string_option): New.
1576 (walk_subclasses): New.
1577 (walk_type): Treat GTY structs that have a "desc" as being the
1578 root of an inheritance hierarchy. Generate a switch on it
1579 within the marking function which walks all subclasses, adding
1580 cases for them via walk_subclasses. For subclasses, visit all
1581 fields of the type (including inherited ones).
1582 (write_func_for_structure): Don't write fns for subclasses, only
1583 for the ultimate base class within an inheritance hierarchy.
1584 Subclasses-marking will be handled by the base class marking functions.
1585 (write_types): Likewise.
1586 (write_local_func_for_structure): Likewise.
1587 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
1588 a "tag" option (and are thus concrete subclasses).
1589 (write_root): Use the marker function for the ultimate base class.
1590 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
1591 (opts_have): Add declaration.
1592
1593 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
1594
1595 * lra-spills.c (lra_final_code_change): Remove useless move insns
1596 originated from moves of pseudos.
1597
1598 2013-10-28 Jeff Law <law@redhat.com>
1599
1600 * lower-subreg.c (resolve_simple_move): Fix comment typo.
1601
1602 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
1603
1604 * df-scan.c (df_collection_rec): Adjust.
1605 (copy_defs): New constant.
1606 (copy_uses): Likewise.
1607 (copy_eq_uses): Likewise.
1608 (copy_mw): Likewise.
1609 (copy_all): Likewise.
1610 (df_insn_rescan): Adjust.
1611 (df_notes_rescan): Likewise.
1612 (df_swap_refs): Likewise.
1613 (df_sort_and_compress_refs): Likewise.
1614 (df_sort_and_compress_mws): Likewise.
1615 (df_install_refs): Likewise.
1616 (df_install_mws): Likewise.
1617 (df_refs_add_to_chains): Add flags parameter controlling which vectors
1618 are coppied.
1619 (df_bb_refs_record): Adjust.
1620 (df_record_entry_block_defs): Likewise.
1621 (df_record_exit_block_defs): Likewise.
1622 (df_refs_verify): Likewise.
1623 (df_mws_verify): Likewise.
1624 (df_insn_refs_verify): Likewise.
1625 (df_bb_verify): Likewise.
1626 * ipa-pure-const.c (finish_state): Remove.
1627 (propagate): Adjust.
1628 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
1629 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
1630 var-tracking.c: Adjust.
1631 * vec.c (stack_vecs): Remove.
1632 (register_stack_vec): Likewise.
1633 (stack_vec_register_index): Likewise.
1634 (unregister_stack_vec): Likewise.
1635 * vec.h (struct va_stack): Remove.
1636 (struct vec<T, A, vl_ptr>): Specialize as
1637 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
1638 allocation strategy compatable with the vl_ptr layout.
1639 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
1640 specialization anyway.
1641 (class stack_vec): New class.
1642 (vec_stack_alloc): Remove.
1643 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
1644
1645 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1646 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1647 Sergey Lega <sergey.s.lega@intel.com>
1648 Anna Tikhonova <anna.tikhonova@intel.com>
1649 Ilya Tocar <ilya.tocar@intel.com>
1650 Andrey Turetskiy <andrey.turetskiy@intel.com>
1651 Ilya Verbin <ilya.verbin@intel.com>
1652 Kirill Yukhin <kirill.yukhin@intel.com>
1653 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1654
1655 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
1656 (*prefetch_avx512pf_<mode>): New.
1657 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
1658 (avx512f_maskcmp<mode>3): Ditto.
1659 (vashrv16si3): Ditto.
1660
1661 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1662 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1663 Sergey Lega <sergey.s.lega@intel.com>
1664 Anna Tikhonova <anna.tikhonova@intel.com>
1665 Ilya Tocar <ilya.tocar@intel.com>
1666 Andrey Turetskiy <andrey.turetskiy@intel.com>
1667 Ilya Verbin <ilya.verbin@intel.com>
1668 Kirill Yukhin <kirill.yukhin@intel.com>
1669 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1670
1671 * config/i386/i386.md (any_truncate): New.
1672 (trunsuffix): Ditto.
1673 * config/i386/predicates.md (const_8_to_9_operand): New.
1674 (const_10_to_11_operand): Ditto.
1675 (const_12_to_13_operand): Ditto.
1676 (const_14_to_15_operand): Ditto.
1677 (const_16_to_19_operand): Ditto.
1678 (const_20_to_23_operand): Ditto.
1679 (const_24_to_27_operand): Ditto.
1680 (const_28_to_31_operand): Ditto.
1681 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
1682 (cvtusi2<ssescalarmodesuffix>32): New.
1683 (cvtusi2<ssescalarmodesuffix>64): Ditto.
1684 (ufloatv16siv16sf2): Ditto.
1685 (avx512f_fix_notruncv16sfv16si): Ditto.
1686 (avx512f_ufix_notruncv16sfv16si): Ditto.
1687 (avx512f_vcvtss2usi): Ditto.
1688 (avx512f_vcvtss2usiq): Ditto.
1689 (avx512f_vcvttss2usi): Ditto.
1690 (avx512f_vcvttss2usiq): Ditto.
1691 (avx512f_vcvtsd2usi): Ditto.
1692 (avx512f_vcvtsd2usiq): Ditto.
1693 (avx512f_vcvttsd2usi): Ditto.
1694 (avx512f_vcvttsd2usiq): Ditto.
1695 (ufloatv8siv8df): Ditto.
1696 (avx512f_cvtdq2pd512_2): Ditto.
1697 (avx512f_cvtpd2dq512): Ditto.
1698 (avx512f_ufix_notruncv8dfv8si): Ditto.
1699 (avx512f_cvtpd2ps512): Ditto.
1700 (vec_unpacks_lo_v16sf): Ditto.
1701 (vec_unpacks_hi_v16sf): Ditto.
1702 (vec_unpacks_float_hi_v16si): Ditto.
1703 (vec_unpacks_float_lo_v16si): Ditto.
1704 (avx512f_unpckhps512): Ditto.
1705 (avx512f_unpcklps512): Ditto.
1706 (avx512f_movshdup512): Ditto.
1707 (avx512f_movsldup512): Ditto.
1708 (vec_extract_lo_v32hi): Ditto.
1709 (vec_extract_hi_v32hi): Ditto.
1710 (vec_extract_lo_v64qi): Ditto.
1711 (vec_extract_hi_v64qi): Ditto.
1712 (avx512f_unpckhpd512): Ditto.
1713 (avx512f_movddup512): Ditto.
1714 (avx512f_unpcklpd512): Ditto.
1715 (*avx512f_unpcklpd512): Ditto.
1716 (avx512f_shufps512_1): Ditto.
1717 (avx512f_shufpd512_1): Ditto.
1718 (avx512f_interleave_highv8di): Ditto.
1719 (avx512f_interleave_lowv8di): Ditto.
1720 (PMOV_DST_MODE): Ditto.
1721 (pmov_src_mode): Ditto.
1722 (pmov_src_lower): Ditto.
1723 (pmov_suff): Ditto.
1724 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
1725 (*avx512f_<code>v8div16qi2): Ditto.
1726 (*avx512f_<code>v8div16qi2_store): Ditto.
1727 (vec_widen_umult_even_v16si): Ditto.
1728 (*vec_widen_umult_even_v16si): Ditto.
1729 (vec_widen_smult_even_v16si): Ditto.
1730 (*vec_widen_smult_even_v16si): Ditto.
1731 (avx512f_interleave_highv16si): Ditto.
1732 (avx512f_interleave_lowv16si): Ditto.
1733 (avx512f_<code>v16qiv16si2): Ditto.
1734 (avx512f_<code>v16hiv16si2): Ditto.
1735 (avx512f_<code>v8qiv8di2): Ditto.
1736 (avx512f_<code>v8hiv8di2): Ditto.
1737 (avx512f_<code>v8siv8di2): Ditto.
1738 (avx512cd_maskb_vec_dupv8di): Ditto.
1739 (avx512cd_maskw_vec_dupv16si): Ditto.
1740 (avx512f_vcvtph2ps512): Ditto.
1741 (avx512f_vcvtps2ph512): Ditto.
1742 (VEC_EXTRACT_MODE): Extened with wider modes.
1743 (VEC_PERM_AVX2): Ditto.
1744 (VEC_PERM_CONST): Ditto.
1745
1746 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
1747
1748 * config/arc/arc.c (arc_ccfsm_post_advance):
1749 Add comment about TYPE_RETURN.
1750
1751 2013-10-28 Bin Cheng <bin.cheng@arm.com>
1752
1753 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
1754 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
1755 (strip_offset): Convert offset to unsigned number.
1756
1757 2013-10-27 Tom de Vries <tom@codesourcery.com>
1758
1759 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
1760 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
1761
1762 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
1763
1764 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
1765 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
1766 (TARGET_BIG_ENDIAN): New macro.
1767 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
1768 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
1769 * config/sh/sh.c: Likewise.
1770 * config/sh/sh.h: Likewise.
1771
1772 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
1773
1774 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
1775 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
1776
1777 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1778
1779 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
1780 Fix formatting.
1781 (cmpstr_t, cmpstrsi): Fix formatting.
1782
1783 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1784
1785 PR target/52483
1786 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
1787 addressing, do not use general_operand for memory operands.
1788
1789 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
1790
1791 Revert:
1792 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1793 * lra-spills.c (lra_final_code_change): Remove useless move insns.
1794
1795 2013-10-26 Jeff Law <law@redhat.com>
1796
1797 * predict.c (PRED_MUDFLAP): Remove.
1798 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
1799
1800 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
1801 (OBJS): Remove tree-nomudflap.o
1802 (GTFILES): Remove tree-mudflap.c
1803 * builtins.c (expand_builtin_alloc): Remove mudflap support.
1804 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
1805 (mfwrap_spec, mflib_spec): Likewise.
1806 (cpp_unique_options, cc1_options, static_specs): Likewise.
1807 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
1808 * passes.def: Likewise.
1809 * toplev.c (compile_file, process_options): Likewise.
1810 * tree-inline.c (copy_tree_r): Likewise.
1811 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
1812 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
1813 (build_constant_desc, output_constant_def_contents): Likewise.
1814 (categorize_decl_for_section): Likewise.
1815 * tree-mudflap.c: Removed.
1816 * tree-mudflap.h: Removed.
1817 * tree-nomudflap.c: Removed.
1818 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
1819 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
1820 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
1821 * config/sol2.h (MFLIB_SPEC): Likewise.
1822 * doc/install.texi: Remove mudflap references.
1823 * doc/passes.texi: Similarly.
1824 * doc/sourcebuild.texi: Similarly.
1825 * doc/invoke.texi: Remove mudlfap related options.
1826
1827 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1828
1829 PR rtl-optimization/58759
1830 * lra-constraints.c (lra_constraints): Remove wrong condition to
1831 remove insn setting up an equivalent pseudo.
1832
1833 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1834
1835 * config/rs6000/rs6000-protos.h
1836 (rs6000_secondary_memory_needed_mode): New prototype.
1837 * config/rs6000/rs6000.c: Include ira.h.
1838 (TARGET_LRA_P): Redefine.
1839 (rs6000_legitimate_offset_address_p): Call
1840 legitimate_constant_pool_address_p in strict mode for LRA.
1841 (rs6000_legitimate_address_p): Ditto.
1842 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
1843 (rs6000_emit_move): Add LRA version of code to generate load/store
1844 of SDmode values.
1845 (rs6000_secondary_memory_needed_mode): New.
1846 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
1847 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
1848 constants, memory, and FP registers.
1849 (rs6000_lra_p): New.
1850 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
1851 * config/rs6000/rs6000.opt (mlra): New option.
1852 * lra-spills.c (lra_final_code_change): Remove useless move insns.
1853
1854 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
1855
1856 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
1857 has_single_use () and not do the conversion if has_single_use ()
1858 returns false for the multiplication result.
1859
1860 2013-10-25 David Malcolm <dmalcolm@redhat.com>
1861
1862 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
1863 to "NODE", since this works on a "tree", not an
1864 "enum tree_code".
1865 (CONSTANT_CLASS_P): Likewise.
1866 (TYPE_P): Likewise.
1867 (DECL_P): Likewise.
1868 (INDIRECT_REF_P): Likewise.
1869 (REFERENCE_CLASS_P): Likewise.
1870 (COMPARISON_CLASS_P): Likewise.
1871 (UNARY_CLASS_P): Likewise.
1872 (BINARY_CLASS_P): Likewise.
1873 (STATEMENT_CLASS_P): Likewise.
1874 (VL_EXP_CLASS_P): Likewise.
1875 (EXPRESSION_CLASS_P): Likewise.
1876 (IS_TYPE_OR_DECL_P): Likewise.
1877
1878 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
1879
1880 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
1881 ADDR_EXPR in the defining statement.
1882
1883 2013-10-25 Richard Biener <rguenther@suse.de>
1884
1885 PR tree-optimization/58626
1886 * tree-loop-distribution.c (enum rdg_dep_type): Remove
1887 anti_dd, output_dd and input_dd.
1888 (struct rdg_edge): Remove level and relation members.
1889 (RDGE_LEVEL, RDGE_RELATION): Remove.
1890 (dot_rdg_1): Adjust.
1891 (create_rdg_edge_for_ddr): Remove.
1892 (create_rdg_edges_for_scalar): Adjust.
1893 (create_edge_for_control_dependence): Likewise.
1894 (create_rdg_edges): Split into ...
1895 (create_rdg_flow_edges): ... this
1896 (create_rdg_cd_edges): ... and this.
1897 (free_rdg): Adjust.
1898 (build_rdg): Likewise, do not compute data dependences or
1899 add edges for them.
1900 (pg_add_dependence_edges): New function.
1901 (pgcmp): Likewise.
1902 (distribute_loop): First apply all non-dependence based
1903 partition mergings. Then compute dependences between partitions
1904 and merge and order partitions according to them.
1905
1906 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
1907
1908 PR rtl-optimization/58831
1909 * alias.c (init_alias_analysis): At the beginning of each iteration,
1910 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
1911
1912 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
1913
1914 * recog.c (search_ofs): New static variable moved from...
1915 (peep2_find_free_register): ...here.
1916 (peephole2_optimize): Initialize it.
1917
1918 2013-10-25 Tobias Burnus <burnus@net-b.de>
1919
1920 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
1921
1922 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
1923
1924 * config/i386/i386.h (TARGET_MPX): New define.
1925 (TARGET_MPX_P): Ditto.
1926
1927 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
1928
1929 * config/i386/constraints.md (B): New.
1930 (Ti): New.
1931 (Tb): New.
1932 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
1933 * config/i386/i386-modes.def (BND32): New.
1934 (BND64): New.
1935 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
1936 * config/i386/i386.c (isa_opts): Add mmpx.
1937 (regclass_map): Add bound registers.
1938 (dbx_register_map): Likewise.
1939 (dbx64_register_map): Likewise.
1940 (svr4_dbx_register_map): Likewise.
1941 (PTA_MPX): New.
1942 (ix86_option_override_internal): Support MPX ISA.
1943 (ix86_conditional_register_usage): Support bound registers.
1944 (print_reg): Likewise.
1945 (ix86_code_end): Add MPX bnd prefix.
1946 (output_set_got): Likewise.
1947 (ix86_output_call_insn): Likewise.
1948 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
1949 (ix86_print_operand_punct_valid_p): Likewise.
1950 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
1951 UNSPEC_BNDMK_ADDR.
1952 (ix86_class_likely_spilled_p): Add bound regs support.
1953 (ix86_hard_regno_mode_ok): Likewise.
1954 (x86_order_regs_for_local_alloc): Likewise.
1955 (ix86_bnd_prefixed_insn_p): New.
1956 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
1957 (FIXED_REGISTERS): Add bound registers.
1958 (CALL_USED_REGISTERS): Likewise.
1959 (REG_ALLOC_ORDER): Likewise.
1960 (HARD_REGNO_NREGS): Likewise.
1961 (TARGET_MPX): New.
1962 (VALID_BND_REG_MODE): New.
1963 (FIRST_BND_REG): New.
1964 (LAST_BND_REG): New.
1965 (reg_class): Add BND_REGS.
1966 (REG_CLASS_NAMES): Likewise.
1967 (REG_CLASS_CONTENTS): Likewise.
1968 (BND_REGNO_P): New.
1969 (ANY_BND_REG_P): New.
1970 (BNDmode): New.
1971 (HI_REGISTER_NAMES): Add bound registers.
1972 * config/i386/i386.md (UNSPEC_BNDMK): New.
1973 (UNSPEC_BNDMK_ADDR): New.
1974 (UNSPEC_BNDSTX): New.
1975 (UNSPEC_BNDLDX): New.
1976 (UNSPEC_BNDLDX_ADDR): New.
1977 (UNSPEC_BNDCL): New.
1978 (UNSPEC_BNDCU): New.
1979 (UNSPEC_BNDCN): New.
1980 (UNSPEC_MPX_FENCE): New.
1981 (BND0_REG): New.
1982 (BND1_REG): New.
1983 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
1984 (length_immediate): Likewise.
1985 (prefix_0f): Likewise.
1986 (memory): Likewise.
1987 (prefix_rep): Check for bnd prefix.
1988 (length_nobnd): New.
1989 (length): Use length_nobnd if specified.
1990 (BND): New.
1991 (bnd_ptr): New.
1992 (BNDCHECK): New.
1993 (bndcheck): New.
1994 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
1995 (*jcc_2): Likewise.
1996 (jump): Likewise.
1997 (simple_return_internal): Likewise.
1998 (simple_return_pop_internal): Likewise.
1999 (*indirect_jump): Add MPX bnd prefix.
2000 (*tablejump_1): Likewise.
2001 (simple_return_internal_long): Likewise.
2002 (simple_return_indirect_internal): Likewise.
2003 (<mode>_mk): New.
2004 (*<mode>_mk): New.
2005 (mov<mode>): New.
2006 (*mov<mode>_internal_mpx): New.
2007 (<mode>_<bndcheck>): New.
2008 (*<mode>_<bndcheck>): New.
2009 (<mode>_ldx): New.
2010 (*<mode>_ldx): New.
2011 (<mode>_stx): New.
2012 (*<mode>_stx): New.
2013 * config/i386/predicates.md (lea_address_operand): Rename to...
2014 (address_no_seg_operand): ... this.
2015 (address_mpx_no_base_operand): New.
2016 (address_mpx_no_index_operand): New.
2017 (bnd_mem_operator): New.
2018 * config/i386/i386.opt (mmpx): New.
2019 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
2020 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
2021
2022 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2023
2024 * mode-classes.def (MODE_POINTER_BOUNDS): New.
2025 * tree.def (POINTER_BOUNDS_TYPE): New.
2026 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
2027 (POINTER_BOUNDS_MODE): New.
2028 (make_pointer_bounds_mode): New.
2029 * machmode.h (POINTER_BOUNDS_MODE_P): New.
2030 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
2031 (layout_type): Support POINTER_BOUNDS_TYPE.
2032 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
2033 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
2034 (type_contains_placeholder_1): Likewise.
2035 * tree.h (POINTER_BOUNDS_TYPE_P): New.
2036 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
2037 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
2038
2039 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2040
2041 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
2042 mode of address computation when calling memory_address_addr_space.
2043
2044 2013-08-24 Richard Henderson <rth@twiddle.net>
2045
2046 PR rtl/58542
2047 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
2048 instead of create_convert_operand_to.
2049 (maybe_emit_sync_lock_test_and_set): Likewise.
2050 (expand_atomic_compare_and_swap): Likewise.
2051 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
2052
2053 2013-08-24 Sriraman Tallam <tmsriram@google.com>
2054
2055 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
2056 Change param_is to use the struct and not the pointer to the struct.
2057
2058 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
2059
2060 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
2061 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2062 Move to targhooks.c.
2063 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
2064 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2065 Relocate from gimplify.c.
2066 * targhooks.h: Add 2 prototypes.
2067 * tree.h: Delete 2 prototypes.
2068
2069 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2070
2071 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
2072 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
2073
2074 2013-10-24 Cong Hou <congh@google.com>
2075
2076 * convert.c (convert_to_real): Guard those unsafe math function
2077 convertions with flag_unsafe_math_optimizations. Handle sqrt()
2078 specially.
2079
2080 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
2081
2082 PR ipa/58712
2083 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
2084 as argument.
2085 (cgraph_create_edge): Use the new argument.
2086 (cgraph_create_indirect_edge): Likewise.
2087
2088 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
2089
2090 * gcc/config/arc/arc.c (arc_ccfsm_post_advance): Also handle
2091 TYPE_UNCOND_BRANCH.
2092 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
2093 changes statep->state.
2094
2095 2013-10-24 Tobias Burnus <burnus@net-b.de>
2096
2097 PR other/33426
2098 * tree-cfg.c (replace_loop_annotate): New function.
2099 (execute_build_cfg): Call it.
2100 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
2101 * internal-fn.c (expand_ANNOTATE): New function.
2102 * internal-fn.def (ANNOTATE): Define as new internal function.
2103 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
2104 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
2105 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
2106 * doc/extend.texi (Pragmas): Document #pragma ivdep.
2107 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
2108
2109 2013-10-17 Ian Bolton <ian.bolton@arm.com>
2110 Marcus Shawcroft <marcus.shawcroft@arm.com>
2111
2112 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
2113 Special case reload SP+C into none GENERAL_REGS.
2114
2115 2013-10-24 Michael Matz <matz@suse.de>
2116
2117 * gengtype.c (is_file_equal): Check that files will be same length.
2118
2119 2013-10-25 Christian Bruel <christian.bruel@st.com>
2120
2121 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
2122 * gcc/config/sh/t-sh (sh-mem.o): New rule.
2123 * gcc/config/sh/sh-mem.cc (expand_block_move): Moved here.
2124 (sh_expand_cmpstr): New function.
2125 * gcc/config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
2126 * gcc/config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
2127 * gcc/config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
2128 (rotlhi3_8): Rename.
2129
2130 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2131
2132 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
2133 * Makefile.in (CXXDEPMODE): Assign and change users.
2134 (CCDEPMODE): Delete.
2135 * configure: Regenerate.
2136
2137 2013-10-23 David Malcolm <dmalcolm@redhat.com>
2138
2139 * gengtype-parse.c (require_without_advance): New.
2140 (type): For GTY-marked types that are not GTY((user)), parse any
2141 base classes, requiring them to be single-inheritance, and not
2142 be templates. For non-GTY-marked types and GTY((user)),
2143 continue to skip over any C++ inheritance specification.
2144 * gengtype-state.c (state_writer::write_state_struct_type):
2145 Write base class of type (if any).
2146 (read_state_struct_type): Read base class of type (if any).
2147 * gengtype.c (new_structure): Add a "base_class" parameter.
2148 (create_optional_field_): Update for new parameter to new_structure.
2149 (adjust_field_rtx_def): Likewise.
2150 (adjust_field_tree_exp): Likewise.
2151 * gengtype.h (struct type): Add "base_class" field to the s
2152 union field.
2153 (new_structure): Add "base" parameter.
2154
2155 2013-10-23 Sriraman Tallam <tmsriram@google.com>
2156
2157 PR target/57756
2158 * config/i386/i386.c (ix86_option_override_internal):
2159 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
2160 (ix86_valid_target_attribute_tree):
2161 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
2162 Change TARGET_SSE to TARGET_SSE_P (opts->...)
2163
2164 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2165
2166 * tree-ssa-loop.h: Remove include files.
2167 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2168 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
2169 * graphite-clast-to-gimple.c: Likewise.
2170 * graphite-scop-detection.c: Likewise.
2171 * graphite-sese-to-poly.c: Likewise.
2172 * ipa-inline-analysis.c: Likewise.
2173 * ipa-pure-const.c: Likewise.
2174 * loop-init.c: Likewise.
2175 * passes.c: Likewise.
2176 * predict.c: Likewise.
2177 * tree-cfg.c: Likewise.
2178 * tree-cfgcleanup.c: Likewise.
2179 * tree-chrec.c: Likewise.
2180 * tree-data-ref.c: Likewise.
2181 * tree-loop-distribution.c: Likewise.
2182 * tree-parloops.c: Likewise.
2183 * tree-predcom.c: Likewise.
2184 * tree-scalar-evolution.c: Likewise.
2185 * tree-ssa-address.c: Likewise.
2186 * tree-ssa.c: Likewise.
2187 * tree-ssa-dce.c: Likewise.
2188 * tree-ssa-loop.c: Likewise.
2189 * tree-ssa-loop-im.c: Likewise.
2190 * tree-ssa-loop-ivcanon.c: Likewise.
2191 * tree-ssa-loop-ivopts.c: Likewise.
2192 * tree-ssa-loop-manip.c: Likewise.
2193 * tree-ssa-loop-niter.c: Likewise.
2194 * tree-ssa-loop-prefetch.c: Likewise.
2195 * tree-ssa-loop-unswitch.c: Likewise.
2196 * tree-ssa-reassoc.c: Likewise.
2197 * tree-vect-data-refs.c: Likewise.
2198 * tree-vect-loop.c: Likewise.
2199 * tree-vect-loop-manip.c: Likewise.
2200 * tree-vectorizer.c: Likewise.
2201 * tree-vect-stmts.c: Likewise.
2202 * tree-vrp.c: Likewise.
2203
2204 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2205
2206 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
2207
2208 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2209
2210 PR tree-optimization/58775
2211 PR tree-optimization/58791
2212 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
2213 (insert_stmt_after): Rewritten, don't move the stmt, but really
2214 insert it.
2215 (get_stmt_uid_with_default): Remove.
2216 (build_and_add_sum): Use insert_stmt_after and
2217 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
2218 (update_range_test): Set uid on stmts added by
2219 force_gimple_operand_gsi. Don't immediately modify statements
2220 in inter-bb optimization, just update oe->op values.
2221 (optimize_range_tests): Return bool whether any changed have been made.
2222 (update_ops): New function.
2223 (struct inter_bb_range_test_entry): New type.
2224 (maybe_optimize_range_tests): Perform statement changes here.
2225 (not_dominated_by, appears_later_in_bb, get_def_stmt,
2226 ensure_ops_are_available): Remove.
2227 (find_insert_point): Rewritten.
2228 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
2229 return LHS of the (new resp. old) stmt. Don't call
2230 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
2231 instead of last, move new stmt at the right place.
2232 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
2233 (negate_value): Likewise. Set uids.
2234 (break_up_subtract_bb): Initialize uids.
2235 (reassociate_bb): Adjust rewrite_expr_tree caller.
2236 (do_reassoc): Don't call renumber_gimple_stmt_uids.
2237
2238 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
2239
2240 PR target/58838
2241 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
2242 TARGET_32BIT final condition.
2243 (mulsi3_internal2 and splitter): Same.
2244
2245 2013-10-23 Jeff Law <law@redhat.com>
2246
2247 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
2248 through joiner blocks with abnormal outgoing edges.
2249
2250 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
2251 Add parameter JOINERS, to allow/disallow threading through joiner
2252 blocks.
2253 (thread_block): New. Call thread_block_1.
2254 (mark_threaded_blocks): Remove code to filter out certain cases
2255 of threading through joiner blocks.
2256 (thread_through_all_blocks): Document how we can have a dangling
2257 edge AUX field and clear it.
2258
2259 2013-10-23 Ian Lance Taylor <iant@google.com>
2260
2261 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
2262 (C++ Dialect Options): Likewise.
2263 (Optimize Options): Likewise.
2264
2265 2013-10-23 Tom de Vries <tom@codesourcery.com>
2266
2267 PR tree-optimization/58805
2268 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
2269
2270 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2271
2272 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
2273 sequence based on get_range_info returned range.
2274
2275 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2276
2277 * tree-ssa.h: Remove all #include's
2278 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2279 * alias.c: Move required includes from tree-ssa.h.
2280 * asan.c: Likewise.
2281 * builtins.c: Likewise.
2282 * calls.c: Likewise.
2283 * cfgexpand.c: Likewise.
2284 * cfghooks.c: Likewise.
2285 * cfgloop.c: Likewise.
2286 * cfgloopmanip.c: Likewise.
2287 * cgraph.c: Likewise.
2288 * cgraphbuild.c: Likewise.
2289 * cgraphclones.c: Likewise.
2290 * cgraphunit.c: Likewise.
2291 * dse.c: Likewise.
2292 * except.c: Likewise.
2293 * expr.c: Likewise.
2294 * final.c: Likewise.
2295 * fold-const.c: Likewise.
2296 * ggc-page.c: Likewise.
2297 * gimple-builder.c: Likewise.
2298 * gimple-fold.c: Likewise.
2299 * gimple-iterator.c: Likewise.
2300 * gimple-low.c: Likewise.
2301 * gimple-pretty-print.c: Likewise.
2302 * gimple-ssa-strength-reduction.c: Likewise.
2303 * gimple-streamer-in.c: Likewise.
2304 * gimple-streamer-out.c: Likewise.
2305 * gimplify.c: Likewise.
2306 * graphite-blocking.c: Likewise.
2307 * graphite-clast-to-gimple.c: Likewise.
2308 * graphite-dependences.c: Likewise.
2309 * graphite-interchange.c: Likewise.
2310 * graphite-optimize-isl.c: Likewise.
2311 * graphite-poly.c: Likewise.
2312 * graphite-scop-detection.c: Likewise.
2313 * graphite-sese-to-poly.c: Likewise.
2314 * graphite.c: Likewise.
2315 * ipa-cp.c: Likewise.
2316 * ipa-inline-analysis.c: Likewise.
2317 * ipa-inline-transform.c: Likewise.
2318 * ipa-inline.c: Likewise.
2319 * ipa-prop.c: Likewise.
2320 * ipa-pure-const.c: Likewise.
2321 * ipa-reference.c: Likewise.
2322 * ipa-split.c: Likewise.
2323 * ipa-utils.c: Likewise.
2324 * loop-init.c: Likewise.
2325 * lto-cgraph.c: Likewise.
2326 * lto-section-in.c: Likewise.
2327 * lto-section-out.c: Likewise.
2328 * lto-streamer-in.c: Likewise.
2329 * lto-streamer-out.c: Likewise.
2330 * lto-streamer.c: Likewise.
2331 * omp-low.c: Likewise.
2332 * passes.c: Likewise.
2333 * predict.c: Likewise.
2334 * print-tree.c: Likewise.
2335 * profile.c: Likewise.
2336 * sese.c: Likewise.
2337 * targhooks.c: Likewise.
2338 * tracer.c: Likewise.
2339 * trans-mem.c: Likewise.
2340 * tree-call-cdce.c: Likewise.
2341 * tree-cfg.c: Likewise.
2342 * tree-cfgcleanup.c: Likewise.
2343 * tree-chrec.c: Likewise.
2344 * tree-complex.c: Likewise.
2345 * tree-data-ref.c: Likewise.
2346 * tree-dfa.c: Likewise.
2347 * tree-eh.c: Likewise.
2348 * tree-emutls.c: Likewise.
2349 * tree-if-conv.c: Likewise.
2350 * tree-inline.c: Likewise.
2351 * tree-into-ssa.c: Likewise.
2352 * tree-loop-distribution.c: Likewise.
2353 * tree-mudflap.c: Likewise.
2354 * tree-nested.c: Likewise.
2355 * tree-nrv.c: Likewise.
2356 * tree-object-size.c: Likewise.
2357 * tree-outof-ssa.c: Likewise.
2358 * tree-parloops.c: Likewise.
2359 * tree-phinodes.c: Likewise.
2360 * tree-predcom.c: Likewise.
2361 * tree-pretty-print.c: Likewise.
2362 * tree-profile.c: Likewise.
2363 * tree-scalar-evolution.c: Likewise.
2364 * tree-sra.c: Likewise.
2365 * tree-ssa-address.c: Likewise.
2366 * tree-ssa-alias.c: Likewise.
2367 * tree-ssa-ccp.c: Likewise.
2368 * tree-ssa-coalesce.c: Likewise.
2369 * tree-ssa-copy.c: Likewise.
2370 * tree-ssa-copyrename.c: Likewise.
2371 * tree-ssa-dce.c: Likewise.
2372 * tree-ssa-dom.c: Likewise.
2373 * tree-ssa-dse.c: Likewise.
2374 * tree-ssa-forwprop.c: Likewise.
2375 * tree-ssa-ifcombine.c: Likewise.
2376 * tree-ssa-live.c: Likewise.
2377 * tree-ssa-loop-ch.c: Likewise.
2378 * tree-ssa-loop-im.c: Likewise.
2379 * tree-ssa-loop-ivcanon.c: Likewise.
2380 * tree-ssa-loop-ivopts.c: Likewise.
2381 * tree-ssa-loop-manip.c: Likewise.
2382 * tree-ssa-loop-niter.c: Likewise.
2383 * tree-ssa-loop-prefetch.c: Likewise.
2384 * tree-ssa-loop-unswitch.c: Likewise.
2385 * tree-ssa-loop.c: Likewise.
2386 * tree-ssa-math-opts.c: Likewise.
2387 * tree-ssa-operands.c: Likewise.
2388 * tree-ssa-phiopt.c: Likewise.
2389 * tree-ssa-phiprop.c: Likewise.
2390 * tree-ssa-pre.c: Likewise.
2391 * tree-ssa-propagate.c: Likewise.
2392 * tree-ssa-reassoc.c: Likewise.
2393 * tree-ssa-sccvn.c: Likewise.
2394 * tree-ssa-sink.c: Likewise.
2395 * tree-ssa-strlen.c: Likewise.
2396 * tree-ssa-structalias.c: Likewise.
2397 * tree-ssa-tail-merge.c: Likewise.
2398 * tree-ssa-ter.c: Likewise.
2399 * tree-ssa-threadedge.c: Likewise.
2400 * tree-ssa-threadupdate.c: Likewise.
2401 * tree-ssa-uncprop.c: Likewise.
2402 * tree-ssa-uninit.c: Likewise.
2403 * tree-ssa.c: Likewise.
2404 * tree-ssanames.c: Likewise.
2405 * tree-stdarg.c: Likewise.
2406 * tree-streamer-in.c: Likewise.
2407 * tree-switch-conversion.c: Likewise.
2408 * tree-tailcall.c: Likewise.
2409 * tree-vect-data-refs.c: Likewise.
2410 * tree-vect-generic.c: Likewise.
2411 * tree-vect-loop-manip.c: Likewise.
2412 * tree-vect-loop.c: Likewise.
2413 * tree-vect-patterns.c: Likewise.
2414 * tree-vect-slp.c: Likewise.
2415 * tree-vect-stmts.c: Likewise.
2416 * tree-vectorizer.c: Likewise.
2417 * tree-vrp.c: Likewise.
2418 * tree.c: Likewise.
2419 * tsan.c: Likewise.
2420 * value-prof.c: Likewise.
2421 * var-tracking.c: Likewise.
2422 * varpool.c: Likewise.
2423 * vtable-verify.c: Likewise.
2424
2425 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2426
2427 * config/tilegx/tilegx.c: Include "tree.h".
2428
2429 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2430
2431 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
2432 the info, not after it.
2433 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
2434 dump_ssaname_info call.
2435 (pp_gimple_stmt_1): Adjust caller.
2436 (dump_phi_nodes): Likewise. Don't print "# " here.
2437
2438 2013-10-22 Jan Hubicka <jh@suse.cz>
2439
2440 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
2441 tuning flag.
2442 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
2443 * i386.c (expand_small_movmem_or_setmem): New function.
2444 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
2445 function.
2446 (alg_usable_p): Add support for value ranges; cleanup.
2447 (ix86_expand_set_or_movmem): Add support for misaligned moves.
2448
2449 2013-10-22 Sterling Augustine <saugustine@google.com>
2450
2451 * doc/invoke.texi: Document -ggnu-pubnames.
2452 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
2453 logic.
2454 * dwarf2out.c: Include gdb/gdb-index.h.
2455 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
2456 debug_generate_pub_sections.
2457 (is_java, output_pubtables, output_pubname): New functions.
2458 (include_pubname_in_output): Handle debug_generate_pub_sections at
2459 level 2.
2460 (size_of_pubnames): Use new local space_for_flags based on
2461 debug_generate_pub_sections.
2462 (output_pubnames): Unify pubnames and pubtypes output logic.
2463 Genericize comments. Call output_pubname.
2464 (dwarf2out_finish): Move logic to output_pubnames and call it.
2465
2466 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
2467
2468 PR target/58779
2469 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
2470 Remove CCCmode handling.
2471 <case LTU>: Return 'c' suffix for CCCmode.
2472 <case GEU>: Return 'nc' suffix for CCCmode.
2473 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
2474 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
2475 (*sub<mode>3_cc_overflow): Ditto.
2476 (*subsi3_zext_cc_overflow): Ditto.
2477
2478 2013-10-22 Steve Ellcey <sellcey@mips.com>
2479
2480 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
2481 (AND (NOT OP1) (NOT OP2)).
2482
2483 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
2484
2485 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2486 meaning of merge-high and merge-low masks for little endian; avoid
2487 use of vector-pack masks for little endian for mismatched modes.
2488
2489 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2490
2491 * config/tilepro/tilepro.c: Include "tree.h".
2492
2493 2013-10-22 Andreas Schwab <schwab@suse.de>
2494
2495 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
2496 with PRE_DEC.
2497
2498 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
2499
2500 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
2501 [Ed Smith-Rowland]): New entries.
2502 ([Stephen M. Webb]): Update.
2503
2504 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
2505
2506 * tree-ssa-ter.h: Remove duplicate copy of file contents.
2507
2508 2013-10-21 Marek Polacek <polacek@redhat.com>
2509
2510 PR middle-end/58809
2511 * fold-const.c (fold_range_test): Return 0 if the type is not
2512 an integral type.
2513
2514 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
2515
2516 * system.h: Move hwint.h include further down.
2517 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
2518 gcc_checking_asserts.
2519 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
2520
2521 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2522
2523 Fix volatile issues in optimize_bit_field_compare.
2524 * fold-const.c (optimize_bit_field_compare): Bail out if
2525 lvolatilep or rvolatilep.
2526
2527 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2528
2529 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
2530 and get_inner_reference returning different pmode for non-volatile
2531 bit-field members dependent on flag_strict_volatile_bitfields.
2532 * stor-layout.c (layout_decl): Remove special handling of
2533 flag_strict_volatile_bitfields.
2534 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
2535 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
2536
2537 2013-10-21 Paulo Matos <pmatos@broadcom.com>
2538
2539 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
2540 calculations.
2541
2542 2013-10-21 Jeff Law <law@redhat.com>
2543
2544 * tree-ssa-threadedge.c (thread_through_normal_block): New
2545 argument VISITED. Remove VISISTED as a local variable. When we
2546 have a threadable jump, verify the destination of the jump has not
2547 been visised.
2548 (thread_across_edge): Allocate VISITED bitmap once at function
2549 scope and use it throughout. Make sure to set appropriate bits in
2550 VISITED for E (start of jump thread path).
2551 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
2552 through a joiner if any edge on the path has a recorded jump thread.
2553
2554 2013-10-21 Ian Lance Taylor <iant@google.com>
2555
2556 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
2557 don't imply that attributes can solve all problems.
2558 (Directory Options): Fix typo.
2559
2560 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2561
2562 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
2563 extend and extend_add.
2564
2565 2013-10-21 Richard Biener <rguenther@suse.de>
2566
2567 PR tree-optimization/58794
2568 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
2569 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
2570
2571 2013-10-21 Richard Biener <rguenther@suse.de>
2572
2573 PR middle-end/58742
2574 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
2575 to (T) X for sign-changing conversions (or no conversion).
2576
2577 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
2578
2579 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
2580
2581 2013-10-20 Jan Hubicka <jh@suse.cz>
2582
2583 * config/i386/i386-tune.def: Add comment; organize into categories
2584
2585 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2586
2587 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
2588 argument. Update function comment.
2589 (expand_set_or_movmem_via_rep): New function combining
2590 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
2591 (expand_movmem_via_rep_mov): Remove.
2592 expand_setmem_via_rep_stos): Remove.
2593 (expand_movmem_epilogue): Update calls correspondingly.
2594 (expand_setmem_epilogue_via_loop): Likewise.
2595 (emit_memset): New.
2596 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
2597 (expand_set_or_movmem_prologue): New function combining
2598 expand_movmem_prologue and expand_setmem_prologue.
2599 (expand_movmem_prologue): Remove.
2600 (expand_setmem_prologue): Remove.
2601 (expand_set_or_movmem_constant_prologue): New function combining
2602 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
2603 (expand_constant_movmem_prologue): Remove.
2604 (expand_constant_setmem_prologue): Remove.
2605 (promote_duplicated_reg): Allow vector-const0 value.
2606 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
2607 and ix86_expand_setmem.
2608 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
2609 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
2610
2611 2013-10-21 Diego Novillo <dnovillo@google.com>
2612
2613 * asan.c: Include tree.h
2614 * bb-reorder.c: Likewise.
2615 * cfgcleanup.c: Likewise.
2616 * cfgloopmanip.c: Likewise.
2617 * data-streamer-in.c: Likewise.
2618 * data-streamer-out.c: Likewise.
2619 * data-streamer.c: Likewise.
2620 * dwarf2cfi.c: Likewise.
2621 * graphite-blocking.c: Likewise.
2622 * graphite-clast-to-gimple.c: Likewise.
2623 * graphite-dependences.c: Likewise.
2624 * graphite-interchange.c: Likewise.
2625 * graphite-optimize-isl.c: Likewise.
2626 * graphite-poly.c: Likewise.
2627 * graphite-scop-detection.c: Likewise.
2628 * graphite-sese-to-poly.c: Likewise.
2629 * graphite.c: Likewise.
2630 * ipa-devirt.c: Likewise.
2631 * ipa-profile.c: Likewise.
2632 * ipa.c: Likewise.
2633 * ira.c: Likewise.
2634 * loop-init.c: Likewise.
2635 * loop-unroll.c: Likewise.
2636 * lower-subreg.c: Likewise.
2637 * lto/lto-object.c: Likewise.
2638 * recog.c: Likewise.
2639 * reginfo.c: Likewise.
2640 * tree-loop-distribution.c: Likewise.
2641 * tree-parloops.c: Likewise.
2642 * tree-ssa-strlen.c: Likewise.
2643 * tree-streamer.c: Likewise.
2644 * value-prof.c: Likewise.
2645 * target-globals.c: Likewise.
2646 * expr.h: Include tree-core.h instead of tree.h.
2647 * gimple.h: Likewise.
2648 * ipa-prop.h: Likewise.
2649 * ipa-utils.h: Likewise.
2650 * lto-streamer.h: Likewise.
2651 * streamer-hooks.h: Likewise.
2652 * ipa-reference.h: Include cgraph.h instead of tree.h.
2653 * cgraph.h: Include basic-block.h instead of tree.h.
2654 * tree-streamer.h: Do not include tree.h.
2655 * genattrtab.c (write_header): Generate inclusion of
2656 tree.h.
2657 * genautomata.c (main): Likewise.
2658 * genemit.c: Likewise.
2659 * genopinit.c: Likewise.
2660 * genoutput.c (output_prologue): Likewise.
2661 * genpeep.c: Likewise.
2662
2663 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2664
2665 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
2666 little endian.
2667 (vec_unpacku_hi_v8hi): Likewise.
2668 (vec_unpacku_lo_v16qi): Likewise.
2669 (vec_unpacku_lo_v8hi): Likewise.
2670
2671 2013-10-20 Jan Hubicka <jh@suse.cz>
2672
2673 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
2674 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
2675 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
2676
2677 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
2678
2679 * config/mips/mips.h (ISA_HAS_WSBH): Define.
2680 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
2681 constants.
2682 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
2683
2684 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
2685
2686 PR target/58603
2687 * system.h: Undef m_slot.
2688
2689 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2690
2691 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
2692 all elements for both endian flavors.
2693
2694 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2695
2696 PR target/58792
2697 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
2698 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
2699 and SI_REG for 64bit SYSV ABI targets.
2700
2701 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2702
2703 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
2704 to multi_reg_return. Clarify that we are skipping USEs of multiple
2705 return registers. Use bool type where appropriate.
2706
2707 2013-10-18 Jan Hubicka <jh@suse.cz>
2708
2709 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
2710 for cold functions.
2711 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
2712 (X86_TUNE_PUSH_MEMORY): Likewise.
2713 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
2714 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
2715 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
2716 New.
2717 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
2718 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
2719 Remove.
2720 (ix86_option_override_internal): Update to use tune features instead
2721 of variables.
2722
2723 2013-10-18 Cong Hou <congh@google.com>
2724
2725 PR tree-optimization/58508
2726 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
2727 statement that contains data refs with zero-step.
2728
2729 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2730
2731 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
2732 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
2733 * gimple-low.c (gimple_check_call_arg,
2734 gimple_check_call_matching_types): Move to cgraph.c.
2735 * gimple-low.h: Remove prototype.
2736 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
2737 Relocate from gimple-low.c.
2738 * cgraph.h: Add prototype. Don't include basic-block.h.
2739 * gimplify.c: Add gimple-low to include list.
2740 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
2741 * tree-eh.c: Add gimple-low to include list.
2742 * tree-nested.c: Likewise.
2743 * cfgexpand.c: Add tree-ssa-address.h to include list.
2744 * expr.c: Likewise.
2745 * gimple-fold.c: Likewise.
2746 * gimple-ssa-strength-reduction.c: Likewise.
2747 * trans-mem.c: Likewise.
2748 * tree-mudflap.c: Likewise.
2749 * tree-ssa-loop-ivopts.c: Likewise.
2750 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
2751 (degenerate_phi_result): Move to tree-phinodes.c.
2752 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
2753 * tree-ssa-threadedge.c: Likewise.
2754 * tree-vrp.c: Likewise.
2755 * tree-phinodes.c (degenerate_phi_result): Relocate here.
2756 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
2757 * tree-phinodes.h (degenerate_phi_result): Add prototype.
2758 * tree-ssa-copy.c: Include tree-ssa-dom.h.
2759 * tree-ssa-forwprop.c: Likewise.
2760 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
2761 pass_data_cleanup_cfg_post_optimizing,
2762 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
2763 * tree-optimize.c: Delete File.
2764 * graphite.c: Include tree-cfgcleanup.h.
2765 * passes.c: Likewise.
2766 * tree-cfg.c: Likewise.
2767 * tree-profile.c: Likewise.
2768 * tree-ssa-dse.c: Likewise.
2769 * tree-ssa-loop-ivcanon.c: Likewise.
2770 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
2771 * tree-outof-ssa.c: Include sbitmap.h.
2772 * tree-ssa-live.c: Likewise.
2773 * tree-ssa-propagate.c: Likewise.
2774 * tree-ssa-structalias.c: Likewise.
2775 * tree-stdarg.c: Likewise.
2776 * Makefile.in (OBJS): Delete tree-optimize.o.
2777 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
2778 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
2779 * varasm.c: Include basic-block.h.
2780 * cfgloop.h: Include function.h instead of basic-block.h
2781 (bb_loop_depth): Move to cfgloop.c.
2782 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
2783
2784 2013-10-18 Teresa Johnson <tejohnson@google.com>
2785
2786 * predict.c (probably_never_executed): Compare frequency-based
2787 count to number of training runs.
2788 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
2789
2790 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2791
2792 * config/arm/arm.c (cortexa9_extra_costs): New table.
2793 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
2794
2795 2013-10-18 Jeff Law <law@redhat.com>
2796
2797 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
2798
2799 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
2800 (register_jump_thread): Add "registered_jump_thread" debug
2801 counter support.
2802 * dbgcnt.def (registered_jump_thread): New debug counter.
2803
2804 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2805
2806 * config/rs6000/rs6000.c: Include cgraph.h.
2807
2808 2013-10-18 Teresa Johnson <tejohnson@google.com>
2809
2810 * tree-ssa-tail-merge.c (replace_block_by): Update edge
2811 weights during merging.
2812
2813 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2814
2815 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
2816 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
2817 * tree-ssa-operands.h: Remove prototype.
2818 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
2819 * gimple.c (virtual_operand_p): Relocate from gimple.c.
2820 * gimple.h: Add prototype.
2821 * gimple-ssa.h: Include tree-ssa-operands.h.
2822 * tree-dump.c: Add tree-cfg.h to include list.
2823 * tree-ssa-alias.c: Add ipa-reference.h to include list.
2824 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
2825 * config/i386/i386.c: Don't include tree-flow.h.
2826 * config/rs6000/rs6000.c: Likewise.
2827
2828 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2829
2830 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
2831
2832 2013-10-18 Richard Biener <rguenther@suse.de>
2833
2834 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
2835 or TYPE_UNSIGNED of integral types.
2836 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
2837 NULL_TREE for zero-precision integral types.
2838
2839 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
2840
2841 * config/aarch64/arm_neon.h
2842 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
2843
2844 2013-10-17 Sriraman Tallam <tmsriram@google.com>
2845
2846 PR target/57756
2847 * opth-gen.awk: Define target_flags_explicit.
2848
2849 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2850
2851 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
2852 fields to the reg_addr array that describes the valid addressing
2853 mode for any register, general purpose registers, floating point
2854 registers, and Altivec registers.
2855 (FIRST_RELOAD_REG_CLASS): Likewise.
2856 (LAST_RELOAD_REG_CLASS): Likewise.
2857 (struct reload_reg_map_type): Likewise.
2858 (reload_reg_map_type): Likewise.
2859 (RELOAD_REG_VALID): Likewise.
2860 (RELOAD_REG_MULTIPLE): Likewise.
2861 (RELOAD_REG_INDEXED): Likewise.
2862 (RELOAD_REG_OFFSET): Likewise.
2863 (RELOAD_REG_PRE_INCDEC): Likewise.
2864 (RELOAD_REG_PRE_MODIFY): Likewise.
2865 (reg_addr): Likewise.
2866 (mode_supports_pre_incdec_p): New helper functions to say whether
2867 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
2868 (mode_supports_pre_modify_p): Likewise.
2869 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
2870 print the valid address mode bits for each mode.
2871 (rs6000_debug_print_mode): Likewise.
2872 (rs6000_debug_reg_global): Likewise.
2873 (rs6000_setup_reg_addr_masks): New function to set up the address
2874 mask bits for each type.
2875 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
2876 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
2877 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
2878 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
2879 PRE_MODIFY are supported.
2880 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
2881 registers, instead of {src,dest}_av_p.
2882 (rs6000_print_options_internal): Tweak the debug output slightly.
2883
2884 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
2885
2886 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
2887 isa attribute.
2888
2889 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
2890
2891 * tree-flow.h (struct omp_region): Move to omp-low.c.
2892 Remove omp_ prototypes and variables.
2893 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
2894 (copy_var_decl): Relocate prototype from tree-flow.h.
2895 * gimple.c (copy_var_decl): Relocate from omp-low.c.
2896 * tree.h: Move prototype to omp-low.h.
2897 * omp-low.h: New File. Relocate prototypes here.
2898 * omp-low.c (struct omp_region): Make local here.
2899 (root_omp_region): Make static.
2900 (copy_var_decl) Move to gimple.c.
2901 (new_omp_region): Make static.
2902 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
2903 * tree-cfg.c: Include omp-low.h.
2904 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
2905 * gimplify.c: Include omp-low.h.
2906 * tree-parloops.c: Likewise.
2907
2908 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
2909
2910 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
2911 operands are in memory, prefer to force non-matched operand 1 to
2912 the register.
2913
2914 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2915
2916 PR target/58673
2917 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
2918 restrict TImode addresses to single indirect registers if both
2919 -mquad-memory and -mvsx-timode are used.
2920 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
2921 we should emit load/store quad. Remove using %y for quad memory
2922 addresses.
2923
2924 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
2925 constraints to allow load/store quad on machines where TImode is
2926 not allowed in VSX registers. Use 'n' instead of 'F' constraint
2927 for TImode to load integer constants.
2928
2929 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2930
2931 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
2932
2933 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
2934
2935 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
2936 handling of STACK_REG.
2937
2938 2013-10-17 Richard Biener <rguenther@suse.de>
2939
2940 PR tree-optimization/58143
2941 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
2942 New function.
2943 (rewrite_to_defined_overflow): Likewise.
2944 (move_computations_dom_walker::before_dom): Rewrite stmts
2945 with undefined signed overflow that are not always executed
2946 into unsigned arithmetic.
2947
2948 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
2949
2950 PR target/57756
2951 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
2952 explicit isa flag to be an options variable, instead of using
2953 global_options_set. Remove define from rs6000.h.
2954 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
2955
2956 * config/rs6000/rs6000.c (rs6000_option_override_internal):
2957 Initialize rs6000_isa_flags_explicit.
2958 (rs6000_function_specific_save): Add gcc_options* parameter, so
2959 that the powerpc builds after the 2013-10-15 changes.
2960 (rs6000_function_specific_restore): Likewise.
2961
2962 2013-10-16 DJ Delorie <dj@redhat.com>
2963
2964 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
2965 op is a REG before checking REGNO.
2966 (rl78_alloc_physical_registers): Verify pattern is a SET before
2967 checking SET_SRC.
2968
2969 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2970
2971 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
2972 endianness.
2973 (vec_unpacks_lo_v4sf): Likewise.
2974 (vec_unpacks_float_hi_v4si): Likewise.
2975 (vec_unpacks_float_lo_v4si): Likewise.
2976 (vec_unpacku_float_hi_v4si): Likewise.
2977 (vec_unpacku_float_lo_v4si): Likewise.
2978
2979 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2980
2981 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
2982 (vsx_concat_v2sf): Likewise.
2983
2984 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
2985
2986 * config/aarch64/aarch64.md
2987 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
2988
2989 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
2990
2991 PR tree-optimization/58697
2992 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
2993 estimated_loop_iterations_int.
2994 (max_stmt_executions_int): Call get_max_loop_iterations_int.
2995 (get_max_loop_iterations_int): New. HWINT version of
2996 get_max_loop_iterations.
2997 * cfgloop.h: Add prototypes.
2998 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
2999 * loop-unroll.c (decide_peel_once_rolling): Call
3000 get_estimated_loop_iterations_int.
3001 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
3002 * tree-ssa-loop-niter.h: Tweak prototypes.
3003
3004 2013-10-16 David Malcolm <dmalcolm@redhat.com>
3005
3006 * gengtype-parse.c (struct_field_seq): Ignore access-control
3007 keywords ("public:" etc).
3008
3009 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
3010
3011 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
3012 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
3013
3014 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
3015
3016 * config/arm/arm.opt (mlra): New option.
3017 * config/arm/arm.c (arm_lra_p): New function.
3018 (TARGET_LRA_P): Define.
3019
3020 2013-10-16 Paulo Matos <pmatos@broadcom.com>
3021
3022 * tree-core.h (tree_code_name): Remove.
3023 * tree.h (get_tree_code_name): New prototype.
3024 * tree.c (tree_code_name): Make static.
3025 (get_tree_code_name): New function.
3026 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
3027 tree_class_check_failed, tree_range_check_failed,
3028 tree_not_class_check_failed, omp_clause_check_failed,
3029 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
3030 wrapper get_tree_code_name instead of calling tree_code_name directly.
3031 * tree-vrp.c (dump_asserts_for): Likewise.
3032 * tree-dump.c (dequeue_and_dump): Likewise.
3033 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
3034 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
3035 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
3036 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
3037 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
3038 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
3039 dump_gimple_omp_for): Likewise.
3040 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
3041 * tree-ssa-pre.c (print_pre_expr): Likewise.
3042 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
3043 * print-tree.c (print_node_brief, print_node): Likewise.
3044 * gimple.c (gimple_check_failed): Likewise.
3045 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
3046 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
3047 * config/mep/mep.c (mep_validate_vliw): Likewise.
3048 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
3049 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
3050
3051 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
3052
3053 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
3054 for AMD bdver3.
3055
3056 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
3057
3058 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
3059 (MULTILIB_MATCHES): Add multilib for -march=v8.
3060
3061 2013-10-15 Sriraman Tallam <tmsriram@google.com>
3062
3063 PR target/57756
3064 * optc-save-gen.awk: Add extra parameter to the save and restore
3065 target calls.
3066 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
3067 * tree.c (build_optimization_node): New parameter. Add extra parameter
3068 to call to cl_optimization_save.
3069 (build_target_option_node): New parameter. Add extra parameter
3070 to call to cl_target_option_save.
3071 * tree.h (build_optimization_node): New parameter.
3072 (build_target_option_node): New parameter.
3073 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
3074 build_optimization_node and build_target_option_node.
3075 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
3076 (handle_pragma_push_options): Ditto.
3077 * toplev.c (process_options): Ditto.
3078 * opts.c (init_options_struct): Check for opts_set non-null.
3079 * target.def (target_option.save): New parameter.
3080 (target_option.restore): New parameter.
3081 * tm.texi: Generate.
3082 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
3083 (ix86_pragma_target_parse): Ditto.
3084 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
3085 parameters.
3086 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
3087 to build_optimization_node and build_target_option_node.
3088 (rs6000_valid_attribute_p): Ditto.
3089 (rs6000_pragma_target_parse): Ditto.
3090 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
3091 data.
3092 * config/i386/i386.h:
3093 TARGET_64BIT_P: New Macro
3094 TARGET_MMX_P: New Macro.
3095 TARGET_3DNOW_P: New Macro.
3096 TARGET_3DNOW_A_P: New Macro.
3097 TARGET_SSE_P: New Macro.
3098 TARGET_SSE2_P: New Macro.
3099 TARGET_SSE3_P: New Macro.
3100 TARGET_SSSE3_P: New Macro.
3101 TARGET_SSE4_1_P: New Macro.
3102 TARGET_SSE4_2_P: New Macro.
3103 TARGET_AVX_P: New Macro.
3104 TARGET_AVX2_P: New Macro.
3105 TARGET_AVX512F_P: New Macro.
3106 TARGET_AVX512PF_P: New Macro.
3107 TARGET_AVX512ER_P: New Macro.
3108 TARGET_AVX512CD_P: New Macro.
3109 TARGET_FMA_P: New Macro.
3110 TARGET_SSE4A_P: New Macro.
3111 TARGET_FMA4_P: New Macro.
3112 TARGET_XOP_P: New Macro.
3113 TARGET_LWP_P: New Macro.
3114 TARGET_ABM_P: New Macro.
3115 TARGET_BMI_P: New Macro.
3116 TARGET_BMI2_P: New Macro.
3117 TARGET_LZCNT_P: New Macro.
3118 TARGET_TBM_P: New Macro.
3119 TARGET_POPCNT_P: New Macro.
3120 TARGET_SAHF_P: New Macro.
3121 TARGET_MOVBE_P: New Macro.
3122 TARGET_CRC32_P: New Macro.
3123 TARGET_AES_P: New Macro.
3124 TARGET_PCLMUL_P: New Macro.
3125 TARGET_CMPXCHG16B_P: New Macro.
3126 TARGET_FSGSBASE_P: New Macro.
3127 TARGET_RDRND_P: New Macro.
3128 TARGET_F16C_P: New Macro.
3129 TARGET_RTM_P: New Macro.
3130 TARGET_HLE_P: New Macro.
3131 TARGET_RDSEED_P: New Macro.
3132 TARGET_PRFCHW_P: New Macro.
3133 TARGET_ADX_P: New Macro.
3134 TARGET_FXSR_P: New Macro.
3135 TARGET_XSAVE_P: New Macro.
3136 TARGET_XSAVEOPT_P: New Macro.
3137 TARGET_LP64_P: New Macro.
3138 TARGET_X32_P: New Macro.
3139 TARGET_FPMATH_DEFAULT_P: New Macro.
3140 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
3141 * config/i386/i386.c (ix86_option_override_internal): New parameters.
3142 opts and opts_set.
3143 Change ix86_tune_string to access opts->x_ix86_tune_string.
3144 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
3145 Change ix86_arch_string to access opts->x_ix86_arch_string.
3146 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
3147 Change ix86_pmode to access opts->x_ix86_pmode.
3148 Change ix86_abi to access opts->x_ix86_abi.
3149 Change ix86_cmodel to access opts->x_ix86_cmodel.
3150 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
3151 Change ix86_isa_flags_explicit to access
3152 opts->x_ix86_isa_flags_explicit.
3153 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
3154 Change ix86_regparm to access opts->x_ix86_regparm.
3155 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
3156 Change ix86_preferred_stack_boundary_arg to access
3157 opts->x_ix86_preferred_stack_boundary_arg.
3158 Change ix86_force_align_arg_pointer to access
3159 opts->x_ix86_force_align_arg_pointer.
3160 Change ix86_incoming_stack_boundar_arg to access
3161 opts->x_ix86_incoming_stack_boundar_arg.
3162 Change ix86_fpmath to access opts->x_ix86_fpmath.
3163 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
3164 Change ix86_recip_name to access opts->x_ix86_recip_name.
3165 Change ix86_stack_protector_guard to access
3166 opts->x_ix86_stack_protector_guard.
3167 Change ix86_tune_memcpy_strategy to access
3168 opts->x_ix86_tune_memcpy_strategy.
3169 Change ix86_tune_memset_strategy to access
3170 opts->x_ix86_tune_memset_strategy.
3171 Change global_options to access opts.
3172 Change global_options_set to access opts_set.
3173 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
3174 Change TARGET_MMX to TARGET_MMX_P (opts->...).
3175 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
3176 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
3177 Change TARGET_SSE to TARGET_SSE_P (opts->...).
3178 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
3179 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
3180 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
3181 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
3182 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
3183 Change TARGET_AVX to TARGET_AVX_P (opts->...).
3184 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
3185 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
3186 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
3187 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
3188 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
3189 Change TARGET_FMA to TARGET_FMA_P (opts->...).
3190 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
3191 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
3192 Change TARGET_XOP to TARGET_XOP_P (opts->...).
3193 Change TARGET_LWP to TARGET_LWP_P (opts->...).
3194 Change TARGET_ABM to TARGET_ABM_P (opts->...).
3195 Change TARGET_BMI to TARGET_BMI_P (opts->...).
3196 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
3197 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
3198 Change TARGET_TBM to TARGET_TBM_P (opts->...).
3199 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
3200 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
3201 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
3202 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
3203 Change TARGET_AES to TARGET_AES_P (opts->...).
3204 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
3205 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
3206 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
3207 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
3208 Change TARGET_F16C to TARGET_F16C_P (opts->...).
3209 Change TARGET_RTM to TARGET_RTM_P (opts->...).
3210 Change TARGET_HLE to TARGET_HLE_P (opts->...).
3211 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
3212 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
3213 Change TARGET_ADX to TARGET_ADX_P (opts->...).
3214 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
3215 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
3216 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
3217 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
3218 Change TARGET_X32 to TARGET_X32_P (opts->...).
3219 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
3220 Change TARGET_FLOAT_RETURNS_IN_80387 to
3221 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
3222 (ix86_function_specific_save): New parameter. Use opts-> fields
3223 to replace global fields.
3224 (ix86_function_specific_restore): Ditto.
3225 (ix86_valid_target_attribute_inner_p): New parameters.
3226 Fix recursive call.
3227 Fix call to ix86_handle_option and set_option.
3228 (ix86_valid_target_attribute_tree): New parameters.
3229 Change global_options to access opts.
3230 Change global_options_set to access opts_set.
3231 Fix call to ix86_valid_target_attribute_inner_p.
3232 Change ix86_tune_string to access opts->x_ix86_tune_string.
3233 Change ix86_arch_string to access opts->x_ix86_arch_string.
3234 Change ix86_fpmath to access opts->x_ix86_fpmath
3235 Fix call to ix86_option_override_internal.
3236 Fix call to ix86_add_new_builtins.
3237 Fix calls to build_optimization_node and build_target_option_node.
3238 (ix86_valid_target_attribute_p): Remove access to global_options.
3239 Use new gcc_options structure func_options.
3240 Fix call to ix86_valid_target_attribute_tree.
3241 Fix call to build_optimization_node.
3242 (get_builtin_code_for_version): Fix call to
3243 ix86_valid_target_attribute_tree.
3244
3245 2013-10-15 David Malcolm <dmalcolm@redhat.com>
3246
3247 * Makefile.in (PICFLAG): New.
3248 (enable_host_shared): New.
3249 (INTERNAL_CFLAGS): Use PICFLAG.
3250 (LIBIBERTY): Use pic build of libiberty.a if configured with
3251 --enable-host-shared.
3252 * configure.ac: Add --enable-host-shared, setting up new
3253 PICFLAG variable.
3254 * configure: Regenerate.
3255 * doc/install.texi (--enable-shared): Add note contrasting it with ...
3256 (--enable-host-shared): New option.
3257
3258 2013-10-15 Richard Biener <rguenther@suse.de>
3259
3260 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
3261 for built-in functions.
3262
3263 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
3264
3265 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
3266 (optimize_range_tests_1): New function,
3267 extracted from optimize_range_tests.
3268 (optimize_range_tests_xor): Similarly.
3269 (optimize_range_tests_diff): New function.
3270 (optimize_range_tests): Use optimize_range_tests_1.
3271
3272 2013-10-15 Cong Hou <congh@google.com>
3273
3274 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
3275 requirement of the reduction pattern so that one operand of the
3276 reduction operation can come from outside of the loop.
3277
3278 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3279
3280 * config/arm/neon-schedgen.ml: Remove.
3281 * config/arm/cortex-a9-neon.md: Remove comment regarding
3282 neon-schedgen.ml.
3283
3284 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3285
3286 * config/arm/types: Remove old neon types.
3287
3288 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3289
3290 * config/arm/cortex-a7.md
3291 (cortex_a7_neon_type): New.
3292 (cortex_a7_neon_mul): Update for new types.
3293 (cortex_a7_neon_mla): Likewise.
3294 (cortex_a7_neon): Likewise.
3295
3296 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3297
3298 * config/arm/cortex-a15-neon.md
3299 (cortex_a15_neon_type): New,
3300
3301 (cortex_a15_neon_int_1): Remove.
3302 (cortex_a15_neon_int_2): Likewise.
3303 (cortex_a15_neon_int_3): Likewise.
3304 (cortex_a15_neon_int_4): Likewise.
3305 (cortex_a15_neon_int_5): Likewise.
3306 (cortex_a15_neon_vqneg_vqabs): Likewise.
3307 (cortex_a15_neon_vmov): Likewise.
3308 (cortex_a15_neon_vaba): Likewise.
3309 (cortex_a15_neon_vaba_qqq): Likewise.
3310 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3311 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3312 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3313 Likewise.
3314 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3315 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3316 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
3317 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3318 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3319 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3320 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3321 (cortex_a15_neon_shift_1): Likewise.
3322 (cortex_a15_neon_shift_2): Likewise.
3323 (cortex_a15_neon_shift_3): Likewise.
3324 (cortex_a15_neon_vshl_ddd): Likewise.
3325 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3326 (cortex_a15_neon_vsra_vrsra): Likewise.
3327 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3328 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3329 (cortex_a15_neon_bp_3cycle): Likewise.
3330 (cortex_a15_neon_ldm_2): Likewise.
3331 (cortex_a15_neon_stm_2): Likewise.
3332 (cortex_a15_neon_mcr): Likewise.
3333 (cortex_a15_neon_mrc): Likewise.
3334 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3335 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3336 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3337 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3338 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3339 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3340 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3341 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3342 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3343 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3344 (cortex_a15_neon_bp_simple): Likewise.
3345 (cortex_a15_neon_bp_2cycle): Likewise.
3346 (cortex_a15_neon_bp_3cycle): Likewise.
3347 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3348 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3349 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3350 (cortex_a15_neon_vld2_4_regs): Likewise.
3351 (cortex_a15_neon_vld3_vld4): Likewise.
3352 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3353 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3354 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
3355 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
3356 (cortex_a15_neon_vst3_vst4): Rename to...
3357 (cortex_a15_neon_vst4): This, update for new attributes.
3358 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
3359 (cortex_a15_neon_vld3_vld4_lane): Likewise.
3360 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3361 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3362 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3363 (cortex_a15_neon_ldm_2): Likewise.
3364 (cortex_a15_neon_stm_2): Likewise.
3365 (cortex_a15_neon_mcr): Likewise.
3366 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3367 (cortex_a15_neon_mrc): Likewise.
3368 (cortex_a15_neon_mrrc): Likewise.
3369
3370 (cortex_a15_neon_abd): New.
3371 (cortex_a15_neon_abd_q): Likewise.
3372 (cortex_a15_neon_aba): Likewise.
3373 (cortex_a15_neon_aba_q): Likewise.
3374 (cortex_a15_neon_acc): Likewise.
3375 (cortex_a15_neon_acc_q): Likewise.
3376 (cortex_a15_neon_arith_basic): Likewise.
3377 (cortex_a15_neon_arith_complex): Likewise.
3378 (cortex_a15_neon_multiply): Likewise.
3379 (cortex_a15_neon_multiply_q): Likewise.
3380 (cortex_a15_neon_mla): Likewise.
3381 (cortex_a15_neon_mla_q): Likewise.
3382 (cortex_a15_neon_sat_mla_long): Likewise.
3383 (cortex_a15_neon_shift_acc): Likewise.
3384 (cortex_a15_neon_shift_imm_basic): Likewise.
3385 (cortex_a15_neon_shift_imm_complex): Likewise.
3386 (cortex_a15_neon_shift_reg_basic): Likewise.
3387 (cortex_a15_neon_shift_reg_basic_q): Likewise.
3388 (cortex_a15_neon_shift_reg_complex): Likewise.
3389 (cortex_a15_neon_shift_reg_complex_q): Likewise.
3390 (cortex_a15_neon_fp_negabs): Likewise
3391 (cortex_a15_neon_fp_arith): Likewise
3392 (cortex_a15_neon_fp_arith_q): Likewise
3393 (cortex_a15_neon_fp_cvt_int): Likewise
3394 (cortex_a15_neon_fp_cvt_int_q): Likewise
3395 (cortex_a15_neon_fp_cvt_16): Likewise
3396 (cortex_a15_neon_fp_mul): Likewise
3397 (cortex_a15_neon_fp_mul_q): Likewise
3398 (cortex_a15_neon_fp_mla): Likewise
3399 (cortex_a15_neon_fp_mla_q): Likewise
3400 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
3401 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
3402 (cortex_a15_neon_bitops): Likewise.
3403 (cortex_a15_neon_bitops_q): Likewise.
3404 (cortex_a15_neon_from_gp): Likewise.
3405 (cortex_a15_neon_from_gp_q): Likewise.
3406 (cortex_a15_neon_tbl3_tbl4): Likewise.
3407 (cortex_a15_neon_zip_q): Likewise.
3408 (cortex_a15_neon_to_gp): Likewise.
3409 (cortex_a15_neon_load_a): Likewise.
3410 (cortex_a15_neon_load_b): Likewise.
3411 (cortex_a15_neon_load_c): Likewise.
3412 (cortex_a15_neon_load_d): Likewise.
3413 (cortex_a15_neon_load_e): Likewise.
3414 (cortex_a15_neon_load_f): Likewise.
3415 (cortex_a15_neon_store_a): Likewise.
3416 (cortex_a15_neon_store_b): Likewise.
3417 (cortex_a15_neon_store_c): Likewise.
3418 (cortex_a15_neon_store_d): Likewise.
3419 (cortex_a15_neon_store_e): Likewise.
3420 (cortex_a15_neon_store_f): Likewise.
3421 (cortex_a15_neon_store_g): Likewise.
3422 (cortex_a15_neon_store_h): Likewise.
3423 (cortex_a15_vfp_to_from_gp): Likewise.
3424
3425 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3426
3427 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
3428
3429 (cortex_a9_neon_vshl_ddd): Remove.
3430 (cortex_a9_neon_vst3_vst4): Likewise.
3431 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3432
3433 (cortex_a9_neon_bit_ops_q): New.
3434
3435 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
3436 (cortex_a9_neon_int_2): Likewise.
3437 (cortex_a9_neon_int_3): Likewise.
3438 (cortex_a9_neon_int_4): Likewise.
3439 (cortex_a9_neon_int_5): Likewise.
3440 (cortex_a9_neon_vqneg_vqabs): Likewise.
3441 (cortex_a9_neon_vmov): Likewise.
3442 (cortex_a9_neon_vaba): Likewise.
3443 (cortex_a9_neon_vaba_qqq): Likewise.
3444 (cortex_a9_neon_shift_1): Likewise.
3445 (cortex_a9_neon_shift_2): Likewise.
3446 (cortex_a9_neon_shift_3): Likewise.
3447 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3448 (cortex_a9_neon_vsra_vrsra): Likewise.
3449 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3450 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3451 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3452 Likewise.
3453 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3454 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3455 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3456 Likewise.
3457 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3458 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3459 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3460 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3461 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3462 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3463 (cortex_a9_neon_fp_vsum): Likewise.
3464 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3465 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3466 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3467 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3468 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3469 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3470 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3471 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3472 (cortex_a9_neon_bp_simple): Likewise.
3473 (cortex_a9_neon_bp_2cycle): Likewise.
3474 (cortex_a9_neon_bp_3cycle): Likewise.
3475 (cortex_a9_neon_ldr): Likewise.
3476 (cortex_a9_neon_str): Likewise.
3477 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3478 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3479 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3480 (cortex_a9_neon_vld2_4_regs): Likewise.
3481 (cortex_a9_neon_vld3_vld4): Likewise.
3482 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3483 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3484 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3485 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3486 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3487 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3488 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3489 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3490 (cortex_a9_neon_mcr): Likewise.
3491 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3492 (cortex_a9_neon_mrc): Likewise.
3493 (cortex_a9_neon_mrrc): Likewise.
3494
3495 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3496
3497 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
3498
3499 (cortex_a8_neon_vshl_ddd): Remove.
3500 (cortex_a8_neon_vst3_vst4): Likewise.
3501 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3502
3503 (cortex_a8_neon_bit_ops_q): New.
3504
3505 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
3506 (cortex_a8_neon_int_2): Likewise..
3507 (cortex_a8_neon_int_3): Likewise.
3508 (cortex_a8_neon_int_5): Likewise.
3509 (cortex_a8_neon_vqneg_vqabs): Likewise.
3510 (cortex_a8_neon_int_4): Likewise.
3511 (cortex_a8_neon_vaba): Likewise.
3512 (cortex_a8_neon_vaba_qqq): Likewise.
3513 (cortex_a8_neon_shift_1): Likewise.
3514 (cortex_a8_neon_shift_2): Likewise.
3515 (cortex_a8_neon_shift_3): Likewise.
3516 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3517 (cortex_a8_neon_vsra_vrsra): Likewise.
3518 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3519 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3520 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3521 Likewise.
3522 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3523 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3524 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3525 Likewise.
3526 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3527 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3528 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3529 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3530 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3531 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3532 (cortex_a8_neon_fp_vsum): Likewise.
3533 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3534 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3535 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3536 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3537 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3538 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3539 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3540 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3541 (cortex_a8_neon_bp_simple): Likewise.
3542 (cortex_a8_neon_bp_2cycle): Likewise.
3543 (cortex_a8_neon_bp_3cycle): Likewise.
3544 (cortex_a8_neon_ldr): Likewise.
3545 (cortex_a8_neon_str): Likewise.
3546 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3547 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3548 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3549 (cortex_a8_neon_vld2_4_regs): Likewise.
3550 (cortex_a8_neon_vld3_vld4): Likewise.
3551 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3552 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3553 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3554 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3555 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3556 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3557 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3558 (cortex_a8_neon_mcr): Likewise.
3559 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3560 (cortex_a8_neon_mrc): Likewise.
3561 (cortex_a8_neon_mrrc): Likewise.
3562
3563 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3564
3565 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
3566 (fp): New.
3567 * config/aarch64/aarch64-simd.md (neon_type): Remove.
3568 (aarch64_simd_dup<mode>): Add "type" attribute.
3569 (aarch64_dup_lane<mode>): Likewise.
3570 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
3571 (*aarch64_simd_mov<mode>): Likewise.
3572 (aarch64_simd_mov_from_<mode>low): Likewise.
3573 (aarch64_simd_mov_from_<mode>high): Likewise.
3574 (orn<mode>3): Likewise.
3575 (bic<mode>3): Likewise.
3576 (add<mode>3): Likewise.
3577 (sub<mode>3): Likewise.
3578 (mul<mode>3): Likewise.
3579 (*aarch64_mul3_elt<mode>): Likewise.
3580 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
3581 (*aarch64_mul3_elt_to_128df): Likewise.
3582 (*aarch64_mul3_elt_to_64v2df): Likewise.
3583 (neg<mode>2): Likewise.
3584 (abs<mode>2): Likewise.
3585 (abd<mode>_3): Likewise.
3586 (aba<mode>_3): Likewise.
3587 (fabd<mode>_3): Likewise.
3588 (*fabd_scalar<mode>3): Likewise.
3589 (and<mode>3): Likewise.
3590 (ior<mode>3): Likewise.
3591 (xor<mode>3): Likewise.
3592 (one_cmpl<mode>2): Likewise.
3593 (aarch64_simd_vec_set<mode>): Likewise.
3594 (aarch64_simd_lshr<mode>): Likewise.
3595 (aarch64_simd_ashr<mode>): Likewise.
3596 (aarch64_simd_imm_shl<mode>): Likewise.
3597 (aarch64_simd_reg_sshl<mode): Likewise.
3598 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
3599 (aarch64_simd_reg_shl<mode>_signed): Likewise.
3600 (aarch64_simd_vec_setv2di): Likewise.
3601 (aarch64_simd_vec_set<mode>): Likewise.
3602 (aarch64_mla<mode>): Likewise.
3603 (*aarch64_mla_elt<mode>): Likewise.
3604 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
3605 (aarch64_mls<mode>): Likewise.
3606 (*aarch64_mls_elt<mode>): Likewise.
3607 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
3608 (<su><maxmin><mode>3): Likewise.
3609 (move_lo_quad_<mode>): Likewise.
3610 (aarch64_simd_move_hi_quad_<mode>): Likewise.
3611 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
3612 (vec_pack_trunc_<mode>): Likewise.
3613 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
3614 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
3615 (*aarch64_<su>mlal_lo<mode>): Likewise.
3616 (*aarch64_<su>mlal_hi<mode>): Likewise.
3617 (*aarch64_<su>mlsl_lo<mode>): Likewise.
3618 (*aarch64_<su>mlsl_hi<mode>): Likewise.
3619 (*aarch64_<su>mlal<mode>): Likewise.
3620 (*aarch64_<su>mlsl<mode>): Likewise.
3621 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
3622 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
3623 (add<mode>3): Likewise.
3624 (sub<mode>3): Likewise.
3625 (mul<mode>3): Likewise.
3626 (div<mode>3): Likewise.
3627 (neg<mode>2): Likewise.
3628 (abs<mode>2): Likewise.
3629 (fma<mode>4): Likewise.
3630 (*aarch64_fma4_elt<mode>): Likewise.
3631 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
3632 (*aarch64_fma4_elt_to_128df): Likewise.
3633 (*aarch64_fma4_elt_to_64v2df): Likewise.
3634 (fnma<mode>4): Likewise.
3635 (*aarch64_fnma4_elt<mode>): Likewise.
3636 (*aarch64_fnma4_elt_<vswap_width_name><mode>
3637 (*aarch64_fnma4_elt_to_128df): Likewise.
3638 (*aarch64_fnma4_elt_to_64v2df): Likewise.
3639 (<frint_pattern><mode>2): Likewise.
3640 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
3641 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
3642 (vec_unpacks_lo_v4sf): Likewise.
3643 (aarch64_float_extend_lo_v2df): Likewise.
3644 (vec_unpacks_hi_v4sf): Likewise.
3645 (aarch64_float_truncate_lo_v2sf): Likewise.
3646 (aarch64_float_truncate_hi_v4sf): Likewise.
3647 (aarch64_vmls<mode>): Likewise.
3648 (<su><maxmin><mode>3): Likewise.
3649 (<maxmin_uns><mode>3): Likewise.
3650 (reduc_<sur>plus_<mode>): Likewise.
3651 (reduc_<sur>plus_v2di): Likewise.
3652 (reduc_<sur>plus_v2si): Likewise.
3653 (reduc_<sur>plus_<mode>): Likewise.
3654 (aarch64_addpv4sf): Likewise.
3655 (clz<mode>2): Likewise.
3656 (reduc_<maxmin_uns>_<mode>): Likewise.
3657 (reduc_<maxmin_uns>_v2di): Likewise.
3658 (reduc_<maxmin_uns>_v2si): Likewise.
3659 (reduc_<maxmin_uns>_<mode>): Likewise.
3660 (reduc_<maxmin_uns>_v4sf): Likewise.
3661 (aarch64_simd_bsl<mode>_internal): Likewise.
3662 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
3663 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
3664 (aarch64_get_lane<mode>): Likewise.
3665 (*aarch64_combinez<mode>): Likewise.
3666 (aarch64_combine<mode>): Likewise.
3667 (aarch64_simd_combine<mode>): Likewise.
3668 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
3669 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
3670 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
3671 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
3672 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
3673 (aarch64_<sur>h<addsub><mode>): Likewise.
3674 (aarch64_<sur><addsub>hn<mode>): Likewise.
3675 (aarch64_<sur><addsub>hn2<mode>): Likewise.
3676 (aarch64_pmul<mode>): Likewise.
3677 (aarch64_<su_optab><optab><mode>): Likewise.
3678 (aarch64_<sur>qadd<mode>): Likewise.
3679 (aarch64_sqmovun<mode>): Likewise.
3680 (aarch64_<sur>qmovn<mode>): Likewise.
3681 (aarch64_s<optab><mode>): Likewise.
3682 (aarch64_sq<r>dmulh<mode>): Likewise.
3683 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3684 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
3685 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3686 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
3687 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3688 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3689 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
3690 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
3691 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
3692 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
3693 (aarch64_sqdmull<mode>): Likewise.
3694 (aarch64_sqdmull_lane<mode>_internal): Likewise.
3695 (aarch64_sqdmull_n<mode>): Likewise.
3696 (aarch64_sqdmull2<mode>_internal): Likewise.
3697 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
3698 (aarch64_sqdmull2_n<mode>_internal): Likewise.
3699 (aarch64_<sur>shl<mode>): Likewise.
3700 (aarch64_<sur>q<r>shl<mode>
3701 (aarch64_<sur>shll_n<mode>): Likewise.
3702 (aarch64_<sur>shll2_n<mode>): Likewise.
3703 (aarch64_<sur>shr_n<mode>): Likewise.
3704 (aarch64_<sur>sra_n<mode>): Likewise.
3705 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
3706 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
3707 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
3708 (aarch64_cm<optab><mode>): Likewise.
3709 (aarch64_cm<optab>di): Likewise.
3710 (aarch64_cm<optab><mode>): Likewise.
3711 (aarch64_cm<optab>di): Likewise.
3712 (aarch64_cmtst<mode>): Likewise.
3713 (aarch64_cmtstdi): Likewise.
3714 (aarch64_cm<optab><mode>): Likewise.
3715 (*aarch64_fac<optab><mode>): Likewise.
3716 (aarch64_addp<mode>): Likewise.
3717 (aarch64_addpdi): Likewise.
3718 (sqrt<mode>2): Likewise.
3719 (vec_load_lanesoi<mode>): Likewise.
3720 (vec_store_lanesoi<mode>): Likewise.
3721 (vec_load_lanesci<mode>): Likewise.
3722 (vec_store_lanesci<mode>): Likewise.
3723 (vec_load_lanesxi<mode>): Likewise.
3724 (vec_store_lanesxi<mode>): Likewise.
3725 (*aarch64_mov<mode>): Likewise.
3726 (aarch64_ld2<mode>_dreg): Likewise.
3727 (aarch64_ld2<mode>_dreg): Likewise.
3728 (aarch64_ld3<mode>_dreg): Likewise.
3729 (aarch64_ld3<mode>_dreg): Likewise.
3730 (aarch64_ld4<mode>_dreg): Likewise.
3731 (aarch64_ld4<mode>_dreg): Likewise.
3732 (aarch64_tbl1<mode>): Likewise.
3733 (aarch64_tbl2v16qi): Likewise.
3734 (aarch64_combinev16qi): Likewise.
3735 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
3736 (aarch64_st2<mode>_dreg): Likewise.
3737 (aarch64_st2<mode>_dreg): Likewise.
3738 (aarch64_st3<mode>_dreg): Likewise.
3739 (aarch64_st3<mode>_dreg): Likewise.
3740 (aarch64_st4<mode>_dreg): Likewise.
3741 (aarch64_st4<mode>_dreg): Likewise.
3742 (*aarch64_simd_ld1r<mode>): Likewise.
3743 (aarch64_frecpe<mode>): Likewise.
3744 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
3745 (aarch64_frecps<mode>): Likewise.
3746
3747 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3748
3749 * config/arm/iterators.md (V_elem_ch): New.
3750 (q): Likewise.
3751 (VQH_type): Likewise.
3752 * config/arm/arm.md (is_neon_type): New.
3753 (conds): Use is_neon_type.
3754 (anddi3_insn): Update type attribute.
3755 (xordi3_insn): Likewise.
3756 (one_cmpldi2): Likewise.
3757 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
3758 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
3759 (*movmisalign<mode>_neon_store): Likewise.
3760 (*movmisalign<mode>_neon_load): Likewise.
3761 (vec_set<mode>_internal): Likewise.
3762 (vec_set<mode>_internal): Likewise.
3763 (vec_setv2di_internal): Likewise.
3764 (vec_extract<mode>): Likewise.
3765 (vec_extract<mode>): Likewise.
3766 (vec_extractv2di): Likewise.
3767 (*add<mode>3_neon): Likewise.
3768 (adddi3_neon): Likewise.
3769 (*sub<mode>3_neon): Likewise.
3770 (subdi3_neon): Likewise.
3771 (fma<VCVTF:mode>4): Likewise.
3772 (fma<VCVTF:mode>4_intrinsic): Likewise.
3773 (*fmsub<VCVTF:mode>4): Likewise.
3774 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3775 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3776 (ior<mode>3): Likewise.
3777 (and<mode>3): Likewise.
3778 (orn<mode>3_neon): Likewise.
3779 (orndi3_neon): Likewise.
3780 (bic<mode>3_neon): Likewise.
3781 (bicdi3_neon): Likewise.
3782 (xor<mode>3): Likewise.
3783 (one_cmpl<mode>2): Likewise.
3784 (abs<mode>2): Likewise.
3785 (neg<mode>2): Likewise.
3786 (negdi2_neon): Likewise.
3787 (*umin<mode>3_neon): Likewise.
3788 (*umax<mode>3_neon): Likewise.
3789 (*smin<mode>3_neon): Likewise.
3790 (*smax<mode>3_neon): Likewise.
3791 (vashl<mode>3): Likewise.
3792 (vashr<mode>3_imm): Likewise.
3793 (vlshr<mode>3_imm): Likewise.
3794 (ashl<mode>3_signed): Likewise.
3795 (ashl<mode>3_unsigned): Likewise.
3796 (neon_load_count): Likewise.
3797 (ashldi3_neon_noclobber): Likewise.
3798 (ashldi3_neon): Likewise.
3799 (signed_shift_di3_neon): Likewise.
3800 (unsigned_shift_di3_neon): Likewise.
3801 (ashrdi3_neon_imm_noclobber): Likewise.
3802 (lshrdi3_neon_imm_noclobber): Likewise.
3803 (<shift>di3_neon): Likewise.
3804 (widen_ssum<mode>3): Likewise.
3805 (widen_usum<mode>3): Likewise.
3806 (quad_halves_<code>v4si): Likewise.
3807 (quad_halves_<code>v4sf): Likewise.
3808 (quad_halves_<code>v8hi): Likewise.
3809 (quad_halves_<code>v16qi): Likewise.
3810 (reduc_splus_v2di): Likewise.
3811 (neon_vpadd_internal<mode>): Likewise.
3812 (neon_vpsmin<mode>): Likewise.
3813 (neon_vpsmax<mode>): Likewise.
3814 (neon_vpumin<mode>): Likewise.
3815 (neon_vpumax<mode>): Likewise.
3816 (*ss_add<mode>_neon): Likewise.
3817 (*us_add<mode>_neon): Likewise.
3818 (*ss_sub<mode>_neon): Likewise.
3819 (*us_sub<mode>_neon): Likewise.
3820 (neon_vadd<mode>_unspec): Likewise.
3821 (neon_vaddl<mode>): Likewise.
3822 (neon_vaddw<mode>): Likewise.
3823 (neon_vhadd<mode>): Likewise.
3824 (neon_vqadd<mode>): Likewise.
3825 (neon_vaddhn<mode>): Likewise.
3826 (neon_vmul<mode>): Likewise.
3827 (neon_vfms<VCVTF:mode>): Likewise.
3828 (neon_vmlal<mode>): Likewise.
3829 (neon_vmls<mode>): Likewise.
3830 (neon_vmlsl<mode>): Likewise.
3831 (neon_vqdmulh<mode>): Likewise.
3832 (neon_vqdmlal<mode>): Likewise.
3833 (neon_vqdmlsl<mode>): Likewise.
3834 (neon_vmull<mode>): Likewise.
3835 (neon_vqdmull<mode>): Likewise.
3836 (neon_vsub<mode>_unspec): Likewise.
3837 (neon_vsubl<mode>): Likewise.
3838 (neon_vsubw<mode>): Likewise.
3839 (neon_vqsub<mode>): Likewise.
3840 (neon_vhsub<mode>): Likewise.
3841 (neon_vsubhn<mode>): Likewise.
3842 (neon_vceq<mode>): Likewise.
3843 (neon_vcge<mode>): Likewise.
3844 (neon_vcgeu<mode>): Likewise.
3845 (neon_vcgt<mode>): Likewise.
3846 (neon_vcgtu<mode>): Likewise.
3847 (neon_vcle<mode>): Likewise.
3848 (neon_vclt<mode>): Likewise.
3849 (neon_vcage<mode>): Likewise.
3850 (neon_vcagt<mode>): Likewise.
3851 (neon_vtst<mode>): Likewise.
3852 (neon_vabd<mode>): Likewise.
3853 (neon_vabdl<mode>): Likewise.
3854 (neon_vaba<mode>): Likewise.
3855 (neon_vabal<mode>): Likewise.
3856 (neon_vmax<mode>): Likewise.
3857 (neon_vmin<mode>): Likewise.
3858 (neon_vpaddl<mode>): Likewise.
3859 (neon_vpadal<mode>): Likewise.
3860 (neon_vpmax<mode>): Likewise.
3861 (neon_vpmin<mode>): Likewise.
3862 (neon_vrecps<mode>): Likewise.
3863 (neon_vrsqrts<mode>): Likewise.
3864 (neon_vqabs<mode>): Likewise.
3865 (neon_vqneg<mode>): Likewise.
3866 (neon_vcls<mode>): Likewise.
3867 (clz<mode>2): Likewise.
3868 (popcount<mode>2): Likewise.
3869 (neon_vrecpe<mode>): Likewise.
3870 (neon_vrsqrte<mode>): Likewise.
3871 (neon_vget_lane<mode>_sext_internal): Likewise.
3872 (neon_vget_lane<mode>_zext_internal): Likewise.
3873 (neon_vdup_n<mode>): Likewise.
3874 (neon_vdup_n<mode>): Likewise.
3875 (neon_vdup_nv2di): Likewise.
3876 (neon_vdup_lane<mode>_interal): Likewise.
3877 (*neon_vswp<mode>): Likewise.
3878 (neon_vcombine<mode>): Likewise.
3879 (float<mode><V_cvtto>2): Likewise.
3880 (floatuns<mode><V_cvtto>2): Likewise.
3881 (fix_trunc<mode><V_cvtto>2): Likewise.
3882 (fixuns_trunc<mode><V_cvtto>2
3883 (neon_vcvt<mode>): Likewise.
3884 (neon_vcvt<mode>): Likewise.
3885 (neon_vcvtv4sfv4hf): Likewise.
3886 (neon_vcvtv4hfv4sf): Likewise.
3887 (neon_vcvt_n<mode>): Likewise.
3888 (neon_vcvt_n<mode>): Likewise.
3889 (neon_vmovn<mode>): Likewise.
3890 (neon_vqmovn<mode>): Likewise.
3891 (neon_vqmovun<mode>): Likewise.
3892 (neon_vmovl<mode>): Likewise.
3893 (neon_vmul_lane<mode>): Likewise.
3894 (neon_vmul_lane<mode>): Likewise.
3895 (neon_vmull_lane<mode>): Likewise.
3896 (neon_vqdmull_lane<mode>): Likewise.
3897 (neon_vqdmulh_lane<mode>): Likewise.
3898 (neon_vqdmulh_lane<mode>): Likewise.
3899 (neon_vmla_lane<mode>): Likewise.
3900 (neon_vmla_lane<mode>): Likewise.
3901 (neon_vmlal_lane<mode>): Likewise.
3902 (neon_vqdmlal_lane<mode>): Likewise.
3903 (neon_vmls_lane<mode>): Likewise.
3904 (neon_vmls_lane<mode>): Likewise.
3905 (neon_vmlsl_lane<mode>): Likewise.
3906 (neon_vqdmlsl_lane<mode>): Likewise.
3907 (neon_vext<mode>): Likewise.
3908 (neon_vrev64<mode>): Likewise.
3909 (neon_vrev32<mode>): Likewise.
3910 (neon_vrev16<mode>): Likewise.
3911 (neon_vbsl<mode>_internal): Likewise.
3912 (neon_vshl<mode>): Likewise.
3913 (neon_vqshl<mode>): Likewise.
3914 (neon_vshr_n<mode>): Likewise.
3915 (neon_vshrn_n<mode>): Likewise.
3916 (neon_vqshrn_n<mode>): Likewise.
3917 (neon_vqshrun_n<mode>): Likewise.
3918 (neon_vshl_n<mode>): Likewise.
3919 (neon_vqshl_n<mode>): Likewise.
3920 (neon_vqshlu_n<mode>): Likewise.
3921 (neon_vshll_n<mode>): Likewise.
3922 (neon_vsra_n<mode>): Likewise.
3923 (neon_vsri_n<mode>): Likewise.
3924 (neon_vsli_n<mode>): Likewise.
3925 (neon_vtbl1v8qi): Likewise.
3926 (neon_vtbl2v8qi): Likewise.
3927 (neon_vtbl3v8qi): Likewise.
3928 (neon_vtbl4v8qi): Likewise.
3929 (neon_vtbl1v16qi): Likewise.
3930 (neon_vtbl2v16qi): Likewise.
3931 (neon_vcombinev16qi): Likewise.
3932 (neon_vtbx1v8qi): Likewise.
3933 (neon_vtbx2v8qi): Likewise.
3934 (neon_vtbx3v8qi): Likewise.
3935 (neon_vtbx4v8qi): Likewise.
3936 (*neon_vtrn<mode>_insn): Likewise.
3937 (*neon_vzip<mode>_insn): Likewise.
3938 (*neon_vuzp<mode>_insn): Likewise.
3939 (neon_vld1<mode>): Likewise.
3940 (neon_vld1_lane<mode>): Likewise.
3941 (neon_vld1_lane<mode>): Likewise.
3942 (neon_vld1_dup<mode>): Likewise.
3943 (neon_vld1_dup<mode>): Likewise.
3944 (neon_vld1_dupv2di): Likewise.
3945 (neon_vst1<mode>): Likewise.
3946 (neon_vst1_lane<mode>): Likewise.
3947 (neon_vst1_lane<mode>): Likewise.
3948 (neon_vld2<mode>): Likewise.
3949 (neon_vld2<mode>): Likewise.
3950 (neon_vld2_lane<mode>): Likewise.
3951 (neon_vld2_lane<mode>): Likewise.
3952 (neon_vld2_dup<mode>): Likewise.
3953 (neon_vst2<mode>): Likewise.
3954 (neon_vst2<mode>): Likewise.
3955 (neon_vst2_lane<mode>): Likewise.
3956 (neon_vst2_lane<mode>): Likewise.
3957 (neon_vld3<mode>): Likewise.
3958 (neon_vld3qa<mode>): Likewise.
3959 (neon_vld3qb<mode>): Likewise.
3960 (neon_vld3_lane<mode>): Likewise.
3961 (neon_vld3_lane<mode>): Likewise.
3962 (neon_vld3_dup<mode>): Likewise.
3963 (neon_vst3<mode>): Likewise.
3964 (neon_vst3qa<mode>): Likewise.
3965 (neon_vst3qb<mode>): Likewise.
3966 (neon_vst3_lane<mode>): Likewise.
3967 (neon_vst3_lane<mode>): Likewise.
3968 (neon_vld4<mode>): Likewise.
3969 (neon_vld4qa<mode>): Likewise.
3970 (neon_vld4qb<mode>): Likewise.
3971 (neon_vld4_lane<mode>): Likewise.
3972 (neon_vld4_lane<mode>): Likewise.
3973 (neon_vld4_dup<mode>): Likewise.
3974 (neon_vst4<mode>): Likewise.
3975 (neon_vst4qa<mode>): Likewise.
3976 (neon_vst4qb<mode>): Likewise.
3977 (neon_vst4_lane<mode>): Likewise.
3978 (neon_vst4_lane<mode>): Likewise.
3979 (neon_vec_unpack<US>_lo_<mode>): Likewise.
3980 (neon_vec_unpack<US>_hi_<mode>): Likewise.
3981 (neon_vec_<US>mult_lo_<mode>): Likewise.
3982 (neon_vec_<US>mult_hi_<mode>): Likewise.
3983 (neon_vec_<US>shiftl_<mode>): Likewise.
3984 (neon_unpack<US>_<mode>): Likewise.
3985 (neon_vec_<US>mult_<mode>): Likewise.
3986 (vec_pack_trunc_<mode>): Likewise.
3987 (neon_vec_pack_trunc_<mode>): Likewise.
3988 (neon_vabd<mode>_2): Likewise.
3989 (neon_vabd<mode>_3): Likewise.
3990
3991 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3992
3993 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
3994 (load_pair): Update type attribute.
3995 (store_pair): Update type attribute.
3996 * config/aarch64/iterators.md (q): New.
3997
3998 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3999
4000 * config/arm/types.md: Add new types for Neon insns.
4001
4002 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4003 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4004 Sergey Lega <sergey.s.lega@intel.com>
4005 Anna Tikhonova <anna.tikhonova@intel.com>
4006 Ilya Tocar <ilya.tocar@intel.com>
4007 Andrey Turetskiy <andrey.turetskiy@intel.com>
4008 Ilya Verbin <ilya.verbin@intel.com>
4009 Kirill Yukhin <kirill.yukhin@intel.com>
4010 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4011
4012 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
4013 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
4014 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
4015 (rcp14<mode>): New.
4016 (srcp14<mode>): Ditto.
4017 (rsqrt14<mode>): Ditto.
4018 (rsqrt14<mode>): Ditto.
4019 (avx512f_vmscalef<mode>): Ditto.
4020 (avx512f_scalef<mode>): Ditto.
4021 (avx512f_getexp<mode>): Ditto.
4022 (avx512f_sgetexp<mode>): Ditto.
4023 (avx512f_fixupimm<mode>): Ditto.
4024 (avx512f_sfixupimm<mode>): Ditto.
4025 (avx512f_rndscale<mode>): Ditto.
4026 (*avx512er_exp2<mode>): Ditto.
4027 (*avx512er_rcp28<mode>): Ditto.
4028 (avx512er_rsqrt28<mode>): Ditto.
4029 (avx512f_getmant<mode>): Ditto.
4030 (avx512f_getmant<mode>): Ditto.
4031 (avx512f_rndscale<mode>): Fix formatting.
4032
4033 2013-10-15 Martin Jambor <mjambor@suse.cz>
4034
4035 * ipa-utils.h (ipa_edge_within_scc): Declare.
4036 * ipa-cp.c (edge_within_scc): Moved...
4037 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
4038
4039 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4040 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4041 Sergey Lega <sergey.s.lega@intel.com>
4042 Anna Tikhonova <anna.tikhonova@intel.com>
4043 Ilya Tocar <ilya.tocar@intel.com>
4044 Andrey Turetskiy <andrey.turetskiy@intel.com>
4045 Ilya Verbin <ilya.verbin@intel.com>
4046 Kirill Yukhin <kirill.yukhin@intel.com>
4047 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4048
4049 * config/i386/predicates.md (const_8_to_15_operand): New.
4050 (const_16_to_31_operand): Ditto.
4051 * config/i386/sse.md (V8FI): New.
4052 (V16FI): Ditto.
4053 (reduc_splus_v8df): Ditto.
4054 (reduc_splus_v16sf): Ditto.
4055 (avx512f_vextract<shuffletype>32x4_1): Ditto.
4056 (vec_extract_hi_<mode>): Ditto.
4057 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
4058 (vec_set_lo_<mode>): Ditto.
4059 (vec_set_hi_<mode>): Ditto.
4060 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
4061 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
4062 (avx512f_pshufd_1): Ditto.
4063 (avx512f_broadcast<mode>): Ditto.
4064 (avx512f_broadcast<mode>): Ditto.
4065 (define_split): Split vec_extract_lo into move.
4066 (ssequartermode): Ditto.
4067 (ssedoublemode): Extened with wider modes.
4068 (vec_extract_lo_<mode>): Ditto.
4069
4070 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4071 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4072 Sergey Lega <sergey.s.lega@intel.com>
4073 Anna Tikhonova <anna.tikhonova@intel.com>
4074 Ilya Tocar <ilya.tocar@intel.com>
4075 Andrey Turetskiy <andrey.turetskiy@intel.com>
4076 Ilya Verbin <ilya.verbin@intel.com>
4077 Kirill Yukhin <kirill.yukhin@intel.com>
4078 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4079
4080 * config/i386/predicates.md (register_or_constm1_operand): New.
4081 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
4082 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
4083 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
4084 UNSPEC_SCATTER_PREFETCH
4085 (VI48_512): New.
4086 (avx512f_ucmp<mode>3): Ditto.
4087 (avx512f_vternlog<mode>): Ditto.
4088 (avx512f_align<mode>): Ditto.
4089 (<shift_insn><mode>3): Ditto.
4090 (avx512f_<rotate>v<mode>): Ditto.
4091 (avx512f_<rotate><mode>): Ditto.
4092 (avx512f_eq<mode>3): Ditto.
4093 (avx512f_eq<mode>3_1): Ditto.
4094 (avx512f_gt<mode>3): Ditto.
4095 (avx512f_testm<mode>3): Ditto.
4096 (avx512f_testnm<mode>3): Ditto.
4097 (avx512pf_gatherpf<mode>): Ditto.
4098 (*avx512pf_gatherpf<mode>_mask): Ditto.
4099 (*avx512pf_gatherpf<mode>): Ditto.
4100 (avx512pf_scatterpf<mode>): Ditto.
4101 (*avx512pf_scatterpf<mode>_mask): Ditto.
4102 (*avx512pf_scatterpf<mode>): Ditto.
4103 (avx512f_vec_dup_gpr<mode>): Ditto.
4104 (clz<mode>2): Ditto.
4105 (conflict<mode>): Ditto.
4106 (REDUC_SMINMAX_MODE): Extened with wider modes.
4107 (reduc_<code>_<mode>): Ditto.
4108 (vlshr<mode>3): Ditto.
4109 (vashl<mode>3): Ditto.
4110
4111 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4112 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4113 Sergey Lega <sergey.s.lega@intel.com>
4114 Anna Tikhonova <anna.tikhonova@intel.com>
4115 Ilya Tocar <ilya.tocar@intel.com>
4116 Andrey Turetskiy <andrey.turetskiy@intel.com>
4117 Ilya Verbin <ilya.verbin@intel.com>
4118 Kirill Yukhin <kirill.yukhin@intel.com>
4119 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4120
4121 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
4122 UNSPEC_SCATTER.
4123 (VI48F_512): New.
4124 (avx512fmaskmode): Ditto.
4125 (bcstscalarsuff): Ditto.
4126 (avx512f_blendm<mode>): Ditto.
4127 (cmp_imm_predicate): Ditto.
4128 (avx512f_cmp<mode>3): Ditto.
4129 (avx512f_vec_dup<mode>): Ditto.
4130 (avx512f_vec_dup_mem<mode>): Ditto.
4131 (avx512f_vpermi2var<mode>3): Ditto.
4132 (avx512f_vpermt2var<mode>3): Ditto.
4133 (vec_init<mode>): Ditto.
4134 (avx512f_gathersi<mode>): Ditto.
4135 (*avx512f_gathersi<mode>): Ditto.
4136 (*avx512f_gathersi<mode>_2): Ditto.
4137 (avx512f_gatherdi<mode>): Ditto.
4138 (*avx512f_gatherdi<mode>): Ditto.
4139 (*avx512f_gatherdi<mode>_2): Ditto.
4140 (avx512f_scattersi<mode>): Ditto.
4141 (*avx512f_scattersi<mode>): Ditto.
4142 (avx512f_scatterdi<mode>): Ditto.
4143 (*avx512f_scatterdi<mode>): Ditto.
4144 (sseintprefix): Extened with wider modes.
4145 (VEC_GATHER_IDXSI): Ditto.
4146 (VEC_GATHER_IDXDI): Ditto.
4147 (VEC_GATHER_SRCDI): Ditto.
4148
4149 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
4150 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4151
4152 * config/arm/t-aprofile: New file.
4153 * config.gcc: Handle --with-multilib-list option.
4154
4155 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
4156
4157 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
4158 the input for r2.
4159
4160 2013-10-15 Richard Biener <rguenther@suse.de>
4161
4162 * tree-loop-distribution.c (build_empty_rdg): Inline into
4163 single user.
4164 (rdg_flag_vertex): Inline into single user.
4165 (rdg_flag_vertex_and_dependent): Likewise.
4166 (build_rdg_partition_for_vertex): Remove processed bitmap.
4167 (rdg_build_partitions): Simplify.
4168
4169 2013-10-15 Richard Biener <rguenther@suse.de>
4170
4171 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
4172 Restructure forwarding through conversions and copies to
4173 avoid performing copy-propagation the wrong way. Adjust
4174 recursion invocations.
4175 (forward_propagate_addr_expr): Add argument stating if we
4176 are recursing from a single-use.
4177 (ssa_forward_propagate_and_combine): Adjust.
4178
4179 2013-10-14 David Malcolm <dmalcolm@redhat.com>
4180
4181 * dumpfile.h (gcc::dump_manager): New class, to hold state
4182 relating to dumpfile management.
4183 (get_dump_file_name): Remove in favor of method of dump_manager.
4184 (dump_initialized_p): Likewise.
4185 (dump_start): Likewise.
4186 (dump_finish): Likewise.
4187 (dump_switch_p): Likewise.
4188 (dump_register): Likewise.
4189 (get_dump_file_info): Likewise.
4190 * context.c (gcc::context::context): Construct the dump_manager
4191 instance.
4192 * context.h (gcc::context::get_dumps): New.
4193 (gcc::context::m_dumps): New.
4194 * coverage.c (coverage_init): Port to dump_manager API.
4195 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
4196 (extra_dump_files_in_use): Likewise.
4197 (extra_dump_files_alloced): Likewise.
4198 (gcc::dump_manager::dump_manager): New.
4199 (dump_register): Convert to...
4200 (gcc::dump_manager::dump_register): ...method, replacing
4201 function-static next_dump with m_next_dump field.
4202 (get_dump_file_info): Convert to...
4203 (gcc::dump_manager::get_dump_file_info): ...method.
4204 (get_dump_file_name): Convert to...
4205 (gcc::dump_manager::get_dump_file_name): ...method.
4206 (dump_start): Convert to...
4207 (gcc::dump_manager::dump_start): ...method.
4208 (dump_finish): Convert to...
4209 (gcc::dump_manager::dump_finish): ...method.
4210 (dump_begin): Replace body with...
4211 (gcc::dump_manager::dump_begin): ...new method.
4212 (dump_phase_enabled_p): Convert to...
4213 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4214 (dump_phase_enabled_p): Convert to...
4215 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4216 (dump_initialized_p): Convert to...
4217 (gcc::dump_manager::dump_initialized_p): ...method.
4218 (dump_flag_name): Replace body with...
4219 (gcc::dump_manager::dump_flag_name): ...new method.
4220 (dump_enable_all): Convert to...
4221 (gcc::dump_manager::dump_enable_all): ...new method.
4222 (opt_info_enable_passes): Convert to...
4223 (gcc::dump_manager::opt_info_enable_passes): ...new method.
4224 (dump_switch_p_1): Convert to...
4225 (gcc::dump_manager::dump_switch_p_1): ...new method.
4226 (dump_switch_p): Convert to...
4227 (gcc::dump_manager::dump_switch_p): ...new method.
4228 (opt_info_switch_p): Port to dump_manager API.
4229 (enable_rtl_dump_file): Likewise.
4230 * opts-global.c (handle_common_deferred_options): Port to new
4231 dump_manager API.
4232 * passes.c (pass_manager::finish_optimization_passes): Likewise.
4233 (pass_manager::register_one_dump_file): Likewise.
4234 (pass_manager::register_pass): Likewise.
4235 (pass_init_dump_file): Likewise.
4236 (pass_fini_dump_file): Likewise.
4237 * statistics.c (statistics_early_init): Likewise.
4238
4239 2013-10-14 Richard Biener <rguenther@suse.de>
4240
4241 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
4242 iterative_hash_canonical_type, gimple_canonical_type_hash,
4243 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
4244 gimple_register_canonical_type, print_gimple_types_stats,
4245 free_gimple_type_tables): Move to lto/lto.c
4246 (gt-gimple.h): Do not include.
4247 * gimple.h (gimple_register_canonical_type,
4248 print_gimple_types_stats, free_gimple_type_tables): Remove.
4249 * Makefile.in (GTFILES): Remove gimple.c.
4250
4251 2013-10-14 Travis Snoozy <quandary@remstate.com>
4252
4253 PR target/58716
4254 * config/msp430/msp430.c (msp430_option_override): Correct thinko
4255 scanning for msp430x targets.
4256
4257 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
4258
4259 PR bootstrap/58509
4260 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
4261 (registers_ok_for_ldd_peep): Move around.
4262 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
4263 * config/sparc/sparc.md (widening peepholes): Use it.
4264
4265 2013-10-14 Richard Biener <rguenther@suse.de>
4266
4267 PR middle-end/58712
4268 PR middle-end/55358
4269 * gimple.c (iterative_hash_canonical_type): Make sure to
4270 record the hash into the correct hashtable slot.
4271
4272 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
4273
4274 PR rtl-optimization/58662
4275 * combine.c (try_combine): Take into account death nodes on I2 when
4276 splitting a PARALLEL of two independent SETs. Fix dump message.
4277
4278 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
4279
4280 PR target/51244
4281 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
4282 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
4283 * config/sh/t-sh (sh_treg_combine.o): New entry.
4284 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
4285 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
4286 (register_sh_passes): New function. Register sh_treg_combine pass.
4287 (sh_option_override): Invoke it.
4288 (sh_canonicalize_comparison): Handle op0_preserve_value.
4289 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
4290 opportunities. Canonicalize branch condition.
4291 (nott): Allow only if pseudos can be created for non-SH2A.
4292
4293 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
4294
4295 PR target/58690
4296 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
4297 (ix86_expand_movmem): Replace copy_addr_to_reg with
4298 ix86_copy_addr_to_reg.
4299 (ix86_expand_setmem): Likewise.
4300
4301 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
4302
4303 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
4304 provided by ix86_fp_compare_mode instead of CCFPUmode.
4305
4306 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
4307
4308 * config/aarch64/arm_neon.h
4309 (vtbx<1,3>_<psu>8): Fix register constriants.
4310
4311 2013-10-11 Jeff Law <law@redhat.com>
4312
4313 PR tree-optimization/58640
4314 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
4315 threading paths that cross over two loop entry points.
4316
4317 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4318
4319 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
4320 handle vector float as well.
4321 (*vsx_le_perm_load_v4si): Likewise.
4322 (*vsx_le_perm_store_v2di): Likewise.
4323 (*vsx_le_perm_store_v4si): Likewise.
4324
4325 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4326
4327 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
4328 directly to circumvent subtract from splat{31} workaround.
4329 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
4330 prototype.
4331 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
4332 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
4333 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
4334 (altivec_vperm_<mode>): Convert to define_insn_and_split to
4335 separate big and little endian logic.
4336 (*altivec_vperm_<mode>_internal): New define_insn.
4337 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
4338 separate big and little endian logic.
4339 (*altivec_vperm_<mode>_uns_internal): New define_insn.
4340 (vec_permv16qi): Add little endian logic.
4341
4342 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
4343
4344 * doc/extend.texi (returns_nonnull): Remove arguments.
4345
4346 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4347 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4348 Sergey Lega <sergey.s.lega@intel.com>
4349 Anna Tikhonova <anna.tikhonova@intel.com>
4350 Ilya Tocar <ilya.tocar@intel.com>
4351 Andrey Turetskiy <andrey.turetskiy@intel.com>
4352 Ilya Verbin <ilya.verbin@intel.com>
4353 Kirill Yukhin <kirill.yukhin@intel.com>
4354 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4355
4356 * config/i386/sse.md (VI48F_256_512): New.
4357 (avx2_permvar<mode>): Change to ...
4358 (<avx2_avx512f>_permvar<mode>): This.
4359
4360 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4361 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4362 Sergey Lega <sergey.s.lega@intel.com>
4363 Anna Tikhonova <anna.tikhonova@intel.com>
4364 Ilya Tocar <ilya.tocar@intel.com>
4365 Andrey Turetskiy <andrey.turetskiy@intel.com>
4366 Ilya Verbin <ilya.verbin@intel.com>
4367 Kirill Yukhin <kirill.yukhin@intel.com>
4368 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4369
4370 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
4371 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
4372 * config/i386/sse.md (VI4_AVX): New.
4373 (sf2simodelower): Ditto.
4374 (sse2_cvtps2dq): Change to ...
4375 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
4376
4377 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4378 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4379 Sergey Lega <sergey.s.lega@intel.com>
4380 Anna Tikhonova <anna.tikhonova@intel.com>
4381 Ilya Tocar <ilya.tocar@intel.com>
4382 Andrey Turetskiy <andrey.turetskiy@intel.com>
4383 Ilya Verbin <ilya.verbin@intel.com>
4384 Kirill Yukhin <kirill.yukhin@intel.com>
4385 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4386
4387 * config/i386/sse.md (V_512): New.
4388 (VI_512): Ditto.
4389 (vcond<V_512:mode><VF_512:mode>): Ditto.
4390 (vcond<V_512:mode><VI_512:mode>): Ditto.
4391 (vcondu<V_512:mode><VI_512:mode>): Ditto.
4392
4393 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4394 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4395 Sergey Lega <sergey.s.lega@intel.com>
4396 Anna Tikhonova <anna.tikhonova@intel.com>
4397 Ilya Tocar <ilya.tocar@intel.com>
4398 Andrey Turetskiy <andrey.turetskiy@intel.com>
4399 Ilya Verbin <ilya.verbin@intel.com>
4400 Kirill Yukhin <kirill.yukhin@intel.com>
4401 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4402
4403 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
4404 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
4405 (FMAMODE): Ditto.
4406 (fma<mode>4): Removed condition.
4407 (fms<mode>4): Ditto.
4408 (fnma<mode>4): Ditto.
4409 (fnms<mode>4): Ditto.
4410 (fma4i_fmadd_<mode>): Ditto.
4411 (*fma_fmadd_<mode>): Ditto.
4412 (*fma_fmsub_<mode>): Ditto.
4413 (*fma_fnmadd_<mode>): Ditto.
4414 (*fma_fnmsub_<mode>): Ditto.
4415 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
4416 (*fma_fmaddsub_<mode>): Ditto.
4417 (*fma_fmsubadd_<mode>): Ditto.
4418 (*fmai_fmadd_<mode>): Ditto.
4419 (*fmai_fmsub_<mode>): Ditto.
4420 (*fmai_fnmadd_<mode>): Ditto.
4421 (*fmai_fnmsub_<mode>): Ditto.
4422
4423 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4424 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4425 Sergey Lega <sergey.s.lega@intel.com>
4426 Anna Tikhonova <anna.tikhonova@intel.com>
4427 Ilya Tocar <ilya.tocar@intel.com>
4428 Andrey Turetskiy <andrey.turetskiy@intel.com>
4429 Ilya Verbin <ilya.verbin@intel.com>
4430 Kirill Yukhin <kirill.yukhin@intel.com>
4431 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4432
4433 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
4434 (VI124_256): Changed to ...
4435 (VI124_256_48_512): This.
4436 (ssepackmode): Extended with wider modes.
4437 (<code><mode>3): Changed iterator.
4438 (*avx2_<code><mode>3): Ditto.
4439 (vec_pack_trunc_<mode>): Ditto.
4440
4441 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4442 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4443 Sergey Lega <sergey.s.lega@intel.com>
4444 Anna Tikhonova <anna.tikhonova@intel.com>
4445 Ilya Tocar <ilya.tocar@intel.com>
4446 Andrey Turetskiy <andrey.turetskiy@intel.com>
4447 Ilya Verbin <ilya.verbin@intel.com>
4448 Kirill Yukhin <kirill.yukhin@intel.com>
4449 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4450
4451 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
4452 (VI8F_256_512): Ditto.
4453 (abs<mode>2): Changed iterator.
4454 (avx2_perm<mode>): Changed to ...
4455 (<avx2_avx512f>_perm<mode>): This.
4456 (avx2_perm<mode>_1): Changed to ...
4457 (<avx2_avx512f>_perm<mode>_1): This.
4458
4459 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4460 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4461 Sergey Lega <sergey.s.lega@intel.com>
4462 Anna Tikhonova <anna.tikhonova@intel.com>
4463 Ilya Tocar <ilya.tocar@intel.com>
4464 Andrey Turetskiy <andrey.turetskiy@intel.com>
4465 Ilya Verbin <ilya.verbin@intel.com>
4466 Kirill Yukhin <kirill.yukhin@intel.com>
4467 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4468
4469 * config/i386/sse.md (VI48_AVX512F): New.
4470 (VI48_AVX2): Changed to ...
4471 (VI48_AVX2_48_AVX512F): This.
4472 (avx2_ashrv<mode>): Changed to ...
4473 (<avx2_avx512f>_ashrv<mode>): This.
4474 (avx2_<shift_insn>v<mode>): Changed to ...
4475 (<avx2_avx512f>_<shift_insn>v<mode>): This.
4476
4477 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4478 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4479 Sergey Lega <sergey.s.lega@intel.com>
4480 Anna Tikhonova <anna.tikhonova@intel.com>
4481 Ilya Tocar <ilya.tocar@intel.com>
4482 Andrey Turetskiy <andrey.turetskiy@intel.com>
4483 Ilya Verbin <ilya.verbin@intel.com>
4484 Kirill Yukhin <kirill.yukhin@intel.com>
4485 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4486
4487 * config/i386/sse.md (VI4_AVX512F): New.
4488 (VI8_AVX2_AVX512F): Ditto.
4489 (mul<mode>3): Extended with wider modes.
4490 (*<sse4_1_avx2>_mul<mode>3): Ditto.
4491 (mul<mode>3): Ditto.
4492 (vec_widen_<s>mult_odd_<mode>): Ditto.
4493
4494 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4495 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4496 Sergey Lega <sergey.s.lega@intel.com>
4497 Anna Tikhonova <anna.tikhonova@intel.com>
4498 Ilya Tocar <ilya.tocar@intel.com>
4499 Andrey Turetskiy <andrey.turetskiy@intel.com>
4500 Ilya Verbin <ilya.verbin@intel.com>
4501 Kirill Yukhin <kirill.yukhin@intel.com>
4502 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4503
4504 * config/i386/sse.md (VI2_AVX512F): New.
4505 (VI124_AVX512F): Ditto.
4506 (sseunpackmode): Extended with wider modes.
4507 (sseunpackfltmode): Ditto.
4508 (vec_unpacks_float_hi_<mode>): Ditto.
4509 (vec_unpacks_float_lo_<mode>): Ditto.
4510 (vec_unpacku_float_hi_<mode>): Ditto.
4511 (vec_unpacku_float_lo_<mode>): Ditto.
4512 (vec_unpacks_lo_<mode>): Ditto.
4513 (vec_unpacks_hi_<mode>): Ditto.
4514 (vec_unpacku_lo_<mode>): Ditto.
4515 (vec_unpacku_hi_<mode>): Ditto.
4516
4517 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4518 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4519 Sergey Lega <sergey.s.lega@intel.com>
4520 Anna Tikhonova <anna.tikhonova@intel.com>
4521 Ilya Tocar <ilya.tocar@intel.com>
4522 Andrey Turetskiy <andrey.turetskiy@intel.com>
4523 Ilya Verbin <ilya.verbin@intel.com>
4524 Kirill Yukhin <kirill.yukhin@intel.com>
4525 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4526
4527 * config/i386/i386.md (multdiv): New.
4528 (multdiv_mnemonic): Ditto.
4529 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
4530 (<sse>_vm<multdiv_mnemonic><mode>3): This.
4531 (<sse>_vmdiv<mode>3): Removed.
4532
4533 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4534 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4535 Sergey Lega <sergey.s.lega@intel.com>
4536 Anna Tikhonova <anna.tikhonova@intel.com>
4537 Ilya Tocar <ilya.tocar@intel.com>
4538 Andrey Turetskiy <andrey.turetskiy@intel.com>
4539 Ilya Verbin <ilya.verbin@intel.com>
4540 Kirill Yukhin <kirill.yukhin@intel.com>
4541 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4542
4543 * config/i386/sse.md (V): Extended with wider modes.
4544 (VF2): Ditto.
4545 (ssehalfvecmode): Ditto.
4546 (i128): Ditto.
4547 (ssepackfltmode): Ditto.
4548 (avx_vec_concat<mode>): Ditto.
4549 (V_256_512): New iterator.
4550 (VF2_512_256): Ditto.
4551 (si2dfmode): New attribute.
4552 (si2dfmodelower): Ditto.
4553 (sf2dfmode): Ditto.
4554 (concat_tg_mode): Ditto.
4555 (floatv4siv4df2): Changed to ...
4556 (float<si2dfmodelower><mode>2): This.
4557 (avx_cvtps2pd256): Changed to ...
4558 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
4559 (vec_pack_trunc_v4df): Changed to ...
4560 (vec_pack_trunc_<mode>): This.
4561 (avx_vpermil<mode>): Changed to ...
4562 (<sse2_avx_avx512f>_vpermil<mode>): This.
4563 (<fixsuffix>fix_truncv8dfv8si2): New.
4564 (vec_pack_sfix_trunc_v8df): Ditto.
4565 (avx512f_rndscale<mode>): Ditto.
4566 (avx512f_roundpd512): Ditto.
4567 (vec_pack_ufix_trunc_<mode>): Updated iterator.
4568
4569 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4570 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4571 Sergey Lega <sergey.s.lega@intel.com>
4572 Anna Tikhonova <anna.tikhonova@intel.com>
4573 Ilya Tocar <ilya.tocar@intel.com>
4574 Andrey Turetskiy <andrey.turetskiy@intel.com>
4575 Ilya Verbin <ilya.verbin@intel.com>
4576 Kirill Yukhin <kirill.yukhin@intel.com>
4577 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4578
4579 * config/i386/i386.md (any_fix): New iterator.
4580 (fixsuffix): New attribute.
4581 * config/i386/sse.md (VF1): Extened with wider modes.
4582 (VI): Ditto.
4583 (VI_AVX2): Ditto.
4584 (VI8): Ditto.
4585 (sseintvecmodelower): Ditto.
4586 (ssescalarmode): Ditto.
4587 (ssescalarnum): Ditto.
4588 (VF1_128_256): New.
4589 (ssexmmmode): Ditto.
4590 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
4591 (<sse>_rcp<mode>2): Change iterator.
4592 (rsqrt<mode>2): Ditto.
4593 (<sse>_rsqrt<mode>2): Ditto.
4594 (avx2_vec_dup<mode>): Ditto.
4595 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
4596 (round<mode>2_sfix): Ditto.
4597 (avx2_pbroadcast<mode>): Ditto.
4598 (*andnot<mode>3): Handle XI mode.
4599 (*<code><mode>3): Ditto.
4600 (AVXTOSSEMODE): Removed.
4601 (avx_vpermil<mode>): Changed to ...
4602 (<sse2_avx_avx512f>_vpermil<mode>): This.
4603
4604 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4605 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4606 Sergey Lega <sergey.s.lega@intel.com>
4607 Anna Tikhonova <anna.tikhonova@intel.com>
4608 Ilya Tocar <ilya.tocar@intel.com>
4609 Andrey Turetskiy <andrey.turetskiy@intel.com>
4610 Ilya Verbin <ilya.verbin@intel.com>
4611 Kirill Yukhin <kirill.yukhin@intel.com>
4612 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4613
4614 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
4615 (<sse>_comi): Ditto.
4616 (<sse>_ucomi): Ditto.
4617 (sse_cvtss2siq_2): Ditto.
4618 (sse2_cvtsd2si): Ditto.
4619 (sse2_cvtsd2siq): Ditto.
4620 (sse2_cvttsd2si): Ditto.
4621 (sse2_cvttsd2siq): Ditto.
4622 (<shift_insn><mode>3): Ditto.
4623 (sse2_cvtsi2sdq): Update constraint and prefix.
4624 (sse_cvtsi2ss): Update prefix.
4625 (sse_cvtsi2ssq): Ditto.
4626
4627 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4628
4629 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
4630 ignore internal calls.
4631
4632 2013-10-11 Richard Biener <rguenther@suse.de>
4633
4634 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
4635 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
4636 TREE_OVERFLOW is set.
4637
4638 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
4639
4640 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
4641
4642 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
4643 * gimple.h: Likewise.
4644
4645 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
4646
4647 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
4648 macros.
4649
4650 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
4651
4652 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
4653 * configure: Regenerate.
4654
4655 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4656
4657 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
4658 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
4659 formatting fixes, use pp_colon instead of pp_character (..., ':'),
4660 similarly pp_right_paren.
4661 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
4662 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
4663 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
4664 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
4665 clauses.
4666 (omp_declare_simd_clauses_equal,
4667 omp_remove_redundant_declare_simd_attrs): New functions.
4668 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
4669 (walk_tree_1): Handle new OpenMP 4.0 clauses.
4670 * tree.h (OMP_LOOP_CHECK): Define.
4671 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
4672 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
4673 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
4674 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
4675 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
4676 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
4677 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
4678 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
4679 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
4680 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
4681 OMP_CLAUSE_SIMDLEN_EXPR): Define.
4682 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
4683 (omp_remove_redundant_declare_simd_attrs): New prototype.
4684 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
4685 GIMPLE_OMP_TEAMS): New codes.
4686 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
4687 * omp-low.c (struct omp_context): Add cancel_label and cancellable
4688 fields.
4689 (target_nesting_level): New variable.
4690 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
4691 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
4692 for collapse > 1 static schedule unless ordered.
4693 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
4694 (determine_parallel_type): Adjust get_ws_args_for caller.
4695 (install_var_field): Handle mask & 4 for double indirection.
4696 (scan_sharing_clauses): Ignore shared clause on teams construct.
4697 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
4698 (create_omp_child_function): If inside target or declare target
4699 constructs, set "omp declare target" attribute on the child function.
4700 (find_combined_for): New function.
4701 (scan_omp_parallel): Handle combined loops.
4702 (scan_omp_target, scan_omp_teams): New functions.
4703 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
4704 restrictions and set ctx->cancellable for cancellable constructs.
4705 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
4706 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
4707 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
4708 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
4709 (omp_clause_aligned_alignment): New function.
4710 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
4711 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
4712 on teams constructs. Handle user defined reductions and new
4713 OpenMP 4.0 clauses.
4714 (lower_reduction_clauses): Don't set placeholder to address of ref
4715 if it has already the right type.
4716 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
4717 (expand_parallel_call): Use the new non-_start suffixed builtins,
4718 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
4719 and GOMP_parallel_end after the call.
4720 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
4721 (expand_omp_for_init_counts): Handle combined loops.
4722 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
4723 loops.
4724 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
4725 end of cancellable loops.
4726 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
4727 Likewise. Handle collapse > 1 loops.
4728 (expand_omp_simd): Handle combined loops.
4729 (expand_omp_for): Add inner_stmt argument, adjust callers of
4730 expand_omp_for* functions, use expand_omp_for_static*chunk even
4731 for collapse > 1 unless ordered.
4732 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
4733 of cancellable sections.
4734 (expand_omp_single): Remove need_barrier variable, just rely on
4735 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
4736 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
4737 (expand_omp_atomic_load, expand_omp_atomic_store,
4738 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
4739 (expand_omp_target): New function.
4740 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
4741 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
4742 (build_omp_regions_1): Immediately close region for
4743 GF_OMP_TARGET_KIND_UPDATE.
4744 (maybe_add_implicit_barrier_cancel): New function.
4745 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
4746 cancellation.
4747 (lower_omp_single): Likewise. Add clobber after the barrier.
4748 (lower_omp_taskgroup): New function.
4749 (lower_omp_for): Handle combined loops. Adjust
4750 lower_rec_input_clauses caller. Handle cancellation.
4751 (lower_depend_clauses): New function.
4752 (lower_omp_taskreg): Lower depend clauses. Adjust
4753 lower_rec_input_clauses caller. Add clobber after the call. Handle
4754 cancellation.
4755 (lower_omp_target, lower_omp_teams): New functions.
4756 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
4757 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
4758 and GOMP_cancellation_point calls.
4759 (lower_omp): Fold stmts inside of target region.
4760 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
4761 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4762 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
4763 (BT_FN_VOID_OMPFN_PTR_UINT,
4764 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
4765 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
4766 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
4767 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
4768 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
4769 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
4770 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
4771 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
4772 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
4773 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
4774 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
4775 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
4776 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
4777 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
4778 BUILT_IN_GOMP_PARALLEL_END.
4779 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
4780 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4781 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
4782 GF_OMP_FOR_KIND_DISTRIBUTE.
4783 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
4784 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
4785 (dump_gimple_omp_return): Print lhs if it has any.
4786 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
4787 gimple_omp_atomic_seq_cst_p.
4788 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
4789 and GIMPLE_OMP_TEAMS.
4790 * langhooks.c (lhd_omp_mappable_type): New function.
4791 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
4792 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
4793 hook.
4794 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
4795 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
4796 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
4797 (struct gimplify_omp_ctx): Add combined_loop field.
4798 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
4799 on stmts inside of target region.
4800 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
4801 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
4802 ORT_TARGET and ORT_TARGET_DATA.
4803 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
4804 Handle GOVD_MAP.
4805 (omp_notice_threadprivate_variable): Complain about threadprivate
4806 variables in target region.
4807 (omp_notice_variable): Complain about vars with non-mappable type
4808 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
4809 (omp_check_private): Ignore ORT_TARGET* regions.
4810 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
4811 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
4812 (find_combined_omp_for): New function.
4813 (gimplify_omp_for): Handle gimplification of combined loops.
4814 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
4815 OMP_TEAMS.
4816 (gimplify_omp_target_update): New function.
4817 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
4818 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
4819 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
4820 (gimplify_body): If fndecl has "omp declare target" attribute, add
4821 implicit ORT_TARGET context around it.
4822 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
4823 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
4824 * tree-nested.c (convert_nonlocal_reference_stmt,
4825 convert_local_reference_stmt, convert_gimple_call): Handle
4826 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4827 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
4828 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
4829 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
4830 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
4831 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
4832 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
4833 BUILT_IN_GOMP_LOOP_END_CANCEL,
4834 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
4835 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
4836 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
4837 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
4838 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
4839 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
4840 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
4841 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
4842 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
4843 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
4844 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
4845 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
4846 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
4847 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
4848 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
4849 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4850 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
4851 gimple_build_omp_teams): New functions.
4852 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
4853 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
4854 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
4855 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4856 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
4857 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
4858 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
4859 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
4860 GF_OMP_ATOMIC_SEQ_CST): New.
4861 (gimple_build_omp_taskgroup, gimple_build_omp_target,
4862 gimple_build_omp_teams): New prototypes.
4863 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
4864 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4865 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
4866 GIMPLE_OMP_SINGLE as end of range.
4867 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
4868 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
4869 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
4870 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
4871 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
4872 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
4873 gimple_omp_target_kind, gimple_omp_target_set_kind,
4874 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
4875 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
4876 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
4877 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
4878 gimple_omp_teams_set_clauses): New inlines.
4879 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
4880 and GIMPLE_OMP_TASKGROUP.
4881 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
4882 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
4883 enum omp_clause_proc_bind_kind): New.
4884 (union omp_clause_subcode): Add depend_kind, map_kind and
4885 proc_bind_kind fields.
4886 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
4887 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4888 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
4889 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
4890 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
4891
4892 2013-10-10 Teresa Johnson <tejohnson@google.com>
4893
4894 * predict.c (tree_estimate_probability): Add new parameter
4895 for estimate_bb_frequencies.
4896 (estimate_bb_frequencies): Add new parameter to force estimation.
4897 (rebuild_frequencies): When max frequency in function is small,
4898 recompute counts from frequencies.
4899 * predict.h (estimate_bb_frequencies): New parameter.
4900
4901 2013-10-10 David Malcolm <dmalcolm@redhat.com>
4902
4903 * ipa-inline.c (ipa_inline): Fix leak of "order" when
4904 optimizations are disabled.
4905
4906 2013-10-10 David Malcolm <dmalcolm@redhat.com>
4907
4908 * coverage.c (coverage_finish): Fix leak of da_file_name.
4909
4910 2013-10-10 Jan Hubicka <jh@suse.cz>
4911
4912 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
4913 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
4914 Bobcat and generic.
4915
4916 2013-10-10 Jakub Jelinek <jakub@redhat.com>
4917
4918 PR middle-end/58670
4919 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
4920 if any labels are in FALLTHRU_BB, use a special label emitted
4921 immediately after the asm goto insn rather than label_rtx
4922 of the LABEL_DECL.
4923 (expand_asm_stmt): Adjust caller.
4924 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
4925 edge if the last insn in predecessor is a jump with single successor,
4926 but it isn't simplejump_p.
4927
4928 2013-10-10 Richard Biener <rguenther@suse.de>
4929
4930 PR tree-optimization/58656
4931 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
4932
4933 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4934
4935 * gimplify.c: Include expr.h and tm_p.h for targets with special
4936 va-arg padding requirements.
4937
4938 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4939
4940 * tree-flow.h: Move some prototypes to gimple.h.
4941 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
4942 * gimple.h: Relocate some prototypes from tree-flow.h
4943 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
4944 Move to gimplify.c.
4945 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
4946 (build_va_arg_indirect_ref): Relocate and make static.
4947 (std_gimplify_va_arg_expr): Relocate here.
4948 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
4949 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
4950
4951 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4952
4953 * doc/md.texi: Document the mnemonic attribute.
4954
4955 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4956
4957 PR target/57377
4958 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
4959 (set_attr_alternative x ...) when searching for user defined
4960 mnemonic attribute.
4961
4962 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4963
4964 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
4965
4966 2013-10-09 Easwaran Raman <eraman@google.com>
4967
4968 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
4969 * cfgexpand.c (defer_stack_allocation): ...use here
4970 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
4971
4972 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
4973
4974 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
4975 (operand_equal_for_value_replacement): New function, extracted from
4976 value_replacement and enhanced to catch more cases.
4977 (value_replacement): Use operand_equal_for_value_replacement.
4978
4979 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4980
4981 * loop-doloop.c (doloop_modify, doloop_optimize): Use
4982 get_max_loop_iterations.
4983
4984 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4985
4986 * config/arm/aarch-common.c (arm_early_load_addr_dep):
4987 Place comment above function.
4988
4989 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4990
4991 * tree-flow.h: Remove all remaining prototypes, enums and structs that
4992 are not related to tree-cfg.c.
4993 * tree-ssa-address.h: New file. Relocate prototypes.
4994 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
4995 (addr_for_mem_ref): New. Combine call to get_address_description and
4996 return addr_for_mem_ref.
4997 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
4998 * tree-ssa-live.h: Adjust prototypes.
4999 * passes.c: Include tree-ssa-live.h.
5000 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
5001 * graphite.c (graphite_transform_loops): Make static.
5002 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5003 make_pass_graphite, pass_data_graphite_transforms,
5004 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
5005 * ipa-pure-const.c (warn_function_noreturn): Make static.
5006 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5007 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5008 Relocate from tree-cfg.c
5009 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
5010 static.
5011 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5012 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5013 Move to ipa-pure-const.c.
5014 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
5015 Relocate from tree-optimize.c.
5016 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
5017 make_pass_fixup_cfg): Move to tree-cfg.c.
5018 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
5019 Relocate some prototypes.
5020 * tree-data-ref.h (tree_check_data_deps) Add prototype.
5021 * tree-dump.c (dump_function_to_file): Remove prototype.
5022 Add tree-flow.h to the include file.
5023 * tree-dump.h: Remove prototype.
5024 * tree-parloops.h: New File. Add prototypes.
5025 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
5026 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
5027 from tree-ssa-loop.c.
5028 * tree-predcom.c (run_tree_predictive_commoning,
5029 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
5030 Relocate here from tree-ssa-loop.c.
5031 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
5032 ssa_name_values.release ().
5033 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
5034 (ssa_name_values): Relocate from tree-flow.h.
5035 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
5036 * tree-ssa-loop.c (run_tree_predictive_commoning,
5037 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
5038 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5039 make_pass_graphite, pass_data_graphite_transforms,
5040 make_pass_graphite_transforms, gate_tree_parallelize_loops,
5041 tree_parallelize_loops, pass_data_parallelize_loops,
5042 make_pass_parallelize_loops): Move to other files.
5043 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
5044 moved here.
5045 * tree.h: Remove prototypes from tree-address.c.
5046
5047 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5048
5049 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
5050 (struct int_tree_map): Move to tree-hasher.h
5051 (SCALE, LABEL, PERCENT): Move to gimple.h
5052 * tree-flow-inline.h: Delete. Move functions to other files.
5053 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
5054 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
5055 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
5056 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
5057 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
5058 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
5059 * tree-hasher.h: Don't include tree-flow.h.
5060 (struct int_tree_map): Relocate from tree-flow.h.
5061 * tree-sra.c (contains_view_convert_expr_p): Relocate from
5062 tree-flow-inline.h and make static.
5063 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
5064 tree-flow-inline.h.
5065 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
5066 tree-flow-inline.h and make static.
5067 * tree.h (is_global_var, may_be_aliased): Relocate from
5068 tree-flow-inline.h.
5069 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
5070 * value-prof.c: No longer include tree-flow-inline.h.
5071 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
5072
5073 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5074
5075 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
5076 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
5077 (enum move_pos): Move to tree-ssa-loop-im.h
5078 * cfgloop.h: Move some prototypes.
5079 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
5080 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
5081 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
5082 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
5083 (loop_containing_stmt): Relocate from tree-flow-inline.h.
5084 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
5085 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
5086 (enum move_pos): Relocate here.
5087 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
5088 tree-ssa-loop.c.
5089 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
5090 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
5091 make_pass_lim): Relocate here from tree-ssa-loop.c.
5092 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
5093 tree-ssa-loop.c.
5094 (loop_edge_to_cancel, unloop_loops): Make static.
5095 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5096 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
5097 (tree_complete_unroll, gate_tree_complete_unroll,
5098 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
5099 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5100 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
5101 * tree-ssa-loop-ivopts.c: Remove local prototypes.
5102 (stmt_invariant_in_loop_p): Remove unused function.
5103 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
5104 * tree-ssa-loop-manip.h: New file. Add prototypes.
5105 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
5106 (gcov_type_to_double_int): Move to cfgloop.h.
5107 (double_int_cmp, bound_index,
5108 estimate_numbers_of_iterations_loop): Make static.
5109 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
5110 (max_loop_iterations): Factor out get_max_loop_iterations.
5111 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
5112 cfgloop.c.
5113 * tree-ssa-loop-niter.h: New file. Add prototypes.
5114 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
5115 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
5116 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
5117 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
5118 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
5119 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
5120 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
5121 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
5122 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
5123 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
5124 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5125 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
5126 pass_data_complete_unroll, make_pass_complete_unroll,
5127 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5128 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
5129 tree-ssa-loop-ivcanon.c.
5130 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
5131 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
5132 tree-ssa-loop-prefetch.c.
5133 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
5134 tree-ssa-loop-im.c.
5135 (get_lsm_tmp_name): Relocate and add suffix parameter.
5136 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
5137 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
5138 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
5139 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
5140 (get_estimated_loop_iterations): Factor out accessor from
5141 estimated_loop_iterations in tree-ssa-loop-niter.c.
5142 (get_max_loop_iterations): Factor out accessor from
5143 _max_loop_iterations in tree-ssa-niter.c.
5144 * loop-unroll.c (decide_unroll_constant_iterations,
5145 decide_unroll_runtime_iterations, decide_peel_simple,
5146 decide_unroll_stupid): Use new get_* accessors.
5147
5148 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
5149
5150 PR tree-optimization/20318
5151 * doc/extend.texi (returns_nonnull): New function attribute.
5152 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
5153 attribute.
5154 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
5155 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
5156
5157 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
5158
5159 PR middle-end/58570
5160 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
5161 false if both components are bitfields.
5162
5163 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5164
5165 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
5166 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
5167 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
5168 * config/aarch64/aarch64.h
5169 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
5170 * config/aarch64/aarch64-simd-builtins.def
5171 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
5172
5173 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5174
5175 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
5176
5177 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5178
5179 * config/aarch64/arm_neon.h (vdiv_f64): Added.
5180
5181 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5182
5183 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
5184 (vneg_f64): New intrinsic.
5185 (vneg_s8): Asm replaced with C.
5186 (vneg_s16): Likewise.
5187 (vneg_s32): Likewise.
5188 (vneg_s64): New intrinsic.
5189 (vnegq_f32): Asm replaced with C.
5190 (vnegq_f64): Likewise.
5191 (vnegq_s8): Likewise.
5192 (vnegq_s16): Likewise.
5193 (vnegq_s32): Likewise.
5194 (vnegq_s64): Likewise.
5195
5196 2013-10-09 Renlin Li <Renlin.Li@arm.com>
5197
5198 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
5199
5200 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5201
5202 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
5203 packed stack special handling.
5204 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
5205 back to fixed stack slots for FPRs saved due to stdarg.
5206
5207 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5208
5209 * config/s390/s390.c (s390_frame_info): Restructure function.
5210
5211 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5212
5213 * config/s390/s390.c (struct s390_frame_layout): New field
5214 gpr_save_slots.
5215 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
5216 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
5217 regs_ever_clobbered to char*.
5218 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
5219 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
5220 (s390_register_info): Enable FPR save slots. Move/Copy some
5221 functionality into ...
5222 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
5223 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
5224 function.
5225 (s390_frame_info): Do gpr slot allocation here now. stdarg does
5226 not imply a stack frame.
5227 (s390_init_frame_layout): Remove variable clobbered_regs.
5228 (s390_update_register_info): Remove function.
5229 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
5230 cannot be used for register renaming.
5231 (s390_hard_regno_scratch_ok): New function.
5232 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
5233 (s390_initial_elimination_offset): Change offset calculation of
5234 the return address pointer.
5235 (save_gprs): Deal with only r6 being saved from the call-saved regs.
5236 (restore_gprs): Set frame related flag.
5237 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
5238 (s390_emit_prologue): Call s390_register_info instead of
5239 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
5240 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
5241 (s390_optimize_prologue): Call s390_optimize_register_info.
5242 Try to remove also FPR slot save/restore INSNs. Remove frame
5243 related flags from restore INSNs.
5244
5245 2013-10-08 DJ Delorie <dj@redhat.com>
5246
5247 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
5248 (movhi): Likewise.
5249
5250 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
5251 avoid conflict with the MI use of %c.
5252 * config/rl78/rl78-real.md: change %c to %C throughout.
5253 * config/rl78/rl78-virt.md: Likewise.
5254
5255 2013-10-08 Jan Hubicka <jh@suse.cz>
5256
5257 * config/i386/i386.c (ix86_option_override_internal): Switch
5258 to SSE math for -ffast-math when target ISA supports SSE2.
5259
5260 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
5261
5262 * tree-flow.h: Remove some prototypes.
5263 * tree.h: Remove some protypes, add a couple.
5264 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
5265 using_eh_for_cleanups_p): Add interface routines for front ends.
5266 * tree-eh.h: New file. Add protoptyes.
5267 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
5268 (add_stmt_to_eh_lp_fn): Make static.
5269 (lower_try_finally): Use new using_eh_for_cleanups_p.
5270 * emit-rtl.c: Include tree-eh.h.
5271 * gimple.h: Include tree-eh.h.
5272
5273 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
5274
5275 PR tree-optimization/58480
5276 * tree-vrp.c (infer_nonnull_range): New function.
5277 (infer_value_range): Call infer_nonnull_range.
5278
5279 2013-10-08 Dehao Chen <dehao@google.com>
5280
5281 PR tree-optimization/58619
5282 * tree-inline.c (copy_phis_for_bb): Combine location data
5283 only if non-null.
5284
5285 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5286
5287 PR target/58423
5288 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
5289 RTX_FRAME_RELATED_P on INSN.
5290
5291 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5292
5293 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
5294 (altivec_expand_vec_perm_const): Call it.
5295
5296 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5297
5298 * config/rs6000/vector.md (mov<mode>): Emit permuted move
5299 sequences for LE VSX loads and stores at expand time.
5300 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
5301 prototype.
5302 * config/rs6000/rs6000.c (rs6000_const_vec): New.
5303 (rs6000_gen_le_vsx_permute): New.
5304 (rs6000_gen_le_vsx_load): New.
5305 (rs6000_gen_le_vsx_store): New.
5306 (rs6000_gen_le_vsx_move): New.
5307 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
5308 (*vsx_le_perm_load_v4si): New.
5309 (*vsx_le_perm_load_v8hi): New.
5310 (*vsx_le_perm_load_v16qi): New.
5311 (*vsx_le_perm_store_v2di): New.
5312 (*vsx_le_perm_store_v4si): New.
5313 (*vsx_le_perm_store_v8hi): New.
5314 (*vsx_le_perm_store_v16qi): New.
5315 (*vsx_xxpermdi2_le_<mode>): New.
5316 (*vsx_xxpermdi4_le_<mode>): New.
5317 (*vsx_xxpermdi8_le_V8HI): New.
5318 (*vsx_xxpermdi16_le_V16QI): New.
5319 (*vsx_lxvd2x2_le_<mode>): New.
5320 (*vsx_lxvd2x4_le_<mode>): New.
5321 (*vsx_lxvd2x8_le_V8HI): New.
5322 (*vsx_lxvd2x16_le_V16QI): New.
5323 (*vsx_stxvd2x2_le_<mode>): New.
5324 (*vsx_stxvd2x4_le_<mode>): New.
5325 (*vsx_stxvd2x8_le_V8HI): New.
5326 (*vsx_stxvd2x16_le_V16QI): New.
5327
5328 2013-10-07 Renlin Li <Renlin.Li@arm.com>
5329
5330 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
5331
5332 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5333
5334 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
5335 loop to work also for 31bit ABI.
5336 Save the stack pointer for frame_size > 0.
5337
5338 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5339
5340 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
5341 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
5342 constraint letters from expanders.
5343 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
5344 retry count to general_operand.
5345 ("tabort"): Give operand 0 a mode.
5346 ("tabort_1"): Add mode and constraint letter for operand 0.
5347 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
5348
5349 2013-10-04 Jeff Law <law@redhat.com>
5350
5351 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
5352
5353 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
5354 out of ...
5355 (thread_across_edge): Here. Call it.
5356
5357 2013-10-04 Cary Coutant <ccoutant@google.com>
5358
5359 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
5360 discarding a location list expression (or a piece of one).
5361
5362 2013-10-03 Jan Hubicka <jh@suse.cz>
5363
5364 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
5365 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
5366 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
5367
5368 2013-10-03 Jan Hubicka <jh@suse.cz>
5369
5370 * config/i386/i386.c (ix86_option_override_internal): Do not enable
5371 accumulate-outgoing-args when producing unwind info.
5372
5373 2013-10-03 Wei Mi <wmi@google.com>
5374
5375 * lra-constraints.c (insert_move_for_subreg): New function
5376 extracted from simplify_operand_subreg.
5377 (simplify_operand_subreg): Add reload for paradoxical subreg.
5378
5379 2013-10-03 Rong Xu <xur@google.com>
5380
5381 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
5382 the candidate of builtin_expect such that we should fix the
5383 size/time estimation.
5384 (estimate_function_body_sizes): Do the acutally size/time fix-up
5385 for builtin_expect.
5386
5387 2013-10-03 Rong Xu <xur@google.com>
5388
5389 * predict.c (tree_predict_by_opcode): Get the probability
5390 for builtin_expect from param builtin_expect_probability.
5391 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
5392 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
5393 * doc/invoke.texi: Add documentation for builtin-expect-probability.
5394
5395 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
5396
5397 PR c++/19476
5398 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
5399 * calls.c (alloca_call_p): Use get_callee_fndecl.
5400 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
5401 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
5402 Likewise.
5403 (vrp_visit_stmt): Remove duplicated code.
5404
5405 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
5406
5407 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
5408 ceildf2, btruncdf2, instead of vsx_* name.
5409
5410 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
5411 iterators to only do V2DF and V4SF here. Move the DF code to
5412 rs6000.md where it is combined with SF mode. Replace <VSv> with
5413 just 'v' since only vector operations are handled with these insns
5414 after moving the DF support to rs6000.md.
5415 (vsx_sub<mode>3): Likewise.
5416 (vsx_mul<mode>3): Likewise.
5417 (vsx_div<mode>3): Likewise.
5418 (vsx_fre<mode>2): Likewise.
5419 (vsx_neg<mode>2): Likewise.
5420 (vsx_abs<mode>2): Likewise.
5421 (vsx_nabs<mode>2): Likewise.
5422 (vsx_smax<mode>3): Likewise.
5423 (vsx_smin<mode>3): Likewise.
5424 (vsx_sqrt<mode>2): Likewise.
5425 (vsx_rsqrte<mode>2): Likewise.
5426 (vsx_fms<mode>4): Likewise.
5427 (vsx_nfma<mode>4): Likewise.
5428 (vsx_copysign<mode>3): Likewise.
5429 (vsx_btrunc<mode>2): Likewise.
5430 (vsx_floor<mode>2): Likewise.
5431 (vsx_ceil<mode>2): Likewise.
5432 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
5433 (vsx_sminsf3): Likewise.
5434 (vsx_fmadf4): Likewise.
5435 (vsx_fmsdf4): Likewise.
5436 (vsx_nfmadf4): Likewise.
5437 (vsx_nfmsdf4): Likewise.
5438 (vsx_cmpdf_internal1): Likewise.
5439
5440 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
5441 simpler to select whether a target has SPE or traditional floating
5442 point support in iterators.
5443 (TARGET_DF_SPE): Likewise.
5444 (TARGET_SF_FPR): Likewise.
5445 (TARGET_DF_FPR): Likewise.
5446 (TARGET_SF_INSN): Macros to say whether floating point support
5447 exists for a given operation for expanders.
5448 (TARGET_DF_INSN): Likewise.
5449
5450 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
5451 combining of SF/DF mode operations, using both traditional and VSX
5452 registers.
5453 (Fvsx): Likewise.
5454 (Ff): Likewise.
5455 (Fv): Likewise.
5456 (Fs): Likewise.
5457 (Ffre): Likewise.
5458 (FFRE): Likewise.
5459 (abs<mode>2): Combine SF/DF modes using traditional floating point
5460 instructions. Add support for using the upper DF registers with
5461 VSX support, and SF registers with power8-vector support. Update
5462 expanders for operations supported by both the SPE and traditional
5463 floating point units.
5464 (abs<mode>2_fpr): Likewise.
5465 (nabs<mode>2): Likewise.
5466 (nabs<mode>2_fpr): Likewise.
5467 (neg<mode>2): Likewise.
5468 (neg<mode>2_fpr): Likewise.
5469 (add<mode>3): Likewise.
5470 (add<mode>3_fpr): Likewise.
5471 (sub<mode>3): Likewise.
5472 (sub<mode>3_fpr): Likewise.
5473 (mul<mode>3): Likewise.
5474 (mul<mode>3_fpr): Likewise.
5475 (div<mode>3): Likewise.
5476 (div<mode>3_fpr): Likewise.
5477 (sqrt<mode>3): Likewise.
5478 (sqrt<mode>3_fpr): Likewise.
5479 (fre<Fs>): Likewise.
5480 (rsqrt<mode>2): Likewise.
5481 (cmp<mode>_fpr): Likewise.
5482 (smax<mode>3): Likewise.
5483 (smin<mode>3): Likewise.
5484 (smax<mode>3_vsx): Likewise.
5485 (smin<mode>3_vsx): Likewise.
5486 (negsf2): Delete SF operations that are merged with DF.
5487 (abssf2): Likewise.
5488 (addsf3): Likewise.
5489 (subsf3): Likewise.
5490 (mulsf3): Likewise.
5491 (divsf3): Likewise.
5492 (fres): Likewise.
5493 (fmasf4_fpr): Likewise.
5494 (fmssf4_fpr): Likewise.
5495 (nfmasf4_fpr): Likewise.
5496 (nfmssf4_fpr): Likewise.
5497 (sqrtsf2): Likewise.
5498 (rsqrtsf_internal1): Likewise.
5499 (smaxsf3): Likewise.
5500 (sminsf3): Likewise.
5501 (cmpsf_internal1): Likewise.
5502 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
5503 (negdf2): Delete DF operations that are merged with SF.
5504 (absdf2): Likewise.
5505 (nabsdf2): Likewise.
5506 (adddf3): Likewise.
5507 (subdf3): Likewise.
5508 (muldf3): Likewise.
5509 (divdf3): Likewise.
5510 (fred): Likewise.
5511 (rsqrtdf_internal1): Likewise.
5512 (fmadf4_fpr): Likewise.
5513 (fmsdf4_fpr): Likewise.
5514 (nfmadf4_fpr): Likewise.
5515 (nfmsdf4_fpr): Likewise.
5516 (sqrtdf2): Likewise.
5517 (smaxdf3): Likewise.
5518 (smindf3): Likewise.
5519 (cmpdf_internal1): Likewise.
5520 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
5521 (btrunc<mode>2): Delete separate expander, and combine with the
5522 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
5523 (btrunc<mode>2_fpr): Likewise.
5524 (ceil<mode>2): Likewise.
5525 (ceil<mode>2_fpr): Likewise.
5526 (floor<mode>2): Likewise.
5527 (floor<mode>2_fpr): Likewise.
5528 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
5529 Add support for using the upper registers with VSX and
5530 power8-vector. Move insns to be closer to the define_expands. On
5531 VSX systems, prefer the traditional form of FMA over the VSX
5532 version, since the traditional form allows the target not to
5533 overlap with the inputs.
5534 (fms<mode>4_fpr): Likewise.
5535 (nfma<mode>4_fpr): Likewise.
5536 (nfms<mode>4_fpr): Likewise.
5537
5538 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5539 Richard Earnshaw <richard.earnshaw@arm.com>
5540
5541 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
5542 (struct mult_cost_table): Likewise.
5543 (struct mem_cost_table): Likewise.
5544 (struct fp_cost_table): Likewise.
5545 (struct vector_cost_table): Likewise.
5546 (cpu_cost_table): Likewise.
5547 * config/arm/arm.opt (mold-rts-costs): New option.
5548 (mnew-generic-costs): Likewise.
5549 * config/arm/arm.c (generic_extra_costs): New table.
5550 (cortexa15_extra_costs): Likewise.
5551 (arm_slowmul_tune): Use NULL as new costs.
5552 (arm_fastmul_tune): Likewise.
5553 (arm_strongarm_tune): Likewise.
5554 (arm_xscale_tune): Likewise.
5555 (arm_9e_tune): Likewise.
5556 (arm_v6t2_tune): Likewise.
5557 (arm_cortex_a5_tune): Likewise.
5558 (arm_cortex_a9_tune): Likewise.
5559 (arm_v6m_tune): Likewise.
5560 (arm_fa726te_tune): Likewise.
5561 (arm_cortex_a15_tune): Use cortex15_extra_costs.
5562 (arm_cortex_tune): Use generict_extra_costs.
5563 (shifter_op_p): New function.
5564 (arm_unspec_cost): Likewise.
5565 (LIBCALL_COST): Define.
5566 (arm_new_rtx_costs): New function.
5567 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
5568 table is available. Use old costs otherwise unless mnew-generic-costs
5569 is specified.
5570 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
5571 (cpu_cost_table): Declare.
5572
5573 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
5574
5575 PR target/58460
5576 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
5577 (*subs_mul_imm_<mode>)
5578 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
5579 (*sub_<shift>_<mode>)
5580 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
5581 Remove k constraint.
5582
5583 2013-10-03 Ian Bolton <ian.bolton@arm.com>
5584
5585 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
5586 code.
5587 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
5588
5589 2013-10-02 Teresa Johnson <tejohnson@google.com>
5590
5591 * predict.c (probably_never_executed): New function.
5592 (probably_never_executed_bb_p): Invoke probably_never_executed.
5593 (probably_never_executed_edge_p): Ditto.
5594 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
5595 Treat profile insanities conservatively.
5596
5597 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
5598
5599 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
5600
5601 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
5602
5603 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
5604 use it. Use smaller increase for scratch. Don't increase reject
5605 for early clobber scratch.
5606 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
5607 setting eliminated regs except setting fp from hfp.
5608 (lra_eliminate): Check lra_insn_recog_data on NULL.
5609
5610 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5611
5612 PR target/58587
5613 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
5614 setting -mvsx-timode by default until the underlying problem is fixed.
5615 (RS6000_CPU, power7 defaults): Likewise.
5616
5617 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5618
5619 * config/x-linux (host-linux.o): Remove header dependencies.
5620 Use $(COMPILE) and $(POSTCOMPILE).
5621 * config/t-linux-android (linux-android.o): Ditto.
5622
5623 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5624
5625 * Makefile.in (expmed.o-warn): Remove.
5626
5627 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5628
5629 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
5630 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
5631
5632 2013-10-02 Teresa Johnson <tejohnson@google.com>
5633
5634 * dojump.c (do_jump_1): Divide probability between
5635 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
5636
5637 2013-10-02 Tom Tromey <tromey@redhat.com>
5638
5639 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
5640
5641 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5642
5643 * tree-flow.h: Remove some prototypes.
5644 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
5645 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
5646 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
5647 mark_virtual_phi_result_for_renaming): Relocate here.
5648 * tree-into-ssa.h: Add prototypes.
5649 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
5650 single_pred_before_succ_order.
5651 (blocks_in_phiopt_order): Rename and move to cfganal.c.
5652 (nonfreeing_call_p) Move to gimple.c.
5653 * cfganal.c (single_pred_before_succ_order): Move and renamed from
5654 tree-ssa-phiopt.c.
5655 * basic-block.h (single_pred_before_succ_order): Add prototype.
5656 * gimple.c (nonfreeing_call_p): Relocate here.
5657 * gimple.h: Add prototype.
5658 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
5659 * tree-ssa-dom.h: New file. Relocate prototypes here.
5660 * tree-ssa.h: Include tree-ssa-dom.h.
5661
5662 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5663
5664 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
5665 Use $(COMPILE) and $(POSTCOMPILE).
5666
5667 * config/alpha/x-alpha (driver-alpha.o): Ditto.
5668
5669 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5670
5671 * tree-flow.h: Remove some prototypes.
5672 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
5673 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5674 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
5675 propagate_tree_value*): Move from here to...
5676 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
5677 propagate_tree_value*): Relocate here.
5678 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5679 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
5680 * gimple-fold.c: Remove gimple-fold.h from include list.
5681 * tree-vrp.c: Remove gimple-fold.h from include list.
5682 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
5683 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
5684 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
5685 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
5686 * sese.c: Add tree-ssa-propagate.h to include list.
5687
5688 2013-10-02 Richard Biener <rguenther@suse.de>
5689
5690 * tree-loop-distribution.c: Include tree-vectorizer.h for
5691 find_loop_location.
5692 (enum partition_kind): Remove PKIND_REDUCTION.
5693 (struct partition_s): Remove has_writes member, add reduction_p member.
5694 (partition_alloc): Adjust.
5695 (partition_builtin_p): Likewise.
5696 (partition_has_writes): Remove.
5697 (partition_reduction_p): New function.
5698 (partition_merge_into): Likewise.
5699 (generate_code_for_partition): Commonize builtin partition
5700 handling tail.
5701 (rdg_cannot_recompute_vertex_p): Remove.
5702 (already_processed_vertex_p): Likewise.
5703 (rdg_flag_vertex): Do not set has_writes.
5704 (classify_partition): Adjust.
5705 (rdg_build_partitions): Do not set has_writes, treat all
5706 partitions as useful.
5707 (distribute_loop): Record number of library calls generated. Adjust.
5708 (tree_loop_distribution): Report number of loops and library
5709 calls generated as opt-info.
5710
5711 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5712
5713 * tree-flow.h: Include new .h files. Move prototypes.
5714 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
5715 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
5716 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
5717 * tree-pretty-print.h: Add prototypes from tree-flow.h.
5718 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
5719 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
5720 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
5721 * tree.h (get_ref_base_and_extent): Move prototype out.
5722 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
5723 tree-dfa.h.
5724 * gimple-low.h: New File. Add prototypes from tree-flow.h.
5725 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
5726 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
5727 * tree-scalar-evolution.c: Include tree.h.
5728 * sese.c: Include tree.h.
5729 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
5730 * dwarf2out.c: Include tree-dfa.h.
5731 * tree-chrec.c: Include tree.h.
5732 * tree-data-ref.c: Include tree.h.
5733
5734 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
5735
5736 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
5737 Fix whitespace.
5738
5739 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5740
5741 * config/t-sol2 (sol2-c.o): Remove header dependencies.
5742 Use $(COMPILE) and $(POSTCOMPILE).
5743 (sol2-cxx.o): Likewise.
5744 (sol2-stubs.o): Likewise.
5745 (sol2.o): Likewise.
5746 * config/x-solaris (host-solaris.o): Likewise.
5747
5748 * config/sparc/t-sparc (sparc.o): Remove.
5749 (sparc-c.o): Remove header dependencies.
5750 Use $(COMPILE) and $(POSTCOMPILE).
5751 * config/sparc/x-sparc: Likewise.
5752
5753 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
5754
5755 * config/arc/arc-opts.h: Add 2013 to Copyright years.
5756 * config/arc/arc700.md: Likewise.
5757 * config/arc/arc-modes.def: Likewise.
5758 * config/arc/arc-simd.h: Likewise.
5759 * config/arc/t-arc-uClibc: Likewise.
5760 * config/arc/t-arc-newlib: Likewise.
5761
5762 2013-10-02 Renlin Li <renlin.li@arm.com>
5763
5764 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
5765 plus_constant.
5766 (aarch64_expand_epilogue): Likewise.
5767
5768 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5769 Yufeng Zhang <yufeng.zhang@arm.com>
5770
5771 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
5772 declaration.
5773 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
5774 'base_in' represent a conversion and legal_cast_p_1 holds; set
5775 'base_in' with the returned value from get_unwidened.
5776
5777 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5778
5779 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
5780 plus_constant is not used.
5781
5782 2013-10-01 Wei Mi <wmi@google.com>
5783
5784 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
5785 * config/i386/i386.md: Add define_peephole2 to
5786 break partial reg stall for cvtss2sd/cvtsd2ss.
5787
5788 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5789
5790 * config/arc/arc.c (pass_arc_ifcvt::clone):
5791 Update for ctxt_ -> m_ctxt change.
5792
5793 2013-10-01 Jeff Law <law@redhat.com>
5794
5795 * tree-ssa-threadupdate.c (struct redirection_data): Delete
5796 outgoing_edge and intermediate_edge fields. Instead store the path.
5797 (redirection_data::hash): Hash on the last edge's destination index.
5798 (redirection_data::equal): Check the entire thread path.
5799 (lookup_redirectio_data): Corresponding changes.
5800 (create_edge_and_update_destination_phis): Likewise.
5801 (thread_single_edge): Likewise.
5802
5803 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5804 Diego Novillo <dnovillo@google.com>
5805
5806 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
5807 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
5808 (vld32wh_insn, vld32wl_insn): Delete commented-out old
5809 versions of these patterns.
5810
5811 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
5812 (__builtin_arc_aligned): Likewise.
5813
5814 * config/arc/arc.md: Expand adc_0 comment stating the intended
5815 purpose and why it isn't ready.
5816 Replace commented out call_value_via_label_mixed with a
5817 plain comment about bl_s.
5818
5819 * config/arc/arc.c (stdio.h): Don't include directly.
5820 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
5821 Remove if (1) condition.
5822 (arc_encode_section_info): Fix comment.
5823
5824 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5825
5826 * config/arc/arc.c (arc_conditional_register_usage):
5827 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
5828 Also set reg_alloc_order for DMA config regs.
5829
5830 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5831 Jeremy Bennett <jeremy.bennett@embecosm.com>
5832
5833 * doc/install.texi (--with-cpu): Mention ARC.
5834 (arc-*-elf32): New paragraph.
5835 (arc-linux-uclibc): Likewise.
5836 * doc/md.texi (Machine Constraints): Add ARC part.
5837 * doc/invoke.texi: (menu): Add ARC Options.
5838 (Machine Dependent Options) <ARC Options>: Add synopsis.
5839 (node ARC Options): Add.
5840 * doc/extend.texi (long_call / short_call attribute): Add ARC.
5841 (ARC Built-in Functions): New section defining
5842 generic ARC built-in functions.
5843 (ARC SIMD Built-in Functions): New section defining SIMD specific
5844 built-in functions.
5845 (Declaring Attributes of Functions): Extended
5846 description of short_call and long_call attributes for ARC and
5847 added index entries.
5848
5849 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
5850 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
5851 Joern Rennecke <joern.rennecke@embecosm.com>
5852 Muhammad Khurram Riaz <khurram.riaz@arc.com>
5853 Brendan Kehoe <brendan@zen.org>
5854 Michael Eager <eager@eagercon.com>
5855 Simon Cook <simon.cook@embecosm.com>
5856 Jeremy Bennett <jeremy.bennett@embecosm.com>
5857
5858 * config/arc, common/config/arc: New directories.
5859
5860 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5861 Brendan Kehoe <brendan@zen.org>
5862 Simon Cook <simon.cook@embecosm.com>
5863
5864 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
5865
5866 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
5867
5868 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
5869 * tree-ssa-coalesce.h: New. Move prototype to here.
5870 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
5871 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
5872 (gimple_can_coalesce_p): Move to...
5873 * gimple.c (gimple_can_coalesce_p): Here.
5874
5875 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
5876
5877 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
5878 (dump_decl_set): Move to gimple.c.
5879 * gimple.h: Don't include tree-ssa-operands.h.
5880 (dump_decl_set): Add prototype.
5881 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
5882 Move to gimple-ssa.h.
5883 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
5884 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
5885 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
5886 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
5887 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
5888 than PHI_ARG_DEF.
5889 (dump_decl_set): Relocate here.
5890 * gimple-ssa.h: New file.
5891 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
5892 Relocate from gimple.h.
5893 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
5894 * tree-ssa-operands.c (swap_ssa_operands): Rename from
5895 swap_tree_operands and remove non-ssa path.
5896 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
5897 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
5898 swap_ssa_operands.
5899 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
5900 vect_is_simple_reduction_1): Use swap_ssa_operands.
5901 * tree-flow.h: Move various prototypes to tree-phinodes.h.
5902 (enum need_phi_state): Move to tree-into-ssa.c.
5903 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
5904 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
5905 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
5906 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
5907 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
5908 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
5909 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
5910 num_imm_uses): Move to ssa-iterators.h.
5911 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
5912 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
5913 tree-phinodes.h.
5914 (op_iter_done, op_iter_next_def, op_iter_next_tree,
5915 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
5916 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
5917 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
5918 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
5919 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
5920 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
5921 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
5922 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
5923 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
5924 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
5925 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
5926 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
5927 (set_phi_nodes): Move to tree-phinodes.h.
5928 * tree-ssa-operands.h (enum ssa_op_iter_type,
5929 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
5930 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
5931 (dump_decl_set): Remove prototype.
5932 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
5933 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
5934 (set_phi_nodes): Relocate from tree-flow-inline.h.
5935 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
5936 tree-flow-inline.h
5937 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
5938 include list. Temporarily add gimple.h to include list.
5939 * ssa-iterators.h: New file.
5940 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
5941 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
5942 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
5943 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
5944 Relocate from tree-ssa-operands.h.
5945 (delink_imm_use, link_imm_use_to_list, link_imm_use,
5946 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
5947 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
5948 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
5949 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
5950 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
5951 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
5952 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
5953 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
5954 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
5955 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
5956 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
5957 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
5958 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
5959 Relocate from tree-flow-inline.h.
5960 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
5961
5962 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
5963
5964 * aarch64-simd.md
5965 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
5966 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
5967 Insert '\t' to output template.
5968 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
5969 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
5970 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
5971 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
5972
5973 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
5974
5975 * doc/install.texi (Host/target specific installation notes for GCC):
5976 Put @anchor before @heading.
5977 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
5978 Use @email for email addresses.
5979
5980 2013-10-01 Jeff Law <law@redhat.com>
5981
5982 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
5983 a vec. Only delete the path if we create one without successfully
5984 registering a jump thread.
5985 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
5986 as a pointer.
5987 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
5988 (paths): New vector of jump threading paths.
5989 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
5990 (THREAD_PATH): New accessor macro for the entire thread path.
5991 (lookup_redirection_data): Get intermediate and final outgoing edge
5992 from the thread path.
5993 (create_edge_and_update_destination_phis): Copy the threading path.
5994 (ssa_fix_duplicate_block_edges): Get edges and block types from the
5995 jump threading path.
5996 (ssa_redirect_edges): Get edges and block types from the jump threading
5997 path. Free the path vector.
5998 (thread_block): Get edges from the jump threading path. Look at the
5999 entire path to see if we thread to a loop exit. If we cancel a jump
6000 thread request, then free the path vector.
6001 (thread_single_edge): Get edges and block types from the jump threading
6002 path. Free the path vector.
6003 (thread_through_loop_header): Get edges and block types from the jump
6004 threading path. Free the path vector.
6005 (mark_threaded_blocks): Iterate over the vector of paths and store
6006 the path on the appropriate edge. Get edges and block types from the
6007 jump threading path.
6008 (mark_threaded_blocks): Get edges and block types from the jump
6009 threading path. Free the path vector.
6010 (thread_through_all_blocks): Use the vector of paths rather than
6011 a vector of 3-edge sets.
6012 (register_jump_thread): Accept pointer to a path vector rather
6013 than the path vector itself. Store the path vector for later use.
6014 Simplify.
6015
6016 2013-10-01 Jakub Jelinek <jakub@redhat.com>
6017 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6018
6019 PR target/58574
6020 * config/s390/s390.c (s390_split_branches): Modify check for table
6021 jump insns.
6022 (s390_chunkify_start): Rearrange table jump insn check in order to
6023 deal with compare and branch insns correctly.
6024
6025 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
6026
6027 PR target/58578
6028 Revert
6029 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
6030 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
6031 define_insn_and_split.
6032 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
6033 (shiftsi3_compare): New pattern.
6034 (rrx): New pattern.
6035 * config/arm/unspecs.md (UNSPEC_RRX): New.
6036
6037 2013-10-01 Alan Modra <amodra@gmail.com>
6038
6039 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
6040 casing inout operands.
6041
6042 2013-10-01 Richard Biener <rguenther@suse.de>
6043
6044 PR tree-optimization/58553
6045 * tree-loop-distribution.c (struct partition_s): Add niter member.
6046 (classify_partition): Populate niter member for the partition
6047 and properly identify whether the relevant store happens before
6048 or after the loop exit.
6049 (generate_memset_builtin): Use niter member from the partition.
6050 (generate_memcpy_builtin): Likewise.
6051
6052 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6053
6054 * vec.h (vec_prefix, vec): Prefix member names with "m_".
6055 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
6056
6057 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6058
6059 * basic-block.h (edge_list): Prefix member names with "m_".
6060 * context.h (context): Likewise.
6061 * domwalk.h (dom_walker): Likewise.
6062 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
6063 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
6064 * hash-table.h (hash_table): Likewise.
6065 * machmode.h (bit_field_mode_iterator): Likewise.
6066 * pass_manager.h (pass_list): Likewise.
6067 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
6068 * tree-pass.h (pass_data): Likewise.
6069 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
6070 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
6071 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
6072 * asan.c (pass_data_asan): Update accordingly.
6073 * cfganal.c (control_dependences::find_control_dependence): Likewise.
6074 (control_dependences::control_dependences): Likewise.
6075 (control_dependences::~control_dependences): Likewise.
6076 (control_dependences::~control_dependences): Likewise.
6077 (control_dependences::get_edges_dependent_on): Likewise.
6078 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
6079 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
6080 * context.c (gcc::context::context): Likewise.
6081 * cprop.c (pass_rtl_cprop::clone): Likewise.
6082 * domwalk.c (dom_walker::walk): Likewise.
6083 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
6084 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
6085 * mode-switching.c (pass_mode_switching::clone): Likewise.
6086 * passes.c (opt_pass::opt_pass): Likewise.
6087 (pass_manager::pass_manager): Likewise.
6088 * predict.c (pass_strip_predict_hints::clone): Likewise.
6089 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
6090 (pass_split_all_insns::clone): Likewise.
6091 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
6092 Likewise.
6093 (bit_field_mode_iterator::next_mode): Likewise.
6094 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
6095 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
6096 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
6097 * tree-complex.c (pass_lower_complex::clone): Likewise.
6098 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
6099 * tree-object-size.c (pass_object_sizes::clone): Likewise.
6100 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
6101 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
6102 (pass_fold_builtins::clone): Likewise.
6103 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
6104 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
6105 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
6106 (pass_cd_dce::clone): Likewise.
6107 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
6108 (pass_phi_only_cprop::clone): Likewise.
6109 * tree-ssa-dse.c (pass_dse::clone): Likewise.
6110 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
6111 * tree-ssa-loop.c (pass_lim::clone): Likewise.
6112 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
6113 * tree-ssa-pre.c (pass_fre::clone): Likewise.
6114 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
6115 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
6116 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
6117 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
6118 * tree-vrp.c (pass_vrp::clone): Likewise.
6119 * tsan.c (pass_tsan::clone): Likewise.
6120
6121 2013-09-30 Jakub Jelinek <jakub@redhat.com>
6122
6123 PR middle-end/58564
6124 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
6125 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
6126
6127 PR middle-end/58564
6128 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
6129 optimization, punt if sign_bit_p looked through any zero extension.
6130
6131 2013-09-30 Teresa Johnson <tejohnson@google.com>
6132
6133 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
6134 Update redirected out edge count in joiner case.
6135 (ssa_redirect_edges): Common the joiner and non-joiner cases
6136 so that joiner case gets profile updates.
6137
6138 2013-09-30 Richard Biener <rguenther@suse.de>
6139
6140 PR tree-optimization/58554
6141 * tree-loop-distribution.c (classify_partition): Require
6142 unconditionally executed stores for memcpy and memset recognition.
6143 (tree_loop_distribution): Calculate dominance info.
6144
6145 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
6146
6147 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
6148 (NO_PROFILE_COUNTERS): Likewise.
6149 (PROFILE_HOOK): Likewise.
6150 (FUNCTION_PROFILER): Likewise.
6151 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
6152
6153 2013-09-30 Iain Sandoe <iain@codesourcery.com>
6154
6155 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
6156 calls and defines in TARGET_MACHO conditional.
6157 (load_macho_picbase_di): Likewise.
6158 (reload_macho_picbase): Likewise.
6159 (reload_macho_picbase_si): Likewise.
6160 (reload_macho_picbase_di): Likewise.
6161 (nonlocal_goto_receiver): Likewise.
6162
6163 2013-09-30 Nick Clifton <nickc@redhat.com>
6164
6165 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
6166 that use the MSP430X ISA.
6167 (msp430_option_override): Scan -mmcu command line option for any
6168 MCU name that supports the MSP430X ISA.
6169 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
6170 -mmcu options which enable the MSP430X ISA.
6171
6172 2013-09-30 Richard Biener <rguenther@suse.de>
6173
6174 PR middle-end/58532
6175 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
6176 before looking for setjmp-like calls.
6177
6178 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6179
6180 PR target/10901
6181 * config/darwin-protos.h (machopic_get_function_picbase): New.
6182 * config/darwin.c (machopic_get_function_picbase): New.
6183 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
6184 label for a new func. (load_macho_picbase_di): Likewise.
6185 (reload_macho_picbase): New expand.
6186 (reload_macho_picbase_si): New insn.
6187 (reload_macho_picbase_di): New insn.
6188 (nonlocal_goto_receiver): New define and split.
6189 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
6190 (unspecv enum): Add UNSPECV_NLGR.
6191
6192 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6193
6194 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
6195 that altivec registers are correctly sized on Darwin.
6196
6197 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6198
6199 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
6200 darwin-driver.o): Use COMPILE and POSTCOMPILE.
6201 * config/x-darwin (host-darwin.o): Likewise.
6202 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
6203 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
6204 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
6205
6206 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
6207
6208 * doc/invoke.texi: Fix usage of @tie{} command.
6209
6210 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
6211
6212 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
6213
6214 2013-09-28 Jan Hubicka <jh@suse.cz>
6215
6216 * config/i386/x86-tune.def: Add documentation for each of the options;
6217 add whitespace.
6218
6219 2013-09-28 Jan Hubicka <jh@suse.cz>
6220
6221 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
6222 generic.
6223 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
6224 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
6225 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
6226 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
6227 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
6228
6229 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
6230
6231 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
6232 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
6233 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
6234 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
6235 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
6236 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
6237 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
6238 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
6239 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
6240 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
6241 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
6242 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
6243 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
6244 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
6245 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
6246 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
6247 gimple.h, godump.c, graphite-clast-to-gimple.c,
6248 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
6249 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
6250 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
6251 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6252 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
6253 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
6254 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
6255 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
6256 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
6257 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
6258 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
6259 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
6260 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
6261 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
6262 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
6263 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
6264 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
6265 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
6266 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
6267 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
6268 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
6269 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
6270 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
6271 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
6272 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6273 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6274 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
6275 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
6276 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
6277 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6278 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6279 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
6280 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
6281 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6282 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6283 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
6284 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
6285 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
6286 whitespace before "(".
6287
6288 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
6289
6290 * expr.h (extract_bit_field): Remove packedp parameter.
6291 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
6292 from forward declaration.
6293 (store_split_bit_field): Remove packedp arg from calls to
6294 extract_fixed_bit_field.
6295 (extract_bit_field_1): Remove packedp parameter and packedp
6296 argument from recursive calls and calls to extract_fixed_bit_field.
6297 (extract_bit_field): Remove packedp parameter and corresponding
6298 arg to extract_bit_field_1.
6299 (extract_fixed_bit_field): Remove packedp parameter. Remove code
6300 to issue warnings.
6301 (extract_split_bit_field): Remove packedp arg from call to
6302 extract_fixed_bit_field.
6303 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
6304 (copy_blkmode_from_reg): Likewise.
6305 (copy_blkmode_to_reg): Likewise.
6306 (read_complex_part): Likewise.
6307 (store_field): Likewise.
6308 (expand_expr_real_1): Likewise.
6309 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
6310 to extract_bit_field.
6311 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
6312 call to extract_bit_field.
6313 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
6314 call to extract_bit_field.
6315 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
6316 and special packedp behavior from -fstrict-volatile-bitfields
6317 documentation.
6318
6319 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6320
6321 * lra-eliminations.c (init_elim_table): Guard value_p.
6322
6323 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6324
6325 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
6326 DFmode, DImode, and SFmode in the upper VSX registers based on the
6327 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
6328 if -mpower8-vector. Combine -mvsx-timode handling with the rest
6329 of the VSX register handling.
6330
6331 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
6332 (f32_sv): Likewise.
6333 (zero_extendsidi2_lfiwzx): Add support for loading into the
6334 Altivec registers with -mpower8-vector. Use wu/wv constraints to
6335 only do VSX memory options on Altivec registers.
6336 (extendsidi2_lfiwax): Likewise.
6337 (extendsfdf2_fpr): Likewise.
6338 (mov<mode>_hardfloat, SF/SD modes): Likewise.
6339 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
6340 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
6341 (movdi_internal64): Likewise.
6342
6343 2013-09-27 Xinliang David Li <davidxl@google.com>
6344
6345 * opts.c (finish_options): Adjust parameters
6346 according to vect cost model.
6347 (common_handle_option): Set dynamic vect cost
6348 model for FDO.
6349 targhooks.c (default_add_stmt_cost): Compute stmt cost
6350 unconditionally.
6351 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
6352 Use helper function.
6353 * tree-vectorizer.h (unlimited_cost_model): New function.
6354 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
6355 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
6356 function.
6357 (vect_enhance_data_refs_alignment): Ditto.
6358 * flag-types.h: New enum.
6359 * common/config/i386/i386-common.c (ix86_option_init_struct):
6360 No need to initialize vect_cost_model flag.
6361 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
6362 unconditionally.
6363
6364 2013-09-27 Diego Novillo <dnovillo@google.com>
6365
6366 * gimple.h (enum ssa_mode): Remove.
6367
6368 2013-09-27 Paulo Matos <pmatos@broadcom.com>
6369
6370 * cfgloop.h (number_of_loops): Fix typo in check for null.
6371
6372 2013-09-27 Jakub Jelinek <jakub@redhat.com>
6373
6374 PR middle-end/58551
6375 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
6376 are children of outermost saved_cfun's loop, and set it up to
6377 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
6378 if loop != loop0 and SESE region contains bbs that belong to loop0.
6379
6380 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6381
6382 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
6383 (binary_scale_code_p, get_base_term, get_index_term): New functions.
6384 (set_address_segment, set_address_base, set_address_index)
6385 (set_address_disp): Accept the argument unconditionally.
6386 (baseness): Remove must_be_base_p and must_be_index_p checks.
6387 (decompose_normal_address): Classify as much as possible in the
6388 main loop.
6389
6390 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6391
6392 * cse.c (count_reg_usage): Handle INT_LIST.
6393 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
6394 * reginfo.c (reg_scan_mark_refs): Likewise.
6395 * reload1.c (eliminate_regs_1): Likewise.
6396
6397 2013-09-27 Iain Sandoe <iain@codesourcery.com>
6398
6399 PR middle-end/58547
6400 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
6401 signed.
6402
6403 2013-09-27 Richard Biener <rguenther@suse.de>
6404
6405 PR tree-optimization/58459
6406 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
6407 restriction not propagating into loops.
6408
6409 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
6410
6411 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
6412 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
6413 * doc/tree-ssa.texi (Walking use-def chains): Delete.
6414
6415 2013-09-26 Richard Biener <rguenther@suse.de>
6416
6417 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
6418
6419 2013-09-26 Richard Biener <rguenther@suse.de>
6420
6421 * alias.h (component_uses_parent_alias_set): Rename to ...
6422 (component_uses_parent_alias_set_from): ... this.
6423 * alias.c (component_uses_parent_alias_set): Rename to ...
6424 (component_uses_parent_alias_set_from): ... this and return
6425 the desired parent.
6426 (reference_alias_ptr_type_1): Use the result from
6427 component_uses_parent_alias_set_from instead of stripping
6428 components one at a time.
6429 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
6430
6431 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6432
6433 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
6434 Move prototypes to...
6435 * tree-ssa-ter.h: New File. Move prototypes here.
6436 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
6437 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
6438 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
6439 from is_replaceable_p.
6440 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
6441 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
6442 (process_replaceable): Use ter_is_replaceable_p.
6443 (find_replaceable_in_bb): Use ter_is_replaceable_p.
6444 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
6445 newly refactored ssa_is_replaceable_p.
6446 * cfgexpand.c: Include tree-outof-ssa.h.
6447 * ssaexpand.h: Delete.
6448
6449 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6450
6451 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
6452 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
6453 tree-ssa.c
6454 (create_gimple_tmp): Delete.
6455 (get_expr_type, build_assign, build_type_cast): Move to...
6456 * gimple-builder.c: New File.
6457 (get_expr_type): Relocate from gimple.c.
6458 (build_assign, build_type_cast): Change to only create ssanames.
6459 * gimple.h: Move prototypes to...
6460 * gimple-builder.h: New File. Here.
6461 * tree-ssa.h: And here.
6462 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
6463 count_uses_and_derefs): Relocate from gimple.c.
6464 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
6465 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
6466 * tree-ssa-math-opts (execute_cse_reciprocals): Use
6467 gimple_replace_ssa_lhs.
6468 * asan.c: Include gimple-builder.h.
6469 * Makefile.in: Add gimple-builder.o.
6470
6471 2013-09-26 Richard Biener <rguenther@suse.de>
6472
6473 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
6474 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
6475 (loe_visit_block): Use gcc_checking_assert.
6476 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
6477 gimple_assign_ssa_name_copy_p.
6478 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
6479 change.
6480
6481 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
6482
6483 * config/rs6000/t-rs6000 (rs6000.o): Remove.
6484 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
6485
6486 2013-09-26 Richard Biener <rguenther@suse.de>
6487
6488 PR tree-optimization/58539
6489 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
6490 the fact that debug statements are not taking part in loop-closed
6491 SSA construction.
6492
6493 2013-09-26 Nick Clifton <nickc@redhat.com>
6494
6495 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
6496 time warning message.
6497 (msp430_print_operand_raw): Delete unused letter parameter.
6498 (TARGET_PRINT_OPERAND_ADDRESS): Define.
6499 (msp430_print_operand_address): New function.
6500 (msp430_print_operand): Move address printing code from here to
6501 new function.
6502 * config/msp430/msp430.md (movsipsi2): Add comment in generated
6503 assembler.
6504 (zero_extendpsisi2): Likewise.
6505 (extendpsisi2): New pattern.
6506 (andneghi3): New pattern.
6507
6508 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
6509
6510 * config/aarch64/aarch64.opt (mlra): New option.
6511 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
6512 (TARGET_LRA_P): Define.
6513
6514 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
6515
6516 * expr.c (expand_assignment): Remove obsolete comment.
6517
6518 2013-09-25 Jeff Law <law@redhat.com>
6519
6520 * tree-flow.h (thread_through_all_blocks): Prototype moved into
6521 tree-ssa-threadupdate.h.
6522 (register_jump_thread): Similarly.
6523 * tree-ssa-threadupdate.h: New header file.
6524 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
6525 * tree-vrp.c: Likewise.
6526 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
6527 (thread_around_empty_blocks): Change type of path vector argument to
6528 an edge,type pair from just an edge. Initialize both elements when
6529 appending to a jump threading path. Tweak references to elements
6530 appropriately.
6531 (thread_across_edge): Similarly. Release memory for the elements
6532 as needed.
6533 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
6534 (dump_jump_thread_path): New function broken out from
6535 register_jump_thread.
6536 (register_jump_thread): Use dump_jump_thread_path. Change type of
6537 path vector entries. Search the path for NULL edges and dump
6538 the path if one is found. Tweak the conversion of path to 3-edge
6539 form to use the block copy type information embedded in the path.
6540
6541 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6542
6543 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
6544
6545 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6546 Vladimir Makarov <vmakarov@redhat.com>
6547
6548 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
6549 from the least significant bit.
6550 (strip_address_mutations): Add bitfield operations handling.
6551 (must_be_index_p): Add shifting and rotate operations handling.
6552 (set_address_base): Use must_be_base_p predicate.
6553 (set_address_index): Use must_be_index_p predicate.
6554
6555 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
6556 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6557 Sergey Lega <sergey.s.lega@intel.com>
6558 Anna Tikhonova <anna.tikhonova@intel.com>
6559 Ilya Tocar <ilya.tocar@intel.com>
6560 Andrey Turetskiy <andrey.turetskiy@intel.com>
6561 Ilya Verbin <ilya.verbin@intel.com>
6562 Kirill Yukhin <kirill.yukhin@intel.com>
6563 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6564
6565 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
6566 Use new names.
6567 (ix86_expand_vector_move_misalign): Support new unaligned load and
6568 stores and use new names.
6569 (CODE_FOR_sse2_storedqu): Rename to ...
6570 (CODE_FOR_sse2_storedquv16qi): ... this.
6571 (CODE_FOR_sse2_loaddqu): Rename to ...
6572 (CODE_FOR_sse2_loaddquv16qi): ... this.
6573 (CODE_FOR_avx_loaddqu256): Rename to ...
6574 (CODE_FOR_avx_loaddquv32qi): ... this.
6575 (CODE_FOR_avx_storedqu256): Rename to ...
6576 (CODE_FOR_avx_storedquv32qi): ... this.
6577 * config/i386/i386.md (fpint_logic): New.
6578 * config/i386/sse.md (VMOVE): Extend for AVX512.
6579 (VF): Ditto.
6580 (VF_128_256): New.
6581 (VF_512): Ditto.
6582 (VI_UNALIGNED_LOADSTORE): Ditto.
6583 (sse2_avx_avx512f): Ditto.
6584 (sse2_avx2): Extend for AVX512.
6585 (sse4_1_avx2): Ditto.
6586 (avx2_avx512f): New.
6587 (sse): Extend for AVX512.
6588 (sse2): Ditto.
6589 (sse4_1): Ditto.
6590 (avxsizesuffix): Ditto.
6591 (sseintvecmode): Ditto.
6592 (ssePSmode): Ditto.
6593 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
6594 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
6595 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6596 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
6597 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6598 (<sse2_avx_avx512f>_storedqu<mode): ... this.
6599 (<sse>_movnt<mode>): Replace constraint "x" with "v".
6600 (STORENT_MODE): Extend for AVX512.
6601 (*absneg<mode>2): Replace constraint "x" with "v".
6602 (*mul<mode>3): Ditto.
6603 (*ieee_smin<mode>3): Ditto.
6604 (*ieee_smax<mode>3): Ditto.
6605 (avx_cmp<mode>3): Replace VF with VF_128_256.
6606 (*<sse>_maskcmp<mode>3_comm): Ditto.
6607 (<sse>_maskcmp<mode>3): Ditto.
6608 (<sse>_andnot<mode>3): Extend for AVX512.
6609 (<code><mode>3, anylogic): Replace VF with VF_128_256.
6610 (<code><mode>3, fpint_logic): New.
6611 (*<code><mode>3): Extend for AVX512.
6612 (avx512flogicsuff): New.
6613 (avx512f_<logic><mode>): Ditto.
6614 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
6615 VF_128_256.
6616 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
6617 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
6618 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
6619 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
6620 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
6621 (xop_vpermil2<mode>3): Ditto.
6622 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
6623 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
6624 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
6625 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
6626
6627 2013-09-25 Tom Tromey <tromey@redhat.com>
6628
6629 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
6630 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
6631 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
6632 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
6633 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
6634 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
6635 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
6636 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
6637 (GRAPHITE_HTAB_H): Remove.
6638
6639 2013-09-25 Tom Tromey <tromey@redhat.com>
6640
6641 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
6642
6643 2013-09-25 Tom Tromey <tromey@redhat.com>
6644
6645 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
6646
6647 2013-09-25 Tom Tromey <tromey@redhat.com>
6648
6649 * config/i386/t-i386 (i386.o): Remove.
6650 (i386-c.o): Use COMPILE and POSTCOMPILE.
6651
6652 2013-09-25 Tom Tromey <tromey@redhat.com>
6653
6654 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
6655
6656 2013-09-25 Tom Tromey <tromey@redhat.com>
6657
6658 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
6659 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
6660 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
6661 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
6662 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
6663 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
6664 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
6665 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
6666 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
6667 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
6668 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
6669 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
6670 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
6671 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
6672 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
6673 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
6674 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
6675 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
6676 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
6677 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
6678 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
6679 (tree-ssa-pre.o, tree-ssa-sccvn.o)
6680 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
6681 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
6682 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
6683 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
6684 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
6685 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
6686 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
6687 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
6688 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
6689 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
6690 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
6691 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
6692 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
6693 (graphite-clast-to-gimple.o, graphite-dependences.o)
6694 (graphite-interchange.o, graphite-poly.o)
6695 (graphite-scop-detection.o, graphite-sese-to-poly.o)
6696 (graphite-optimize-isl.o, tree-vect-loop.o)
6697 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
6698 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
6699 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
6700 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
6701 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
6702 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
6703 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
6704 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
6705 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
6706 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
6707 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
6708 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
6709 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
6710 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
6711 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
6712 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
6713 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
6714 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
6715 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
6716 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
6717 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
6718 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
6719 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
6720 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
6721 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
6722 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
6723 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
6724 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
6725 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
6726 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
6727 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
6728 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
6729 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
6730 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
6731 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
6732 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
6733 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
6734 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
6735 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
6736 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
6737 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
6738 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
6739 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
6740 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
6741 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
6742 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
6743 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
6744 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
6745 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
6746 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
6747 (gcov-dump.o): Remove.
6748 (default-c.o): Use COMPILE and POSTCOMPILE.
6749 (CFLAGS-gcc.o): New variable.
6750 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
6751
6752 2013-09-25 Tom Tromey <tromey@redhat.com>
6753
6754 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
6755 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
6756 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
6757 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
6758 (c-family/c-pragma.o, c-family/c-pretty-print.o)
6759 (c-family/c-semantics.o, c-family/c-ada-spec.o)
6760 (c-family/array-notation-common.o, c-family/stub-objc.o)
6761 (c-family/c-ubsan.o): Remove.
6762
6763 2013-09-25 Tom Tromey <tromey@redhat.com>
6764
6765 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
6766
6767 2013-09-25 Tom Tromey <tromey@redhat.com>
6768
6769 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
6770 to add -DENABLE_SHARED_LIBGCC.
6771 (gcc.o): Don't use subshell.
6772
6773 2013-09-25 Tom Tromey <tromey@redhat.com>
6774
6775 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
6776 * configure.ac: Don't invoke AM_PROG_CC_C_O.
6777 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
6778 * configure, config.in: Rebuild.
6779
6780 2013-09-25 Tom Tromey <tromey@redhat.com>
6781
6782 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
6783 (COMPILE, POSTCOMPILE): New variables.
6784 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
6785 (DEPFILES): New variable.
6786 Include ".Po" files.
6787 * configure.ac: Add checks for dependency checking.
6788 * configure, aclocal.m4: Regenerate.
6789
6790 2013-09-25 Tom Tromey <tromey@redhat.com>
6791
6792 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
6793 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
6794
6795 2013-09-25 Tom Tromey <tromey@redhat.com>
6796
6797 * Makefile.in (generated_files): Add options.h,
6798 target-hooks-def.h, insn-opinit.h,
6799 common/common-target-hooks-def.h, pass-instances.def,
6800 c-family/c-target-hooks-def.h.
6801
6802 2013-09-25 Jeff Law <law@redhat.com>
6803
6804 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
6805 than foo[foo.length () - 1] to access last member in a vec.
6806 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
6807
6808 2013-09-25 Richard Biener <rguenther@suse.de>
6809
6810 PR middle-end/58521
6811 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
6812
6813 2013-09-25 Jan Hubicka <jh@suse.cz>
6814
6815 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
6816 test.
6817
6818 2013-09-25 Marek Polacek <polacek@redhat.com>
6819
6820 PR sanitizer/58420
6821 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
6822 when determining the type name.
6823
6824 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
6825
6826 * config/sh/sh.md: Fix formatting.
6827
6828 2013-09-24 Xinliang David Li <davidxl@google.com>
6829
6830 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
6831 max peel iterations parameter.
6832 * param.def: New parameter.
6833 * doc/invoke.texi: Document New parameter.
6834
6835 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
6836
6837 * gimple-pretty-print.c: Various whitespace tweaks.
6838 * tree-core.h: Likewise.
6839 * tree-pretty-print.c: Likewise.
6840 * tree-ssa-alias.c: Likewise.
6841 * tree-ssa-copy.c: Likewise.
6842 * tree-ssanames.c: Likewise.
6843 * tree-ssanames.h: Likewise.
6844 * tree-vrp.c: Likewise.
6845
6846 2013-09-24 Alan Modra <amodra@gmail.com>
6847
6848 PR middle-end/57134
6849 PR middle-end/57586
6850 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
6851 for output operands that disallow regs. Don't use EXPAND_WRITE on
6852 inout operands.
6853
6854 2013-09-24 Richard Biener <rguenther@suse.de>
6855
6856 PR middle-end/58513
6857 * tree.c (reference_alias_ptr_type): Move ...
6858 * alias.c (reference_alias_ptr_type): ... here and implement
6859 in terms of the new reference_alias_ptr_type_1.
6860 (ref_all_alias_ptr_type_p): New helper.
6861 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
6862 use ref_all_alias_ptr_type_p.
6863 (get_deref_alias_set): Add flag_strict_aliasing check here.
6864 (reference_alias_ptr_type_1): New function, split out from ...
6865 (get_alias_set): ... here.
6866 (alias_ptr_types_compatible_p): New function.
6867 * alias.h (reference_alias_ptr_type): Declare.
6868 (alias_ptr_types_compatible_p): Likewise.
6869 * tree.h (reference_alias_ptr_type): Remove.
6870 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
6871 to compare MEM_REF alias types.
6872
6873 2013-09-24 Richard Biener <rguenther@suse.de>
6874
6875 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
6876
6877 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6878
6879 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
6880 reload helper function arrays into a single array reg_addr.
6881 (reload_fpr_gpr): Likewise.
6882 (reload_gpr_vsx): Likewise.
6883 (reload_vsx_gpr): Likewise.
6884 (struct rs6000_reg_addr): Likewise.
6885 (reg_addr): Likewise.
6886 (rs6000_debug_reg_global): Change rs6000_vector_reload,
6887 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
6888 (rs6000_init_hard_regno_mode_ok): Likewise.
6889 (rs6000_secondary_reload_direct_move): Likewise.
6890 (rs6000_secondary_reload): Likewise.
6891
6892 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
6893 constraints: wu, ww, and wy. Repurpose wv constraint added during
6894 power8 changes. Put wg constraint in alphabetical order.
6895
6896 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
6897 for future work to add ISA 2.07 VSX single precision support.
6898 (-mvsx-scalar-double): Change default from -1 to 1, update
6899 documentation comment.
6900 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
6901 (-mupper-regs-df): New debug switch to control whether DF values
6902 can go in the traditional Altivec registers.
6903 (-mupper-regs-sf): New debug switch to control whether SF values
6904 can go in the traditional Altivec registers.
6905
6906 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
6907 and wy constraints.
6908 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
6909 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
6910 Add new constraints, wu/ww/wy. Repurpose wv constraint.
6911 (rs6000_debug_legitimate_address_p): Print if we are running
6912 before, during, or after reload.
6913 (rs6000_secondary_reload): Add a comment.
6914 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
6915
6916 * config/rs6000/constraints.md (wa constraint): Sort w<x>
6917 constraints. Update documentation string.
6918 (wd constraint): Likewise.
6919 (wf constraint): Likewise.
6920 (wg constraint): Likewise.
6921 (wn constraint): Likewise.
6922 (ws constraint): Likewise.
6923 (wt constraint): Likewise.
6924 (wx constraint): Likewise.
6925 (wz constraint): Likewise.
6926 (wu constraint): New constraint for ISA 2.07 SFmode scalar
6927 instructions.
6928 (ww constraint): Likewise.
6929 (wy constraint): Likewise.
6930 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
6931 in the previous submissions.
6932 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
6933
6934 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
6935
6936 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
6937 (REG_BR_PROB): Say that the probability is stored in an int_list.
6938 * reg-notes.def: Update commentary to mention INT_LIST.
6939 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
6940 (INT_LIST): New rtx.
6941 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
6942 * rtlanal.c (int_reg_note_p): New function.
6943 (alloc_reg_note): Assert that the note does not have an int argument.
6944 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
6945 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
6946 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
6947 rather than an INSN_LIST. Handle INT_LIST.
6948 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
6949 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
6950 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
6951 Manipulate them as ints rather than rtxes.
6952 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
6953 * regmove.c (copy_src_to_dest): Likewise.
6954 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
6955
6956 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
6957 into the cases that need it.
6958 * config/arm/arm.c (arm_unwind_emit): Likewise.
6959
6960 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
6961 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
6962 * loop-doloop.c (add_test, doloop_modify): Likewise.
6963 * loop-unswitch.c (compare_and_jump_seq): Likewise.
6964 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
6965 * predict.c (combine_predictions_for_insn): Likewise.
6966 * print-rtl.c (print_rtx): Handle INT_LIST.
6967 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
6968 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
6969 * config/arm/arm.c (emit_unlikely_jump): Likewise.
6970 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
6971 (ix86_split_fp_branch, predict_jump): Likewise.
6972 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
6973 * config/sh/sh.c (expand_cbranchsi4): Likewise.
6974 * config/spu/spu.c (ea_load_store_inline): Likewise.
6975
6976 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
6977 value of a REG_BR_PROB note.
6978 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
6979 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
6980 * emit-rtl.c (try_split): Likewise.
6981 * predict.c (br_prob_note_reliable_p): Likewise.
6982 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
6983 * reorg.c (mostly_true_jump): Likewise.
6984 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
6985 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
6986 * config/i386/i386.c (ix86_print_operand): Likewise.
6987 * config/ia64/ia64.c (ia64_print_operand): Likewise.
6988 * config/mmix/mmix.c (mmix_print_operand): Likewise.
6989 * config/rs6000/rs6000.c (output_cbranch): Likewise.
6990 * config/s390/s390.c (s390_expand_tbegin): Likewise.
6991 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
6992 * config/sparc/sparc.c (output_cbranch): Likewise.
6993 * config/spu/spu.c (get_branch_target): Likewise.
6994 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
6995 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
6996
6997 2013-09-23 Jan Hubicka <jh@suse.cz>
6998
6999 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
7000 for ipa-devirt.
7001 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
7002 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
7003 of external calls
7004 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
7005 (gimple_fold_call): Dump inconsistent devirtualizations; add
7006 sanity check for type based devirtualizations.
7007 * ipa-prop.c: Include ipa-utils.h
7008 (ipa_intraprocedural_devirtualization): Add sanity check.
7009 (try_make_edge_direct_virtual_call): Likewise.
7010
7011 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
7012
7013 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
7014 assignment statements.
7015
7016 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
7017
7018 * gimple-pretty-print.c (dump_ssaname_info): New function.
7019 (dump_gimple_phi): Call it.
7020 (pp_gimple_stmt_1): Likewise.
7021 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
7022 (range_info_def): Declare.
7023 * tree-pretty-print.c (pp_double_int): New function.
7024 (dump_generic_node): Call it.
7025 * tree-pretty-print.h (pp_double_int): Declare.
7026 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
7027 * tree-ssanames.h (range_info_def): New structure.
7028 (value_range_type): Move definition here.
7029 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
7030 Declare.
7031 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
7032 initialization.
7033 (set_range_info): New function.
7034 (get_range_info): Likewise.
7035 (duplicate_ssa_name_range_info): Likewise.
7036 (duplicate_ssa_name_fn): Check pointer type and call
7037 duplicate_ssa_name_range_info.
7038 * tree-ssa-copy.c (fini_copy_prop): Likewise.
7039 * tree-vrp.c (value_range_type): Remove definition, now in
7040 tree-ssanames.h.
7041 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
7042 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
7043 (SSA_NAME_RANGE_INFO): New macro.
7044
7045 2013-09-23 Richard Biener <rguenther@suse.de>
7046
7047 PR tree-optimization/58464
7048 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7049 (phi_trans_add): Change to add conditionally on being not present.
7050 (phi_translate_1): Remove recursion detection here.
7051 (phi_translate): Pre-seed the cache with NULL to catch
7052 recursion here in a more generic way.
7053 (bitmap_find_leader): Adjust comment.
7054 (get_representative_for): Dump value-numbers.
7055 (create_expression_by_pieces): Likewise.
7056 (insert_into_preds_of_block): Likewise.
7057
7058 2013-09-23 Christian Bruel <christian.bruel@st.com>
7059
7060 PR target/58475
7061 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
7062 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
7063
7064 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
7065
7066 Revert r202780:
7067 2013-09-20 Renlin Li <renlin.li@arm.com>
7068
7069 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
7070 plus_constant.
7071 (aarch64_expand_epilogue): Likewise.
7072 (aarch64_legitimize_reload_address): Likewise.
7073
7074 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
7075
7076 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
7077 NULL_TREE before pushing them onto the vector. Likewise for labels.
7078
7079 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
7080
7081 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
7082 comparison operators when -fno-trapping-math is in effect.
7083 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
7084 comparison operators in TFmode and assert that unsupported operators
7085 cannot reach here.
7086 (ia64_print_operand): Likewise.
7087
7088 2013-09-21 Jan Hubicka <jh@suse.cz>
7089
7090 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
7091 (sse_typeless_stores): Enable for core
7092 (sse_load0_by_pxor): Likewise.
7093 (four_jump_limit): Disable for core.
7094 (pad_returns): Likewise.
7095 (avoid_vector_decode): Likewise.
7096 (fuse_cmp_and_branch): Enable for cores.
7097 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
7098
7099 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7100
7101 PR middle-end/56791
7102 * config/pa/pa.c (pa_option_override): Disable auto increment and
7103 decrement instructions until reload is completed.
7104
7105 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
7106 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
7107 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
7108
7109 2013-09-20 DJ Delorie <dj@redhat.com>
7110 Nick Clifton <nickc@redhat.com>
7111
7112 * config/rl78/rl78.c: Various whitespace and comment tweaks.
7113 (need_to_save): Save bank 0 on interrupts.
7114 (characterize_address): Strip far address wrappers.
7115 (rl78_as_legitimate_address): Likewise.
7116 (transcode_memory_rtx): Likewise.
7117 (rl78_peep_movhi_p): Disable this peephole after devirt.
7118 (rl78_propogate_register_origins): Forget all origins when a
7119 CLOBBER is seen.
7120 * config/rl78/rl78-virt.md: Various whitespace tweaks.
7121 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
7122 peephole2's.
7123 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
7124 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
7125 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
7126 relocating.
7127 * config/rl78/constraints.md: Various whitespace and paren tweaks.
7128
7129 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7130
7131 * config/pa/pa.md: In "scc" insn patterns, change output template to
7132 handle const0_rtx in reg_or_0_operand operands.
7133
7134 2013-09-20 Martin Husemann <martin@NetBSD.org>
7135
7136 PR target/56875
7137 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
7138 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
7139
7140 2013-09-20 Richard Biener <rguenther@suse.de>
7141
7142 PR middle-end/58484
7143 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
7144 remembering SSA name version and block index.
7145 (new_scev_info_str): Adjust.
7146 (hash_scev_info): Likewise. Also hash the block index.
7147 (eq_scev_info): Adjust.
7148 (find_var_scev_info): Likewise.
7149 (struct instantiate_cache_entry): Remove.
7150 (struct instantiate_cache_type): Use a htab to map name, block
7151 to chrec.
7152 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7153 (get_instantiated_value_entry): Likewise.
7154 (hash_idx_scev_info, eq_idx_scev_info): New functions.
7155 (instantiate_scev_name): Adjust.
7156
7157 2013-09-20 Jeff Law <law@redhat.com>
7158
7159 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
7160
7161 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
7162
7163 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
7164 Call aarch64_simd_expand_args to update op[argc].
7165
7166 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7167
7168 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
7169 plugin argument.
7170
7171 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7172
7173 * gengtype.c (file_rules): Added rule for *.cc files.
7174 (get_output_file_with_visibility): Give fatal message when no
7175 rules found.
7176
7177 2013-09-20 Renlin Li <renlin.li@arm.com>
7178
7179 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
7180 (aarch64_expand_epilogue): Likewise.
7181 (aarch64_legitimize_reload_address): Likewise.
7182
7183 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
7184
7185 PR middle-end/57748
7186 * expr.c (expand_assignment): Remove misalignp code path.
7187
7188 2013-09-20 Marek Polacek <polacek@redhat.com>
7189
7190 PR sanitizer/58413
7191 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
7192 TYPE_PRECISION. Add asserts.
7193
7194 2013-09-20 Richard Biener <rguenther@suse.de>
7195
7196 PR tree-optimization/58453
7197 * tree-loop-distribution.c (distribute_loop): Apply the cost
7198 model for -ftree-loop-distribute-patterns, too.
7199
7200 2013-09-20 Richard Biener <rguenther@suse.de>
7201
7202 PR middle-end/58473
7203 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
7204 make type comparison less strict.
7205
7206 2013-09-20 Alan Modra <amodra@gmail.com>
7207
7208 * configure: Regenerate.
7209 * aclocal.m4: Regenerate.
7210
7211 2013-09-20 Marek Polacek <polacek@redhat.com>
7212
7213 PR other/58467
7214 * doc/extend.texi: Document that attribute used is meant to be used
7215 on variables with static storage duration.
7216
7217 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7218
7219 PR tree-optimization/58472
7220 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
7221 simd_lane_access set inv_p = false.
7222 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
7223 the simduid magic VAR_DECL.
7224
7225 2013-09-19 Jan Hubicka <jh@suse.cz>
7226
7227 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
7228
7229 2013-09-17 Jeff Law <law@redhat.com>
7230
7231 * tree-ssa-dom.c (record_temporary_equivalences): New function
7232 split out of dom_opt_dom_walker::after_dom_children.
7233 (dom_opt_dom_walker::thread_across_edge): Move common code
7234 in here from dom_opt_dom_walker::after_dom_children.
7235 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
7236
7237 2013-09-19 Jan Hubicka <jh@suse.cz>
7238
7239 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
7240 (TARGET_GENERIC): Use PROCESOR_GENERIC
7241 (enum processor_type): Unify generic32 and 64.
7242 * i386.md (cpu): Likewise.
7243 * x86-tune.def (use_leave): Enable for generic32.
7244 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
7245 * athlon.md: Change generic64 to generic in all occurences.
7246 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
7247 (ix86_target_macros_internal): Likewise.
7248 * driver-i386.c (host_detect_local_cpu): Likewise.
7249 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
7250 to ..
7251 (generic_memcpy, generic_memset, generic_cost): This one.
7252 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
7253 (m_GENERIC32, m_GENERIC64): Remove.
7254 (m_GENERIC): Turn into one flag.
7255 (processor_target): Unify generic tunnings.
7256 (ix86_option_override_internal): Replace generic32/64 by generic.
7257 (ix86_issue_rate): Likewise.
7258 (ix86_adjust_cost): Likewise.
7259
7260 2013-09-19 Jan Hubicka <jh@suse.cz>
7261
7262 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
7263 of speculative flag.
7264
7265 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7266
7267 * omp-low.c (expand_omp_sections): Always pass len - 1 to
7268 GOMP_sections_start, even if !exit_reachable.
7269
7270 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7271
7272 * lra-constraints.c (need_for_all_save_p): Use macro
7273 HARD_REGNO_CALL_PART_CLOBBERED.
7274 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
7275 set up pseudo conflict hard regs.
7276
7277 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7278
7279 PR target/58452
7280 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
7281 operands.
7282
7283 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7284
7285 PR rtl-optimization/58438
7286 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
7287 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
7288 from previous subpasses.
7289
7290 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
7291
7292 * arm.c (arm_get_frame_offsets): Validate architecture supports
7293 LDRD/STRD before accepting the tuning preference.
7294 (arm_expand_prologue): Likewise.
7295 (arm_expand_epilogue): Likewise.
7296
7297 2013-09-18 Richard Biener <rguenther@suse.de>
7298
7299 PR tree-optimization/58417
7300 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
7301 have chrecs with symbols defined in the loop as operands.
7302 (chrec_fold_multiply): Likewise.
7303 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
7304 parameters before folding binary operations.
7305 (struct instantiate_cache_entry_hasher): Remove.
7306 (struct instantiate_cache_type): Use a pointer-map.
7307 (instantiate_cache_type::instantiate_cache_type): New function.
7308 (instantiate_cache_type::get): Likewise.
7309 (instantiate_cache_type::set): Likewise.
7310 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7311 (get_instantiated_value_entry): Likewise.
7312 (global_cache): New global.
7313 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
7314 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
7315 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
7316 (instantiate_scev_name): Adjust.
7317 (instantiate_scev): Construct global instead of local cache.
7318 (resolve_mixers): Likewise.
7319
7320 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
7321 Paolo Carlini <paolo.carlini@oracle.com>
7322
7323 PR c++/58458
7324 * doc/implement-cxx.texi: Fix references to the C++ standards.
7325
7326 2013-09-18 Jakub Jelinek <jakub@redhat.com>
7327
7328 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
7329 * tree-vect-data-refs.c (vect_analyze_data_refs): For
7330 simd_lane_access drs, update also DR_ALIGNED_TO.
7331
7332 2013-09-18 Marek Polacek <polacek@redhat.com>
7333
7334 PR sanitizer/58411
7335 * doc/extend.texi: Document no_sanitize_undefined attribute.
7336 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
7337 no_sanitize_undefined attribute.
7338
7339 2013-09-18 Nick Clifton <nickc@redhat.com>
7340
7341 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
7342 (ASM_DECLARE_FUNCTION_NAME): Define.
7343
7344 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
7345
7346 * compare-elim.c (find_comparison_dom_walker): New class
7347 (find_comparisons_in_bb): Rename to
7348 find_comparison_dom_walker::before_dom_children
7349 (find_comparisons): Adjust
7350 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
7351 adjust.
7352 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
7353 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
7354 (init_walk_dominator_tree): Remove declaration.
7355 (fini_walk_dominator_tree): Remove declaration.
7356 * fwprop.c (single_def_use_dom_walker): New class
7357 (single_def_use_enter_block): Convert to
7358 single_def_use_dom_walker::before_dom_children.
7359 (single_def_use_leave_block): Convert to
7360 single_def_use_dom_walker::after_dom_children.
7361 (build_single_def_use_links): Adjust.
7362 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
7363 class.
7364 (find_candidates_in_block): Convert to
7365 find_candidates_dom_walker::before_dom_children.
7366 (execute_strength_reduction): Adjust.
7367 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
7368 (sese_dom_walker): New class.
7369 (sese_dom_walker::sese_dom_walker): New constructor.
7370 (sese_dom_walker::~sese_dom_walker): New destructor.
7371 (build_sese_conditions_before): Convert to
7372 sese_dom_walker::before_dom_children.
7373 (build_sese_conditions_after): Convert to
7374 sese_dom_walker::after_dom_children.
7375 (build_poly_scop): Adjust
7376 * tree-into-ssa.c (rewrite_dom_walker): New class
7377 (rewrite_enter_block): Convert to
7378 rewrite_dom_walker::before_dom_children.
7379 (rewrite_leave_block): Convert to
7380 rewrite_dom_walker::after_dom_children.
7381 (rewrite_update_dom_walker): New class.
7382 (rewrite_update_enter_block): Convert to
7383 rewrite_update_dom_walker::before_dom_children.
7384 (rewrite_update_leave_block): Convert to
7385 rewrite_update_dom_walker::after_dom_children.
7386 (rewrite_blocks, rewrite_into_ssa): Adjust.
7387 (mark_def_dom_walker): New class.
7388 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
7389 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
7390 (mark_def_sites_blocks): Convert to
7391 mark_def_dom_walker::before_dom_children.
7392 (mark_def_site_blocks): Remove.
7393 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
7394 (tree_ssa_dominator_optimize): Adjust.
7395 (dom_thread_across_edge): Convert to method
7396 dom_opt_dom_walker::thread_across_edge.
7397 (dom_opt_enter_block): Convert to member function
7398 dom_opt_dom_walker::before_dom_children.
7399 (dom_opt_leave_block): Convert to member function
7400 dom_opt_dom_walker::after_dom_children.
7401 * tree-ssa-dse.c (dse_dom_walker): New class.
7402 (dse_enter_block): Convert to member function
7403 dse_dom_walker::before_dom_children.
7404 (tree_ssa_dse): Adjust.
7405 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
7406 (determine_invariantness_stmt): Convert to method
7407 invariantness_dom_walker::before_dom_children.
7408 (determine_invariantness): Remove
7409 (move_computations_dom_walker): New class.
7410 (move_computations_stmt): Convert to method
7411 move_computations_dom_walker::before_dom_children.
7412 (move_computations, tree_ssa_lim): Adjust.
7413 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
7414 (nt_init_block): Convert to method
7415 notrappping_dom_walker::before_dom_children.
7416 (nt_fini_block): Convert to method
7417 method nontrapping_dom_walker::after_dom_children.
7418 (get_non_trapping): Adjust.
7419 * tree-ssa-pre.c (eliminate_dom_walker): New class.
7420 (eliminate_bb): Convert to method
7421 eliminate_dom_walker::before_dom_children.
7422 (eliminate_leave_block): Convert to method
7423 eliminate_dom_walker::after_dom_children.
7424 (eliminate): Adjust.
7425 * tree-ssa-strlen.c (strlen_dom_walker): New class.
7426 (strlen_enter_block): Convert to method
7427 strlen_dom_walker::before_dom_children.
7428 (strlen_leave_block): Convert to method
7429 method strlen_dom_walker::after_dom_children.
7430 (tree_ssa_strlen): Adjust.
7431 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
7432 (tree_ssa_uncprop): Adjust.
7433 (uncprop_leave_block): Convert to method
7434 uncprop_dom_walker::after_dom_children.
7435 (uncprop_leave_block): Convert to method
7436 uncprop_dom_walker::before_dom_children.
7437
7438 2013-09-18 Bin Cheng <bin.cheng@arm.com>
7439
7440 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
7441 branch in same basic block. Check both src and dest of the move insn.
7442
7443 2013-09-17 Nick Clifton <nickc@redhat.com>
7444
7445 * config/rl78/rl78-real.md (bf): New pattern.
7446 (bt): New pattern.
7447 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
7448 (rl78_print_operand): Do not put a # before a %B.
7449 * config/rl78/rl78.opt: Tweak doc strings.
7450
7451 2013-09-17 DJ Delorie <dj@redhat.com>
7452
7453 * config/rl78/constraints.md (Wcv): Allow up to $r31.
7454 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
7455 (rl78_option_override): Likewise, if -mallregs.
7456 (is_virtual_register): Likewise.
7457 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
7458 (REGNO_OK_FOR_BASE_P): Likewise.
7459 * config/rl78/rl78.opt (-mallregs): New.
7460
7461 2013-09-17 Nick Clifton <nickc@redhat.com>
7462
7463 * config/rl78/rl78.c (need_to_save): Change return type to bool.
7464 For interrupt functions: save all call clobbered registers if the
7465 interrupt handler is not a leaf function.
7466 (rl78_expand_prologue): Always recompute the frame information.
7467 For interrupt functions: only select bank 0 if one of the bank 0
7468 registers is going to be psuhed.
7469
7470 2013-09-17 DJ Delorie <dj@redhat.com>
7471
7472 * config/rl78/constraints.md: For each W* constraint, rename to C*
7473 and create a W* constraint that checks for an optional ES: prefix
7474 pattern also.
7475 * config/rl78/rl78.md (UNS_ES_ADDR): New.
7476 (es_addr): New. Used to wrap far addresses.
7477 * config/rl78/rl78-protos.h (rl78_es_addr): New.
7478 (rl78_es_base): New.
7479 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
7480 wrapped far addresses.
7481 (rl78_print_operand_1): Unwrap far addresses before processing.
7482 (rl78_lo16): Wrap far addresses in unspecs.
7483 (rl78_es_addr): New.
7484 (rl78_es_base): New.
7485 (insn_ok_now): Check for not-yet-wrapped far addresses.
7486 (transcode_memory_rtx): Properly re-wrap far addresses.
7487
7488 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
7489
7490 * config/sparc/t-rtems: Add leon3 multilibs.
7491
7492 2013-09-17 Cong Hou <congh@google.com>
7493
7494 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
7495 when checking the dot production pattern. The type of rhs operand
7496 of multiply is now checked correctly.
7497
7498 2013-09-17 Jeff Law <law@redhat.com>
7499
7500 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
7501 edge implied equivalences into successor phis.
7502 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
7503 here from tree-ssa-threadedge.c.
7504 (mark_threaded_blocks): When threading through a joiner, if both
7505 successors of the joiner's clone reach the same block, verify the
7506 PHI arguments are equal. If not, cancel the jump threading request.
7507 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
7508 tree-ssa-threadupdate.c
7509 (thread_across_edge): Don't check PHI argument equality when
7510 threading through joiner block here.
7511
7512 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
7513
7514 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
7515 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
7516 (warn_uninit, warn_uninitialized_vars,
7517 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7518 Move to tree-ssa-uninit.c.
7519 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
7520 (has_undefined_value_p): New. Pass dependant parts of
7521 ssa_undefined_value_p.
7522 (uninit_undefined_value_p): Use has_undefined_value_p.
7523 (warn_uninit, warn_uninitialized_vars,
7524 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7525 Move from tree-ssa.c.
7526 * tree-ssa.h: Adjust prototypes.
7527
7528 2013-09-17 Jan Hubicka <jh@suse.cz>
7529
7530 PR middle-end/58332
7531 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
7532 * ipa-inline.c (can_inline_edge_p): Do not downgrade
7533 FUNCTION_NOT_OPTIMIZED.
7534 * ipa-inline-analysis.c (compute_inline_parameters): Function
7535 not optimized is not inlinable unless it is alwaysinline.
7536 (inline_analyze_function): Force calls in not optimized
7537 function not inlinable.
7538
7539 2013-09-17 Jan Hubicka <jh@suse.cz>
7540
7541 PR middle-end/58329
7542 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
7543 to return NULL.
7544 * ipa.c (function_and_variable_visibility): Likewise.
7545 * ipa-profile.c (ipa_profile): Likewise.
7546
7547 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
7548
7549 PR ipa/58398
7550 * cgraph.c (cgraph_function_body_availability): Check for ifunc
7551 attribute, and don't inline the resolver in this case.
7552
7553 2013-09-17 Teresa Johnson <tejohnson@google.com>
7554
7555 * coverage.c (get_coverage_counts): Add missing newline.
7556
7557 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7558
7559 PR tree-optimization/58088
7560 * fold-const.c (mask_with_trailing_zeros): New function.
7561 (fold_binary_loc): Make sure we don't recurse infinitely
7562 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
7563 Use mask_with_trailing_zeros where appropriate.
7564
7565 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
7566
7567 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
7568 of 'prev' var to get better distance estimation.
7569
7570 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
7571
7572 * tree-inline.h (struct copy_body_data): Add transform_parameter.
7573 * tree-inline.c (is_parameter_of): New predicate.
7574 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
7575 a parameter has been remapped.
7576 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
7577 (optimize_inline_calls): Initialize transform_parameter.
7578 (copy_gimple_seq_and_replace_locals): Likewise.
7579 (tree_function_versioning): Likewise.
7580 (maybe_inline_call_in_expr): Likewise.
7581
7582 2013-09-17 Nick Clifton <nickc@redhat.com>
7583
7584 * config/msp430/msp430-protos.h: Add prototypes for new functions.
7585 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
7586 interrupt handlers.
7587 (is_attr_func): New function.
7588 (msp430_is_interrupt_func): New function.
7589 (is_naked_func): New function.
7590 (is_reentrant_func): New function.
7591 (is_critical_func): New function.
7592 (msp430_start_function): Add annotations for function attributes.
7593 (msp430_attr): New function.
7594 (msp430_attribute_table): New.
7595 (msp430_function_section): New function.
7596 (TARGET_ASM_FUNCTION_SECTION): Define.
7597 (msp430_builtin): New enum.
7598 (msp430_init_builtins): New function.
7599 (msp430_builtin_devl): New function.
7600 (msp430_expand_builtin): New function.
7601 (TARGET_INIT_BUILTINS): Define.
7602 (TARGET_EXPAND_BUILTINS): Define.
7603 (TARGET_BUILTIN_DECL): Define.
7604 (msp430_expand_prologue): Add support for naked, interrupt,
7605 critical and reentrant functions.
7606 (msp430_expand_epilogue): Likewise.
7607 (msp430_print_operand): Handle 'O' character.
7608 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
7609 NO_TRAMPOLINES.
7610 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
7611 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
7612 (pushm): Use a 'n' rather than an 'i' constraint.
7613 (msp_return): Add generation of the interrupt return instruction.
7614 (disable_interrupts): New pattern.
7615 (enable_interrupts): New pattern.
7616 (push_intr_state): New pattern.
7617 (pop_intr_state): New pattern.
7618 (bic_SR): New pattern.
7619 (bis_SR): New pattern.
7620 * doc/extend.texi: Document MSP430 function attributes and builtin
7621 functions.
7622
7623 2013-09-17 Richard Biener <rguenther@suse.de>
7624
7625 PR tree-optimization/58432
7626 * tree-loop-distribution.c (tree_loop_distribution): Also
7627 scan PHIs for outside loop uses and seed a partition from them.
7628
7629 2013-09-17 Bin Cheng <bin.cheng@arm.com>
7630
7631 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
7632 (restructure_reference): Call backtrace_base_for_ref.
7633
7634 2013-09-17 Alan Modra <amodra@gmail.com>
7635
7636 PR target/57589
7637 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
7638 patch.
7639
7640 2013-09-16 DJ Delorie <dj@redhat.com>
7641
7642 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
7643 vregs location for RL78/G10.
7644 (rl78_expand_prologue): Avoid SEL on G10.
7645 (rl78_expand_epilogue): Likewise.
7646 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
7647 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
7648 __RL78_G10__ when appropriate.
7649 (ASM_SPEC): Pass -mg10 along to the assembler.
7650 * config/rl78/rl78.md (sel_rb): Disable for G10.
7651 * config/rl78/rl78.opt: Add -mg10 option.
7652 * config/rl78/t-rl78: Add -mg10 multilib.
7653
7654 2013-09-16 Xinliang David Li <davidxl@google.com>
7655
7656 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
7657 * omp-low.c (omp_max_vf): Ditto.
7658 (expand_omp_simd): Ditto.
7659 * tree-vectorizer.c (vectorize_loops): Ditto.
7660 (gate_vect_slp): Ditto.
7661 (gate_increase_alignment): Ditto.
7662 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
7663 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
7664 (gate_tree_vectorize): Name change.
7665 (tree_vectorize): Ditto.
7666 (pass_vectorize::gate): Call new function.
7667 (pass_vectorize::execute): Ditto.
7668 * opts.c: O3 default setting change.
7669 (finish_options): Check new flag.
7670 * doc/invoke.texi: Document new flags.
7671 * common.opt: New flags.
7672
7673 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
7674
7675 * doc/tm.texi.in (Cond Exec Macros): Remove node.
7676 (Condition Code): Don't reference it.
7677 * doc/tm.texi: Regenerate.
7678
7679 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
7680
7681 PR middle-end/58418
7682 * lra-constraints.c (undo_optional_reloads): Consider all optional
7683 reload even if it did not get a hard reg.
7684
7685 2013-09-16 Teresa Johnson <tejohnson@google.com>
7686
7687 * dumpfile.c (dump_loc): Remove newline emission.
7688 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
7689 emission to dump_printf_loc calls where missing.
7690 (vect_mark_for_runtime_alias_test): Ditto.
7691 (vect_analyze_data_ref_dependence): Ditto.
7692 (vect_analyze_data_ref_dependences): Ditto.
7693 (vect_slp_analyze_data_ref_dependence): Ditto.
7694 (vect_slp_analyze_data_ref_dependences): Ditto.
7695 (vect_compute_data_ref_alignment): Ditto.
7696 (vect_update_misalignment_for_peel): Ditto.
7697 (vect_verify_datarefs_alignment): Ditto.
7698 (vector_alignment_reachable_p): Ditto.
7699 (vect_get_data_access_cost): Ditto.
7700 (vect_enhance_data_refs_alignment): Ditto.
7701 (vect_find_same_alignment_drs): Ditto.
7702 (vect_analyze_data_refs_alignment): Ditto.
7703 (vect_analyze_group_access): Ditto.
7704 (vect_analyze_data_ref_access): Ditto.
7705 (vect_analyze_data_ref_accesses): Ditto.
7706 (vect_prune_runtime_alias_test_list): Ditto.
7707 (vect_analyze_data_refs): Ditto.
7708 (vect_create_addr_base_for_vector_ref): Ditto.
7709 (vect_create_data_ref_ptr): Ditto.
7710 (vect_grouped_store_supported): Ditto.
7711 (vect_grouped_load_supported): Ditto.
7712 * value-prof.c (check_counter): Ditto.
7713 (check_ic_target): Ditto.
7714 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
7715 (vect_recog_widen_mult_pattern): Ditto.
7716 (vect_recog_widen_sum_pattern): Ditto.
7717 (vect_recog_over_widening_pattern): Ditto.
7718 (vect_recog_widen_shift_pattern): Ditto.
7719 (vect_recog_rotate_pattern): Ditto.
7720 (vect_recog_vector_vector_shift_pattern): Ditto.
7721 (vect_recog_divmod_pattern): Ditto.
7722 (vect_recog_mixed_size_cond_pattern): Ditto.
7723 (vect_recog_bool_pattern): Ditto.
7724 (vect_pattern_recog_1): Ditto.
7725 (vect_pattern_recog): Ditto.
7726 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
7727 (vect_is_simple_iv_evolution): Ditto.
7728 (vect_analyze_scalar_cycles_1): Ditto.
7729 (vect_get_loop_niters): Ditto.
7730 (vect_analyze_loop_1): Ditto.
7731 (vect_analyze_loop_form): Ditto.
7732 (vect_analyze_loop_operations): Ditto.
7733 (vect_analyze_loop_2): Ditto.
7734 (vect_analyze_loop): Ditto.
7735 (report_vect_op): Ditto.
7736 (vect_is_slp_reduction): Ditto.
7737 (vect_is_simple_reduction_1): Ditto.
7738 (vect_get_known_peeling_cost): Ditto.
7739 (vect_estimate_min_profitable_iters): Ditto.
7740 (vect_model_reduction_cost): Ditto.
7741 (vect_model_induction_cost): Ditto.
7742 (get_initial_def_for_induction): Ditto.
7743 (vect_create_epilog_for_reduction): Ditto.
7744 (vectorizable_reduction): Ditto.
7745 (vectorizable_induction): Ditto.
7746 (vectorizable_live_operation): Ditto.
7747 (vect_loop_kill_debug_uses): Ditto.
7748 (vect_transform_loop): Ditto.
7749 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
7750 (vect_stmt_relevant_p): Ditto.
7751 (process_use): Ditto.
7752 (vect_mark_stmts_to_be_vectorized): Ditto.
7753 (vect_model_simple_cost): Ditto.
7754 (vect_model_promotion_demotion_cost): Ditto.
7755 (vect_model_store_cost): Ditto.
7756 (vect_get_store_cost): Ditto.
7757 (vect_model_load_cost): Ditto.
7758 (vect_get_load_cost): Ditto.
7759 (vect_init_vector_1): Ditto.
7760 (vect_get_vec_def_for_operand): Ditto.
7761 (vect_finish_stmt_generation): Ditto.
7762 (vectorizable_call): Ditto.
7763 (vectorizable_conversion): Ditto.
7764 (vectorizable_assignment): Ditto.
7765 (vectorizable_shift): Ditto.
7766 (vectorizable_operation): Ditto.
7767 (vectorizable_store): Ditto.
7768 (vectorizable_load): Ditto.
7769 (vectorizable_condition): Ditto.
7770 (vect_analyze_stmt): Ditto.
7771 (vect_transform_stmt): Ditto.
7772 (vect_is_simple_use): Ditto.
7773 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
7774 (vect_can_advance_ivs_p): Ditto.
7775 (vect_update_ivs_after_vectorizer): Ditto.
7776 (vect_do_peeling_for_loop_bound): Ditto.
7777 (vect_gen_niters_for_prolog_loop): Ditto.
7778 (vect_update_inits_of_drs): Ditto.
7779 (vect_create_cond_for_alias_checks): Ditto.
7780 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
7781 (vect_build_slp_tree_1): Ditto.
7782 (vect_supported_load_permutation_p): Ditto.
7783 (vect_analyze_slp_instance): Ditto.
7784 (vect_analyze_slp): Ditto.
7785 (vect_make_slp_decision): Ditto.
7786 (vect_detect_hybrid_slp): Ditto.
7787 (vect_bb_vectorization_profitable_p): Ditto.
7788 (vect_slp_analyze_bb_1): Ditto.
7789 (vect_update_slp_costs_according_to_vf): Ditto.
7790 (vect_get_mask_element): Ditto.
7791 (vect_transform_slp_perm_load): Ditto.
7792 (vect_schedule_slp_instance): Ditto.
7793 (vect_schedule_slp): Ditto.
7794 (vect_slp_transform_bb): Ditto.
7795 * profile.c (read_profile_edge_counts): Ditto.
7796 (compute_branch_probabilities): Ditto.
7797 * coverage.c (get_coverage_counts): Ditto.
7798
7799 2013-09-16 Diego Novillo <dnovillo@google.com>
7800
7801 * tree-core.h: Add missing comment lines from refactoring of tree.h.
7802
7803 2013-09-16 Jan Hubicka <jh@suse.cz>
7804
7805 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
7806 abstract functions; for static functions check the presence of body.
7807
7808 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7809
7810 * config/aarch64/aarch64-simd-builtins.def (fma): New.
7811 * config/aarch64/aarch64-simd.md
7812 (aarch64_mla_elt<mode>): New.
7813 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7814 (aarch64_mls_elt<mode>): Likewise.
7815 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7816 (aarch64_fma4_elt<mode>): Likewise.
7817 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7818 (aarch64_fma4_elt_to_128v2df): Likewise.
7819 (aarch64_fma4_elt_to_64df): Likewise.
7820 (fnma<mode>4): Likewise.
7821 (aarch64_fnma4_elt<mode>): Likewise.
7822 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7823 (aarch64_fnma4_elt_to_128v2df): Likewise.
7824 (aarch64_fnma4_elt_to_64df): Likewise.
7825 * config/aarch64/iterators.md (VDQSF): New.
7826 * config/aarch64/arm_neon.h
7827 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
7828 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
7829
7830 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7831
7832 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
7833 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
7834 (aarch64_mul3_elt_to_128df): Likewise.
7835 (aarch64_mul3_elt_to_64v2df): Likewise.
7836 * config/aarch64/iterators.md (VEL): Also handle DFmode.
7837 (VMUL): New.
7838 (VMUL_CHANGE_NLANES) Likewise.
7839 (h_con): Likewise.
7840 (f): Likewise.
7841 * config/aarch64/arm_neon.h
7842 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
7843
7844 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7845
7846 * config/aarch64/arm_neon.h
7847 (vcvtx_high_f32_f64): Fix parameters.
7848
7849 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7850 Uros Bizjak <ubizjak@gmail.com>
7851
7852 * config/alpha.c: Include tree-ssanames.h.
7853
7854 2013-09-16 Richard Biener <rguenther@suse.de>
7855
7856 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
7857 (dot_rdg_1): Handle control_dd.
7858 (create_edge_for_control_dependence): New function.
7859 (create_rdg_edges): Add control dependences if asked for.
7860 (build_rdg): Likewise.
7861 (generate_loops_for_partition): If there are not necessary
7862 control stmts remove all their dependencies.
7863 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
7864 (distribute_loop): Pass on control dependences.
7865 (tree_loop_distribution): Compute control dependences and remove
7866 restriction on number of loop nodes.
7867
7868 2013-09-16 Jakub Jelinek <jakub@redhat.com>
7869
7870 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
7871 for internal calls.
7872
7873 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
7874
7875 * cse.c (try_const_anchors): Punt on CC modes.
7876
7877 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7878
7879 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
7880
7881 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
7882
7883 PR target/58382
7884 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
7885 calls to word_mode.
7886
7887 2013-09-14 Iain Sandoe <iain@codesourcery.com>
7888
7889 PR target/48094
7890 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
7891 seen.
7892 (darwin_objc1_section): Likewise.
7893 (darwin_file_end): Emit Image Info section when required.
7894
7895 2013-09-14 Jan Hubicka <jh@suse.cz>
7896
7897 * tree-into-ssa.c (gate_into_ssa): New.
7898 (pass_data_build_ssa): Use it.
7899 * cgraph.h (expand_thunk): Update prototype.
7900 * cgraphunit.c (analyze_function): Expand thunks early.
7901 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
7902 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
7903 set lowered flag; do not add new function.
7904 (assemble_thunks_and_aliases): Update.
7905 * tree-ssa.c (gate_init_datastructures): New gate.
7906 (pass_data_init_datastructures): Use it.
7907
7908 2013-09-14 Iain Sandoe <iain@codesourcery.com>
7909
7910 PR target/58269
7911 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
7912 xmm register set described in the psABI.
7913
7914 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
7915
7916 * dwarf2out.c (should_emit_struct_debug): Add check
7917 for type_decl variable is not NULL.
7918
7919 2013-09-13 Jacek Caban <jacek@codeweavers.com>
7920
7921 * config.gcc: Use new winnt-c.c target hooks
7922 * config/t-winnt: New file
7923 * config/winnt-c.c: New file
7924 * doc/tm.texi.in: Document new hook
7925 * doc/tm.texi: Regenerated
7926
7927 2013-09-13 Jan Hubicka <jh@suse.cz>
7928
7929 PR middle-end/58094
7930 * ipa-inline.c (check_callers): New function.
7931 (check_caller_edge): Remove.
7932 (want_inline_function_to_all_callers_p): Also permit alises that are
7933 called dirrectly.
7934 (inline_to_all_callers): Terminate the walk when devirtualization
7935 introduce new calls.
7936
7937 2013-09-13 Jan Hubicka <jh@suse.cz>
7938
7939 * ipa-inline-analysis.c (struct growth_data): Add node.
7940 (do_estimate_growth_1): Fix detection of recursion.
7941
7942 2013-09-13 Jakub Jelinek <jakub@redhat.com>
7943
7944 PR tree-optimization/58392
7945 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
7946 to avoid shadowing of outer loop variable. If
7947 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
7948 replace_by_duplicate_decl simduid of loops that have it set and
7949 set dest_cfun->has_simduid_loops and/or
7950 dest_cfun->has_force_vect_loops.
7951 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
7952 instead of maybe_lookup_decl.
7953 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
7954 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
7955 call. Copy over force_vect and copy and remap simduid. Set
7956 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
7957 (copy_cfg_body): Remove blocks_to_copy argument. Use
7958 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
7959 caller. Don't set cfun->has_simduid_loops and/or
7960 cfun->has_force_vect_loops here.
7961 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
7962 caller.
7963 (expand_call_inline, tree_function_versioning): Adjust copy_body
7964 callers.
7965
7966 2013-09-13 Martin Jambor <mjambor@suse.cz>
7967
7968 PR bootstrap/58388
7969 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
7970 the assert if the edge was a speculative one.
7971
7972 2013-09-13 Richard Biener <rguenther@suse.de>
7973
7974 * tree-data-ref.h (known_dependences_p): Move here ...
7975 * tree-loop-distribution.c (known_dependences_p): ... from here.
7976 (dump_rdg_component, debug_rdg_component): Remove.
7977 (dump_rdg): Adjust.
7978 (generate_loops_for_partition): Use gimple_uid instead of
7979 relying on matching stmt visit order.
7980 (rdg_build_partitions): Take starting stmt vector.
7981 (ldist_gen): Merge into ...
7982 (distribute_loop): ... this function. Do not compute starting
7983 vertices vector.
7984 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
7985
7986 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7987
7988 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
7989 Set type attribute correctly. Set predicable_short_it attribute.
7990 (cmpsi_shiftsi): Remove %? from output template.
7991
7992 2013-09-13 Richard Biener <rguenther@suse.de>
7993
7994 * tree-loop-distribution.c (struct rdg_component,
7995 rdg_defs_used_in_other_loops_p, free_rdg_components,
7996 rdg_build_components): Remove.
7997 (stmts_from_loop): Do not record virtual PHIs.
7998 (generate_loops_for_partition): Skip virtual PHIs.
7999 (build_rdg_partition_for_component): Rename to ...
8000 (build_rdg_partition_for_vertex): ... this and adjust.
8001 (rdg_build_partitions): Take a vector of starting vertices
8002 instead of components. Remove unnecessary leftover handling.
8003 (ldist_gen): Do not build components or record other stores.
8004 (distribute_loop): Do not distribute loops containing stmts
8005 with side-effects.
8006
8007 2013-09-13 Christian Bruel <christian.bruel@st.com>
8008
8009 PR target/58314
8010 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
8011
8012 2013-09-13 Kai Tietz <ktietz@redhat.com>
8013
8014 * config.gcc: Separate cases for mingw and cygwin targets,
8015 and add 64-bit cygwin target case.
8016
8017 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
8018 dll-export inline-functions.
8019 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
8020
8021
8022 2013-09-13 Jeff Law <law@redhat.com>
8023
8024 PR middle-end/58387
8025 Revert:
8026 2013-09-06 Jeff Law <law@redhat.com>
8027
8028 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8029 edge implied equivalences into successor phis.
8030
8031 2013-09-12 DJ Delorie <dj@redhat.com>
8032
8033 * config/rl78/rl78-virt.md: Change from | to \; for asm line
8034 separators.
8035
8036 2013-09-12 Brooks Moses <bmoses@google.com>
8037
8038 PR driver/42955
8039 * Makefile.in: Do not install driver binaries in $(target)/bin.
8040
8041 2013-09-12 DJ Delorie <dj@redhat.com>
8042
8043 * config/rl78/rl78.opt (mrelax): New.
8044 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
8045 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
8046
8047 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
8048 between SP and FP.
8049 (rl78_expand_epilogue): Likewise.
8050
8051 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
8052
8053 PR middle-end/58335
8054 * lra-eliminations.c (remove_reg_equal_offset_note): New.
8055 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
8056 pointer elimination with using remove_reg_equal_offset_note.
8057
8058 2013-09-12 DJ Delorie <dj@redhat.com>
8059
8060 * config/msp430/: New port.
8061 * config.gcc (msp430): Added.
8062 * doc/invoke.texi: Document MSP430 options.
8063 * doc/install.texi: Document msp430-elf
8064 * doc/md.texi: Document msp430-elf
8065 * doc/contrib.texi: Document msp430-elf
8066
8067 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
8068 PSImode.
8069
8070 2013-09-12 Martin Jambor <mjambor@suse.cz>
8071
8072 PR ipa/58389
8073 * ipa-prop.c (remove_described_reference): Give up if the edge in the
8074 reference descriptor is NULL.
8075 (ipa_edge_removal_hook): If owning a reference descriptor, set its
8076 edge to NULL.
8077
8078 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8079
8080 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
8081 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
8082 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
8083 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
8084 make_temp_ssa_name): move to tree-ssanames.h
8085 * tree-ssa-alias.h: Move prototype.
8086 * tree-ssa.h: Include tree-ssanames.h.
8087 * tree-ssanames.c (FREE_SSANAMES): Move to here.
8088 * tree-ssanames.h: New. Move items from tree-flow*.h
8089 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
8090
8091 2013-09-12 Richard Biener <rguenther@suse.de>
8092
8093 PR tree-optimization/58404
8094 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
8095 propagate non-invariant addresses into dereferences wrapped
8096 in component references.
8097
8098 2013-09-12 Richard Biener <rguenther@suse.de>
8099
8100 PR tree-optimization/58402
8101 * passes.def: Move pass_late_warn_uninitialized later.
8102
8103 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8104
8105 * tree-ssa.h: New. Move content from tree-flow.h and
8106 tree-flow-inline.h.
8107 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
8108 Move prototypes belonging to tree-ssa.c.
8109 * tree-flow-inline.h (redirect_edge_var_map_def,
8110 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
8111 tree-ssa.h.
8112 * gimple.h: Adjust prototypes.
8113 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
8114 to...
8115 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
8116 * tree.h: Move prototype to tree-ssa.h.
8117 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
8118 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
8119 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
8120 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
8121 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
8122 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
8123 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
8124 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
8125 graphite-blocking.c, graphite-clast-to-gimple.c,
8126 graphite-dependences.c, graphite-interchange.c,
8127 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
8128 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
8129 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
8130 ipa-reference.c, ipa-split.c, ipa-utils.c,
8131 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
8132 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
8133 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
8134 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
8135 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
8136 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
8137 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
8138 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
8139 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
8140 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
8141 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
8142 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
8143 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
8144 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
8145 value-prof.c, var-tracking.c,
8146 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
8147
8148 2013-09-12 Richard Biener <rguenther@suse.de>
8149
8150 PR tree-optimization/58396
8151 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
8152 (build_rdg): Take a loop-nest parameter, fix memleaks.
8153 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
8154
8155 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8156
8157 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
8158 for SLM.
8159
8160 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
8161
8162 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
8163 The FMA instruction names should have a 'v' prefix.
8164
8165 2013-09-12 Richard Biener <rguenther@suse.de>
8166
8167 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
8168 (dot_rdg): Use popen instead of system in optional code.
8169 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
8170 (already_processed_vertex_p): Adjust.
8171 (has_anti_or_output_dependence, predecessor_has_mem_write,
8172 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
8173 rdg_flag_uses): Remove.
8174 (rdg_flag_vertex): Simplify.
8175 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
8176 remove recursion.
8177 (build_rdg_partition_for_component): Process the first vertex
8178 of a component only.
8179 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
8180
8181 2013-09-12 Alan Modra <amodra@gmail.com>
8182
8183 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
8184
8185 2013-09-11 DJ Delorie <dj@redhat.com>
8186 Nick Clifton <nickc@redhat.com>
8187
8188 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
8189 (rl78_stack_based_mem): New.
8190 * config/rl78/constraints.md (Iv08): New.
8191 (Iv16): New.
8192 (Iv24): New.
8193 (Is09): New.
8194 (Is17): New.
8195 (Is25): New.
8196 (ISsi): New.
8197 (IShi): New.
8198 (ISqi): New.
8199 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
8200 (movhi): Likewise.
8201 (movsi): Change from expand to insn-and-split.
8202 (ashrsi3): Clobber AX.
8203 (lshrsi3): New.
8204 (ashlsi3): New.
8205 (cbranchsi4): New.
8206 * config/rl78/rl78.md (CC_REG): Fix.
8207 (addsi3): Allow memory and immediate operands.
8208 (addsi3_internal): Split into...
8209 (addsi3_internal_virt): ...new, and ...
8210 (addsi3_internal_real): ...new.
8211 (subsi): New.
8212 (subsi3_internal_virt): New.
8213 (subsi3_internal_real): New.
8214 (mulsi3): Add memory operand.
8215 (mulsi3_rl78): Likewise.
8216 (mulsi3_g13): Likewise.
8217 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
8218 (cbranchqi4_real): Add more constraint options.
8219 (cbranchhi4_real): Expand pattern.
8220 (cbranchhi4_real_signed): New.
8221 (cbranchhi4_real_inverted): New.
8222 (cbranchsi4_real_lt): New.
8223 (cbranchsi4_real_ge): New.
8224 (cbranchsi4_real_signed): New.
8225 (cbranchsi4_real): New.
8226 (peephole2): New.
8227 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
8228 constant shifts.
8229 (lshrsi3_virt): Likewise.
8230 (ashlsi3_virt): Likewise.
8231 (cbranchqi4_virt_signed): New.
8232 (cbranchhi4_virt_signed): New.
8233 (cbranchsi4_virt): New.
8234 * config/rl78/rl78.c: Whitespace fixes throughout.
8235 (move_elim_pass): New.
8236 (pass_data_rl78_move_elim): New.
8237 (pass_rl78_move_elim): New.
8238 (make_pass_rl78_move_elim): New.
8239 (rl78_devirt_info): Run devirt earlier.
8240 (rl78_move_elim_info): New.
8241 (rl78_asm_file_start): Register it.
8242 (rl78_split_movsi): New.
8243 (rl78_as_legitimate_address): Allow virtual base registers when
8244 appropriate.
8245 (rl78_addr_space_convert): Remove spurious debug stuff.
8246 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
8247 (rl78_print_operand): More cases for not printing '#'.
8248 (rl78_expand_compare): Remove most of the logic.
8249 (content_memory): New.
8250 (clear_content_memory): New.
8251 (get_content_index): New.
8252 (get_content_name): New.
8253 (display_content_memory): New.
8254 (update_content): New.
8255 (record_content): New.
8256 (already_contains): New.
8257 (insn_ok_now): Re-recog insns with virtual registers.
8258 (add_postponed_content_update): New.
8259 (process_postponed_content_update): New.
8260 (gen_and_emit_move): New.
8261 (transcode_memory_rtx): Record new location content.
8262 Use gen_and_emit_move.
8263 (force_into_acc): New.
8264 (move_to_acc): Use gen_and_emit_move.
8265 (move_from_acc): Likewise.
8266 (move_acc_to_reg): Likewise.
8267 (move_to_x): Likewise.
8268 (move_to_hl): Likewise.
8269 (move_to_de): Likewise.
8270 (rl78_alloc_physical_registers_op1): Record location content.
8271 (has_constraint): New.
8272 (rl78_alloc_physical_registers_op2): Record location content.
8273 Optimize use of HL.
8274 (rl78_alloc_physical_registers_ro1): Likewise.
8275 (rl78_alloc_physical_registers_cmp): Likewise.
8276 (rl78_alloc_physical_registers_umul): Likewise.
8277 (rl78_alloc_address_registers_macax): New.
8278 (rl78_alloc_physical_registers): Initialize and set location
8279 content memory as needed.
8280 (rl78_reorg): Make sure split2 is called.
8281 (rl78_rtx_costs): New.
8282
8283 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
8284
8285 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
8286 for (not (neg ...)) and (neg (not ...)) cases.
8287
8288 2013-09-11 Richard Biener <rguenther@suse.de>
8289
8290 PR middle-end/58377
8291 * passes.def: Split critical edges before late uninit warning passes.
8292 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
8293
8294 2013-09-11 Jakub Jelinek <jakub@redhat.com>
8295
8296 PR tree-optimization/58385
8297 * fold-const.c (build_range_check): If both low and high are NULL,
8298 use omit_one_operand_loc to preserve exp side-effects.
8299
8300 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8301
8302 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
8303
8304 2013-09-11 Richard Biener <rguenther@suse.de>
8305
8306 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
8307 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
8308 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
8309 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
8310 stmts_from_loop, known_dependences_p, build_empty_rdg,
8311 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
8312 * tree-loop-distribution.c: ... here.
8313 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
8314 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
8315 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
8316 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
8317 * tree-loop-distribution.c: ... here.
8318 * tree-loop-distribution.c: Include gimple-pretty-print.h.
8319 (struct partition_s): Add loops member.
8320 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
8321 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
8322 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
8323
8324 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8325 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8326 Sergey Lega <sergey.s.lega@intel.com>
8327 Anna Tikhonova <anna.tikhonova@intel.com>
8328 Ilya Tocar <ilya.tocar@intel.com>
8329 Andrey Turetskiy <andrey.turetskiy@intel.com>
8330 Ilya Verbin <ilya.verbin@intel.com>
8331 Kirill Yukhin <kirill.yukhin@intel.com>
8332 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8333
8334 * config/i386/constraints.md (k): New.
8335 (Yk): Ditto.
8336 * config/i386/i386.c (const regclass_map): Add new mask registers.
8337 (dbx_register_map): Ditto.
8338 (dbx64_register_map): Ditto.
8339 (svr4_dbx_register_map): Ditto.
8340 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
8341 disabled.
8342 (ix86_preferred_reload_class): Disable constants for mask registers.
8343 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
8344 (ix86_hard_regno_mode_ok): Support new mask registers.
8345 (x86_order_regs_for_local_alloc): Ditto.
8346 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
8347 (FIXED_REGISTERS): Add new mask registers.
8348 (CALL_USED_REGISTERS): Ditto.
8349 (REG_ALLOC_ORDER): Ditto.
8350 (VALID_MASK_REG_MODE): New.
8351 (FIRST_MASK_REG): Ditto.
8352 (LAST_MASK_REG): Ditto.
8353 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
8354 (MAYBE_MASK_CLASS_P): New.
8355 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
8356 (REG_CLASS_CONTENTS): Ditto.
8357 (MASK_REGNO_P): New.
8358 (ANY_MASK_REG_P): Ditto.
8359 (HI_REGISTER_NAMES): Add new mask registers.
8360 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
8361 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
8362 mask registers.
8363 (attribute "type"): Add mskmov, msklog.
8364 (attribute "length_immediate"): Support them.
8365 (attribute "memory"): Ditto.
8366 (attribute "prefix_0f"): Ditto.
8367 (*movhi_internal): Support new mask registers.
8368 (*movqi_internal): Ditto.
8369 (define_split): Split out clobber pattern is a logic
8370 insn on mask registers.
8371 (*k<logic><mode>): New.
8372 (*andhi_1): Extend to support mask regs.
8373 (*andqi_1): Extend to support mask regs.
8374 (kandn<mode>): New.
8375 (define_split): Split and-not to and and not if operands
8376 are not mask regs.
8377 (*<code><mode>_1): Separate HI mode to new pattern...
8378 (*<code>hi_1): This.
8379 (*<code>qi_1): Extend to support mask regs.
8380 (kxnor<mode>): New.
8381 (kortestzhi): Ditto.
8382 (kortestchi): Ditto.
8383 (kunpckhi): Ditto.
8384 (*one_cmpl<mode>2_1): Remove HImode and handle it...
8385 (*one_cmplhi2_1): ...Here, now with mask registers support.
8386 (*one_cmplqi2_1): Support new mask registers.
8387 (HI/QImode arithmetics splitter): Don't split if mask registers
8388 are used.
8389 (HI/QImode not splitter): Ditto.
8390 * config/i386/predicated.md (mask_reg_operand): New.
8391 (general_reg_operand): Ditto.
8392
8393 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8394
8395 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
8396 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
8397
8398 2013-09-10 Jeff Law <law@redhat.com>
8399
8400 PR tree-optimization/58380
8401 * tree-ssa-threadupdate.c (thread_block): Recognize another case
8402 of threading through a buried loop header.
8403
8404 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
8405 return value for single successor case.
8406
8407 2013-09-10 Jan Hubicka <jh@suse.cz>
8408
8409 * ipa-devirt.c (ipa_devirt): Enable with LTO.
8410
8411 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
8412
8413 PR target/58361
8414 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
8415 support conditional execution.
8416 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
8417
8418 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
8419
8420 * lra.c (lra): Clear lra_optional_reload_pseudos before every
8421 constraint pass.
8422 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
8423 Check destination too to check move insn.
8424 (undo_optional_reloads): Add check that the original peudo did not
8425 changed its allocation and the optional reload was inherited on last
8426 inheritance pass. Break loop after deciding to keep optional reload.
8427 (lra_undo_inheritance): Add check that inherited pseudo still in
8428 memory.
8429
8430 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
8431
8432 * config/aarch64/aarch64.md (generic_sched): New.
8433 * config/aarch64/aarch64-generic.md (load): Make conditional
8434 on generic_sched attribute.
8435 (nonload): Likewise.
8436
8437 2013-09-10 Jan Hubicka <jh@suse.cz>
8438
8439 * lto-cgraph.c: Include ipa-utils.h.
8440 (compute_ltrans_boundary): Also add possible targets into the boundary.
8441
8442 2013-09-10 Jan Hubicka <jh@suse.cz>
8443
8444 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
8445 VAR_DECL of vtable rather than full expression.
8446
8447 2013-09-10 Jan Hubicka <jh@suse.cz>
8448 Paolo Carlini <paolo.carlini@oracle.com>
8449
8450 * cgraphunit.c (analyze_functions): Save input_location, set it
8451 to UNKNOWN_LOCATION and restore it at the end.
8452
8453 2013-09-10 Martin Jambor <mjambor@suse.cz>
8454
8455 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
8456 represented by a thunk.
8457
8458 2013-09-10 Jeff Law <law@redhat.com>
8459
8460 PR tree-optimization/58343
8461 * tree-ssa-threadupdate.c (thread_block): Identify and disable
8462 jump threading requests through loop headers buried in the middle
8463 of a jump threading path.
8464
8465 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
8466 in return value/type.
8467
8468 2013-09-10 Jakub Jelinek <jakub@redhat.com>
8469
8470 PR rtl-optimization/58365
8471 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
8472 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
8473 it differs.
8474
8475 2013-09-10 Richard Biener <rguenther@suse.de>
8476
8477 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
8478 * tree-data-ref.c (create_rdg_vertices): Collect all data
8479 references, signal failure to the caller, use data-ref API.
8480 (build_rdg): Compute data references only once. Maintain lifetime
8481 of data references and data dependences from within RDG.
8482 (free_rdg): Free dependence relations.
8483 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
8484 inventing extra dependences.
8485 (distribute_loop): Update for RDG API changes.
8486
8487 2013-09-10 Kai Tietz <ktietz@redhat.com>
8488
8489 * doc/invoke.texi (fms-extensions): Document changed
8490 behavior for ms-abi targets.
8491 * config/i386/i386.c (ix86_option_override_internal):
8492 Set default value of option -fms-extension for ms-abi targets.
8493
8494 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
8495
8496 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
8497
8498 2013-09-10 Alan Modra <amodra@gmail.com>
8499
8500 PR target/58330
8501 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
8502
8503 2013-09-10 Alan Modra <amodra@gmail.com>
8504
8505 * config/rs6000/predicates.md (add_cint_operand): New.
8506 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
8507 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
8508 using add_cint_operand.
8509 (largetoc_high_plus_aix): Likewise.
8510
8511 2013-09-09 Jakub Jelinek <jakub@redhat.com>
8512
8513 PR tree-optimization/58364
8514 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
8515 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
8516 the current range can't be an unconditional true or false.
8517
8518 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
8519
8520 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
8521
8522 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
8523
8524 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
8525
8526 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
8527
8528 PR c++/43452
8529 * doc/invoke.texi (-Wdelete-incomplete): Document it.
8530
8531 2013-09-09 Ian Bolton <ian.bolton@arm.com>
8532
8533 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
8534 NO_REGS for immediate that can't be moved directly into FP_REGS.
8535
8536 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8537
8538 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
8539 comparison with negated operand.
8540 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
8541 RTL form.
8542
8543 2013-09-09 Richard Biener <rguenther@suse.de>
8544
8545 PR middle-end/58326
8546 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
8547 of a subloop record all its block as affecting loop-closed SSA form.
8548
8549 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8550
8551 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
8552 of an rtx/bitpos pair.
8553 (store_fixed_bit_field): Update accordingly.
8554
8555 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8556
8557 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
8558 GEN_INT.
8559 * builtins.c (expand_errno_check): Likewise.
8560 * dwarf2cfi.c (init_return_column_size): Likewise.
8561 * except.c (sjlj_mark_call_sites): Likewise.
8562 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
8563 * lra-constraints.c (emit_inc): Likewise.
8564 * ree.c (combine_set_extension): Likewise.
8565 * regmove.c (fixup_match_2): Likewise.
8566 * reload1.c (inc_for_reload): Likewise.
8567
8568 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8569
8570 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
8571 (force_to_mode, simplify_shift_const_1, simplify_comparison):
8572 Use gen_int_mode with the mode of the associated simplify_* call.
8573 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
8574 * expmed.c (expand_shift_1): Likewise.
8575 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8576 * loop-iv.c (iv_number_of_iterations): Likewise.
8577 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8578 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
8579
8580 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8581
8582 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
8583 of the associated expand_* call.
8584 (asan_emit_stack_protection): Likewise.
8585 * builtins.c (round_trampoline_addr): Likewise.
8586 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
8587 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
8588 (emit_store_flag): Likewise.
8589 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
8590 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
8591 Likewise.
8592 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8593 * ifcvt.c (noce_try_store_flag_constants): Likewise.
8594 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8595 * modulo-sched.c (generate_prolog_epilog): Likewise.
8596 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
8597 (expand_ctz, expand_ffs, expand_unop): Likewise.
8598
8599 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8600
8601 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
8602 of the associated gen_rtx_* call.
8603 * caller-save.c (init_caller_save): Likewise.
8604 * combine.c (find_split_point, make_extraction): Likewise.
8605 (make_compound_operation): Likewise.
8606 * dwarf2out.c (mem_loc_descriptor): Likewise.
8607 * explow.c (plus_constant, probe_stack_range): Likewise.
8608 * expmed.c (expand_mult_const): Likewise.
8609 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
8610 * reload1.c (init_reload): Likewise.
8611 * valtrack.c (cleanup_auto_inc_dec): Likewise.
8612 * var-tracking.c (adjust_mems): Likewise.
8613 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
8614 rather than gen_rtx_fmt_ee.
8615
8616 2013-09-09 Jan Hubicka <jh@suse.cz>
8617
8618 PR middle-end/58294
8619 * value-prof.c (gimple_ic): Copy also abnormal edges.
8620
8621 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8622
8623 * asan.c (asan_shadow_cst): Use gen_int_mode.
8624
8625 2013-09-08 Jan Hubicka <jh@suse.cz>
8626
8627 * ipa-profile.c: Add toplevel comment.
8628 (ipa_propagate_frequency_1): Be more conservative when profile is read.
8629 (contains_hot_call_p): New function.
8630 (ipa_propagate_frequency): Set frequencies based on counts when
8631 profile is read.
8632 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
8633 profile; do not tamper with profile after inlining if it is read.
8634
8635 2013-09-08 Jan Hubicka <jh@suse.cz>
8636
8637 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
8638 speculative edges.
8639
8640 2013-09-08 Jan Hubicka <jh@suse.cz>
8641
8642 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
8643 summary generation.
8644
8645 2013-09-08 Jeff Law <law@redhat.com>
8646
8647 PR bootstrap/58340
8648 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
8649 of 'found'.
8650
8651 2013-09-08 Andi Kleen <ak@linux.intel.com>
8652
8653 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
8654
8655 2013-09-08 Jan Hubicka <jh@suse.cz>
8656
8657 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
8658
8659 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
8660
8661 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
8662 for non-debug insns.
8663 * lra.c (new_insn_reg): Take the containing insn as a parameter.
8664 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
8665 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
8666 accordingly.
8667
8668 2013-09-08 Jan Hubicka <jh@suse.cz>
8669
8670 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
8671 targets and devirtualize to BUILT_IN_UNREACHABLE.
8672 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
8673 * ipa.c (walk_polymorphic_call_targets): New function.
8674 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
8675 functions; use the new timevar.
8676 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
8677 was removed from the program.
8678 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
8679 not consider it in the walk when its vtable is dead.
8680 (possible_polymorphic_call_targets_1): Pass anonymous flag to
8681 record_binfo.
8682 (devirt_variable_node_removal_hook): New function.
8683 (possible_polymorphic_call_targets): Also register
8684 devirt_variable_node_removal_hook.
8685 (ipa_devirt): Do not do non-speculative devirtualization.
8686 (gate_ipa_devirt): One execute if devirtualizing speculatively.
8687
8688 2013-09-08 Jan Hubicka <jh@suse.cz>
8689
8690 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
8691 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
8692 varpool_remove_variable_insertion_hook): Declare.
8693 * varpool.c (varpool_node_hook_list): New structure.
8694 (first_varpool_node_removal_hook,
8695 first_varpool_variable_insertion_hook): New variables.
8696 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
8697 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
8698 varpool_remove_variable_insertion_hook,
8699 varpool_call_variable_insertion_hooks): New functions.
8700 (varpool_remove_node): Use it.
8701
8702 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
8703
8704 PR c++/54941
8705 * diagnostic.c (diagnostic_build_prefix): When s.file is
8706 "<built-in>" don't output line and column numbers.
8707
8708 2013-09-06 Jan Hubicka <jh@suse.cz>
8709
8710 * cgraphunit.c (expand_thunk): Get body before touching arguments.
8711 * lto-streamer-out.c: Stream thunks, too.
8712 * lto-streamer-in.c (input_function): Pop cfun here
8713 (lto_read_body): Instead of here.
8714
8715 2013-09-06 Caroline Tice <cmtice@google.com>
8716
8717 * doc/install.texi: Add documentation for the --enable-vtable-verify
8718 and the --disable-libvtv configure options.
8719
8720 2013-09-06 Jeff Law <law@redhat.com>
8721
8722 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8723 edge implied equivalences into successor phis.
8724
8725 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
8726
8727 * resource.c (mark_referenced_resources): Handle COND_EXEC.
8728
8729 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
8730
8731 * resource.c (mark_target_live_regs): Compute resources taking
8732 into account if a call is predicated or not.
8733
8734 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
8735
8736 * toplev.c (output_stack_usage): Be prepared for suffixes created by
8737 the compiler in the function names.
8738
8739 2013-09-06 Jan Hubicka <jh@suse.cz>
8740
8741 PR middle-end/58094
8742 * ipa-inline.c (has_caller_p): New function.
8743 (want_inline_function_to_all_callers_p): Use it.
8744 (sum_callers, inline_to_all_callers): Break out from ...
8745 (ipa_inline): ... here.
8746
8747 2013-09-06 Jan Hubicka <jh@suse.cz>
8748
8749 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
8750 only when AVX is enabled.
8751
8752 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8753
8754 * config/aarch64/aarch64.md
8755 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
8756 is fpsimd_<load/store>2.
8757 (load_pair<mode>): Likewise.
8758 (store_pair<mode>): Likewise.
8759
8760 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8761
8762 * config/arm/types.md (type): Add "mrs" type.
8763 * config/aarch64/aarch64.md
8764 (aarch64_load_tp_hard): Make type "mrs".
8765 * config/arm/arm.md
8766 (load_tp_hard): Make type "mrs".
8767 * config/arm/cortex-a15.md: Update with new attributes.
8768 * config/arm/cortex-a5.md: Update with new attributes.
8769 * config/arm/cortex-a53.md: Update with new attributes.
8770 * config/arm/cortex-a7.md: Update with new attributes.
8771 * config/arm/cortex-a8.md: Update with new attributes.
8772 * config/arm/cortex-a9.md: Update with new attributes.
8773 * config/arm/cortex-m4.md: Update with new attributes.
8774 * config/arm/cortex-r4.md: Update with new attributes.
8775 * config/arm/fa526.md: Update with new attributes.
8776 * config/arm/fa606te.md: Update with new attributes.
8777 * config/arm/fa626te.md: Update with new attributes.
8778 * config/arm/fa726te.md: Update with new attributes.
8779
8780 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8781
8782 * config/aarch64/aarch64.md
8783 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
8784 (*movtf_aarch64): Likewise.
8785 * config/arm/arm.md
8786 (thumb1_movdi_insn): Use "multiple" for type where more than one
8787 instruction is used for a move.
8788 (*arm32_movhf): Likewise.
8789 (*thumb_movdf_insn): Likewise.
8790
8791 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8792
8793 * config/arm/types.md (type): Rename fcpys to fmov.
8794 * config/arm/vfp.md
8795 (*arm_movsi_vfp): Rename type fcpys as fmov.
8796 (*thumb2_movsi_vfp): Likewise
8797 (*movhf_vfp_neon): Likewise
8798 (*movhf_vfp): Likewise
8799 (*movsf_vfp): Likewise
8800 (*thumb2_movsf_vfp): Likewise
8801 (*movsfcc_vfp): Likewise
8802 (*thumb2_movsfcc_vfp): Likewise
8803 * config/aarch64/aarch64-simd.md
8804 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
8805 * config/aarch64/aarch64.md
8806 (*movsi_aarch64): Replace type mov_reg with fmovs.
8807 (*movdi_aarch64): Likewise
8808 (*movsf_aarch64): Likewise
8809 (*movdf_aarch64): Likewise
8810 * config/arm/arm.c
8811 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
8812 * config/arm/iwmmxt.md
8813 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
8814 * config/arm/arm1020e.md: Update with new attributes.
8815 * config/arm/cortex-a15-neon.md: Update with new attributes.
8816 * config/arm/cortex-a5.md: Update with new attributes.
8817 * config/arm/cortex-a53.md: Update with new attributes.
8818 * config/arm/cortex-a7.md: Update with new attributes.
8819 * config/arm/cortex-a8-neon.md: Update with new attributes.
8820 * config/arm/cortex-a9.md: Update with new attributes.
8821 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8822 * config/arm/cortex-r4f.md: Update with new attributes.
8823 * config/arm/marvell-pj4.md: Update with new attributes.
8824 * config/arm/vfp11.md: Update with new attributes.
8825
8826 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8827
8828 * config/aarch64/aarch64.md
8829 (*madd<mode>): Fix type attribute.
8830 (*maddsi_uxtw): Likewise.
8831 (*msub<mode>): Likewise.
8832 (*msubsi_uxtw): Likewise.
8833 (<su_optab>maddsidi4): Likewise.
8834 (<su_optab>msubsidi4): Likewise.
8835
8836 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8837
8838 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
8839 * config/arm/arm.md (core_cycles): Remove fdiv.
8840 * config/arm/vfp.md:
8841 (*sqrtsf2_vfp): Update for attribute changes.
8842 (*sqrtdf2_vfp): Likewise.
8843 * config/aarch64/aarch64.md:
8844 (sqrt<mode>2): Update for attribute changes.
8845 * config/arm/arm1020e.md: Update with new attributes.
8846 * config/arm/cortex-a15-neon.md: Update with new attributes.
8847 * config/arm/cortex-a5.md: Update with new attributes.
8848 * config/arm/cortex-a53.md: Update with new attributes.
8849 * config/arm/cortex-a7.md: Update with new attributes.
8850 * config/arm/cortex-a8-neon.md: Update with new attributes.
8851 * config/arm/cortex-a9.md: Update with new attributes.
8852 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8853 * config/arm/cortex-r4f.md: Update with new attributes.
8854 * config/arm/marvell-pj4.md: Update with new attributes.
8855 * config/arm/vfp11.md: Update with new attributes.
8856
8857 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8858
8859 * config/arm/types.md
8860 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
8861 * config/aarch64/aarch64.md
8862 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
8863 new attributes.
8864 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
8865 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
8866 (float<GPI:mode><GPF:mode>2): Likewise.
8867 * config/arm/vfp.md
8868 (*truncsisf2_vfp): Update with new attributes.
8869 (*truncsidf2_vfp): Likewise.
8870 (fixuns_truncsfsi2): Likewise.
8871 (fixuns_truncdfsi2): Likewise.
8872 (*floatsisf2_vfp): Likewise.
8873 (*floatsidf2_vfp): Likewise.
8874 (floatunssisf2): Likewise.
8875 (floatunssidf2): Likewise.
8876 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
8877 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
8878 * config/arm/arm1020e.md: Update with new attributes.
8879 * config/arm/cortex-a15-neon.md: Update with new attributes.
8880 * config/arm/cortex-a5.md: Update with new attributes.
8881 * config/arm/cortex-a53.md: Update with new attributes.
8882 * config/arm/cortex-a7.md: Update with new attributes.
8883 * config/arm/cortex-a8-neon.md: Update with new attributes.
8884 * config/arm/cortex-a9.md: Update with new attributes.
8885 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8886 * config/arm/cortex-r4f.md: Update with new attributes.
8887 * config/arm/marvell-pj4.md: Update with new attributes.
8888 * config/arm/vfp11.md: Update with new attributes.
8889
8890 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8891
8892 * config/aarch64/arm_neon.h
8893 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
8894 (vqtbx<1,2,3,4><q>_s8): Likewise.
8895
8896 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8897
8898 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
8899 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
8900 (add<mode>3): Add type attribute.
8901 (add<mode>3): Likewise.
8902 (usadd<mode>3): Likewise.
8903 (ssadd<mode>3): Likewise.
8904 (sub<mode>3): Likewise.
8905 (sub<mode>3): Likewise.
8906 (ussub<mode>3): Likewise.
8907 (sssub<mode>3): Likewise.
8908 (ssmulsa3): Likewise.
8909 (usmulusa3): Likewise.
8910 (arm_usatsihi): Likewise.
8911 * config/arm/vfp.md
8912 (*movdi_vfp): Add types for all instructions.
8913 (*movdi_vfp_cortexa8): Likewise.
8914 (*movhf_vfp_neon): Likewise.
8915 (*movhf_vfp): Likewise.
8916 (*movdf_vfp): Likewise.
8917 (*thumb2_movdf_vfp): Likewise.
8918 (*thumb2_movdfcc_vfp): Likewise.
8919 * config/arm/arm.md: Add type attribute to all insn patterns.
8920 (*thumb1_adddi3): Add type attribute.
8921 (*arm_adddi3): Likewise.
8922 (*adddi_sesidi_di): Likewise.
8923 (*adddi_zesidi_di): Likewise.
8924 (*thumb1_addsi3): Likewise.
8925 (addsi3_compare0): Likewise.
8926 (*addsi3_compare0_scratch): Likewise.
8927 (*compare_negsi_si): Likewise.
8928 (cmpsi2_addneg): Likewise.
8929 (*addsi3_carryin_<optab>): Likewise.
8930 (*addsi3_carryin_alt2_<optab>): Likewise.
8931 (*addsi3_carryin_clobercc_<optab>): Likewise.
8932 (*subsi3_carryin): Likewise.
8933 (*subsi3_carryin_const): Likewise.
8934 (*subsi3_carryin_compare): Likewise.
8935 (*subsi3_carryin_compare_const): Likewise.
8936 (*arm_subdi3): Likewise.
8937 (*thumb_subdi3): Likewise.
8938 (*subdi_di_zesidi): Likewise.
8939 (*subdi_di_sesidi): Likewise.
8940 (*subdi_zesidi_di): Likewise.
8941 (*subdi_sesidi_di): Likewise.
8942 (*subdi_zesidi_ze): Likewise.
8943 (thumb1_subsi3_insn): Likewise.
8944 (*arm_subsi3_insn): Likewise.
8945 (*anddi3_insn): Likewise.
8946 (*anddi_zesidi_di): Likewise.
8947 (*anddi_sesdi_di): Likewise.
8948 (*ne_zeroextracts): Likewise.
8949 (*ne_zeroextracts): Likewise.
8950 (*ite_ne_zeroextr): Likewise.
8951 (*ite_ne_zeroextr): Likewise.
8952 (*anddi_notdi_di): Likewise.
8953 (*anddi_notzesidi): Likewise.
8954 (*anddi_notsesidi): Likewise.
8955 (andsi_notsi_si): Likewise.
8956 (thumb1_bicsi3): Likewise.
8957 (*iordi3_insn): Likewise.
8958 (*iordi_zesidi_di): Likewise.
8959 (*iordi_sesidi_di): Likewise.
8960 (*thumb1_iorsi3_insn): Likewise.
8961 (*xordi3_insn): Likewise.
8962 (*xordi_zesidi_di): Likewise.
8963 (*xordi_sesidi_di): Likewise.
8964 (*arm_xorsi3): Likewise.
8965 (*andsi_iorsi3_no): Likewise.
8966 (*smax_0): Likewise.
8967 (*smax_m1): Likewise.
8968 (*arm_smax_insn): Likewise.
8969 (*smin_0): Likewise.
8970 (*arm_smin_insn): Likewise.
8971 (*arm_umaxsi3): Likewise.
8972 (*arm_uminsi3): Likewise.
8973 (*minmax_arithsi): Likewise.
8974 (*minmax_arithsi_): Likewise.
8975 (*satsi_<SAT:code>): Likewise.
8976 (arm_ashldi3_1bit): Likewise.
8977 (arm_ashrdi3_1bit): Likewise.
8978 (arm_lshrdi3_1bit): Likewise.
8979 (*arm_negdi2): Likewise.
8980 (*thumb1_negdi2): Likewise.
8981 (*arm_negsi2): Likewise.
8982 (*thumb1_negsi2): Likewise.
8983 (*negdi_extendsid): Likewise.
8984 (*negdi_zero_extend): Likewise.
8985 (*arm_abssi2): Likewise.
8986 (*thumb1_abssi2): Likewise.
8987 (*arm_neg_abssi2): Likewise.
8988 (*thumb1_neg_abss): Likewise.
8989 (one_cmpldi2): Likewise.
8990 (extend<mode>di2): Likewise.
8991 (*compareqi_eq0): Likewise.
8992 (*arm_extendhisi2addsi): Likewise.
8993 (*arm_movdi): Likewise.
8994 (*thumb1_movdi_insn): Likewise.
8995 (*arm_movt): Likewise.
8996 (*thumb1_movsi_insn): Likewise.
8997 (pic_add_dot_plus_four): Likewise.
8998 (pic_add_dot_plus_eight): Likewise.
8999 (tls_load_dot_plus_eight): Likewise.
9000 (*thumb1_movhi_insn): Likewise.
9001 (*thumb1_movsf_insn): Likewise.
9002 (*movdf_soft_insn): Likewise.
9003 (*thumb_movdf_insn): Likewise.
9004 (cbranchsi4_insn): Likewise.
9005 (cbranchsi4_scratch): Likewise.
9006 (*negated_cbranchsi4): Likewise.
9007 (*tbit_cbranch): Likewise.
9008 (*tlobits_cbranch): Likewise.
9009 (*tstsi3_cbranch): Likewise.
9010 (*cbranchne_decr1): Likewise.
9011 (*addsi3_cbranch): Likewise.
9012 (*addsi3_cbranch_scratch): Likewise.
9013 (*arm_cmpdi_insn): Likewise.
9014 (*arm_cmpdi_unsig): Likewise.
9015 (*arm_cmpdi_zero): Likewise.
9016 (*thumb_cmpdi_zero): Likewise.
9017 (*deleted_compare): Likewise.
9018 (*mov_scc): Likewise.
9019 (*mov_negscc): Likewise.
9020 (*mov_notscc): Likewise.
9021 (*cstoresi_eq0_thumb1_insn): Likewise.
9022 (cstoresi_nltu_thumb1): Likewise.
9023 (cstoresi_ltu_thu): Likewise.
9024 (thumb1_addsi3_addgeu): Likewise.
9025 (*arm_jump): Likewise.
9026 (*thumb_jump): Likewise.
9027 (*check_arch2): Likewise.
9028 (arm_casesi_internal): Likewise.
9029 (thumb1_casesi_dispatch): Likewise.
9030 (*arm_indirect_jump): Likewise.
9031 (*thumb1_indirect_jump): Likewise.
9032 (nop): Likewise.
9033 (*and_scc): Likewise.
9034 (*ior_scc): Likewise.
9035 (*compare_scc): Likewise.
9036 (*cond_move): Likewise.
9037 (*cond_arith): Likewise.
9038 (*cond_sub): Likewise.
9039 (*cmp_ite0): Likewise.
9040 (*cmp_ite1): Likewise.
9041 (*cmp_and): Likewise.
9042 (*cmp_ior): Likewise.
9043 (*ior_scc_scc): Likewise.
9044 (*ior_scc_scc_cmp): Likewise.
9045 (*and_scc_scc): Likewise.
9046 (*and_scc_scc_cmp): Likewise.
9047 (*and_scc_scc_nod): Likewise.
9048 (*negscc): Likewise.
9049 (movcond_addsi): Likewise.
9050 (movcond): Likewise.
9051 (*ifcompare_plus_move): Likewise.
9052 (*if_plus_move): Likewise.
9053 (*ifcompare_move_plus): Likewise.
9054 (*if_move_plus): Likewise.
9055 (*ifcompare_arith_arith): Likewise.
9056 (*if_arith_arith): Likewise.
9057 (*ifcompare_arith_move): Likewise.
9058 (*if_arith_move): Likewise.
9059 (*ifcompare_move_arith): Likewise.
9060 (*if_move_arith): Likewise.
9061 (*ifcompare_move_not): Likewise.
9062 (*if_move_not): Likewise.
9063 (*ifcompare_not_move): Likewise.
9064 (*if_not_move): Likewise.
9065 (*ifcompare_shift_move): Likewise.
9066 (*if_shift_move): Likewise.
9067 (*ifcompare_move_shift): Likewise.
9068 (*if_move_shift): Likewise.
9069 (*ifcompare_shift_shift): Likewise.
9070 (*ifcompare_not_arith): Likewise.
9071 (*ifcompare_arith_not): Likewise.
9072 (*if_arith_not): Likewise.
9073 (*ifcompare_neg_move): Likewise.
9074 (*if_neg_move): Likewise.
9075 (*ifcompare_move_neg): Likewise.
9076 (*if_move_neg): Likewise.
9077 (prologue_thumb1_interwork): Likewise.
9078 (*cond_move_not): Likewise.
9079 (*sign_extract_onebit): Likewise.
9080 (*not_signextract_onebit): Likewise.
9081 (stack_tie): Likewise.
9082 (align_4): Likewise.
9083 (align_8): Likewise.
9084 (consttable_end): Likewise.
9085 (consttable_1): Likewise.
9086 (consttable_2): Likewise.
9087 (consttable_4): Likewise.
9088 (consttable_8): Likewise.
9089 (consttable_16): Likewise.
9090 (*thumb1_tablejump): Likewise.
9091 (prefetch): Likewise.
9092 (force_register_use): Likewise.
9093 (thumb_eh_return): Likewise.
9094 (load_tp_hard): Likewise.
9095 (load_tp_soft): Likewise.
9096 (tlscall): Likewise.
9097 (*arm_movtas_ze): Likewise.
9098 (*arm_rev): Likewise.
9099 (*arm_revsh): Likewise.
9100 (*arm_rev16): Likewise.
9101 * config/arm/thumb2.md
9102 (*thumb2_smaxsi3): Likewise.
9103 (*thumb2_sminsi3): Likewise.
9104 (*thumb32_umaxsi3): Likewise.
9105 (*thumb2_uminsi3): Likewise.
9106 (*thumb2_negdi2): Likewise.
9107 (*thumb2_abssi2): Likewise.
9108 (*thumb2_neg_abss): Likewise.
9109 (*thumb2_movsi_insn): Likewise.
9110 (tls_load_dot_plus_four): Likewise.
9111 (*thumb2_movhi_insn): Likewise.
9112 (*thumb2_mov_scc): Likewise.
9113 (*thumb2_mov_negs): Likewise.
9114 (*thumb2_mov_negs): Likewise.
9115 (*thumb2_mov_nots): Likewise.
9116 (*thumb2_mov_nots): Likewise.
9117 (*thumb2_movsicc_): Likewise.
9118 (*thumb2_movsfcc_soft_insn): Likewise.
9119 (*thumb2_indirect_jump): Likewise.
9120 (*thumb2_and_scc): Likewise.
9121 (*thumb2_ior_scc): Likewise.
9122 (*thumb2_ior_scc_strict_it): Likewise.
9123 (*thumb2_cond_move): Likewise.
9124 (*thumb2_cond_arith): Likewise.
9125 (*thumb2_cond_ari): Likewise.
9126 (*thumb2_cond_sub): Likewise.
9127 (*thumb2_negscc): Likewise.
9128 (*thumb2_movcond): Likewise.
9129 (thumb2_casesi_internal): Likewise.
9130 (thumb2_casesi_internal_pic): Likewise.
9131 (*thumb2_alusi3_short): Likewise.
9132 (*thumb2_mov<mode>_shortim): Likewise.
9133 (*thumb2_addsi_short): Likewise.
9134 (*thumb2_subsi_short): Likewise.
9135 (thumb2_addsi3_compare0): Likewise.
9136 (*thumb2_cbz): Likewise.
9137 (*thumb2_cbnz): Likewise.
9138 (*thumb2_one_cmplsi2_short): Likewise.
9139 (*thumb2_negsi2_short): Likewise.
9140 (*orsi_notsi_si): Likewise.
9141 * config/arm/arm1020e.md: Update with new attributes.
9142 * config/arm/arm1026ejs.md: Update with new attributes.
9143 * config/arm/arm1136jfs.md: Update with new attributes.
9144 * config/arm/arm926ejs.md: Update with new attributes.
9145 * config/arm/cortex-a15.md: Update with new attributes.
9146 * config/arm/cortex-a5.md: Update with new attributes.
9147 * config/arm/cortex-a53.md: Update with new attributes.
9148 * config/arm/cortex-a7.md: Update with new attributes.
9149 * config/arm/cortex-a8.md: Update with new attributes.
9150 * config/arm/cortex-a9.md: Update with new attributes.
9151 * config/arm/cortex-m4.md: Update with new attributes.
9152 * config/arm/cortex-r4.md: Update with new attributes.
9153 * config/arm/fa526.md: Update with new attributes.
9154 * config/arm/fa606te.md: Update with new attributes.
9155 * config/arm/fa626te.md: Update with new attributes.
9156 * config/arm/fa726te.md: Update with new attributes.
9157
9158 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9159
9160 * config/aarch64/aarch64-simd.md
9161 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
9162 <vwx> iterator to ensure correct register choice.
9163 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
9164 (aarch64_sqdmull_n<mode>): Likewise.
9165 (aarch64_sqdmull2_n<mode>_internal): Likewise.
9166 * config/aarch64/arm_neon.h
9167 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
9168 (vml<as><q>_n_<su>16): Likewise.
9169 (vml<as>l_high_lane<q>_<su>16): Likewise.
9170 (vml<as>l_high_n_<su>16): Likewise.
9171 (vml<as>l_lane<q>_<su>16): Likewise.
9172 (vml<as>l_n_<su>16): Likewise.
9173 (vmul<q>_lane<q>_<su>16): Likewise.
9174 (vmul<q>_n_<su>16): Likewise.
9175 (vmull_lane<q>_<su>16): Likewise.
9176 (vmull_n_<su>16): Likewise.
9177 (vmull_high_lane<q>_<su>16): Likewise.
9178 (vmull_high_n_<su>16): Likewise.
9179 (vqrdmulh<q>_n_s16): Likewise.
9180
9181 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
9182
9183 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
9184 have the correct lane parameter.
9185
9186 2013-09-06 Richard Biener <rguenther@suse.de>
9187
9188 * cfganal.c (control_dependences::~control_dependences):
9189 Properly free all of the vector.
9190
9191 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
9192
9193 PR target/58269
9194 * config/i386/i386.c (ix86_conditional_register_usage):
9195 Proper initialize extended SSE registers.
9196
9197 2013-09-06 Jan Hubicka <jh@suse.cz>
9198
9199 PR tree-optimization/58311
9200 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
9201
9202 2013-09-06 Jan Hubicka <jh@suse.cz>
9203
9204 * Makefile.in (tree-sra.o): Update dependencies.
9205 * tree-sra.c: Include ipa-utils.h
9206 (scan_function): Use recursive_call_p.
9207 (has_caller_p): New function.
9208 (cgraph_for_node_and_aliases): Count also callers of aliases.
9209
9210 2013-09-06 Jan Hubicka <jh@suse.cz>
9211
9212 PR middle-end/58094
9213 * cgraph.h (symtab_semantically_equivalent_p): Declare.
9214 * tree-tailcall.c: Include ipa-utils.h.
9215 (find_tail_calls): Use it.
9216 * ipa-pure-const.c (check_call): Likewise.
9217 * ipa-utils.c (recursive_call_p): New function.
9218 * ipa-utils.h (recursive_call_p): Dclare.
9219 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
9220 (symtab_semantically_equivalent_p): New function.
9221 * Makefile.in (tree-tailcall.o): Update dependencies.
9222
9223 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9224
9225 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
9226 non-inlinable part.
9227
9228 2013-09-06 Richard Biener <rguenther@suse.de>
9229
9230 * lto-streamer.h (lto_global_var_decls): Remove.
9231 * Makefile.in (OBJS): Remove lto-symtab.o.
9232 (lto-symtab.o): Remove.
9233 (GTFILES): Remove lto-symtab.c
9234 * lto-symtab.c: Move to lto/
9235
9236 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9237
9238 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
9239 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
9240 (UNSPEC_FPINT_RINT): New constant definitions.
9241 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
9242 definition with 2 attributes.
9243 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
9244 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
9245 definitions.
9246
9247 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9248
9249 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
9250 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
9251 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
9252 attribute to "z196_alone".
9253 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
9254 "zEC12_simple".
9255
9256 2013-09-06 Richard Biener <rguenther@suse.de>
9257
9258 * basic-block.h (class control_dependences): New.
9259 * tree-ssa-dce.c (control_dependence_map): Remove.
9260 (cd): New global.
9261 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
9262 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
9263 find_pdom, find_control_dependence, find_all_control_dependences):
9264 Move to cfganal.c.
9265 (mark_control_dependent_edges_necessary,
9266 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
9267 tree_dce_done, perform_tree_ssa_dce): Adjust.
9268 * cfganal.c (set_control_dependence_map_bit,
9269 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
9270 find_all_control_dependences): Move from tree-ssa-dce.c and
9271 implement as methods of control_dependences class.
9272 (control_dependences::control_dependences): New.
9273 (control_dependences::~control_dependences): Likewise.
9274 (control_dependences::get_edges_dependent_on): Likewise.
9275 (control_dependences::get_edge): Likewise.
9276
9277 2013-09-04 Jan Hubicka <jh@suse.cz>
9278
9279 * tree.c (types_same_for_odr): Drop overactive check.
9280 * ipa-devirt.c (hash_type_name): Likewise.
9281
9282 2013-09-04 Jan Hubicka <jh@suse.cz>
9283
9284 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
9285 (analyze_functions): ... here.
9286
9287 2013-09-04 Jan Hubicka <jh@suse.cz>
9288
9289 PR middle-end/58201
9290 * cgraphunit.c (analyze_functions): Clear AUX fields
9291 after processing; initialize assembler name has.
9292
9293 2013-09-05 Jeff Law <law@redhat.com>
9294
9295 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
9296 from thread_around_empty_block. Record threading path into PATH.
9297 Recurse if threading through the initial block is successful.
9298 (thread_across_edge): Corresponding changes to slightly simplify.
9299
9300 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9301
9302 * config/aarch64/aarch64.md
9303 (type): Remove frecpe, frecps, frecpx.
9304 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
9305 fix to be a TARGET_SIMD instruction.
9306 (aarch64_frecps): Remove.
9307 * config/aarch64/aarch64-simd.md
9308 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
9309 (aarch64_frecps<mode>): Handle all float/vector of float modes.
9310
9311 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9312 Sofiane Naci <sofiane.naci@arm.com>
9313
9314 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
9315 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
9316 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
9317 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
9318 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
9319 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
9320 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
9321 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
9322 into "clz, "rbit". Rename "shift" to "shift_imm".
9323 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
9324 changes. Update for attribute changes all occurrences of arlo_* and
9325 shift* types.
9326 * config/arm/arm-fixed.md: Update for attribute changes
9327 all occurrences of arlo_* types.
9328 * config/arm/thumb2.md: Update for attribute changes all occurrences
9329 of arlo_* types.
9330 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
9331 (cortexa7_older_only): Likewise.
9332 (cortexa7_younger): Likewise.
9333 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9334 (1020alu_shift_op): Likewise.
9335 (1020alu_shift_reg_op): Likewise.
9336 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9337 (alu_shift_op): Likewise.
9338 (alu_shift_reg_op): Likewise.
9339 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9340 (11_alu_shift_op): Likewise.
9341 (11_alu_shift_reg_op): Likewise.
9342 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9343 (9_alu_shift_reg_op): Likewise.
9344 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
9345 attribute changes.
9346 (cortex_a15_alu_shift): Likewise.
9347 (cortex_a15_alu_shift_reg): Likewise.
9348 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
9349 attribute changes.
9350 (cortex_a5_alu_shift): Likewise.
9351 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
9352 attribute changes.
9353 (cortex_a53_alu_shift): Likewise.
9354 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
9355 attribute changes.
9356 (cortex_a7_alu_reg): Likewise.
9357 (cortex_a7_alu_shift): Likewise.
9358 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
9359 attribute changes.
9360 (cortex_a8_alu_shift): Likewise.
9361 (cortex_a8_alu_shift_reg): Likewise.
9362 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9363 (cortex_a9_dp_shift): Likewise.
9364 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
9365 attribute changes.
9366 * config/arm/cortex-r4.md
9367 (cortex_r4_alu): Update for attribute changes.
9368 (cortex_r4_mov): Likewise.
9369 (cortex_r4_alu_shift_reg): Likewise.
9370 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9371 (526_alu_shift_op): Likewise.
9372 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9373 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9374 (626te_alu_shift_op): Likewise.
9375 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
9376 (726te_alu_shift_op): Likewise.
9377 (726te_alu_shift_reg_op): Likewise.
9378 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9379 (mp626_alu_shift_op): Likewise.
9380 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
9381 (pj4_alu_conds): Likewise.
9382 (pj4_shift): Likewise.
9383 (pj4_shift_conds): Likewise.
9384 (pj4_alu_shift): Likewise.
9385 (pj4_alu_shift_conds): Likewise.
9386 * config/aarch64/aarch64.md: Update for attribute change
9387 all occurrences of arlo_* and shift* types.
9388
9389 2013-09-05 Mike Stump <mikestump@comcast.net>
9390
9391 * tree.h: Move documentation for tree_function_decl to tree-core.h
9392 with the declaration.
9393
9394 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
9395
9396 PR target/58139
9397 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
9398 looking for widest mode.
9399
9400 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
9401
9402 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
9403
9404 2013-09-05 Richard Biener <rguenther@suse.de>
9405
9406 PR tree-optimization/58137
9407 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
9408 Do not create vectors of pointers.
9409 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
9410 types for the components of the vector initializer.
9411 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
9412 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
9413
9414 2013-09-05 Martin Jambor <mjambor@suse.cz>
9415
9416 * ipa-prop.c (remove_described_reference): Accept missing references,
9417 return false if that hppens, otherwise return true.
9418 (cgraph_node_for_jfunc): New function.
9419 (try_decrement_rdesc_refcount): Likewise.
9420 (try_make_edge_direct_simple_call): Use them.
9421 (ipa_edge_removal_hook): Remove references from rdescs.
9422 (ipa_edge_duplication_hook): Clone rdescs and their references
9423 when the new edge has the same caller as the old one.
9424 * cgraph.c (cgraph_resolve_speculation): Remove speculative
9425 reference before removing any edges.
9426
9427 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
9428
9429 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
9430 initial store.
9431 * thumb2.md (thumb2_storewb_parisi): New pattern.
9432
9433 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
9434
9435 * config/aarch64/aarch64-option-extensions.def: Add
9436 AARCH64_OPT_EXTENSION of 'crc'.
9437 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
9438 (AARCH64_ISA_CRC): Ditto.
9439 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
9440 description of the CRC extension.
9441
9442 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
9443
9444 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
9445 * config/rs6000/linux.h: Ditto.
9446 * alpha/linux.h: Ditto.
9447 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
9448 no_c99_libc_has_function.
9449 * config/c6x/uclinux-elf.h: Ditto.
9450 * config/lm32/uclinux-elf.h: Ditto.
9451 * config/m68k/uclinux.h: Ditto.
9452 * config/moxie/uclinux.h: Ditto.
9453 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
9454 (crisv32-*-linux*, cris-*-linux*): Ditto.
9455 * config/bfin/bfin.c: Include "tm_p.h".
9456
9457 2013-09-05 Richard Biener <rguenther@suse.de>
9458
9459 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
9460 check for a definition without a basic-block.
9461
9462 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9463 Sofiane Naci <sofiane.naci@arm.com>
9464
9465 * config/aarch64/aarch64.md
9466 (*movti_aarch64): Rename r_2_f and f_2_r.
9467 (*movsf_aarch64): Likewise.
9468 (*movdf_aarch64): Likewise.
9469 (*movtf_aarch64): Likewise.
9470 (aarch64_movdi_<mode>low): Likewise.
9471 (aarch64_movdi_<mode>high): Likewise.
9472 (aarch64_mov<mode>high_di): Likewise.
9473 (aarch64_mov<mode>low_di): Likewise.
9474 (aarch64_movtilow_tilow): Likewise.
9475 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
9476 values to config/arm/types.md
9477 (attribute "conds"): Update for attribute change.
9478 (anddi3_insn): Likewise.
9479 (iordi3_insn): Likewise.
9480 (xordi3_insn): Likewise.
9481 (one_cmpldi2): Likewise.
9482 * config/arm/types.md (type): Add Neon types.
9483 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
9484 use "type" attribute.
9485 (movmisalign<mode>_neon_store): Likewise.
9486 (movmisalign<mode>_neon_load): Likewise.
9487 (vec_set<mode>_internal): Likewise.
9488 (vec_setv2di_internal): Likewise.
9489 (vec_extract<mode>): Likewise.
9490 (vec_extractv2di): Likewise.
9491 (add<mode>3_neon): Likewise.
9492 (adddi3_neon): Likewise.
9493 (sub<mode>3_neon): Likewise.
9494 (subdi3_neon): Likewise.
9495 (mul<mode>3_neon): Likewise.
9496 (mul<mode>3add<mode>_neon): Likewise.
9497 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
9498 (fma<VCVTF:mode>4)): Likewise.
9499 (fma<VCVTF:mode>4_intrinsic): Likewise.
9500 (fmsub<VCVTF:mode>4)): Likewise.
9501 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
9502 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
9503 (ior<mode>3): Likewise.
9504 (and<mode>3): Likewise.
9505 (anddi3_neon): Likewise.
9506 (orn<mode>3_neon): Likewise.
9507 (orndi3_neon): Likewise.
9508 (bic<mode>3_neon): Likewise.
9509 (bicdi3_neon): Likewise.
9510 (xor<mode>3): Likewise.
9511 (one_cmpl<mode>2): Likewise.
9512 (abs<mode>2): Likewise.
9513 (neg<mode>2): Likewise.
9514 (umin<mode>3_neon): Likewise.
9515 (umax<mode>3_neon): Likewise.
9516 (smin<mode>3_neon): Likewise.
9517 (smax<mode>3_neon): Likewise.
9518 (vashl<mode>3): Likewise.
9519 (vashr<mode>3_imm): Likewise.
9520 (vlshr<mode>3_imm): Likewise.
9521 (ashl<mode>3_signed): Likewise.
9522 (ashl<mode>3_unsigned): Likewise.
9523 (neon_load_count): Likewise.
9524 (ashldi3_neon_noclobber): Likewise.
9525 (signed_shift_di3_neon): Likewise.
9526 (unsigned_shift_di3_neon): Likewise.
9527 (ashrdi3_neon_imm_noclobber): Likewise.
9528 (lshrdi3_neon_imm_noclobber): Likewise.
9529 (widen_ssum<mode>3): Likewise.
9530 (widen_usum<mode>3): Likewise.
9531 (quad_halves_<code>v4si): Likewise.
9532 (quad_halves_<code>v4sf): Likewise.
9533 (quad_halves_<code>v8hi): Likewise.
9534 (quad_halves_<code>v16qi): Likewise.
9535 (reduc_splus_v2di): Likewise.
9536 (neon_vpadd_internal<mode>): Likewise.
9537 (neon_vpsmin<mode>): Likewise.
9538 (neon_vpsmax<mode>): Likewise.
9539 (neon_vpumin<mode>): Likewise.
9540 (neon_vpumax<mode>): Likewise.
9541 (ss_add<mode>_neon): Likewise.
9542 (us_add<mode>_neon): Likewise.
9543 (ss_sub<mode>_neon): Likewise.
9544 (us_sub<mode>_neon): Likewise.
9545 (neon_vadd<mode>_unspec): Likewise.
9546 (neon_vaddl<mode>): Likewise.
9547 (neon_vaddw<mode>): Likewise.
9548 (neon_vhadd<mode>): Likewise.
9549 (neon_vqadd<mode>): Likewise.
9550 (neon_vaddhn<mode>): Likewise.
9551 (neon_vmul<mode>): Likewise.
9552 (neon_vmla<mode>): Likewise.
9553 (neon_vmlal<mode>): Likewise.
9554 (neon_vmls<mode>): Likewise.
9555 (neon_vmlsl<mode>): Likewise.
9556 (neon_vqdmulh<mode>): Likewise.
9557 (neon_vqdmlal<mode>): Likewise.
9558 (neon_vqdmlsl<mode>): Likewise.
9559 (neon_vmull<mode>): Likewise.
9560 (neon_vqdmull<mode>): Likewise.
9561 (neon_vsub<mode>_unspec): Likewise.
9562 (neon_vsubl<mode>): Likewise.
9563 (neon_vsubw<mode>): Likewise.
9564 (neon_vqsub<mode>): Likewise.
9565 (neon_vhsub<mode>): Likewise.
9566 (neon_vsubhn<mode>): Likewise.
9567 (neon_vceq<mode>): Likewise.
9568 (neon_vcge<mode>): Likewise.
9569 (neon_vcgeu<mode>): Likewise.
9570 (neon_vcgt<mode>): Likewise.
9571 (neon_vcgtu<mode>): Likewise.
9572 (neon_vcle<mode>): Likewise.
9573 (neon_vclt<mode>): Likewise.
9574 (neon_vcage<mode>): Likewise.
9575 (neon_vcagt<mode>): Likewise.
9576 (neon_vtst<mode>): Likewise.
9577 (neon_vabd<mode>): Likewise.
9578 (neon_vabdl<mode>): Likewise.
9579 (neon_vaba<mode>): Likewise.
9580 (neon_vabal<mode>): Likewise.
9581 (neon_vmax<mode>): Likewise.
9582 (neon_vmin<mode>): Likewise.
9583 (neon_vpaddl<mode>): Likewise.
9584 (neon_vpadal<mode>): Likewise.
9585 (neon_vpmax<mode>): Likewise.
9586 (neon_vpmin<mode>): Likewise.
9587 (neon_vrecps<mode>): Likewise.
9588 (neon_vrsqrts<mode>): Likewise.
9589 (neon_vqabs<mode>): Likewise.
9590 (neon_vqneg<mode>): Likewise.
9591 (neon_vcls<mode>): Likewise.
9592 (clz<mode>2): Likewise.
9593 (popcount<mode>2): Likewise.
9594 (neon_vrecpe): Likewise.
9595 (neon_vrsqrte): Likewise.
9596 (neon_vget_lane<mode>_sext_internal): Likewise.
9597 (neon_vget_lane<mode>_zext_internal): Likewise.
9598 (neon_vdup_n<mode>): Likewise.
9599 (neon_vdup_nv2di): Likewise.
9600 (neon_vdpu_lane<mode>_internal): Likewise.
9601 (neon_vswp<mode>): Likewise.
9602 (float<mode><V_cvtto>2): Likewise.
9603 (floatuns<mode><V_cvtto>2): Likewise.
9604 (fix_trunc<mode><V_cvtto>)2): Likewise
9605 (fixuns_trunc<mode><V_cvtto)2): Likewise.
9606 (neon_vcvt<mode>): Likewise.
9607 (neon_vcvtv4sfv4hf): Likewise.
9608 (neon_vcvtv4hfv4sf): Likewise.
9609 (neon_vcvt_n<mode>): Likewise.
9610 (neon_vmovn<mode>): Likewise.
9611 (neon_vqmovn<mode>): Likewise.
9612 (neon_vqmovun<mode>): Likewise.
9613 (neon_vmovl<mode>): Likewise.
9614 (neon_vmul_lane<mode>): Likewise.
9615 (neon_vmull_lane<mode>): Likewise.
9616 (neon_vqdmull_lane<mode>): Likewise.
9617 (neon_vqdmulh_lane<mode>): Likewise.
9618 (neon_vmla_lane<mode>): Likewise.
9619 (neon_vmlal_lane<mode>): Likewise.
9620 (neon_vqdmlal_lane<mode>): Likewise.
9621 (neon_vmls_lane<mode>): Likewise.
9622 (neon_vmlsl_lane<mode>): Likewise.
9623 (neon_vqdmlsl_lane<mode>): Likewise.
9624 (neon_vext<mode>): Likewise.
9625 (neon_vrev64<mode>): Likewise.
9626 (neon_vrev32<mode>): Likewise.
9627 (neon_vrev16<mode>): Likewise.
9628 (neon_vbsl<mode>_internal): Likewise.
9629 (neon_vshl<mode>): Likewise.
9630 (neon_vqshl<mode>): Likewise.
9631 (neon_vshr_n<mode>): Likewise.
9632 (neon_vshrn_n<mode>): Likewise.
9633 (neon_vqshrn_n<mode>): Likewise.
9634 (neon_vqshrun_n<mode>): Likewise.
9635 (neon_vshl_n<mode>): Likewise.
9636 (neon_vqshl_n<mode>): Likewise.
9637 (neon_vqshlu_n<mode>): Likewise.
9638 (neon_vshll_n<mode>): Likewise.
9639 (neon_vsra_n<mode>): Likewise.
9640 (neon_vsri_n<mode>): Likewise.
9641 (neon_vsli_n<mode>): Likewise.
9642 (neon_vtbl1v8qi): Likewise.
9643 (neon_vtbl2v8qi): Likewise.
9644 (neon_vtbl3v8qi): Likewise.
9645 (neon_vtbl4v8qi): Likewise.
9646 (neon_vtbx1v8qi): Likewise.
9647 (neon_vtbx2v8qi): Likewise.
9648 (neon_vtbx3v8qi): Likewise.
9649 (neon_vtbx4v8qi): Likewise.
9650 (neon_vtrn<mode>_internal): Likewise.
9651 (neon_vzip<mode>_internal): Likewise.
9652 (neon_vuzp<mode>_internal): Likewise.
9653 (neon_vld1<mode>): Likewise.
9654 (neon_vld1_lane<mode>): Likewise.
9655 (neon_vld1_dup<mode>): Likewise.
9656 (neon_vld1_dupv2di): Likewise.
9657 (neon_vst1<mode>): Likewise.
9658 (neon_vst1_lane<mode>): Likewise.
9659 (neon_vld2<mode>): Likewise.
9660 (neon_vld2_lane<mode>): Likewise.
9661 (neon_vld2_dup<mode>): Likewise.
9662 (neon_vst2<mode>): Likewise.
9663 (neon_vst2_lane<mode>): Likewise.
9664 (neon_vld3<mode>): Likewise.
9665 (neon_vld3qa<mode>): Likewise.
9666 (neon_vld3qb<mode>): Likewise.
9667 (neon_vld3_lane<mode>): Likewise.
9668 (neon_vld3_dup<mode>): Likewise.
9669 (neon_vst3<mode>): Likewise.
9670 (neon_vst3qa<mode>): Likewise.
9671 (neon_vst3qb<mode>): Likewise.
9672 (neon_vst3_lane<mode>): Likewise.
9673 (neon_vld4<mode>): Likewise.
9674 (neon_vld4qa<mode>): Likewise.
9675 (neon_vld4qb<mode>): Likewise.
9676 (neon_vld4_lane<mode>): Likewise.
9677 (neon_vld4_dup<mode>): Likewise.
9678 (neon_vst4<mode>): Likewise.
9679 (neon_vst4qa<mode>): Likewise.
9680 (neon_vst4qb<mode>): Likewise.
9681 (neon_vst4_lane<mode>): Likewise.
9682 (neon_vec_unpack<US>_lo_<mode>): Likewise.
9683 (neon_vec_unpack<US>_hi_<mode>): Likewise.
9684 (neon_vec_<US>mult_lo_<mode>): Likewise.
9685 (neon_vec_<US>mult_hi_<mode>): Likewise.
9686 (neon_vec_<US>shiftl_<mode>): Likewise.
9687 (neon_unpack<US>_<mode>): Likewise.
9688 (neon_vec_<US>mult_<mode>): Likewise.
9689 (vec_pack_trunc_<mode>): Likewise.
9690 (neon_vec_pack_trunk_<mode>): Likewise.
9691 (neon_vabd<mode>_2): Likewise.
9692 (neon_vabd<mode>_3): Likewise.
9693 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
9694 (thumb2_movsi_vfp): Likewise.
9695 (movdi_vfp): Likewise.
9696 (movdi_vfp_cortexa8): Likewise.
9697 (movhf_vfp_neon): Likewise.
9698 (movhf_vfp): Likewiwse.
9699 (movsf_vfp): Likewiwse.
9700 (thumb2_movsf_vfp): Likewiwse.
9701 (movdf_vfp): Likewise.
9702 (thumb2_movdf_vfp): Likewise.
9703 (movsfcc_vfp): Likewise.
9704 (thumb2_movsfcc_vfp): Likewise.
9705 (movdfcc_vfp): Likewise.
9706 (thumb2_movdfcc_vfp): Likewise.
9707 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
9708 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
9709 (v10_v2c): Likewise.
9710 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
9711 attribute change.
9712 (cortex_a15_neon_int_2): Likewise.
9713 (cortex_a15_neon_int_3): Likewise.
9714 (cortex_a15_neon_int_4): Likewise.
9715 (cortex_a15_neon_int_5): Likewise.
9716 (cortex_a15_neon_vqneg_vqabs): Likewise.
9717 (cortex_a15_neon_vmov): Likewise.
9718 (cortex_a15_neon_vaba): Likewise.
9719 (cortex_a15_neon_vaba_qqq): Likewise.
9720 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9721 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9722 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
9723 scalar_64_32_long_scalar): Likewise.
9724 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9725 (cortex_a15_neon_mla_qqq_8_16): Likewise.
9726 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9727 lotype_qdd_64_32_long): Likewise.
9728 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9729 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9730 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
9731 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9732 (cortex_a15_neon_shift_1): Likewise.
9733 (cortex_a15_neon_shift_2): Likewise.
9734 (cortex_a15_neon_shift_3): Likewise.
9735 (cortex_a15_neon_vshl_ddd): Likewise.
9736 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9737 (cortex_a15_neon_vsra_vrsra): Likewise.
9738 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
9739 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
9740 (cortex_a15_neon_fp_vmul_ddd): Likewise.
9741 (cortex_a15_neon_fp_vmul_qqd): Likewise.
9742 (cortex_a15_neon_fp_vmla_ddd): Likewise.
9743 (cortex_a15_neon_fp_vmla_qqq): Likewise.
9744 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
9745 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
9746 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9747 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9748 (cortex_a15_neon_bp_simple): Likewise.
9749 (cortex_a15_neon_bp_2cycle): Likewise.
9750 (cortex_a15_neon_bp_3cycle): Likewise.
9751 (cortex_a15_neon_vld1_1_2_regs): Likewise.
9752 (cortex_a15_neon_vld1_3_4_regs): Likewise.
9753 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9754 (cortex_a15_neon_vld2_4_regs): Likewise.
9755 (cortex_a15_neon_vld3_vld4): Likewise.
9756 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9757 (cortex_a15_neon_vst1_3_4_regs): Likewise.
9758 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
9759 (cortex_a15_neon_vst3_vst4): Likewise.
9760 (cortex_a15_neon_vld1_vld2_lane): Likewise.
9761 (cortex_a15_neon_vld3_vld4_lane" 10
9762 (cortex_a15_neon_vst1_vst2_lane): Likewise.
9763 (cortex_a15_neon_vst3_vst4_lane): Likewise.
9764 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
9765 (cortex_a15_neon_ldm_2): Likewise.0
9766 (cortex_a15_neon_stm_2): Likewise.
9767 (cortex_a15_neon_mcr): Likewise.
9768 (cortex_a15_neon_mcr_2_mcrr): Likewise.
9769 (cortex_a15_neon_mrc): Likewise.
9770 (cortex_a15_neon_mrrc): Likewise.
9771 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9772 change.
9773 (cortex_a15_alu_shift): Likewise.
9774 (cortex_a15_alu_shift_reg): Likewise.
9775 (cortex_a15_mult32): Likewise.
9776 (cortex_a15_mult64): Likewise.
9777 (cortex_a15_block): Likewise.
9778 (cortex_a15_branch): Likewise.
9779 (cortex_a15_load1): Likewise.
9780 (cortex_a15_load3): Likewise.
9781 (cortex_a15_store1): Likewise.
9782 (cortex_a15_store3): Likewise.
9783 (cortex_a15_call): Likewise.
9784 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
9785 (cortex_a5_f2r): Likewise.
9786 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
9787 change.
9788 (cortex_a53_f2r): Likewise.
9789 * config/arm/cortex-a7.md
9790 (cortex_a7_branch): Update for attribute change.
9791 (cortex_a7_call): Likewise.
9792 (cortex_a7_alu_imm): Likewise.
9793 (cortex_a7_alu_reg): Likewise.
9794 (cortex_a7_alu_shift): Likewise.
9795 (cortex_a7_mul): Likewise.
9796 (cortex_a7_load1): Likewise.
9797 (cortex_a7_store1): Likewise.
9798 (cortex_a7_load2): Likewise.
9799 (cortex_a7_store2): Likewise.
9800 (cortex_a7_load3): Likewise.
9801 (cortex_a7_store3): Likewise.
9802 (cortex_a7_load4): Likewise.
9803 (cortex_a7_store4): Likewise.
9804 (cortex_a7_fpalu): Likewise.
9805 (cortex_a7_fconst): Likewise.
9806 (cortex_a7_fpmuls): Likewise.
9807 (cortex_a7_neon_mul): Likewise.
9808 (cortex_a7_fpmacs): Likewise.
9809 (cortex_a7_neon_mla: Likewise.
9810 (cortex_a7_fpmuld: Likewise.
9811 (cortex_a7_fpmacd: Likewise.
9812 (cortex_a7_fpfmad: Likewise.
9813 (cortex_a7_fdivs: Likewise.
9814 (cortex_a7_fdivd: Likewise.
9815 (cortex_a7_r2f: Likewise.
9816 (cortex_a7_f2r: Likewise.
9817 (cortex_a7_f_flags: Likewise.
9818 (cortex_a7_f_loads: Likewise.
9819 (cortex_a7_f_loadd: Likewise.
9820 (cortex_a7_f_stores: Likewise.
9821 (cortex_a7_f_stored: Likewise.
9822 (cortex_a7_neon): Likewise.
9823 * config/arm/cortex-a8-neon.md
9824 (cortex_a8_neon_mrc): Update for attribute change.
9825 (cortex_a8_neon_mrrc): Likewise.
9826 (cortex_a8_neon_int_1): Likewise.
9827 (cortex_a8_neon_int_2): Likewise.
9828 (cortex_a8_neon_int_3): Likewise.
9829 (cortex_a8_neon_int_4): Likewise.
9830 (cortex_a8_neon_int_5): Likewise.
9831 (cortex_a8_neon_vqneg_vqabs): Likewise.
9832 (cortex_a8_neon_vmov): Likewise.
9833 (cortex_a8_neon_vaba): Likewise.
9834 (cortex_a8_neon_vaba_qqq): Likewise.
9835 (cortex_a8_neon_vsma): Likewise.
9836 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9837 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9838 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
9839 long_scalar): Likewise.
9840 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9841 (cortex_a8_neon_mla_qqq_8_16): Likewise.
9842 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9843 long_scalar_qdd_64_32_long): Likewise.
9844 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9845 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9846 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
9847 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9848 (cortex_a8_neon_shift_1): Likewise.
9849 (cortex_a8_neon_shift_2): Likewise.
9850 (cortex_a8_neon_shift_3): Likewise.
9851 (cortex_a8_neon_vshl_ddd): Likewise.
9852 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9853 (cortex_a8_neon_vsra_vrsra): Likewise.
9854 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
9855 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
9856 (cortex_a8_neon_fp_vsum): Likewise.
9857 (cortex_a8_neon_fp_vmul_ddd): Likewise.
9858 (cortex_a8_neon_fp_vmul_qqd): Likewise.
9859 (cortex_a8_neon_fp_vmla_ddd): Likewise.
9860 (cortex_a8_neon_fp_vmla_qqq): Likewise.
9861 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
9862 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
9863 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9864 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9865 (cortex_a8_neon_bp_simple): Likewise.
9866 (cortex_a8_neon_bp_2cycle): Likewise.
9867 (cortex_a8_neon_bp_3cycle): Likewise.
9868 (cortex_a8_neon_ldr): Likewise.
9869 (cortex_a8_neon_str): Likewise.
9870 (cortex_a8_neon_vld1_1_2_regs): Likewise.
9871 (cortex_a8_neon_vld1_3_4_regs): Likewise.
9872 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9873 (cortex_a8_neon_vld2_4_regs): Likewise.
9874 (cortex_a8_neon_vld3_vld4): Likewise.
9875 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9876 (cortex_a8_neon_vst1_3_4_regs): Likewise.
9877 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
9878 (cortex_a8_neon_vst3_vst4): Likewise.
9879 (cortex_a8_neon_vld1_vld2_lane): Likewise.
9880 (cortex_a8_neon_vld3_vld4_lane): Likewise.
9881 (cortex_a8_neon_vst1_vst2_lane): Likewise.
9882 (cortex_a8_neon_vst3_vst4_lane): Likewise.
9883 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
9884 (cortex_a8_neon_mcr): Likewise.
9885 (cortex_a8_neon_mcr_2_mcrr): Likewise.
9886 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
9887 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
9888 change.
9889 (ca9_neon_mrrc): Likewise.
9890 (cortex_a9_neon_int_1): Likewise.
9891 (cortex_a9_neon_int_2): Likewise.
9892 (cortex_a9_neon_int_3): Likewise.
9893 (cortex_a9_neon_int_4): Likewise.
9894 (cortex_a9_neon_int_5): Likewise.
9895 (cortex_a9_neon_vqneg_vqabs): Likewise.
9896 (cortex_a9_neon_vmov): Likewise.
9897 (cortex_a9_neon_vaba): Likewise.
9898 (cortex_a9_neon_vaba_qqq): Likewise.
9899 (cortex_a9_neon_vsma): Likewise.
9900 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9901 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9902 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
9903 long_scalar): Likewise.
9904 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9905 (cortex_a9_neon_mla_qqq_8_16): Likewise.
9906 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9907 long_scalar_qdd_64_32_long): Likewise.
9908 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9909 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9910 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
9911 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9912 (cortex_a9_neon_shift_1): Likewise.
9913 (cortex_a9_neon_shift_2): Likewise.
9914 (cortex_a9_neon_shift_3): Likewise.
9915 (cortex_a9_neon_vshl_ddd): Likewise.
9916 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9917 (cortex_a9_neon_vsra_vrsra): Likewise.
9918 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
9919 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
9920 (cortex_a9_neon_fp_vsum): Likewise.
9921 (cortex_a9_neon_fp_vmul_ddd): Likewise.
9922 (cortex_a9_neon_fp_vmul_qqd): Likewise.
9923 (cortex_a9_neon_fp_vmla_ddd): Likewise.
9924 (cortex_a9_neon_fp_vmla_qqq): Likewise.
9925 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
9926 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
9927 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9928 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9929 (cortex_a9_neon_bp_simple): Likewise.
9930 (cortex_a9_neon_bp_2cycle): Likewise.
9931 (cortex_a9_neon_bp_3cycle): Likewise.
9932 (cortex_a9_neon_ldr): Likewise.
9933 (cortex_a9_neon_str): Likewise.
9934 (cortex_a9_neon_vld1_1_2_regs): Likewise.
9935 (cortex_a9_neon_vld1_3_4_regs): Likewise.
9936 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9937 (cortex_a9_neon_vld2_4_regs): Likewise.
9938 (cortex_a9_neon_vld3_vld4): Likewise.
9939 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9940 (cortex_a9_neon_vst1_3_4_regs): Likewise.
9941 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
9942 (cortex_a9_neon_vst3_vst4): Likewise.
9943 (cortex_a9_neon_vld1_vld2_lane): Likewise.
9944 (cortex_a9_neon_vld3_vld4_lane): Likewise.
9945 (cortex_a9_neon_vst1_vst2_lane): Likewise.
9946 (cortex_a9_neon_vst3_vst4_lane): Likewise.
9947 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
9948 (cortex_a9_neon_mcr): Likewise.
9949 (cortex_a9_neon_mcr_2_mcrr): Likewise.
9950 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
9951 (cortex_a9_fps): Likewise.
9952 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
9953 change.
9954 (cortex_m4_fmuls): Likewise.
9955 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
9956 change.
9957 (cortex_r4_mrc): Likewise.
9958 * config/arm/iterators.md: Update comment referring to neon_type.
9959 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
9960 (iwmmxt_movsi_insn): Likewise.
9961 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
9962 attribute change.
9963 (pj4_core_to_vfp): Likewise.
9964 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
9965 attribute change.
9966 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
9967 (vfp_fstore): Likewise.
9968 * doc/md.texi: Change references to neon_type to refer to type.
9969
9970 2013-09-04 Dodji Seketeli <dodji@redhat.com>
9971
9972 * tree.h (DECL_BUILT_IN): Fix typo in comment.
9973
9974 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
9975
9976 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
9977 lglobl if not weak.
9978
9979 2013-09-04 Easwaran Raman <eraman@google.com>
9980
9981 PR middle-end/57370
9982 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
9983 (build_and_add_sum): Use it.
9984 (appears_later_in_bb): Simplify code.
9985
9986 2013-09-04 Teresa Johnson <tejohnson@google.com>
9987
9988 * dumpfile.c (dump_finish): Don't close stderr/stdout.
9989
9990 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
9991
9992 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
9993
9994 2013-09-04 Jan Hubicka <jh@suse.cz>
9995
9996 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
9997 * ipa-devirt.c: Include diganostic.h
9998 (odr_type_d): Add types and types_set.
9999 (hash_type_name): Work for types with vtables during LTO.
10000 (odr_hasher::remove): Fix comment; destroy types_set.
10001 (add_type_duplicate): New function,
10002 (get_odr_type): Use it.
10003 (dump_type_inheritance_graph): Dump type duplicates.
10004 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
10005 graph.
10006 * tree.c (types_same_for_odr): Give exact answers on types with
10007 virtual tables.
10008
10009 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10010
10011 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
10012 explaining their differences.
10013
10014 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
10015
10016 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
10017
10018 2013-09-03 Jeff Law <law@redhat.com>
10019
10020 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
10021 when not threading through a joiner block. Pass joiner/no joiner
10022 state to register_jump_thread.
10023 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
10024 state from argument rather than implying on path length.
10025 Dump the entire jump thread path into debugging dump.
10026 * tree-flow.h (register_jump_thread): Update prototype.
10027
10028 2013-08-29 Xinliang David Li <davidxl@google.com>
10029
10030 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10031 Remove a trivial gcc_assert.
10032
10033 2013-08-29 Xinliang David Li <davidxl@google.com>
10034
10035 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
10036 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
10037 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10038 Delay base decl alignment adjustment.
10039 * tree-vectorizer.c (vect_destroy_datarefs): New function.
10040 * tree-vectorizer.h: New data structure.
10041 (set_dr_misalignment): New function.
10042 (dr_misalignment): Ditto.
10043 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
10044 (vectorizable_load): Ditto.
10045 (ensure_base_align): New function.
10046 (vectorize_loops): Add dbg_cnt support.
10047 (execute_vect_slp): Ditto.
10048 * dbgcnt.def: New debug counter.
10049 * Makefile: New dependency.
10050
10051 2013-09-03 Meador Inge <meadori@codesourcery.com>
10052
10053 Revert:
10054
10055 2013-08-30 Meador Inge <meadori@codesourcery.com>
10056
10057 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10058
10059 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
10060
10061 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
10062 function descriptor.
10063
10064 2013-09-03 Richard Biener <rguenther@suse.de>
10065
10066 * tree-affine.c (add_elt_to_tree): Fix association issue,
10067 avoid useless converts and make sure to always return a
10068 properly typed result.
10069
10070 2013-09-03 Richard Biener <rguenther@suse.de>
10071
10072 PR middle-end/57656
10073 * fold-const.c (negate_expr_p): Fix division case.
10074 (negate_expr): Likewise.
10075
10076 2013-09-03 Richard Biener <rguenther@suse.de>
10077
10078 PR lto/58285
10079 * tree-streamer-out.c: Include tm.h.
10080 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
10081
10082 2013-09-03 Jan Hubicka <jh@suse.cz>
10083
10084 * tree-profile.c (tree_profiling): Cleanup CFG when done.
10085
10086 2013-09-03 Alan Modra <amodra@gmail.com>
10087
10088 * config.gcc (powerpc*-*-linux*): Add support for little-endian
10089 multilibs to big-endian target and vice versa.
10090 * config/rs6000/t-linux64: Use := assignment on all vars.
10091 (MULTILIB_EXTRA_OPTS): Remove fPIC.
10092 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
10093 * config/rs6000/t-linux64le: New file.
10094 * config/rs6000/t-linux64bele: New file.
10095 * config/rs6000/t-linux64lebe: New file.
10096
10097 2013-09-02 Jan Hubicka <jh@suse.cz>
10098
10099 * ipa-inline-transform.c (inline_transform): Do not
10100 optimize_inline_calls when not optimizing.
10101
10102 2013-09-02 Jan Hubicka <jh@suse.cz>
10103
10104 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
10105 duplicated nodes for assembler names.
10106 * symtab.c (symtab_unregister_node): Do not attempt to unlink
10107 hard registers from assembler name hash.
10108
10109 2013-09-02 Jan Hubicka <jh@suse.cz>
10110
10111 * ipa-split.c (execute_split_functions): Split externally visible
10112 functions called once.
10113
10114 2013-09-02 Martin Jambor <mjambor@suse.cz>
10115
10116 PR ipa/58106
10117 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
10118 linked list. When finding the correct duplicate, also consider also
10119 the caller in additon to its inlined_to node.
10120
10121 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
10122
10123 * config/aarch64/aarch64-simd-builtins.def
10124 (dup_lane_scalar): Remove.
10125 * config/aarch64/aarch64-simd.md
10126 (aarch64_simd_dup): Add 'w->w' alternative.
10127 (aarch64_dup_lane<mode>): Allow for VALL.
10128 (aarch64_dup_lane_scalar<mode>): Remove.
10129 (aarch64_dup_lane_<vswap_width_name><mode>): New.
10130 (aarch64_get_lane_signed<mode>): Add w->w altenative.
10131 (aarch64_get_lane_unsigned<mode>): Likewise.
10132 (aarch64_get_lane<mode>): Likewise.
10133 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
10134 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
10135 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
10136 (VCON): Change container of V2SF.
10137 (vswap_width_name): Likewise.
10138 * config/aarch64/arm_neon.h
10139 (__aarch64_vdup_lane_any): New.
10140 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10141 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
10142 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10143
10144 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
10145
10146 PR middle-end/56382
10147 * expr.c (emit_move_complex): Do not move complex FP values as parts if
10148 the source or the destination is a single hard register.
10149
10150 2013-09-02 Richard Biener <rguenther@suse.de>
10151
10152 PR middle-end/57511
10153 * tree-scalar-evolution.c (instantiate_scev_name): Allow
10154 non-linear SCEVs.
10155
10156 2013-09-02 Richard Biener <rguenther@suse.de>
10157
10158 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
10159 arithmetic to sizetype.
10160
10161 2013-09-02 Bin Cheng <bin.cheng@arm.com>
10162
10163 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
10164 Find auto-increment use both before and after candidate.
10165
10166 2013-09-02 Marek Polacek <polacek@redhat.com>
10167
10168 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
10169
10170 2013-09-01 Jan Hubicka <jh@suse.cz>
10171
10172 * Makefile.in: Add ipa-profile.o
10173 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
10174 * cgraph.c (struct cgraph_propagate_frequency_data,
10175 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
10176 ipa-profile.c; replace cgraph_ by ipa_ prefix.
10177 * cgraph.h (cgraph_propagate_frequency): Remove.
10178 * ipa-inline-analysis.c: Include ipa-utils.h;
10179 drop duplicated cfgloop.h.
10180 (inline_update_callee_summaries): Update.
10181 * ipa-profile.c: New file.
10182 * ipa-utils.h (ipa_propagate_frequency): Declare.
10183 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
10184 data-streamer.h, value-prof.h.
10185 (symtab_remove_unreachable_nodes): Update profile.
10186 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
10187 account_time_size, cmp_counts, dump_histogram,
10188 ipa_profile_generate_summary, ipa_profile_write_summary,
10189 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
10190 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
10191 Move to ipa-profile.c.
10192
10193 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
10194
10195 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
10196
10197 2013-09-01 Jan Hubicka <jh@suse.cz>
10198
10199 * common.opt (fdevirtualize-speculatively): New function.
10200 * invoke.texi (fdevirtualize-speculatively): Document.
10201 * ipa-devirt.c: Include ipa-inline.h
10202 (likely_target_p): New function.
10203 (ipa_devirt): New function.
10204 (gate_ipa_devirt): New function.
10205 (pass_data_ipa_devirt): New static var.
10206 (pass_ipa_devirt): Likewise.
10207 (make_pass_ipa_devirt): New function.
10208 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
10209 (common_handle_option): Disable devirtualization when
10210 value range profiling is available.
10211 * passes.def (pass_ipa_devirt): Add.
10212 * timever.def (TV_IPA_DEVIRT): New timevar.
10213 * tree-pass.h (make_pass_ipa_devirt):
10214
10215 2013-09-01 Iain Sandoe <iain@codesourcery.com>
10216
10217 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
10218 include sanitize(undefined).
10219
10220 2013-08-31 Diego Novillo <dnovillo@google.com>
10221
10222 * Makefile.in (TREE_CORE_H): Define.
10223 (TREE_H): Use.
10224 (GTFILES): Add tree-core.h.
10225 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
10226 size the array.
10227 * tree-core.h: New file.
10228 Move all data structures, enum, typedefs, global
10229 declarations and constants from ...
10230 * tree.h: ... here.
10231
10232 2013-08-31 Jan Hubicka <jh@suse.cz>
10233
10234 * bulitins.c (expand_builtin): Do not early exit for gcov
10235 instrumented functions.
10236
10237 2013-08-31 Marek Polacek <polacek@redhat.com>
10238
10239 * ubsan.c: Include tm_p.h.
10240
10241 2013-08-31 Jan Hubicka <jh@suse.cz>
10242
10243 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
10244 warning.
10245
10246 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
10247 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
10248 * tree-cfg.c (verify_gimple_label): ... here.
10249 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
10250 (ipa_merge_profiles): New function.
10251 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
10252 (lto_input_function_body): Likewise.
10253 * ipa-utils.h (ipa_merge_profiles): Declare.
10254 * lto-streamer.h (lto_input_function_body): Update prototype.
10255 (emit_label_in_global_context_p): Remove.
10256 * lto-symtab.c: Include ipa-utils.h
10257 (lto_cgraph_replace_node): Use ipa_merge_profiles.
10258
10259 2013-08-31 Jan Hubicka <jh@suse.cz>
10260
10261 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
10262
10263 2013-08-31 Jan Hubicka <jh@suse.cz>
10264
10265 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
10266
10267 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
10268
10269 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
10270 "cmp" RTX before signed_comparison_operator check to account
10271 for "code" changes.
10272
10273 2013-08-30 Jan Hubicka <jh@suse.cz>
10274
10275 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
10276 (detect_type_change_1): Rename to ...
10277 (detect_type_change): ... this one; early return on non-polymorphic
10278 types.
10279 (detect_type_change_ssa): Add comp_type parameter; update
10280 use of detect_type_change.
10281 (compute_complex_assign_jump_func): Add param_type parameter;
10282 update use of detect_type_change_ssa.
10283 (compute_complex_ancestor_jump_func): Likewise.
10284 (ipa_get_callee_param_type): New function.
10285 (ipa_compute_jump_functions_for_edge): Compute parameter type;
10286 update calls to the jump function computation functions.
10287
10288 2013-08-30 Teresa Johnson <tejohnson@google.com>
10289 Steven Bosscher <steven@gcc.gnu.org>
10290
10291 * cfgrtl.c (fixup_new_cold_bb): New routine.
10292 (commit_edge_insertions): Invoke fixup_partitions.
10293 (find_partition_fixes): New routine.
10294 (fixup_partitions): Ditto.
10295 (verify_hot_cold_block_grouping): Update comments.
10296 (rtl_verify_edges): Invoke find_partition_fixes.
10297 (rtl_verify_bb_pointers): Update comments.
10298 (rtl_verify_bb_layout): Ditto.
10299 * basic-block.h (probably_never_executed_edge_p): Declare.
10300 (fixup_partitions): Ditto.
10301 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
10302 * bb-reorder.c (sanitize_hot_paths): New function.
10303 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
10304 sanitize_hot_paths.
10305 * predict.c (probably_never_executed_edge_p): New routine.
10306 * cfg.c (check_bb_profile): Add partition insanity warnings.
10307
10308 2013-08-30 Meador Inge <meadori@codesourcery.com>
10309
10310 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10311
10312 2013-08-30 Marek Polacek <polacek@redhat.com>
10313
10314 * Makefile.in (ubsan.o): Add.
10315 (c-family/c-ubsan.o): Add.
10316 (builtins.o): Add ubsan.h dependency.
10317 * ubsan.h: New file.
10318 * ubsan.c: New file.
10319 * common.opt: Add -fsanitize=undefined option.
10320 (flag_sanitize): Add variable.
10321 (fsanitize=): Add option. Add Driver.
10322 (fsanitize=thread): Remove option.
10323 (fsanitize=address): Likewise.
10324 (static-libubsan): New option.
10325 * doc/invoke.texi: Document the new flag and -static-libubsan.
10326 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
10327 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
10328 * builtin-attrs.def (ATTR_COLD): Define.
10329 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10330 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
10331 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
10332 * flag-types.h (sanitize_code): New enum.
10333 * opts.c (common_handle_option): Parse command line arguments
10334 of -fsanitize=. Add -fsanitize=unreachable option.
10335 * varasm.c (get_variable_section): Adjust.
10336 (assemble_noswitch_variable): Likewise.
10337 (assemble_variable): Likewise.
10338 (output_constant_def_contents): Likewise.
10339 (categorize_decl_for_section): Likewise.
10340 (place_block_symbol): Likewise.
10341 (output_object_block): Likewise.
10342 * builtins.def: Likewise.
10343 * toplev.c (compile_file): Likewise.
10344 (process_options): Likewise.
10345 * cppbuiltin.c: Likewise.
10346 * tsan.c (tsan_pass): Likewise.
10347 (tsan_gate): Likewise.
10348 (tsan_gate_O0): Likewise.
10349 * cfgexpand.c (partition_stack_vars): Likewise.
10350 (expand_stack_vars): Likewise.
10351 (defer_stack_allocation): Likewise.
10352 (expand_used_vars): Likewise.
10353 * cfgcleanup.c (old_insns_match_p): Likewise.
10354 * asan.c (asan_finish_file): Likewise.
10355 (asan_instrument): Likewise.
10356 (gate_asan): Likewise.
10357 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
10358 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10359 (asan_global_struct): Use pointer_sized_int_node instead
10360 calling build_nonstandard_integer_type.
10361 (initialize_sanitizer_builtins): Likewise.
10362 (asan_finish_file): Likewise.
10363 * gcc.c: Document %{%:function(args):X}.
10364 (static_spec_functions): Add sanitize.
10365 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
10366 store funcval != NULL there.
10367 (do_spec_1): Adjust handle_spec_function caller.
10368 (handle_braces): Allow %:function(args) as condition.
10369 (sanitize_spec_function): New function.
10370 (ADD_STATIC_LIBUBSAN_LIBS): Define.
10371 (LIBUBSAN_SPEC): Likewise.
10372 (LIBUBSAN_EARLY_SPEC): Likewise.
10373 (SANITIZER_SPEC): Handle libubsan.
10374 (SANITIZER_EARLY_SPEC): Likewise.
10375 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
10376 instead of fsanitize=address.
10377 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
10378 instead of fsanitize=address*.
10379 * builtins.c: Include ubsan.h.
10380 (fold_builtin_0): Instrument __builtin_unreachable.
10381 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
10382 instead of flag_asan.
10383 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
10384 (pointer_sized_int_node): Define.
10385 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
10386
10387 2013-08-30 Mike Stump <mikestump@comcast.net>
10388
10389 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
10390 with RE patterns.
10391
10392 2013-08-29 Jan Hubicka <jh@suse.cz>
10393
10394 * cgraph.c (cgraph_function_body_availability): Handle weakref
10395 correctly.
10396 * passes.def: Remove pass_fixup_cfg.
10397 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
10398 track when we need to remove functions.
10399 (gate_ipa_inline): Execute inlining always; add comment why.
10400 (pass_data_ipa_inline): Remove TODO_remove_functions.
10401 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
10402 do not produce summaries.
10403 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
10404 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
10405 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
10406 constant pool and vtable.
10407
10408 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
10409
10410 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
10411 New arm_neon.h's internal macros to specify 64-bit constants.
10412 Avoid using stdint.h's macros.
10413
10414 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
10415
10416 * recog.c (verify_changes): Verify that changes[i].old is non-zero
10417 before applying REG_P.
10418
10419 2013-08-30 Jakub Jelinek <jakub@redhat.com>
10420
10421 PR tree-optimization/58277
10422 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
10423 after seeing too many stmts with vdef in between dombb and current
10424 bb, invalidate everything.
10425
10426 2013-08-30 Richard Biener <rguenther@suse.de>
10427
10428 * fold-const.c (fold_single_bit_test): Fix overflow test.
10429
10430 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
10431
10432 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
10433 and which can live in a register, always retrieve the value on entry.
10434 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
10435 passed by invisible reference specially.
10436 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
10437 (vt_add_function_parameter): Correctly deal with a parameter passed by
10438 invisible reference.
10439
10440 2013-08-30 Jan Hubicka <jh@suse.cz>
10441
10442 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
10443
10444 2013-08-30 Richard Biener <rguenther@suse.de>
10445
10446 PR tree-optimization/58228
10447 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
10448 allow invariant loads in nested loop vectorization.
10449
10450 2013-08-30 Richard Biener <rguenther@suse.de>
10451
10452 PR tree-optimization/58223
10453 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
10454 (has_anti_or_output_dependence): ... this and adjust to also
10455 look for output dependences.
10456 (mark_nodes_having_upstream_mem_writes): Adjust.
10457 (rdg_flag_uses): Likewise.
10458
10459 2013-08-30 Richard Biener <rguenther@suse.de>
10460
10461 PR tree-optimization/58010
10462 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
10463 assert that we have a loop-closed PHI.
10464
10465 2013-08-29 Jan Hubicka <jh@suse.cz>
10466
10467 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
10468 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
10469 * lto-section-in.c (lto_free_function_in_decl_state): New function.
10470 (lto_free_function_in_decl_state_for_node): New function.
10471
10472 2013-08-29 Xinliang David Li <davidxl@google.com>
10473
10474 * loop-unroll.c (report_unroll_peel): Minor message change.
10475 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
10476 Emit alignment peeling message with default -fopt-info.
10477 (vect_loop_versioning): Emit loop version info message.
10478 * tree-vectorizer.c (vectorize_loops): Minor message change.
10479 (execute_vect_slp): Ditto.
10480
10481 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
10482
10483 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
10484 of the clone from the DECL_NAME of the original function.
10485
10486 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
10487
10488 * passes.c (register_pass): Add overload.
10489 * tree-pass.h (register_pass): Forward declare it. Add comment.
10490
10491 2013-08-29 Jan Hubicka <jh@suse.cz>
10492
10493 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
10494 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
10495 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
10496 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10497 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
10498 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
10499 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10500 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
10501
10502 2013-08-29 Teresa Johnson <tejohnson@google.com>
10503
10504 * dumpfile.c (dump_loc): Output column number.
10505 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
10506 * doc/invoke.texi: Document optall -fopt-info flag.
10507 * profile.c (read_profile_edge_counts): Use new dump framework.
10508 (compute_branch_probabilities): Ditto.
10509 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
10510 when pass not in any opt group.
10511 * pass_manager.h (pass_manager::get_pass_profile): New method.
10512 * value-prof.c (check_counter): Use new dump framework.
10513 (check_ic_target): Ditto.
10514 * coverage.c (get_coverage_counts): Ditto.
10515 (coverage_init): Setup new dump framework.
10516
10517 2013-08-29 Richard Biener <rguenther@suse.de>
10518
10519 PR tree-optimization/58246
10520 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
10521 handle the dominance check inside a basic-block.
10522
10523 2013-08-29 Richard Biener <rguenther@suse.de>
10524
10525 PR middle-end/57287
10526 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
10527 of default defs that appear in abnormal PHI nodes.
10528
10529 2013-08-29 Richard Biener <rguenther@suse.de>
10530
10531 PR tree-optimization/57685
10532 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
10533 single-use operands to avoid exponential complexity.
10534
10535 2013-08-28 Dehao Chen <dehao@google.com>
10536
10537 * ipa-inline.c (edge_badness): Fix integer underflow.
10538
10539 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
10540
10541 * gtm-builtins.def (_ITM_free): Declare leaf.
10542
10543 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10544
10545 PR target/58067
10546 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
10547 (*tls_local_dynamic_base_64_largepic): Likewise.
10548 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
10549 Remove predicate from call operand.
10550 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
10551 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
10552
10553 2013-08-28 Jeff Law <law@redhat.com>
10554
10555 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
10556 checks for the number of predecessors and successors allowed.
10557 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
10558 which require copying a joiner block if there is a request which
10559 is a subpath that requires no joiner block copying.
10560
10561 2013-08-28 Jan Hubicka <jh@suse.cz>
10562
10563 * lto-streamer-out.c (DFS_write_tree_body): Drop
10564 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
10565 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
10566 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
10567 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
10568 Do not read DECL_ERROR_ISSUED.
10569 (unpack_ts_decl_with_vis_value_fields): Do not read
10570 DECL_DEFER_OUTPUT.
10571 (lto_input_ts_binfo_tree_pointers): Do not read
10572 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
10573 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
10574 write DECL_ERROR_ISSUED..
10575 (pack_ts_decl_with_vis_value_fields): Do not write
10576 DECL_DEFER_OUTPUT.
10577 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
10578 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
10579 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
10580 * tree.h (tree_decl_common): Update comment.
10581 (DECL_ERROR_ISSUED): Remove.
10582
10583 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10584
10585 PR middle-end/58257
10586 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
10587
10588 2013-08-28 Jan Hubicka <jh@suse.cz>
10589
10590 * builtins.def (free): Declare leaf.
10591
10592 2013-08-27 David Malcolm <dmalcolm@redhat.com>
10593
10594 * gdbhooks.py: New.
10595 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
10596 * configure: Regenerate.
10597
10598 2013-08-27 Martin Jambor <mjambor@suse.cz>
10599
10600 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
10601 (ipa_ancestor_jf_data): Likewise.
10602 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
10603 (ipa_get_jf_pass_through_type_preserved): New function.
10604 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
10605 (ipa_get_jf_ancestor_type_preserved): New function.
10606 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
10607 (ipa_get_jf_ancestor_result): Likewise.
10608 (propagate_vals_accross_pass_through): Use
10609 ipa_get_jf_pass_through_result to do all the value mappings.
10610 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
10611 type_preserved flag.
10612 (ipa_set_jf_cst_copy): New function.
10613 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
10614 (ipa_set_jf_arith_pass_through): Likewise.
10615 (ipa_set_ancestor_jf): Likewise.
10616 (compute_complex_assign_jump_func): Set type_preserved instead of
10617 punting.
10618 (ipa_compute_jump_functions_for_edge): Likewise.
10619 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
10620 (update_jump_functions_after_inlining): Update type_preserved.
10621 Explicitely create jump functions when combining one with pass_through.
10622 (ipa_write_jump_function): Stream the type_preserved flags.
10623 (ipa_read_jump_function): Likewise.
10624
10625 2013-08-27 Jakub Jelinek <jakub@redhat.com>
10626 Aldy Hernandez <aldyh@redhat.com>
10627
10628 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
10629 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
10630 * function.h (struct function): Add has_force_vect_loops and
10631 has_simduid_loops.
10632 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
10633 * gimple.c (gimple_build_omp_critical): Add KIND argument and
10634 handle it.
10635 * gimple.def: Update CLAUSES comments.
10636 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
10637 (gimple_build_omp_for): Add argument to prototype.
10638 (gimple_omp_for_kind): New.
10639 (gimple_omp_for_set_kind): New.
10640 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
10641 GOVD_DATA_SHARE_CLASS.
10642 (enum omp_region_type): Add ORT_SIMD.
10643 (gimple_add_tmp_var): Handle ORT_SIMD.
10644 (gimplify_var_or_parm_decl): Same.
10645 (is_gimple_stmt): Same.
10646 (omp_firstprivatize_variable): Same.
10647 (omp_add_variable): Only use splay_tree_insert if lookup failed.
10648 (omp_notice_variable): Handle ORT_SIMD.
10649 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
10650 (omp_check_private): Handle ORT_SIMD.
10651 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10652 OMP_CLAUSE_SAFELEN.
10653 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
10654 Handle OMP_CLAUSE_LASTPRIVATE.
10655 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10656 OMP_CLAUSE_SAFELEN.
10657 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
10658 (gimplify_expr): Handle OMP_SIMD.
10659 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
10660 (expand_GOMP_SIMD_VF): New.
10661 (expand_GOMP_SIMD_LAST_LANE): New.
10662 * internal-fn.def (GOMP_SIMD_LANE): New.
10663 (GOMP_SIMD_VF): New.
10664 (GOMP_SIMD_LAST_LANE): New.
10665 * omp-low.c: Include target.h.
10666 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
10667 OMP_CLAUSE_SAFELEN.
10668 (check_omp_nesting_restrictions): Same.
10669 (omp_max_vf): New.
10670 (lower_rec_simd_input_clauses): New.
10671 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
10672 OMP_CLAUSE_LINEAR.
10673 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
10674 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
10675 (expand_omp_build_assign): New.
10676 (expand_omp_for_init_counts): New.
10677 (expand_omp_for_init_vars): New.
10678 (extract_omp_for_update_vars): New.
10679 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
10680 and rewrite accordingly.
10681 (expand_omp_simd): New.
10682 (expand_omp_for): Use expand_omp_simd.
10683 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
10684 (lower_omp_for): Do not lower the body.
10685 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
10686 in their own loops.
10687 * tree-flow.h (find_omp_clause): Remove prototype.
10688 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
10689 forcing vectorization of the loop, or if flag_tree_vectorize.
10690 (gate_tree_if_conversion): Similarly.
10691 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
10692 gimple_build_omp_for.
10693 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
10694 * tree-parloops (create_parallel_loop): Pass kind argument to
10695 gimple_build_omp_for.
10696 * tree-pretty-print.c (dump_omp_clause): Add cases for
10697 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
10698 OMP_CLAUSE__SIMDUID_.
10699 (dump_generic_node): Handle OMP_SIMD.
10700 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
10701 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
10702 unroll OMP_SIMD loops here.
10703 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
10704 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
10705 loop->safelen.
10706 (vect_analyze_data_refs): Handle simd loops.
10707 * tree-vect-loop.c (vectorizable_live_operation): Handle
10708 IFN_GOMP_SIMD*.
10709 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
10710 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
10711 (vectorizable_load): Same.
10712 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
10713 (struct simduid_to_vf): New.
10714 (simduid_to_vf::hash): New.
10715 (simduid_to-vf::equal): New.
10716 (struct simd_array_to_simduid): New.
10717 (simd_array_to_simduid::hash): New.
10718 (simd_array_to_simduid::equal): New.
10719 (adjust_simduid_builtins): New.
10720 (struct note_simd_array_uses_struct): New.
10721 (note_simd_array_uses_cb): New.
10722 (note_simd_array_uses): New.
10723 (vectorize_loops): Handle simd hints and adjust simd builtins
10724 accordingly.
10725 * tree-vectorizer.h (struct _stmt_vec_info): Add
10726 simd_lane_access_p field.
10727 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
10728 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
10729 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
10730 (omp_clause_code_name): Same.
10731 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
10732 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
10733 * tree.def (OMP_SIMD): New entry.
10734 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
10735 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
10736 (OMP_CLAUSE_DECL): Adjust range for new clauses.
10737 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
10738 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
10739 (OMP_CLAUSE_LINEAR_STEP): New.
10740 (OMP_CLAUSE_SAFELEN_EXPR): New.
10741 (OMP_CLAUSE__SIMDUID__DECL): New.
10742 (find_omp_clause): New prototype.
10743
10744 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
10745
10746 * config/i386/driver-i386.c (host_detect_local_cpu): Update
10747 Haswell processor detection.
10748
10749 2013-08-27 Christian Widmer <shadow@umbrox.de>
10750
10751 PR target/57927
10752 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
10753 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
10754 AVX2 capable processors.
10755
10756 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
10757
10758 * config/aarch64/arm_neon.h: Replace all inline asm implementations
10759 of vget_low_* with implementations in terms of other intrinsics.
10760
10761 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
10762
10763 PR middle-end/57219
10764 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
10765 values to -1, 0 and 1.
10766
10767 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
10768
10769 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
10770 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
10771 (<optab><mode>3_insn): Remove.
10772 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
10773 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
10774 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
10775 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
10776 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
10777 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
10778 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
10779 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
10780 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
10781 (ror<mode>3_insn): Likewise.
10782 * config/aarch64/predicates.md (aarch64_simd_register): New.
10783
10784 2013-08-27 Richard Biener <rguenther@suse.de>
10785
10786 PR tree-optimization/57521
10787 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
10788 one edge is non-critical.
10789 (find_phi_replacement_condition): Make sure to use a non-critical
10790 edge. Cleanup and remove old bug workarounds.
10791 (bb_postdominates_preds): Remove.
10792 (if_convertible_loop_p_1): Do not compute post-dominators.
10793 (combine_blocks): Do not free post-dominators.
10794 (main_tree_if_conversion): Likewise.
10795 (pass_data_if_conversion): Add TODO_verify_ssa.
10796
10797 2013-08-27 DJ Delorie <dj@redhat.com>
10798
10799 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
10800
10801 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
10802
10803 * function.c (assign_parm_find_data_types): Set passed_mode and
10804 nominal_mode to the TYPE_MODE of nominal_type for the built
10805 pointer type in case of the struct-pass-by-reference.
10806
10807 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10808
10809 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
10810 (UINT16_TYPE): Change default to "unsigned int".
10811
10812 * config/avr/avr.opt (mfract-convert-truncate): New option.
10813 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
10814 is set, round negative fractional integers according to n1169
10815 when converting to integer types.
10816
10817 2013-08-26 Jan Hubicka <jh@suse.cz>
10818
10819 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
10820 methods can not be called indirectly when their address is not taken.
10821
10822 2013-08-26 Jan Hubicka <jh@suse.cz>
10823
10824 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
10825 ctor_for_folding.
10826
10827 2013-08-26 Jan Hubicka <jh@suse.cz>
10828
10829 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
10830 can be unshared.
10831
10832 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10833
10834 * reload.c (find_valid_class): Allow classes that do not include
10835 FIRST_PSEUDO_REGISTER - 1.
10836
10837 2013-08-26 Jan Hubicka <jh@suse.cz>
10838
10839 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
10840 fix edge count/frequency when speculation failed; fix type check
10841 for the direct call.
10842
10843 2013-08-26 Jan Hubicka <jh@suse.cz>
10844
10845 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
10846
10847 2013-08-26 Jan Hubicka <jh@suse.cz>
10848
10849 * ipa-inline-transform.c (inline_transform): Be ready for basic block
10850 to be changed by edge redirection.
10851
10852 2013-08-26 Jan Hubicka <jh@suse.cz>
10853
10854 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
10855 formating; add sanity check.
10856 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
10857 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
10858 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
10859 (resolve_noninline_speculation): Update callee keys, too.
10860
10861 2013-08-26 Jan Hubicka <jh@suse.cz>
10862
10863 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
10864 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
10865
10866 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10867
10868 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
10869 into proper place.
10870
10871 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
10872
10873 * config/i386/i386.c (ix86_debug_options): Remove prototype.
10874 (x86_64_elf_select_section): Ditto.
10875 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
10876 arguments.
10877 (ix86_pass_by_reference): Ditto.
10878 (output_set_got): Ditto.
10879 (ix86_unary_operator_ok): Ditto.
10880 (ix86_expand_builtin): Ditto.
10881
10882 2013-08-23 Jan Hubicka <jh@suse.cz>
10883
10884 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
10885
10886 2013-08-23 Jan Hubicka <jh@suse.cz>
10887
10888 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
10889 (tree_decl_with_vis): Add FINAL field.
10890
10891 2013-08-23 Jeff Law <law@redhat.com>
10892
10893 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
10894 the debugging dump when the expression is fully redundant.
10895
10896 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
10897
10898 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
10899 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
10900 * pretty-print.c (pp_formatted_text_data): Likewise.
10901 (pp_write_text_to_stream): Likewise.
10902 (pp_write_text_as_dot_label_to_stream): Likewise.
10903 (pp_append_r): Likewise.
10904 (pp_format): Likewise.
10905 (pp_flush): Likewise.
10906 (pp_clear_output_area): Likewise.
10907 (pp_append_text): Likewise.
10908 (pp_formatted_text): Likewise.
10909 (pp_remaining_character_count_for_line): Likewise.
10910 (pp_newline): Likewise.
10911 (pp_character): Likewise.
10912 (output_buffer::~output_buffer): Define.
10913 (pretty_printer::~pretty_printer): Destruct output buffer.
10914 * pretty-print.h (output_buffer::~output_buffer): Declare.
10915 (pretty_printer::~pretty_printer): Declare virtual.
10916
10917 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
10918
10919 PR other/57324
10920 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
10921 HOST_WIDE_INT_M1U): New macros.
10922 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
10923 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
10924 unsigned -1 for lshift.
10925 * cse.c (cse_insn): Likewise.
10926 * double-int.c (rshift_double, lshift_double): Likewise.
10927 * builtins.c (fold_builtin_bitop): Likewise.
10928 * combine.c (force_to_mode): Likewise.
10929 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
10930 * simplify-rtx.c (simplify_const_unary_operation,
10931 simplify_const_binary_operation): Likewise.
10932 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
10933 check_va_list_escapes): Likewise.
10934 * rtlanal.c (nonzero_bits1): Likewise.
10935 * expmed.c (expand_smod_pow2): Likewise.
10936 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
10937
10938 2013-08-23 Jan Hubicka <jh@suse.cz>
10939
10940 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
10941 as having address taken.
10942
10943 2013-08-23 Jan Hubicka <jh@suse.cz>
10944
10945 * ipa-utils.h (method_class_type): Declare.
10946 * ipa-devirt.c (method_class_type): Export.
10947
10948 * cgraphunit.c (analyze_functions): Do basic devirtualization;
10949 do not walk base classes of anonymous types.
10950
10951 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
10952
10953 PR rtl-optimization/58220
10954 PR regression/58221
10955 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
10956 handle SEQUENCE insns properly.
10957
10958 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
10959
10960 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
10961 definition.
10962 (pp_newline_and_indent): Likewise.
10963 (pp_separate_with): Likewise.
10964 * pretty-print.c (pp_newline_and_flush): Define.
10965 (pp_newline_and_indent): Likewise.
10966 (pp_separate_with): Likewise.
10967
10968 2013-08-23 Jakub Jelinek <jakub@redhat.com>
10969
10970 PR target/58218
10971 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
10972 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
10973
10974 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
10975
10976 * config/i386/predicates.md (ext_sse_reg_operand): New.
10977 * config/i386/i386.md (*movti_internal): Use
10978 predicate to determine if EVEX is needed.
10979 (*movsi_internal): Ditto.
10980 (*movdf_internal): Ditto.
10981 (*movsf_internal): Ditto.
10982 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
10983
10984 2013-08-23 Jakub Jelinek <jakub@redhat.com>
10985
10986 PR tree-optimization/58209
10987 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
10988 (find_tail_calls): Give up for pointer result types if m is non-NULL.
10989 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
10990 emit POINTER_PLUS_EXPR.
10991 (create_tailcall_accumulator): For pointer result type accumulate in
10992 sizetype type.
10993
10994 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
10995
10996 * configure.ac: Add backslashes missing from the last change.
10997 * configure: Regenerate.
10998
10999 2013-08-22 Jan Hubicka <jh@susue.cz>
11000
11001 * ipa.c (function_and_variable_visibility): First remember function
11002 was global and then make it local.
11003
11004 2013-08-22 Julian Brown <julian@codesourcery.com>
11005
11006 * configure.ac: Add aarch64 to list of arches which use "nop" in
11007 debug_line test.
11008 * configure: Regenerate.
11009
11010 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11011
11012 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
11013 gnu_libc_has_function.
11014 * config/s390/tpf.h: Likewise.
11015
11016 2013-08-22 Jan Hubicka <jh@susue.cz>
11017
11018 * timevar.c (validate_phases): Add cast.
11019
11020 2013-08-22 Jan Hubicka <jh@susue.cz>
11021
11022 * timevar.c (validate_phases): Use size_t for memory.
11023 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
11024
11025 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
11026
11027 * pretty-print.h (output_buffer::output_buffer): Declare.
11028 (pretty_printer::pretty_printer): Likewise.
11029 (pp_construct): Remove.
11030 * pretty-print.c (output_buffer::output_buffer): Define.
11031 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
11032 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
11033 (print_gimple_expr): Likewise.
11034 (print_gimple_seq): Likewise.
11035 (gimple_dump_bb): Likewise.
11036 * sched-vis.c (dump_value_slim): Likewise.
11037 (dump_insn_slim): Likewise.
11038 (dump_rtl_slim): Likewise.
11039 (str_pattern_slim): Likewise.
11040 * tree-mudflap.c (mf_varname_tree): Likewise.
11041 * graph.c (print_graph_cfg): Likewise.
11042 (start_graph_dump): Likewise.
11043 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
11044 placement-new.
11045 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
11046 pretty printer initialization.
11047 * coretypes.h (diagnostic_context): Remove superflous type alias
11048 declaration.
11049 (pretty_printer): Likewise. Declare directly as a class.
11050 (pretty_print_info): Remove declaration as class.
11051 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
11052 and pp_clear_output_area.
11053 (asan_add_global): Likewise.
11054
11055 2013-08-22 Jan Hubicka <jh@suse.cz>
11056
11057 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
11058 * ipa-utils.h (update_type_inheritance_graph): Declare.
11059 (possible_polymorphic_call_target_p): Declare.
11060 (possible_polymorphic_call_target_p): New.
11061 * ipa-devirt.c: Update toplevel comments.
11062 (cached_polymorphic_call_targets): Move up.
11063 (odr_type_d): Move ID down.
11064 (polymorphic_type_binfo_p): Update comment.
11065 (odr_hasher::remove): Likewise;
11066 (get_odr_type): Set anonymous_namespace.
11067 (dump_odr_type): Dump it.
11068 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
11069 (maybe_record_node): Record node in cached_polymorphic_call_targets.
11070 (record_binfo): Add comment.
11071 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
11072 built.
11073 (devirt_node_removal_hook): Do not iCE when cache is freed.
11074 (possible_polymorphic_call_target_p): New predicate.
11075 (update_type_inheritance_graph): New function.
11076
11077 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
11078 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
11079 Sergey Lega <sergey.s.lega@intel.com>
11080 Anna Tikhonova <anna.tikhonova@intel.com>
11081 Ilya Tocar <ilya.tocar@intel.com>
11082 Andrey Turetskiy <andrey.turetskiy@intel.com>
11083 Ilya Verbin <ilya.verbin@intel.com>
11084 Kirill Yukhin <kirill.yukhin@intel.com>
11085 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11086
11087 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
11088 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
11089 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
11090 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
11091 (OPTION_MASK_ISA_AVX2_UNSET): Update.
11092 (OPTION_MASK_ISA_AVX512F_UNSET): New.
11093 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
11094 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
11095 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
11096 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
11097 OPT_mavx512pf, OPT_mavx512er cases.
11098 * config/i386/constraints.md (v): New constraint.
11099 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
11100 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
11101 (bit_AVX512CD): New.
11102 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11103 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
11104 * config/i386/i386-c.c (ix86_target_macros_internal):
11105 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
11106 __AVX512PF__.
11107 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
11108 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
11109 * config/i386/i386.c (regclass_map, dbx_register_map)
11110 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
11111 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
11112 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
11113 -mavx512pf options.
11114 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
11115 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
11116 -mavx512pf options. Fix formatting.
11117 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
11118 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
11119 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
11120 -mavx512cd, -mavx512pf options.
11121 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
11122 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
11123 (ix86_preferred_output_reload_class): Replace SSE_REGS with
11124 ALL_SSE_REGS.
11125 (ix86_hard_regno_mode_ok): Support 512-bit registers.
11126 (ix86_set_reg_reg_cost): Ditto.
11127 (x86_order_regs_for_local_alloc): Ditto.
11128 (MAX_VECT_LEN): Extend to 64-byte.
11129 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
11130 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
11131 (TARGET_AVX512ER, TARGET_AVX512CD): New.
11132 (BIGGEST_ALIGNMENT): Extend to 512-bits.
11133 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
11134 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
11135 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
11136 (SSE_REG_MODE_P): Support new modes.
11137 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
11138 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
11139 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
11140 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
11141 (REG_CLASS_CONTENTS): Add new registers.
11142 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
11143 (EXT_REX_SSE_REGNO_P): New.
11144 (HI_REGISTER_NAMES): Add new registers.
11145 * config/i386/i386.md: Define constants for new registers.
11146 (mode): Add new 512-bit modes.
11147 (prefix): Support evex prefix.
11148 (isa): Support avx512f, noavx512f, fma_avx512f.
11149 (ssemodesuffix): Add new 512-bit modes.
11150 (movxi): New.
11151 (*movxi_internal_avx512f): Ditto.
11152 (*movdi_internal): Replace constraint "x" with the new constraint "v".
11153 Support MODE_XI.
11154 (*movsi_internal): Likewise.
11155 (*movdf_internal): Likewise.
11156 (*movsf_internal): Likewise.
11157 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
11158 (<code><mode>3): Likewise.
11159 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
11160 New.
11161 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
11162 with the new constraint "v".
11163 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
11164 modes.
11165 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
11166 with the new constraint "v".
11167 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
11168 (<sse2>_storedqu<avxsizesuffix>): Likewise.
11169 (*<plusminus_insn><mode>3): Likewise.
11170 (<sse>_vm<plusminus_insn><mode>3): Likewise.
11171 (*mul<mode>3): Likewise.
11172 (<sse>_vmmul<mode>3): Likewise.
11173 (<sse>_div<mode>3): Likewise.
11174 (<sse>_vmdiv<mode>3): Likewise.
11175 (<sse>_sqrt<mode>2): Likewise.
11176 (<sse>_vmsqrt<mode>2): Likewise.
11177 (*<code><mode>3_finite): Likewise.
11178 (*<code><mode>3) <smaxmin>: Likewise.
11179 (<sse>_vm<code><mode>3): Likewise.
11180 (*<code><mode>3) <any_logic>: Likewise.
11181 (*fma_fmadd_<mode>): Likewise.
11182 (*fma_fmsub_<mode>): Likewise.
11183 (*fma_fnmadd_<mode>): Likewise.
11184 (*fma_fnmsub_<mode>): Likewise.
11185 (*fma_fmaddsub_<mode>): Likewise.
11186 (*fma_fmsubadd_<mode>): Likewise.
11187 (*fmai_fmadd_<mode>): Likewise.
11188 (*fmai_fmsub_<mode>): Likewise.
11189 (*fmai_fnmadd_<mode>): Likewise.
11190 (*fmai_fnmsub_<mode>): Likewise.
11191 (sse_cvtsi2ss): Likewise.
11192 (sse_cvtsi2ssq): Likewise.
11193 (sse_cvtss2si): Likewise.
11194 (sse_cvtss2si_2): Likewise.
11195 (sse_cvtss2siq): Likewise.
11196 (sse_cvtss2siq_2): Likewise.
11197 (sse_cvttss2si): Likewise.
11198 (sse_cvtss2siq_2): Likewise.
11199 (float<sseintvecmodelower><mode>2): Likewise.
11200 (sse2_cvtsd2si_2): Likewise.
11201 (sse2_cvtsd2siq_2): Likewise.
11202 (*<plusminus_insn><mode>3): Likewise.
11203 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
11204 (*<sse4_1_avx2>_mul<mode>3): Likewise.
11205 (ashr<mode>3): Likewise.
11206 (<shift_insn><mode>3): Likewise.
11207 (avx2_<code><mode>3): Likewise.
11208 (*avx2_<code><mode>3): Likewise.
11209 (*andnot<mode>3): Likewise.
11210 (*<code><mode>3) <any_logic>: Likewise.
11211 (abs<mode>2): Likewise.
11212 (avx2_permvar<mode>): Likewise.
11213 (avx2_perm<mode>_1): Likewise.
11214 (*avx_vpermilp<mode>): Likewise.
11215 (avx_vpermilvar<mode>3): Likewise.
11216 (avx2_ashrv<mode>): Likewise.
11217 (avx2_<shift_insn>v<mode>): Likewise.
11218 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
11219 -mavx512cd.
11220 * doc/rtl.texi: Document XImode.
11221
11222 2013-08-21 Jeff Law <law@redhat.com>
11223
11224 * tree-flow.h (register_jump_thread): Pass vector of edges
11225 instead of each important edge.
11226 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
11227 thread path into a vector and pass that to register_jump_thread.
11228 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
11229 passed in edge vector to the current 3-edge form.
11230
11231 Revert:
11232 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11233
11234 * dce.c (fini_dce): Call df_analyze again just in case
11235 delete_unmarked_insns removed anything.
11236
11237 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
11238
11239 * reload.h (struct reg_equivs): Rename to ..
11240 (struct reg_equivs_s): .. this.
11241
11242 2013-08-20 Martin Liska <marxin.liska@gmail.com>
11243
11244 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
11245
11246 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11247
11248 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
11249
11250 2013-08-21 Jeff Law <law@redhat.com>
11251
11252 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
11253 simplify assignments too. If the RHS collapses to a singleton
11254 range, then return the value for the range.
11255
11256 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
11257
11258 * config/i386/sse.md (V16): Rename to...
11259 (VMOVE): this.
11260 (mov<mode>): Update iterator name.
11261 (*mov<mode>_internal): Ditto.
11262 (push<mode>1): Ditto.
11263 (movmisalign<mode>): Ditto.
11264
11265 2013-08-20 Jan Hubicka <jh@suse.cz>
11266
11267 PR bootstrap/58186
11268 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
11269 entry for direct edges.
11270 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
11271
11272 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11273
11274 Revert my last two changes, r201865 and r201864:
11275
11276 Revert r201865:
11277 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11278
11279 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11280 instances can own GC refs.
11281
11282 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11283 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11284 (gcc::context::gt_pch_nx): Likewise.
11285 (gcc::context::gt_pch_nx): Likewise.
11286 * ggc.h (gt_ggc_mx <T>): New.
11287 (gt_pch_nx_with_op <T>): New.
11288 (gt_pch_nx <T>): New.
11289 * passes.c (opt_pass::gt_ggc_mx): New.
11290 (opt_pass::gt_pch_nx): New.
11291 (opt_pass::gt_pch_nx_with_op): New.
11292 (pass_manager::gt_ggc_mx): New.
11293 (pass_manager::gt_pch_nx): New.
11294 (pass_manager::gt_pch_nx_with_op): New.
11295 (pass_manager::operator new): Use
11296 ggc_internal_cleared_alloc_stat rather than xcalloc.
11297 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11298 (pass_manager::gt_ggc_mx): New.
11299 (pass_manager::gt_pch_nx): New.
11300 (pass_manager::gt_pch_nx_with_op): New.
11301 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11302 (opt_pass::operator new): New.
11303 (opt_pass::gt_ggc_mx): New.
11304 (opt_pass::gt_pch_nx): New.
11305 (opt_pass::gt_pch_nx_with_op): New.
11306
11307 Revert r201864:
11308 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11309
11310 * Makefile.in (GTFILES): Add context.h.
11311 * context.c (gcc::context::operator new): New.
11312 (gcc::context::gt_ggc_mx): New.
11313 (gcc::context::gt_pch_nx): New.
11314 (gcc::context::gt_pch_nx): New.
11315 * context.h (gcc::context): Add GTY((user)) marking.
11316 (gcc::context::operator new): New.
11317 (gcc::context::gt_ggc_mx): New.
11318 (gcc::context::gt_pch_nx): New.
11319 (gcc::context::gt_pch_nx): New.
11320 (g): Add GTY marking.
11321 (gt_ggc_mx (gcc::context *)): New.
11322 (gt_pch_nx (gcc::context *)): New.
11323 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11324 void *cookie)): New.
11325 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11326
11327 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11328
11329 * dce.c (fini_dce): Call df_analyze again just in case
11330 delete_unmarked_insns removed anything.
11331
11332 2013-08-20 Teresa Johnson <tejohnson@google.com>
11333
11334 PR rtl-optimizations/57451
11335 * final.c (reemit_insn_block_notes): Prevent lexical blocks
11336 from crossing split section boundaries.
11337
11338 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11339
11340 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
11341 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
11342 with MULTLIB_DEFAULTS.
11343
11344 2013-08-20 Nick Clifton <nickc@redhat.com>
11345
11346 * target.def (narrow_volatile_bitfield): Note that the default
11347 value is false, not !TARGET_STRICT_ALIGN.
11348 * doc/tm.texi: Regenerate.
11349
11350 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
11351
11352 Fix LIB_SPEC for systems without libpthread.
11353
11354 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
11355 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
11356 for Android.
11357 * config/i386/linux-common.h: Likewise.
11358 * config/mips/linux-common.h: Likewise.
11359
11360 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11361
11362 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
11363 checks.
11364
11365 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11366
11367 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11368 instances can own GC refs.
11369
11370 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11371 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11372 (gcc::context::gt_pch_nx): Likewise.
11373 (gcc::context::gt_pch_nx): Likewise.
11374 * ggc.h (gt_ggc_mx <T>): New.
11375 (gt_pch_nx_with_op <T>): New.
11376 (gt_pch_nx <T>): New.
11377 * passes.c (opt_pass::gt_ggc_mx): New.
11378 (opt_pass::gt_pch_nx): New.
11379 (opt_pass::gt_pch_nx_with_op): New.
11380 (pass_manager::gt_ggc_mx): New.
11381 (pass_manager::gt_pch_nx): New.
11382 (pass_manager::gt_pch_nx_with_op): New.
11383 (pass_manager::operator new): Use
11384 ggc_internal_cleared_alloc_stat rather than xcalloc.
11385 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11386 (pass_manager::gt_ggc_mx): New.
11387 (pass_manager::gt_pch_nx): New.
11388 (pass_manager::gt_pch_nx_with_op): New.
11389 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11390 (opt_pass::operator new): New.
11391 (opt_pass::gt_ggc_mx): New.
11392 (opt_pass::gt_pch_nx): New.
11393 (opt_pass::gt_pch_nx_with_op): New.
11394
11395 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11396
11397 * Makefile.in (GTFILES): Add context.h.
11398 * context.c (gcc::context::operator new): New.
11399 (gcc::context::gt_ggc_mx): New.
11400 (gcc::context::gt_pch_nx): New.
11401 (gcc::context::gt_pch_nx): New.
11402 * context.h (gcc::context): Add GTY((user)) marking.
11403 (gcc::context::operator new): New.
11404 (gcc::context::gt_ggc_mx): New.
11405 (gcc::context::gt_pch_nx): New.
11406 (gcc::context::gt_pch_nx): New.
11407 (g): Add GTY marking.
11408 (gt_ggc_mx (gcc::context *)): New.
11409 (gt_pch_nx (gcc::context *)): New.
11410 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11411 void *cookie)): New.
11412 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11413
11414 2013-08-20 Alan Modra <amodra@gmail.com>
11415
11416 PR target/57865
11417 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
11418 (rs6000_emit_epilogue): Likewise.
11419
11420 2013-08-19 Dehao Chen <dehao@google.com>
11421
11422 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
11423
11424 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
11425 Jakub Jelinek <jakub@redhat.com>
11426
11427 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
11428 (BUILT_IN_FABSD64): Likewise.
11429 (BUILT_IN_FABSD128): Likewise.
11430 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
11431 (fold_builtin_1): Likewise.
11432 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
11433 destination and source operands.
11434 (*abstd2_fpr): Likewise.
11435 (*nabstd2_fpr): Likewise.
11436
11437 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
11438
11439 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
11440 JUMP_P and INSN_P.
11441
11442 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
11443
11444 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
11445 incomplete.
11446
11447 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
11448
11449 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
11450 * builtins.c (default_libc_has_function): New.
11451 (gnu_libc_has_function): Ditto.
11452 (no_c99_libc_has_function): Ditto.
11453 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
11454 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
11455 (fold_builtin_sincos): Likewise.
11456 (fold_builtin_cexp): Likewise.
11457 * builtins.def (DEF_C94_BUILTIN): Likewise.
11458 (DEF_C99_BUILTIN): Likewise.
11459 (DEF_C99_C90RES_BUILTIN): Likewise.
11460 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
11461 definitions to using this define.
11462 * config/darwin-protos.h (darwin_libc_has_function): New.
11463 * config/darwin.c (darwin_libc_has_function): Ditto.
11464 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
11465 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
11466 * config/darwin.h: Ditto.
11467 * config/elfos.h: Ditto.
11468 * config/freebsd.h: Ditto.
11469 * config/i386/cygming.h: Ditto.
11470 * config/i386/djgpp.h: Ditto.
11471 * config/i386/i386-interix.h: Ditto.
11472 * config/microblaze/microblaze.h: Ditto.
11473 * config/mmix/mmix.h: Ditto.
11474 * config/gnu-user.h: Ditto.
11475 * config/ia64/hpux.h: Ditto.
11476 * config/pa/pa-hpux.h: Ditto.
11477 * config/pdp11/pdp11.h: Ditto.
11478 * config/picochip/picochip.h: Ditto.
11479 * config/linux.h: Ditto.
11480 * config/netbsd.h: Ditto.
11481 * config/openbsd.h: Ditto.
11482 * config/rs6000/aix43.h: Ditto.
11483 * config/rs6000/aix51.h: Ditto.
11484 * config/rs6000/aix52.h: Ditto.
11485 * config/rs6000/aix53.h: Ditto.
11486 * config/rs6000/aix61.h: Ditto.
11487 * config/rs6000/darwin.h: Ditto.
11488 * config/rs6000/linux.h: Ditto.
11489 * config/rs6000/linux64.h: Ditto.
11490 * config/s390/tpf.h: Ditto.
11491 * config/sol2-10.h: Ditto.
11492 * config/sol2.h: Ditto.
11493 * config/vms/vms.h: Ditto.
11494 * config/vxworks.h: Ditto.
11495 * config/linux-android.c (linux_android_libc_has_function):
11496 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
11497 * config/linux-protos.h (linux_android_libc_has_function):
11498 New declaration.
11499 * config/i386/i386.c (ix86_libc_has_function): New.
11500 * config/i386/i386-protos.h
11501 (ix86_libc_has_function): New declaration.
11502 * config/i386/i386.md
11503 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
11504 ("isinf<mode>2): Likewise.
11505 * convert.c (convert_to_integer): Using new target hook
11506 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
11507 TARGET_C99_FUNCTIONS.
11508 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
11509 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
11510 * coretypes.h (function_class): New enum for different
11511 classes of functions.
11512 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
11513 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
11514 (TARGET_HAS_SINCOS): Likewise.
11515 (TARGET_LIBC_HAS_FUNCTION): New.
11516 * doc/tm.texi: Regenerated.
11517 * targhooks.h (default_libc_has_function): New declaration.
11518 (no_c99_libc_has_function): Ditto.
11519 (gnu_libc_has_function): Ditto.
11520 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
11521 and TARGET_HAS_SINCOS.
11522
11523 2013-08-18 Jan Hubicka <jh@suse.cz>
11524
11525 * Makeifle-in (ipa-devirt.o): New.
11526 (GTFILES): Add ipa-utils.h and ipa-devirt.c
11527 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
11528 (analyze_functions): Look into possible targets of polymorphic call.
11529 * dumpfile.c (dump_files): Add type-inheritance dump.
11530 * dumpfile.h (TDI_inheritance): New.
11531 * ipa-devirt.c: New file.
11532 * ipa-utils.h (odr_type_d): Forward declare.
11533 (odr_type): New type.
11534 (build_type_inheritance_graph): Declare.
11535 (possible_polymorphic_call_targets): Declare and introduce inline
11536 variant when only edge is pased.
11537 (dump_possible_polymorphic_call_targets): Likewise.
11538 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
11539 * tree.c (type_in_anonymous_namespace_p): Break out from ...
11540 (types_same_for_odr): ... here.
11541 * tree.h (type_in_anonymous_namespace_p): Declare.
11542
11543 2013-08-18 Jakub Jelinek <jakub@redhat.com>
11544
11545 PR tree-optimization/58006
11546 * tree-parloops.c (take_address_of): Don't ICE if get_name
11547 returns NULL.
11548 (eliminate_local_variables_stmt): Remove clobber stmts.
11549
11550 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
11551
11552 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
11553 error message is issued for an alias to undefined symbol.
11554
11555 2013-08-18 Jan Hubicka <jh@suse.cz>
11556
11557 * cgraph.c (cgraph_create_indirect_edge): Discover
11558 polymorphic calls and record basic info into indirect_info.
11559 * gimple-fold.c (gimple_fold_call): When doing BINFO based
11560 devirtualization, ignore objc function calls.
11561 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
11562 call with no parm index info.
11563 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
11564 * tree.c (virtual_method_call_p): New function.
11565 * tree.h (virtual_method_call_p): Declare.
11566
11567 2013-08-16 Jan Hubicka <jh@suse.cz>
11568
11569 PR middle-end/58179
11570 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
11571
11572 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
11573
11574 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
11575 attribute.
11576
11577 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11578
11579 * gengtype.c (type_for_name): Add special-case support for
11580 locating types within the "gcc::" namespace.
11581 (open_base_files): Emit a "using namespace gcc" directive.
11582
11583 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11584
11585 PR target/58160
11586 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
11587 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
11588
11589 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
11590 array instead of each individual operand as a separate argument.
11591 (emit_fusion_gpr_load): Likewise.
11592 (expand_fusion_gpr_load): Add new function declaration.
11593
11594 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
11595 signature to have the operands passed as an array, instead of as
11596 separate arguments. Allow ZERO_EXTEND to be in the memory
11597 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
11598 depend on the register live/dead flags when peepholes are run.
11599 (expand_fusion_gpr_load): New function to be called from the
11600 peephole2 pass, to change the register that addis sets to be the
11601 target register.
11602 (emit_fusion_gpr_load): Change the calling signature to have the
11603 operands passed as an array, instead of as separate arguments.
11604 Allow ZERO_EXTEND to be in the memory address, and also
11605 SIGN_EXTEND if -mpower8-fusion-sign.
11606
11607 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
11608 unspec enumeration.
11609 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
11610 adjust the register addis loads up in the peephole2 pass. Do not
11611 depend on the register live/dead state when the peephole pass is done.
11612
11613 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11614
11615 * gengtype.c (create_user_defined_type): Ensure that the kind
11616 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
11617 declaration is seen before the GTY((user)) marking.
11618
11619 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
11620
11621 PR target/58105
11622 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
11623
11624 2013-08-16 Jan Hubicka <jh@suse.cz>
11625
11626 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
11627 arugment expected_type.
11628 (gimple_fold_call): Use it.
11629 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
11630 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
11631 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
11632 (try_make_edge_direct_virtual_call): Likewise.
11633 * tree.c (obj_type_ref_class): New.
11634 * tree.h (obj_type_ref_class): Use it.
11635
11636 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
11637
11638 * sched-vis.c (rtl_slim_pp_initialized): Remove.
11639 (rtl_slim_pp): Likewise.
11640 (init_rtl_slim_pretty_print): Likewise.
11641 (dump_value_slim): Don't call it. Use local pretty printer.
11642 (dump_insn_slim): Likewise.
11643 (dump_rtl_slim): Likewise.
11644 (str_pattern_slim): Likewise.
11645 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
11646 Simplify.
11647
11648 2013-08-16 Jakub Jelinek <jakub@redhat.com>
11649
11650 PR tree-optimization/58164
11651 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
11652 walk gimple_goto_dest of GIMPLE_GOTO.
11653
11654 PR tree-optimization/58165
11655 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
11656 bi_call must be the last stmt in a bb, don't split_block, instead
11657 use fallthru edge from it and give up if there is none.
11658 Release conds vector when returning early.
11659
11660 2013-08-14 Xinliang David Li <davidxl@google.com>
11661
11662 * config/i386/i386.c (ix86_option_override_internal):
11663 Remove unused variable and field.
11664
11665 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11666
11667 PR target/57949
11668 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
11669 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
11670 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
11671 and Linux, correct BLKmode alignment when 128-bit alignment is
11672 required and compatibility flag is not set.
11673 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
11674 for zero-size arguments when compatibility flag is not set.
11675
11676 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11677
11678 PR tree-optimization/58145
11679 * tree-sra.c (build_ref_for_offset): If prev_base has
11680 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
11681
11682 2013-08-14 Xinliang David Li <davidxl@google.com>
11683
11684 * config/i386/i386.c (ix86_option_override_internal):
11685 Fix uninitialized variable error.
11686
11687 2013-08-14 Xinliang David Li <davidxl@google.com>
11688
11689 * config/i386/i386.opt: Define two new options.
11690 * config/i386/x86-tune.def: Add arch selector field in macros.
11691 * config/i386/i386.h: Adjust macro definition.
11692 * config/i386/i386.c (ix86_option_override_internal):
11693 Refactor the code.
11694 (parse_mtune_ctrl_str): New function.
11695 (set_ix86_tune_features): New function.
11696 (ix86_function_specific_restore): Call the new helper function.
11697
11698 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
11699
11700 PR rtl-optimization/57662
11701 * sel-sched.c (code_motion_process_successors): When the current insn
11702 is removed after the recursive traversal, break from the loop.
11703 Add comments and debug printouts.
11704
11705 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11706 Alexandre Oliva <aoliva@redhat.com>
11707
11708 PR target/58067
11709 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
11710 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
11711 there also UNSPEC_PLTOFF.
11712
11713 2013-08-14 Marek Polacek <polacek@redhat.com>
11714
11715 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
11716 NUM_CONDITIONS bit positions.
11717
11718 2013-08-13 Cary Coutant <ccoutant@google.com>
11719
11720 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
11721 (attr_checksum): Hash vector contents instead of pointer.
11722 (attr_checksum_ordered): Likewise.
11723
11724 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
11725
11726 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
11727 when Pmode != word_mode. Add length_address attribute.
11728 (sse3_monitor_<mode>): Merge from sse3_monitor and
11729 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
11730 Pmode != word_mode. Update insn length attribute.
11731 * config/i386/i386.c (ix86_option_override_internal): Update
11732 ix86_gen_monitor selection for merged sse3_monitor insn.
11733
11734 2013-08-13 Julian Brown <julian@codesourcery.com>
11735
11736 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
11737 perform invalid legitimization on greater-than-word-size modes for
11738 TARGET_E500_DOUBLE.
11739
11740 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
11741
11742 * ira.c (setup_class_translate_array): Use aclass instead of cl
11743 for classes not fully covered by allocno classes.
11744
11745 2013-08-13 Jakub Jelinek <jakub@redhat.com>
11746
11747 PR tree-optimization/57661
11748 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
11749 * tree-inline.c (tree_function_versioning): Initialize it.
11750 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
11751 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
11752 that is not being copied.
11753
11754 PR sanitizer/56417
11755 * asan.c (instrument_strlen_call): Fix typo in comment.
11756 Use char * type even for the lhs of POINTER_PLUS_EXPR.
11757
11758 2013-08-13 Steve Ellcey <sellcey@mips.com>
11759
11760 * config/mips/mips.md (prefetch): Use lw instead of ld on
11761 loongson in 32bit mode.
11762
11763 2013-08-13 Nick Clifton <nickc@redhat.com>
11764
11765 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
11766
11767 2013-08-13 Jan Hubicka <jh@suse.cz>
11768
11769 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
11770 introduced edge; fix typo in sanity check.
11771 (cgraph_resolve_speculation): Export; improve diagnostic.
11772 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
11773 speculation at type mismatch.
11774 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
11775 (cgraph_resolve_speculation): Declare.
11776 (symtab_can_be_discarded): New function.
11777 * value-prof.c (gimple_ic_transform): Remove actual transform code.
11778 * ipa-inline-transform.c (speculation_removed): New global var.
11779 (clone_inlined_nodes): See if speculation can be removed.
11780 (inline_call): If speculations was removed, we growths may not match.
11781 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
11782 (speculation_useful_p): New function.
11783 (resolve_noninline_speculation): New function.
11784 (inline_small_functions): Resolve useless speculations.
11785 * ipa-inline.h (speculation_useful_p): Declare
11786 * ipa.c (can_replace_by_local_alias): Simplify.
11787 (ipa_profile): Produce speculative calls in non-lto, too;
11788 add simple cost model; produce local aliases.
11789
11790 2013-08-13 David Malcolm <dmalcolm@redhat.com>
11791
11792 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
11793 PASS_MANAGER_H.
11794
11795 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
11796
11797 * config/i386/i386.c (ix86_function_versions): Use error + inform.
11798
11799 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
11800
11801 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
11802 iterator instead of X87MODEF.
11803
11804 2013-08-12 Perez Read <netfirewall@gmail.com>
11805
11806 PR target/58132
11807 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
11808 operand 0 for intel asm alternative.
11809 (*movabs<mode>_2): Ditto for operand 1.
11810
11811 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
11812
11813 * config/aarch64/arm_none.h
11814 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
11815
11816 2013-08-12 Nick Clifton <nickc@redhat.com>
11817
11818 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
11819
11820 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
11821
11822 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
11823 expand for QI/HImode operand to produce more effictive code for
11824 unsigned char(short) --> float(double) conversion.
11825
11826 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
11827
11828 * doc/invoke.texi: Mention that -ftls-model does not force the final
11829 model.
11830
11831 2013-08-12 Marek Polacek <polacek@redhat.com>
11832 Marc Glisse <marc.glisse@inria.fr>
11833
11834 PR tree-optimization/57980
11835 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
11836 when creating -1 constant.
11837
11838 2013-08-10 Jan Hubicka <jh@suse.cz>
11839
11840 Workaround binutils PR14342.
11841 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
11842 (gimple_init_edge_profiler): Likewise.
11843 (gimple_gen_ic_func_profiler): Likewise.
11844
11845 2013-08-09 Jan Hubicka <jh@suse.cz>
11846
11847 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
11848
11849 2013-08-09 Xinliang David Li <davidxl@google.com>
11850
11851 * config/i386/stringop.def: New file.
11852 * config/i386/stringop.opt: New file.
11853 * config/i386/i386-opts.h: Include stringopt.def.
11854 * config/i386/i386.opt: Include stringopt.opt.
11855 * config/i386/i386.c (ix86_option_override_internal):
11856 Override default size based stringop inline strategies with options.
11857 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
11858 New function.
11859
11860 2013-08-09 Jan Hubicka <jh@suse.cz>
11861
11862 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
11863
11864 2013-08-09 Jan Hubicka <jh@suse.cz>
11865
11866 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
11867 CGRAPH_FREQ_MAX.
11868 (dump_cgraph_node): Dump profile-id.
11869 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
11870 and common_target_probability.
11871 * lto-cgraph.c (lto_output_edge): Stream common targets.
11872 (lto_output_node): Stream profile ids.
11873 (input_node): Stream profile ids.
11874 (input_edge): Stream common targets.
11875 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
11876 * ipa.c: Include value-prof.h
11877 (ipa_profile_generate_summary): Turn indirect call statement histograms
11878 into common targets.
11879 (ipa_profile): Turn common targets into speculative edges.
11880
11881 2013-08-09 Jan Hubicka <jh@suse.cz>
11882
11883 * cgraph.h (cgraph_node): Add profile_id.
11884 * value-prof.c (cgraph_node_map): Turn into pointer_map.
11885 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
11886 (del_node_map): Update.
11887 (find_func_by_funcdef_no): Replace by ...
11888 (find_func_by_profile_id): ... this one.
11889 (gimple_ic_transform): Do not remove useful histograms when
11890 speculation is not done; dump info when indirect call removal
11891 can happen at LTO.
11892 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
11893 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
11894 (__gcov_indirect_call_profiler_v2): .. this one.
11895 * profile.h (init_node_map): Update.
11896 * coverage.c (coverage_compute_profile_id): New function.
11897 * coverage.h (coverage_compute_profile_id): Declare.
11898 * tree-profile.c (init_ic_make_global_vars): Make
11899 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
11900 (gimple_init_edge_profiler): Update prototype of
11901 __gcov_indirect_call_profiler.
11902 (gimple_gen_ic_func_profiler): Simplify.
11903 (tree_profiling): Use init_node_map
11904
11905 2013-08-09 Jan Hubicka <jh@suse.cz>
11906
11907 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
11908 non-speculative refs.
11909 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
11910 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
11911 (cgraph_set_call_stmt): Likewise.
11912 (cgraph_create_edge_1): Fix release checking compilatoin;
11913 clear lto_stmt_uid.
11914 (cgraph_free_edge): Free indirect info.
11915 (cgraph_turn_edge_to_speculative): New function.
11916 (cgraph_speculative_call_info): New function.
11917 (cgraph_make_edge_direct): Return direct edge; handle speculation.
11918 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
11919 (dump_cgraph_node): Dump speculation.
11920 (verify_edge_count_and_frequency): Accept speculative edges.
11921 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
11922 (verify_cgraph_node): Handle speculation.
11923 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
11924 (cgraph_set_call_stmt): Update prototype.
11925 (cgraph_make_edge_direct): Update prototype.
11926 (cgraph_speculative_call_info): Declare.
11927 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
11928 to change; update call of ipa_find_references.
11929 * ipa-ref.c (ipa_record_reference): Fix return value; clear
11930 lto_stmt_uid and speculative flags.
11931 (ipa_dump_references): Dump speculation.
11932 (ipa_clone_references): Clone speculative flag.
11933 (ipa_clone_referring): Likewise.
11934 (ipa_clone_ref): New function.
11935 (ipa_find_reference): Look into lto_stmt_uids
11936 (ipa_clear_stmts_in_references): Do not clear speculative calls.
11937 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
11938 (ipa_find_reference): Update declaration.
11939 (ipa_clone_ref): Declare.
11940 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
11941 stream speculative flag.
11942 (lto_output_ref): Stream statements uids and speculation.
11943 (input_ref): Likewise.
11944 (input_edge): Stream speuclation.
11945 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
11946 (cgraph_set_call_stmt_including_clones): Handle speculation.
11947 * ipa-inline.c (heap_edge_removal_hook): New function.
11948 (inline_small_functions): Register it.
11949 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
11950 also initialize refs.
11951 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
11952 edge to change.
11953 (try_make_edge_direct_simple_call): Likewise.
11954 (try_make_edge_direct_simple_call): Likewise.
11955 (update_indirect_edges_after_inlining): Likewise.
11956 (remove_described_reference): Look proper lto_stmt_uid.
11957 (propagate_controlled_uses): Likewise.
11958 (propagate_controlled_uses): Liekwise.
11959 * tree-inline.c (copy_bb): Copy speculative edges.
11960 (redirect_all_calls): New function.
11961 (copy_cfg_body): Do redirection after loop info is updated.
11962 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
11963
11964 2013-08-09 Jan Hubicka <jh@suse.cz>
11965
11966 * lto-streamer-out.c (output_function): Renumber PHIs.
11967 * lto-streamer-in.c (input_function): Likewise.
11968
11969 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
11970
11971 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
11972 (get_lane_unsigned): Likewise.
11973 (dup_lane_scalar): Likewise.
11974 (get_lane): enable for VALL.
11975 * config/aarch64/aarch64-simd.md
11976 (aarch64_dup_lane_scalar<mode>): Remove.
11977 (aarch64_get_lane_signed<mode>): Likewise.
11978 (aarch64_get_lane_unsigned<mode>): Likewise.
11979 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
11980 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
11981 (aarch64_get_lane<mode>): Enable for all vector modes.
11982 (aarch64_get_lanedi): Remove misleading constraints.
11983 * config/aarch64/arm_neon.h
11984 (__aarch64_vget_lane_any): Define.
11985 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
11986 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
11987 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
11988 * config/aarch64/iterators.md (VDQQH): New.
11989 (VDQQHS): Likewise.
11990 (vwcore): Likewise.
11991
11992 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
11993
11994 * configure.ac: Add GAS check for LEON instructions on SPARC.
11995 * configure: Regenerate.
11996 * config.in: Likewise.
11997 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
11998 sparc*-*-* block.
11999 * config/sparc/sparc.opt (LEON, LEON3): New masks.
12000 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
12001 for LEON or LEON3.
12002 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
12003 (AS_LEON_FLAG): New macro.
12004 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
12005 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
12006 Deal with LEON and LEON3 for the memory model.
12007 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
12008 (atomic_compare_and_swap<mode>_1): Likewise.
12009 (*atomic_compare_and_swap<mode>_1): Likewise.
12010
12011 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12012
12013 * config/arm/neon.md (vcond): Fix floating-point vector
12014 comparisons against 0.
12015
12016 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
12017
12018 * lra-constraints.c (emit_spill_move): Remove assert.
12019 (process_alt_operands): Add more debugging
12020 output. Increase reject for spilling into memory. Decrease
12021 reject for reloading scratch.
12022 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
12023
12024 2013-08-08 Steve Ellcey <sellcey@mips.com>
12025
12026 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
12027 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
12028 micromips incompatible. Add nan2008.
12029 (MULTILIB_DIRNAMES): Add nan2008.
12030 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12031 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
12032 and micromips incompatible. Add nan2008.
12033 (MULTILIB_DIRNAMES): Add nan2008.
12034 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12035
12036 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
12037
12038 PR rtl-optimization/58079
12039 * combine.c (combine_simplify_rtx): Avoid using SUBST if
12040 simplify_comparison has widened a comparison with an integer.
12041
12042 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12043
12044 * config/arm/neon.md (movmisalign<mode>): Disable when we
12045 don't allow unaligned accesses.
12046 (*movmisalign<mode>_neon_store): Likewise.
12047 (*movmisalign<mode>_neon_load): Likewise.
12048 (*movmisalign<mode>_neon_store): Likewise.
12049 (*movmisalign<mode>_neon_load): Likewise.
12050
12051 2013-08-08 Jan Hubicka <jh@suse.cz>
12052
12053 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
12054 (make_pass_rebuild_cgraph_edges): Also clear references.
12055 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
12056 * ipa-inline-transform.c (inline_transform): Remove all references
12057 after inlining.
12058 * cgraphunit.c (expand_function): Remove all references after
12059 expansion.
12060 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
12061 (ipa_find_reference): Rewrite to iterator.
12062 (remove_stmt_references): Likewise.
12063 (ipa_clear_stmts_in_references): New function.
12064 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
12065 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
12066 clear references.
12067 * ipa-split.c (split_function): Remove references in split function.
12068
12069 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12070
12071 PR target/57431
12072 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
12073 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
12074
12075 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12076
12077 PR target/56979
12078 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
12079 suggested mode for the assignment isn't compatible with the
12080 registers required.
12081
12082 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12083
12084 PR target/58065
12085 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
12086
12087 2013-08-07 Xinliang David Li <davidxl@google.com>
12088
12089 * config/i386/i386.opt: New option -mtune-ctrl=.
12090 * config/i386/x86-tune.def: New file.
12091 * config/i386/i386.h: include x86-tune.def.
12092 * config/i386/i386.c (ix86_option_override_internal):
12093 Parsing -mtune-ctrl= option and set tune features.
12094
12095 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
12096
12097 PR other/12081
12098 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
12099 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
12100 to gen_2arg_fn_t.
12101
12102 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12103
12104 * rtl.h (update_alignments): Declare.
12105 * final.c (grow_label_align): New function extracted from...
12106 (shorten_branches): ...here. Call it.
12107 (update_alignments): New function.
12108 * reorg.c (sibling_labels): New variable.
12109 (get_label_before): Add SIBLING parameter. If it is non-zero, push
12110 the new label along with it onto the sibling_labels vector.
12111 (fill_simple_delay_slots): Adjust call to get_label_before.
12112 (fill_slots_from_thread): Likewise.
12113 (relax_delay_slots): Likewise.
12114 (make_return_insns): Likewise.
12115 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
12116
12117 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12118
12119 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
12120 document its semantics.
12121 (diagnostic_report_diagnostic): Adjust accordingly.
12122
12123 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12124
12125 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
12126 (sparc_option_override): ...and port to new C++ pass API.
12127 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
12128
12129 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
12130
12131 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
12132
12133 2013-08-06 Caroline Tice <cmtice@google.com>
12134
12135 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
12136 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
12137 * tree-pass.h: Add pass_vtable_verify.
12138 * varasm.c (assemble_variable): Add code to properly set the comdat
12139 section and name for the .vtable_map_vars section.
12140 (assemble_vtyv_preinit_initializer): New function.
12141 (default_sectin_type_flags): Make sure .vtable_map_vars section has
12142 LINK_ONCE flag.
12143 * output.h: Add function decl for assemble_vtv_preinit_initializer.
12144 * vtable-verify.c: New file.
12145 * vtable-verify.h: New file.
12146 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
12147 initialiation levels.
12148 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
12149 * passes.def: Insert pass_vtable_verify.
12150 * aclocal.m4: Reorder includes.
12151 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
12152 -fvtv-counts options.
12153 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
12154 as appropriate, if -fvtable-verify=... is used.
12155 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
12156 -fvtable-verify=... is used.
12157 * Makefile.in (OBJS): Add vtable-verify.o to list.
12158 (vtable-verify.o): Add new build rule.
12159 (GTFILES): Add vtable-verify.c to list.
12160 * common.opt (fvtable-verify=): New flag.
12161 (vtv_priority): Values for fvtable-verify= flag.
12162 (fvtv-counts): New flag.
12163 (fvtv-debug): New flag.
12164 * tree.h (save_vtable_map_decl): New extern function decl.
12165
12166 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12167
12168 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
12169 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
12170 (pass_data_rl78_devirt): ...new pass_data instance and...
12171 (make_pass_rl78_devirt): ...new function.
12172 (rl78_asm_file_start): Port pass registration to new C++ API.
12173
12174 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12175
12176 * coretypes.h (rtl_opt_pass): Add.
12177 (gcc::context): Add.
12178 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
12179 (epiphany_init): Port to new C++ pass API.
12180 (epiphany_optimize_mode_switching): Likewise.
12181 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
12182 (pass_manager::get_pass_mode_switching): New.
12183 (pass_manager::get_pass_peephole2): New.
12184 * mode-switching.c (pass_mode_switching): Add clone method.
12185 * recog.c (pass_peephole2): Add clone method.
12186 (pass_split_all_insns): Add clone method.
12187
12188 2013-08-06 David Malcolm <dmalcolm@redhat.com>
12189
12190 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
12191 (mips_option_override): ...here, porting to new C++ API for passes.
12192
12193 2013-08-06 Jan Hubicka <jh@suse.cz>
12194
12195 * cgraph.c (cgraph_get_body): New function based on lto.c
12196 implementation.
12197 * cgraph.h (cgraph_get_body): Declare.
12198 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
12199 LTO paths.
12200 * cgraphunit.c (expand_function): Get body prior expanding.
12201 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
12202 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
12203 really need.
12204 * passes.c (do_per_function_toporder): Get body.
12205 * tree-inline.c (expand_call_inline): Get body prior inlining it.
12206 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
12207
12208 2013-08-06 Martin Jambor <mjambor@suse.cz>
12209
12210 PR fortran/57987
12211 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
12212 is not re-finalized. Rename second parameter to no_collect.
12213
12214 2013-08-06 Martin Jambor <mjambor@suse.cz>
12215
12216 PR middle-end/58041
12217 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
12218 MEM_REF has proper alignment information.
12219
12220 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
12221
12222 PR other/12081
12223 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
12224 class insn_gen_fn.
12225 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
12226 rtx (*) (rtx, ...) with insn_gen_fn.
12227 * genoutput.c (output_insn_data): Cast gen_? function pointers to
12228 insn_gen_fn::stored_funcptr. Add initializer braces.
12229
12230 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12231
12232 Rewrite how instances of passes are cloned to remove assumptions
12233 about their sizes (thus allowing pass subclasses to have
12234 additional data fields, albeit non-GC-managed ones at this point).
12235
12236 * passes.c (make_pass_instance): Now that passes have clone
12237 methods, rewrite this function to eliminate XNEW and memcpy
12238 calls that used hardcoded sizes. Since this function no longer
12239 creates pass instances, rename it to...
12240 (add_pass_instance): ...this. Document the old way that passes were
12241 numbered and flagged, and rework this function to continue using it.
12242 (next_pass_1): Add an initial_pass argument for use by
12243 add_pass_instance.
12244 (position_pass): When adding multiple instances of a pass, use
12245 the pass's clone method, rather than relying on the XNEW/memcpy
12246 within the former make_pass_instance (now add_pass_instance).
12247 (pass_manager::pass_manager): When invoking next_pass_1, also supply
12248 the initial instance of the current pass within the pass manager.
12249
12250 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12251
12252 This is the automated part of the conversion of passes from C
12253 structs to C++ classes.
12254
12255 Patch autogenerated by refactor_passes.py from
12256 https://github.com/davidmalcolm/gcc-refactoring-scripts
12257 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
12258
12259 * asan.c (pass_asan): Convert from a global struct to a subclass of
12260 gimple_opt_pass along with...
12261 (pass_data_asan): ...new pass_data instance and...
12262 (make_pass_asan): ...new function.
12263 (pass_asan_O0): Convert from a global struct to a subclass of
12264 gimple_opt_pass along with...
12265 (pass_data_asan_O0): ...new pass_data instance and...
12266 (make_pass_asan_O0): ...new function.
12267 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
12268 subclass of rtl_opt_pass along with...
12269 (pass_data_inc_dec): ...new pass_data instance and...
12270 (make_pass_inc_dec): ...new function.
12271 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
12272 a subclass of rtl_opt_pass along with...
12273 (pass_data_reorder_blocks): ...new pass_data instance and...
12274 (make_pass_reorder_blocks): ...new function.
12275 (pass_duplicate_computed_gotos): Convert from a global struct to a
12276 subclass of rtl_opt_pass along with...
12277 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
12278 (make_pass_duplicate_computed_gotos): ...new function.
12279 (pass_partition_blocks): Convert from a global struct to a subclass of
12280 rtl_opt_pass along with...
12281 (pass_data_partition_blocks): ...new pass_data instance and...
12282 (make_pass_partition_blocks): ...new function.
12283 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
12284 struct to a subclass of rtl_opt_pass along with...
12285 (pass_data_branch_target_load_optimize1): ...new pass_data instance
12286 and...
12287 (make_pass_branch_target_load_optimize1): ...new function.
12288 (pass_branch_target_load_optimize2): Convert from a global struct to a
12289 subclass of rtl_opt_pass along with...
12290 (pass_data_branch_target_load_optimize2): ...new pass_data instance
12291 and...
12292 (make_pass_branch_target_load_optimize2): ...new function.
12293 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
12294 of rtl_opt_pass along with...
12295 (pass_data_jump): ...new pass_data instance and...
12296 (make_pass_jump): ...new function.
12297 (pass_jump2): Convert from a global struct to a subclass of
12298 rtl_opt_pass along with...
12299 (pass_data_jump2): ...new pass_data instance and...
12300 (make_pass_jump2): ...new function.
12301 * cfgexpand.c (pass_expand): Convert from a global struct to a
12302 subclass of rtl_opt_pass along with...
12303 (pass_data_expand): ...new pass_data instance and...
12304 (make_pass_expand): ...new function.
12305 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
12306 of rtl_opt_pass along with...
12307 (pass_data_free_cfg): ...new pass_data instance and...
12308 (make_pass_free_cfg): ...new function.
12309 (pass_into_cfg_layout_mode): Convert from a global struct to a
12310 subclass of rtl_opt_pass along with...
12311 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
12312 (make_pass_into_cfg_layout_mode): ...new function.
12313 (pass_outof_cfg_layout_mode): Convert from a global struct to a
12314 subclass of rtl_opt_pass along with...
12315 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
12316 (make_pass_outof_cfg_layout_mode): ...new function.
12317 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
12318 struct to a subclass of gimple_opt_pass along with...
12319 (pass_data_build_cgraph_edges): ...new pass_data instance and...
12320 (make_pass_build_cgraph_edges): ...new function.
12321 (pass_rebuild_cgraph_edges): Convert from a global struct to a
12322 subclass of gimple_opt_pass along with...
12323 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
12324 (make_pass_rebuild_cgraph_edges): ...new function.
12325 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
12326 subclass of gimple_opt_pass along with...
12327 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
12328 and...
12329 (make_pass_remove_cgraph_callee_edges): ...new function.
12330 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
12331 struct to a subclass of rtl_opt_pass along with...
12332 (pass_data_stack_adjustments): ...new pass_data instance and...
12333 (make_pass_stack_adjustments): ...new function.
12334 * combine.c (pass_combine): Convert from a global struct to a subclass
12335 of rtl_opt_pass along with...
12336 (pass_data_combine): ...new pass_data instance and...
12337 (make_pass_combine): ...new function.
12338 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
12339 global struct to a subclass of rtl_opt_pass along with...
12340 (pass_data_compare_elim_after_reload): ...new pass_data instance
12341 and...
12342 (make_pass_compare_elim_after_reload): ...new function.
12343 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
12344 of rtl_opt_pass along with...
12345 (pass_data_rtl_cprop): ...new pass_data instance and...
12346 (make_pass_rtl_cprop): ...new function.
12347 * cse.c (pass_cse): Convert from a global struct to a subclass of
12348 rtl_opt_pass along with...
12349 (pass_data_cse): ...new pass_data instance and...
12350 (make_pass_cse): ...new function.
12351 (pass_cse2): Convert from a global struct to a subclass of
12352 rtl_opt_pass along with...
12353 (pass_data_cse2): ...new pass_data instance and...
12354 (make_pass_cse2): ...new function.
12355 (pass_cse_after_global_opts): Convert from a global struct to a
12356 subclass of rtl_opt_pass along with...
12357 (pass_data_cse_after_global_opts): ...new pass_data instance and...
12358 (make_pass_cse_after_global_opts): ...new function.
12359 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
12360 of rtl_opt_pass along with...
12361 (pass_data_ud_rtl_dce): ...new pass_data instance and...
12362 (make_pass_ud_rtl_dce): ...new function.
12363 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
12364 rtl_opt_pass along with...
12365 (pass_data_fast_rtl_dce): ...new pass_data instance and...
12366 (make_pass_fast_rtl_dce): ...new function.
12367 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
12368 a subclass of rtl_opt_pass along with...
12369 (pass_data_df_initialize_opt): ...new pass_data instance and...
12370 (make_pass_df_initialize_opt): ...new function.
12371 (pass_df_initialize_no_opt): Convert from a global struct to a
12372 subclass of rtl_opt_pass along with...
12373 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
12374 (make_pass_df_initialize_no_opt): ...new function.
12375 (pass_df_finish): Convert from a global struct to a subclass of
12376 rtl_opt_pass along with...
12377 (pass_data_df_finish): ...new pass_data instance and...
12378 (make_pass_df_finish): ...new function.
12379 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
12380 rtl_opt_pass along with...
12381 (pass_data_rtl_dse1): ...new pass_data instance and...
12382 (make_pass_rtl_dse1): ...new function.
12383 (pass_rtl_dse2): Convert from a global struct to a subclass of
12384 rtl_opt_pass along with...
12385 (pass_data_rtl_dse2): ...new pass_data instance and...
12386 (make_pass_rtl_dse2): ...new function.
12387 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
12388 subclass of rtl_opt_pass along with...
12389 (pass_data_dwarf2_frame): ...new pass_data instance and...
12390 (make_pass_dwarf2_frame): ...new function.
12391 * except.c (pass_set_nothrow_function_flags): Convert from a global
12392 struct to a subclass of rtl_opt_pass along with...
12393 (pass_data_set_nothrow_function_flags): ...new pass_data instance
12394 and...
12395 (make_pass_set_nothrow_function_flags): ...new function.
12396 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
12397 subclass of rtl_opt_pass along with...
12398 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
12399 and...
12400 (make_pass_convert_to_eh_region_ranges): ...new function.
12401 * final.c (pass_compute_alignments): Convert from a global struct to a
12402 subclass of rtl_opt_pass along with...
12403 (pass_data_compute_alignments): ...new pass_data instance and...
12404 (make_pass_compute_alignments): ...new function.
12405 (pass_final): Convert from a global struct to a subclass of
12406 rtl_opt_pass along with...
12407 (pass_data_final): ...new pass_data instance and...
12408 (make_pass_final): ...new function.
12409 (pass_shorten_branches): Convert from a global struct to a subclass of
12410 rtl_opt_pass along with...
12411 (pass_data_shorten_branches): ...new pass_data instance and...
12412 (make_pass_shorten_branches): ...new function.
12413 (pass_clean_state): Convert from a global struct to a subclass of
12414 rtl_opt_pass along with...
12415 (pass_data_clean_state): ...new pass_data instance and...
12416 (make_pass_clean_state): ...new function.
12417 * function.c (pass_instantiate_virtual_regs): Convert from a global
12418 struct to a subclass of rtl_opt_pass along with...
12419 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
12420 (make_pass_instantiate_virtual_regs): ...new function.
12421 (pass_leaf_regs): Convert from a global struct to a subclass of
12422 rtl_opt_pass along with...
12423 (pass_data_leaf_regs): ...new pass_data instance and...
12424 (make_pass_leaf_regs): ...new function.
12425 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
12426 subclass of rtl_opt_pass along with...
12427 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
12428 and...
12429 (make_pass_thread_prologue_and_epilogue): ...new function.
12430 (pass_match_asm_constraints): Convert from a global struct to a
12431 subclass of rtl_opt_pass along with...
12432 (pass_data_match_asm_constraints): ...new pass_data instance and...
12433 (make_pass_match_asm_constraints): ...new function.
12434 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
12435 subclass of rtl_opt_pass along with...
12436 (pass_data_rtl_fwprop): ...new pass_data instance and...
12437 (make_pass_rtl_fwprop): ...new function.
12438 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
12439 rtl_opt_pass along with...
12440 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
12441 (make_pass_rtl_fwprop_addr): ...new function.
12442 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
12443 rtl_opt_pass along with...
12444 (pass_data_rtl_pre): ...new pass_data instance and...
12445 (make_pass_rtl_pre): ...new function.
12446 (pass_rtl_hoist): Convert from a global struct to a subclass of
12447 rtl_opt_pass along with...
12448 (pass_data_rtl_hoist): ...new pass_data instance and...
12449 (make_pass_rtl_hoist): ...new function.
12450 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
12451 subclass of gimple_opt_pass along with...
12452 (pass_data_lower_cf): ...new pass_data instance and...
12453 (make_pass_lower_cf): ...new function.
12454 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
12455 from a global struct to a subclass of gimple_opt_pass along with...
12456 (pass_data_strength_reduction): ...new pass_data instance and...
12457 (make_pass_strength_reduction): ...new function.
12458 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
12459 of rtl_opt_pass along with...
12460 (pass_data_rtl_ifcvt): ...new pass_data instance and...
12461 (make_pass_rtl_ifcvt): ...new function.
12462 (pass_if_after_combine): Convert from a global struct to a subclass of
12463 rtl_opt_pass along with...
12464 (pass_data_if_after_combine): ...new pass_data instance and...
12465 (make_pass_if_after_combine): ...new function.
12466 (pass_if_after_reload): Convert from a global struct to a subclass of
12467 rtl_opt_pass along with...
12468 (pass_data_if_after_reload): ...new pass_data instance and...
12469 (make_pass_if_after_reload): ...new function.
12470 * init-regs.c (pass_initialize_regs): Convert from a global struct to
12471 a subclass of rtl_opt_pass along with...
12472 (pass_data_initialize_regs): ...new pass_data instance and...
12473 (make_pass_initialize_regs): ...new function.
12474 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
12475 of ipa_opt_pass_d along with...
12476 (pass_data_ipa_cp): ...new pass_data instance and...
12477 (make_pass_ipa_cp): ...new function.
12478 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
12479 global struct to a subclass of gimple_opt_pass along with...
12480 (pass_data_inline_parameters): ...new pass_data instance and...
12481 (make_pass_inline_parameters): ...new function.
12482 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
12483 subclass of gimple_opt_pass along with...
12484 (pass_data_early_inline): ...new pass_data instance and...
12485 (make_pass_early_inline): ...new function.
12486 (pass_ipa_inline): Convert from a global struct to a subclass of
12487 ipa_opt_pass_d along with...
12488 (pass_data_ipa_inline): ...new pass_data instance and...
12489 (make_pass_ipa_inline): ...new function.
12490 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
12491 struct to a subclass of gimple_opt_pass along with...
12492 (pass_data_local_pure_const): ...new pass_data instance and...
12493 (make_pass_local_pure_const): ...new function.
12494 (pass_ipa_pure_const): Convert from a global struct to a subclass of
12495 ipa_opt_pass_d along with...
12496 (pass_data_ipa_pure_const): ...new pass_data instance and...
12497 (make_pass_ipa_pure_const): ...new function.
12498 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
12499 to a subclass of ipa_opt_pass_d along with...
12500 (pass_data_ipa_reference): ...new pass_data instance and...
12501 (make_pass_ipa_reference): ...new function.
12502 * ipa-split.c (pass_split_functions): Convert from a global struct to
12503 a subclass of gimple_opt_pass along with...
12504 (pass_data_split_functions): ...new pass_data instance and...
12505 (make_pass_split_functions): ...new function.
12506 (pass_feedback_split_functions): Convert from a global struct to a
12507 subclass of gimple_opt_pass along with...
12508 (pass_data_feedback_split_functions): ...new pass_data instance and...
12509 (make_pass_feedback_split_functions): ...new function.
12510 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
12511 global struct to a subclass of simple_ipa_opt_pass along with...
12512 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
12513 instance and...
12514 (make_pass_ipa_function_and_variable_visibility): ...new function.
12515 (pass_ipa_free_inline_summary): Convert from a global struct to a
12516 subclass of simple_ipa_opt_pass along with...
12517 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
12518 (make_pass_ipa_free_inline_summary): ...new function.
12519 (pass_ipa_whole_program_visibility): Convert from a global struct to a
12520 subclass of ipa_opt_pass_d along with...
12521 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
12522 and...
12523 (make_pass_ipa_whole_program_visibility): ...new function.
12524 (pass_ipa_profile): Convert from a global struct to a subclass of
12525 ipa_opt_pass_d along with...
12526 (pass_data_ipa_profile): ...new pass_data instance and...
12527 (make_pass_ipa_profile): ...new function.
12528 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
12529 ipa_opt_pass_d along with...
12530 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
12531 (make_pass_ipa_cdtor_merge): ...new function.
12532 * ira.c (pass_ira): Convert from a global struct to a subclass of
12533 rtl_opt_pass along with...
12534 (pass_data_ira): ...new pass_data instance and...
12535 (make_pass_ira): ...new function.
12536 (pass_reload): Convert from a global struct to a subclass of
12537 rtl_opt_pass along with...
12538 (pass_data_reload): ...new pass_data instance and...
12539 (make_pass_reload): ...new function.
12540 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
12541 subclass of rtl_opt_pass along with...
12542 (pass_data_cleanup_barriers): ...new pass_data instance and...
12543 (make_pass_cleanup_barriers): ...new function.
12544 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
12545 of rtl_opt_pass along with...
12546 (pass_data_loop2): ...new pass_data instance and...
12547 (make_pass_loop2): ...new function.
12548 (pass_rtl_loop_init): Convert from a global struct to a subclass of
12549 rtl_opt_pass along with...
12550 (pass_data_rtl_loop_init): ...new pass_data instance and...
12551 (make_pass_rtl_loop_init): ...new function.
12552 (pass_rtl_loop_done): Convert from a global struct to a subclass of
12553 rtl_opt_pass along with...
12554 (pass_data_rtl_loop_done): ...new pass_data instance and...
12555 (make_pass_rtl_loop_done): ...new function.
12556 (pass_rtl_move_loop_invariants): Convert from a global struct to a
12557 subclass of rtl_opt_pass along with...
12558 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
12559 (make_pass_rtl_move_loop_invariants): ...new function.
12560 (pass_rtl_unswitch): Convert from a global struct to a subclass of
12561 rtl_opt_pass along with...
12562 (pass_data_rtl_unswitch): ...new pass_data instance and...
12563 (make_pass_rtl_unswitch): ...new function.
12564 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
12565 subclass of rtl_opt_pass along with...
12566 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
12567 and...
12568 (make_pass_rtl_unroll_and_peel_loops): ...new function.
12569 (pass_rtl_doloop): Convert from a global struct to a subclass of
12570 rtl_opt_pass along with...
12571 (pass_data_rtl_doloop): ...new pass_data instance and...
12572 (make_pass_rtl_doloop): ...new function.
12573 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
12574 a subclass of rtl_opt_pass along with...
12575 (pass_data_lower_subreg): ...new pass_data instance and...
12576 (make_pass_lower_subreg): ...new function.
12577 (pass_lower_subreg2): Convert from a global struct to a subclass of
12578 rtl_opt_pass along with...
12579 (pass_data_lower_subreg2): ...new pass_data instance and...
12580 (make_pass_lower_subreg2): ...new function.
12581 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
12582 struct to a subclass of ipa_opt_pass_d along with...
12583 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
12584 (make_pass_ipa_lto_gimple_out): ...new function.
12585 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
12586 of ipa_opt_pass_d along with...
12587 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
12588 (make_pass_ipa_lto_finish_out): ...new function.
12589 * mode-switching.c (pass_mode_switching): Convert from a global struct
12590 to a subclass of rtl_opt_pass along with...
12591 (pass_data_mode_switching): ...new pass_data instance and...
12592 (make_pass_mode_switching): ...new function.
12593 * modulo-sched.c (pass_sms): Convert from a global struct to a
12594 subclass of rtl_opt_pass along with...
12595 (pass_data_sms): ...new pass_data instance and...
12596 (make_pass_sms): ...new function.
12597 * omp-low.c (pass_expand_omp): Convert from a global struct to a
12598 subclass of gimple_opt_pass along with...
12599 (pass_data_expand_omp): ...new pass_data instance and...
12600 (make_pass_expand_omp): ...new function.
12601 (pass_lower_omp): Convert from a global struct to a subclass of
12602 gimple_opt_pass along with...
12603 (pass_data_lower_omp): ...new pass_data instance and...
12604 (make_pass_lower_omp): ...new function.
12605 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
12606 of gimple_opt_pass along with...
12607 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
12608 (make_pass_diagnose_omp_blocks): ...new function.
12609 * passes.c (pass_early_local_passes): Convert from a global struct to
12610 a subclass of simple_ipa_opt_pass along with...
12611 (pass_data_early_local_passes): ...new pass_data instance and...
12612 (make_pass_early_local_passes): ...new function.
12613 (pass_all_early_optimizations): Convert from a global struct to a
12614 subclass of gimple_opt_pass along with...
12615 (pass_data_all_early_optimizations): ...new pass_data instance and...
12616 (make_pass_all_early_optimizations): ...new function.
12617 (pass_all_optimizations): Convert from a global struct to a subclass
12618 of gimple_opt_pass along with...
12619 (pass_data_all_optimizations): ...new pass_data instance and...
12620 (make_pass_all_optimizations): ...new function.
12621 (pass_all_optimizations_g): Convert from a global struct to a subclass
12622 of gimple_opt_pass along with...
12623 (pass_data_all_optimizations_g): ...new pass_data instance and...
12624 (make_pass_all_optimizations_g): ...new function.
12625 (pass_rest_of_compilation): Convert from a global struct to a subclass
12626 of rtl_opt_pass along with...
12627 (pass_data_rest_of_compilation): ...new pass_data instance and...
12628 (make_pass_rest_of_compilation): ...new function.
12629 (pass_postreload): Convert from a global struct to a subclass of
12630 rtl_opt_pass along with...
12631 (pass_data_postreload): ...new pass_data instance and...
12632 (make_pass_postreload): ...new function.
12633 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
12634 subclass of rtl_opt_pass along with...
12635 (pass_data_gcse2): ...new pass_data instance and...
12636 (make_pass_gcse2): ...new function.
12637 * postreload.c (pass_postreload_cse): Convert from a global struct to
12638 a subclass of rtl_opt_pass along with...
12639 (pass_data_postreload_cse): ...new pass_data instance and...
12640 (make_pass_postreload_cse): ...new function.
12641 * predict.c (pass_profile): Convert from a global struct to a subclass
12642 of gimple_opt_pass along with...
12643 (pass_data_profile): ...new pass_data instance and...
12644 (make_pass_profile): ...new function.
12645 (pass_strip_predict_hints): Convert from a global struct to a subclass
12646 of gimple_opt_pass along with...
12647 (pass_data_strip_predict_hints): ...new pass_data instance and...
12648 (make_pass_strip_predict_hints): ...new function.
12649 * recog.c (pass_peephole2): Convert from a global struct to a subclass
12650 of rtl_opt_pass along with...
12651 (pass_data_peephole2): ...new pass_data instance and...
12652 (make_pass_peephole2): ...new function.
12653 (pass_split_all_insns): Convert from a global struct to a subclass of
12654 rtl_opt_pass along with...
12655 (pass_data_split_all_insns): ...new pass_data instance and...
12656 (make_pass_split_all_insns): ...new function.
12657 (pass_split_after_reload): Convert from a global struct to a subclass
12658 of rtl_opt_pass along with...
12659 (pass_data_split_after_reload): ...new pass_data instance and...
12660 (make_pass_split_after_reload): ...new function.
12661 (pass_split_before_regstack): Convert from a global struct to a
12662 subclass of rtl_opt_pass along with...
12663 (pass_data_split_before_regstack): ...new pass_data instance and...
12664 (make_pass_split_before_regstack): ...new function.
12665 (pass_split_before_sched2): Convert from a global struct to a subclass
12666 of rtl_opt_pass along with...
12667 (pass_data_split_before_sched2): ...new pass_data instance and...
12668 (make_pass_split_before_sched2): ...new function.
12669 (pass_split_for_shorten_branches): Convert from a global struct to a
12670 subclass of rtl_opt_pass along with...
12671 (pass_data_split_for_shorten_branches): ...new pass_data instance
12672 and...
12673 (make_pass_split_for_shorten_branches): ...new function.
12674 * ree.c (pass_ree): Convert from a global struct to a subclass of
12675 rtl_opt_pass along with...
12676 (pass_data_ree): ...new pass_data instance and...
12677 (make_pass_ree): ...new function.
12678 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
12679 subclass of rtl_opt_pass along with...
12680 (pass_data_stack_regs): ...new pass_data instance and...
12681 (make_pass_stack_regs): ...new function.
12682 (pass_stack_regs_run): Convert from a global struct to a subclass of
12683 rtl_opt_pass along with...
12684 (pass_data_stack_regs_run): ...new pass_data instance and...
12685 (make_pass_stack_regs_run): ...new function.
12686 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
12687 subclass of rtl_opt_pass along with...
12688 (pass_data_cprop_hardreg): ...new pass_data instance and...
12689 (make_pass_cprop_hardreg): ...new function.
12690 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
12691 subclass of rtl_opt_pass along with...
12692 (pass_data_reginfo_init): ...new pass_data instance and...
12693 (make_pass_reginfo_init): ...new function.
12694 * regmove.c (pass_regmove): Convert from a global struct to a subclass
12695 of rtl_opt_pass along with...
12696 (pass_data_regmove): ...new pass_data instance and...
12697 (make_pass_regmove): ...new function.
12698 * regrename.c (pass_regrename): Convert from a global struct to a
12699 subclass of rtl_opt_pass along with...
12700 (pass_data_regrename): ...new pass_data instance and...
12701 (make_pass_regrename): ...new function.
12702 * reorg.c (pass_delay_slots): Convert from a global struct to a
12703 subclass of rtl_opt_pass along with...
12704 (pass_data_delay_slots): ...new pass_data instance and...
12705 (make_pass_delay_slots): ...new function.
12706 (pass_machine_reorg): Convert from a global struct to a subclass of
12707 rtl_opt_pass along with...
12708 (pass_data_machine_reorg): ...new pass_data instance and...
12709 (make_pass_machine_reorg): ...new function.
12710 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
12711 of rtl_opt_pass along with...
12712 (pass_data_sched): ...new pass_data instance and...
12713 (make_pass_sched): ...new function.
12714 (pass_sched2): Convert from a global struct to a subclass of
12715 rtl_opt_pass along with...
12716 (pass_data_sched2): ...new pass_data instance and...
12717 (make_pass_sched2): ...new function.
12718 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
12719 to a subclass of rtl_opt_pass along with...
12720 (pass_data_stack_ptr_mod): ...new pass_data instance and...
12721 (make_pass_stack_ptr_mod): ...new function.
12722 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
12723 to a subclass of rtl_opt_pass along with...
12724 (pass_data_rtl_store_motion): ...new pass_data instance and...
12725 (make_pass_rtl_store_motion): ...new function.
12726 * tracer.c (pass_tracer): Convert from a global struct to a subclass
12727 of gimple_opt_pass along with...
12728 (pass_data_tracer): ...new pass_data instance and...
12729 (make_pass_tracer): ...new function.
12730 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
12731 to a subclass of gimple_opt_pass along with...
12732 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
12733 (make_pass_diagnose_tm_blocks): ...new function.
12734 (pass_lower_tm): Convert from a global struct to a subclass of
12735 gimple_opt_pass along with...
12736 (pass_data_lower_tm): ...new pass_data instance and...
12737 (make_pass_lower_tm): ...new function.
12738 (pass_tm_init): Convert from a global struct to a subclass of
12739 gimple_opt_pass along with...
12740 (pass_data_tm_init): ...new pass_data instance and...
12741 (make_pass_tm_init): ...new function.
12742 (pass_tm_mark): Convert from a global struct to a subclass of
12743 gimple_opt_pass along with...
12744 (pass_data_tm_mark): ...new pass_data instance and...
12745 (make_pass_tm_mark): ...new function.
12746 (pass_tm_edges): Convert from a global struct to a subclass of
12747 gimple_opt_pass along with...
12748 (pass_data_tm_edges): ...new pass_data instance and...
12749 (make_pass_tm_edges): ...new function.
12750 (pass_tm_memopt): Convert from a global struct to a subclass of
12751 gimple_opt_pass along with...
12752 (pass_data_tm_memopt): ...new pass_data instance and...
12753 (make_pass_tm_memopt): ...new function.
12754 (pass_ipa_tm): Convert from a global struct to a subclass of
12755 simple_ipa_opt_pass along with...
12756 (pass_data_ipa_tm): ...new pass_data instance and...
12757 (make_pass_ipa_tm): ...new function.
12758 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
12759 subclass of gimple_opt_pass along with...
12760 (pass_data_call_cdce): ...new pass_data instance and...
12761 (make_pass_call_cdce): ...new function.
12762 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
12763 subclass of gimple_opt_pass along with...
12764 (pass_data_build_cfg): ...new pass_data instance and...
12765 (make_pass_build_cfg): ...new function.
12766 (pass_split_crit_edges): Convert from a global struct to a subclass of
12767 gimple_opt_pass along with...
12768 (pass_data_split_crit_edges): ...new pass_data instance and...
12769 (make_pass_split_crit_edges): ...new function.
12770 (pass_warn_function_return): Convert from a global struct to a
12771 subclass of gimple_opt_pass along with...
12772 (pass_data_warn_function_return): ...new pass_data instance and...
12773 (make_pass_warn_function_return): ...new function.
12774 (pass_warn_function_noreturn): Convert from a global struct to a
12775 subclass of gimple_opt_pass along with...
12776 (pass_data_warn_function_noreturn): ...new pass_data instance and...
12777 (make_pass_warn_function_noreturn): ...new function.
12778 (pass_warn_unused_result): Convert from a global struct to a subclass
12779 of gimple_opt_pass along with...
12780 (pass_data_warn_unused_result): ...new pass_data instance and...
12781 (make_pass_warn_unused_result): ...new function.
12782 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
12783 a subclass of gimple_opt_pass along with...
12784 (pass_data_merge_phi): ...new pass_data instance and...
12785 (make_pass_merge_phi): ...new function.
12786 * tree-complex.c (pass_lower_complex): Convert from a global struct to
12787 a subclass of gimple_opt_pass along with...
12788 (pass_data_lower_complex): ...new pass_data instance and...
12789 (make_pass_lower_complex): ...new function.
12790 (pass_lower_complex_O0): Convert from a global struct to a subclass of
12791 gimple_opt_pass along with...
12792 (pass_data_lower_complex_O0): ...new pass_data instance and...
12793 (make_pass_lower_complex_O0): ...new function.
12794 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
12795 subclass of gimple_opt_pass along with...
12796 (pass_data_lower_eh): ...new pass_data instance and...
12797 (make_pass_lower_eh): ...new function.
12798 (pass_refactor_eh): Convert from a global struct to a subclass of
12799 gimple_opt_pass along with...
12800 (pass_data_refactor_eh): ...new pass_data instance and...
12801 (make_pass_refactor_eh): ...new function.
12802 (pass_lower_resx): Convert from a global struct to a subclass of
12803 gimple_opt_pass along with...
12804 (pass_data_lower_resx): ...new pass_data instance and...
12805 (make_pass_lower_resx): ...new function.
12806 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
12807 of gimple_opt_pass along with...
12808 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
12809 (make_pass_lower_eh_dispatch): ...new function.
12810 (pass_cleanup_eh): Convert from a global struct to a subclass of
12811 gimple_opt_pass along with...
12812 (pass_data_cleanup_eh): ...new pass_data instance and...
12813 (make_pass_cleanup_eh): ...new function.
12814 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
12815 to a subclass of simple_ipa_opt_pass along with...
12816 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
12817 (make_pass_ipa_lower_emutls): ...new function.
12818 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
12819 a subclass of gimple_opt_pass along with...
12820 (pass_data_if_conversion): ...new pass_data instance and...
12821 (make_pass_if_conversion): ...new function.
12822 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
12823 subclass of gimple_opt_pass along with...
12824 (pass_data_build_ssa): ...new pass_data instance and...
12825 (make_pass_build_ssa): ...new function.
12826 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
12827 global struct to a subclass of gimple_opt_pass along with...
12828 (pass_data_loop_distribution): ...new pass_data instance and...
12829 (make_pass_loop_distribution): ...new function.
12830 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
12831 subclass of gimple_opt_pass along with...
12832 (pass_data_mudflap_1): ...new pass_data instance and...
12833 (make_pass_mudflap_1): ...new function.
12834 (pass_mudflap_2): Convert from a global struct to a subclass of
12835 gimple_opt_pass along with...
12836 (pass_data_mudflap_2): ...new pass_data instance and...
12837 (make_pass_mudflap_2): ...new function.
12838 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
12839 subclass of gimple_opt_pass along with...
12840 (pass_data_mudflap_1): ...new pass_data instance and...
12841 (make_pass_mudflap_1): ...new function.
12842 (pass_mudflap_2): Convert from a global struct to a subclass of
12843 gimple_opt_pass along with...
12844 (pass_data_mudflap_2): ...new pass_data instance and...
12845 (make_pass_mudflap_2): ...new function.
12846 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
12847 gimple_opt_pass along with...
12848 (pass_data_nrv): ...new pass_data instance and...
12849 (make_pass_nrv): ...new function.
12850 (pass_return_slot): Convert from a global struct to a subclass of
12851 gimple_opt_pass along with...
12852 (pass_data_return_slot): ...new pass_data instance and...
12853 (make_pass_return_slot): ...new function.
12854 * tree-object-size.c (pass_object_sizes): Convert from a global struct
12855 to a subclass of gimple_opt_pass along with...
12856 (pass_data_object_sizes): ...new pass_data instance and...
12857 (make_pass_object_sizes): ...new function.
12858 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
12859 global struct to a subclass of gimple_opt_pass along with...
12860 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
12861 and...
12862 (make_pass_cleanup_cfg_post_optimizing): ...new function.
12863 (pass_fixup_cfg): Convert from a global struct to a subclass of
12864 gimple_opt_pass along with...
12865 (pass_data_fixup_cfg): ...new pass_data instance and...
12866 (make_pass_fixup_cfg): ...new function.
12867 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
12868 (make_pass_mudflap_1): ...new function.
12869 (pass_mudflap_2): Replace declaration with that of...
12870 (make_pass_mudflap_2): ...new function.
12871 (pass_asan): Replace declaration with that of...
12872 (make_pass_asan): ...new function.
12873 (pass_asan_O0): Replace declaration with that of...
12874 (make_pass_asan_O0): ...new function.
12875 (pass_tsan): Replace declaration with that of...
12876 (make_pass_tsan): ...new function.
12877 (pass_tsan_O0): Replace declaration with that of...
12878 (make_pass_tsan_O0): ...new function.
12879 (pass_lower_cf): Replace declaration with that of...
12880 (make_pass_lower_cf): ...new function.
12881 (pass_refactor_eh): Replace declaration with that of...
12882 (make_pass_refactor_eh): ...new function.
12883 (pass_lower_eh): Replace declaration with that of...
12884 (make_pass_lower_eh): ...new function.
12885 (pass_lower_eh_dispatch): Replace declaration with that of...
12886 (make_pass_lower_eh_dispatch): ...new function.
12887 (pass_lower_resx): Replace declaration with that of...
12888 (make_pass_lower_resx): ...new function.
12889 (pass_build_cfg): Replace declaration with that of...
12890 (make_pass_build_cfg): ...new function.
12891 (pass_early_tree_profile): Replace declaration with that of...
12892 (make_pass_early_tree_profile): ...new function.
12893 (pass_cleanup_eh): Replace declaration with that of...
12894 (make_pass_cleanup_eh): ...new function.
12895 (pass_sra): Replace declaration with that of...
12896 (make_pass_sra): ...new function.
12897 (pass_sra_early): Replace declaration with that of...
12898 (make_pass_sra_early): ...new function.
12899 (pass_early_ipa_sra): Replace declaration with that of...
12900 (make_pass_early_ipa_sra): ...new function.
12901 (pass_tail_recursion): Replace declaration with that of...
12902 (make_pass_tail_recursion): ...new function.
12903 (pass_tail_calls): Replace declaration with that of...
12904 (make_pass_tail_calls): ...new function.
12905 (pass_tree_loop): Replace declaration with that of...
12906 (make_pass_tree_loop): ...new function.
12907 (pass_tree_loop_init): Replace declaration with that of...
12908 (make_pass_tree_loop_init): ...new function.
12909 (pass_lim): Replace declaration with that of...
12910 (make_pass_lim): ...new function.
12911 (pass_tree_unswitch): Replace declaration with that of...
12912 (make_pass_tree_unswitch): ...new function.
12913 (pass_predcom): Replace declaration with that of...
12914 (make_pass_predcom): ...new function.
12915 (pass_iv_canon): Replace declaration with that of...
12916 (make_pass_iv_canon): ...new function.
12917 (pass_scev_cprop): Replace declaration with that of...
12918 (make_pass_scev_cprop): ...new function.
12919 (pass_empty_loop): Replace declaration with that of...
12920 (make_pass_empty_loop): ...new function.
12921 (pass_record_bounds): Replace declaration with that of...
12922 (make_pass_record_bounds): ...new function.
12923 (pass_graphite): Replace declaration with that of...
12924 (make_pass_graphite): ...new function.
12925 (pass_graphite_transforms): Replace declaration with that of...
12926 (make_pass_graphite_transforms): ...new function.
12927 (pass_if_conversion): Replace declaration with that of...
12928 (make_pass_if_conversion): ...new function.
12929 (pass_loop_distribution): Replace declaration with that of...
12930 (make_pass_loop_distribution): ...new function.
12931 (pass_vectorize): Replace declaration with that of...
12932 (make_pass_vectorize): ...new function.
12933 (pass_slp_vectorize): Replace declaration with that of...
12934 (make_pass_slp_vectorize): ...new function.
12935 (pass_complete_unroll): Replace declaration with that of...
12936 (make_pass_complete_unroll): ...new function.
12937 (pass_complete_unrolli): Replace declaration with that of...
12938 (make_pass_complete_unrolli): ...new function.
12939 (pass_parallelize_loops): Replace declaration with that of...
12940 (make_pass_parallelize_loops): ...new function.
12941 (pass_loop_prefetch): Replace declaration with that of...
12942 (make_pass_loop_prefetch): ...new function.
12943 (pass_iv_optimize): Replace declaration with that of...
12944 (make_pass_iv_optimize): ...new function.
12945 (pass_tree_loop_done): Replace declaration with that of...
12946 (make_pass_tree_loop_done): ...new function.
12947 (pass_ch): Replace declaration with that of...
12948 (make_pass_ch): ...new function.
12949 (pass_ccp): Replace declaration with that of...
12950 (make_pass_ccp): ...new function.
12951 (pass_phi_only_cprop): Replace declaration with that of...
12952 (make_pass_phi_only_cprop): ...new function.
12953 (pass_build_ssa): Replace declaration with that of...
12954 (make_pass_build_ssa): ...new function.
12955 (pass_build_alias): Replace declaration with that of...
12956 (make_pass_build_alias): ...new function.
12957 (pass_build_ealias): Replace declaration with that of...
12958 (make_pass_build_ealias): ...new function.
12959 (pass_dominator): Replace declaration with that of...
12960 (make_pass_dominator): ...new function.
12961 (pass_dce): Replace declaration with that of...
12962 (make_pass_dce): ...new function.
12963 (pass_dce_loop): Replace declaration with that of...
12964 (make_pass_dce_loop): ...new function.
12965 (pass_cd_dce): Replace declaration with that of...
12966 (make_pass_cd_dce): ...new function.
12967 (pass_call_cdce): Replace declaration with that of...
12968 (make_pass_call_cdce): ...new function.
12969 (pass_merge_phi): Replace declaration with that of...
12970 (make_pass_merge_phi): ...new function.
12971 (pass_split_crit_edges): Replace declaration with that of...
12972 (make_pass_split_crit_edges): ...new function.
12973 (pass_pre): Replace declaration with that of...
12974 (make_pass_pre): ...new function.
12975 (pass_profile): Replace declaration with that of...
12976 (make_pass_profile): ...new function.
12977 (pass_strip_predict_hints): Replace declaration with that of...
12978 (make_pass_strip_predict_hints): ...new function.
12979 (pass_lower_complex_O0): Replace declaration with that of...
12980 (make_pass_lower_complex_O0): ...new function.
12981 (pass_lower_complex): Replace declaration with that of...
12982 (make_pass_lower_complex): ...new function.
12983 (pass_lower_vector): Replace declaration with that of...
12984 (make_pass_lower_vector): ...new function.
12985 (pass_lower_vector_ssa): Replace declaration with that of...
12986 (make_pass_lower_vector_ssa): ...new function.
12987 (pass_lower_omp): Replace declaration with that of...
12988 (make_pass_lower_omp): ...new function.
12989 (pass_diagnose_omp_blocks): Replace declaration with that of...
12990 (make_pass_diagnose_omp_blocks): ...new function.
12991 (pass_expand_omp): Replace declaration with that of...
12992 (make_pass_expand_omp): ...new function.
12993 (pass_expand_omp_ssa): Replace declaration with that of...
12994 (make_pass_expand_omp_ssa): ...new function.
12995 (pass_object_sizes): Replace declaration with that of...
12996 (make_pass_object_sizes): ...new function.
12997 (pass_strlen): Replace declaration with that of...
12998 (make_pass_strlen): ...new function.
12999 (pass_fold_builtins): Replace declaration with that of...
13000 (make_pass_fold_builtins): ...new function.
13001 (pass_stdarg): Replace declaration with that of...
13002 (make_pass_stdarg): ...new function.
13003 (pass_early_warn_uninitialized): Replace declaration with that of...
13004 (make_pass_early_warn_uninitialized): ...new function.
13005 (pass_late_warn_uninitialized): Replace declaration with that of...
13006 (make_pass_late_warn_uninitialized): ...new function.
13007 (pass_cse_reciprocals): Replace declaration with that of...
13008 (make_pass_cse_reciprocals): ...new function.
13009 (pass_cse_sincos): Replace declaration with that of...
13010 (make_pass_cse_sincos): ...new function.
13011 (pass_optimize_bswap): Replace declaration with that of...
13012 (make_pass_optimize_bswap): ...new function.
13013 (pass_optimize_widening_mul): Replace declaration with that of...
13014 (make_pass_optimize_widening_mul): ...new function.
13015 (pass_warn_function_return): Replace declaration with that of...
13016 (make_pass_warn_function_return): ...new function.
13017 (pass_warn_function_noreturn): Replace declaration with that of...
13018 (make_pass_warn_function_noreturn): ...new function.
13019 (pass_cselim): Replace declaration with that of...
13020 (make_pass_cselim): ...new function.
13021 (pass_phiopt): Replace declaration with that of...
13022 (make_pass_phiopt): ...new function.
13023 (pass_forwprop): Replace declaration with that of...
13024 (make_pass_forwprop): ...new function.
13025 (pass_phiprop): Replace declaration with that of...
13026 (make_pass_phiprop): ...new function.
13027 (pass_tree_ifcombine): Replace declaration with that of...
13028 (make_pass_tree_ifcombine): ...new function.
13029 (pass_dse): Replace declaration with that of...
13030 (make_pass_dse): ...new function.
13031 (pass_nrv): Replace declaration with that of...
13032 (make_pass_nrv): ...new function.
13033 (pass_rename_ssa_copies): Replace declaration with that of...
13034 (make_pass_rename_ssa_copies): ...new function.
13035 (pass_sink_code): Replace declaration with that of...
13036 (make_pass_sink_code): ...new function.
13037 (pass_fre): Replace declaration with that of...
13038 (make_pass_fre): ...new function.
13039 (pass_check_data_deps): Replace declaration with that of...
13040 (make_pass_check_data_deps): ...new function.
13041 (pass_copy_prop): Replace declaration with that of...
13042 (make_pass_copy_prop): ...new function.
13043 (pass_vrp): Replace declaration with that of...
13044 (make_pass_vrp): ...new function.
13045 (pass_uncprop): Replace declaration with that of...
13046 (make_pass_uncprop): ...new function.
13047 (pass_return_slot): Replace declaration with that of...
13048 (make_pass_return_slot): ...new function.
13049 (pass_reassoc): Replace declaration with that of...
13050 (make_pass_reassoc): ...new function.
13051 (pass_rebuild_cgraph_edges): Replace declaration with that of...
13052 (make_pass_rebuild_cgraph_edges): ...new function.
13053 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
13054 (make_pass_remove_cgraph_callee_edges): ...new function.
13055 (pass_build_cgraph_edges): Replace declaration with that of...
13056 (make_pass_build_cgraph_edges): ...new function.
13057 (pass_local_pure_const): Replace declaration with that of...
13058 (make_pass_local_pure_const): ...new function.
13059 (pass_tracer): Replace declaration with that of...
13060 (make_pass_tracer): ...new function.
13061 (pass_warn_unused_result): Replace declaration with that of...
13062 (make_pass_warn_unused_result): ...new function.
13063 (pass_diagnose_tm_blocks): Replace declaration with that of...
13064 (make_pass_diagnose_tm_blocks): ...new function.
13065 (pass_lower_tm): Replace declaration with that of...
13066 (make_pass_lower_tm): ...new function.
13067 (pass_tm_init): Replace declaration with that of...
13068 (make_pass_tm_init): ...new function.
13069 (pass_tm_mark): Replace declaration with that of...
13070 (make_pass_tm_mark): ...new function.
13071 (pass_tm_memopt): Replace declaration with that of...
13072 (make_pass_tm_memopt): ...new function.
13073 (pass_tm_edges): Replace declaration with that of...
13074 (make_pass_tm_edges): ...new function.
13075 (pass_split_functions): Replace declaration with that of...
13076 (make_pass_split_functions): ...new function.
13077 (pass_feedback_split_functions): Replace declaration with that of...
13078 (make_pass_feedback_split_functions): ...new function.
13079 (pass_strength_reduction): Replace declaration with that of...
13080 (make_pass_strength_reduction): ...new function.
13081 (pass_ipa_lower_emutls): Replace declaration with that of...
13082 (make_pass_ipa_lower_emutls): ...new function.
13083 (pass_ipa_function_and_variable_visibility): Replace declaration with
13084 that of...
13085 (make_pass_ipa_function_and_variable_visibility): ...new function.
13086 (pass_ipa_tree_profile): Replace declaration with that of...
13087 (make_pass_ipa_tree_profile): ...new function.
13088 (pass_early_local_passes): Replace declaration with that of...
13089 (make_pass_early_local_passes): ...new function.
13090 (pass_ipa_whole_program_visibility): Replace declaration with that
13091 of...
13092 (make_pass_ipa_whole_program_visibility): ...new function.
13093 (pass_ipa_lto_gimple_out): Replace declaration with that of...
13094 (make_pass_ipa_lto_gimple_out): ...new function.
13095 (pass_ipa_increase_alignment): Replace declaration with that of...
13096 (make_pass_ipa_increase_alignment): ...new function.
13097 (pass_ipa_inline): Replace declaration with that of...
13098 (make_pass_ipa_inline): ...new function.
13099 (pass_ipa_free_lang_data): Replace declaration with that of...
13100 (make_pass_ipa_free_lang_data): ...new function.
13101 (pass_ipa_free_inline_summary): Replace declaration with that of...
13102 (make_pass_ipa_free_inline_summary): ...new function.
13103 (pass_ipa_cp): Replace declaration with that of...
13104 (make_pass_ipa_cp): ...new function.
13105 (pass_ipa_reference): Replace declaration with that of...
13106 (make_pass_ipa_reference): ...new function.
13107 (pass_ipa_pure_const): Replace declaration with that of...
13108 (make_pass_ipa_pure_const): ...new function.
13109 (pass_ipa_pta): Replace declaration with that of...
13110 (make_pass_ipa_pta): ...new function.
13111 (pass_ipa_lto_finish_out): Replace declaration with that of...
13112 (make_pass_ipa_lto_finish_out): ...new function.
13113 (pass_ipa_tm): Replace declaration with that of...
13114 (make_pass_ipa_tm): ...new function.
13115 (pass_ipa_profile): Replace declaration with that of...
13116 (make_pass_ipa_profile): ...new function.
13117 (pass_ipa_cdtor_merge): Replace declaration with that of...
13118 (make_pass_ipa_cdtor_merge): ...new function.
13119 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
13120 of...
13121 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13122 (pass_init_datastructures): Replace declaration with that of...
13123 (make_pass_init_datastructures): ...new function.
13124 (pass_fixup_cfg): Replace declaration with that of...
13125 (make_pass_fixup_cfg): ...new function.
13126 (pass_expand): Replace declaration with that of...
13127 (make_pass_expand): ...new function.
13128 (pass_instantiate_virtual_regs): Replace declaration with that of...
13129 (make_pass_instantiate_virtual_regs): ...new function.
13130 (pass_rtl_fwprop): Replace declaration with that of...
13131 (make_pass_rtl_fwprop): ...new function.
13132 (pass_rtl_fwprop_addr): Replace declaration with that of...
13133 (make_pass_rtl_fwprop_addr): ...new function.
13134 (pass_jump): Replace declaration with that of...
13135 (make_pass_jump): ...new function.
13136 (pass_jump2): Replace declaration with that of...
13137 (make_pass_jump2): ...new function.
13138 (pass_lower_subreg): Replace declaration with that of...
13139 (make_pass_lower_subreg): ...new function.
13140 (pass_cse): Replace declaration with that of...
13141 (make_pass_cse): ...new function.
13142 (pass_fast_rtl_dce): Replace declaration with that of...
13143 (make_pass_fast_rtl_dce): ...new function.
13144 (pass_ud_rtl_dce): Replace declaration with that of...
13145 (make_pass_ud_rtl_dce): ...new function.
13146 (pass_rtl_dce): Replace declaration with that of...
13147 (make_pass_rtl_dce): ...new function.
13148 (pass_rtl_dse1): Replace declaration with that of...
13149 (make_pass_rtl_dse1): ...new function.
13150 (pass_rtl_dse2): Replace declaration with that of...
13151 (make_pass_rtl_dse2): ...new function.
13152 (pass_rtl_dse3): Replace declaration with that of...
13153 (make_pass_rtl_dse3): ...new function.
13154 (pass_rtl_cprop): Replace declaration with that of...
13155 (make_pass_rtl_cprop): ...new function.
13156 (pass_rtl_pre): Replace declaration with that of...
13157 (make_pass_rtl_pre): ...new function.
13158 (pass_rtl_hoist): Replace declaration with that of...
13159 (make_pass_rtl_hoist): ...new function.
13160 (pass_rtl_store_motion): Replace declaration with that of...
13161 (make_pass_rtl_store_motion): ...new function.
13162 (pass_cse_after_global_opts): Replace declaration with that of...
13163 (make_pass_cse_after_global_opts): ...new function.
13164 (pass_rtl_ifcvt): Replace declaration with that of...
13165 (make_pass_rtl_ifcvt): ...new function.
13166 (pass_into_cfg_layout_mode): Replace declaration with that of...
13167 (make_pass_into_cfg_layout_mode): ...new function.
13168 (pass_outof_cfg_layout_mode): Replace declaration with that of...
13169 (make_pass_outof_cfg_layout_mode): ...new function.
13170 (pass_loop2): Replace declaration with that of...
13171 (make_pass_loop2): ...new function.
13172 (pass_rtl_loop_init): Replace declaration with that of...
13173 (make_pass_rtl_loop_init): ...new function.
13174 (pass_rtl_move_loop_invariants): Replace declaration with that of...
13175 (make_pass_rtl_move_loop_invariants): ...new function.
13176 (pass_rtl_unswitch): Replace declaration with that of...
13177 (make_pass_rtl_unswitch): ...new function.
13178 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
13179 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13180 (pass_rtl_doloop): Replace declaration with that of...
13181 (make_pass_rtl_doloop): ...new function.
13182 (pass_rtl_loop_done): Replace declaration with that of...
13183 (make_pass_rtl_loop_done): ...new function.
13184 (pass_web): Replace declaration with that of...
13185 (make_pass_web): ...new function.
13186 (pass_cse2): Replace declaration with that of...
13187 (make_pass_cse2): ...new function.
13188 (pass_df_initialize_opt): Replace declaration with that of...
13189 (make_pass_df_initialize_opt): ...new function.
13190 (pass_df_initialize_no_opt): Replace declaration with that of...
13191 (make_pass_df_initialize_no_opt): ...new function.
13192 (pass_reginfo_init): Replace declaration with that of...
13193 (make_pass_reginfo_init): ...new function.
13194 (pass_inc_dec): Replace declaration with that of...
13195 (make_pass_inc_dec): ...new function.
13196 (pass_stack_ptr_mod): Replace declaration with that of...
13197 (make_pass_stack_ptr_mod): ...new function.
13198 (pass_initialize_regs): Replace declaration with that of...
13199 (make_pass_initialize_regs): ...new function.
13200 (pass_combine): Replace declaration with that of...
13201 (make_pass_combine): ...new function.
13202 (pass_if_after_combine): Replace declaration with that of...
13203 (make_pass_if_after_combine): ...new function.
13204 (pass_ree): Replace declaration with that of...
13205 (make_pass_ree): ...new function.
13206 (pass_partition_blocks): Replace declaration with that of...
13207 (make_pass_partition_blocks): ...new function.
13208 (pass_match_asm_constraints): Replace declaration with that of...
13209 (make_pass_match_asm_constraints): ...new function.
13210 (pass_regmove): Replace declaration with that of...
13211 (make_pass_regmove): ...new function.
13212 (pass_split_all_insns): Replace declaration with that of...
13213 (make_pass_split_all_insns): ...new function.
13214 (pass_fast_rtl_byte_dce): Replace declaration with that of...
13215 (make_pass_fast_rtl_byte_dce): ...new function.
13216 (pass_lower_subreg2): Replace declaration with that of...
13217 (make_pass_lower_subreg2): ...new function.
13218 (pass_mode_switching): Replace declaration with that of...
13219 (make_pass_mode_switching): ...new function.
13220 (pass_sms): Replace declaration with that of...
13221 (make_pass_sms): ...new function.
13222 (pass_sched): Replace declaration with that of...
13223 (make_pass_sched): ...new function.
13224 (pass_ira): Replace declaration with that of...
13225 (make_pass_ira): ...new function.
13226 (pass_reload): Replace declaration with that of...
13227 (make_pass_reload): ...new function.
13228 (pass_clean_state): Replace declaration with that of...
13229 (make_pass_clean_state): ...new function.
13230 (pass_branch_prob): Replace declaration with that of...
13231 (make_pass_branch_prob): ...new function.
13232 (pass_value_profile_transformations): Replace declaration with that
13233 of...
13234 (make_pass_value_profile_transformations): ...new function.
13235 (pass_postreload_cse): Replace declaration with that of...
13236 (make_pass_postreload_cse): ...new function.
13237 (pass_gcse2): Replace declaration with that of...
13238 (make_pass_gcse2): ...new function.
13239 (pass_split_after_reload): Replace declaration with that of...
13240 (make_pass_split_after_reload): ...new function.
13241 (pass_branch_target_load_optimize1): Replace declaration with that
13242 of...
13243 (make_pass_branch_target_load_optimize1): ...new function.
13244 (pass_thread_prologue_and_epilogue): Replace declaration with that
13245 of...
13246 (make_pass_thread_prologue_and_epilogue): ...new function.
13247 (pass_stack_adjustments): Replace declaration with that of...
13248 (make_pass_stack_adjustments): ...new function.
13249 (pass_peephole2): Replace declaration with that of...
13250 (make_pass_peephole2): ...new function.
13251 (pass_if_after_reload): Replace declaration with that of...
13252 (make_pass_if_after_reload): ...new function.
13253 (pass_regrename): Replace declaration with that of...
13254 (make_pass_regrename): ...new function.
13255 (pass_cprop_hardreg): Replace declaration with that of...
13256 (make_pass_cprop_hardreg): ...new function.
13257 (pass_reorder_blocks): Replace declaration with that of...
13258 (make_pass_reorder_blocks): ...new function.
13259 (pass_branch_target_load_optimize2): Replace declaration with that
13260 of...
13261 (make_pass_branch_target_load_optimize2): ...new function.
13262 (pass_leaf_regs): Replace declaration with that of...
13263 (make_pass_leaf_regs): ...new function.
13264 (pass_split_before_sched2): Replace declaration with that of...
13265 (make_pass_split_before_sched2): ...new function.
13266 (pass_compare_elim_after_reload): Replace declaration with that of...
13267 (make_pass_compare_elim_after_reload): ...new function.
13268 (pass_sched2): Replace declaration with that of...
13269 (make_pass_sched2): ...new function.
13270 (pass_stack_regs): Replace declaration with that of...
13271 (make_pass_stack_regs): ...new function.
13272 (pass_stack_regs_run): Replace declaration with that of...
13273 (make_pass_stack_regs_run): ...new function.
13274 (pass_df_finish): Replace declaration with that of...
13275 (make_pass_df_finish): ...new function.
13276 (pass_compute_alignments): Replace declaration with that of...
13277 (make_pass_compute_alignments): ...new function.
13278 (pass_duplicate_computed_gotos): Replace declaration with that of...
13279 (make_pass_duplicate_computed_gotos): ...new function.
13280 (pass_variable_tracking): Replace declaration with that of...
13281 (make_pass_variable_tracking): ...new function.
13282 (pass_free_cfg): Replace declaration with that of...
13283 (make_pass_free_cfg): ...new function.
13284 (pass_machine_reorg): Replace declaration with that of...
13285 (make_pass_machine_reorg): ...new function.
13286 (pass_cleanup_barriers): Replace declaration with that of...
13287 (make_pass_cleanup_barriers): ...new function.
13288 (pass_delay_slots): Replace declaration with that of...
13289 (make_pass_delay_slots): ...new function.
13290 (pass_split_for_shorten_branches): Replace declaration with that of...
13291 (make_pass_split_for_shorten_branches): ...new function.
13292 (pass_split_before_regstack): Replace declaration with that of...
13293 (make_pass_split_before_regstack): ...new function.
13294 (pass_convert_to_eh_region_ranges): Replace declaration with that
13295 of...
13296 (make_pass_convert_to_eh_region_ranges): ...new function.
13297 (pass_shorten_branches): Replace declaration with that of...
13298 (make_pass_shorten_branches): ...new function.
13299 (pass_set_nothrow_function_flags): Replace declaration with that of...
13300 (make_pass_set_nothrow_function_flags): ...new function.
13301 (pass_dwarf2_frame): Replace declaration with that of...
13302 (make_pass_dwarf2_frame): ...new function.
13303 (pass_final): Replace declaration with that of...
13304 (make_pass_final): ...new function.
13305 (pass_rtl_seqabstr): Replace declaration with that of...
13306 (make_pass_rtl_seqabstr): ...new function.
13307 (pass_release_ssa_names): Replace declaration with that of...
13308 (make_pass_release_ssa_names): ...new function.
13309 (pass_early_inline): Replace declaration with that of...
13310 (make_pass_early_inline): ...new function.
13311 (pass_inline_parameters): Replace declaration with that of...
13312 (make_pass_inline_parameters): ...new function.
13313 (pass_update_address_taken): Replace declaration with that of...
13314 (make_pass_update_address_taken): ...new function.
13315 (pass_convert_switch): Replace declaration with that of...
13316 (make_pass_convert_switch): ...new function.
13317 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
13318 to a subclass of simple_ipa_opt_pass along with...
13319 (pass_data_ipa_tree_profile): ...new pass_data instance and...
13320 (make_pass_ipa_tree_profile): ...new function.
13321 * tree-sra.c (pass_sra_early): Convert from a global struct to a
13322 subclass of gimple_opt_pass along with...
13323 (pass_data_sra_early): ...new pass_data instance and...
13324 (make_pass_sra_early): ...new function.
13325 (pass_sra): Convert from a global struct to a subclass of
13326 gimple_opt_pass along with...
13327 (pass_data_sra): ...new pass_data instance and...
13328 (make_pass_sra): ...new function.
13329 (pass_early_ipa_sra): Convert from a global struct to a subclass of
13330 gimple_opt_pass along with...
13331 (pass_data_early_ipa_sra): ...new pass_data instance and...
13332 (make_pass_early_ipa_sra): ...new function.
13333 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
13334 subclass of gimple_opt_pass along with...
13335 (pass_data_ccp): ...new pass_data instance and...
13336 (make_pass_ccp): ...new function.
13337 (pass_fold_builtins): Convert from a global struct to a subclass of
13338 gimple_opt_pass along with...
13339 (pass_data_fold_builtins): ...new pass_data instance and...
13340 (make_pass_fold_builtins): ...new function.
13341 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
13342 subclass of gimple_opt_pass along with...
13343 (pass_data_copy_prop): ...new pass_data instance and...
13344 (make_pass_copy_prop): ...new function.
13345 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
13346 global struct to a subclass of gimple_opt_pass along with...
13347 (pass_data_rename_ssa_copies): ...new pass_data instance and...
13348 (make_pass_rename_ssa_copies): ...new function.
13349 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
13350 subclass of gimple_opt_pass along with...
13351 (pass_data_dce): ...new pass_data instance and...
13352 (make_pass_dce): ...new function.
13353 (pass_dce_loop): Convert from a global struct to a subclass of
13354 gimple_opt_pass along with...
13355 (pass_data_dce_loop): ...new pass_data instance and...
13356 (make_pass_dce_loop): ...new function.
13357 (pass_cd_dce): Convert from a global struct to a subclass of
13358 gimple_opt_pass along with...
13359 (pass_data_cd_dce): ...new pass_data instance and...
13360 (make_pass_cd_dce): ...new function.
13361 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
13362 subclass of gimple_opt_pass along with...
13363 (pass_data_dominator): ...new pass_data instance and...
13364 (make_pass_dominator): ...new function.
13365 (pass_phi_only_cprop): Convert from a global struct to a subclass of
13366 gimple_opt_pass along with...
13367 (pass_data_phi_only_cprop): ...new pass_data instance and...
13368 (make_pass_phi_only_cprop): ...new function.
13369 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
13370 subclass of gimple_opt_pass along with...
13371 (pass_data_dse): ...new pass_data instance and...
13372 (make_pass_dse): ...new function.
13373 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
13374 a subclass of gimple_opt_pass along with...
13375 (pass_data_forwprop): ...new pass_data instance and...
13376 (make_pass_forwprop): ...new function.
13377 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
13378 struct to a subclass of gimple_opt_pass along with...
13379 (pass_data_tree_ifcombine): ...new pass_data instance and...
13380 (make_pass_tree_ifcombine): ...new function.
13381 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
13382 subclass of gimple_opt_pass along with...
13383 (pass_data_ch): ...new pass_data instance and...
13384 (make_pass_ch): ...new function.
13385 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
13386 subclass of gimple_opt_pass along with...
13387 (pass_data_tree_loop): ...new pass_data instance and...
13388 (make_pass_tree_loop): ...new function.
13389 (pass_tree_loop_init): Convert from a global struct to a subclass of
13390 gimple_opt_pass along with...
13391 (pass_data_tree_loop_init): ...new pass_data instance and...
13392 (make_pass_tree_loop_init): ...new function.
13393 (pass_lim): Convert from a global struct to a subclass of
13394 gimple_opt_pass along with...
13395 (pass_data_lim): ...new pass_data instance and...
13396 (make_pass_lim): ...new function.
13397 (pass_tree_unswitch): Convert from a global struct to a subclass of
13398 gimple_opt_pass along with...
13399 (pass_data_tree_unswitch): ...new pass_data instance and...
13400 (make_pass_tree_unswitch): ...new function.
13401 (pass_predcom): Convert from a global struct to a subclass of
13402 gimple_opt_pass along with...
13403 (pass_data_predcom): ...new pass_data instance and...
13404 (make_pass_predcom): ...new function.
13405 (pass_vectorize): Convert from a global struct to a subclass of
13406 gimple_opt_pass along with...
13407 (pass_data_vectorize): ...new pass_data instance and...
13408 (make_pass_vectorize): ...new function.
13409 (pass_graphite): Convert from a global struct to a subclass of
13410 gimple_opt_pass along with...
13411 (pass_data_graphite): ...new pass_data instance and...
13412 (make_pass_graphite): ...new function.
13413 (pass_graphite_transforms): Convert from a global struct to a subclass
13414 of gimple_opt_pass along with...
13415 (pass_data_graphite_transforms): ...new pass_data instance and...
13416 (make_pass_graphite_transforms): ...new function.
13417 (pass_check_data_deps): Convert from a global struct to a subclass of
13418 gimple_opt_pass along with...
13419 (pass_data_check_data_deps): ...new pass_data instance and...
13420 (make_pass_check_data_deps): ...new function.
13421 (pass_iv_canon): Convert from a global struct to a subclass of
13422 gimple_opt_pass along with...
13423 (pass_data_iv_canon): ...new pass_data instance and...
13424 (make_pass_iv_canon): ...new function.
13425 (pass_scev_cprop): Convert from a global struct to a subclass of
13426 gimple_opt_pass along with...
13427 (pass_data_scev_cprop): ...new pass_data instance and...
13428 (make_pass_scev_cprop): ...new function.
13429 (pass_record_bounds): Convert from a global struct to a subclass of
13430 gimple_opt_pass along with...
13431 (pass_data_record_bounds): ...new pass_data instance and...
13432 (make_pass_record_bounds): ...new function.
13433 (pass_complete_unroll): Convert from a global struct to a subclass of
13434 gimple_opt_pass along with...
13435 (pass_data_complete_unroll): ...new pass_data instance and...
13436 (make_pass_complete_unroll): ...new function.
13437 (pass_complete_unrolli): Convert from a global struct to a subclass of
13438 gimple_opt_pass along with...
13439 (pass_data_complete_unrolli): ...new pass_data instance and...
13440 (make_pass_complete_unrolli): ...new function.
13441 (pass_parallelize_loops): Convert from a global struct to a subclass
13442 of gimple_opt_pass along with...
13443 (pass_data_parallelize_loops): ...new pass_data instance and...
13444 (make_pass_parallelize_loops): ...new function.
13445 (pass_loop_prefetch): Convert from a global struct to a subclass of
13446 gimple_opt_pass along with...
13447 (pass_data_loop_prefetch): ...new pass_data instance and...
13448 (make_pass_loop_prefetch): ...new function.
13449 (pass_iv_optimize): Convert from a global struct to a subclass of
13450 gimple_opt_pass along with...
13451 (pass_data_iv_optimize): ...new pass_data instance and...
13452 (make_pass_iv_optimize): ...new function.
13453 (pass_tree_loop_done): Convert from a global struct to a subclass of
13454 gimple_opt_pass along with...
13455 (pass_data_tree_loop_done): ...new pass_data instance and...
13456 (make_pass_tree_loop_done): ...new function.
13457 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
13458 struct to a subclass of gimple_opt_pass along with...
13459 (pass_data_cse_reciprocals): ...new pass_data instance and...
13460 (make_pass_cse_reciprocals): ...new function.
13461 (pass_cse_sincos): Convert from a global struct to a subclass of
13462 gimple_opt_pass along with...
13463 (pass_data_cse_sincos): ...new pass_data instance and...
13464 (make_pass_cse_sincos): ...new function.
13465 (pass_optimize_bswap): Convert from a global struct to a subclass of
13466 gimple_opt_pass along with...
13467 (pass_data_optimize_bswap): ...new pass_data instance and...
13468 (make_pass_optimize_bswap): ...new function.
13469 (pass_optimize_widening_mul): Convert from a global struct to a
13470 subclass of gimple_opt_pass along with...
13471 (pass_data_optimize_widening_mul): ...new pass_data instance and...
13472 (make_pass_optimize_widening_mul): ...new function.
13473 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
13474 subclass of gimple_opt_pass along with...
13475 (pass_data_phiopt): ...new pass_data instance and...
13476 (make_pass_phiopt): ...new function.
13477 (pass_cselim): Convert from a global struct to a subclass of
13478 gimple_opt_pass along with...
13479 (pass_data_cselim): ...new pass_data instance and...
13480 (make_pass_cselim): ...new function.
13481 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
13482 subclass of gimple_opt_pass along with...
13483 (pass_data_phiprop): ...new pass_data instance and...
13484 (make_pass_phiprop): ...new function.
13485 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
13486 subclass of gimple_opt_pass along with...
13487 (pass_data_pre): ...new pass_data instance and...
13488 (make_pass_pre): ...new function.
13489 (pass_fre): Convert from a global struct to a subclass of
13490 gimple_opt_pass along with...
13491 (pass_data_fre): ...new pass_data instance and...
13492 (make_pass_fre): ...new function.
13493 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
13494 subclass of gimple_opt_pass along with...
13495 (pass_data_reassoc): ...new pass_data instance and...
13496 (make_pass_reassoc): ...new function.
13497 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
13498 subclass of gimple_opt_pass along with...
13499 (pass_data_sink_code): ...new pass_data instance and...
13500 (make_pass_sink_code): ...new function.
13501 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
13502 subclass of gimple_opt_pass along with...
13503 (pass_data_strlen): ...new pass_data instance and...
13504 (make_pass_strlen): ...new function.
13505 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
13506 struct to a subclass of gimple_opt_pass along with...
13507 (pass_data_build_alias): ...new pass_data instance and...
13508 (make_pass_build_alias): ...new function.
13509 (pass_build_ealias): Convert from a global struct to a subclass of
13510 gimple_opt_pass along with...
13511 (pass_data_build_ealias): ...new pass_data instance and...
13512 (make_pass_build_ealias): ...new function.
13513 (pass_ipa_pta): Convert from a global struct to a subclass of
13514 simple_ipa_opt_pass along with...
13515 (pass_data_ipa_pta): ...new pass_data instance and...
13516 (make_pass_ipa_pta): ...new function.
13517 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
13518 subclass of gimple_opt_pass along with...
13519 (pass_data_uncprop): ...new pass_data instance and...
13520 (make_pass_uncprop): ...new function.
13521 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
13522 global struct to a subclass of gimple_opt_pass along with...
13523 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
13524 (make_pass_late_warn_uninitialized): ...new function.
13525 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
13526 to a subclass of gimple_opt_pass along with...
13527 (pass_data_init_datastructures): ...new pass_data instance and...
13528 (make_pass_init_datastructures): ...new function.
13529 (pass_early_warn_uninitialized): Convert from a global struct to a
13530 subclass of gimple_opt_pass along with...
13531 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
13532 (make_pass_early_warn_uninitialized): ...new function.
13533 (pass_update_address_taken): Convert from a global struct to a
13534 subclass of gimple_opt_pass along with...
13535 (pass_data_update_address_taken): ...new pass_data instance and...
13536 (make_pass_update_address_taken): ...new function.
13537 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
13538 struct to a subclass of gimple_opt_pass along with...
13539 (pass_data_release_ssa_names): ...new pass_data instance and...
13540 (make_pass_release_ssa_names): ...new function.
13541 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
13542 subclass of gimple_opt_pass along with...
13543 (pass_data_stdarg): ...new pass_data instance and...
13544 (make_pass_stdarg): ...new function.
13545 * tree-switch-conversion.c (pass_convert_switch): Convert from a
13546 global struct to a subclass of gimple_opt_pass along with...
13547 (pass_data_convert_switch): ...new pass_data instance and...
13548 (make_pass_convert_switch): ...new function.
13549 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
13550 to a subclass of gimple_opt_pass along with...
13551 (pass_data_tail_recursion): ...new pass_data instance and...
13552 (make_pass_tail_recursion): ...new function.
13553 (pass_tail_calls): Convert from a global struct to a subclass of
13554 gimple_opt_pass along with...
13555 (pass_data_tail_calls): ...new pass_data instance and...
13556 (make_pass_tail_calls): ...new function.
13557 * tree-vect-generic.c (pass_lower_vector): Convert from a global
13558 struct to a subclass of gimple_opt_pass along with...
13559 (pass_data_lower_vector): ...new pass_data instance and...
13560 (make_pass_lower_vector): ...new function.
13561 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
13562 gimple_opt_pass along with...
13563 (pass_data_lower_vector_ssa): ...new pass_data instance and...
13564 (make_pass_lower_vector_ssa): ...new function.
13565 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
13566 to a subclass of gimple_opt_pass along with...
13567 (pass_data_slp_vectorize): ...new pass_data instance and...
13568 (make_pass_slp_vectorize): ...new function.
13569 (pass_ipa_increase_alignment): Convert from a global struct to a
13570 subclass of simple_ipa_opt_pass along with...
13571 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
13572 (make_pass_ipa_increase_alignment): ...new function.
13573 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
13574 gimple_opt_pass along with...
13575 (pass_data_vrp): ...new pass_data instance and...
13576 (make_pass_vrp): ...new function.
13577 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
13578 subclass of simple_ipa_opt_pass along with...
13579 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
13580 (make_pass_ipa_free_lang_data): ...new function.
13581 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
13582 gimple_opt_pass along with...
13583 (pass_data_tsan): ...new pass_data instance and...
13584 (make_pass_tsan): ...new function.
13585 (pass_tsan_O0): Convert from a global struct to a subclass of
13586 gimple_opt_pass along with...
13587 (pass_data_tsan_O0): ...new pass_data instance and...
13588 (make_pass_tsan_O0): ...new function.
13589 * var-tracking.c (pass_variable_tracking): Convert from a global
13590 struct to a subclass of rtl_opt_pass along with...
13591 (pass_data_variable_tracking): ...new pass_data instance and...
13592 (make_pass_variable_tracking): ...new function.
13593 * web.c (pass_web): Convert from a global struct to a subclass of
13594 rtl_opt_pass along with...
13595 (pass_data_web): ...new pass_data instance and...
13596 (make_pass_web): ...new function.
13597 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
13598 declaration with that of...
13599 (make_pass_mode_switch_use): ...new function.
13600 (pass_resolve_sw_modes): Replace declaration with that of...
13601 (make_pass_resolve_sw_modes): ...new function.
13602 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
13603 from a global struct to a subclass of rtl_opt_pass along with...
13604 (pass_data_mode_switch_use): ...new pass_data instance and...
13605 (make_pass_mode_switch_use): ...new function.
13606 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
13607 from a global struct to a subclass of rtl_opt_pass along with...
13608 (pass_data_resolve_sw_modes): ...new pass_data instance and...
13609 (make_pass_resolve_sw_modes): ...new function.
13610 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
13611 struct to a subclass of rtl_opt_pass along with...
13612 (pass_data_insert_vzeroupper): ...new pass_data instance and...
13613 (make_pass_insert_vzeroupper): ...new function.
13614 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
13615 global struct to a subclass of rtl_opt_pass along with...
13616 (pass_data_work_around_errata): ...new pass_data instance and...
13617 (make_pass_work_around_errata): ...new function.
13618 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
13619 struct to a subclass of rtl_opt_pass along with...
13620 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
13621 (make_pass_mips_machine_reorg2): ...new function.
13622
13623 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13624
13625 * passes.c (pass_manager::operator new): New.
13626
13627 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13628
13629 Handwritten part of conversion of passes to C++ classes.
13630
13631 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
13632 (toplev.o): Add dep on PASS_MANAGER_H.
13633 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
13634 of early local pases to reflect this moving from a global to a
13635 member of gcc::pass_manager.
13636 (cgraph_add_new_function): Likewise.
13637 * lto-cgraph.c (lto_output_node): Update for conversion of
13638 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
13639 * passes.c (opt_pass::clone): New.
13640 (opt_pass::gate): New.
13641 (opt_pass::execute): New.
13642 (opt_pass::opt_pass): New.
13643 (pass_manager::execute_early_local_passes): New.
13644 (pass_manager::execute_pass_mode_switching): new.
13645 (finish_optimization_passes): Convert to...
13646 (pass_manager::finish_optimization_passes): ...this.
13647 (finish_optimization_passes): Update for conversion of passes to
13648 C++ classes.
13649 (register_dump_files_1): Use has_gate since we cannot portably
13650 check a vtable entry against NULL.
13651 (dump_one_pass): Likewise.
13652 (ipa_write_summaries_2): Likewise.
13653 (ipa_write_optimization_summaries_1): Likewise.
13654 (ipa_read_summaries_1): Likewise.
13655 (ipa_read_optimization_summaries_1): Likewise.
13656 (execute_ipa_stmt_fixups): Likewise.
13657 (pass_manager::pass_manager): Rewrite pass-creation, invoking
13658 pass-creation functions rather than wiring up globals, and
13659 storing the results in fields of pass_manager generated using
13660 pass-instances.def.
13661 (pass_manager::dump_profile_report): Update for conversion of
13662 passes to C++ classes.
13663 (pass_manager::execute_ipa_summary_passes): Likewise.
13664 (execute_one_ipa_transform_pass): Likewise.
13665 (execute_one_pass): Use has_gate and has_execute since we cannot
13666 portably check a vtable entry against NULL.
13667 * pass_manager.h (pass_manager::finish_optimization_passes): New.
13668 (pass_manager): Use pass-instances.def to add fields for the
13669 various pass instances.
13670 * toplev.c (finalize): Update for move of
13671 finish_optimization_passes to a method of gcc::pass_manager.
13672 * toplev.h (finish_optimization_passes): Move to method of class
13673 pass_manager.
13674 * tree-pass.h (struct pass_data): New.
13675 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
13676 (opt_pass::gate): Convert to virtual function.
13677 (opt_pass::~opt_pass): New.
13678 (opt_pass::clone): New.
13679 (opt_pass::execute): Convert to virtual function.
13680 (opt_pass::opt_pass): New.
13681 (opt_pass::ctxt_): new.
13682 (gimple_opt_pass): Convert to subclass of opt_pass.
13683 (gimple_opt_pass::gimple_opt_pass): New.
13684 (rtl_opt_pass): Convert to subclass of opt_pass.
13685 (rtl_opt_pass::rtl_opt_pass): New.
13686 (ipa_opt_pass_d): Convert to subclass of opt_pass.
13687 (ipa_opt_pass_d::ipa_opt_pass_d): New.
13688 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
13689 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
13690 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
13691 invocation of pass_mode_switching to reflect this moving from a
13692 global to a member of gcc::pass_manager.
13693 (ix86_option_override): Rework how pass_insert_vzeroupper is
13694 added to the pass_manager to reflect autogenerated changes.
13695 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
13696
13697 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
13698
13699 PR rtl-optimization/57708
13700 * recog.c (peep2_find_free_register): Validate all regs in a
13701 multi-reg mode.
13702
13703 2013-08-05 Jan Hubicka <jh@suse.cz>
13704
13705 PR lto/57602
13706 * cgraph.c (verify_cgraph_node): Accept local flags from other
13707 partitions.
13708 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
13709 (function_and_variable_visibility): Likewise.
13710 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
13711
13712 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13713
13714 * graph.c (init_graph_slim_pretty_print): Remove.
13715 (print_graph_cfg): Do not call it. Use local pretty printer.
13716 (start_graph_dump): Likewise.
13717
13718 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13719
13720 * gimple-pretty-print.c (buffer): Remove.
13721 (initialized): Likewise.
13722 (maybe_init_pretty_print): Likewise.
13723 (print_gimple_stmt): Do not call it. Use non-static local
13724 pretty_printer variable.
13725 (print_gimple_expr): Likewise.
13726 (print_gimple_seq): Likewise.
13727 (gimple_dump_bb): Likewise.
13728
13729 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13730
13731 * asan.c (asan_pp): Remove.
13732 (asan_pp_initialized): Likewise.
13733 (asan_pp_initialize): Likewise.
13734 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
13735 (asan_emit_stack_protection): Tidy. Use local pretty printer.
13736 (asan_add_global): Likewise.
13737
13738 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
13739
13740 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
13741 * diagnostic.h (diagnostic_flush_buffer): Adjust.
13742 * pretty-print.c (pp_formatted_text_data): Likewise.
13743 (pp_indent): Rename from pp_base_indent.
13744 (pp_format): Rename from pp_base_format.
13745 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
13746 (pp_format_verbatim): Rename from pp_base_format_verbatim.
13747 (pp_flush): Rename from pp_base_flush.
13748 (pp_set_line_maximum_length): Rename from
13749 pp_base_set_line_maximum_length.
13750 (pp_clear_output_area): Rename from pp_base_clear_output_area.
13751 (pp_set_prefix): Rename from pp_base_set_prefix.
13752 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
13753 (pp_emit_prefix): Rename from pp_base_emit_prefix.
13754 (pp_append_text): Rename from pp_base_append_text.
13755 (pp_formatted_text): Rename from pp_base_formatted_text.
13756 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
13757 (pp_remaining_character_count_for_line): Rename from
13758 pp_base_remaining_character_count_for_line.
13759 (pp_newline): Rename from pp_base_newline.
13760 (pp_character): Rename from pp_base_character.
13761 (pp_string): Rename from pp_base_string.
13762 (pp_maybe_space): Rename from pp_base_maybe_space.
13763 * asan.c (asan_pp_string): Adjust.
13764 (asan_emit_stack_protection): Likewise.
13765 (asan_add_global): Likewise.
13766 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
13767 * tree-mudflap.c (mf_varname_tree): Likewise.
13768 * tree-pretty-print.c (pp_tree_identifier): Rename from
13769 pp_base_tree_identifier.
13770 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
13771 Declare as function.
13772
13773 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13774
13775 * pretty-print.h (pp_bar_bar): New.
13776 (pp_ampersand_ampersand): Likewise.
13777 (pp_less_equal): Likewise.
13778 (pp_greater_equal): Likewise.
13779 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
13780 printer functions instead of pp_string or operators and punctuators.
13781 (dump_gimple_call): Likewise.
13782 (dump_gimple_omp_for): Likewise.
13783 (dump_gimple_transaction): Likewise.
13784 (dump_gimple_phi): Likewise.
13785 (pp_gimple_stmt_1): Likewise.
13786 * sched-vis.c (print_insn): Likewise.
13787 * tree-mudflap.c (mf_varname_tree): Likewise.
13788 * tree-pretty-print.c (dump_block_node): Likewise.
13789 (dump_generic_node): Likewise.
13790
13791 2013-08-02 Jan Hubicka <jh@suse.cz>
13792
13793 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
13794 boundaries.
13795 * lto-streamer-out.c (tree_is_indexable): Results decls and
13796 parm decls are not indexable.
13797 (DFS_write_tree_body): Do not follow args and results.
13798 (hash_tree): Likewise.
13799 (output_functions): Rearrange so struct function is needed
13800 only when real body is output; be able to also ouptut abstract
13801 functions; output DECL_ARGUMENTS and DECL_RESULT.
13802 (lto_output): When not in WPA, ale store abstract functions.
13803 (write_symbol): Do not care about RESULT_DECL.
13804 (output_symbol_p): Handle correctly sbtract decls.
13805 * lto-streamer-in.c (input_function): Rearrange so struct
13806 function can be NULL at entry; allow streaming of
13807 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
13808 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
13809 sanity check during LTO.
13810 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
13811 RESULT_DECl and DECL_ARGUMENTS.
13812 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
13813 Likewise.
13814
13815 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13816
13817 * pretty-print.h (pp_underscore): New.
13818 (pp_comma): Tidy.
13819 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
13820 printer functions instead of pp_character.
13821 (dump_binary_rhs): Likewise.
13822 (dump_ternary_rhs): Likewise.
13823 (dump_gimple_call_args): Likewise.
13824 (pp_points_to_solution): Likewise.
13825 (dump_gimple_call): Likewise.
13826 (dump_gimple_switch): Likewise.
13827 (dump_gimple_cond): Likewise.
13828 (dump_gimple_bind): Likewise.
13829 (dump_gimple_try): Likewise.
13830 (dump_gimple_omp_for): Likewise.
13831 (dump_gimple_omp_continue): Likewise.
13832 (dump_gimple_omp_single): Likewise.
13833 (dump_gimple_omp_sections): Likewise.
13834 (dump_gimple_omp_block): Likewise.
13835 (dump_gimple_omp_critical): Likewise.
13836 (dump_gimple_transaction): Likewise.
13837 (dump_gimple_asm): Likewise.
13838 (dump_gimple_phi): Likewise.
13839 (dump_gimple_omp_parallel): Likewise.
13840 (dump_gimple_omp_task): Likewise.
13841 (dump_gimple_omp_atomic_load): Likewise.
13842 (dump_gimple_omp_atomic_store): Likewise.
13843 (dump_gimple_mem_ops): Likewise.
13844 (pp_gimple_stmt_1): Likewise.
13845 (pp_cfg_jump): Likewise.
13846 (dump_implicit_edges): Likewise.
13847 (gimple_dump_bb_for_graph): Likewise.
13848 * graph.c (draw_cfg_node): Likewise.
13849 * langhooks.c (lhd_print_error_function): Likewise.
13850 * sched-vis.c (print_exp): Likewise.
13851 (print_value): Likewise.
13852 (print_pattern): Likewise.
13853 (print_insn): Likewise.
13854 (rtl_dump_bb_for_graph): Likewise.
13855 * tree-pretty-print.c (dump_function_declaration): Likewise.
13856 (dump_array_domain): Likewise.
13857 (dump_omp_clause): Likewise.
13858 (dump_location): Likewise.
13859 (dump_generic_node): Likewise.
13860 (print_struct_decl): Likewise.
13861 * diagnostic.c (diagnostic_show_locus): Use pp_space.
13862
13863 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
13864
13865 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
13866 candidate table when replacing a candidate statement.
13867 (replace_rhs_if_not_dup): Likewise.
13868 (replace_one_candidate): Likewise.
13869
13870 2013-08-02 Jan Hubicka <jh@suse.cz>
13871 Martin Liska <marxin.liska@gmail.com>
13872
13873 * cgraphunit.c (add_new_function): Fix logic when adding from
13874 late IPA pass.
13875 (assemble_thunk): Rename to ...
13876 (expand_thunk); .. this one; export; get it working with
13877 general functions; make produced gimple valid.
13878 * cgraph.h (expand_thunk): Declare.
13879
13880 2013-08-02 Jan Hubicka <jh@suse.cz>
13881
13882 * ipa-cp.c (gather_context_independent_values): Use
13883 ipa_get_param_move_cost.
13884 (get_replacement_map): Remove PARAM; move parameter folding
13885 into tree-inline.c
13886 (create_specialized_node): Update.
13887 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
13888 assert that we have gimple body; update move_cost.
13889 (count_formal_params): Assert that we have gimple body.
13890 (ipa_dump_param): New function.
13891 (ipa_alloc_node_params): Break out from ...
13892 (ipa_initialize_node_params): ... here.
13893 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
13894 (ipa_write_node_info): Stream move costs.
13895 (ipa_read_node_info): Read move costs.
13896 (ipa_update_after_lto_read): Do not recompute node params.
13897 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
13898 (ipa_get_param): Check we are not in WPA.
13899 (ipa_get_param_move_cost): New.
13900 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
13901 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
13902 parm numbers to be present.
13903
13904 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
13905
13906 PR rtl-optimization/58048
13907 * lra-constraints.c (process_alt_operands): Don't check asm
13908 operand on register.
13909
13910 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
13911
13912 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
13913 the implied StoreLoad barrier for atomic operations if before.
13914
13915 2013-08-02 Jan Hubicka <jh@suse.cz>
13916 Martin Liska <marxin.liska@gmail.com>
13917
13918 * cgraph.c (cgraph_function_body_availability): Do not check
13919 cgraph flags.
13920 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
13921 symtab_node_availability): Declare.
13922 * ipa.c (can_replace_by_local_alias): New.
13923 (function_and_variable_visibility): Use it.
13924 * symtab.c (symtab_for_node_and_aliases,
13925 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
13926
13927 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
13928
13929 PR rtl-optimization/57963
13930 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
13931 (lra_constraints): Use them.
13932
13933 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
13934
13935 * config/arm/types.md (define_attr "type"): Add "load_acq"
13936 and "store_rel".
13937 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
13938 changes.
13939 (cortex_a53_store1): Likewise.
13940
13941 2013-08-01 Jan Hubicka <jh@suse.cz>
13942
13943 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
13944 partitions are not needed.
13945
13946 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
13947
13948 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
13949 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
13950 MAYBE_NON_Q_CLASS_P where appropriate.
13951
13952 2013-08-01 Jan Hubicka <jh@suse.cz>
13953
13954 * cgraph.h (release_function_body): Declare.
13955 * tree.c (free_lang_data_in_decl): Free, parameters and return values
13956 of unused delcarations.
13957
13958 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13959
13960 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
13961 RTL form when subtracting a constant.
13962
13963 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13964
13965 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
13966 Generate canonical plus rtx with negated immediate instead of minus
13967 where appropriate.
13968 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
13969
13970 2013-08-01 Jan Hubicka <jh@suse.cz>
13971
13972 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
13973 (cgraph_release_function_body): Likewise.
13974 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
13975 * cgraph.h (cgrpah_node): Rename abstract_and_needed
13976 to used_as_abstract_origin.
13977 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
13978 symbols used as abstract origins.
13979 * cgraphunit.c (analyze_functions): Update.
13980 * ipa.c (symtab_remove_unreachable_nodes): Recompute
13981 used_as_abstract_origin.
13982 * tree-inline.c (tree_function_versioning): Update
13983 used_as_abstract_origin; be ready for DECL_RESULT and
13984 DECL_ARGUMENTS to be NULL.
13985
13986 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
13987 for abstract functions.
13988 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
13989 real symbols.
13990
13991 2013-08-01 Jan Hubicka <jh@suse.cz>
13992
13993 * profile.c (compute_value_histograms): Fix thinko.
13994
13995 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
13996
13997 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
13998 aarch-common-protos.h to extra_headers.
13999 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14000 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
14001 * config/aarch64/t-aarch64 (aarch-common.o): Define.
14002
14003 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14004
14005 * config/aarch64/aarch64.md (define_attr "type"): Delete.
14006 Include "../arm/types.md". Define "type" attribute for all patterns.
14007 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
14008 attribute changes.
14009
14010 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
14011
14012 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
14013 to support power8 load fusion.
14014 (fusion_gpr_mem_load): Likewise.
14015
14016 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
14017
14018 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
14019 declarations for power8 load fusion.
14020 (emit_fusion_gpr_load): Likewise.
14021
14022 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
14023 tuning for power8, turn on fusion mode by default. Turn on sign
14024 extending fusion mode if normal fusion mode is on, and we are at
14025 -O2 or -O3.
14026 (fusion_gpr_load_p): New function, return true if we can fuse an
14027 addis instruction with a dependent load to a GPR.
14028 (emit_fusion_gpr_load): Emit the instructions for power8 load
14029 fusion to GPRs.
14030
14031 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
14032 (VSX load fusion peepholes): New peepholes to fuse together an
14033 addi instruction with a VSX load instruction.
14034
14035 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
14036 peepholes to fuse an addis instruction with a load to a GPR base
14037 register. If we are supporting sign extending fusions, convert
14038 sign extending loads to zero extending loads and add an explicit
14039 sign extension.
14040
14041 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14042
14043 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
14044 aarch-common-protos.h to extra_headers.
14045 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14046 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
14047 (arm_early_store_addr_dep): Likewise.
14048 (arm_no_early_alu_shift_dep): Likewise.
14049 (arm_no_early_alu_shift_value_dep): Likewise.
14050 (arm_no_early_mul_dep): Likewise.
14051 (arm_no_early_store_addr_dep): Likewise.
14052 (arm_mac_accumulator_is_mul_result): Likewise.
14053 (arm_mac_accumulator_is_result): Likewise.
14054 * config/arm/aarch-common.c: ... here. New file.
14055 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
14056 here to ...
14057 (arm_early_store_addr_dep): Likewise.
14058 (arm_no_early_alu_shift_dep): Likewise.
14059 (arm_no_early_alu_shift_value_dep): Likewise.
14060 (arm_no_early_mul_dep): Likewise.
14061 (arm_no_early_store_addr_dep): Likewise.
14062 (arm_mac_accumulator_is_mul_result): Likewise.
14063 (arm_mac_accumulator_is_result): Likewise.
14064 * config/arm/aarch-common-protos.h: ... here. New file.
14065 * config/arm/t-arm (aarch-common.o): Define.
14066
14067 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14068
14069 * config/arm/arm.md: Include new file "types.md".
14070 (define_attr "type"): Move from here to ...
14071 (define_attr "mul32"): Likewise.
14072 (define_attr "mul64"): Likewise.
14073 * config/arm/types.md: ... here. New file.
14074
14075 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
14076
14077 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
14078 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
14079
14080 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
14081
14082 * gen-pass-instances.awk: Fix offset of substr().
14083
14084 2013-07-31 David Malcolm <dmalcolm@redhat.com>
14085
14086 * Makefile.in (pass-instances.def): New.
14087 (passes.o): Replace dependency on passes.def with one on
14088 pass-instances.def
14089
14090 * gen-pass-instances.awk: New.
14091
14092 * passes.c (pass_manager::pass_manager): Use pass-instances.def
14093 rather than passes.def, updating local definition of NEXT_PASS
14094 macro to add an extra NUM parameter (currently unused).
14095
14096 2013-07-30 David Malcolm <dmalcolm@redhat.com>
14097
14098 * Makefile.in (PASS_MANAGER_H): New.
14099 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
14100 (passes.o): Likewise.
14101 (statistics.o): Likewise.
14102 (cgraphunit.o): Likewise.
14103 (context.o): Depend on PASS_MANAGER_H.
14104
14105 * pass_manager.h: New.
14106
14107 * cgraphunit.c (cgraph_add_new_function): Update for moves
14108 of globals to fields of pass_manager.
14109 (analyze_function): Likewise.
14110 (expand_function): Likewise.
14111 (ipa_passes): Likewise.
14112 (compile): Likewise.
14113
14114 * context.c (context::context): New.
14115 * context.h (context::context): New.
14116 (context::get_passes): New.
14117 (context::passes_): New.
14118
14119 * lto-cgraph.c (input_node): Update for moves of globals to
14120 fields of pass_manager.
14121
14122 * passes.c (all_passes): Remove, in favor of a field of the
14123 same name within the new class pass_manager.
14124 (all_small_ipa_passes): Likewise.
14125 (all_lowering_passes): Likewise.
14126 (all_regular_ipa_passes): Likewise.
14127 (all_late_ipa_passes): Likewise.
14128 (all_lto_gen_passes): Likewise.
14129 (passes_by_id): Likewise.
14130 (passes_by_id_size): Likewise.
14131 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14132 the new class pass_manager.
14133 (set_pass_for_id): Convert to...
14134 (pass_manager::set_pass_for_id): ...method.
14135 (get_pass_for_id): Convert to...
14136 (pass_manager::get_pass_for_id): ...method.
14137 (register_one_dump_file): Move body of implementation into...
14138 (pass_manager::register_one_dump_file): ...here.
14139 (register_dump_files_1): Convert to...
14140 (pass_manager::register_dump_files_1): ...method.
14141 (register_dump_files): Convert to...
14142 (pass_manager::register_dump_files): ...method.
14143 (create_pass_tab): Update for moves of globals to fields of
14144 pass_manager.
14145 (dump_passes): Move body of implementation into...
14146 (pass_manager::dump_passes): ...here.
14147 (register_pass): Move body of implementation into...
14148 (pass_manager::register_pass): ...here.
14149 (init_optimization_passes): Convert into...
14150 (pass_manager::pass_manager): ...constructor for new
14151 pass_manager class, and initialize the pass_lists array.
14152 (check_profile_consistency): Update for moves of globals to
14153 fields of pass_manager.
14154 (dump_profile_report): Move body of implementation into...
14155 (pass_manager::dump_profile_report): ...here.
14156 (ipa_write_summaries_1): Update for moves of pass lists from
14157 being globals to fields of pass_manager.
14158 (ipa_write_optimization_summaries): Likewise.
14159 (ipa_read_summaries): Likewise.
14160 (ipa_read_optimization_summaries): Likewise.
14161 (execute_all_ipa_stmt_fixups): Likewise.
14162
14163 * statistics.c (statistics_fini): Update for moves of globals to
14164 fields of pass_manager.
14165
14166 * toplev.c (general_init): Replace call to
14167 init_optimization_passes with construction of the pass_manager
14168 instance.
14169
14170 * tree-pass.h (all_passes): Remove, in favor of a field of the
14171 same name within the new class pass_manager.
14172 (all_small_ipa_passes): Likewise.
14173 (all_lowering_passes): Likewise.
14174 (all_regular_ipa_passes): Likewise.
14175 (all_lto_gen_passes): Likewise.
14176 (all_late_ipa_passes): Likewise.
14177 (passes_by_id): Likewise.
14178 (passes_by_id_size): Likewise.
14179 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14180 the new class pass_manager.
14181 (get_pass_for_id): Remove.
14182
14183 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14184
14185 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
14186 configs.
14187
14188 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14189
14190 * arm.md (mulhi3): New expand pattern.
14191
14192 2013-07-30 Jan Hubicka <jh@suse.cz>
14193 Martin Liska <marxin.liska@gmail.com>
14194
14195 * profile.c (compute_value_histograms): Do not ICE when
14196 there is mismatch only on some counters.
14197
14198 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14199
14200 PR rtl-optimization/57637
14201 * function.c (move_insn_for_shrink_wrap): Also check the
14202 GEN set of the LIVE problem for the liveness analysis
14203 if it exists, otherwise give up.
14204
14205 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14206
14207 PR tree-optimization/57993
14208 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
14209 replaced statement in the candidate table.
14210 (phi_add_costs): Return infinite cost when the hidden basis does
14211 not dominate all phis on which the candidate is dependent.
14212 (replace_one_candidate): Record replaced statement in the
14213 candidate table.
14214
14215 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
14216
14217 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
14218 (ashlv2si3): New expander.
14219 (*ashlv2si3_i): New define_insn_and_split.
14220 * predicates.md (float_operation): Allow patterns with three
14221 basic sub-patterns.
14222
14223 PR rtl-optimization/58021
14224 * mode-switching.c (create_pre_exit): Always split off preceding
14225 insns if we are not at the basic block head.
14226
14227 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
14228
14229 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
14230 (UCLIBC_DYNAMIC_LINKER): New macro.
14231 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
14232 `-mnan=2008'.
14233 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
14234 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
14235 `-mnan=2008'.
14236 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
14237 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
14238 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
14239 for SF and DF modes. Use ieee_quad_format for TF mode.
14240 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
14241 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
14242 (mips_option_override): Handle `-mnan=legacy'.
14243 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
14244 `-mabs=2008' and `-mnan=2008'.
14245 (OPTION_DEFAULT_SPECS): Add "nan" default.
14246 (ASM_SPEC): Handle `-mnan='.
14247 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
14248 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
14249 comment accordingly.
14250 (neg<mode>2): Likewise.
14251 * config/mips/mips.opt (mabs, mnan): New options.
14252 * doc/install.texi (Configuration): Document `--with-nan=' option.
14253 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
14254 `-mnan=' options.
14255 (MIPS Options): Document them.
14256 * config.gcc <mips*-*-*>: Handle `--with-nan='.
14257 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
14258 * configure: Regenerate.
14259 * config.in: Regenerate.
14260
14261 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14262
14263 * config/i386/i386.md (float post-reload splitters): Do not check
14264 for subregs of SSE registers.
14265
14266 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14267 H.J. Lu <hongjiu.lu@intel.com>
14268
14269 PR target/57954
14270 PR target/57988
14271 * config/i386/i386.md (post-reload splitter
14272 to avoid partial SSE reg dependency stalls): New pattern.
14273
14274 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
14275
14276 * config/s390/s390.md ("movcc"): Swap load and store instructions.
14277
14278 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
14279
14280 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14281 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
14282
14283 2013-07-26 Cary Coutant <ccoutant@google.com>
14284
14285 * dwarf2out.c (die_checksum_ordered): Don't include template
14286 instantiations in signature.
14287 (is_template_parameter): New function.
14288 (is_template_instantiation): New function.
14289 (generate_skeleton_bottom_up): Don't include template instantiations
14290 in type unit DIE.
14291 (generate_skeleton): Likewise.
14292 (break_out_comdat_types): Move recursive call to break out nested
14293 types earlier.
14294 (prune_unused_types_mark_generic_parms_dies): Call
14295 is_template_parameter.
14296
14297 2013-07-26 Ian Bolton <ian.bolton@arm.com>
14298
14299 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
14300 uses vector registers.
14301 * config/aarch64/iterators.md: Add attributes rtn and vas.
14302
14303 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14304 Richard Earnshaw <richard.earnshaw@arm.com>
14305
14306 * combine.c (simplify_comparison): Re-canonicalize operands
14307 where appropriate.
14308 * config/arm/arm.md (movcond_addsi): New splitter.
14309
14310 2013-07-25 Sterling Augustine <saugustine@google.com>
14311
14312 * dwarf2out.c (size_of_pubnames): Move code to...
14313 (include_pubname_in_output): ...here. New.
14314 (want_pubnames): Rearrange.
14315 (output_pubnames): Call include_pubname_in_output. Move assertion.
14316
14317 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14318
14319 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
14320
14321 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14322
14323 PR target/38836
14324 * doc/extend.texi: Remove obsolete builtins. Fix
14325 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
14326
14327 2013-07-25 Jan Hubicka <jh@suse.cz>
14328
14329 * cgraph.c (release_function_body): Break out from ...
14330 (cgraph_release_function_body): ... this one; also release DECL_RESULT
14331 and DECL_ARGUMENTS.
14332 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
14333 old_tree in the map.
14334 (create_specialized_node): Update.
14335 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
14336 into index.
14337 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
14338 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
14339 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
14340 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
14341 DECL_RESULT.
14342
14343 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14344
14345 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
14346 addsi3_carryin_alt2_<optab>): Correct output template.
14347
14348 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14349
14350 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
14351 Adjust for arm_restrict_it.
14352 Remove trailing whitespace.
14353
14354 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
14355
14356 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
14357 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
14358
14359 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
14360
14361 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
14362
14363 PR rtl-optimization/57960
14364 * lra-constraints.c (process_alt_operands): Use the right mode
14365 when checking strict_low.
14366
14367 2013-07-25 Jan Hubicka <jh@suse.cz>
14368
14369 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
14370 * cgraph.c (cgraph_remove_node): Do not release function body
14371 when in cgraph streaming.
14372 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
14373 in other partitions are not considered reachable; fix handling of
14374 clones.
14375
14376 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14377
14378 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
14379
14380 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14381
14382 PR target/19599
14383 PR target/57731
14384 PR target/57837
14385 * config/arm/arm.md ("*sibcall_insn): Replace use of
14386 Ss with US. Adjust output for v5 and v4t.
14387 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
14388
14389 * config/arm/constraints.md ("Ss"): Rename to US.
14390
14391 2013-07-25 Terry Guo <terry.guo@arm.com>
14392
14393 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
14394 shift_add/shift_sub0/shift_sub1 RTXs.
14395
14396 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
14397 Anton Blanchard <anton@au1.ibm.com>
14398
14399 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
14400 (altivec_vpks<VI_char>ss): Likewise.
14401 (altivec_vpks<VI_char>us): Likewise.
14402 (altivec_vpku<VI_char>us): Likewise.
14403 (altivec_vpku<VI_char>um): Likewise.
14404
14405 2013-07-24 David Malcolm <dmalcolm@redhat.com>
14406
14407 Introduce context class.
14408
14409 * Makefile.in (CONTEXT_H): New.
14410 (OBJS): Add context.o.
14411 (toplev.o): Add CONTEXT_H to dependencies.
14412 (context.o): New.
14413
14414 * toplev.c (general_init): Create the singleton gcc::context instance.
14415
14416 * context.c: New.
14417
14418 * context.h: New.
14419
14420 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
14421
14422 PR rtl-optimization/57968
14423 * mode-switching.c (create_pre_exit): Allow instructions that
14424 don't set a return register to need a non-exit mode.
14425
14426 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14427 Anton Blanchard <anton@au1.ibm.com>
14428
14429 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
14430 operands to vperm for little endian.
14431 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
14432 of lvsl to create the control mask for a vperm for little endian.
14433
14434 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14435 Anton Blanchard <anton@au1.ibm.com>
14436
14437 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
14438 two operands for little-endian.
14439
14440 2013-07-23 Steve Ellcey <sellcey@mips.com>
14441
14442 * config/mips/mips.c (mips_case_values_threshold): New.
14443 (TARGET_CASE_VALUES_THRESHOLD): Define.
14444
14445 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14446 Anton Blanchard <anton@au1.ibm.com>
14447
14448 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
14449 selection of field for vector splat in little endian mode.
14450
14451 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
14452
14453 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
14454 expanders to rs6000.md.
14455 (ior<mode>3): Likewise.
14456 (and<mode>3): Likewise.
14457 (one_cmpl<mode>2): Likewise.
14458 (nor<mode>3): Likewise.
14459 (andc<mode>3): Likewise.
14460 (eqv<mode>3): Likewise.
14461 (nand<mode>3): Likewise.
14462 (orc<mode>3): Likewise.
14463
14464 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
14465 declaration.
14466
14467 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
14468 to split multi-word logical operations.
14469 (rs6000_split_logical_di): Likewise.
14470 (rs6000_split_logical): Likewise.
14471
14472 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
14473 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
14474 and allow TImode operations in 32-bit.
14475 (vsx_and<mode>3_64bit): Likewise.
14476 (vsx_ior<mode>3_32bit): Likewise.
14477 (vsx_ior<mode>3_64bit): Likewise.
14478 (vsx_xor<mode>3_32bit): Likewise.
14479 (vsx_xor<mode>3_64bit): Likewise.
14480 (vsx_one_cmpl<mode>2_32bit): Likewise.
14481 (vsx_one_cmpl<mode>2_64bit): Likewise.
14482 (vsx_nor<mode>3_32bit): Likewise.
14483 (vsx_nor<mode>3_64bit): Likewise.
14484 (vsx_andc<mode>3_32bit): Likewise.
14485 (vsx_andc<mode>3_64bit): Likewise.
14486 (vsx_eqv<mode>3_32bit): Likewise.
14487 (vsx_eqv<mode>3_64bit): Likewise.
14488 (vsx_nand<mode>3_32bit): Likewise.
14489 (vsx_nand<mode>3_64bit): Likewise.
14490 (vsx_orc<mode>3_32bit): Likewise.
14491 (vsx_orc<mode>3_64bit): Likewise.
14492
14493 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
14494 logical types in GPRs.
14495
14496 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
14497 logical insns to rs6000.md, and allow TImode operations in
14498 32-bit.
14499 (altivec_ior<mode>3): Likewise.
14500 (altivec_xor<mode>3): Likewise.
14501 (altivec_one_cmpl<mode>2): Likewise.
14502 (altivec_nor<mode>3): Likewise.
14503 (altivec_andc<mode>3): Likewise.
14504
14505 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
14506 attributes for moving the 128-bit logical operations into
14507 rs6000.md.
14508 (BOOL_REGS_OUTPUT): Likewise.
14509 (BOOL_REGS_OP1): Likewise.
14510 (BOOL_REGS_OP2): Likewise.
14511 (BOOL_REGS_UNARY): Likewise.
14512 (BOOL_REGS_AND_CR0): Likewise.
14513 (one_cmpl<mode>2): Add support for DI logical operations on
14514 32-bit, splitting the operations to 32-bit.
14515 (anddi3): Likewise.
14516 (iordi3): Likewise.
14517 (xordi3): Likewise.
14518 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
14519 changes to combine the 32/64-bit code, allow logical operations on
14520 TI mode in 32-bit, and to use similar match_operator patterns like
14521 scalar mode uses. Combine the Altivec and VSX code for logical
14522 operations, and move it here.
14523 (ior<mode>3, 128-bit types): Likewise.
14524 (xor<mode>3, 128-bit types): Likewise.
14525 (one_cmpl<mode>3, 128-bit types): Likewise.
14526 (nor<mode>3, 128-bit types): Likewise.
14527 (andc<mode>3, 128-bit types): Likewise.
14528 (eqv<mode>3, 128-bit types): Likewise.
14529 (nand<mode>3, 128-bit types): Likewise.
14530 (orc<mode>3, 128-bit types): Likewise.
14531 (and<mode>3_internal): Likewise.
14532 (bool<mode>3_internal): Likewise.
14533 (boolc<mode>3_internal1): Likewise.
14534 (boolc<mode>3_internal2): Likewise.
14535 (boolcc<mode>3_internal1): Likewise.
14536 (boolcc<mode>3_internal2): Likewise.
14537 (eqv<mode>3_internal1): Likewise.
14538 (eqv<mode>3_internal2): Likewise.
14539 (one_cmpl1<mode>3_internal): Likewise.
14540
14541 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14542
14543 * config/microblaze/microblaze.c (microblaze_expand_prologue):
14544 Rename flag_stack_usage to flag_stack_usage_info.
14545
14546 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14547
14548 * config/microblaze/sync.md: New file.
14549 * config/microblaze/microblaze.md: Include sync.md
14550 * config/microblaze/microblaze.c: Add print_operand 'y'.
14551 * config/microblaze/constraints.md: Add memory_contraint
14552 'Q' which is a single register.
14553
14554 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
14555
14556 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
14557
14558 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14559
14560 * reload.c (find_reloads): Exit loop once we find this operand
14561 cannot be reloaded somehow for this alternative.
14562
14563 * reload.c (find_reloads): Exit loop once we find a hard register.
14564
14565 * rtlanal.c (computed_jump_p): Exit loop once we find label
14566 reference is used.
14567
14568 * i386.c (ix86_pad_returns): Exit loop after setting replace.
14569
14570 * cfgloopmanip.c (remove_path): Exit loop after setting
14571 irred_invalidated.
14572
14573 * gensupport.c (subst_dup): Avoid loop if code is not
14574 MATCH_DUP nor MATCH_OP_DUP.
14575
14576 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
14577
14578 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
14579
14580 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14581
14582 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
14583 true for SP_REGNUM if mode == ptr_mode.
14584 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
14585 with value R0_REGNUM + 31.
14586
14587 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14588
14589 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
14590 pad pointer-typed argument downward.
14591
14592 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14593
14594 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
14595 and __ILP32__ when the ILP32 model is in use.
14596
14597 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14598
14599 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
14600 (aarch64_load_symref_appropriately): In the case of
14601 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
14602 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
14603 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
14604 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
14605 if 'mode' doesn't equal to 'ptr_mode'.
14606 (aarch64_output_mi_thunk): Add an assertion on the alignment of
14607 'vcall_offset'; change to call aarch64_emit_move differently depending
14608 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
14609 to calculate the upper bound of 'vcall_offset'.
14610 (aarch64_cannot_force_const_mem): Change to also return true if
14611 mode != ptr_mode.
14612 (aarch64_legitimize_reload_address): In the case of large
14613 displacements, add new local variable 'xmode' and an assertion
14614 based on it; change to use 'xmode' to generate the new rtx and
14615 reload.
14616 (aarch64_asm_trampoline_template): Change to generate the template
14617 differently depending on TARGET_ILP32 or not; change to use
14618 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
14619 (aarch64_trampoline_size): Removed.
14620 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
14621 and replace immediate literals with it. Change to use 'ptr_mode'
14622 instead of 'DImode' and call convert_memory_address if the mode
14623 of 'fnaddr' doesn't equal to 'ptr_mode'.
14624 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
14625 to output symbol.
14626 (aarch64_elf_asm_destructor): Likewise.
14627 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
14628 on TARGET_ILP32 instead of aarch64_trampoline_size.
14629 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
14630 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
14631 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
14632 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
14633 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
14634 (storewb_pair<GPI:mode>_<P:mode>): ... this.
14635 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
14636 depending on the value of 'mode'.
14637 (add_losym_<mode>): New.
14638 (ldr_got_small_<mode>): New, based on ldr_got_small.
14639 (ldr_got_small): Remove.
14640 (ldr_got_small_sidi): New.
14641 * config/aarch64/iterators.md (P): New.
14642 (PTR): Change to 'ptr_mode' in the condition.
14643
14644 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14645
14646 * config.gcc (aarch64*-*-*): Support --with-abi.
14647 (aarch64*-*-elf): Support --with-multilib-list.
14648 (aarch64*-*-linux*): Likewise.
14649 (supported_defaults): Add abi to aarch64*-*-*.
14650 * configure.ac: Mention AArch64 for --with-multilib-list.
14651 * configure: Re-generated.
14652 * config/aarch64/biarchilp32.h: New file.
14653 * config/aarch64/biarchlp64.h: New file.
14654 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
14655 (ABI_SPEC): Ditto.
14656 (MULTILIB_DEFAULTS): Ditto.
14657 (DRIVER_SELF_SPECS): Ditto.
14658 (ASM_SPEC): Update to also substitute -mabi.
14659 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
14660 file whose name depends on -mabi= and -mbig-endian.
14661 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
14662 TARGET_ILP32.
14663 (POINTER_SIZE): New define.
14664 (POINTERS_EXTEND_UNSIGNED): Ditto.
14665 (enum aarch64_abi_type): New enumeration tag.
14666 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
14667 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
14668 (TARGET_ILP32): New define.
14669 * config/aarch64/aarch64.opt (mabi): New.
14670 (aarch64_abi): New.
14671 (ilp32, lp64): New values for -mabi.
14672 * config/aarch64/t-aarch64 (comma): New define.
14673 (MULTILIB_OPTIONS): Ditto.
14674 (MULTILIB_DIRNAMES): Ditto.
14675 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
14676 * doc/invoke.texi: Document -mabi for AArch64.
14677
14678 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
14679
14680 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
14681
14682 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14683 Anton Blanchard <anton@au1.ibm.com>
14684
14685 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
14686 endianness when selecting field to splat.
14687
14688 2013-07-22 Eric Christopher <echristo@gmail.com>
14689
14690 * dwarf2out.c (die_odr_checksum): New function to use
14691 CHECKSUM_ macros and ULEB128 for DIE tag.
14692 (generate_type_signature): Use.
14693
14694 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
14695
14696 * config.gcc (sparc*-*-*): Accept leon3 processor.
14697 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
14698 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
14699 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
14700 * config/sparc/sparc.opt (enum processor_type): Add leon3.
14701 (mfix-ut699): Adjust comment.
14702 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
14703 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
14704 (CPP_CPU_SPEC): Likewise.
14705 (ASM_CPU_SPEC): Likewise.
14706 * config/sparc/sparc.c (leon3_cost): New constant.
14707 (sparc_option_override): Add leon3 support.
14708 (mem_ref): New function.
14709 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
14710 (sparc_do_work_around_errata): Look into the instruction in the delay
14711 slot and adjust accordingly. Add fix for the data cache nullify issues
14712 of the UT699. Change insertion position for the NOP.
14713 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
14714 (leon3_load): New reservation.
14715 (leon_store): Bump latency to 2.
14716 (grfpu): New automaton.
14717 (grfpu_alu): New unit.
14718 (grfpu_ds): Likewise.
14719 (leon_fp_alu): Adjust.
14720 (leon_fp_mult): Delete.
14721 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
14722 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
14723 * config/sparc/sparc.md (cpu): Add leon3.
14724 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
14725 (swapsi): Likewise.
14726 (atomic_test_and_set): Likewise.
14727 (ldstub): Likewise.
14728
14729 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
14730
14731 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
14732 default for R5900 targets.
14733 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
14734 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
14735 * config/mips/mips.c (mips_option_override): Report an error for
14736 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
14737 for -march=r5900 -mhard-float.
14738
14739 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14740
14741 * df-problems.c (can_move_insns_across): Exit loop once we
14742 find a non-fixed, non-global register.
14743
14744 * ipa-pure-const.c (propagate_nothrow): Exit loop after
14745 setting can_throw.
14746
14747 * omega.c (omega_eliminate_red): Break after setting red_found.
14748 (omega_problem_has_red_equations): Similarly after setting found.
14749 (omega_query_variable): Similarly after setting coupled.
14750
14751 2013-07-22 Marek Polacek <polacek@redhat.com>
14752
14753 * gimplify.c: Don't include gimple.h twice.
14754
14755 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14756
14757 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
14758 instead of TARGET_THUMB1.
14759 (Pz): New constraint.
14760 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
14761 encodings.
14762 (compare_negsi_si): Likewise.
14763 (compare_addsi2_op0): Likewise.
14764 (compare_addsi2_op1): Likewise.
14765 (addsi3_carryin_<optab>): Likewise.
14766 (addsi3_carryin_alt2_<optab>): Likewise.
14767 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
14768 for arm_restrict_it.
14769 (subsi3_carryin): Likewise.
14770 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
14771 (minmax_arithsi): Disable for arm_restrict_it.
14772 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
14773 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
14774 (satsi_<SAT:code>_shift): Likewise.
14775 (arm_shiftsi3): Add alternative for 16-bit encoding.
14776 (arm32_movhf): Disable for arm_restrict_it.
14777 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
14778 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
14779
14780 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
14781
14782 * config/arm/arm.md (attribute "insn"): Delete.
14783 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
14784 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
14785 (not_shiftsi): Update for attribute change.
14786 (not_shiftsi_compare0): Likewise.
14787 (not_shiftsi_compare0_scratch): Likewise.
14788 (arm_one_cmplsi2): Likewise.
14789 (thumb1_one_cmplsi2): Likewise.
14790 (notsi_compare0): Likewise.
14791 (notsi_compare0_scratch): Likewise.
14792 (thumb1_movdi_insn): Likewise.
14793 (arm_movsi_insn): Likewise.
14794 (movhi_insn_arch4): Likewise.
14795 (movhi_bytes): Likewise.
14796 (arm_movqi_insn): Likewise.
14797 (thumb1_movqi_insn): Likewise.
14798 (arm32_movhf): Likewise.
14799 (thumb1_movhf): Likewise.
14800 (arm_movsf_soft_insn): Likewise.
14801 (thumb1_movsf_insn): Likewise.
14802 (thumb_movdf_insn): Likewise.
14803 (movsicc_insn): Likewise.
14804 (movsfcc_soft_insn): Likewise.
14805 (and_scc): Likewise.
14806 (cond_move): Likewise.
14807 (if_move_not): Likewise.
14808 (if_not_move): Likewise.
14809 (if_shift_move): Likewise.
14810 (if_move_shift): Likewise.
14811 (if_shift_shift): Likewise.
14812 (if_not_arith): Likewise.
14813 (if_arith_not): Likewise.
14814 (cond_move_not): Likewise.
14815 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
14816 (neon_mov<mode>): Likewise.
14817 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
14818 (thumb2_movsi_vfp): Likewise.
14819 (movsf_vfp): Likewise.
14820 (thumb2_movsf_vfp): Likewise.
14821 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
14822 change.
14823 (cortexa7_older_only): Likewise.
14824 (cortexa7_younger): Likewise.
14825 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
14826 (1020alu_shift_op): Likewise.
14827 (1020alu_shift_reg_op): Likewise.
14828 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
14829 (alu_shift_op): Likewise.
14830 (alu_shift_reg_op): Likewise.
14831 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
14832 (11_alu_shift_op): Likewise.
14833 (11_alu_shift_reg_op): Likewise.
14834 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
14835 (9_alu_shift_reg_op): Likewise.
14836 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
14837 change.
14838 (cortex_a15_alu_shift): Likewise.
14839 (cortex_a15_alu_shift_reg): Likewise.
14840 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
14841 (cortex_a5_alu_shift): Likewise.
14842 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
14843 change.
14844 (cortex_a53_alu_shift): Likewise.
14845 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
14846 change.
14847 (cortex_a7_alu_reg): Likewise.
14848 (cortex_a7_alu_shift): Likewise.
14849 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
14850 (cortex_a8_alu_shift): Likewise.
14851 (cortex_a8_alu_shift_reg): Likewise.
14852 (cortex_a8_mov): Likewise.
14853 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
14854 (cortex_a9_dp_shift): Likewise.
14855 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
14856 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
14857 (cortex_r4_mov): Likewise.
14858 (cortex_r4_alu_shift): Likewise.
14859 (cortex_r4_alu_shift_reg): Likewise.
14860 * config/arm/fa526.md (526_alu_op): Update for attribute change.
14861 (526_alu_shift_op): Likewise.
14862 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
14863 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
14864 (626te_alu_shift_op): Likewise.
14865 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
14866 (726te_alu_op): Likewise.
14867 (726te_alu_shift_op): Likewise.
14868 (726te_alu_shift_reg_op): Likewise.
14869 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
14870 (mp626_alu_shift_op): Likewise.
14871 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
14872 (pj4_alu_e1_conds): Likewise.
14873 (pj4_alu): Likewise.
14874 (pj4_alu_conds): Likewise.
14875 (pj4_shift): Likewise.
14876 (pj4_shift_conds): Likewise.
14877 (pj4_alu_shift): Likewise.
14878 (pj4_alu_shift_conds): Likewise.
14879
14880 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14881
14882 * config/arm/predicates.md (shiftable_operator_strict_it):
14883 New predicate.
14884 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
14885 Disable cond_exec version for arm_restrict_it.
14886 (thumb2_smaxsi3): Convert to generate cond_exec.
14887 (thumb2_sminsi3): Likewise.
14888 (thumb32_umaxsi3): Likewise.
14889 (thumb2_uminsi3): Likewise.
14890 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
14891 (thumb2_neg_abssi2): Likewise.
14892 (thumb2_mov_scc): Add alternative for 16-bit encoding.
14893 (thumb2_movsicc_insn): Adjust alternatives.
14894 (thumb2_mov_negscc): Disable for arm_restrict_it.
14895 (thumb2_mov_negscc_strict_it): New pattern.
14896 (thumb2_mov_notscc_strict_it): New pattern.
14897 (thumb2_mov_notscc): Disable for arm_restrict_it.
14898 (thumb2_ior_scc): Likewise.
14899 (thumb2_ior_scc_strict_it): New pattern.
14900 (thumb2_cond_move): Adjust for arm_restrict_it.
14901 (thumb2_cond_arith): Disable for arm_restrict_it.
14902 (thumb2_cond_arith_strict_it): New pattern.
14903 (thumb2_cond_sub): Adjust for arm_restrict_it.
14904 (thumb2_movcond): Likewise.
14905 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
14906 (thumb2_zero_extendhisi2_v6): Likewise.
14907 (thumb2_zero_extendqisi2_v6): Likewise.
14908 (orsi_notsi_si): Likewise.
14909 (orsi_not_shiftsi_si): Likewise.
14910
14911 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
14912
14913 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
14914 instruction sequence is 1 byte shorter.
14915
14916 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
14917
14918 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
14919 it is not needed after split.
14920
14921 2013-07-20 Iain Sandoe <iain@codesourcery.com>
14922
14923 PR target/51784
14924 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
14925 second label for nonlocal goto receivers. Don't output pic base labels
14926 unless we're producing PIC; mark that action unreachable().
14927 (ix86_save_reg): If the function contains a nonlocal label, save the
14928 PIC base reg.
14929 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
14930 * config/darwin.c (emitted_pic_label_num): New GTY.
14931 (update_pic_label_number_if_needed): New.
14932 (machopic_output_function_base_name): Adjust for nonlocal receiver
14933 case.
14934 (machopic_should_output_picbase_label): New.
14935 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
14936 (nonlocal_goto_receiver): New insn and split.
14937
14938 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
14939
14940 * config/aarch64/aarch64-builtins.c
14941 (aarch64_fold_builtin): Fold abs in all modes.
14942 * config/aarch64/aarch64-simd-builtins.def
14943 (abs): Enable for all modes.
14944 * config/aarch64/arm_neon.h
14945 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
14946 (vabs_f64): Add missing intrinsic.
14947
14948 2013-07-19 Ian Bolton <ian.bolton@arm.com>
14949
14950 * config/aarch64/arm_neon.h (vabs_s64): New function
14951
14952 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
14953
14954 PR target/57516
14955 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
14956 * config/avr/avr.md (adjust_len): Add `round'.
14957 * config/avr/avr-protos.h (avr_out_round): New prototype.
14958 (avr_out_plus): Add `out_label' argument.
14959 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
14960 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
14961 Handle the case where `insn' is just a pattern.
14962 (avr_out_bitop): Handle the case where `insn' is just a pattern.
14963 (avr_out_round): New function.
14964 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
14965
14966 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
14967
14968 * config/microblaze/microblaze.c (microblaze_expand_prologue):
14969 Add check for flag_stack_usage to handle -fstack-usage support
14970
14971 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
14972
14973 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
14974 interaction for new Power8 flags and VSX.
14975
14976 2013-07-18 Sriraman Tallam <tmsriram@google.com>
14977
14978 PR middle-end/57698
14979 * tree-inline.c (expand_call_inline): Emit errors during
14980 early_inlining only if optimization is not turned on.
14981
14982 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14983
14984 * passes.def: New.
14985
14986 * passes.c (init_optimization_passes): Move the construction of
14987 the pass hierarchy into a new passes.def file.
14988
14989 * Makefile.in (passes.o): Add dependency on passes.def.
14990
14991 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14992
14993 * passes.c (init_optimization_passes): Introduce macros for
14994 constructing the tree of passes (INSERT_PASSES_AFTER,
14995 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
14996 TERMINATE_PASS_LIST).
14997
14998 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
14999 Wei Mi <wmi@google.com>
15000
15001 PR rtl-optimization/57878
15002 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
15003 top.
15004 (reload_pseudo_compare_func): Check nregs first for reload
15005 pseudos.
15006
15007 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15008
15009 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
15010
15011 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
15012
15013 * read-rtl.c (validate_const_int): Once an invalid character is
15014 seen, quit the loop.
15015
15016 * gengtype.c (write_roots): Similarly once we find the "deletable"
15017 or "if_marked" option.
15018
15019 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15020
15021 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
15022 "xtab" and "sat". Move value "clz" from here to ...
15023 (attriubte "type"): ... here.
15024 (satsi_<SAT:code>): Delete "insn" attribute.
15025 (satsi_<SAT:code>_shift): Likewise.
15026 (arm_zero_extendqisi2addsi): Likewise.
15027 (arm_extendqisi2addsi): Likewise.
15028 (clzsi2): Update for attribute changes.
15029 (rbitsi2): Likewise.
15030 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
15031 attribute.
15032 (arm_usatsihi): Likewise.
15033 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15034
15035 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15036
15037 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
15038 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
15039 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
15040 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
15041 in alphabetical order.
15042 (attribute "core_cycles"): Update for attribute changes.
15043 (arm_addsi3): Likewise.
15044 (addsi3_compare0): Likewise.
15045 (addsi3_compare0_scratch): Likewise.
15046 (addsi3_compare_op1): Likewise.
15047 (addsi3_compare_op2): Likewise.
15048 (compare_addsi2_op0): Likewise.
15049 (compare_addsi2_op1): Likewise.
15050 (addsi3_carryin_shift_<optab>): Likewise.
15051 (subsi3_carryin_shift): Likewise.
15052 (rsbsi3_carryin_shift): Likewise.
15053 (arm_subsi3_insn): Likewise.
15054 (subsi3_compare0): Likewise.
15055 (subsi3_compare): Likewise.
15056 (arm_andsi3_insn): Likewise.
15057 (thumb1_andsi3_insn): Likewise.
15058 (andsi3_compare0): Likewise.
15059 (andsi3_compare0_scratch): Likewise.
15060 (zeroextractsi_compare0_scratch
15061 (andsi_not_shiftsi_si): Likewise.
15062 (iorsi3_insn): Likewise.
15063 (iorsi3_compare0): Likewise.
15064 (iorsi3_compare0_scratch): Likewise.
15065 (arm_xorsi3): Likewise.
15066 (thumb1_xorsi3_insn): Likewise.
15067 (xorsi3_compare0): Likewise.
15068 (xorsi3_compare0_scratch): Likewise.
15069 (satsi_<SAT:code>_shift): Likewise.
15070 (rrx): Likewise.
15071 (arm_shiftsi3): Likewise.
15072 (shiftsi3_compare0): Likewise.
15073 (not_shiftsi): Likewise.
15074 (not_shiftsi_compare0): Likewise.
15075 (not_shiftsi_compare0_scratch): Likewise.
15076 (arm_one_cmplsi2): Likewise.
15077 (thumb_one_complsi2): Likewise.
15078 (notsi_compare0): Likewise.
15079 (notsi_compare0_scratch): Likewise.
15080 (thumb1_zero_extendhisi2): Likewise.
15081 (arm_zero_extendhisi2): Likewise.
15082 (arm_zero_extendhisi2_v6): Likewise.
15083 (arm_zero_extendhisi2addsi): Likewise.
15084 (thumb1_zero_extendqisi2): Likewise.
15085 (thumb1_zero_extendqisi2_v6): Likewise.
15086 (arm_zero_extendqisi2): Likewise.
15087 (arm_zero_extendqisi2_v6): Likewise.
15088 (arm_zero_extendqisi2addsi): Likewise.
15089 (thumb1_extendhisi2): Likewise.
15090 (arm_extendhisi2): Likewise.
15091 (arm_extendhisi2_v6): Likewise.
15092 (arm_extendqisi): Likewise.
15093 (arm_extendqisi_v6): Likewise.
15094 (arm_extendqisi2addsi): Likewise.
15095 (thumb1_extendqisi2): Likewise.
15096 (thumb1_movdi_insn): Likewise.
15097 (arm_movsi_insn): Likewise.
15098 (movsi_compare0): Likewise.
15099 (movhi_insn_arch4): Likewise.
15100 (movhi_bytes): Likewise.
15101 (arm_movqi_insn): Likewise.
15102 (thumb1_movqi_insn): Likewise.
15103 (arm32_movhf): Likewise.
15104 (thumb1_movhf): Likewise.
15105 (arm_movsf_soft_insn): Likewise.
15106 (thumb1_movsf_insn): Likewise.
15107 (movdf_soft_insn): Likewise.
15108 (thumb_movdf_insn): Likewise.
15109 (arm_cmpsi_insn): Likewise.
15110 (cmpsi_shiftsi): Likewise.
15111 (cmpsi_shiftsi_swp): Likewise.
15112 (arm_cmpsi_negshiftsi_si): Likewise.
15113 (movsicc_insn): Likewise.
15114 (movsfcc_soft_insn): Likewise.
15115 (arith_shiftsi): Likewise.
15116 (arith_shiftsi_compare0
15117 (arith_shiftsi_compare0_scratch
15118 (sub_shiftsi): Likewise.
15119 (sub_shiftsi_compare0
15120 (sub_shiftsi_compare0_scratch
15121 (and_scc): Likewise.
15122 (cond_move): Likewise.
15123 (if_plus_move): Likewise.
15124 (if_move_plus): Likewise.
15125 (if_move_not): Likewise.
15126 (if_not_move): Likewise.
15127 (if_shift_move): Likewise.
15128 (if_move_shift): Likewise.
15129 (if_shift_shift): Likewise.
15130 (if_not_arith): Likewise.
15131 (if_arith_not): Likewise.
15132 (cond_move_not): Likewise.
15133 (thumb1_ashlsi3): Set type attribute.
15134 (thumb1_ashrsi3): Likewise.
15135 (thumb1_lshrsi3): Likewise.
15136 (thumb1_rotrsi3): Likewise.
15137 (shiftsi3_compare0_scratch): Likewise.
15138 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
15139 (neon_mov<mode>): Likewise.
15140 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
15141 attribute changes.
15142 (thumb2_movsi_insn): Likewise.
15143 (thumb2_cmpsi_neg_shiftsi): Likewise.
15144 (thumb2_extendqisi_v6): Likewise.
15145 (thumb2_zero_extendhisi2_v6): Likewise.
15146 (thumb2_zero_extendqisi2_v6): Likewise.
15147 (thumb2_shiftsi3_short): Likewise.
15148 (thumb2_addsi3_compare0_scratch): Likewise.
15149 (orsi_not_shiftsi_si): Likewise.
15150 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
15151 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
15152 changes.
15153 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
15154 (1020alu_shift_op): Likewise.
15155 (1020alu_shift_reg_op): Likewise.
15156 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
15157 (alu_shift_op): Likewise.
15158 (alu_shift_reg_op): Likewise.
15159 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
15160 (11_alu_shift_op): Likewise.
15161 (11_alu_shift_reg_op): Likewise.
15162 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
15163 (9_alu_shift_reg_op): Likewise.
15164 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15165 changes.
15166 (cortex_a15_alu_shift): Likewise.
15167 (cortex_a15_alu_shift_reg): Likewise.
15168 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
15169 changes.
15170 (cortex_a5_alu_shift): Likewise.
15171 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
15172 changes.
15173 (cortex_a53_alu_shift): Likewise.
15174 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15175 changes.
15176 (cortex_a7_alu_reg): Likewise.
15177 (cortex_a7_alu_shift): Likewise.
15178 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
15179 changes.
15180 (cortex_a8_alu_shift): Likewise.
15181 (cortex_a8_alu_shift_reg): Likewise.
15182 (cortex_a8_mov): Likewise.
15183 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
15184 (cortex_a9_dp_shift): Likewise.
15185 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
15186 changes.
15187 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
15188 changes.
15189 (cortex_r4_mov): Likewise.
15190 (cortex_r4_alu_shift): Likewise.
15191 (cortex_r4_alu_shift_reg): Likewise.
15192 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
15193 (526_alu_shift_op): Likewise.
15194 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
15195 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
15196 (626te_alu_shift_op): Likewise.
15197 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
15198 (726te_alu_op): Likewise.
15199 (726te_alu_shift_op): Likewise.
15200 (726te_alu_shift_reg_op): Likewise.
15201 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
15202 (mp626_alu_shift_op): Likewise.
15203 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
15204 (pj4_alu_e1_conds): Likewise.
15205 (pj4_alu): Likewise.
15206 (pj4_alu_conds): Likewise.
15207 (pj4_shift): Likewise.
15208 (pj4_shift_conds): Likewise.
15209 (pj4_alu_shift): Likewise.
15210 (pj4_alu_shift_conds): Likewise.
15211 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15212 changes.
15213 (cortexa7_older_only): Likewise.
15214 (cortexa7_younger): Likewise.
15215
15216 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15217
15218 * ipa-pure-const.c (generate_summary): Rename to...
15219 (pure_const_generate_summary): ... this.
15220
15221 2013-07-17 Iain Sandoe <iain@codesourcery.com>
15222
15223 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
15224
15225 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
15226
15227 PR target/57909
15228 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
15229 usage in HI mode.
15230
15231 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15232
15233 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
15234 enabled without -march=zEC12.
15235 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
15236 flags to be set.
15237
15238 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15239
15240 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
15241 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
15242 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
15243 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
15244 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
15245 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
15246 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
15247 ISA_HAS_FP4.
15248 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
15249 and ISA_HAS_NMADD3_NMSUB3.
15250 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
15251 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
15252 (nmsub4<mode>, nmsub3<mode>): Likewise.
15253 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
15254
15255 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15256
15257 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
15258 TARGET_MIPS5400 checking.
15259
15260 2013-07-16 Jakub Jelinek <jakub@redhat.com>
15261 Peter Bergner <bergner@vnet.ibm.com>
15262
15263 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
15264 registers in the comment.
15265 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
15266 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
15267 rather than FIRST_PSEUDO_REGISTERS.
15268
15269 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
15270
15271 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
15272 enable extra ISA flags with TARGET_HTM.
15273
15274 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15275
15276 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
15277 Fix comment typos.
15278
15279 2013-07-15 Cong Hou <congh@google.com>
15280
15281 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
15282 in compare function for sorting.
15283
15284 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
15285
15286 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
15287 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
15288 * config/rs6000/rs6000.opt: Add -mhtm option.
15289 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
15290 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
15291 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
15292 __HTM__ if the HTM instructions are available.
15293 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
15294 htm_spr_reg_operand): New define_predicates.
15295 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
15296 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
15297 Include htm.md.
15298 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
15299 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
15300 HTM builtin functions.
15301 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
15302 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
15303 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
15304 (rs6000_builtin_mask_calculate): Likewise.
15305 (rs6000_option_override_internal): Likewise.
15306 (bdesc_htm): Add new HTM builtin support.
15307 (htm_spr_num): New function.
15308 (htm_spr_regno): Likewise.
15309 (rs6000_htm_spr_icode): Likewise.
15310 (htm_expand_builtin): Likewise.
15311 (htm_init_builtins): Likewise.
15312 (rs6000_expand_builtin): Add support for HTM builtin functions.
15313 (rs6000_init_builtins): Likewise.
15314 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
15315 option.
15316 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
15317 (TARGET_HTM, MASK_HTM): Define macros.
15318 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
15319 (FIXED_REGISTERS): Likewise.
15320 (CALL_USED_REGISTERS): Likewise.
15321 (CALL_REALLY_USED_REGISTERS): Likewise.
15322 (REG_ALLOC_ORDER): Likewise.
15323 (enum reg_class): Likewise.
15324 (REG_CLASS_NAMES): Likewise.
15325 (REG_CLASS_CONTENTS): Likewise.
15326 (REGISTER_NAMES): Likewise.
15327 (ADDITIONAL_REGISTER_NAMES): Likewise.
15328 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
15329 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
15330 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
15331 * config/rs6000/htm.md: New file.
15332 * config/rs6000/htmintrin.h: New file.
15333 * config/rs6000/htmxlintrin.h: New file.
15334
15335 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
15336
15337 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15338 Define SYMBOL_TINY_GOT, update comment.
15339 * config/aarch64/aarch64.c
15340 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
15341 (aarch64_expand_mov_immediate): Likewise.
15342 (aarch64_print_operand): Likewise.
15343 (aarch64_classify_symbol): Likewise.
15344 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
15345 (ldr_got_tiny): Define.
15346
15347 2013-07-13 Tobias Grosser <tobias@grosser.es>
15348
15349 PR tree-optimization/54094
15350 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
15351 scheduling dimension for the parallelism check from the polyhedral
15352 information in the AST.
15353 * graphite-dependences.c (carries_deps): Do not assume the schedule is
15354 in 2D + 1 form.
15355
15356 2013-07-13 Jason Merrill <jason@redhat.com>
15357
15358 * print-tree.c (debug_vec_tree): Use debug_raw.
15359 (debug_raw (vec<tree, va_gc> &)): New.
15360 (debug_raw (vec<tree, va_gc> *)): New.
15361 * tree.h: Declare them.
15362
15363 2013-07-13 Bin Cheng <bin.cheng@arm.com>
15364
15365 * ifcvt.c (ifcvt_after_combine): New static variable.
15366 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
15367 for size.
15368 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
15369 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
15370 rest_of_handle_if_after_reload): Pass new argument for if_convert.
15371
15372 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
15373
15374 * config/mips/mips.c (mips_expand_call): Remove empty statement.
15375
15376 2013-07-12 Michael Matz <matz@suse.de>
15377
15378 PR middle-end/55771
15379 * convert.c (convert_to_real): Reject non-float inner types.
15380
15381 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
15382
15383 * config/aarch64/aarch64-protos.h
15384 (aarch64_simd_immediate_valid_for_move): Remove.
15385 * config/aarch64/aarch64.c (simd_immediate_info): New member.
15386 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
15387 cases.
15388 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
15389
15390 2013-07-11 Steve Ellcey <sellcey@mips.com>
15391
15392 * config/mips/mips.c (mips_conditional_register_usage): Do not
15393 use t[0-7] registers in MIPS16 mode when optimizing for size.
15394
15395 2013-07-11 Sriraman Tallam <tmsriram@google.com>
15396
15397 * config/i386/i386.c (dispatch_function_versions): Fix array
15398 indexing of function_version_info to match actual_versions.
15399
15400 2013-07-11 Teresa Johnson <tejohnson@google.com>
15401
15402 * vec.h (struct va_gc): Move release out-of-line.
15403 (va_gc::release): Call ggc_free on released vec.
15404
15405 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15406
15407 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
15408 Require GOT register as additional operand in UNSPEC.
15409 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
15410 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15411 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15412 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
15413 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
15414 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
15415 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
15416
15417 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
15418
15419 PR target/57631
15420 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
15421 name seen by assembler/linker rather if available.
15422
15423 2013-07-11 Andreas Schwab <schwab@suse.de>
15424
15425 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
15426
15427 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
15428
15429 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
15430
15431 2013-07-10 Joseph Myers <joseph@codesourcery.com>
15432
15433 * doc/tm.texi.in: Move hook documentation to ....
15434 * target.def: ... here.
15435
15436 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
15437 text on @hook line.
15438 * doc/tm.texi: Regenerate.
15439
15440 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
15441
15442 PR c++/57869
15443 * doc/invoke.texi: Document -Wconditionally-supported.
15444
15445 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15446
15447 PR target/57844
15448 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
15449 of my_fp.
15450
15451 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15452
15453 PR target/57506
15454 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
15455 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
15456 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
15457 Remove duplicate devices.
15458 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
15459 * config/avr/t-multilib: Regenerate.
15460 * config/avr/avr-tables.opt: Regenerate.
15461 * doc/avr-mmcu.texi: Regenerate.
15462
15463 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15464
15465 PR target/56987
15466 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
15467
15468 2013-07-10 Graham Stott <graham.stott@btinternet.com>
15469
15470 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
15471 the cost of MULT when optimizing for size.
15472
15473 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
15474
15475 * config/cr16/cr16-protos.h: Don't include target.h.
15476
15477 2013-07-09 Joseph Myers <joseph@codesourcery.com>
15478
15479 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
15480 adjust register size for TDmode and TFmode for VSX registers.
15481
15482 2013-07-08 Kai Tietz <ktietz@redhat.com>
15483
15484 PR target/56892
15485 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
15486 hook_bool_const_tree_true.
15487
15488 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15489
15490 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
15491 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
15492 * config/s390/s390.md: Define FPR*_REGNUM constants.
15493 Fix FPR2_REGNUM constant (18 -> 17).
15494 ("*trunc<BFP:mode><DFP_ALL:mode>2")
15495 ("*trunc<DFP_ALL:mode><BFP:mode>2")
15496 ("trunc<BFP:mode><DFP_ALL:mode>2")
15497 ("trunc<DFP_ALL:mode><BFP:mode>2")
15498 ("*extend<BFP:mode><DFP_ALL:mode>2")
15499 ("*extend<DFP_ALL:mode><BFP:mode>2")
15500 ("extend<BFP:mode><DFP_ALL:mode>2")
15501 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
15502 FPR4_REGNUM.
15503
15504 2013-07-08 Graham Stott <graham.stott@btinternet.com>
15505
15506 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
15507
15508 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15509
15510 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
15511 and cfun_fpr_bit_p to cfun_fpr_save_p.
15512 (s390_frame_area, s390_register_info, s390_frame_info)
15513 (s390_emit_prologue, s390_emit_epilogue)
15514 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
15515 register numbers.
15516 * config/s390/s390.h: Define *_REGNUM macros for floating point
15517 register numbers.
15518
15519 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
15520
15521 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
15522
15523 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
15524
15525 PR rtl-optimization/57786
15526 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
15527 and break out of the loop when it is set to false.
15528
15529 2013-07-08 Jakub Jelinek <jakub@redhat.com>
15530
15531 PR target/57819
15532 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
15533 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
15534 (const_int 63)) 0)).
15535 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
15536 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
15537 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
15538
15539 PR rtl-optimization/57829
15540 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
15541 mask bits outside of mode are just sign-extension from mode to HWI.
15542
15543 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
15544
15545 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
15546 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
15547 adjust_address instead of change_address to keep info about alignment.
15548 (emit_strmov): Remove.
15549 (emit_memmov): New function.
15550 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
15551 (expand_movmem_epilogue): Likewise and return updated rtx for
15552 destination.
15553 (expand_constant_movmem_prologue): Likewise and return updated rtx for
15554 destination and source.
15555 (decide_alignment): Refactor, handle vector_loop.
15556 (ix86_expand_movmem): Likewise.
15557 (ix86_expand_setmem): Likewise.
15558 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
15559
15560 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
15561
15562 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
15563 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
15564
15565 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15566
15567 * config/i386/sse.md (sse_movlhps): Change alternative 3
15568 of operand 2 to "m".
15569
15570 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15571
15572 PR target/57807
15573 * config/i386/sse.md (iptr): New mode attribute.
15574 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
15575 (<sse>_vm<plusminus_insn><mode>3): Ditto.
15576 (<sse>_vmmul<mode>3): Ditto.
15577 (<sse>_vmdiv<mode>3): Ditto.
15578 (sse_vmrcpv4sf2): Ditto.
15579 (<sse>_vmsqrt<mode>2): Ditto.
15580 (sse_vmrsqrtv4sf2): Ditto.
15581 (<sse>_vm<code><mode>3): Ditto.
15582 (avx_vmcmp<mode>3): Ditto.
15583 (<sse>_vmmaskcmp<mode>3): Ditto.
15584 (<sse>_comi): Ditto.
15585 (<sse>_ucomi): Ditto.
15586 (*xop_vmfrcz_<mode>): Ditto.
15587 (*fmai_fmadd_<mode>): Ditto.
15588 (*fmai_fmsub_<mode>): Ditto.
15589 (*fmai_fnmadd_<mode>): Ditto.
15590 (*fmai_fnmsub_<mode>): Ditto.
15591 (*fma4i_vmfmadd_<mode>): Ditto.
15592 (*fma4i_vmfmsub_<mode>): Ditto.
15593 (*fma4i_vmfnmadd_<mode>): Ditto.
15594 (*fma4i_vmfnmsub_<mode>): Ditto.
15595 (*xop_vmfrcz_<mode>): Ditto.
15596 (sse_cvtps2pi): Ditto.
15597 (sse_cvttps2pi): Ditto.
15598 (sse_cvtss2si): Ditto.
15599 (sse_cvtss2si_2): Ditto.
15600 (sse_cvtss2siq_2): Ditto.
15601 (sse_cvttss2si): Ditto.
15602 (sse_cvttss2siq): Ditto.
15603 (sse_cvtsd2si): Ditto.
15604 (sse_cvtsd2si_2): Ditto.
15605 (sse_cvtsd2siq_2): Ditto.
15606 (sse_cvttsd2si): Ditto.
15607 (sse_cvttsd2siq): Ditto.
15608 (sse_cvtsd2ss): Ditto.
15609 (sse_cvtss2sd): Ditto.
15610 (avx2_pbroadcast<mode>): Ditto.
15611 (avx2_pbroadcast<mode>_1): Ditto.
15612 (*avx_vperm_broadcast_v4sf): Ditto.
15613
15614 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
15615 (sse_movlhps): Ditto.
15616 (sse_storehps): Ditto.
15617 (sse_loadhps): Ditto.
15618 (sse_storelps): Ditto.
15619 (sse_loadlps): Ditto.
15620 (*vec_concatv4sf): Ditto.
15621 (*vec_interleave_highv2df): Ditto.
15622 (*vec_interleave_lowv2df): Ditto.
15623 (*vec_extractv2df_1_sse): Ditto.
15624 (*vec_extractv2df_0_sse): Ditto.
15625 (sse2_storelpd): Ditto.
15626 (sse2_loadlpd): Ditto.
15627 (sse2_movsd): Ditto.
15628 (*vec_concatv4si): Ditto.
15629 (vec_concatv2di): Ditto.
15630
15631 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
15632 for Intel asm dialect.
15633 (mmx_punpcklwd): Ditto.
15634 (mmx_punpckldq): Ditto.
15635
15636 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
15637 for intel assembler dialect.
15638
15639 2013-07-06 Jakub Jelinek <jakub@redhat.com>
15640
15641 PR target/29776
15642 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
15643 for BUILT_IN_C{LZ,LRSB}*.
15644 * tree.h (CASE_INT_FN): Add FN##IMAX case.
15645 * tree-vrp.c (extract_range_basic): Handle
15646 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
15647 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
15648 fall thru to code calling set_value*.
15649 * builtins.c (expand_builtin): Remove *IMAX cases.
15650 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
15651 if width is bigger than 2*HWI.
15652
15653 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
15654
15655 PR rtl-optimization/55342
15656 * lra-int.h (lra_subreg_reload_pseudos): New.
15657 * lra.c: Add undoing optional reloads to the block diagram.
15658 (lra_subreg_reload_pseudos): New.
15659 (lra_optional_reload_pseudos): Change comments.
15660 (lra): Init and clear lra_subreg_reload_pseudos. Clear
15661 lra_optional_reload_pseudos after undo transformations.
15662 * lra-assigns.c (pseudo_prefix_title): New.
15663 (lra_setup_reg_renumber): Use it.
15664 (spill_for): Ditto. Check subreg reload pseudos too.
15665 (assign_by_spills): Consider subreg reload pseudos too.
15666 * lra-constraints.c (simplify_operand_subreg): Use
15667 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
15668 (curr_insn_transform): Recognize and do optional reloads.
15669 (undo_optional_reloads): New.
15670 (lra_undo_inheritance): Call undo_optional_reloads.
15671
15672 2013-07-05 Thomas Quinot <quinot@adacore.com>
15673
15674 * tree-complex.c (expand_complex_operations_1): Fix typo.
15675
15676 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
15677
15678 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
15679 (tune_params): New member 'const vec_costs'.
15680 * config/aarch64/aarch64.c (generic_vector_cost): New.
15681 (generic_tunings): New member 'generic_vector_cost'.
15682 (aarch64_builtin_vectorization_cost): New.
15683 (aarch64_add_stmt_cost): New.
15684 (TARGET_VECTORIZE_ADD_STMT_COST): New.
15685 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
15686
15687 2013-07-03 Jakub Jelinek <jakub@redhat.com>
15688
15689 PR target/57777
15690 * config/i386/predicates.md (vsib_address_operand): Disallow
15691 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
15692
15693 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
15694
15695 PR middle-end/55030
15696 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
15697 expand_builtin_setjmp_receiver.
15698 (expand_label): Adjust, call expand_builtin_setjmp_receiver
15699 with NULL for the label parameter.
15700 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
15701 the frame-pointer. Adjust comments.
15702 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
15703 only if LABEL is non-NULL.
15704
15705 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
15706
15707 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
15708 (ARM_ABI_AAPCS64): Ditto.
15709 (arm_abi): Ditto.
15710 (ARM_DEFAULT_ABI): Ditto.
15711
15712 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
15713
15714 * config/aarch64/aarch64-builtins.c
15715 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
15716 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
15717 (st1): Likewise.
15718 * config/aarch64/aarch64-simd.md
15719 (aarch64_ld1<VALL:mode>): New.
15720 (aarch64_st1<VALL:mode>): Likewise.
15721 * config/aarch64/arm_neon.h
15722 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
15723
15724 2013-07-02 Sriraman Tallam <tmsriram@google.com>
15725
15726 * config/i386/i386.c (gate_insert_vzeroupper): Check if
15727 target ISA is AVX.
15728 (ix86_option_override_internal):Turn on all -mavx target flags by
15729 default as they are dependent on AVX anyway.
15730
15731 2013-07-02 Cary Coutant <ccoutant@google.com>
15732
15733 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
15734 deterministic hash.
15735 (loc_checksum_ordered): Likewise.
15736 (hash_loc_operands): Remove inline keyword.
15737
15738 2013-07-02 Jakub Jelinek <jakub@redhat.com>
15739
15740 PR tree-optimization/57741
15741 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
15742 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
15743 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
15744 Allow REAL_CST step_exprs if flag_associative_math.
15745 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
15746
15747 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15748
15749 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
15750
15751 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15752
15753 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
15754
15755 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15756
15757 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
15758 encoding.
15759 (iorsi3_insn): Likewise.
15760 (arm_xorsi3): Likewise.
15761
15762 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
15763
15764 * arm.md (attribute "wtype"): Delete. Move attribute values from here
15765 to ...
15766 (attribute "type"): ... here, and prefix with "wmmx_".
15767 (attribute "core_cycles"): Update for attribute changes.
15768 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
15769 (tbcstv4hi): Likewise.
15770 (tbcstv2si): Likewise.
15771 (iwmmxt_iordi3): Likewise.
15772 (iwmmxt_xordi3): Likewise.
15773 (iwmmxt_anddi3): Likewise.
15774 (iwmmxt_nanddi3): Likewise.
15775 (iwmmxt_arm_movdi): Likewise.
15776 (iwmmxt_movsi_insn): Likewise.
15777 (mov<mode>_internal): Likewise.
15778 (and<mode>3_iwmmxt): Likewise.
15779 (ior<mode>3_iwmmxt): Likewise.
15780 (xor<mode>3_iwmmxt): Likewise.
15781 (add<mode>3_iwmmxt): Likewise.
15782 (ssaddv8qi3): Likewise.
15783 (ssaddv4hi3): Likewise.
15784 (ssaddv2si3): Likewise.
15785 (usaddv8qi3): Likewise.
15786 (usaddv4hi3): Likewise.
15787 (usaddv2si3): Likewise.
15788 (sub<mode>3_iwmmxt): Likewise.
15789 (sssubv8qi3): Likewise.
15790 (sssubv4hi3): Likewise.
15791 (sssubv2si3): Likewise.
15792 (ussubv8qi3): Likewise.
15793 (ussubv4hi3): Likewise.
15794 (ussubv2si3): Likewise.
15795 (mulv4hi3_iwmmxt): Likewise.
15796 (smulv4hi3_highpart): Likewise.
15797 (umulv4hi3_highpart): Likewise.
15798 (iwmmxt_wmacs): Likewise.
15799 (iwmmxt_wmacsz): Likewise.
15800 (iwmmxt_wmacu): Likewise.
15801 (iwmmxt_wmacuz): Likewise.
15802 (iwmmxt_clrdi): Likewise.
15803 (iwmmxt_clrv8qi): Likewise.
15804 (iwmmxt_clr4hi): Likewise.
15805 (iwmmxt_clr2si): Likewise.
15806 (iwmmxt_uavgrndv8qi3): Likewise.
15807 (iwmmxt_uavgrndv4hi3): Likewise.
15808 (iwmmxt_uavgv8qi3): Likewise.
15809 (iwmmxt_uavgv4hi3): Likewise.
15810 (iwmmxt_tinsrb): Likewise.
15811 (iwmmxt_tinsrh): Likewise.
15812 (iwmmxt_tinsrw): Likewise.
15813 (iwmmxt_textrmub): Likewise.
15814 (iwmmxt_textrmsb): Likewise.
15815 (iwmmxt_textrmuh): Likewise.
15816 (iwmmxt_textrmsh): Likewise.
15817 (iwmmxt_textrmw): Likewise.
15818 (iwmxxt_wshufh): Likewise.
15819 (eqv8qi3): Likewise.
15820 (eqv4hi3): Likewise.
15821 (eqv2si3): Likewise.
15822 (gtuv8qi3): Likewise.
15823 (gtuv4hi3): Likewise.
15824 (gtuv2si3): Likewise.
15825 (gtv8qi3): Likewise.
15826 (gtv4hi3): Likewise.
15827 (gtv2si3): Likewise.
15828 (smax<mode>3_iwmmxt): Likewise.
15829 (umax<mode>3_iwmmxt): Likewise.
15830 (smin<mode>3_iwmmxt): Likewise.
15831 (umin<mode>3_iwmmxt): Likewise.
15832 (iwmmxt_wpackhss): Likewise.
15833 (iwmmxt_wpackwss): Likewise.
15834 (iwmmxt_wpackdss): Likewise.
15835 (iwmmxt_wpackhus): Likewise.
15836 (iwmmxt_wpackwus): Likewise.
15837 (iwmmxt_wpackdus): Likewise.
15838 (iwmmxt_wunpckihb): Likewise.
15839 (iwmmxt_wunpckihh): Likewise.
15840 (iwmmxt_wunpckihw): Likewise.
15841 (iwmmxt_wunpckilb): Likewise.
15842 (iwmmxt_wunpckilh): Likewise.
15843 (iwmmxt_wunpckilw): Likewise.
15844 (iwmmxt_wunpckehub): Likewise.
15845 (iwmmxt_wunpckehuh): Likewise.
15846 (iwmmxt_wunpckehuw): Likewise.
15847 (iwmmxt_wunpckehsb): Likewise.
15848 (iwmmxt_wunpckehsh): Likewise.
15849 (iwmmxt_wunpckehsw): Likewise.
15850 (iwmmxt_wunpckelub): Likewise.
15851 (iwmmxt_wunpckeluh): Likewise.
15852 (iwmmxt_wunpckeluw): Likewise.
15853 (iwmmxt_wunpckelsb): Likewise.
15854 (iwmmxt_wunpckelsh): Likewise.
15855 (iwmmxt_wunpckelsw): Likewise.
15856 (ror<mode>3): Likewise.
15857 (ashr<mode>3_iwmmxt): Likewise.
15858 (lshr<mode>3_iwmmxt): Likewise.
15859 (ashl<mode>3_iwmmxt): Likewise.
15860 (ror<mode>3_di): Likewise.
15861 (ashr<mode>3_di): Likewise.
15862 (lshr<mode>3_di): Likewise.
15863 (ashl<mode>3_di): Likewise.
15864 (iwmmxt_wmadds): Likewise.
15865 (iwmmxt_wmaddu): Likewise.
15866 (iwmmxt_tmia): Likewise.
15867 (iwmmxt_tmiaph): Likewise.
15868 (iwmmxt_tmiabb): Likewise.
15869 (iwmmxt_tmiatb): Likewise.
15870 (iwmmxt_tmiabt): Likewise.
15871 (iwmmxt_tmiatt): Likewise.
15872 (iwmmxt_tmovmskb): Likewise.
15873 (iwmmxt_tmovmskh): Likewise.
15874 (iwmmxt_tmovmskw): Likewise.
15875 (iwmmxt_waccb): Likewise.
15876 (iwmmxt_wacch): Likewise.
15877 (iwmmxt_waccw): Likewise.
15878 (iwmmxt_waligni): Likewise.
15879 (iwmmxt_walignr): Likewise.
15880 (iwmmxt_walignr0): Likewise.
15881 (iwmmxt_walignr1): Likewise.
15882 (iwmmxt_walignr2): Likewise.
15883 (iwmmxt_walignr3): Likewise.
15884 (iwmmxt_wsadb): Likewise.
15885 (iwmmxt_wsadh): Likewise.
15886 (iwmmxt_wsadbz): Likewise.
15887 (iwmmxt_wsadhz): Likewise.
15888 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
15889 (iwmmxt_wabsdiffb): Likewise.
15890 (iwmmxt_wabsdiffh): Likewise.
15891 (iwmmxt_wabsdiffw): Likewise.
15892 (iwmmxt_waddsubhx): Likewise
15893 (iwmmxt_wsubaddhx): Likewise.
15894 (addc<mode>3): Likewise.
15895 (iwmmxt_avg4): Likewise.
15896 (iwmmxt_avg4r): Likewise.
15897 (iwmmxt_wmaddsx): Likewise.
15898 (iwmmxt_wmaddux): Likewise.
15899 (iwmmxt_wmaddsn): Likewise.
15900 (iwmmxt_wmaddun): Likewise.
15901 (iwmmxt_wmulwsm): Likewise.
15902 (iwmmxt_wmulwum): Likewise.
15903 (iwmmxt_wmulsmr): Likewise.
15904 (iwmmxt_wmulumr): Likewise.
15905 (iwmmxt_wmulwsmr): Likewise.
15906 (iwmmxt_wmulwumr): Likewise.
15907 (iwmmxt_wmulwl): Likewise.
15908 (iwmmxt_wqmulm): Likewise.
15909 (iwmmxt_wqmulwm): Likewise.
15910 (iwmmxt_wqmulmr): Likewise.
15911 (iwmmxt_wqmulwmr): Likewise.
15912 (iwmmxt_waddbhusm): Likewise.
15913 (iwmmxt_waddbhusl): Likewise.
15914 (iwmmxt_wqmiabb): Likewise.
15915 (iwmmxt_wqmiabt): Likewise.
15916 (iwmmxt_wqmiatb): Likewise.
15917 (iwmmxt_wqmiatt): Likewise.
15918 (iwmmxt_wqmiabbn): Likewise.
15919 (iwmmxt_wqmiabtn): Likewise.
15920 (iwmmxt_wqmiatbn): Likewise.
15921 (iwmmxt_wqmiattn): Likewise.
15922 (iwmmxt_wmiabb): Likewise.
15923 (iwmmxt_wmiabt): Likewise.
15924 (iwmmxt_wmiatb): Likewise.
15925 (iwmmxt_wmiatt): Likewise.
15926 (iwmmxt_wmiabbn): Likewise.
15927 (iwmmxt_wmiabtn): Likewise.
15928 (iwmmxt_wmiatbn): Likewise.
15929 (iwmmxt_wmiattn): Likewise.
15930 (iwmmxt_wmiawbb): Likewise.
15931 (iwmmxt_wmiawbt): Likewise.
15932 (iwmmxt_wmiawtb): Likewise.
15933 (iwmmxt_wmiawtt): Likewise.
15934 (iwmmxt_wmiawbbn): Likewise.
15935 (iwmmxt_wmiawbtn): Likewise.
15936 (iwmmxt_wmiawtbn): Likewise.
15937 (iwmmxt_wmiawttn): Likewise.
15938 (iwmmxt_wmerge): Likewise.
15939 (iwmmxt_tandc<mode>3): Likewise.
15940 (iwmmxt_torc<mode>3): Likewise.
15941 (iwmmxt_torvsc<mode>3): Likewise.
15942 (iwmmxt_textrc<mode>3): Likewise.
15943 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
15944 (wmmxt_pack): Likewise.
15945 (wmmxt_mult_c1): Likewise.
15946 (wmmxt_mult_c2): Likewise.
15947 (wmmxt_alu_c1): Likewise.
15948 (wmmxt_alu_c2): Likewise.
15949 (wmmxt_alu_c3): Likewise.
15950 (wmmxt_transfer_c1): Likewise.
15951 (wmmxt_transfer_c2): Likewise.
15952 (wmmxt_transfer_c3): Likewise.
15953 (marvell_f_iwmmxt_wstr): Likewise.
15954 (marvell_f_iwmmxt_wldr): Likewise.
15955
15956 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
15957
15958 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
15959
15960 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15961
15962 Revert:
15963 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15964 * lra-constraints.c (need_for_split_p): Check call used hard regs
15965 living through calls.
15966
15967 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
15968 call used regs for call insn.
15969
15970 2013-06-28 Jakub Jelinek <jakub@redhat.com>
15971
15972 PR target/57736
15973 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
15974 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
15975
15976 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15977
15978 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
15979
15980 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15981
15982 * lra-constraints.c (need_for_split_p): Check call used hard regs
15983 living through calls.
15984
15985 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
15986
15987 PR target/57744
15988 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
15989 to tie with any other modes. Eliminate Altivec vector mode tests,
15990 since these are a subset of ALTIVEC or VSX vector modes. Simplify
15991 code, to return 0 if testing MODE2 for a condition, if we've
15992 already tested MODE1 for the same condition.
15993
15994 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
15995
15996 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
15997 layout.
15998
15999 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16000
16001 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
16002 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
16003
16004 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16005
16006 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
16007 Define.
16008 (aarch64_symbolic_constant_p): Remove.
16009 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
16010 static. Fix line length and white space.
16011 (aarch64_symbolic_constant_p): Remove.
16012 * config/aarch64/predicates.md (aarch64_valid_symref):
16013 Use aarch64_classify_symbol_expression.
16014
16015 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16016
16017 * config/arm/constraints.md (Ts): New constraint.
16018 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
16019 16-bit encodings.
16020 (compare_scc): Use "Ts" constraint for operand 0.
16021 (ior_scc_scc): Likewise.
16022 (and_scc_scc): Likewise.
16023 (and_scc_scc_nodom): Likewise.
16024 (ior_scc_scc_cmp): Likewise for operand 7.
16025 (and_scc_scc_cmp): Likewise.
16026 * config/arm/thumb2.md (thumb2_movsi_insn):
16027 Add alternatives for 16-bit encodings.
16028 (thumb2_movhi_insn): Likewise.
16029 (thumb2_movsicc_insn): Likewise.
16030 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
16031 (thumb2_negscc): Use "Ts" constraint.
16032 Move mvn instruction outside cond_exec block.
16033 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
16034 for 16-bit encodings.
16035
16036 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16037
16038 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
16039 encoding.
16040 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
16041 (mulsi3subsi): Likewise.
16042 (mulsidi3adddi): Likewise.
16043 (mulsidi3_v6): Likewise.
16044 (umulsidi3_v6): Likewise.
16045 (umulsidi3adddi_v6): Likewise.
16046 (smulsi3_highpart_v6): Likewise.
16047 (umulsi3_highpart_v6): Likewise.
16048 (mulhisi3tb): Likewise.
16049 (mulhisi3bt): Likewise.
16050 (mulhisi3tt): Likewise.
16051 (maddhisi4): Likewise.
16052 (maddhisi4tb): Likewise.
16053 (maddhisi4tt): Likewise.
16054 (maddhidi4): Likewise.
16055 (maddhidi4tb): Likewise.
16056 (maddhidi4tt): Likewise.
16057 (zeroextractsi_compare0_scratch): Likewise.
16058 (insv_zero): Likewise.
16059 (insv_t2): Likewise.
16060 (anddi_notzesidi_di): Likewise.
16061 (anddi_notsesidi_di): Likewise.
16062 (andsi_notsi_si): Likewise.
16063 (iordi_zesidi_di): Likewise.
16064 (xordi_zesidi_di): Likewise.
16065 (andsi_iorsi3_notsi): Likewise.
16066 (smax_0): Likewise.
16067 (smax_m1): Likewise.
16068 (smin_0): Likewise.
16069 (not_shiftsi): Likewise.
16070 (unaligned_loadsi): Likewise.
16071 (unaligned_loadhis): Likewise.
16072 (unaligned_loadhiu): Likewise.
16073 (unaligned_storesi): Likewise.
16074 (unaligned_storehi): Likewise.
16075 (extv_reg): Likewise.
16076 (extzv_t2): Likewise.
16077 (divsi3): Likewise.
16078 (udivsi3): Likewise.
16079 (arm_zero_extendhisi2addsi): Likewise.
16080 (arm_zero_extendqisi2addsi): Likewise.
16081 (compareqi_eq0): Likewise.
16082 (arm_extendhisi2_v6): Likewise.
16083 (arm_extendqisi2addsi): Likewise.
16084 (arm_movt): Likewise.
16085 (thumb2_ldrd): Likewise.
16086 (thumb2_ldrd_base): Likewise.
16087 (thumb2_ldrd_base_neg): Likewise.
16088 (thumb2_strd): Likewise.
16089 (thumb2_strd_base): Likewise.
16090 (thumb2_strd_base_neg): Likewise.
16091 (arm_negsi2): Add alternative for 16-bit encoding.
16092 (arm_one_cmplsi2): Likewise.
16093
16094 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16095
16096 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
16097 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
16098 (movdfcc): Likewise.
16099 * config/arm/vfp.md (*thumb2_movsf_vfp):
16100 Disable predication for arm_restrict_it.
16101 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
16102 (*thumb2_movdfcc_vfp): Likewise.
16103 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
16104 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
16105 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
16106 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
16107 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
16108 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
16109 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
16110 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
16111 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
16112 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
16113 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
16114 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
16115 Disable predication for arm_restrict_it.
16116
16117 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
16118
16119 * config/i386/bmiintrin.h (_bextr_u32): New.
16120 (_bextr_u64): Ditto.
16121
16122 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
16123
16124 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
16125 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
16126 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
16127 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
16128 * config/mips/n32-elf.h: ...this new file.
16129
16130 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
16131
16132 PR target/57224
16133 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
16134 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
16135
16136 2013-06-27 Catherine Moore <clm@codesourcery.com>
16137
16138 * config/mips/mips-tables.opt: Regenerate.
16139 * config/mips/mips-cpus.def: Add m14ke and m14kec.
16140 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
16141 * doc/invoke.texi: Add -m14kc.
16142
16143 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16144
16145 PR target/57623
16146 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
16147 constraints of operand 1 and 2.
16148
16149 PR target/57623
16150 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
16151 to match RTL canonicalization. Swap predicates and
16152 constraints of operand 1 and 2.
16153
16154 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
16155
16156 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
16157 Process OP_INOUT regs for splitting too.
16158
16159 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16160
16161 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
16162 decl before the loop, initialize to NULL.
16163 (vectorizable_load): Initialize ptr_incr to NULL.
16164
16165 2013-06-27 Martin Jambor <mjambor@suse.cz>
16166
16167 PR lto/57208
16168 * ipa-ref.h (ipa_maybe_record_reference): Declare.
16169 * ipa-ref.c (ipa_maybe_record_reference): New function.
16170 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
16171 * ipa-cp.c (create_specialized_node): Record potential references from
16172 aggvals.
16173 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
16174
16175 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16176
16177 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
16178 parameter 'mode' of type 'enum machine_mode mode'; change to pass
16179 'mode' to force_reg.
16180 (aarch64_add_offset): Update calls to aarch64_force_temporary.
16181 (aarch64_expand_mov_immediate): Likewise.
16182
16183 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16184
16185 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
16186 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
16187
16188 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16189
16190 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
16191 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16192 (struct machine_function): Add tbegin_p.
16193 (s390_canonicalize_comparison): Fold CC mode compares to
16194 conditional jump if possible.
16195 (s390_emit_jump): Return the emitted jump.
16196 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
16197 Handle CCRAWmode compares.
16198 (s390_option_override): Default to -mhtm if available.
16199 (s390_reg_clobbered_rtx): Handle floating point regs as well.
16200 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
16201 FPRs instead of df_regs_ever_live_p.
16202 (s390_optimize_nonescaping_tx): New function.
16203 (s390_init_frame_layout): Extend clobbered_regs array to cover
16204 FPRs as well.
16205 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
16206 (s390_expand_tbegin): New function.
16207 (enum s390_builtin): New enum definition.
16208 (code_for_builtin): New array definition.
16209 (s390_init_builtins): New function.
16210 (s390_expand_builtin): New function.
16211 (TARGET_INIT_BUILTINS): Define.
16212 (TARGET_EXPAND_BUILTIN): Define.
16213 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
16214 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
16215 (s390_alc_comparison): Likewise.
16216 * config/s390/s390-modes.def: Add CCRAWmode.
16217 * config/s390/s390.h (processor_flags): Add PF_TX.
16218 (TARGET_CPU_HTM): Define macro.
16219 (TARGET_HTM): Define macro.
16220 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
16221 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
16222 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16223 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
16224 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
16225 (TBEGIN_MASK, TBEGINC_MASK): New constants.
16226 ("*cc_to_int"): Move up.
16227 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
16228 constants other than 0.
16229 ("*ccraw_to_int"): New insn and splitter definition.
16230 ("tbegin", "tbegin_nofloat", "tbegin_retry")
16231 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
16232 ("tx_assist"): New expander.
16233 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
16234 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
16235 * config/s390/s390.opt: Add -mhtm option.
16236 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
16237 * config/s390/htmxlintrin.h: New file.
16238 * config/s390/htmintrin.h: New file.
16239 * config/s390/s390intrin.h: New file.
16240 * doc/extend.texi: Document htm builtins.
16241 * config.gcc: Add the new header files to extra_headers.
16242
16243 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
16244
16245 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
16246 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
16247
16248 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
16249 Pat Haugen <pthaugen@us.ibm.com>
16250 Peter Bergner <bergner@vnet.ibm.com>
16251
16252 * config/rs6000/power8.md: New.
16253 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
16254 setting for power8 entry.
16255 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
16256 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
16257 test for Power4/Power5 only.
16258 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
16259 support.
16260 (force_new_group): Adjust comment.
16261 * config/rs6000/rs6000.md: Include power8.md.
16262
16263 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
16264
16265 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
16266 * config/arm/arm-protos.h (arm_max_conditional_execute): New
16267 declaration.
16268 (tune_params): Update comment.
16269 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
16270 (arm_max_conditional_execute): New function.
16271 (thumb2_final_prescan_insn): Use max_insn_skipped and
16272 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
16273
16274 2013-06-25 Jakub Jelinek <jakub@redhat.com>
16275
16276 PR tree-optimization/57705
16277 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
16278 SSA_NAME step, provided that it is not defined inside the loop.
16279 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
16280 (get_initial_def_for_induction): Handle SSA_NAME IV step.
16281
16282 2013-06-25 Martin Jambor <mjambor@suse.cz>
16283
16284 PR middle-end/57670
16285 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
16286 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
16287 calls in the dump.
16288 (ipa_note_param_call): Initialize member_ptr flag.
16289 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
16290 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
16291 (ipa_write_indirect_edge_info): Stream member_ptr flag.
16292 (ipa_read_indirect_edge_info): Likewise.
16293
16294 2013-06-25 Richard Biener <rguenther@suse.de>
16295
16296 PR middle-end/56977
16297 * passes.c (init_optimization_passes): Move pass_fold_builtins
16298 and pass_dce earlier with -Og.
16299
16300 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
16301
16302 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
16303 <BIT_FIELD_REF>: Remove trailing TAB.
16304 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
16305 remove blank line.
16306
16307 2013-06-24 Martin Jambor <mjambor@suse.cz>
16308
16309 PR tree-optimization/57358
16310 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
16311 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
16312 (ipa_analyze_params_uses): Generate pessimistic info when true.
16313
16314 2013-06-24 Martin Jambor <mjambor@suse.cz>
16315
16316 PR tree-optimization/57539
16317 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
16318 global.inlined_to of the new node to it. All callers changed.
16319 * ipa-inline-transform.c (clone_inlined_nodes): New variable
16320 inlining_into, pass it to cgraph_clone_node.
16321 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
16322 ipa_free_edge_args_substructures.
16323 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
16324 rdesc linked list. Do not assert rdesc edges have inlined caller.
16325 Assert we have found an rdesc in the rdesc list.
16326
16327 2013-06-24 Richard Biener <rguenther@suse.de>
16328
16329 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
16330 (pointer_set_lookup): Declare.
16331 (class pointer_map): New template class implementing a
16332 generic pointer to T map.
16333 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
16334 pointer_map<T>::contains, pointer_map<T>::insert,
16335 pointer_map<T>::traverse): New functions.
16336 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
16337 (pointer_set_lookup): New function.
16338 (pointer_set_contains): Use pointer_set_lookup.
16339 (pointer_set_insert): Likewise.
16340 (insert_aux): Remove.
16341 (struct pointer_map_t): Embed a pointer_set_t.
16342 (pointer_map_create): Adjust.
16343 (pointer_map_destroy): Likewise.
16344 (pointer_map_contains): Likewise.
16345 (pointer_map_insert): Likewise.
16346 (pointer_map_traverse): Likewise.
16347 * tree-streamer.h (struct streamer_tree_cache_d): Use a
16348 pointer_map<unsigned> instead of a pointer_map_t.
16349 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
16350 (streamer_tree_cache_lookup): Likewise.
16351 (streamer_tree_cache_create): Likewise.
16352 (streamer_tree_cache_delete): Likewise.
16353 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
16354 pointer_map<unsigned> instead of a pointer_map_t.
16355 (lto_init_tree_ref_encoder): Adjust.
16356 (lto_destroy_tree_ref_encoder): Likewise.
16357 * lto-section-out.c (lto_output_decl_index): Likewise.
16358 (lto_record_function_out_decl_state): Likewise.
16359 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
16360
16361 2013-06-24 Richard Biener <rguenther@suse.de>
16362
16363 PR tree-optimization/57488
16364 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
16365
16366 2013-06-24 Alan Modra <amodra@gmail.com>
16367
16368 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
16369 (gen_easy_altivec_constant): Likewise.
16370 * config/rs6000/predicates.md (easy_vector_constant_add_self,
16371 easy_vector_constant_msb): Likewise.
16372
16373 2013-06-23 Jakub Jelinek <jakub@redhat.com>
16374
16375 PR target/57688
16376 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
16377 add missing return true.
16378
16379 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
16380
16381 PR target/52483
16382 * config/sh/predicates.md (general_extend_operand): Invoke
16383 general_movsrc_operand for memory operands.
16384 (general_movsrc_operand): Allow reg+reg addressing, do not use
16385 general_operand for memory operands.
16386
16387 2013-06-23 Sriraman Tallam <tmsriram@google.com>
16388
16389 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
16390 when current target options does not apply.
16391 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
16392 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
16393 * config/i386/bmiintrin.h: Pass appropriate target
16394 attributes to header.
16395 * config/i386/mmintrin.h: Ditto.
16396 * config/i386/nmmintrin.h: Ditto.
16397 * config/i386/avx2intrin.h: Ditto.
16398 * config/i386/fxsrintrin.h: Ditto.
16399 * config/i386/tbmintrin.h: Ditto.
16400 * config/i386/xsaveintrin.h: Ditto.
16401 * config/i386/f16cintrin.h: Ditto.
16402 * config/i386/xtestintrin.h: Ditto.
16403 * config/i386/xsaveoptintrin.h: Ditto.
16404 * config/i386/bmi2intrin.h: Ditto.
16405 * config/i386/lzcntintrin.h: Ditto.
16406 * config/i386/smmintrin.h: Ditto.
16407 * config/i386/wmmintrin.h: Ditto.
16408 * config/i386/x86intrin.h: Remove all header include guards.
16409 * config/i386/prfchwintrin.h: Ditto.
16410 * config/i386/pmmintrin.h: Ditto.
16411 * config/i386/tmmintrin.h: Ditto.
16412 * config/i386/xmmintrin.h: Ditto.
16413 * config/i386/popcntintrin.h: Ditto.
16414 * config/i386/rdseedintrin.h: Ditto.
16415 * config/i386/ammintrin.h: Ditto.
16416 * config/i386/emmintrin.h: Ditto.
16417 * config/i386/immintrin.h: Remove all header include guards.
16418 * config/i386/fma4intrin.h: Ditto.
16419 * config/i386/lwpintrin.h: Ditto.
16420 * config/i386/xopintrin.h: Ditto.
16421 * config/i386/ia32intrin.h: Ditto.
16422 * config/i386/avxintrin.h: Ditto.
16423 * config/i386/rtmintrin.h: Ditto.
16424 * config/i386/fmaintrin.h: Ditto.
16425 * config/i386/mm3dnow.h: Ditto.
16426
16427 2013-06-22 Sriraman Tallam <tmsriram@google.com>
16428
16429 * common/config/i386/i386-common.c: Handle LZCNT.
16430
16431 2013-06-22 Andi Kleen <ak@linux.intel.com>
16432
16433 * doc/extend.texi: Use __atomic_store_n instead of
16434 __atomic_store in HLE example.
16435
16436 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
16437
16438 * config/sh/sh.c: Remove <cstdlib> workaround.
16439
16440 2013-06-21 Andi Kleen <ak@linux.intel.com>
16441
16442 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
16443
16444 2013-06-21 Andi Kleen <ak@linux.intel.com>
16445
16446 * doc/extend.texi: Document that __atomic_clear and
16447 __atomic_test_and_set should only be used with bool.
16448
16449 2013-06-20 Jan Hubicka <jh@suse.cz>
16450
16451 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
16452 types_same_for_odr.
16453 * tree.c (decls_same_for_odr): New function.
16454 (same_for_edr): New function.
16455 (types_same_for_odr): New function.
16456 (get_binfo_at_offset): Use it.
16457 * tree.h (types_same_for_odr): Declare.
16458
16459 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
16460 Jason Merrill <jason@redhat.com>
16461
16462 * system.h: Include <cstdlib> as well as <stdlib.h>.
16463
16464 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
16465
16466 PR target/57655
16467 * config/i386/i386.c (construct_container): Report error if
16468 long double is used with disabled x87 float returns.
16469
16470 2013-06-20 Jan Hubicka <jh@suse.cz>
16471
16472 * lto-cgraph.c (input_symtab): Do not set cgraph state.
16473
16474 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
16475
16476 PR rtl-optimization/57425
16477 PR rtl-optimization/57569
16478 * alias.c (write_dependence_p): Remove parameters mem_mode and
16479 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
16480 Changed all callers.
16481 (canon_anti_dependence): Get comments and semantics in sync.
16482 Add parameter mem_canonicalized. Changed all callers.
16483 * rtl.h (canon_anti_dependence): Update prototype.
16484
16485 2013-06-20 Richard Biener <rguenther@suse.de>
16486
16487 * data-streamer-in.c (streamer_read_uhwi): Optimize single
16488 byte case, inline streamer_read_uchar and defer section
16489 overrun check.
16490
16491 2013-06-20 Richard Biener <rguenther@suse.de>
16492
16493 PR tree-optimization/57584
16494 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
16495 SSA names into the expanded expression that take part in
16496 abnormal coalescing.
16497
16498 2013-06-19 Sharad Singhai <singhai@google.com>
16499
16500 * gcov.c (print_usage): Handle new option.
16501 (process_args): Ditto.
16502 (get_gcov_intermediate_filename): New function.
16503 (output_intermediate_file): New function.
16504 (output_gcov_file): New function
16505 (generate_results): Handle new option.
16506 (release_function): Relase demangled name.
16507 (read_graph_file): Handle demangled name.
16508 (output_lines): Ditto.
16509 * doc/gcov.texi: Document gcov intermediate format.
16510
16511 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
16512
16513 PR bootstrap/57604
16514 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
16515 (lra_emit_add): Use the functions. Add comment about Y as an
16516 address segment.
16517
16518 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
16519
16520 PR driver/57652
16521 * collect2.c (collect_atexit): New.
16522 (collect_exit): Delete.
16523 (main): Register collect_atexit with atexit.
16524 (collect_wait): Change collect_exit to exit.
16525 (do_wait): Same.
16526 * collect2.h (collect_exit): Delete.
16527 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
16528
16529 2013-06-19 Wei Mi <wmi@google.com>
16530
16531 PR rtl-optimization/57518
16532 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
16533 if regno is used in paradoxical subreg.
16534 (update_equiv_regs): Check pdx_subregs[regno] before
16535 set a reg to be equivalent with a mem.
16536
16537 2013-06-19 Matthias Klose <doko@ubuntu.com>
16538
16539 PR driver/57651
16540 * file-find.h (find_a_file): Add a mode parameter.
16541 * file-find.c (find_a_file): Likewise.
16542 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
16543 with X_OK for the executables.
16544 * collect2.c (main): Call find_a_file with X_OK.
16545
16546 2013-06-19 Steve Ellcey <sellcey@mips.com>
16547
16548 PR target/56942
16549 * config/mips/mips.md (casesi_internal_mips16_<mode>):
16550 Use NEXT_INSN instead of next_real_insn.
16551
16552 2013-06-19 Jan Hubicka <jh@suse.cz>
16553
16554 * cgraph.h (const_value_known_p): Replace by ...
16555 (ctor_for_folding): .. this one.
16556 * cgraphunit.c (process_function_and_variable_attributes): Use it.
16557 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
16558 * expr.c (expand_expr_real_1): Likewise.
16559 (string_constant): Likewise.
16560 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
16561 * ipa.c (process_references): Likewise.
16562 (symtab_remove_unreachable_nodes): Likewise.
16563 * ipa-inline-analysis.c (param_change_prob): Likewise.
16564 * gimple-fold.c (canonicalize_constructor_val): Likewise.
16565 (get_base_constructor): Likwise.
16566 * varpool.c (varpool_remove_node): Likewise.
16567 (varpool_remove_initializer): LIkewise.
16568 (dump_varpool_node): LIkwise.
16569 (const_value_known_p): Rewrite to ...
16570 (ctor_for_folding): ... this one.
16571
16572 2013-06-19 Jakub Jelinek <jakub@redhat.com>
16573
16574 PR driver/57651
16575 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
16576 PERSONALITY in $PATH derived prefixes.
16577
16578 2013-06-19 Jeff Law <law@redhat.com>
16579
16580 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
16581 in comment.
16582
16583 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
16584 (simplify_bitwise_binary): Use it to simpify certain binary ops on
16585 booleans.
16586
16587 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
16588
16589 * config/arm/vfp.md: Move VFP instruction classification documentation
16590 to ...
16591 * config/arm/arm.md: ... here. Update instruction classification
16592 documentation.
16593
16594 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
16595
16596 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
16597 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
16598 pattern. Use more efficient sequences on ARMv5 and Thumb2.
16599
16600 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
16601
16602 PR target/57609
16603 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16604 with NEXT_INSN. Use tablejump_p to check for jump table data
16605 insns.
16606
16607 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
16608
16609 PR c++/56544
16610 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
16611 that now in C++ the value is correct per the C++ standards.
16612
16613 2013-06-19 Richard Biener <rguenther@suse.de>
16614
16615 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
16616 for global context.
16617
16618 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16619
16620 Revert:
16621 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16622
16623 PR target/57609
16624 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16625 with next_active_insn.
16626
16627 2013-06-18 Sriraman Tallam <tmsriram@google.com>
16628
16629 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
16630 functions are inlined during failures to flag an error.
16631 * tree-inline.c (expand_call_inline): Allow the error to be flagged
16632 in early inline pass.
16633
16634 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
16635
16636 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
16637 in comments.
16638
16639 2013-06-18 Julian Brown <julian@codesourcery.com>
16640
16641 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
16642 Permit virtual register pre-reload if !strict.
16643 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
16644 change.
16645 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
16646 prototype.
16647 * config/arm/neon.md (movmisalign<mode>): Use
16648 neon_perm_struct_or_reg_operand instead of
16649 neon_struct_or_register_operand.
16650 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
16651 neon_permissive_struct_operand instead of neon_struct_operand.
16652 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
16653 neon_vector_mem_operand.
16654 * config/arm/predicates.md (neon_struct_operand): Adjust call to
16655 neon_vector_mem_operand.
16656 (neon_permissive_struct_operand): New.
16657 (neon_struct_or_register_operand): Rename to...
16658 (neon_perm_struct_or_reg_operand): This. Adjust call to
16659 neon_vector_mem_operand.
16660
16661 2013-06-18 Richard Biener <rguenther@suse.de>
16662
16663 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
16664 * lto-streamer.h: Include pointer-set.h.
16665 (struct lto_decl_slot): Remove.
16666 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
16667 Remove next_index entry.
16668 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16669 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16670 (lto_init_tree_ref_encoder): Adjust.
16671 (lto_destroy_tree_ref_encoder): Likewise.
16672 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16673 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16674 (lto_output_decl_index): Adjust.
16675 (lto_new_out_decl_state): Likewise.
16676 (lto_record_function_out_decl_state): Likewise.
16677 * lto-streamer-out.c (copy_function): Likewise.
16678
16679 2013-06-18 Richard Biener <rguenther@suse.de>
16680
16681 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
16682 * cgraphunit.c: Include cfgloop.h.
16683 (init_lowered_empty_function): Initialize the loop tree.
16684 (assemble_thunk): Insert new BBs into loops.
16685
16686 2013-06-18 Richard Biener <rguenther@suse.de>
16687
16688 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
16689 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
16690 the map from cache entry to cache index optional.
16691 (streamer_tree_cache_replace_tree): Adjust accordingly.
16692 (streamer_tree_cache_append): Likewise.
16693 (streamer_tree_cache_delete): Likewise.
16694 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
16695 streamer cache map from cache entry to cache index.
16696 * lto-streamer-out.c (create_output_block): Adjust.
16697
16698 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
16699
16700 * config/arm/arm.md (attribute "insn"): Move multiplication and
16701 division attributes to...
16702 (attribute "type"): ... here. Remove mult.
16703 (attribute "mul32"): New attribute.
16704 (attribute "mul64"): Add umaal.
16705 (*arm_mulsi3): Update attributes.
16706 (*arm_mulsi3_v6): Likewise.
16707 (*thumb_mulsi3): Likewise.
16708 (*thumb_mulsi3_v6): Likewise.
16709 (*mulsi3_compare0): Likewise.
16710 (*mulsi3_compare0_v6): Likewise.
16711 (*mulsi_compare0_scratch): Likewise.
16712 (*mulsi_compare0_scratch_v6): Likewise.
16713 (*mulsi3addsi): Likewise.
16714 (*mulsi3addsi_v6): Likewise.
16715 (*mulsi3addsi_compare0): Likewise.
16716 (*mulsi3addsi_compare0_v6): Likewise.
16717 (*mulsi3addsi_compare0_scratch): Likewise.
16718 (*mulsi3addsi_compare0_scratch_v6): Likewise.
16719 (*mulsi3subsi): Likewise.
16720 (*mulsidi3adddi): Likewise.
16721 (*mulsi3addsi_v6): Likewise.
16722 (*mulsidi3adddi_v6): Likewise.
16723 (*mulsidi3_nov6): Likewise.
16724 (*mulsidi3_v6): Likewise.
16725 (*umulsidi3_nov6): Likewise.
16726 (*umulsidi3_v6): Likewise.
16727 (*umulsidi3adddi): Likewise.
16728 (*umulsidi3adddi_v6): Likewise.
16729 (*smulsi3_highpart_nov6): Likewise.
16730 (*smulsi3_highpart_v6): Likewise.
16731 (*umulsi3_highpart_nov6): Likewise.
16732 (*umulsi3_highpart_v6): Likewise.
16733 (mulhisi3): Likewise.
16734 (*mulhisi3tb): Likewise.
16735 (*mulhisi3bt): Likewise.
16736 (*mulhisi3tt): Likewise.
16737 (maddhisi4): Likewise.
16738 (*maddhisi4tb): Likewise.
16739 (*maddhisi4tt): Likewise.
16740 (maddhidi4): Likewise.
16741 (*maddhidi4tb): Likewise.
16742 (*maddhidi4tt): Likewise.
16743 (divsi3): Likewise.
16744 (udivsi3): Likewise.
16745 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
16746 (thumb2_mulsi_short_compare0): Likewise.
16747 (thumb2_mulsi_short_compare0_scratch): Likewise.
16748 * config/arm/arm1020e.md (1020mult1): Update attribute change.
16749 (1020mult2): Likewise.
16750 (1020mult3): Likewise.
16751 (1020mult4): Likewise.
16752 (1020mult5): Likewise.
16753 (1020mult6): Likewise.
16754 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
16755 change.
16756 (cortex_a15_mult64): Likewise.
16757 (cortex_a15_sdiv): Likewise.
16758 (cortex_a15_udiv): Likewise.
16759 * config/arm/arm1026ejs.md (mult1): Update attribute change.
16760 (mult2): Likewise.
16761 (mult3): Likewise.
16762 (mult4): Likewise.
16763 (mult5): Likewise.
16764 (mult6): Likewise.
16765 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
16766 (pj4_ir_div): Likewise.
16767 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
16768 (11_mult2): Likewise.
16769 (11_mult3): Likewise.
16770 (11_mult4): Likewise.
16771 (11_mult5): Likewise.
16772 (11_mult6): Likewise.
16773 (11_mult7): Likewise.
16774 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
16775 (cortex_a8_mla): Likewise.
16776 (cortex_a8_mull): Likewise.
16777 (cortex_a8_smulwy): Likewise.
16778 (cortex_a8_smlald): Likewise.
16779 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
16780 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
16781 (cortex_r4_mul_3): Likewise.
16782 (cortex_r4_mla_4): Likewise.
16783 (cortex_r4_mla_3): Likewise.
16784 (cortex_r4_smlald): Likewise.
16785 (cortex_r4_mull): Likewise.
16786 (cortex_r4_sdiv): Likewise.
16787 (cortex_r4_udiv): Likewise.
16788 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
16789 (cortex_a7_idiv): Likewise.
16790 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
16791 (9_mult2): Likewise.
16792 (9_mult3): Likewise.
16793 (9_mult4): Likewise.
16794 (9_mult5): Likewise.
16795 (9_mult6): Likewise.
16796 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
16797 (cortex_a53_sdiv): Likewise.
16798 (cortex_a53_udiv): Likewise.
16799 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
16800 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
16801 (mp626_mult2): Likewise.
16802 (mp626_mult3): Likewise.
16803 (mp626_mult4): Likewise.
16804 * config/arm/fa526.md (526_mult1): Update attribute change.
16805 (526_mult2): Likewise.
16806 * config/arm/arm-generic.md (mult): Update attribute change.
16807 (mult_ldsched_strongarm): Likewise.
16808 (mult_ldsched): Likewise.
16809 (multi_cycle): Likewise.
16810 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
16811 * config/arm/fa606te.md (606te_mult1): Update attribute change.
16812 (606te_mult2): Likewise.
16813 (606te_mult3): Likewise.
16814 (606te_mult4): Likewise.
16815 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
16816 (cortex_a9_mac16): Likewise.
16817 (cortex_a9_multiply): Likewise.
16818 (cortex_a9_mac): Likewise.
16819 (cortex_a9_multiply_long): Likewise.
16820 * config/arm/fa626te.md (626te_mult1): Update attribute change.
16821 (626te_mult2): Likewise.
16822 (626te_mult3): Likewise.
16823 (626te_mult4): Likewise.
16824
16825 2013-06-18 Richard Biener <rguenther@suse.de>
16826
16827 PR lto/57334
16828 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
16829
16830 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16831
16832 PR target/57609
16833 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16834 with next_active_insn.
16835
16836 2013-06-18 Alan Modra <amodra@gmail.com>
16837
16838 * config/rs6000/rs6000.h (enum data_align): New.
16839 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
16840 (DATA_ABI_ALIGNMENT): Define.
16841 (CONSTANT_ALIGNMENT): Correct comment.
16842 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
16843 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
16844
16845 2013-06-17 David Malcolm <dmalcolm@redhat.com>
16846
16847 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
16848 ATTRIBUTE_UNUSED marking.
16849
16850 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
16851
16852 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
16853 alternative and update.
16854 (aarch64_dup_lanedi): Delete.
16855 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
16856 * config/aarch64/aarch64-simd-builtins.def: Update.
16857
16858 2013-06-17 Richard Biener <rguenther@suse.de>
16859
16860 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
16861 (lto_input_scc): Declare.
16862 (lto_input_tree_1): Likewise.
16863 (struct lto_stats_d): Add num_tree_bodies_output and
16864 num_pickle_refs_output.
16865 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
16866 (lto_read_tree_1): Split out from ...
16867 (lto_read_tree): ... this.
16868 (lto_input_scc): New function.
16869 (lto_input_tree_1): Split out from ...
16870 (lto_input_tree): ... this. Handle LTO_tree_scc.
16871 (lto_data_in_create): Create the streamer cache without hashes.
16872 * lto-streamer-out.c (create_output_block): Create the streamer
16873 cache with hashes when not doing WPA.
16874 (lto_write_tree_1): Split out from ...
16875 (lto_write_tree): ... this.
16876 (get_symbol_initial_value): New function.
16877 (lto_output_tree_1): Split out from ...
16878 (lto_output_tree): ... this. Write trees as series of SCCs
16879 using a DFS walk via DFS_write_tree.
16880 (struct sccs, struct scc_entry): New types.
16881 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
16882 (DFS_write_tree_body): New function.
16883 (DFS_write_tree): Likewise.
16884 (hash_tree): Likewise.
16885 (scc_entry_compare): Likewise.
16886 (hash_scc): Likewise.
16887 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
16888 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
16889 TREE_CHAIN as regular reference.
16890 (streamer_read_integer_cst): Remove.
16891 (streamer_get_pickled_tree): Adjust.
16892 * tree-streamer-out.c (streamer_write_chain): Disable streaming
16893 of DECL_EXTERNALs in BLOCK_VARS for now.
16894 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
16895 reference.
16896 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
16897 Add hash value argument and record that if hashes are recorded
16898 in the cache.
16899 (streamer_tree_cache_insert_1): Adjust.
16900 (streamer_tree_cache_insert): Likewise.
16901 (streamer_tree_cache_insert_at): Rename to ...
16902 (streamer_tree_cache_replace_tree): ... this and adjust.
16903 (streamer_tree_cache_append): Adjust.
16904 (record_common_node): Likewise.
16905 (streamer_tree_cache_create): Add argument whether to
16906 record hash values together with trees.
16907 (streamer_tree_cache_delete): Adjust.
16908 * tree-streamer.h (struct streamer_tree_cache_d): Add
16909 vector of hashes.
16910 (streamer_read_integer_cst): Remove.
16911 (streamer_tree_cache_insert): Adjust.
16912 (streamer_tree_cache_append): Likewise.
16913 (streamer_tree_cache_insert_at): Rename to ...
16914 (streamer_tree_cache_replace_tree): ... this and adjust.
16915 (streamer_tree_cache_create): Add argument whether to record hashes.
16916 (streamer_tree_cache_get): Rename to ...
16917 (streamer_tree_cache_get_tree): ... this.
16918 (streamer_tree_cache_get_hash): New function.
16919 * tree.c (cache_integer_cst): New function.
16920 * tree.h (cache_integer_cst): Declare.
16921 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
16922 * lto-symtab.c (lto_varpool_replace_node): Only release
16923 DECL_INITIAL of non-prevailing decls.
16924 * varpool.c (varpool_remove_initializer): Do not release
16925 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
16926
16927 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
16928
16929 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
16930 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
16931 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
16932 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
16933 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
16934 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
16935 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
16936 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
16937 instead of TARGET_64BIT.
16938 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
16939 Require ISA_HAS_<D>DIV.
16940
16941 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
16942
16943 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
16944 (mips*-*-linux*): Move default with_llsc setting to where other
16945 defaults are set.
16946 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
16947 with_arch block.
16948 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
16949 Likewise. Remove default with_tune setting. Move default float
16950 setting to its own block. Handle with_llsc in the same block as above.
16951
16952 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
16953
16954 PR rtl-optimization/57425
16955 PR rtl-optimization/57569
16956 * alias.c (write_dependence_p): Add new parameters mem_mode,
16957 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
16958 Changed all callers.
16959 (canon_anti_dependence): New function.
16960 * cse.c (check_dependence): Use canon_anti_dependence.
16961 * cselib.c (cselib_invalidate_mem): Likewise.
16962 * rtl.h (canon_anti_dependence): Declare.
16963
16964 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
16965
16966 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
16967 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
16968 ".set mips3" for 64-bit targets.
16969
16970 2013-06-15 Dehao Chen <dehao@google.com>
16971
16972 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
16973 * gimple-low.c (gimple_check_call_matching_types): Likewise.
16974 (gimple_check_call_args): Likewise.
16975 * value-prof.c (check_ic_target): Likewise.
16976 * ipa-inline.c (early_inliner): Likewise.
16977 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
16978 * cgraph.c (cgraph_create_edge_1): Likewise.
16979 (cgraph_make_edge_direct): Likewise.
16980
16981 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
16982
16983 PR target/57615
16984 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
16985 rs6000_output_move_128bit to handle emitting quad memory
16986 operations. Set attribute length to 8 bytes.
16987
16988 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
16989
16990 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
16991 New pattern.
16992 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
16993 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
16994 (aarch64_<su>mlsl<mode>): Likewise.
16995
16996 2013-06-14 Mike Stump <mikestump@comcast.net>
16997
16998 * Makefile.in (TARGET_H): Add insn-codes.h.
16999
17000 2013-06-14 Alan Modra <amodra@gmail.com>
17001
17002 PR middle-end/57134
17003 PR middle-end/57586
17004 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
17005 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
17006 bitfield expansion when EXPAND_MEMORY.
17007 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
17008
17009 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
17010
17011 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
17012 test for clearing quad memory on 32-bit later.
17013
17014 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17015
17016 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
17017 (fold_negate_expr): Likewise.
17018 (fold_real_zero_addition_p): Handle vectors.
17019 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
17020
17021 2013-06-14 Alan Modra <amodra@gmail.com>
17022
17023 * varasm.c (force_const_mem): Revert 2013-06-07 change.
17024
17025 2013-06-13 Jan Hubicka <jh@suse.cz>
17026
17027 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17028 Local comdats are not externally visible.
17029 * symtab.c (dump_symtab_base): Dump externally visible.
17030 (verify_symtab_base): Verify back links in the symtab hash.
17031
17032 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17033
17034 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
17035 CONVERT_EXPR as equal nodes.
17036
17037 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17038
17039 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
17040
17041 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17042
17043 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
17044 Generalize to complex and vector.
17045 * tree.c (build_all_ones_cst): New function.
17046 * tree.h (build_all_ones_cst): Declare it.
17047
17048 2013-06-13 Alan Modra <amodra@gmail.com>
17049
17050 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
17051 * config/rs6000/rs6000.md (signbittf2): New insn.
17052 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
17053 (abstf2_internal, cmptf_internal2): Likewise.
17054 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
17055
17056 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
17057 Pat Haugen <pthaugen@us.ibm.com>
17058 Peter Bergner <bergner@vnet.ibm.com>
17059
17060 * config/rs6000/rs6000.c (emit_load_locked): Add support for
17061 power8 byte, half-word, and quad-word atomic instructions.
17062 (emit_store_conditional): Likewise.
17063 (rs6000_expand_atomic_compare_and_swap): Likewise.
17064 (rs6000_expand_atomic_op): Likewise.
17065
17066 * config/rs6000/sync.md (larx): Add new modes for power8.
17067 (stcx): Likewise.
17068 (AINT): New mode iterator to include TImode as well as normal
17069 integer modes on power8.
17070 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
17071 that VSX registers are not considered. Use AINT mode iterator
17072 instead of INT1 to allow inclusion of quad word atomic operations
17073 on power8.
17074 (load_locked<mode>): Likewise.
17075 (store_conditional<mode>): Likewise.
17076 (atomic_compare_and_swap<mode>): Likewise.
17077 (atomic_exchange<mode>): Likewise.
17078 (atomic_nand<mode>): Likewise.
17079 (atomic_fetch_<fetchop_name><mode>): Likewise.
17080 (atomic_nand_fetch<mode>): Likewise.
17081 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
17082 each type.
17083 (ATOMIC): On power8, add QImode, HImode modes.
17084 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
17085 modes that promote to SImode.
17086 (load_lockedti): Convert TImode arguments to PTImode, so that we
17087 get a guaranteed even/odd register pair.
17088 (load_lockedpti): Likewise.
17089 (store_conditionalti): Likewise.
17090 (store_conditionalpti): Likewise.
17091
17092 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
17093 atomic load/store instructions.
17094 (HSI): Likewise.
17095
17096 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
17097
17098 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
17099 loads.
17100 (insn_count): New attribute, with most cases extracted from...
17101 (length): ...here. Redefine most cases in terms of insn_count.
17102 (single_insn): Delete.
17103 (can_delay): Use insn_count to check for single instructions.
17104 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
17105 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
17106 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
17107 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
17108 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
17109 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
17110 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
17111 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
17112 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
17113 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
17114 rather than "length".
17115 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
17116 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
17117 Use "insn_count" rather than "length".
17118 * config/mips/mips-dsp.md
17119 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
17120 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
17121 length attributes.
17122
17123 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
17124
17125 PR tree-optimization/57361
17126 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
17127
17128 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
17129
17130 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
17131 to split.
17132 (aarch64_simd_combine<mode>): New instruction expansion.
17133 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
17134 function prototype.
17135 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
17136 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
17137
17138 2013-06-12 Jan Hubicka <jh@suse.cz>
17139
17140 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
17141 decl has when in streaming stage.
17142 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
17143 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
17144
17145 2013-06-12 Roland Stigge <stigge@antcom.de>
17146
17147 PR target/57578
17148 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
17149
17150 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17151
17152 PR tree-optimization/57537
17153 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
17154 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
17155
17156 2013-06-12 Richard Biener <rguenther@suse.de>
17157
17158 * data-streamer.h (streamer_write_char_stream): CSE
17159 obs->current_pointer.
17160 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
17161 streamer_write_char_stream manually and optimize the resulting loop.
17162 (streamer_write_hwi_stream): Likewise.
17163
17164 2013-06-12 Jan Hubicka <jh@suse.cz>
17165
17166 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
17167 * cgraph.h (varpool_create_empty_node): Declare.
17168 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
17169 duplicated nodes.
17170 * symtab.c (symtab_unregister_node): Be lax about missin entries
17171 in node hash.
17172 (symtab_get_node): Update comment.
17173 * varpool.c (varpool_create_empty_node): Break out from ...
17174 (varpool_node_for_decl): ... here.
17175 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
17176
17177 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
17178
17179 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
17180 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
17181 part. Use straight-line flow at the end.
17182 <COMPONENT_REF>: Remove superfluous else.
17183 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
17184
17185 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17186
17187 PR target/56564
17188 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
17189 target hook even for !TREE_PUBLIC decls. If no resolution info
17190 is available, return false for common and external decls.
17191
17192 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17193
17194 * config/rl78/constraints.md (U): New constraint.
17195 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
17196 valloc attribute.
17197
17198 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17199
17200 PR target/57589
17201 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
17202 to allow returning address to AT_PLATFORM name.
17203
17204 2013-06-11 Jan Hubicka <jh@suse.cz>
17205
17206 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
17207 * cgraph.h (symtab_node_base): Add weakref flag.
17208 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
17209 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
17210 (output_weakrefs): Use weakref flag.
17211 * fold-const.c (simple_operand_p): Handle WEAK.
17212 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
17213 * ipa.c (varpool_externally_visible_p): Drop weakref.
17214 (function_and_variable_visibility): Update comment; fix weakref
17215 sanity checks; do not clear DECL_WEAK on them.
17216 * lto-cgraph.c (lto_output_node): update.
17217 (lto_output_varpool_node): Update.
17218 (input_overwrite_node): Update.
17219 (input_node): Update.
17220 (input_varpool_node): Update.
17221 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
17222 (lto_symtab_merge_symbols): Add sanity check.
17223 (lto_symtab_prevailing_decl): Do not special case weakrefs.
17224 * passes.c (rest_of_decl_compilation): Set static flag, too.
17225 * symtab.c (dump_symtab_base): Dump weakref.
17226 (verify_symtab_base): Sanity check weakrefs.
17227 (symtab_make_decl_local): Remove duplicated code.
17228 (symtab_alias_ultimate_target): Simplify.
17229 * varpool.c (varpool_create_variable_alias): Set weakref flag.
17230
17231 2013-06-11 Tom de Vries <tom@codesourcery.com>
17232
17233 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
17234 sequence_vect == NULL.
17235
17236 2013-06-11 DJ Delorie <dj@redhat.com>
17237
17238 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
17239 (rl78_unwind_word_mode): New.
17240
17241 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17242
17243 * final.c (debug_prefix_maps): Make static.
17244
17245 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17246
17247 * function.c (initial_trampoline): Remove stray copy.
17248
17249 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
17250
17251 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
17252
17253 2013-06-11 Martin Jambor <mjambor@suse.cz>
17254
17255 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
17256 within bounds at the beginning of the function.
17257
17258 2013-06-11 Alan Modra <amodra@gmail.com>
17259
17260 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
17261 reporting.
17262 (get_named_section): Don't NULL !DECL_P decl.
17263
17264 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
17265
17266 * doc/invoke.texi (core-avx2): Document.
17267 (slm): Likewise.
17268 (atom): Updated with MOVBE.
17269
17270 2013-06-11 Richard Biener <rguenther@suse.de>
17271
17272 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
17273
17274 2013-06-11 Anton Blanchard <anton@samba.org>
17275
17276 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
17277 correct shift value in little-endian mode.
17278
17279 2013-06-11 Jakub Jelinek <jakub@redhat.com>
17280
17281 PR target/56564
17282 * varasm.c (get_variable_align): Move #endif to the right place.
17283
17284 2013-06-10 Cary Coutant <ccoutant@google.com>
17285
17286 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
17287 for hash so that hash table traversal order is deterministic.
17288
17289 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
17290 Pat Haugen <pthaugen@us.ibm.com>
17291 Peter Bergner <bergner@vnet.ibm.com>
17292
17293 * config/rs6000/vector.md (GPR move splitter): Do not split moves
17294 of vectors in GPRS if they are direct moves or quad word load or
17295 store moves.
17296
17297 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
17298 declaration.
17299 (direct_move_p): Likewise.
17300 (quad_load_store_p): Likewise.
17301
17302 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
17303 classes into bins based on the physical register type.
17304 (reg_class_to_reg_type): Likewise.
17305 (IS_STD_REG_TYPE): Likewise.
17306 (IS_FP_VECT_REG_TYPE): Likewise.
17307 (reload_fpr_gpr): Arrays to determine what insn to use if we can
17308 use direct move instructions.
17309 (reload_gpr_vsx): Likewise.
17310 (reload_vsx_gpr): Likewise.
17311 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
17312 information that is a simplification of register classes. Also
17313 precalculate direct move reload helpers.
17314 (direct_move_p): New function to return true if the operation can
17315 be done as a direct move instruciton.
17316 (quad_load_store_p): New function to return true if the operation
17317 is a quad memory operation.
17318 (rs6000_legitimize_address): If quad memory, only allow register
17319 indirect for TImode addresses.
17320 (rs6000_legitimate_address_p): Likewise.
17321 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
17322 (rs6000_reload_register_type): Likewise.
17323 (register_to_reg_type): Return register type.
17324 (rs6000_secondary_reload_simple_move): New helper function for
17325 secondary reload and secondary memory needed to identify anything
17326 that is a simple move, and does not need reloading.
17327 (rs6000_secondary_reload_direct_move): New helper function for
17328 secondary reload to identify cases that can be done with several
17329 instructions via the direct move instructions.
17330 (rs6000_secondary_reload_move): New helper function for secondary
17331 reload to identify moves between register types that can be done.
17332 (rs6000_secondary_reload): Add support for quad memory operations
17333 and for direct move.
17334 (rs6000_secondary_memory_needed): Likewise.
17335 (rs6000_debug_secondary_memory_needed): Change argument names.
17336 (rs6000_output_move_128bit): New function to return the move to
17337 use for 128-bit moves, including knowing about the various
17338 limitations of quad memory operations.
17339
17340 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
17341 memory operations. call rs6000_output_move_128bit for the actual
17342 instruciton(s) to generate.
17343 (vsx_movti_64bit): Likewise.
17344
17345 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
17346 (UNSPEC_P8V_MTVSRWZ): Likewise.
17347 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
17348 (UNSPEC_P8V_MTVSRD): Likewise.
17349 (UNSPEC_P8V_XXPERMDI): Likewise.
17350 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
17351 (UNSPEC_FUSION_GPR): Likewise.
17352 (FMOVE128_GPR): New iterator for direct move.
17353 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
17354 (f32_sv): Likewise.
17355 (f32_dm): Likewise.
17356 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
17357 loads and direct move instructions.
17358 (zero_extendsidi2_lfiwzx): Likewise.
17359 (extendsidi2_lfiwax): Likewise.
17360 (extendsidi2_nocell): Likewise.
17361 (floatsi<mode>2_lfiwax): Likewise.
17362 (lfiwax): Likewise.
17363 (floatunssi<mode>2_lfiwzx): Likewise.
17364 (lfiwzx): Likewise.
17365 (fix_trunc<mode>_stfiwx): Likewise.
17366 (fixuns_trunc<mode>_stfiwx): Likewise.
17367 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
17368 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
17369 (parity<mode>2_cmpb): Set length/type attr.
17370 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
17371 for 'mr.' to fast_compare.
17372 (bpermd_<mode>): Change type attr to popcnt.
17373 (p8_fmrgow_<mode>): New insns for power8 direct move support.
17374 (p8_mtvsrwz_1): Likewise.
17375 (p8_mtvsrwz_2): Likewise.
17376 (reload_fpr_from_gpr<mode>): Likewise.
17377 (p8_mtvsrd_1): Likewise.
17378 (p8_mtvsrd_2): Likewise.
17379 (p8_xxpermdi_<mode>): Likewise.
17380 (reload_vsx_from_gpr<mode>): Likewise.
17381 (reload_vsx_from_gprsf): Likewise.
17382 (p8_mfvsrd_3_<mode>): LIkewise.
17383 (reload_gpr_from_vsx<mode>): Likewise.
17384 (reload_gpr_from_vsxsf): Likewise.
17385 (p8_mfvsrd_4_disf): Likewise.
17386 (multi-word GPR splits): Do not split direct moves or quad memory
17387 operations.
17388
17389 2013-06-10 David Malcolm <dmalcolm@redhat.com>
17390
17391 * tree-into-ssa.c (interesting_blocks): Make static.
17392
17393 2013-06-10 Jakub Jelinek <jakub@redhat.com>
17394
17395 PR target/56564
17396 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
17397 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
17398 Use DATA_ABI_ALIGNMENT for that case instead if defined.
17399 (get_variable_align): New function.
17400 (get_variable_section, emit_bss, emit_common,
17401 assemble_variable_contents, place_block_symbol): Use
17402 get_variable_align instead of DECL_ALIGN.
17403 (assemble_noswitch_variable): Add align argument, use it
17404 instead of DECL_ALIGN.
17405 (assemble_variable): Adjust caller. Use get_variable_align
17406 instead of DECL_ALIGN.
17407 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
17408 caller.
17409 (DATA_ABI_ALIGNMENT): Define.
17410 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
17411 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
17412 opt is false, only return the psABI mandated alignment increase.
17413 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
17414 (DATA_ABI_ALIGNMENT): ... this.
17415 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
17416 (DATA_ABI_ALIGNMENT): ... this.
17417 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
17418 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
17419 (DATA_ABI_ALIGNMENT): ... this.
17420 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
17421 * doc/tm.texi: Regenerated.
17422
17423 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
17424
17425 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
17426 cmp_code to construct REG_EQUAL note.
17427
17428 2013-06-09 Jakub Jelinek <jakub@redhat.com>
17429
17430 PR target/57568
17431 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
17432 that operands[2] doesn't overlap with operands[0].
17433
17434 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
17435 Jan Hubicka <jh@suse.cz>
17436
17437 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
17438 hack to mark symbols as used.
17439
17440 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
17441
17442 PR rtl-optimization/57559
17443 * lra-constraints.c (process_alt_operands): Don't discourage
17444 memory with known offset for offsetable memory constraint.
17445 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
17446
17447 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
17448
17449 * varasm.c (struct oc_local_state): Reorder fields.
17450 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
17451 and adjust accordingly.
17452 (output_constructor): Reorder initialization code and adjust call to
17453 output_constructor_bitfield.
17454
17455 2013-06-07 Jan Hubicka <jh@suse.cz>
17456
17457 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
17458
17459 2013-06-07 David Malcolm <dmalcolm@redhat.com>
17460
17461 * tree-object-size.c (unknown): Make const.
17462
17463 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17464
17465 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
17466 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
17467 for last alternative in the cpu_facility attribute.
17468
17469 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17470
17471 PR target/56315
17472 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
17473 (xordi3): Change operand 2 constraint to arm_xordi_operand.
17474 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
17475 * config/arm/constraints.md (Dg): New constraint.
17476 * config/arm/neon.md (xordi3_neon): Remove.
17477 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
17478 * config/arm/predicates.md (arm_xordi_operand): New predicate.
17479
17480 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17481
17482 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
17483 Clean up alternatives.
17484
17485 2013-06-07 Alan Modra <amodra@gmail.com>
17486
17487 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
17488 va_list_gpr_size.
17489
17490 2013-06-07 Alan Modra <amodra@gmail.com>
17491
17492 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
17493
17494 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17495
17496 * config/arm/constraints.md (Df): New constraint.
17497 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
17498 Correct length attribute for last two alternatives.
17499
17500 2013-06-07 Alan Modra <amodra@gmail.com>
17501
17502 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
17503 override user -mfp-in-toc.
17504 (offsettable_ok_by_alignment): Consider just the current access
17505 rather than the whole object, unless BLKmode. Handle
17506 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
17507 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
17508 for -mcmodel=medium.
17509 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
17510 override user -mfp-in-toc or -msum-in-toc. Default to
17511 -mno-fp-in-toc for -mcmodel=medium.
17512
17513 2013-06-06 DJ Delorie <dj@redhat.com>
17514
17515 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
17516 TARGET_VALID_POINTER_MODE.
17517
17518 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
17519 Pat Haugen <pthaugen@us.ibm.com>
17520 Peter Bergner <bergner@vnet.ibm.com>
17521
17522 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
17523 Document new power8 builtins.
17524
17525 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
17526 condition code register, to allow 128-bit logical operations to be
17527 done in the VSX or GPR registers.
17528 (nor<mode>3): Use the canonical form for nor.
17529 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
17530 vclz*, and vpopcnt* vector instructions.
17531 (nand<mode>3): Likewise.
17532 (orc<mode>3): Likewise.
17533 (clz<mode>2): LIkewise.
17534 (popcount<mode>2): Likewise.
17535
17536 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
17537 that only the GPRs are recognized.
17538
17539 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
17540 support for new power8 builtins.
17541
17542 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
17543 builtin functions.
17544 (xscvdpspn): Likewise.
17545 (vclz): Likewise.
17546 (vclzb): Likewise.
17547 (vclzh): Likewise.
17548 (vclzw): Likewise.
17549 (vclzd): Likewise.
17550 (vpopcnt): Likewise.
17551 (vpopcntb): Likewise.
17552 (vpopcnth): Likewise.
17553 (vpopcntw): Likewise.
17554 (vpopcntd): Likewise.
17555 (vgbbd): Likewise.
17556 (vmrgew): Likewise.
17557 (vmrgow): Likewise.
17558 (eqv): Likewise.
17559 (eqv_v16qi3): Likewise.
17560 (eqv_v8hi3): Likewise.
17561 (eqv_v4si3): Likewise.
17562 (eqv_v2di3): Likewise.
17563 (eqv_v4sf3): Likewise.
17564 (eqv_v2df3): Likewise.
17565 (nand): Likewise.
17566 (nand_v16qi3): Likewise.
17567 (nand_v8hi3): Likewise.
17568 (nand_v4si3): Likewise.
17569 (nand_v2di3): Likewise.
17570 (nand_v4sf3): Likewise.
17571 (nand_v2df3): Likewise.
17572 (orc): Likewise.
17573 (orc_v16qi3): Likewise.
17574 (orc_v8hi3): Likewise.
17575 (orc_v4si3): Likewise.
17576 (orc_v2di3): Likewise.
17577 (orc_v4sf3): Likewise.
17578 (orc_v2df3): Likewise.
17579
17580 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
17581 allow power8 quad mode in 64-bit.
17582 (rs6000_builtin_vectorized_function): Add support to vectorize
17583 ISA 2.07 count leading zeros, population count builtins.
17584 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
17585 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
17586 (builtin_function_type): Add vgbbd builtin function which takes an
17587 unsigned argument.
17588 (altivec_expand_vec_perm_const): Add support for new power8 merge
17589 instructions.
17590
17591 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
17592 that does not include TImdoe for use with 32-bit.
17593 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
17594 instructions.
17595 (UNSPEC_VSX_CVDPSPN): Likewise.
17596 (vsx_xscvdpspn): Likewise.
17597 (vsx_xscvspdpn): Likewise.
17598 (vsx_xscvdpspn_scalar): Likewise.
17599 (vsx_xscvspdpn_directmove): Likewise.
17600 (vsx_and<mode>3): Split logical operations into 32-bit and
17601 64-bit. Add support to do logical operations on TImode as well as
17602 VSX vector types. Allow logical operations to be done in either
17603 VSX registers or in general purpose registers in 64-bit mode. Add
17604 splitters if GPRs were used. For AND, add clobber of CCmode to
17605 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
17606 encoding.
17607 (vsx_and<mode>3_32bit): Likewise.
17608 (vsx_and<mode>3_64bit): Likewise.
17609 (vsx_ior<mode>3): Likewise.
17610 (vsx_ior<mode>3_32bit): Likewise.
17611 (vsx_ior<mode>3_64bit): Likewise.
17612 (vsx_xor<mode>3): Likewise.
17613 (vsx_xor<mode>3_32bit): Likewise.
17614 (vsx_xor<mode>3_64bit): Likewise.
17615 (vsx_one_cmpl<mode>2): Likewise.
17616 (vsx_one_cmpl<mode>2_32bit): Likewise.
17617 (vsx_one_cmpl<mode>2_64bit): Likewise.
17618 (vsx_nor<mode>3): Likewise.
17619 (vsx_nor<mode>3_32bit): Likewise.
17620 (vsx_nor<mode>3_64bit): Likewise.
17621 (vsx_andc<mode>3): Likewise.
17622 (vsx_andc<mode>3_32bit): Likewise.
17623 (vsx_andc<mode>3_64bit): Likewise.
17624 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
17625 and xxlorc instructions.
17626 (vsx_eqv<mode>3_64bit): Likewise.
17627 (vsx_nand<mode>3_32bit): Likewise.
17628 (vsx_nand<mode>3_64bit): Likewise.
17629 (vsx_orc<mode>3_32bit): Likewise.
17630 (vsx_orc<mode>3_64bit): Likewise.
17631
17632 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
17633
17634 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
17635 instruction.
17636 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
17637 (p8_vmrgow): Likewise.
17638 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
17639 GPRs to be split under VSX.
17640 (p8v_clz<mode>2): Add power8 count leading zero support.
17641 (p8v_popcount<mode>2): Add power8 population count support.
17642 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
17643 support.
17644
17645 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
17646 instruction.
17647
17648 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
17649 builtin functions.
17650 (vec_nand): Likewise.
17651 (vec_vclz): Likewise.
17652 (vec_vclzb): Likewise.
17653 (vec_vclzd): Likewise.
17654 (vec_vclzh): Likewise.
17655 (vec_vclzw): Likewise.
17656 (vec_vgbbd): Likewise.
17657 (vec_vmrgew): Likewise.
17658 (vec_vmrgow): Likewise.
17659 (vec_vpopcnt): Likewise.
17660 (vec_vpopcntb): Likewise.
17661 (vec_vpopcntd): Likewise.
17662 (vec_vpopcnth): Likewise.
17663 (vec_vpopcntw): Likewise.
17664
17665 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17666
17667 PR rtl-optimization/57468
17668 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
17669 spilled pseudos.
17670
17671 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17672
17673 PR rtl-optimization/57459
17674 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
17675 type when setting live regs.
17676
17677 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17678
17679 * config/s390/s390.opt (mlra): New option.
17680 * config/s390/s390.c (s390_decompose_address): Check displacement
17681 for all registers for LRA.
17682 (s390_secondary_reload): Don't used secondary reloads for LRA.
17683 (s390_lra_p): New function.
17684 (TARGET_LRA_P): Define.
17685 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
17686 of attribute cpu_facility to zarch for the last alternative.
17687 (*cmpmem_short): Ditto.
17688
17689 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
17690
17691 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
17692 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
17693 (arm_expand_prologue): Likewise.
17694
17695 2013-06-06 Teresa Johnson <tejohnson@google.com>
17696
17697 PR c++/53743
17698 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
17699 as this is now done by redirect_edge_and_branch_force.
17700 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
17701 barriers, and fix interaction with splitting.
17702 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
17703 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
17704 reflect changes made in the routine.
17705 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
17706 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
17707 since this is called in cfglayout mode, and replace partition fixup
17708 with assert as that is now done by force_nonfallthru_and_redirect.
17709 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
17710 already be marked with region crossing note.
17711 (insert_section_boundary_note): Make non-static, gate on flag
17712 has_bb_partition, rewrite to also check for multiple partitions.
17713 (rest_of_handle_reorder_blocks): Remove call to
17714 insert_section_boundary_note, now done later during free_cfg.
17715 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
17716 * bb-reorder.h (insert_section_boundary_note): Declare.
17717 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
17718 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
17719 invoke insert_section_boundary_note.
17720 (try_redirect_by_replacing_jump): Remove unnecessary
17721 check for region crossing note.
17722 (fixup_partition_crossing): New function.
17723 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
17724 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
17725 in non-cfglayout mode.
17726 (force_nonfallthru_and_redirect): Fixup partition boundaries,
17727 remove old code that tried to do this. Emit barrier correctly
17728 when we are in cfglayout mode.
17729 (last_bb_in_partition): New function.
17730 (rtl_split_edge): Correctly fixup partition boundaries.
17731 (commit_one_edge_insertion): Remove old code that tried to
17732 fixup region crossing edge since this is now handled in
17733 split_block, and set up insertion point correctly since
17734 block may now end in a jump.
17735 (verify_hot_cold_block_grouping): Guard against checking when not in
17736 linearized RTL mode.
17737 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
17738 notes.
17739 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
17740 rtl_verify_flow_info, so not called in cfglayout mode.
17741 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
17742 (fixup_reorder_chain): Remove old code that attempted to fixup region
17743 crossing note as this is now handled in force_nonfallthru_and_redirect.
17744 (duplicate_insn_chain): Don't duplicate switch section notes.
17745 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
17746 note.
17747 * basic-block.h (emit_barrier_after_bb): Declare.
17748
17749 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17750
17751 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
17752 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
17753 arm_usatsihi): Adjust alternatives for arm_restrict_it.
17754
17755 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17756
17757 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
17758 where appropriate.
17759 * config/arm/ldmstm.md: Regenerate.
17760
17761 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17762
17763 * config/arm/sync.md (atomic_loaddi_1):
17764 Disable predication for arm_restrict_it.
17765 (arm_load_exclusive<mode>): Likewise.
17766 (arm_load_exclusivesi): Likewise.
17767 (arm_load_exclusivedi): Likewise.
17768 (arm_load_acquire_exclusive<mode>): Likewise.
17769 (arm_load_acquire_exclusivesi): Likewise.
17770 (arm_load_acquire_exclusivedi): Likewise.
17771 (arm_store_exclusive<mode>): Likewise.
17772 (arm_store_exclusive<mode>): Likewise.
17773 (arm_store_release_exclusivedi): Likewise.
17774 (arm_store_release_exclusive<mode>): Likewise.
17775
17776 2013-06-06 Richard Biener <rguenther@suse.de>
17777
17778 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
17779 after LTO_null.
17780 (lto_tag_is_tree_code_p): Adjust.
17781 (lto_tag_is_gimple_code_p): Likewise.
17782 (lto_gimple_code_to_tag): Likewise.
17783 (lto_tag_to_gimple_code): Likewise.
17784 (lto_tree_code_to_tag): Likewise.
17785 (lto_tag_to_tree_code): Likewise.
17786 * data-streamer.h (streamer_write_hwi_in_range): Use
17787 uhwi streaming to stream the normalized range.
17788 (streamer_read_hwi_in_range): Likewise.
17789
17790 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17791
17792 * config/arm/arm.md (enabled_for_depr_it): New attribute.
17793 (predicable_short_it): Likewise.
17794 (predicated): Likewise.
17795 (enabled): Handle above.
17796 (define_cond_exec): Set predicated attribute to yes.
17797
17798 2013-06-05 Mike Stump <mikestump@comcast.net>
17799
17800 * gdbinit.in (__FUNCTION__): Add.
17801
17802 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
17803
17804 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
17805 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
17806
17807 2013-06-05 Jan Hubicka <jh@suse.cz>
17808
17809 * varasm.c (mark_decl_referenced): Revert the removal until targets
17810 are fixed.
17811
17812 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
17813
17814 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
17815 instead of mark_decl_referenced.
17816
17817 2013-06-05 Jan Hubicka <jh@suse.cz>
17818
17819 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
17820 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
17821 and symtab_used_from_object_file_p.
17822 (cgraph_make_node_local_1): Clear forced_by_abi.
17823 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
17824 * cgraph.h (symtab_node_base): Add forced_by_abi.
17825 (decide_is_variable_needed): Remove.
17826 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
17827 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
17828 (decide_is_symbol_needed): ... this one; handle symbols in general;
17829 always analyze virtuals; honnor forced_by_abi.
17830 (cgraph_finalize_function): Update.
17831 (varpool_finalize_decl): Update.
17832 (symbol_defined_and_needed): Remove.
17833 (analyze_functions): Update.
17834 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17835 output_refs, input_overwrite_node): Handle forced_by_abi.
17836 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
17837 (address_taken_from_non_vtable_p): ... this one.
17838 (comdat_can_be_unshared_p_1): New function.
17839 (cgraph_comdat_can_be_unshared_p): Rename to ...
17840 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
17841 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
17842 (function_and_variable_visibility): Clear forced_by_abi as needed.
17843 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
17844 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
17845 * symtab.c (dump_symtab_base): Dump forced_by_abi.
17846 * varpool.c (decide_is_variable_needed): Remove.
17847
17848 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17849
17850 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
17851 (arm_option_override): Override arm_restrict_it where appropriate.
17852 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
17853 * config/arm/arm.opt (mrestrict-it): New command-line option.
17854 * doc/invoke.texi: Document -mrestrict-it.
17855
17856 2013-06-05 David Malcolm <dmalcolm@redhat.com>
17857
17858 * tsan.c (tsan_atomic_table): Make const.
17859
17860 2013-06-05 Richard Biener <rguenther@suse.de>
17861
17862 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
17863 index associated with the tree we are supposed to replace.
17864 * tree-streamer-out.c (pack_ts_base_value_fields): Output
17865 TREE_ASM_WRITTEN as zero for everything but SSA names.
17866
17867 2013-06-05 David Malcolm <dmalcolm@redhat.com>
17868
17869 * tree-ssa-structalias.c (call_stmt_vars): Make static.
17870
17871 2013-06-04 Jan Hubicka <jh@suse.cz>
17872
17873 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
17874 (input_node, input_varpool_node): Handle correctly external same
17875 body aliases.
17876 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
17877 nodes at ltrans stage.
17878
17879 2013-06-04 Jan Hubicka <jh@suse.cz>
17880
17881 * ipa-inline.c (update_caller_keys): Fix availability test.
17882 (update_callee_keys): Likewise.
17883 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
17884 to follow ELF standard.
17885
17886 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
17887
17888 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
17889 (mips64r5900el-*-elf*): New configurations.
17890 * config/mips/mips-cpus.def (r5900): New processor.
17891 * config/mips/mips-tables.opt: Regenerate.
17892 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
17893 (mips_issue_rate): Handle PROCESSOR_R5900.
17894 (mips_reorg_process_insns): Force reorder mode for the R5900.
17895 * config/mips/mips.h (TARGET_MIPS5900): Define.
17896 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
17897 TARGET_MIPS5900.
17898 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
17899 TARGET_MIPS5900.
17900 * config/mips/mips.md (processor): Add r5900.
17901 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
17902
17903 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17904
17905 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
17906 into function to generate MOVI instruction.
17907 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
17908 (aarch64_preferred_simd_mode): Turn into wrapper.
17909 (aarch64_output_scalar_simd_mov_immediate): New function.
17910 * config/aarch64/aarch64-protos.h: Add prototype for above.
17911
17912 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17913
17914 * config/aarch64/aarch64.c (simd_immediate_info): Remove
17915 element_char member.
17916 (sizetochar): Return signed char.
17917 (aarch64_simd_valid_immediate): Remove elchar and other
17918 unnecessary variables.
17919 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
17920 Calculate element_char as required.
17921 * config/aarch64/aarch64-protos.h: Update and move prototype
17922 for aarch64_output_simd_mov_immediate.
17923 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
17924 Update arguments.
17925
17926 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17927
17928 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
17929 information completed by aarch64_simd_valid_immediate.
17930 (aarch64_legitimate_constant_p): Update arguments.
17931 (aarch64_simd_valid_immediate): Work with struct rather than many
17932 pointers.
17933 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
17934 (aarch64_simd_make_constant): Update arguments.
17935 (aarch64_output_simd_mov_immediate): Work with struct rather than
17936 many pointers. Output immediate directly rather than as operand.
17937 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
17938 Update prototype.
17939 * config/aarch64/constraints.md (Dn): Update arguments.
17940
17941 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17942
17943 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
17944 longer static.
17945 (aarch64_simd_immediate_valid_for_move): Remove.
17946 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
17947 (aarch64_simd_make_constant): Update call.
17948 (aarch64_output_simd_mov_immediate): Update call.
17949 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
17950 Add prototype.
17951 * config/aarch64/constraints.md (Dn): Update call.
17952
17953 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17954
17955 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
17956 return type to bool for prototype.
17957 (aarch64_legitimate_constant_p): Check for true instead of not -1.
17958 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
17959 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
17960
17961 2013-06-04 Catherine Moore <clm@codesourcery.com>
17962
17963 * config/mips/mips.opt (meva): New.
17964 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
17965 (ASM_SPEC): Handle -meva.
17966 * doc/invoke.texi (meva): Document.
17967
17968 2013-06-04 Alan Modra <amodra@gmail.com>
17969
17970 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
17971 constant output.
17972
17973 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17974
17975 * rtl.def: Add extra fourth optional field to define_cond_exec.
17976 * gensupport.c (process_one_cond_exec): Process attributes from
17977 define_cond_exec.
17978 * doc/md.texi: Document fourth field in define_cond_exec.
17979
17980 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
17981
17982 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
17983 out the processing order as in store_bit_field_1.
17984
17985 2013-06-04 Jan Hubicka <jh@suse.cz>
17986
17987 PR middle-end/57500
17988 * cgraphunit.c (cgraph_process_same_body_aliases): Create
17989 non-VAR_DECL node if it does not exist yet.
17990
17991 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
17992
17993 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
17994 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
17995 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
17996 target_cpu_default setting.
17997
17998 2013-06-03 Teresa Johnson <tejohnson@google.com>
17999
18000 * dumpfile.c (opt_info_switch_p): Change -fopt-info
18001 default to -fopt-info=optimized instead of all.
18002 * doc/invoke.texi: Ditto.
18003 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
18004 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
18005 (execute_vect_slp): Emit BB vectorization success under
18006 MSG_OPTIMIZED_LOCATIONS.
18007 * tree-vect-slp.c (vect_slp_transform_bb): Change
18008 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
18009 * tree-vect-loop.c (vect_transform_loop): Ditto.
18010
18011 2013-06-03 Jason Merrill <jason@redhat.com>
18012
18013 PR c++/57415
18014 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
18015 Use TARGET_EXPR for C++.
18016
18017 2013-06-03 Jakub Jelinek <jakub@redhat.com>
18018
18019 PR rtl-optimization/57268
18020 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
18021 if DEBUG_INSN_P (insn).
18022
18023 Reapply
18024 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18025
18026 PR rtl-optimization/57268
18027 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18028 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18029
18030 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18031
18032 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
18033 (ix86_avoid_lea_for_addr): Likewise.
18034 (exact_dependency_1): Likewise.
18035 (ix86_adjust_cost): Likewise.
18036 (swap_top_of_ready_list): Fix formatting and !reload_completed check
18037 removed.
18038 (do_reorder_for_imul): Fix typo, formatting and
18039 !reload_completed check removed.
18040 (ix86_sched_reorder): Fix typo and formatting.
18041 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
18042 list.
18043
18044 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
18045
18046 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
18047
18048 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
18049
18050 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
18051 <STRING_CST>: Likewise.
18052 <VECTOR_CST>: Likewise.
18053
18054 2013-06-01 Janus Weil <janus@gcc.gnu.org>
18055 Mikael Morin <mikael@gcc.gnu.org>
18056
18057 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
18058 * config.in: Regenerated.
18059 * configure: Regenerated.
18060
18061 2013-06-01 Jan Hubicka <jh@suse.cz>
18062
18063 PR middle-end/57366
18064 * cgraphunit.c (compile): When weakref is not supported,
18065 set up transparent aliases before final output pass.
18066 * varasm.c (assemble_alias): Do not try to do it here.
18067
18068 2013-06-01 Jan Hubicka <jh@suse.cz>
18069
18070 PR middle-end/57467
18071 * passes.c (for_per_function): Skip unanalyzed functions.
18072
18073 2013-06-01 Jan Hubicka <jh@suse.cz>
18074
18075 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
18076 (lto_symtab_merge_symbols_1): ... this one.
18077 (lto_symtab_merge_cgraph_nodes): Rename to ...
18078 (lto_symtab_merge_symbols): ... this one; simplify.
18079 * cgraph.c (same_body_aliases_done): Rename to ...
18080 (cpp_implicit_aliases_done): ... this one.
18081 (cgraph_create_function_alias): Update.
18082 (cgraph_same_body_alias): Update.
18083 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
18084 (verify_edge_corresponds_to_fndecl): Simplify.
18085 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
18086 (cgraph_node): Remove same_body_alias.
18087 (varpool_node): Remove alias_of and extra_name_alias.
18088 (same_body_aliases_done): Rename to ..
18089 (cpp_implicit_aliases_done): ... this one.
18090 (symtab_alias_ultimate_target): Add default parameter.
18091 (symtab_resolve_alias): New function.
18092 (fixup_same_cpp_alias_visibility): Declare.
18093 (cgraph_function_node): Add default parameter.
18094 (cgraph_node_asm_name): Likewise.
18095 (cgraph_function_or_thunk_node): Add default parameter; do
18096 not ICE when it is NULL.
18097 (varpool_variable_node): Likewise.
18098 * tree-emutls.c (create_emultls_var): Update.
18099 (ipa_lower_emutls): Update.
18100 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
18101 (cgraph_reset_node): Reset alias info.
18102 (cgraph_finalize_function): Update.
18103 (fixup_same_cpp_alias_visibility): Move to symtab.c.
18104 (analyze_function): Simplify.
18105 (cgraph_process_same_body_aliases): Simplify.
18106 (analyze_functions): Fixup same body aliases.
18107 (handle_alias_pairs): Simplify.
18108 (assemble_thunk): Update.
18109 (assemble_thunks_and_aliases): Update.
18110 (output_weakrefs): Rewrite.
18111 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
18112 (lto_output_varpool_node): Likewise.
18113 (compute_ltrans_boundary): Remve assert.
18114 (get_alias_symbol): New functoin.
18115 (input_node): Rewrite alias handling.
18116 (input_varpool_node): Likewise.
18117 * ipa-pure-const.c (propagate_pure_const): Fix formating.
18118 * ipa.c (process_references): Handle weakrefs correctly.
18119 (symtab_remove_unreachable_nodes): Likewise.
18120 * trans-mem.c (get_cg_data): Update.
18121 (ipa_tm_create_version_alias): Update.
18122 (ipa_tm_execute): Update.
18123 * symtab.c (dump_symtab_base): Dump aliases.
18124 (verify_symtab_base): Verify aliases.
18125 (symtab_node_availability): New function.
18126 (symtab_alias_ultimate_target): Simplify.
18127 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
18128 handle all the fixup cases.
18129 (symtab_resolve_alias): New function.
18130 * passes.c (ipa_write_summaries): Handle weakrefs.
18131 * varpool.c (varpool_analyze_node): Simplify.
18132 (assemble_aliases): Update.
18133 (varpool_create_variable_alias): Simplify.
18134 (varpool_extra_name_alias): Simplify.
18135 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
18136 (lto_symtab_merge_symbols): ... this one.
18137
18138 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
18139
18140 Revert
18141 PR rtl-optimization/57268
18142 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18143 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18144
18145 2013-06-01 Tobias Burnus <burnus@net-b.de>
18146
18147 Partially reverted:
18148 2013-05-31 Tobias Burnus <burnus@net-b.de>
18149
18150 PR middle-end/57073
18151 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
18152 further up.
18153
18154 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18155
18156 PR rtl-optimization/57268
18157 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18158 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18159
18160 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18161
18162 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
18163 unordered comparison operators when -fno-trapping-math is in effect
18164 on the e500.
18165 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
18166 and implement unordered comparison operators properly on the e500.
18167
18168 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18169
18170 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
18171 for constant scalar integers.
18172 (simplify_relational_operation_1): Likewise.
18173
18174 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
18175
18176 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
18177 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
18178 Fix comment.
18179
18180 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18181 Igor Zamyatin <igor.zamyatin@intel.com>
18182
18183 Silvermont (SLM) architecture performance tuning.
18184 * config/i386/i386.h (enum ix86_tune_indices): Add
18185 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
18186 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
18187
18188 * config/i386/i386.c (initial_ix86_tune_features)
18189 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
18190 (ix86_lea_outperforms): Handle Silvermont tuning.
18191 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
18192 call.
18193 (ix86_use_lea_for_mov): Likewise.
18194 (ix86_avoid_lea_for_addr): Likewise.
18195 (ix86_lea_for_add_ok): Likewise.
18196 (exact_dependency_1): New function.
18197 (exact_store_load_dependency): Likewise.
18198 (ix86_adjust_cost): Handle Silvermont tuning.
18199 (do_reoder_for_imul): Likewise.
18200 (swap_top_of_ready_list): New function.
18201 (ix86_sched_reorder): Changed to handle Silvermont tuning.
18202
18203 * config/i386/i386.md (peepholes that split memory operand in fp
18204 converts): New.
18205
18206 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18207
18208 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
18209 Remove un-necessary braces.
18210
18211 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18212
18213 * config/aarch64/aarch64.c (aarch64_classify_symbol):
18214 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
18215
18216 2013-05-31 Tobias Burnus <burnus@net-b.de>
18217
18218 PR middle-end/57073
18219 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
18220
18221 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18222
18223 PR target/56315
18224 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
18225 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
18226 * config/arm/neon.md (iordi3_neon): Remove.
18227 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
18228 * config/arm/predicates.md (imm_for_neon_logic_operand):
18229 Move to earlier in the file.
18230 (neon_logic_op2): Likewise.
18231 (arm_iordi_operand_neon): New predicate.
18232
18233 2013-05-31 Richard Biener <rguenther@suse.de>
18234
18235 PR tree-optimization/57478
18236 PR tree-optimization/57453
18237 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
18238 are life as well.
18239
18240 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18241
18242 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
18243 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
18244
18245 2013-05-30 Tobias Burnus <burnus@net-b.de>
18246 Thomas Koenig <tkoenig@gcc.gnu.org>
18247
18248 PR middle-end/57073
18249 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
18250 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
18251
18252 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
18253
18254 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
18255
18256 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
18257
18258 * target.def (register_usage_leveling_p): New hook.
18259 * targhooks.c (default_register_usage_leveling_p): New.
18260 * targhooks.h (default_register_usage_leveling_p): New prototype.
18261 * lra-assigns.c (register_usage_leveling_p): Use the hook.
18262 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
18263 * doc/tm.texi: Update.
18264 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
18265
18266 2013-05-30 Ian Bolton <ian.bolton@arm.com>
18267
18268 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
18269 (*insv_reg<mode>): New define_insn.
18270
18271 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
18272
18273 PR rtl-optimization/57439
18274 * postreload.c (move2add_valid_value_p): Check that we have
18275 a zero subreg_regno_offset when accessing the register in
18276 the requested mode.
18277
18278 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18279 Igor Zamyatin <igor.zamyatin@intel.com>
18280
18281 Silvermont (SLM) architecture pipeline model, tuning and
18282 insn selection.
18283 * config.gcc: Add slm config options and target.
18284
18285 * config/i386/slm.md: New.
18286
18287 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
18288
18289 * config/i386/i386-c.c (ix86_target_macros_internal): New case
18290 PROCESSOR_SLM.
18291 (ix86_target_macros_internal): Likewise.
18292
18293 * config/i386/i386.c (slm_cost): New cost.
18294 (m_SLM): New macro flag.
18295 (initial_ix86_tune_features): Set m_SLM.
18296 (x86_accumulate_outgoing_args): Likewise.
18297 (x86_arch_always_fancy_math_387): Likewise.
18298 (processor_target_table): Add slm cost.
18299 (cpu_names): Add slm cpu name.
18300 (x86_option_override_internal): Set SLM ISA.
18301 (ix86_issue_rate): New case PROCESSOR_SLM.
18302 (ia32_multipass_dfa_lookahead): Likewise.
18303 (fold_builtin_cpu): Add slm.
18304
18305 * config/i386/i386.h (TARGET_SLM): New target macro.
18306 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
18307 (processor_type): Add PROCESSOR_SLM.
18308
18309 * config/i386/i386.md (cpu): Add new value "slm".
18310 (slm.md): Include slm.md.
18311
18312 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
18313 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18314
18315 * config/arm/arm-protos.h: Add and update function protos.
18316 * config/arm/arm.c (use_simple_return_p): New added.
18317 (thumb2_expand_return): Check simple_return flag.
18318 * config/arm/arm.md: Add simple_return and conditional simple_return.
18319 * config/arm/iterators.md: Add iterator for return and simple_return.
18320
18321 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18322
18323 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
18324 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
18325 (arm_emit_vfp_multi_reg_pop): Likewise.
18326 (thumb2_emit_ldrd_pop): Likewise.
18327 (arm_expand_epilogue): Add misc REG_CFA notes.
18328 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
18329
18330 2013-05-29 Lawrence Crowl <crowl@google.com>
18331
18332 * config/arm/t-arm: Update for below.
18333
18334 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
18335 Change type to hash_table. Update dependent calls and types.
18336
18337 * config/i386/t-cygming: Update for below.
18338
18339 * config/i386/t-interix: Update for below.
18340
18341 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
18342 Change type to hash_table. Update dependent calls and types.
18343 (i386_find_on_wrapper_list::wrappers): Likewise.
18344
18345 * config/ia64/t-ia64: Update for below.
18346
18347 * config/ia64/ia64.c (bundle_state_table):
18348 Change type to hash_table. Update dependent calls and types.
18349
18350 * config/mips/mips.c (mips_reorg_process_insns::htab):
18351 Change type to hash_table. Update dependent calls and types.
18352
18353 * config/sol2.c (solaris_comdat_htab):
18354 Change type to hash_table. Update dependent calls and types.
18355
18356 * config/t-sol2: Update for above.
18357
18358 2013-05-29 Teresa Johnson <tejohnson@google.com>
18359
18360 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
18361 functions are not yet marked as defined.
18362
18363 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
18364 Pat Haugen <pthaugen@us.ibm.com>
18365 Peter Bergner <bergner@vnet.ibm.com>
18366
18367 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
18368 instructions.
18369 (VEC_A): Likewise.
18370 (VEC_C): Likewise.
18371 (vrotl<mode>3): Likewise.
18372 (vashl<mode>3): Likewise.
18373 (vlshr<mode>3): Likewise.
18374 (vashr<mode>3): Likewise.
18375
18376 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18377 support for power8 V2DI builtins.
18378
18379 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
18380 power8 V2DI builtins.
18381 (vupkhsw): Likewise.
18382 (vupklsw): Likewise.
18383 (vaddudm): Likewise.
18384 (vminsd): Likewise.
18385 (vmaxsd): Likewise.
18386 (vminud): Likewise.
18387 (vmaxud): Likewise.
18388 (vpkudum): Likewise.
18389 (vpksdss): Likewise.
18390 (vpkudus): Likewise.
18391 (vpksdus): Likewise.
18392 (vrld): Likewise.
18393 (vsld): Likewise.
18394 (vsrd): Likewise.
18395 (vsrad): Likewise.
18396 (vsubudm): Likewise.
18397 (vcmpequd): Likewise.
18398 (vcmpgtsd): Likewise.
18399 (vcmpgtud): Likewise.
18400 (vcmpequd_p): Likewise.
18401 (vcmpgtsd_p): Likewise.
18402 (vcmpgtud_p): Likewise.
18403 (vupkhsw): Likewise.
18404 (vupklsw): Likewise.
18405 (vaddudm): Likewise.
18406 (vmaxsd): Likewise.
18407 (vmaxud): Likewise.
18408 (vminsd): Likewise.
18409 (vminud): Likewise.
18410 (vpksdss): Likewise.
18411 (vpksdus): Likewise.
18412 (vpkudum): Likewise.
18413 (vpkudus): Likewise.
18414 (vrld): Likewise.
18415 (vsld): Likewise.
18416 (vsrad): Likewise.
18417 (vsrd): Likewise.
18418 (vsubudm): Likewise.
18419
18420 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
18421 support for power8 V2DI instructions.
18422
18423 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
18424 power8 V2DI instructions. Combine pack and unpack insns to use an
18425 iterator for each mode. Check whether a particular mode supports
18426 Altivec instructions instead of just checking TARGET_ALTIVEC.
18427 (UNSPEC_VPKUWUM): Likewise.
18428 (UNSPEC_VPKSHSS): Likewise.
18429 (UNSPEC_VPKSWSS): Likewise.
18430 (UNSPEC_VPKUHUS): Likewise.
18431 (UNSPEC_VPKSHUS): Likewise.
18432 (UNSPEC_VPKUWUS): Likewise.
18433 (UNSPEC_VPKSWUS): Likewise.
18434 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
18435 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
18436 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
18437 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
18438 (UNSPEC_VUPKHSB): Likewise.
18439 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
18440 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
18441 (UNSPEC_VUPKHSH): Likewise.
18442 (UNSPEC_VUPKLSB): Likewise.
18443 (UNSPEC_VUPKLSH): Likewise.
18444 (VI2): Likewise.
18445 (VI_char): Likewise.
18446 (VI_scalar): Likewise.
18447 (VI_unit): Likewise.
18448 (VP): Likewise.
18449 (VP_small): Likewise.
18450 (VP_small_lc): Likewise.
18451 (VU_char): Likewise.
18452 (add<mode>3): Likewise.
18453 (altivec_vaddcuw): Likewise.
18454 (altivec_vaddu<VI_char>s): Likewise.
18455 (altivec_vadds<VI_char>s): Likewise.
18456 (sub<mode>3): Likewise.
18457 (altivec_vsubcuw): Likewise.
18458 (altivec_vsubu<VI_char>s): Likewise.
18459 (altivec_vsubs<VI_char>s): Likewise.
18460 (altivec_vavgs<VI_char>): Likewise.
18461 (altivec_vcmpbfp): Likewise.
18462 (altivec_eq<mode>): Likewise.
18463 (altivec_gt<mode>): Likewise.
18464 (altivec_gtu<mode>): Likewise.
18465 (umax<mode>3): Likewise.
18466 (smax<mode>3): Likewise.
18467 (umin<mode>3): Likewise.
18468 (smin<mode>3): Likewise.
18469 (altivec_vpkuhum): Likewise.
18470 (altivec_vpkuwum): Likewise.
18471 (altivec_vpkshss): Likewise.
18472 (altivec_vpkswss): Likewise.
18473 (altivec_vpkuhus): Likewise.
18474 (altivec_vpkshus): Likewise.
18475 (altivec_vpkuwus): Likewise.
18476 (altivec_vpkswus): Likewise.
18477 (altivec_vpks<VI_char>ss): Likewise.
18478 (altivec_vpks<VI_char>us): Likewise.
18479 (altivec_vpku<VI_char>us): Likewise.
18480 (altivec_vpku<VI_char>um): Likewise.
18481 (altivec_vrl<VI_char>): Likewise.
18482 (altivec_vsl<VI_char>): Likewise.
18483 (altivec_vsr<VI_char>): Likewise.
18484 (altivec_vsra<VI_char>): Likewise.
18485 (altivec_vsldoi_<mode>): Likewise.
18486 (altivec_vupkhsb): Likewise.
18487 (altivec_vupkhs<VU_char>): Likewise.
18488 (altivec_vupkls<VU_char>): Likewise.
18489 (altivec_vupkhsh): Likewise.
18490 (altivec_vupklsb): Likewise.
18491 (altivec_vupklsh): Likewise.
18492 (altivec_vcmpequ<VI_char>_p): Likewise.
18493 (altivec_vcmpgts<VI_char>_p): Likewise.
18494 (altivec_vcmpgtu<VI_char>_p): Likewise.
18495 (abs<mode>2): Likewise.
18496 (vec_unpacks_hi_v16qi): Likewise.
18497 (vec_unpacks_hi_v8hi): Likewise.
18498 (vec_unpacks_lo_v16qi): Likewise.
18499 (vec_unpacks_hi_<VP_small_lc>): Likewise.
18500 (vec_unpacks_lo_v8hi): Likewise.
18501 (vec_unpacks_lo_<VP_small_lc>): Likewise.
18502 (vec_pack_trunc_v8h): Likewise.
18503 (vec_pack_trunc_v4si): Likewise.
18504 (vec_pack_trunc_<mode>): Likewise.
18505
18506 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
18507 V2DI builtins.
18508 (vec_vmaxsd): Likewise.
18509 (vec_vmaxud): Likewise.
18510 (vec_vminsd): Likewise.
18511 (vec_vminud): Likewise.
18512 (vec_vpksdss): Likewise.
18513 (vec_vpksdus): Likewise.
18514 (vec_vpkudum): Likewise.
18515 (vec_vpkudus): Likewise.
18516 (vec_vrld): Likewise.
18517 (vec_vsld): Likewise.
18518 (vec_vsrad): Likewise.
18519 (vec_vsrd): Likewise.
18520 (vec_vsubudm): Likewise.
18521 (vec_vupkhsw): Likewise.
18522 (vec_vupklsw): Likewise.
18523
18524 2013-05-29 Jan Hubicka <jh@suse.cz>
18525
18526 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
18527 flags; reorder rest of fields in more consistent way.
18528 (varpool_node): Remove analyzed, finalized and alias.
18529 (cgraph_ndoe): Likewise.
18530 (symtab_alias_ultimate_target): New function.
18531 (cgraph_function_node): Move offline.
18532 (cgraph_reset_node): Declare.
18533 (cgraph_comdat_can_be_unshared_p): Remove.
18534 (varpool_remove_initializer): Declare.
18535 (varpool_first_defined_variable, varpool_next_defined_variable
18536 cgraph_first_defined_function, cgraph_next_defined_function): Update.
18537 (cgraph_function_with_gimple_body_p): Update.
18538 (varpool_all_refs_explicit_p): Update.
18539 (symtab_alias_target): New function.
18540 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
18541 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
18542 (cgraph_function_or_thunk_node): Simplify using
18543 symtab_alias_ultimate_target.
18544 (varpool_variable_node): Likewise.
18545 * cgraph.c (cgraph_create_function_alias): Update.
18546 (cgraph_add_thunk): Update.
18547 (cgraph_remove_node): Update.
18548 (dump_cgraph_node): Do not dump removed flags.
18549 (cgraph_function_body_availability): Update.
18550 (cgraph_propagate_frequency): Update.
18551 (verify_cgraph_node): Check sanity of local flag.
18552 (cgraph_function_node): Move here from cgraph.h; revamp for
18553 cgraph_function_or_thunk_node.
18554 * lto-symtab.c (lto_varpool_replace_node): Update.
18555 (lto_symtab_resolve_can_prevail_p): Update.
18556 (lto_symtab_merge_cgraph_nodes): Update.
18557 * ipa-cp.c (determine_versionability, initialize_node_lattices,
18558 propagate_constants_accross_call, devirtualization_time_bonus,
18559 ipcp_propagate_stage): Update.
18560 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
18561 * ipa-inline-transform.c (clone_inlined_nodes,
18562 preserve_function_body_p): Update.
18563 * ipa-reference.c (propagate): Update.
18564 (write_node_summary_p): Update.
18565 * toplev.c (wrapup_global_declaration_2): Update.
18566 * cgraphunit.c (cgraph_analyze_function): Rename to ...
18567 (analyze_function) ... this one.
18568 (cgraph_process_new_functions): Update.
18569 (cgraph_reset_node): Export.
18570 (cgraph_finalize_function): Update.
18571 (cgraph_add_new_function): Update.
18572 (process_function_and_variable_attributes): Update.
18573 (varpool_finalize_decl): Update.
18574 (symbol_finalized): Remove.
18575 (symbol_finalized_and_needed): Rename to ...
18576 (symbol_defined_and_needed): ... update.
18577 (cgraph_analyze_functions): Update.
18578 (handle_alias_pairs): Update.
18579 (mark_functions_to_output): Update.
18580 (assemble_thunk): Update.
18581 (output_in_order): Update.
18582 (output_weakrefs): Update.
18583 (finalize_compilation_unit): Update.
18584 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
18585 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
18586 input_node, input_varpool_node): Update.
18587 * dbxout.c (dbxout_expand_expr): Update.
18588 * cgraphclones.c (cgraph_clone_node): Update.
18589 (cgraph_copy_node_for_versioning): Update.
18590 (cgraph_materialize_clone): Update.
18591 (cgraph_materialize_all_clones): Update.
18592 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
18593 propagate_pure_const, propagate_nothrow): Update.
18594 * lto-streamer-out.c (lto_output, write_symbol): Update.
18595 * ipa-utils.c (ipa_reverse_postorder): Update.
18596 * ipa-inline.c (can_inline_edge_p): Update.
18597 (update_caller_keys, ipa_inline): Update.
18598 * dwarf2out.c (reference_to_unused,
18599 premark_types_used_by_global_vars_helper): Update.
18600 * tree-eh.c (tree_could_trap_p): Update.
18601 * ipa-split.c (consider_split, execute_split_functions): Update.
18602 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
18603 has_addr_references_p): Update; move ahead in file for better
18604 readability.
18605 (process_references): Simplify.
18606 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
18607 bodies are removed.
18608 (cgraph_comdat_can_be_unshared_p): Make static.
18609 (cgraph_externally_visible_p): Update.
18610 (varpool_externally_visible_p): Update.
18611 (function_and_variable_visibility): Update.
18612 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
18613 ipa_tm_mark_force_output_node): Update.
18614 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
18615 estimate_edge_devirt_benefit, inline_generate_summary,
18616 inline_write_summary): Update.
18617 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
18618 * ipa-prop.c (ipa_compute_jump_functions): Update.
18619 (ipa_print_node_params, ipa_prop_read_section,
18620 ipa_update_after_lto_read, read_replacements_section): Update.
18621 * varasm.c (mark_decl_referenced): Update.
18622 (assemble_alias, dump_tm_clone_pairs): Update.
18623 * tree-inline.c (copy_bb): Update.
18624 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
18625 Update.
18626 * symtab.c (dump_symtab_base): Print new flags.
18627 (verify_symtab_base): Verify new flags.
18628 (symtab_alias_ultimate_target): New function.
18629 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
18630 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
18631 Update.
18632 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
18633 Update.
18634 * i386.c (ix86_get_function_versions_dispatcher,
18635 ix86_generate_version_dispatcher_body): Update.
18636 (fold_builtin_cpu): Use varpool_add_new_variable.
18637 * varpool.c (varpool_remove_initializer): Break out from ...
18638 (varpool_remove_node): ... this one.
18639 (dump_varpool_node, varpool_node_for_asm,
18640 cgraph_variable_initializer_availability, varpool_analyze_node,
18641 varpool_assemble_decl, varpool_remove_unreferenced_decls,
18642 varpool_finalize_named_section_flags, varpool_create_variable_alias):
18643 Update.
18644
18645 2013-05-29 Jan Hubicka <jh@suse.cz>
18646
18647 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
18648
18649 2013-05-29 Easwaran Raman <eraman@google.com>
18650
18651 PR tree-optimization/57442
18652 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
18653 when control exits the main loop.
18654
18655 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
18656
18657 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
18658 and RX600.
18659 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
18660 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
18661 * rx/t-rx: Add rx100 under multi library matches option for nofpu
18662 option.
18663
18664 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18665
18666 PR tree-optimization/57441
18667 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
18668 Don't limit size of incr_vec to number of candidates.
18669
18670 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
18671
18672 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
18673 and mips16 directories.
18674 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
18675 (MULTILIB_DIRNAMES): Ditto.
18676 (MULTILIB_EXCEPTIONS): Add new exceptions.
18677 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
18678 (MULTILIB_DIRNAMES): Ditto.
18679 (MULTILIB_EXCEPTIONS): Add new exceptions.
18680
18681 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18682 Marcus Shawcroft <marcus.shawcroft@arm.com>
18683
18684 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
18685 SYMBOL_TINY_ABSOLUTE.
18686 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
18687 SYMBOL_TINY_ABSOLUTE.
18688 (aarch64_expand_mov_immediate): Likewise.
18689 (aarch64_classify_symbol): Likewise.
18690 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
18691 Permit SYMBOL_TINY_ABSOLUTE.
18692 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
18693
18694 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18695 Marcus Shawcroft <marcus.shawcroft@arm.com>
18696
18697 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
18698 Refactor if/switch. Replace gcc_assert with if.
18699
18700 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
18701
18702 * config/i386/i386.c (initial_ix86_tune_features): Enable
18703 FP Reassociation for AMD bdver1 and bdver2.
18704
18705 2013-05-29 Martin Jambor <mjambor@suse.cz>
18706
18707 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
18708 and IMAGPART_EXPR do not occur within other handled_components.
18709
18710 2013-05-29 Richard Biener <rguenther@suse.de>
18711
18712 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
18713 access on whether the use is in the BB we currently try to
18714 vectorize.
18715 (vect_bb_vectorization_profitable_p): Pass the BB we currently
18716 vectorize to vect_bb_slp_scalar_cost.
18717
18718 2013-05-29 Richard Biener <rguenther@suse.de>
18719
18720 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
18721 computing scalar cost offsetted by stmts that are kept live
18722 by scalar uses.
18723 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
18724 for computation of scalar cost.
18725
18726 2013-05-28 Steve Ellcey <sellcey@mips.com>
18727
18728 * config/mips/mips-cpus.def (mips32r2): Change processor type.
18729
18730 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
18731
18732 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
18733 array notation built-in reduction functions.
18734 * doc/passes.texi (Passes): Added documentation about changes done
18735 for Cilk Plus.
18736 * doc/invoke.texi (C Dialect Options): Added documentation about
18737 the -fcilkplus flag.
18738 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
18739 (BUILTINS_DEF): Depend on cilkplus.def.
18740 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
18741 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
18742 * cilkplus.def: New file.
18743
18744 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
18745
18746 PR rtl-optimization/57439
18747 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
18748
18749 2013-05-28 Easwaran Raman <eraman@google.com>
18750
18751 PR tree-optimization/57337
18752 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
18753 (find_insert_point): Correctly identify the insertion point
18754 when two statements with the same UID is compared.
18755
18756 2013-05-28 Richard Biener <rguenther@suse.de>
18757
18758 PR tree-optimization/56787
18759 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
18760 from the list of data references.
18761 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
18762 clobbers.
18763 (vect_analyze_loop_operations): Likewise.
18764 (vect_transform_loop): Remove clobbers.
18765
18766 2013-05-28 Martin Jambor <mjambor@suse.cz>
18767
18768 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
18769 and REALPART_EXPRs have scalar type.
18770
18771 2013-05-28 Richard Biener <rguenther@suse.de>
18772
18773 PR tree-optimization/57411
18774 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
18775 virtual operands.
18776 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
18777 virtual operand propagation.
18778
18779 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18780
18781 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
18782 destination register for bmasksi_vis.
18783 (vector_init_bshuffle): Likewise.
18784 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
18785
18786 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18787
18788 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
18789 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
18790 mode if the instruction isn't available in the original mode.
18791 * config/sparc/sparc.opt (mfix-ut699): New option.
18792 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
18793 (divdf3): Turn into expander.
18794 (divdf3_nofix): New insn.
18795 (divdf3_fix): Likewise.
18796 (divsf3): Disable if -mfix-ut699.
18797 (sqrtdf2): Turn into expander.
18798 (sqrtdf2_nofix): New insn.
18799 (sqrtdf2_fix): Likewise.
18800 (sqrtsf2): Disable if -mfix-ut699.
18801
18802 2013-05-27 Richard Biener <rguenther@suse.de>
18803
18804 PR middle-end/57412
18805 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
18806 block for the new loop.
18807
18808 2013-05-27 Richard Biener <rguenther@suse.de>
18809
18810 PR tree-optimization/57343
18811 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
18812 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
18813 (number_of_iterations_cond): Do not build the folded tree.
18814
18815 2013-05-27 Richard Biener <rguenther@suse.de>
18816
18817 Revert
18818 PR middle-end/57381
18819 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
18820 OEP_CONSTANT_ADDRESS_OF retained.
18821
18822 PR tree-optimization/57417
18823 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
18824 for unchanged base.
18825 (set_ssa_val_to): Compare addresses using
18826 get_addr_base_and_unit_offset.
18827
18828 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
18829
18830 PR rtl-optimization/56833
18831 * postreload.c (move2add_record_mode): New function.
18832 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
18833 (move2add_use_add2_insn): Use move2add_record_sym_value.
18834 (move2add_use_add3_insn): Likewise.
18835 (reload_cse_move2add): Use move2add_valid_value_p and
18836 move2add_record_mode. Invalidate call-clobbered and REG_INC
18837 affected regs by setting reg_mode to VOIDmode.
18838 (move2add_note_store): Don't pretend the inside of a SUBREG is
18839 the actual destination. Invalidate single/leading registers by
18840 setting reg_mode to VOIDmode.
18841 Use move2add_record_sym_value, move2add_valid_value_p and
18842 move2add_record_mode.
18843
18844 2013-05-27 Richard Biener <rguenther@suse.de>
18845
18846 PR tree-optimization/57396
18847 * tree-affine.c (double_int_constant_multiple_p): Properly
18848 return false for val == 0 and div != 0.
18849
18850 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18851
18852 * config/mips/mips.h: Use #elif in preprocessor conditions.
18853
18854 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18855
18856 PR target/53916
18857 * config/mips/constraints.md (kl): New constraint.
18858 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
18859 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
18860 constraint for operand 0. Split after CSE for MIPS16. Emit a move
18861 from LO for MIPS16.
18862 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
18863
18864 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18865
18866 PR target/55777
18867 * config/mips/mips.c (mips_can_inline_p): New function.
18868 (TARGET_CAN_INLINE_P): Define.
18869
18870 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
18871
18872 * sched-int.h (ds_t, dw_t): Make unsigned int.
18873 Fix documentation that describes how all the ds_t bits are used.
18874 Reserve the last bit for delayed-branch scheduling.
18875 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
18876 (BITS_PER_DEP_WEAK): Fix definition and documentation.
18877 (gen_dep_weak_1): Remove prototype.
18878 * sched-deps.c (get_dep_weak_1): Make static.
18879 * target.def (speculate_insn, needs_block_p, gen_spec_check,
18880 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
18881 * doc/tm.texi: Regenerate.
18882 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
18883
18884 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
18885
18886 PR debug/56950
18887 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
18888
18889 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
18890 Sandra Loosemore <sandra@codesourcery.com>
18891
18892 * config.gcc (powerpc-*): Allow native for with-cpu.
18893
18894 2013-05-24 Jeff Law <law@redhat.com>
18895
18896 PR tree-optimization/57124
18897 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
18898 conversion feeding a condition if the range has an overflow
18899 if -fstrict-overflow. Add warnings for when we do make the
18900 transformation.
18901
18902 2013-05-24 Dehao Chen <dehao@google.com>
18903
18904 * tree-cfg.c (locus_discrim_map): Fix the typo.
18905 (locus_discrim_hasher): Likewise.
18906 (locus_discrim_hasher::hash): Likewise.
18907 (locus_discrim_hasher::equal): Likewise.
18908
18909 2013-05-24 Martin Jambor <mjambor@suse.cz>
18910
18911 PR tree-optimization/57294
18912 * cgraph.h (ipa_record_stmt_references): Declare.
18913 * cgraphbuild.c (ipa_record_stmt_references): New function.
18914 (build_cgraph_edges): Use ipa_record_stmt_references.
18915 (rebuild_cgraph_edges): Likewise.
18916 (cgraph_rebuild_references): Likewise.
18917 * ipa-prop.c (ipa_modify_call_arguments): Discard references
18918 associated with the old statement and build references from the
18919 newly built statements.
18920 * ipa-ref.c (ipa_remove_stmt_references): New function.
18921 * ipa-ref.h (ipa_remove_stmt_references): Declare.
18922
18923 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
18924
18925 * lra-constraints.c (emit_spill_move): Use smaller mode for
18926 mem-mem moves.
18927 (check_and_process_move): Consider mem-reg moves for secondary
18928 too.
18929 (curr_insn_transform): Don't lose insns emitted before for
18930 secondary memory moves.
18931 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
18932 reg set up in the current insn.
18933
18934 2013-05-24 Dehao Chen <dehao@google.com>
18935
18936 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
18937 hash function.
18938 (locus_descrim_hasher::equal): Likewise.
18939 (build_gimple_cfg): New discriminator assignment algorithm.
18940 (make_edges): Likewise.
18941 (next_discriminator_for_locus): Likewise.
18942 (same_line_p): Likewise.
18943 (assign_discriminators): Likewise.
18944 (make_cond_expr_edges): Likewise.
18945 (make_gimple_switch_edges): Likewise.
18946 (make_goto_expr_edges): Likewise.
18947 (make_gimple_asm_edges): Likewise.
18948
18949 2013-05-24 Ian Bolton <ian.bolton@arm.com>
18950
18951 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
18952 X format specifier to only display bottom 16 bits.
18953 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
18954 immediate to match for operand 2, since it will be masked.
18955
18956 2013-05-24 Richard Biener <rguenther@suse.de>
18957
18958 PR tree-optimization/57287
18959 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
18960 all SSA names that occur in abnormal PHIs.
18961
18962 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
18963
18964 PR tree-ssa/57385
18965 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
18966 that index is not negative.
18967
18968 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
18969
18970 PR rtl-optimization/55177
18971 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
18972 (simplify_byte_swapping_operation): New.
18973 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
18974 (simplify_relational_operation_1): Deal with BSWAP.
18975
18976 2013-05-23 Richard Henderson <rth@redhat.com>
18977
18978 PR target/56742
18979 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
18980 (ix86_reorg): Call it.
18981
18982 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
18983
18984 PR target/57379
18985 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
18986 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
18987 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
18988
18989 2013-05-23 Christian Bruel <christian.bruel@st.com>
18990
18991 PR debug/57351
18992 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
18993
18994 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18995 Marcus Shawcroft <marcus.shawcroft@arm.com>
18996
18997 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
18998 * config/aarch64/constraints.md (Usa): Remove.
18999 * doc/md.texi (AArch64 Usa): Remove.
19000
19001 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19002 Marcus Shawcroft <marcus.shawcroft@arm.com>
19003
19004 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
19005 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
19006 * config/aarch64/predicates.md (aarch64_const_address): Remove.
19007 (aarch64_mov_operand): Use aarch64_mov_operand_p.
19008
19009 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
19010
19011 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
19012 instruction (AdvSIMD).
19013 * config/aarch64/aarch64-builtins.c
19014 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
19015 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
19016
19017 2013-05-23 Martin Jambor <mjambor@suse.cz>
19018
19019 PR middle-end/57347
19020 * tree.h (contains_bitfld_component_ref_p): Declare.
19021 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
19022 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
19023 caller.
19024 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
19025 not access a bit-field. Assert all final offsets are byte-aligned.
19026
19027 2013-05-23 Richard Biener <rguenther@suse.de>
19028
19029 PR tree-optimization/57380
19030 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
19031 least one invariant or re-used load.
19032 * passes.c (init_optimization_passes): Move pass_phiprop before
19033 pass_forwprop.
19034
19035 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
19036
19037 * config/aarch64/aarch64-simd.md
19038 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
19039
19040 2013-05-23 Richard Biener <rguenther@suse.de>
19041
19042 PR middle-end/57381
19043 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19044 OEP_CONSTANT_ADDRESS_OF retained.
19045
19046 2013-05-23 Jakub Jelinek <jakub@redhat.com>
19047
19048 PR middle-end/57344
19049 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
19050 don't lower unit. Handle unit not being always BITS_PER_WORD.
19051
19052 2013-05-23 Richard Biener <rguenther@suse.de>
19053
19054 PR rtl-optimization/57341
19055 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
19056 instead of true_dependence.
19057
19058 2013-05-22 David Malcolm <dmalcolm@redhat.com>
19059
19060 * bb-reorder.c (branch_threshold): Make const.
19061 (exec_threshold): Ditto.
19062
19063 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19064 Pat Haugen <pthaugen@us.ibm.com>
19065 Peter Bergner <bergner@vnet.ibm.com>
19066
19067 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
19068 documentation for the power8 crypto builtins.
19069
19070 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
19071
19072 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
19073 macros for defining power8 builtin functions.
19074 (BU_P8V_AV_2): Likewise.
19075 (BU_P8V_AV_P): Likewise.
19076 (BU_P8V_VSX_1): Likewise.
19077 (BU_P8V_OVERLOAD_1): Likewise.
19078 (BU_P8V_OVERLOAD_2): Likewise.
19079 (BU_CRYPTO_1): Likewise.
19080 (BU_CRYPTO_2): Likewise.
19081 (BU_CRYPTO_3): Likewise.
19082 (BU_CRYPTO_OVERLOAD_1): Likewise.
19083 (BU_CRYPTO_OVERLOAD_2): Likewise.
19084 (XSCVSPDP): Fix typo, point to the correct instruction.
19085 (VCIPHER): Add power8 crypto builtins.
19086 (VCIPHERLAST): Likewise.
19087 (VNCIPHER): Likewise.
19088 (VNCIPHERLAST): Likewise.
19089 (VPMSUMB): Likewise.
19090 (VPMSUMH): Likewise.
19091 (VPMSUMW): Likewise.
19092 (VPERMXOR_V2DI): Likewise.
19093 (VPERMXOR_V4SI: Likewise.
19094 (VPERMXOR_V8HI: Likewise.
19095 (VPERMXOR_V16QI: Likewise.
19096 (VSHASIGMAW): Likewise.
19097 (VSHASIGMAD): Likewise.
19098 (VPMSUM): Likewise.
19099 (VPERMXOR): Likewise.
19100 (VSHASIGMA): Likewise.
19101
19102 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
19103 __CRYPTO__ if the crypto instructions are available.
19104 (altivec_overloaded_builtins): Add support for overloaded power8
19105 builtins.
19106
19107 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
19108 support for power8 crypto builtins.
19109 (builtin_function_type): Likewise.
19110 (altivec_init_builtins): Add support for builtins that take vector
19111 long long (V2DI) arguments.
19112
19113 * config/rs6000/crypto.md: New file, define power8 crypto
19114 instructions.
19115
19116 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19117 Pat Haugen <pthaugen@us.ibm.com>
19118 Peter Bergner <bergner@vnet.ibm.com>
19119
19120 * doc/invoke.texi (Option Summary): Add power8 options.
19121 (RS/6000 and PowerPC Options): Likewise.
19122
19123 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
19124 constraints.md instead of rs6000.h. Reorder w* constraints. Add
19125 wm, wn, wr documentation.
19126
19127 * config/rs6000/constraints.md (wm): New constraint for VSX
19128 registers if direct move instructions are enabled.
19129 (wn): New constraint for no registers.
19130 (wq): New constraint for quad word even GPR registers.
19131 (wr): New constraint if 64-bit instructions are enabled.
19132 (wv): New constraint if power8 vector instructions are enabled.
19133 (wQ): New constraint for quad word memory locations.
19134
19135 * config/rs6000/predicates.md (const_0_to_15_operand): New
19136 constraint for 0..15 for crypto instructions.
19137 (gpc_reg_operand): If VSX allow registers in VSX registers as well
19138 as GPR and floating point registers.
19139 (int_reg_operand): New predicate to match only GPR registers.
19140 (base_reg_operand): New predicate to match base registers.
19141 (quad_int_reg_operand): New predicate to match even GPR registers
19142 for quad memory operations.
19143 (vsx_reg_or_cint_operand): New predicate to allow vector logical
19144 operations in both GPR and VSX registers.
19145 (quad_memory_operand): New predicate for quad memory operations.
19146 (reg_or_indexed_operand): New predicate for direct move support.
19147
19148 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
19149 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
19150 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
19151 (POWERPC_MASKS): Add power8 options.
19152 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
19153 various options.
19154
19155 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
19156 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
19157
19158 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
19159 (-mpower8-fusion): New power8 options.
19160 (-mpower8-fusion-sign): Likewise.
19161 (-mpower8-vector): Likewise.
19162 (-mcrypto): Likewise.
19163 (-mdirect-move): Likewise.
19164 (-mquad-memory): Likewise.
19165
19166 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
19167 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
19168 registers.
19169 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
19170 (rs6000_debug_vector_unit): Add p8_vector.
19171 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
19172 definitions. Also print fusion state.
19173 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
19174 (rs6000_builtin_mask_calculate): Add power8 builtin support.
19175 (rs6000_option_override_internal): Add support for power8.
19176 (rs6000_common_init_builtins): Add debugging for skipped builtins
19177 if -mdebug=builtin.
19178 (rs6000_adjust_cost): Add power8 support.
19179 (rs6000_issue_rate): Likewise.
19180 (insn_must_be_first_in_group): Likewise.
19181 (insn_must_be_last_in_group): Likewise.
19182 (force_new_group): Likewise.
19183 (rs6000_register_move_cost): Likewise.
19184 (rs6000_opt_masks): Likewise.
19185
19186 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
19187 power8 capable assembler, default to power7 options.
19188 (TARGET_DIRECT_MOVE): Likewise.
19189 (TARGET_CRYPTO): Likewise.
19190 (TARGET_P8_VECTOR): Likewise.
19191 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
19192 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
19193 (VECTOR_MEM_P8_VECTOR_P): Likewise.
19194 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
19195 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
19196 (TARGET_XSCVDPSPN): Likewise.
19197 (TARGET_XSCVSPDPN): Likewsie.
19198 (TARGET_SYNC_HI_QI): Likewise.
19199 (TARGET_SYNC_TI): Likewise.
19200 (MASK_CRYPTO): Likewise.
19201 (MASK_DIRECT_MOVE): Likewise.
19202 (MASK_P8_FUSION): Likewise.
19203 (MASK_P8_VECTOR): Likewise.
19204 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
19205 temporary used by some of the direct move instructions to get two FP
19206 temporary registers does not force creation of a stack frame.
19207 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
19208 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
19209 that any VSX registers are tieable, even if they are also an
19210 Altivec vector mode.
19211 (r6000_reg_class_enum): Add wm, wr, wv constraints.
19212 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
19213 (RS6000_BTM_CRYPTO): Likewise.
19214 (RS6000_BTM_COMMON): Likewise.
19215
19216 * config/rs6000/rs6000.md (cpu attribute): Add power8.
19217 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
19218 (enum rs6000_vector): Add power8 vector support.
19219
19220 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19221
19222 PR target/19599
19223 PR target/57340
19224 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
19225 (any_sibcall_could_use_r3): this and handle indirect calls.
19226 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
19227
19228 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19229
19230 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
19231
19232 2013-05-22 Richard Biener <rguenther@suse.de>
19233
19234 PR middle-end/57349
19235 * profile.c (branch_prob): Do not split blocks that are
19236 abnormally receiving from ECF_RETURNS_TWICE functions.
19237
19238 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19239
19240 * recog.c (offsettable_address_addr_space_p): Fix calculation of
19241 address mode. Move pointer mode initialization to the same place.
19242
19243 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
19244
19245 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
19246 while it has any effect.
19247
19248 2013-05-21 Easwaran Raman <eraman@google.com>
19249
19250 PR tree-optimization/57322
19251 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
19252 UID of the statement added to the BB to be 1.
19253
19254 2013-05-21 Jakub Jelinek <jakub@redhat.com>
19255
19256 PR tree-optimization/57331
19257 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
19258 of conversion from pointer type to integral type with integer.
19259
19260 2013-05-21 Martin Jambor <mjambor@suse.cz>
19261
19262 PR lto/57289
19263 * ipa-prop.c (ipa_read_node_info): Process param_used and
19264 controlled_uses in the same order as when writing.
19265
19266 2013-05-21 Magnus Granberg <baldrick@free.fr>
19267
19268 PR plugins/56754
19269 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
19270
19271 2013-05-21 Richard Biener <rguenther@suse.de>
19272
19273 PR tree-optimization/57318
19274 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
19275 estimate stmts with side-effects as likely eliminated.
19276
19277 2013-05-21 Richard Biener <rguenther@suse.de>
19278
19279 PR tree-optimization/57330
19280 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
19281 preserve the call stmts fntype.
19282
19283 2013-05-21 Richard Biener <rguenther@suse.de>
19284
19285 PR tree-optimization/57303
19286 * tree-ssa-sink.c (statement_sink_location): Improve killing
19287 stmt detection and properly handle self-assignments.
19288
19289 2013-05-21 Christian Bruel <christian.bruel@st.com>
19290
19291 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
19292 spanning registers. LEAF_REG_REMAP is supported only for contiguous
19293 registers. Set register size out of the PARALLEL loop.
19294
19295 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
19296
19297 PR target/56547
19298 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
19299 (*fmasf4, *fmasf4_media): New insns.
19300
19301 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19302
19303 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
19304 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
19305 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
19306 (mips_idiv_insns): Update the comments to say that the returned
19307 instruction counts are in units of BASE_INSN_LENGTH.
19308 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
19309 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
19310 using 2 rather than 4 as the length of indirect MIPS16 and
19311 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
19312 length of a NOP. Don't divide MIPS16 lengths by 2.
19313 (mips16_split_long_branches): Assume a branch is long if the
19314 length is greater than 4 rather than 8.
19315 * config/mips/mips.md (length): Give MIPS16 lengths directly,
19316 rather than multiplying them by 2. Multiply instruction counts
19317 by BASE_INSN_LENGTH rather than 4.
19318 (*jump_mips16, tls_get_tp_mips16_<mode>)
19319 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
19320
19321 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19322
19323 * config/mips/mips.md (extended_mips16): Remove branch case.
19324 (length): Remove duplicated extended_mips16 test.
19325
19326 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19327
19328 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
19329
19330 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
19331
19332 * recog.h (Recog_data): Rename to...
19333 (recog_data_d): ...this.
19334 (recog_data): Update accordingly.
19335 * recog.c (recog_data): Likewise.
19336 * reload.c (save_recog_data): Likewise.
19337 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
19338 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
19339
19340 2013-05-17 Julian Brown <julian@codesourcery.com>
19341
19342 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
19343 found in a REG_EQUAL note, invalidate it.
19344
19345 2013-05-17 Easwaran Raman <eraman@google.com>
19346
19347 * tree-ssa-reassoc.c (find_insert_point): New function.
19348 (insert_stmt_after): Likewise.
19349 (get_def_stmt): Likewise.
19350 (ensure_ops_are_available): Likewise.
19351 (not_dominated_by): Likewise.
19352 (rewrite_expr_tree): Do not move statements beyond what is
19353 necessary. Remove call to swap_ops_for_binary_stmt...
19354 (reassociate_bb): ... and move it here.
19355 (build_and_add_sum): Assign UIDs for new statements.
19356 (linearize_expr): Likewise.
19357 (do_reassoc): Renumber gimple statement UIDs.
19358
19359 2013-05-17 Jan Hubicka <jh@suse.cz>
19360
19361 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
19362 weakrefs.
19363 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
19364 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
19365 weakrefs.
19366 (output_weakrefs): Update.
19367
19368 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
19369 Martin Jambor <mjambor@suse.cz>
19370
19371 PR middle-end/57276
19372 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
19373 value that corresponds to the given aggval is found in values vector.
19374
19375 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
19376
19377 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
19378 sse, sse2, sse3, ssse3 and sse4a flags to options.
19379
19380 2013-05-17 David Malcolm <dmalcolm@redhat.com>
19381
19382 * gengtype-state.c: (s_expr_writer): New class, to handle
19383 prettifying of output layout of s-expressions.
19384 (state_writer): New class, to write out gtype.state.
19385 (state_written_type_count): Move this variable into member data of
19386 state_writer.
19387 (s_expr_writer::s_expr_writer): New code: constructor for new class
19388 (state_writer::state_writer(): ditto
19389 (s_expr_writer::write_new_line): New function
19390 (s_expr_writer::write_any_indent): ditto
19391 (s_expr_writer::begin_s_expr): ditto
19392 (s_expr_writer::end_s_expr): ditto
19393 (write_state_fileloc): convert to method of state_writer...
19394 (state_writer:: write_state_fileloc): ...and use methods of
19395 s_expr_writer to write indentation into the gtype.state output file
19396 to visually represent the hierarchical structure of the list
19397 structures
19398 (write_state_fields): ditto, renaming to...
19399 (state_writer::write_state_fields)
19400 (write_state_a_string): ditto, renaming to...
19401 (state_writer::write_state_a_string)
19402 (write_state_string_option): ditto, renaming to...
19403 (state_writer::write_state_string_option)
19404 (write_state_type_option): ditto, renaming to...
19405 (state_writer::write_state_type_option)
19406 (write_state_nested_option): ditto, renaming to...
19407 (state_writer::write_state_nested_option)
19408 (write_state_option): ditto, renaming to...
19409 (state_writer::write_state_option)
19410 (write_state_options): ditto, renaming to...
19411 (state_writer::write_state_options)
19412 (write_state_lang_bitmap): ditto, renaming to...
19413 (state_writer::write_state_lang_bitmap)
19414 (write_state_version): ditto, renaming to...
19415 (state_writer::write_state_version)
19416 (write_state_scalar_type): ditto, renaming to...
19417 (state_writer::write_state_scalar_type)
19418 (write_state_string_type): ditto, renaming to...
19419 (state_writer::write_state_string_type)
19420 (write_state_undefined_type): ditto, renaming to...
19421 (state_writer::write_state_undefined_type)
19422 (write_state_struct_union_type): ditto, renaming to...
19423 (state_writer::write_state_struct_union_type)
19424 (write_state_struct_type): ditto, renaming to...
19425 (state_writer::write_state_struct_type)
19426 (write_state_user_struct_type): ditto, renaming to...
19427 (state_writer::write_state_user_struct_type)
19428 (write_state_lang_struct_type): ditto, renaming to...
19429 (state_writer::write_state_lang_struct_type)
19430 (write_state_param_struct_type): ditto, renaming to...
19431 (state_writer::write_state_param_struct_type)
19432 (write_state_pointer_type): ditto, renaming to...
19433 (state_writer::write_state_pointer_type)
19434 (write_state_array_type): ditto, renaming to...
19435 (state_writer::write_state_array_type)
19436 (write_state_gc_used): ditto, renaming to...
19437 (state_writer::write_state_gc_used)
19438 (write_state_common_type_content): ditto, renaming to...
19439 (state_writer::write_state_common_type_content)
19440 (write_state_type): ditto, renaming to...
19441 (state_writer::write_state_type)
19442 (write_state_pair_list): ditto, renaming to...
19443 (state_writer::write_state_pair_list)
19444 (write_state_pair): ditto, renaming to...
19445 (state_writer::write_state_pair)
19446 (write_state_typedefs): ditto, renaming to...
19447 (state_writer::write_state_typedefs)
19448 (write_state_structures): ditto, renaming to...
19449 (state_writer::write_state_structures)
19450 (write_state_param_structs): ditto, renaming to...
19451 (state_writer::write_state_param_structs)
19452 (write_state_variables): ditto, renaming to...
19453 (state_writer::write_state_variables)
19454 (write_state_srcdir): ditto, renaming to...
19455 (state_writer::write_state_srcdir)
19456 (write_state_files_list): ditto, renaming to...
19457 (state_writer::write_state_files_list)
19458 (write_state_languages): ditto, renaming to...
19459 (state_writer::write_state_languages)
19460 (write_state): create a state_writer instance and use it when
19461 writing out the state file
19462
19463 2013-05-17 Mike Stump <mikestump@comcast.net>
19464
19465 PR rtl-optimization/57304
19466 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
19467 accessing DF_REF_REAL_LOC.
19468
19469 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19470
19471 PR rtl-optimization/57281
19472 PR rtl-optimization/57300
19473 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
19474 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
19475 what the other splitter did if the registers are dead.
19476
19477 2013-05-17 Richard Biener <rguenther@suse.de>
19478
19479 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
19480 MEM_REF offsets.
19481
19482 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19483
19484 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
19485 linking.
19486
19487 2013-05-17 Marek Polacek <polacek@redhat.com>
19488
19489 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
19490 length when doing non-zero store of storing '\0' to '\0'.
19491
19492 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19493
19494 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
19495 vect_external_def oprnd1 with loop_vinfo, try to emit
19496 optional cast, negation and and stmts on the loop preheader
19497 edge instead of into the pattern def seq.
19498
19499 PR tree-optimization/57051
19500 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
19501 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
19502
19503 2013-05-16 Nick Clifton <nickc@redhat.com>
19504
19505 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
19506 (rl78_is_naked_func): New function.
19507 (rl78_expand_prologue): Skip prologue generation for naked functions.
19508 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
19509 * doc/extend.texi (naked): Add RL78 to the list of processors
19510 that supports this attribute.
19511
19512 2013-05-16 Jeff Law <law@redhat.com>
19513
19514 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
19515
19516 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19517
19518 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
19519 cache parameters using detect_caches_amd also for CYRIX,
19520 NSC and TM2 signatures.
19521
19522 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19523 Dzianis Kahanovich <mahatma@eu.by>
19524
19525 PR target/45359
19526 PR target/46396
19527 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
19528 VIA/Centaur processors and determine their cache parameters
19529 using detect_caches_amd.
19530
19531 2013-05-16 Teresa Johnson <tejohnson@google.com>
19532
19533 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
19534 (rtl_verify_edges): New function.
19535 (rtl_verify_bb_insns): Ditto.
19536 (rtl_verify_bb_pointers): Ditto.
19537 (rtl_verify_bb_insn_chain): Ditto.
19538 (rtl_verify_fallthru): Ditto.
19539 (rtl_verify_bb_layout): Ditto.
19540 (rtl_verify_flow_info_1): Outline checks into new functions.
19541 (rtl_verify_flow_info): Ditto.
19542
19543 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
19544
19545 * cfghooks.c (copy_bbs): Add update_dominance argument.
19546 * cfghooks.h (copy_bbs): Update prototype.
19547 * tree-cfg.c (gimple_duplicate_sese_region):
19548 Add update_dominance argument.
19549 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
19550 * tree-ssa-loop-ch.c (copy_loop_headers): Update
19551 gimple_duplicate_sese_region call.
19552 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
19553 Update copy_bbs call.
19554 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
19555 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
19556
19557 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19558
19559 * tree-vectorizer.h (NUM_PATTERNS): Increment.
19560 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
19561 vect_recog_rotate_pattern.
19562 (vect_recog_rotate_pattern): New function.
19563
19564 2013-05-16 Jason Merrill <jason@redhat.com>
19565
19566 * Makefile.in (LLINKER): New variable.
19567 (mostlyclean): Remove link mutex.
19568 * configure.ac: Handle --enable-link-mutex.
19569 * lock-and-run.sh: New script.
19570
19571 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19572
19573 PR target/19599
19574 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
19575 for NULL decl.
19576
19577 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19578
19579 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
19580
19581 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
19582
19583 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
19584 * config/arm/arm.c (next_consecutive_mem): New function.
19585 (gen_movmem_ldrd_strd): Likewise.
19586 * config/arm/arm.md (movmemqi): Update condition and code.
19587 (unaligned_loaddi, unaligned_storedi): New patterns.
19588
19589 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19590
19591 * config.gcc: Obsolete *-*-solaris2.9*.
19592 * doc/install.texi (Specific, *-*-solaris2*): Document it.
19593
19594 2013-05-16 Richard Biener <rguenther@suse.de>
19595
19596 * passes.c (init_optimization_passes): Move pass_parallelize_loops
19597 earlier, after GRAPHITE transforms and IV canonicalization.
19598
19599 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19600
19601 * omp-low.c (extract_omp_for_data): For collapsed loops,
19602 if at least one of the loops is known at compile time to
19603 iterate zero times, set count to 0.
19604 (expand_omp_regimplify_p): New function.
19605 (expand_omp_for_generic): For collapsed loops, if at least
19606 one of the loops isn't known to iterate at least once,
19607 add runtime check with setting count to 0.
19608 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
19609 For unsigned types if it isn't known at compile time that
19610 the loop will iterate at least once, add runtime check to bypass
19611 the whole loop if initial condition isn't true.
19612
19613 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
19614
19615 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
19616
19617 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
19618
19619 PR middle-end/57286
19620 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
19621 transformations to avoid an infinite loop.
19622
19623 2013-05-16 Marek Polacek <polacek@redhat.com>
19624
19625 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
19626
19627 2013-05-15 Leif Ekblad <leif@rdos.net>
19628
19629 * config/i386/i386.c (ix86_decompose_address): Use
19630 DEFAULT_TLS_SEG_REG to access TLS segment register.
19631 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
19632 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
19633 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
19634
19635 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
19636
19637 PR target/57260
19638 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
19639 sibling calls to functions that would normally be lazily bound,
19640 unless $gp is call-clobbered.
19641
19642 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
19643
19644 * config/i386/i386.c (ix86_option_override_internal): Update
19645 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
19646 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
19647 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
19648 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
19649 of TARGET_3DNOW.
19650 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
19651
19652 2013-05-15 Andreas Schwab <schwab@suse.de>
19653
19654 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
19655 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
19656 third operand.
19657
19658 2013-05-15 Teresa Johnson <tejohnson@google.com>
19659
19660 * loop-unroll.c (report_unroll_peel): Check decision before
19661 emitting unroll/peel message.
19662
19663 2013-05-15 Teresa Johnson <tejohnson@google.com>
19664
19665 * function.h (has_bb_partition): New rtl_data flag.
19666 (bb_reorder_complete): Ditto.
19667 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
19668 instead of flag_reorder_blocks_and_partition.
19669 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
19670 with some enhancements.
19671 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
19672 * bb-reorder.c (connect_traces): Check for has_bb_partition
19673 instead of flag_reorder_blocks_and_partition.
19674 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
19675 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
19676 verify_hot_cold_block_grouping.
19677 (partition_hot_cold_basic_blocks): Set has_bb_partition.
19678
19679 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19680
19681 PR target/19599
19682 * config/arm/predicates.md (call_insn_operand): New predicate.
19683 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
19684 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
19685 if insn is not a tail call.
19686 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
19687 registers.
19688 * config/arm/arm.h (enum reg_class): New caller save register class.
19689 (REG_CLASS_NAMES): Likewise.
19690 (REG_CLASS_CONTENTS): Likewise.
19691 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
19692 without decls.
19693
19694 2013-05-15 Richard Biener <rguenther@suse.de>
19695
19696 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
19697 of MSG_OPTIMIZED_LOCATIONS.
19698 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
19699 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
19700 message.
19701 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
19702 of MSG_OPTIMIZED_LOCATIONS.
19703 (execute_vect_slp): Likewise.
19704 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
19705 (vect_create_cond_for_alias_checks): Likewise.
19706 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
19707 (vect_recog_widen_mult_pattern): Likewise.
19708 (vect_recog_widen_sum_pattern): Likewise.
19709 (vect_recog_over_widening_pattern): Likewise.
19710 (vect_recog_widen_shift_pattern): Likewise.
19711 (vect_recog_vector_vector_shift_pattern): Likewise.
19712 (vect_recog_divmod_pattern): Likewise.
19713 (vect_recog_mixed_size_cond_pattern): Likewise.
19714 (vect_recog_bool_pattern): Likewise.
19715 (vect_pattern_recog_1): Likewise.
19716
19717 2013-05-15 Martin Jambor <mjambor@suse.cz>
19718
19719 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
19720 non-functions to builtin_unreachable.
19721 * ipa-inline-transform.c (inline_call): Do not assert estimates were
19722 correct when new direct edges were discovered.
19723
19724 2013-05-15 Martin Jambor <mjambor@suse.cz>
19725
19726 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
19727 header, print symbol order instead of node uid, print more information
19728 about indirect edge targets.
19729 (ipa_make_edge_direct_to_target): Print symbol order instead of node
19730 uids.
19731 (ipa_make_edge_direct_to_target): Likewise.
19732 (remove_described_reference): Likewise.
19733 (propagate_controlled_uses): Likewise.
19734 (ipa_print_node_params): Also print symbol order.
19735 (ipcp_transform_function): Print symbol order instead of node uids.
19736 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
19737 (cgraph_get_create_real_symbol_node): Likewise.
19738 * ipa-cp.c (print_lattice): Likewise.
19739 (print_all_lattices): Likewise.
19740 (determine_versionability): Likewise.
19741 (initialize_node_lattices): Likewise.
19742 (estimate_local_effects): Likewise.
19743 (update_profiling_info): Likewise.
19744 (create_specialized_node): Likewise.
19745 (perhaps_add_new_callers): Likewise.
19746 (decide_about_value): Likewise.
19747 (decide_whether_version_node): Likewise.
19748 (identify_dead_nodes): Likewise.
19749 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
19750 (dump_inline_summary): Likewise.
19751 (estimate_node_size_and_time): Likewise.
19752 (inline_analyze_function): Likewise.
19753 * ipa-inline.c (report_inline_failed_reason): Likewise.
19754 (want_early_inline_function_p): Likewise.
19755 (edge_badness): Likewise.
19756 (update_edge_key): Likewise.
19757 (inline_small_functions): Likewise. Add dumping of order to two other
19758 dumps.
19759 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
19760 instead of node uids.
19761 (propagate_pure_const): Likewise.
19762 (propagate_pure_const): Likewise.
19763 * ipa-utils.c (dump_cgraph_node_set): Likewise.
19764 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
19765 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
19766 of node uids.
19767 * tree-pretty-print.c (dump_function_header): Likewise.
19768 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
19769 Print symbol order instead of node uids.
19770
19771 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19772
19773 * config/s390/s390.c (s390_register_move_cost): Don't impose the
19774 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
19775
19776 2013-05-15 Richard Biener <rguenther@suse.de>
19777
19778 PR tree-optimization/57275
19779 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
19780 return value for fail to do runtime alias checks for gather loads.
19781
19782 2013-05-15 Jan Hubicka <jh@suse.cz>
19783
19784 PR lto/57038
19785 PR lto/47375
19786 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
19787 weakrefs are not external.
19788 (lto_symtab_merge_decls): Fix thinko when dealing with
19789 non-lto_symtab decls.
19790 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
19791 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
19792 * varpool.c (dump_varpool_node): Dump more flags.
19793
19794 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19795
19796 * config/i386/i386.c (processor_alias_table): Add instruction
19797 FSGSBASE for AMD bdver3 architecture.
19798
19799 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19800
19801 * tree.c (warn_deprecated_use): Print file:line using locus color.
19802 * diagnostic.c (diagnostic_report_current_module): Print file:line
19803 and file:line:column using locus color.
19804
19805 2013-05-14 Mike Stump <mikestump@comcast.net>
19806
19807 * gdbinit.in: Add __null.
19808
19809 2013-05-14 Mike Stump <mikestump@comcast.net>
19810
19811 * recog.h: Rename struct recog_data to Recog_data.
19812 * recog.c: Likewise.
19813 * reload.c (can_reload_into): Likewise.
19814 * config/picochip/picochip.c: Likewise.
19815
19816 2013-05-14 Mike Stump <mikestump@comcast.net>
19817
19818 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
19819
19820 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
19821
19822 * resource.h (struct resources): Remove unch_memory member.
19823 (CLEAR_RESOURCE): Don't clear unch_memory.
19824 * resource.c (mark_referenced_resources): Don't set it.
19825 (mark_set_resources): Likewise.
19826 (mark_target_live_regs): Don't clear it.
19827 (init_resource_info): Likewise.
19828 * reorg.c (resource_conflicts_p): Don't compare it.
19829 (redundant_insn): Don't set it.
19830
19831 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
19832 Remove prototypes.
19833 * emit-rtl.c (next_label): Remove unused function.
19834 (skip_consecutive_labels, link_cc0_insns): Move to ...
19835 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
19836 only place where these functions are used, and make them static.
19837
19838 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
19839
19840 * fold-const.c (fold_negate_expr): Handle vectors.
19841 (fold_truth_not_expr): Make it static.
19842 (fold_invert_truthvalue): New static function.
19843 (invert_truthvalue_loc): Handle vectors. Do not call
19844 fold_truth_not_expr directly.
19845 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
19846 <TRUTH_NOT_EXPR>: Do not cast to boolean.
19847 (fold_comparison): Handle vector constants.
19848 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
19849 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
19850 * tree.h (fold_truth_not_expr): Remove declaration.
19851
19852 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
19853
19854 * config/aarch64/aarch64-simd.md
19855 (aarch64_vcond_internal<mode>): Rename to...
19856 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
19857 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
19858 float modes. Clarify all iterator modes.
19859 (vcond<mode><mode>): Use new name for vcond expanders.
19860 (vcond<v_cmp_result><mode>): Likewise.
19861 (vcondu<mode><mode>: Likewise.
19862 * config/aarch64/iterators.md (VDQF_COND): New.
19863
19864 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
19865
19866 PR bootstrap/57266
19867 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
19868 variable for the shift amount. Check that we shift by non-negative
19869 amounts.
19870
19871 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
19872
19873 PR target/42017
19874 * config/arm/arm.h (EPILOGUE_USES): Only return true
19875 for LR_REGNUM after epilogue_completed.
19876
19877 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
19878
19879 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
19880 is error_mark_node.
19881
19882 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19883
19884 PR target/57261
19885 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
19886 and Solaris 11+/x86 with gld.
19887 * configure: Regenerate.
19888
19889 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19890
19891 * expmed.c (expand_shift_1): Canonicalize rotates by
19892 constant bitsize / 2 to bitsize - 1.
19893 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
19894 case ROTATERT>: Likewise.
19895
19896 Revert:
19897 2013-05-10 Jakub Jelinek <jakub@redhat.com>
19898
19899 * config/i386/i386.md (rotateinv): New code attr.
19900 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
19901 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
19902 roll $31, %eax, etc.
19903
19904 2013-05-14 Richard Biener <rguenther@suse.de>
19905
19906 PR middle-end/57235
19907 * tree-eh.c (sink_clobbers): Give up for successors with
19908 multiple predecessors and no virtual uses.
19909
19910 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
19911
19912 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
19913 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
19914
19915 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19916
19917 PR middle-end/57251
19918 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
19919 the case when both op0 and op1 have VOIDmode.
19920
19921 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
19922
19923 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
19924 in multiply-accumulate mode.
19925
19926 2013-05-13 Guozhi Wei <carrot@google.com>
19927
19928 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
19929
19930 2013-05-13 Kai Tietz <ktietz@redhat.com>
19931
19932 PR target/56975
19933 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
19934 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
19935 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
19936 * config/i386/i386.c (ix86_option_override_internal): Likewise.
19937 (ix86_expand_prologue): Likewise.
19938 (ix86_expand_split_stack_prologue): Likewise.
19939 (legitimate_pic_address_disp_p): Likewise.
19940 (legitimize_pic_address): Likewise.
19941 (legitimize_tls_address): Likewise.
19942 (legitimize_pe_coff_symbol): Likewise.
19943 (output_pic_addr_const): Likewise.
19944 (construct_plt_address): Likewise.
19945 (ix86_expand_call): Likewise.
19946 (x86_output_mi_thunk): Likewise.
19947 (x86_function_profiler): Likewise.
19948
19949 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
19950
19951 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
19952 similar switch cases.
19953 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
19954 (aarch64_simd_mov_to_<mode>low): Delete.
19955 (aarch64_simd_mov_to_<mode>high): Delete.
19956 (move_lo_quad_<mode>): Add w<-r alternative.
19957 (aarch64_simd_move_hi_quad_<mode>): Likewise.
19958 (aarch64_simd_mov_from_*): Update type attribute.
19959 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
19960 statement.
19961
19962 2013-05-13 Jan Hubicka <jh@suse.cz>
19963
19964 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
19965 * config/i386/i386.c (ix86_compute_frame_layout,
19966 ix86_expand_epilogue, emit_i387_cw_initialization,
19967 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
19968 ix86_local_alignment): Fix use of size/speed predicates.
19969
19970 2013-05-13 Jakub Jelinek <jakub@redhat.com>
19971
19972 PR tree-optimization/45216
19973 PR tree-optimization/57157
19974 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
19975 the (-Y) & (B - 1) variant if OP is |.
19976 * expmed.c (expand_shift_1): For rotations by const0_rtx just
19977 return shifted. Use (-op1) & (prec - 1) as other_amount
19978 instead of prec - op1.
19979
19980 2013-05-13 Martin Jambor <mjambor@suse.cz>
19981
19982 PR middle-end/42371
19983 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
19984 (ipa_constant_data): New type.
19985 (ipa_jump_func): Use ipa_constant_data to hold information about
19986 constant jump functions.
19987 (ipa_get_jf_constant): Adjust to jump function type changes.
19988 (ipa_get_jf_constant_rdesc): New function.
19989 (ipa_param_descriptor): New field controlled_uses.
19990 (ipa_get_controlled_uses): New function.
19991 (ipa_set_controlled_uses): Likewise.
19992 * ipa-ref.h (ipa_find_reference): Declare.
19993 * ipa-prop.c (ipa_cst_ref_desc): New type.
19994 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
19995 changes.
19996 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
19997 New parameter cs. Adjust all callers.
19998 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
19999 (remove_described_reference): New function.
20000 (jfunc_rdesc_usable): Likewise.
20001 (try_make_edge_direct_simple_call): Decrement controlled use count,
20002 attempt to remove reference if it hits zero.
20003 (combine_controlled_uses_counters): New function.
20004 (propagate_controlled_uses): Likewise.
20005 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
20006 (ipa_edge_duplication_hook): Duplicate reference descriptions.
20007 (ipa_print_node_params): Print described use counter.
20008 (ipa_write_jump_function): Adjust to jump function type changes.
20009 (ipa_read_jump_function): New parameter CS, pass it to
20010 ipa_set_jf_constant. Adjust caller.
20011 (ipa_write_node_info): Stream controlled use count
20012 (ipa_read_node_info): Likewise.
20013 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
20014 asserting.
20015 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
20016 count. Remove cloning-added reference if it reaches zero.
20017 * ipa-ref.c (ipa_find_reference): New function.
20018
20019 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20020
20021 * config/i386/i386.c (processor_target_table): Modified default
20022 alignment values for AMD BD and BT architectures.
20023
20024 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
20025
20026 * tree-vect-generic.c (uniform_vector_p): Move ...
20027 * tree.c (uniform_vector_p): ... here.
20028 * tree.h (uniform_vector_p): Declare it.
20029 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
20030 into a scalar.
20031
20032 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20033
20034 PR tree-optimization/57230
20035 * tree-ssa-strlen.c (handle_char_store): Record length for
20036 array store from STRING_CST.
20037
20038 PR tree-optimization/57230
20039 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
20040 check.
20041
20042 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
20043
20044 * config/epiphany/epiphany.c (epiphany_init): Check size of
20045 NUM_MODES_FOR_MODE_SWITCHING.
20046 (epiphany_expand_prologue):
20047 Remove CONFIG_REGNUM initial value handling code.
20048 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
20049 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
20050 (emit_set_fp_mode, epiphany_mode_after): Likewise.
20051 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
20052 Don't return 1 for FP_MODE_NONE.
20053 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
20054 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
20055 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
20056 * config/epiphany/epiphany.md (save_config): New pattern.
20057
20058 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
20059
20060 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
20061
20062 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
20063
20064 * config/i386/i386.md (memory): Handle sseishft1.
20065 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
20066 (*vec_extractv2di_1): Ditto.
20067
20068 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
20069
20070 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
20071 saved registers.
20072
20073 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
20074
20075 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
20076 Add mthumb/march=armv7-a multilib.
20077 Add mthumb/march=armv7-r multilib.
20078 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
20079
20080 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
20081
20082 * config/v850/t-rtems: Add more multilibs.
20083
20084 2013-05-10 Richard Biener <rguenther@suse.de>
20085
20086 PR tree-optimization/57214
20087 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
20088 not propagate from SSA names that occur in abnormal PHI nodes.
20089
20090 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
20091
20092 * stor-layout.c (element_precision): New function.
20093 * machmode.h (element_precision): Declare it.
20094 * tree.c (build_minus_one_cst): New function.
20095 (element_precision): Likewise.
20096 * tree.h (build_minus_one_cst): Declare new function.
20097 (element_precision): Likewise.
20098 * fold-const.c (operand_equal_p): Use element_precision.
20099 (fold_binary_loc): Handle vector types.
20100 * convert.c (convert_to_integer): Use element_precision.
20101 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
20102 separately.
20103
20104 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
20105
20106 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
20107 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
20108 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
20109 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20110 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
20111 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
20112 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
20113 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20114 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
20115 (Uuw8): New constraints.
20116 (Usb4): Move into alphabetical order.
20117 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
20118 (sd8_operand, ub8_operand, uw8_operand): New predicates.
20119 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
20120 previously unnamed patterns.
20121 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
20122 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
20123 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
20124 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
20125 of set_attr_alternative/if_then_else. Use extended_mips16 instead
20126 of specific lengths.
20127
20128 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20129
20130 * config/i386/i386.md (rotateinv): New code attr.
20131 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20132 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20133 roll $31, %eax, etc.
20134
20135 PR tree-optimization/45216
20136 PR tree-optimization/57157
20137 * tree-ssa-forwprop.c (simplify_rotate): New function.
20138 (ssa_forward_propagate_and_combine): Call it.
20139
20140 2013-05-10 Richard Biener <rguenther@suse.de>
20141
20142 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
20143 disable peeling when we version for aliasing.
20144 (vector_alignment_reachable_p): Honor explicit user alignment.
20145 (vect_supportable_dr_alignment): Likewise.
20146 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
20147 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
20148 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
20149 then peeling to arrange for the cost-model check to come first.
20150
20151 2013-05-10 Alan Modra <amodra@gmail.com>
20152
20153 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
20154 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
20155 * configure: Regenerate.
20156
20157 2013-05-10 Alan Modra <amodra@gmail.com>
20158
20159 PR target/55033
20160 * varasm.c (default_elf_select_section): Move !DECL_P check..
20161 (get_named_section): ..to here before calling get_section_name.
20162 Adjust assertion.
20163 (default_section_type_flags): Add DECL_P check.
20164 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
20165 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
20166
20167 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
20168
20169 * config/epiphany/epiphany.c (epiphany_expand_prologue):
20170 When using gen_stack_adjust_str with a register offset, add a
20171 REG_FRAME_RELATED_EXPR note.
20172
20173 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
20174
20175 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
20176 (*vec_extractv4si_zext_mem): Ditto.
20177 (*vec_extractv2di): Add 0->x and x->x alternatives.
20178 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
20179 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
20180
20181 2013-05-09 Jason Merrill <jason@redhat.com>
20182
20183 N3639 C++1y VLA support
20184 * gimplify.c (gimplify_vla_decl): Don't touch an existing
20185 DECL_VALUE_EXPR.
20186
20187 * tree.c (build_constructor_va): New.
20188 * tree.h: Declare it.
20189
20190 2013-05-09 Martin Jambor <mjambor@suse.cz>
20191
20192 PR lto/57084
20193 * gimple-fold.c (canonicalize_constructor_val): Call
20194 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
20195
20196 2013-05-09 Jan Hubicka <jh@suse.cz>
20197 Richard Biener <rguenther@suse.de>
20198
20199 PR lto/54095
20200 * symtab.c (symtab_make_decl_local): Do not add private names.
20201
20202 2013-05-09 Jan Hubicka <jh@suse.cz>
20203
20204 PR lto/54095
20205 * symtab.c (insert_to_assembler_name_hash): Handle clones.
20206 (unlink_from_assembler_name_hash): Likewise.
20207 (symtab_prevail_in_asm_name_hash, symtab_register_node,
20208 symtab_unregister_node, symtab_initialize_asm_name_hash,
20209 change_decl_assembler_name): Update.
20210
20211 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
20212
20213 * config/aarch64/aarch64.md: New movtf split.
20214 (*movtf_aarch64): Update.
20215 (aarch64_movdi_tilow): Handle TF modes and rename to
20216 aarch64_movdi_<mode>low.
20217 (aarch64_movdi_tihigh): Handle TF modes and rename to
20218 aarch64_movdi_<mode>high
20219 (aarch64_movtihigh_di): Handle TF modes and rename to
20220 aarch64_mov<mode>high_di
20221 (aarch64_movtilow_di): Handle TF modes and rename to
20222 aarch64_mov<mode>low_di
20223 (aarch64_movtilow_tilow): Remove spurious whitespace.
20224 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
20225 splits.
20226 (aarch64_print_operand): Update.
20227
20228 2013-05-09 Alan Modra <amodra@gmail.com>
20229
20230 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
20231 powerpc64le.
20232 * configure: Regenerate.
20233
20234 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20235
20236 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
20237 splitter preparation statements.
20238 * config/i386/sse.md (*vec_extract* splitters): Ditto.
20239 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
20240 adjust_address_nv.
20241
20242 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20243
20244 * gimple-ssa-strength-reduction.c (count_candidates): Change
20245 return value to int.
20246 (analyze_candidates_and_replace): Change type of length to int.
20247
20248 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20249
20250 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
20251 (*vec_extract<mode>): Use VI12_128 mode iterator.
20252 (*vec_extract<mode>_mem): Ditto.
20253 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
20254 attribute.
20255
20256 2013-05-08 Diego Novillo <dnovillo@google.com>
20257
20258 PR bootstrap/54659
20259
20260 Revert:
20261 2012-08-17 Diego Novillo <dnovillo@google.com>
20262
20263 PR bootstrap/54281
20264 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
20265 * config.in: Regenerate.
20266 * configure: Regenerate.
20267 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
20268
20269 2013-05-08 Jan Hubicka <jh@suse.cz>
20270
20271 PR lto/54095
20272 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
20273 * cgraph.h (symtab_node_base): Add unique_name.
20274 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20275 input_overwrite_node, input_varpool_node): Stream unique_name.
20276 * cgraphclones.c (cgraph_create_virtual_clone,
20277 cgraph_function_versioning): Set unique_name.
20278 * ipa.c (function_and_variable_visibility): Set unique_name.
20279
20280 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20281
20282 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
20283 (alloc_cand_and_find_basis): Restrict conditional candidate
20284 processing to CAND_MULTs.
20285
20286 2013-05-08 Jan Hubicka <jh@suse.cz>
20287
20288 PR lto/54095
20289 lto-symtab.c (lto_symtab_symbol_p): New function.
20290 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
20291 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
20292 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
20293 Skip static symbols.
20294
20295 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
20296
20297 PR tree-optimization/57200
20298 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
20299 Only call inform if the preceding warning_at returns true.
20300
20301 2013-05-07 Han Shen <shenhan@google.com>
20302
20303 * cfgexpand.c (record_or_union_type_has_array_p): New function.
20304 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
20305 * common.opt (fstack-protector-strong): New option.
20306 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
20307 * doc/invoke.texi (Optimization Options): Document
20308 "-fstack-protector-strong".
20309 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
20310
20311 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20312
20313 * config/mips/mips.c (mips_machine_reorg2): Return 0.
20314
20315 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
20316
20317 * ira.c (update_equiv_regs): Add insn having equiv memory even if
20318 it is not lhs of the insn.
20319 (setup_reg_equiv): Remove insn having equiv memory which it is not
20320 lhs of the insn.
20321 * lra-constraints.c (process_address): Try to improve generation
20322 code for address base + disp.
20323 (lra_constraints): Make correct the code for checking insn setting
20324 up backward equivalence. Remove insn only if it is in the init
20325 insn list.
20326 * lra-eliminations.c (update_reg_eliminate): Change return value.
20327 (lra_eliminate): Use the result.
20328
20329 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
20330
20331 * config/i386/sse.md (ssescalarnummask): New mode attribute.
20332 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
20333 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
20334 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
20335 register target operands.
20336 (*vec_extractv8hi_sse2): New pattern.
20337 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
20338 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
20339 (*vec_extract<mode>_mem): New insn and split pattern.
20340
20341 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
20342
20343 * config/arm/arm.c (arm_asan_shadow_offset): New function.
20344 (TARGET_ASAN_SHADOW_OFFSET): Define.
20345 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
20346 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
20347
20348 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20349
20350 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
20351 (incr_vec_index): Return -1 if increment not found.
20352 (create_add_on_incoming_edge): Assert if increment not found.
20353 (record_increment): Limit number of increments recorded.
20354 (all_phi_incrs_profitable): Return false if an increment not found.
20355 (replace_profitable_candidates): Don't process increments that were
20356 not recorded.
20357 (analyze_candidates_and_replace): Limit size of incr_vec.
20358
20359 2013-05-07 Richard Biener <rguenther@suse.de>
20360
20361 * calls.c (special_function_p): setjmp-like functions are leaf.
20362 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
20363 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
20364
20365 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
20366
20367 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
20368 (aarch64_simd_mov<mode>): New expander.
20369 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
20370 (aarch64_simd_mov_to_<mode>high): Likewise.
20371 (aarch64_simd_mov_from_<mode>low): Likewise.
20372 (aarch64_simd_mov_from_<mode>high): Likewise.
20373 (aarch64_dup_lane<mode>): Update.
20374 (aarch64_dup_lanedi): New instruction pattern.
20375 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
20376 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
20377
20378 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20379
20380 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
20381 (replace_mult_candidate): Remove unnecessary argument; remove
20382 unnecessary parameter from call to introduce_cast_before_cand.
20383 (replace_unconditional_candidate): Remove unnecessary parameter
20384 from call to replace_mult_candidate.
20385 (replace_conditional_candidate): Likewise.
20386 (insert_initializers): Use make_temp_ssa_name.
20387 (introduce_cast_before_cand): Remove unnecessary argument; use
20388 make_temp_ssa_name.
20389 (replace_one_candidate): Remove unnecessary argument; remove
20390 unnecessary parameter from calls to introduce_cast_before_cand.
20391 (replace_profitable_candidates): Remove unnecessary parameters
20392 from calls to replace_one_candidate.
20393
20394 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20395
20396 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
20397 phi def as possibly hiding a basis for a CAND_ADD whose operands
20398 have been commuted in the analysis.
20399 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
20400
20401 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20402
20403 * config/aarch64/aarch64.md
20404 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
20405 shift value between 0-4.
20406
20407 2013-05-07 Richard Biener <rguenther@suse.de>
20408
20409 * double-int.h (rshift): New overload.
20410 * double-int.c (rshift): New function.
20411 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
20412 (create_reference_ops_from_ref): Remove.
20413 (vn_reference_insert): Use shared ops for constructing the
20414 reference and copy it.
20415
20416 2013-05-07 Richard Biener <rguenther@suse.de>
20417
20418 PR middle-end/57190
20419 * tree-eh.c (sink_clobbers): Properly propagate
20420 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
20421
20422 2013-05-07 Jakub Jelinek <jakub@redhat.com>
20423
20424 PR tree-optimization/57149
20425 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
20426 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
20427 collect_phi_def_edges, execute_late_warn_uninitialized): Use
20428 uninit_undefined_value_p instead of ssa_undefined_value_p.
20429
20430 PR debug/57184
20431 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
20432 for modifier == EXPAND_INITIALIZER.
20433
20434 2013-05-07 Anton Blanchard <anton@samba.org>
20435
20436 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
20437 for powerpc64 little endian.
20438 * configure: Regenerate.
20439
20440 2013-05-06 Graham Stott <grahams@btinternet.com>
20441
20442 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
20443 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
20444 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
20445 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
20446
20447 2013-05-06 Graham Stott <grahams@btinternet.com>
20448
20449 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
20450 codes which allow non-lvalues.
20451
20452 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
20453
20454 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
20455 components are all 1s.
20456 (integer_minus_onep): New function.
20457 * tree.h (integer_minus_onep): Declare it.
20458 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
20459 integer_minus_onep instead of integer_all_onesp.
20460
20461 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20462
20463 PR target/52933
20464 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
20465 variations of these patterns.
20466
20467 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20468
20469 * config/i386/i386.md (isa): Add x64_sse4 member.
20470 (enabled): Handle x64_sse4.
20471 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
20472 instruction for 64bit SSE4_1 targets. Update insn attributes.
20473 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
20474 instruction for SSE4_1 targets. Update insn attributes.
20475 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20476 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
20477 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
20478 const_1 selector.
20479 (*vec_extractv4si): Rename from *sse4_1_pextrd.
20480 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
20481 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
20482
20483 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20484
20485 PR target/57108
20486 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
20487
20488 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
20489
20490 * final.c (do_assembler_dialects): Don't handle curly braces and
20491 vertical bar escaped by % as dialect delimiters.
20492 (output_asm_insn): Print curly braces and vertical bar if escaped
20493 by % and ASSEMBLER_DIALECT defined.
20494 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
20495 * doc/tm.texi: Regenerated.
20496
20497 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20498
20499 * config/mips/mips.c: Include tree-pass.h.
20500 (mips_reorg): Split in pre- and post-dbr_schedule parts.
20501 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
20502 (pass_mips_machine_reorg2): New machine specific pass.
20503 (insert_pass_mips_machine_reorg2): New pass plugin definition.
20504 (mips_option_override): Register the new pass.
20505 * rtl.h (cleanup_barriers): Remove prototype.
20506 (dbr_schedule): Likewise.
20507 * jump.c (cleanup_barriers): Make static.
20508 * reorg.c (dbr_schedule): Likewise.
20509
20510 2013-05-06 Richard Biener <rguenther@suse.de>
20511
20512 PR tree-optimization/57185
20513 * tree-parloops.c (add_field_for_reduction): Handle anonymous
20514 SSA names properly.
20515
20516 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20517
20518 PR target/57106
20519 * config/i386/i386.c (add_parameter_dependencies): Add dependence
20520 between "first_arg" and "insn", not "last" and "insn".
20521
20522 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20523
20524 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
20525 (find_candidates_in_block): Re-enable slsr_process_phi.
20526 (create_phi_basis): Fix double counting of candidate adjustment.
20527
20528 2013-05-06 Richard Biener <rguenther@suse.de>
20529
20530 PR middle-end/57147
20531 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
20532 the edge is also fallthru, preserve it and just clear the
20533 abnormal flag.
20534 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
20535 also complex, preserve that and just clear the fallthru flag.
20536 * tree-inline.c (update_ssa_across_abnormal_edges): Also
20537 update virtual operands.
20538
20539 2013-05-06 Alan Modra <amodra@gmail.com>
20540
20541 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
20542 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
20543 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
20544 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
20545 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20546 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
20547
20548 2013-05-06 Alan Modra <amodra@gmail.com>
20549
20550 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
20551 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
20552 (DEFAULT_ASM_ENDIAN): Define.
20553 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20554 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
20555 Update -K PIC clause from sysv4.h.
20556 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
20557 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
20558
20559 2013-05-06 Alan Modra <amodra@gmail.com>
20560
20561 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
20562 twice for little-endian.
20563 (ashrdi3_no_power, ashrdi3): Support little-endian.
20564
20565 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20566
20567 PR target/55303
20568 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
20569 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
20570 related expanders.
20571 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
20572 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
20573 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
20574 New predicates.
20575
20576 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
20577 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20578
20579 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
20580 * config/pa/pa.opt: Make mbig-switch a no-op.
20581 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
20582 (CASE_VECTOR_MODE): Always return SImode.
20583 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
20584 for the !TARGET_BIG_SWITCH case.
20585 * config/pa/pa-linux.h: Likewise.
20586 * config/pa/pa-openbsd.h: Likewise.
20587 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
20588 * config/pa/pa.md (short_jump): Remove define_insn.
20589 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
20590 (casesi0): Remove define_insn.
20591 (type): Remove btable_branch.
20592 (pa_combine_type): Likewise.
20593 (in_nullified_branch_delay): Likewise.
20594 (in_call_delay): Likewise.
20595 (define_delay): Likewise.
20596 (define_insn_reservation "Z3"): Likewise.
20597 (define_insn_reservation "Z4"): Likewise.
20598 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
20599 (pa_adjust_insn_length): Remove adjustment for btable branches.
20600 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
20601 and mno-big-switch
20602
20603 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
20604
20605 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20606 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
20607 Add m->r,x alternatives.
20608 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
20609 splitters using SWI48x mode iterator.
20610 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
20611 TARGET_64BIT. Add m->x alternative.
20612 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
20613 Add o->x alternative. Enable for TARGET_SSE.
20614 (sse_storeq): Remove expander.
20615 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
20616 with memory input operand.
20617 (*vec_extractv2di_1 splitter): New.
20618 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
20619 * config/i386/i386.md (ssevecmodelower): New mode attribute.
20620
20621 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20622
20623 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
20624 (INT_LOWPART): Delete.
20625 (extract_MB): Adjust.
20626 (extract_ME): Adjust.
20627 (print_operand): Adjust.
20628
20629 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20630
20631 * config/rs6000/predicates.md (reg_or_add_cint_operand,
20632 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20633 (reg_or_logical_cint_operand, easy_fp_constant,
20634 logical_const_operand): Delete "CONST_DOUBLE" case.
20635 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
20636 "HOST_BITS_PER_WIDE_INT == 64" test.
20637 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
20638 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
20639 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
20640 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
20641 test.
20642 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
20643 CONST_DOUBLE DImode/VOIDmode case.
20644 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
20645 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20646 CONST_DOUBLE VOIDmode case.
20647 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20648 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
20649 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
20650 Delete CONST_DOUBLE case.
20651 (splitters for mov FMOVE64 const_double): Delete
20652 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20653 "HOST_BITS_PER_WIDE_INT >= 64" test.
20654 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
20655 case.
20656 (mov DI const_double): Delete.
20657
20658 2013-05-04 Jakub Jelinek <jakub@redhat.com>
20659
20660 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
20661 on op shows all bits zero in mode of a lowpart subreg, return zero.
20662
20663 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
20664
20665 PR target/57150
20666 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
20667 to save TFmode registers and DImode to save TImode registers for
20668 caller save operations.
20669 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
20670 mark being partially clobbered since they only use the first
20671 double word.
20672
20673 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
20674 and TDmode only use the upper 64-bits of each VSX register.
20675
20676 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20677
20678 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
20679 (find_candidates_in_block): Disable slsr_process_phi.
20680
20681 2013-05-03 Guozhi Wei <carrot@google.com>
20682
20683 * coverage.c (coverage_obj_init): Move the construction of gcov
20684 constructor to ...
20685 (build_init_ctor): ... here.
20686
20687 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20688
20689 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
20690 (slsr_cand_d): Redefine def_phi.
20691 (stride_status, phi_adjust_status, count_phis_status): New enums.
20692 (find_phi_def): New.
20693 (find_basis_for_base_expr): New.
20694 (find_basis_for_candidate): Handle hidden bases.
20695 (alloc_cand_and_find_basis): Handle phi candidates.
20696 (slsr_process_phi): New.
20697 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
20698 (create_mul_imm_cand): Likewise.
20699 (create_add_ssa_cand): Exclude phi base candidates.
20700 (create_add_imm_cand): Likewise.
20701 (slsr_process_cast): Likewise.
20702 (slsr_process_copy): Likewise.
20703 (find_candidates_in_block): Handle phi candidates.
20704 (dump_candidate): Likewise.
20705 (unconditional_cands): Delete.
20706 (unconditional_cands_with_known_stride_p): Delete.
20707 (phi_dependent_cand_p): New.
20708 (cand_increment): Handle phi-dependent candidates.
20709 (replace_dependent): Delete.
20710 (replace_mult_candidate): New.
20711 (replace_unconditional_candidate): New.
20712 (incr_vec_index): Move to avoid forward reference.
20713 (create_add_on_incoming_edge): New.
20714 (create_phi_basis): New.
20715 (replace_dependents): Delete.
20716 (replace_conditional_candidate): New.
20717 (phi_add_costs): New.
20718 (replace_uncond_cands_and_profitable_phis): New.
20719 (record_increment): Handle phi adjustments.
20720 (record_phi_increments): New.
20721 (record_increments): Handle phi adjustments.
20722 (phi_incr_cost): New.
20723 (lowest_cost_path): Handle phis.
20724 (total_savings): Likewise.
20725 (analyze_increments): Likewise.
20726 (ncd_with_phi): New.
20727 (ncd_of_cand_and_phis): New.
20728 (nearest_common_dominator_for_cands): Handle phi increments.
20729 (all_phi_incrs_profitable): New.
20730 (replace_profitable_candidates): Handle phi-dependent candidates.
20731 (analyze_candidates_and_replace): Likewise.
20732
20733 2013-05-03 Teresa Johnson <tejohnson@google.com>
20734
20735 PR bootstrap/57154
20736 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
20737 do not exceed REG_BR_PROB_BASE.
20738
20739 2013-05-03 Jeff Law <law@redhat.com>
20740
20741 PR tree-optimization/57144
20742 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
20743 operand of the condition will bit into the new type when eliminating
20744 a cast feeding a condition.
20745
20746 2013-05-03 Jakub Jelinek <jakub@redhat.com>
20747
20748 PR rtl-optimization/57130
20749 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
20750 of COMPARE as in_code to the recursive call if needed.
20751
20752 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
20753
20754 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
20755 (enabled): Handle new members.
20756 * config/i386/sse.md (*vec_concatv2si): Merge from
20757 *vec_concatv2si_sse2 and vec_concatv2si_sse.
20758 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
20759
20760 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
20761
20762 PR tree-optimization/57027
20763 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
20764 for fnms opportunity, check we got the prerequisite kind
20765 of tree / gimple before using accessor functions.
20766
20767 2013-05-03 Richard Biener <rguenther@suse.de>
20768
20769 * double-int.h (lshift): New overload without precision
20770 and arith argument.
20771 (operator *=, operator +=, operator -=): Move ...
20772 * double-int.c (operator *=, operator +=, operator -=): ... here
20773 and implement more efficiently.
20774 (mul_double_with_sign): Remove.
20775 (lshift_double): Adjust to take unsinged shift argument, push
20776 dispatching code to callers.
20777 (mul_double_wide_with_sign): Add early out for callers that
20778 are not interested in high parts or overflow.
20779 (lshift): New function.
20780 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
20781 dispatch code here.
20782 (lrotate, rrotate): Use logical shifts.
20783 * expr.c (get_inner_reference): Use lshift.
20784 * fixed-value.c (do_fixed_divide): Likewise.
20785 * tree-dfa.c (get_ref_base_and_extent): Likewise.
20786 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
20787 (indirect_refs_may_alias_p): Likewise.
20788 (stmt_kills_ref_p_1): Likewise.
20789
20790 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20791
20792 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
20793
20794 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20795
20796 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
20797 scalar form of FABD instruction.
20798
20799 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
20800
20801 * lra-constraints.c (process_alt_operands): Add checking alt
20802 number to choose the best alternative.
20803
20804 2013-05-02 Richard Biener <rguenther@suse.de>
20805
20806 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
20807 bitmap and its handling.
20808 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
20809
20810 2013-05-02 Richard Biener <rguenther@suse.de>
20811
20812 PR middle-end/57140
20813 * tree-inline.c (copy_loops): Properly handle removed loops.
20814 (copy_cfg_body): Mark destination loops for fixup if source
20815 loops needed fixup.
20816
20817 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
20818
20819 PR target/56732
20820 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
20821 generating simple_return for naked functions.
20822
20823 2013-05-02 Martin Jambor <mjambor@suse.cz>
20824
20825 PR middle-end/56988
20826 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
20827 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
20828 flags match.
20829 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
20830 ipa_agg_replacement_value structures.
20831 (known_aggs_to_agg_replacement_list): Likewise.
20832 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
20833 (read_agg_replacement_chain): Likewise.
20834 (ipcp_transform_function): Also check that by_ref flags match.
20835
20836 2013-05-02 Richard Biener <rguenther@suse.de>
20837
20838 * graphds.h (struct graph): Add obstack member.
20839 * graphds.c (new_graph): Initialize obstack and allocate
20840 vertices from it.
20841 (add_edge): Allocate edge from the obstack.
20842 (free_graph): Free the obstack instead of all edges and vertices.
20843
20844 2013-05-02 Teresa Johnson <tejohnson@google.com>
20845
20846 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
20847 divides.
20848 * cfg.c (update_bb_profile_for_threading): Ditto.
20849 * tree-inline.c (copy_bb): Ditto.
20850 (copy_edges_for_bb): Ditto.
20851 (initialize_cfun): Ditto.
20852 (copy_cfg_body): Ditto.
20853 (expand_call_inline): Ditto.
20854 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
20855 (estimate_node_size_and_time): Ditto.
20856 (inline_merge_summary): Ditto.
20857 * cgraphclones.c (cgraph_clone_edge): Ditto.
20858 (cgraph_clone_node): Ditto.
20859 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20860 (compute_trg_info): Ditto.
20861
20862 2013-05-02 Ian Bolton <ian.bolton@arm.com>
20863
20864 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
20865 S reg when fp attribute set.
20866 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
20867
20868 2013-05-02 Ian Bolton <ian.bolton@arm.com>
20869
20870 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
20871 New pattern.
20872 (*and_one_cmplsi3_compare0_uxtw): Likewise.
20873 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
20874 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
20875
20876 2013-05-02 Richard Biener <rguenther@suse.de>
20877
20878 * tree-scalar-evolution.c (scev_info_hasher): Remove.
20879 (struct instantiate_cache_entry): New type.
20880 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
20881 (struct instantiate_cache_type): New type.
20882 (set_instantiated_value, get_instantiated_value): Remove.
20883 (get_instantiated_value_entry): New function.
20884 (instantiate_scev_name): Use the new cache and adjust.
20885 (instantiate_scev_poly): Adjust.
20886 (instantiate_scev_binary): Likewise.
20887 (instantiate_array_ref): Likewise.
20888 (instantiate_scev_convert): Likewise.
20889 (instantiate_scev_not): Likewise.
20890 (instantiate_scev_3): Likewise.
20891 (instantiate_scev_2): Likewise.
20892 (instantiate_scev_r): Likewise.
20893 (instantiate_scev): Likewise.
20894 (resolve_mixers): Likewise.
20895
20896 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
20897
20898 PR target/57091
20899 * lra-constraints.c (best_small_class_operands_num): Remove.
20900 (process_alt_operands): Remove small_class_operands_num. Take
20901 small classes operands into losers and only if the operand is not
20902 matched. Modify debugging output.
20903 (curr_insn_transform): Remove best_small_class_operands_num.
20904 Print insn name.
20905
20906 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20907
20908 * config/aarch64/aarch64-builtins.c
20909 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
20910 * config/aarch64/aarch64-simd-builtins.def
20911 (reduc_splus_): Add new modes.
20912 (reduc_uplus_): New.
20913 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
20914 (reduc_uplus_v4sf): Likewise.
20915 (reduc_splus_v4sf): Likewise.
20916 (aarch64_addv<mode>): Likewise.
20917 (reduc_uplus_<mode>): Likewise.
20918 (reduc_splus_<mode>): Likewise.
20919 (aarch64_addvv2di): Likewise.
20920 (reduc_uplus_v2di): Likewise.
20921 (reduc_splus_v2di): Likewise.
20922 (aarch64_addvv2si): Likewise.
20923 (reduc_uplus_v2si): Likewise.
20924 (reduc_splus_v2si): Likewise.
20925 (reduc_<sur>plus_<mode>): New.
20926 (reduc_<sur>plus_v2di): Likewise.
20927 (reduc_<sur>plus_v2si): Likewise.
20928 (reduc_<sur>plus_v4sf): Likewise.
20929 (aarch64_addpv4sf): Likewise.
20930 * config/aarch64/arm_neon.h
20931 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
20932 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
20933 add UNSPEC_SADDV, UNSPEC_UADDV.
20934 (SUADDV): New.
20935 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
20936
20937 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20938
20939 * config/aarch64/arm_neon.h
20940 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
20941
20942 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20943
20944 * config/aarch64/aarch64-builtins
20945 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
20946
20947 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20948
20949 * config/aarch64/aarch64-simd-builtins.def
20950 (reduc_smax_): New.
20951 (reduc_smin_): Likewise.
20952 (reduc_umax_): Likewise.
20953 (reduc_umin_): Likewise.
20954 (reduc_smax_nan_): Likewise.
20955 (reduc_smin_nan_): Likewise.
20956 (fmax): Remove.
20957 (fmin): Likewise.
20958 (smax): Update for V2SF, V4SF and V2DF modes.
20959 (smin): Likewise.
20960 (smax_nan): New.
20961 (smin_nan): Likewise.
20962 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
20963 (<su><maxmin><mode>3): ...This, refactor.
20964 (s<maxmin><mode>3): New.
20965 (<maxmin_uns><mode>3): Likewise.
20966 (reduc_<maxmin_uns>_<mode>): Refactor.
20967 (reduc_<maxmin_uns>_v4sf): Likewise.
20968 (reduc_<maxmin_uns>_v2si): Likewise.
20969 (aarch64_<fmaxmin><mode>: Remove.
20970 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
20971 new builtin names.
20972 (vmin<q>_f<32,64>): Likewise.
20973 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
20974 (FMAXMIN): New.
20975 (su): Add mappings for smax, smin, umax, umin.
20976 (maxmin): New.
20977 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
20978 (FMAXMIN): Rename as...
20979 (FMAXMIN_UNS): ...This.
20980 (maxminv): Remove.
20981 (fmaxminv): Likewise.
20982 (fmaxmin): Likewise.
20983 (maxmin_uns): New.
20984 (maxmin_uns_op): Likewise.
20985
20986 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20987
20988 * config/aarch64/arm_neon.h
20989 (vac<ge, gt><sd>_f<32, 64>): Rename to...
20990 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
20991 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
20992
20993 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20994
20995 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
20996 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
20997
20998 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20999
21000 * config/aarch64/aarch64-simd.md
21001 (vcond<mode>_internal): Handle special cases for constant masks.
21002 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
21003 (vcondu<mode><mode>): Likewise.
21004 (vcond<v_cmp_result><mode>): New.
21005
21006 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21007
21008 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
21009 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
21010 * config/aarch64/aarch64-simd-builtins.def
21011 (cmeq): Update to BUILTIN_VALLDI.
21012 (cmgt): Likewise.
21013 (cmge): Likewise.
21014 (cmle): Likewise.
21015 (cmlt): Likewise.
21016 * config/aarch64/arm_neon.h
21017 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
21018 to builtins or C as appropriate.
21019
21020 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21021
21022 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
21023 (cmgeu): ...This.
21024 (cmhi): Rename to...
21025 (cmgtu): ...This.
21026 * config/aarch64/aarch64-simd.md
21027 (simd_mode): Add SF.
21028 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
21029 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
21030 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
21031 (cstore<mode>_neg): ...This.
21032 * config/aarch64/iterators.md
21033 (VALLF): new.
21034 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
21035 (COMPARISONS): New.
21036 (UCOMPARISONS): Likewise.
21037 (optab): Add missing comparisons.
21038 (n_optab): New.
21039 (cmp_1): Likewise.
21040 (cmp_2): Likewise.
21041 (CMP): Likewise.
21042 (cmp): Remove.
21043 (VCMP_S): Likewise.
21044 (VCMP_U): Likewise.
21045 (V_cmp_result): Add DF, SF modes.
21046 (v_cmp_result): Likewise.
21047 (v): Likewise.
21048 (vmtype): Likewise.
21049 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
21050
21051 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
21052
21053 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
21054 define_insn to define_insn_and_split.
21055 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
21056 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
21057 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
21058 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
21059 (thumb2_negscc): Likewise.
21060
21061 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21062
21063 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
21064
21065 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21066
21067 * config/arm/thumb2.md: Remove trailing whitespaces.
21068
21069 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21070
21071 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
21072 Use gen_int_mode rather than GEN_INT.
21073
21074 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
21075
21076 * value-prof.c (stream_in_histogram_value): Remove the strayed
21077 debug_gimple_stmt.
21078
21079 2013-04-30 Richard Biener <rguenther@suse.de>
21080
21081 PR middle-end/57122
21082 * cfghooks.c (split_edge): Properly check for the loop latch edge.
21083
21084 2013-04-30 Richard Biener <rguenther@suse.de>
21085
21086 PR middle-end/57107
21087 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
21088
21089 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
21090
21091 PR rtl-optimization/56957
21092 PR rtl-optimization/57105
21093 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
21094 variable. Use just INSN_UID for determining whether an insn
21095 should be only disconnected from the insn stream.
21096 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
21097
21098 2013-04-30 Jakub Jelinek <jakub@redhat.com>
21099
21100 PR tree-optimization/57104
21101 * tsan.c (instrument_expr): Don't instrument accesses to
21102 DECL_HARD_REGISTER VAR_DECLs.
21103
21104 2013-04-30 Richard Biener <rguenther@suse.de>
21105
21106 * function.h (loops_for_fn): New inline function.
21107 (set_loops_for_fn): Likewise.
21108 * cfgloop.h (place_new_loop): Add struct function parameter.
21109 (get_loop): Likewise.
21110 (get_loops): Likewise.
21111 (number_of_loops): Likewise.
21112 (fel_next): Adjust.
21113 (fel_init): Likewise.
21114 * cfg.c (get_loop_copy): Adjust.
21115 * cfgloop.c (flow_loops_dump): Likewise.
21116 (record_loop_exits): Likewise.
21117 (verify_loop_structure): Likewise.
21118 * cfgloopanal.c (mark_irreducible_loops): Likewise.
21119 (estimate_reg_pressure_cost): Likewise.
21120 (mark_loop_exit_edges): Likewise.
21121 * cfgloopmanip.c (place_new_loop): Likewise.
21122 (add_loop): Likewise.
21123 (duplicate_loop): Likewise.
21124 * graph.c (draw_cfg_nodes): Likewise.
21125 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
21126 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
21127 (extract_affine_chrec): Likewise.
21128 (build_scop_iteration_domain): Likewise.
21129 * graphite.c (graphite_initialize): Likewise.
21130 * ira-build.c (create_loop_tree_nodes): Likewise.
21131 (more_one_region_p): Likewise.
21132 (rebuild_regno_allocno_maps): Likewise.
21133 (mark_loops_for_removal): Likewise.
21134 (mark_all_loops_for_removal): Likewise.
21135 (remove_unnecessary_regions): Likewise.
21136 (ira_build): Likewise.
21137 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21138 * loop-init.c (fix_loop_structure): Likewise.
21139 (gate_rtl_move_loop_invariants): Likewise.
21140 (gate_rtl_unswitch): Likewise.
21141 (gate_rtl_unroll_and_peel_loops): Likewise.
21142 (rtl_doloop): Likewise.
21143 * lto-streamer-in.c (input_cfg): Likewise.
21144 * lto-streamer-out.c (output_cfg): Likewise.
21145 * modulo-sched.c (sms_schedule): Likewise.
21146 * predict.c (tree_estimate_probability): Likewise.
21147 (tree_estimate_probability_driver): Likewise.
21148 (estimate_loops): Likewise.
21149 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
21150 (move_sese_region_to_fn): Likewise.
21151 (debug_loop_num): Likewise.
21152 * tree-chrec.c (chrec_evaluate): Likewise.
21153 (hide_evolution_in_other_loops_than_loop): Likewise.
21154 (chrec_component_in_loop_num): Likewise.
21155 (reset_evolution_in_loop): Likewise.
21156 (evolution_function_is_invariant_rec_p): Likewise.
21157 * tree-if-conv.c (main_tree_if_conversion): Likewise.
21158 * tree-inline.c (copy_loops): Likewise.
21159 (copy_cfg_body): Likewise.
21160 (tree_function_versioning): Likewise.
21161 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
21162 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
21163 Likewise.
21164 (add_to_evolution_1): Likewise.
21165 (scev_const_prop): Likewise.
21166 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
21167 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
21168 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
21169 (tree_ssa_lim_initialize): Likewise.
21170 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
21171 (verify_loop_closed_ssa): Likewise.
21172 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
21173 (tree_ssa_loop_im): Likewise.
21174 (tree_ssa_loop_unswitch): Likewise.
21175 (tree_vectorize): Likewise.
21176 (check_data_deps): Likewise.
21177 (tree_ssa_loop_ivcanon): Likewise.
21178 (tree_ssa_loop_bounds): Likewise.
21179 (tree_complete_unroll): Likewise.
21180 (tree_complete_unroll_inner): Likewise.
21181 (tree_parallelize_loops): Likewise.
21182 (tree_ssa_loop_prefetch): Likewise.
21183 (tree_ssa_loop_ivopts): Likewise.
21184 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
21185 * tree-vectorizer.c (vectorize_loops): Likewise.
21186
21187 2013-04-29 Mike Frysinger <vapier@gentoo.org>
21188
21189 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
21190 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
21191 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
21192 with EABI_LINK_SPEC.
21193
21194 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21195
21196 PR target/44578
21197 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
21198 alternative.
21199
21200 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
21201
21202 PR target/57097
21203 * lra-constraints.c (process_alt_operands): Discourage a bit more
21204 using memory for pseudos. Print cost dump for alternatives.
21205 Modify cost values for conflicts with early clobbers.
21206 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
21207
21208 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21209
21210 PR target/57098
21211 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
21212
21213 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21214
21215 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
21216 from/to S register.
21217 (movdi_aarch64): Support LDR/STR from/to D register.
21218
21219 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21220
21221 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
21222 or higher by default.
21223
21224 2013-04-29 Richard Biener <rguenther@suse.de>
21225
21226 PR middle-end/57075
21227 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
21228 even if not adding abnormal edges for calls that can make
21229 abnormal gotos.
21230
21231 2013-04-29 Richard Biener <rguenther@suse.de>
21232
21233 PR middle-end/57103
21234 * tree-cfg.c (move_stmt_op): Fix condition under which to update
21235 TREE_BLOCK.
21236 (move_stmt_r): Remove redundant checking.
21237
21238 2013-04-29 Teresa Johnson <tejohnson@google.com>
21239
21240 PR bootstrap/57077
21241 * basic-block.h (apply_scale): New function.
21242 (apply_probability): Use apply_scale.
21243 * gimple-streamer-in.c (input_bb): Ditto.
21244 * lto-streamer-in.c (input_cfg): Ditto.
21245 * lto-cgraph.c (merge_profile_summaries): Ditto.
21246 * tree-optimize.c (execute_fixup_cfg): Ditto.
21247 * tree-inline.c (copy_bb): Update comment to use apply_scale.
21248 (copy_edges_for_bb): Ditto.
21249 (copy_cfg_body): Ditto.
21250
21251 2013-04-29 Tom de Vries <tom@codesourcery.com>
21252
21253 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
21254 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
21255 (tail_merge_optimize): Handle current_loops == NULL.
21256
21257 2013-04-26 Jeff Law <law@redhat.com>
21258
21259 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
21260 (simplify_cond_using_ranges): Generalize code to simplify
21261 COND_EXPRs where one argument is a constant and the other
21262 is an SSA_NAME created by an integral type conversion.
21263
21264 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21265
21266 * config/arm/arm.md (store_minmaxsi): Use only when
21267 optimize_insn_for_size_p.
21268
21269 2013-04-29 Christian Bruel <christian.bruel@st.com>
21270
21271 PR target/57108
21272 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
21273
21274 2013-04-29 Richard Biener <rguenther@suse.de>
21275
21276 PR middle-end/57089
21277 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
21278 loop tree make sure to schedule a fixup for the child as well.
21279 (expand_omp_for_generic): Properly add loops.
21280 (expand_omp_for_static_nochunk): Likewise.
21281 (expand_omp_for_static_chunk): Likewise.
21282 (expand_omp_for): For the degenerate case fixup loops.
21283 (expand_omp_sections): Fix default bb placement in loops.
21284 (expand_omp_atomic_pipeline): Properly add loops.
21285
21286 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21287
21288 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
21289
21290 2013-04-29 Tom de Vries <tom@codesourcery.com>
21291
21292 * tree-ssa-tail-merge.c: Update header comment.
21293
21294 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21295
21296 * config/aarch64/arm_neon.h
21297 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
21298 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
21299 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
21300 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
21301 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
21302 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
21303 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
21304 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
21305
21306 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21307
21308 * config/aarch64/aarch64-simd.md
21309 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
21310 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
21311 fix_trunc, fixuns_trunc.
21312 (ftrunc<VDQF:mode>2): New.
21313 * config/aarch64/iterators.md (optab): Add fix, fixuns.
21314 (fix_trunc_optab): New.
21315
21316 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21317
21318 * config/aarch64/aarch64-builtins.c
21319 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
21320 iceilf, lround, iroundf.
21321
21322 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21323
21324 PR target/54349
21325 * config/i386/i386.h (enum ix86_tune_indices)
21326 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
21327 New, split from X86_TUNE_INTER_UNIT_MOVES.
21328 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
21329 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
21330 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
21331 (TARGET_INTER_UNIT_MOVES): Remove.
21332 * config/i386/i386.c (initial_ix86_tune_features): Update.
21333 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
21334 (ix86_expand_convert_uns_didf_sse): Use
21335 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21336 (ix86_expand_vector_init_one_nonzero): Ditto.
21337 (ix86_expand_vector_init_interleave): Ditto.
21338 (inline_secondary_memory_needed): Return true for moves from SSE class
21339 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
21340 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
21341 * config/i386/constraints.md (Yi, Ym): Depend on
21342 TARGET_INTER_UNIT_MOVES_TO_VEC.
21343 (Yj, Yn): New constraints.
21344 * config/i386/i386.md (*movdi_internal): Change constraints of
21345 operand 1 from Yi to Yj and from Ym to Yn.
21346 (*movsi_internal): Ditto.
21347 (*movdf_internal): Ditto.
21348 (*movsf_internal): Ditto.
21349 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
21350 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21351 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
21352 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
21353 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
21354 * config/i386/sse.md (movdi_to_sse): Ditto.
21355 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
21356 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
21357 TARGET_INTER_UNIT_MOVES.
21358 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
21359 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
21360 instead of TARGET_INTER_UNIT_MOVES.
21361 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
21362 operand 1 from Yi to Yj and from Ym to Yn.
21363
21364 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21365
21366 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
21367 (float_truncate_hi_): Likewise.
21368 (float_extend_lo_): Likewise.
21369 (float_truncate_lo_): Likewise.
21370 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
21371 (aarch64_float_extend_lo_v2df): Likewise.
21372 (vec_unpacks_hi_v4sf): Likewise.
21373 (aarch64_float_truncate_lo_v2sf): Likewise.
21374 (aarch64_float_truncate_hi_v4sf): Likewise.
21375 (vec_pack_trunc_v2df): Likewise.
21376 (vec_pack_trunc_df): Likewise.
21377
21378 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21379
21380 * config/aarch64/aarch64-builtins.c
21381 (aarch64_fold_builtin): Fold float conversions.
21382 * config/aarch64/aarch64-simd-builtins.def
21383 (floatv2si, floatv4si, floatv2di): New.
21384 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
21385 * config/aarch64/aarch64-simd.md
21386 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
21387 * config/aarch64/iterators.md (FLOATUORS): New.
21388 (optab): Add float, floatuns.
21389 (su_optab): Likewise.
21390
21391 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21392
21393 * config/aarch64/aarch64-builtins.c
21394 (aarch64_builtin_vectorized_function): Use new names for
21395 fcvt builtins.
21396 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
21397 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
21398 (fcvtzu): Split as...
21399 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
21400 (fcvtas): Split as...
21401 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
21402 (fcvtau): Split as...
21403 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
21404 (fcvtps): Split as...
21405 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
21406 (fcvtpu): Split as...
21407 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
21408 (fcvtms): Split as...
21409 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
21410 (fcvtmu): Split as...
21411 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
21412 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
21413 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
21414 (lfrintnusf, lfrintnudf): Likewise.
21415 * config/aarch64/aarch64-simd.md
21416 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
21417 define_insn.
21418 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
21419 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
21420 (fcvt_pattern): Likewise.
21421
21422 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21423
21424 * config/aarch64/aarch64-simd.md
21425 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
21426 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
21427
21428 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21429
21430 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
21431 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
21432 (vrnd<a,m,n,p>_f32): Implement using builtins.
21433 (vrnd<i,x><q>_f<32, 64>): New.
21434
21435 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21436
21437 * config/aarch64/aarch64-builtins.c
21438 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
21439 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
21440 (frintz): Rename to...
21441 (btrunc): ...this.
21442 (frintp): Rename to...
21443 (ceil): ...this.
21444 (frintm): Rename to...
21445 (floor): ...this.
21446 (frinti): Rename to...
21447 (nearbyint): ...this.
21448 (frintx): Rename to...
21449 (rint): ...this.
21450 (frinta): Rename to...
21451 (round): ...this.
21452 * config/aarch64/aarch64-simd.md
21453 (aarch64_frint<frint_suffix><mode>): Delete.
21454 (<frint_pattern><mode>2): Convert to insn.
21455 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
21456 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
21457 (frint_pattern): Likewise.
21458 (frint_suffix): Likewise.
21459
21460 2013-04-29 Richard Biener <rguenther@suse.de>
21461
21462 PR tree-optimization/57081
21463 * loop-init.c: Include tree-flow.h.
21464 (loop_optimizer_finalize): Free number of iteration estimates.
21465 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
21466
21467 2013-04-29 Jakub Jelinek <jakub@redhat.com>
21468
21469 PR tree-optimization/57083
21470 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
21471 non-singleton shift count range, zero extend low_bound for uns case.
21472
21473 * config/i386/predicates.md (general_vector_operand): New predicate.
21474 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
21475 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
21476 if they aren't nonimmediate operands. If their original values
21477 satisfy const_vector_equal_evenodd_p, don't shift them.
21478 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
21479 predicates. For the SSE4.1 case force operands[{1,2}] into registers
21480 if not nonimmediate_operand.
21481 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
21482 instead of register_operand.
21483 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
21484
21485 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
21486
21487 * stor-layout.c (finalize_size_functions): Allocate a structure and
21488 reset cfun before dumping the functions.
21489
21490 2013-04-27 Jakub Jelinek <jakub@redhat.com>
21491
21492 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
21493
21494 PR target/56866
21495 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
21496 use xop_pmacsdqh if uns_p.
21497 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
21498 the immediate rotate count.
21499
21500 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
21501
21502 * rtl.h (struct rtx_def): Add comment for field jump.
21503 (LRA_SUBREG_P): New macro.
21504 * recog.c (register_operand): Check LRA_SUBREG_P.
21505 * lra.c (lra): Add note at the end of RTL code. Align non-empty
21506 stack frame.
21507 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
21508 (lra_final_code_change): Skip subreg change for operators.
21509 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
21510 if there are no operand changes.
21511 * lra-constraints.c (curr_insn_set): New.
21512 (match_reload): Set LRA_SUBREG_P.
21513 (emit_spill_move): Ditto.
21514 (check_and_process_move): Use curr_insn_set. Process only single
21515 set insns. Don't initialize sec_mem_p and change_p.
21516 (simplify_operand_subreg): Use LRA_SUBREG_P.
21517 (reg_in_class_p): New function.
21518 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
21519 of #ifdef. Add code to remove cycling.
21520 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
21521 non-null disp. Reload inner instead of disp when base and index
21522 are null. Try to put lo_sum into register.
21523 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
21524 (check_and_process_move): Move code for move cost check to
21525 simple_move_p. Remove equiv_substitution.
21526 (simple_move_p): New function.
21527 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
21528 curr_insn_set. Call check_and_process_move only for single set
21529 insns. Use the new function. Move call of check_and_process_move
21530 after operand equiv substitution and address process.
21531
21532 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21533
21534 PR go/57045
21535 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
21536 with nonlocal goto receivers or returns twice calls, ignore
21537 unininitialized values from abnormal edges to nl goto receiver
21538 or returns twice call.
21539
21540 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21541
21542 PR tree-optimization/57051
21543 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
21544 and VEC_RSHIFT_EXPR if shift count is a multiple of element
21545 bitsize.
21546
21547 2013-04-26 Richard Biener <rguenther@suse.de>
21548
21549 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
21550 (expand_omp_taskreg): Likewise. Mark loops for fixup.
21551 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
21552 (fixup_loop_arrays_after_move): New function.
21553 (move_sese_region_to_fn): Properly outline the loop tree parts
21554 of the SESE region.
21555
21556 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
21557
21558 * config/i386/i386.md (type, unit): Fix long lines.
21559
21560 2013-04-26 Richard Biener <rguenther@suse.de>
21561
21562 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
21563 (lto-streamer-out.o): Likewise.
21564 * cfgloop.c (init_loops_structure): Export, add struct function
21565 argument and adjust.
21566 (flow_loops_find): Adjust.
21567 * cfgloop.h (enum loop_estimation): Add EST_LAST.
21568 (init_loops_structure): Declare.
21569 * lto-streamer-in.c: Include cfgloop.h.
21570 (input_cfg): Input the loop tree.
21571 * lto-streamer-out.c: Include cfgloop.h.
21572 (output_cfg): Output the loop tree.
21573 (output_struct_function_base): Do not drop PROP_loops.
21574
21575 2013-03-26 Richard Biener <rguenther@suse.de>
21576
21577 * tree-cfg.c (execute_build_cfg): Build the loop tree.
21578 (pass_build_cfg): Provide PROP_loops.
21579 (move_sese_region_to_fn): Remove loops that are outlined into fn
21580 for now.
21581 * tree-inline.c: Include cfgloop.h.
21582 (initialize_cfun): Do not drop PROP_loops.
21583 (copy_loops): New function.
21584 (copy_cfg_body): Copy loop structure.
21585 (tree_function_versioning): Initialize destination loop tree.
21586 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
21587 (pass_parallelize_loops): Do IL verification.
21588 * loop-init.c (loop_optimizer_init): Fixup loops if required.
21589 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
21590 the CFG make sure we fixup loops as well.
21591 * tree-ssa-tail-merge.c: Include cfgloop.h.
21592 (replace_block_by): When merging loop latches mark loops for fixup.
21593 * lto-streamer-out.c (output_struct_function_base): Drop
21594 PROP_loops for now.
21595 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
21596 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
21597 * ipa-split.c: Include cfgloop.h.
21598 (split_function): Add the new return block to the loop tree root.
21599 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
21600 whether we have removed the forwarder block.
21601 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
21602 * cfgloop.h (place_new_loop): Declare.
21603 * cfgloopmanip.c (place_new_loop): Export.
21604 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
21605 (tree-switch-conversion.o): Likewise.
21606 (tree-complex.o): Likewise.
21607 (tree-inline.o): Likewise.
21608 (tree-ssa-tailmerge.o): Likewise.
21609 (ipa-split.o): Likewise.
21610 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
21611 (tree-ssa-copy.o): Likewise.
21612 * tree-switch-conversion.c: Include cfgloop.h
21613 (process_switch): If we emit a bit-test cascade, schedule loops
21614 for fixup.
21615 * tree-complex.c: Include cfgloop.h.
21616 (expand_complex_div_wide): Properly add new basic-blocks to loops.
21617 * asan.c: Include cfgloop.h.
21618 (create_cond_insert_point): Properly add new basic-blocks to
21619 loops, schedule loop fixup.
21620 * cfgloop.c (verify_loop_structure): Check that looks are not
21621 marked for fixup.
21622 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
21623 to loops.
21624 (expand_omp_for_generic): Likewise.
21625 (expand_omp_sections): Likewise.
21626 (expand_omp_atomic_pipeline): Schedule loops for fixup.
21627 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
21628 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
21629 is initialized, not when loops are present.
21630 * tree-parloops.c (parallelize_loops): Remove checking here.
21631 * passes.c (init_optimization_passes): Schedule a copy-propagation
21632 pass before complete unrolling of inner loops.
21633
21634 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21635
21636 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
21637 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
21638 (colorize_init): Add argument to _WIN32 version.
21639 * toplev.c: Include diagnostic-color.h.
21640 (process_options): Default to -fdiagnostics-color=auto if
21641 GCC_COLORS env var is in the environment.
21642 * common.opt (fdiagnostics-color=): Add Var and Init.
21643 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
21644 env var is in the environment, the default is auto rather than never.
21645
21646 * diagnostic.h (file_name_as_prefix): Add context argument.
21647 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
21648 the string as locus.
21649 * langhooks.c (lhd_print_error_function): Adjust caller.
21650
21651 2013-04-25 Lawrence Crowl <crowl@google.com>
21652
21653 * var-tracking.c (shared_hash_def::htab):
21654 Change type to hash_table. Update dependent calls and types.
21655
21656 2013-04-25 Lawrence Crowl <crowl@google.com>
21657
21658 * Makefile.in: Update as needed below.
21659
21660 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
21661 Move declaration to after the type's method definitons.
21662
21663 * attribs.c (htab_t scoped_attributes::attribute_hash):
21664 Change type to hash_table. Update dependent calls and types.
21665
21666 * bitmap.c (htab_t bitmap_desc_hash):
21667 Change type to hash_table. Update dependent calls and types.
21668
21669 * cselib.c (htab_t cselib_hash_table):
21670 Change type to hash_table. Update dependent calls and types.
21671
21672 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
21673 (hash_string_slot_node): Move implementation into lto-streamer.h
21674 struct string_slot_hasher.
21675 (eq_string_slot_node): Likewise.
21676
21677 * data-streamer-out.c: Update output_block::string_hash_table
21678 dependent calls and types.
21679
21680 * dwarf2cfi.c (htab_t trace_index):
21681 Change type to hash_table. Update dependent calls and types.
21682
21683 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
21684 Change type to hash_table. Update dependent calls and types.
21685 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
21686 (htab_t optimize_external_refs::map): Likewise.
21687 (htab_t output_comp_unit::extern_map): Likewise.
21688 (htab_t output_comdat_type_unit::extern_map): Likewise.
21689 (htab_t output_macinfo::macinfo_htab): Likewise.
21690 (htab_t optimize_location_lists::htab): Likewise.
21691 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
21692
21693 * except.c (htab_t ehspec_hash_type):
21694 Change type to hash_table. Update dependent calls and types.
21695 (assign_filter_values::ttypes): Likewise.
21696 (assign_filter_values::ehspec): Likewise.
21697 (sjlj_assign_call_site_values::ar_hash): Likewise.
21698 (convert_to_eh_region_ranges::ar_hash): Likewise.
21699
21700 * gcse.c (htab_t pre_ldst_table):
21701 Change type to hash_table. Update dependent calls and types.
21702
21703 * ggc-common.c (htab_t saving_htab):
21704 Change type to hash_table. Update dependent calls and types.
21705 (htab_t loc_hash): Likewise.
21706 (htab_t ptr_hash): Likewise.
21707 (call_count): Rename ggc_call_count.
21708 (call_alloc): Rename ggc_call_alloc.
21709 (loc_descriptor): Rename make_loc_descriptor.
21710 (add_statistics): Rename ggc_add_statistics.
21711
21712 * ggc-common.c (saving_htab):
21713 Change type to hash_table. Update dependent calls and types.
21714
21715 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
21716 (push_gimplify_context): Likewise.
21717 (pop_gimplify_context): Likewise.
21718 (struct gimple_temp_hash_elt): Added.
21719 (struct gimplify_hasher): Likewise.
21720 (struct gimplify_ctx.temp_htab):
21721 Change type to hash_table. Update dependent calls and types.
21722
21723 * gimple-fold.c: Include gimplify-ctx.h.
21724
21725 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
21726 Change type to hash_table. Update dependent calls and types.
21727 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
21728 avoid potential global name collision.
21729
21730 * gimplify.c: Include gimplify-ctx.h.
21731 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
21732 (htab_t gimplify_ctx::temp_htab):
21733 Update dependent calls and types for new type hash_table.
21734 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
21735 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
21736
21737 * gimplify-ctx.h: New.
21738 (struct gimple_temp_hash_elt): Move from gimplify.c.
21739 (class gimplify_hasher): New.
21740 (struct gimplify_ctx): Move from gimple.h.
21741 (htab_t gimplify_ctx::temp_htab):
21742 Change type to hash_table. Update dependent calls and types.
21743
21744 * graphite-clast-to-gimple.c: Include graphite-htab.h.
21745 (htab_t ivs_params::newivs_index):
21746 Change type to hash_table. Update dependent calls and types.
21747 (htab_t ivs_params::params_index): Likewise.
21748 (htab_t print_generated_program::params_index): Likewise.
21749 (htab_t gloog::newivs_index): Likewise.
21750 (htab_t gloog::params_index): Likewise.
21751
21752 * graphite.c: Include graphite-htab.h.
21753 4htab_t graphite_transform_loops::bb_pbb_mapping):
21754 Change type to hash_table. Update dependent calls and types.
21755
21756 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
21757 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
21758 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
21759
21760 * graphite-dependences.c: Include graphite-htab.h.
21761 (loop_is_parallel_p): Change hash table type of parameter.
21762
21763 * graphite-htab.h: New.
21764 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
21765 (extern find_pbb_via_hash): Move from graphite-poly.h.
21766 (extern loop_is_parallel_p): Move from graphite-poly.h.
21767 (extern get_loop_body_pbbs): Move from graphite-poly.h.
21768
21769 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
21770 (extern loop_is_parallel_p): Move to graphite-htab.h.
21771 (extern get_loop_body_pbbs): Move to graphite-htab.h.
21772
21773 * haifa-sched.c (htab_t delay_htab):
21774 Change type to hash_table. Update dependent calls and types.
21775 (htab_t delay_htab_i2): Likewise.
21776
21777 * ira-color.c (htab_t allocno_hard_regs_htab):
21778 Change type to hash_table. Update dependent calls and types.
21779
21780 * ira-costs.c (htab_t cost_classes_htab):
21781 Change type to hash_table. Update dependent calls and types.
21782
21783 * loop-invariant.c (htab_t merge_identical_invariants::eq):
21784 Change type to hash_table. Update dependent calls and types.
21785
21786 * loop-iv.c (htab_t bivs):
21787 Change type to hash_table. Update dependent calls and types.
21788
21789 * loop-unroll.c (htab_t opt_info::insns_to_split):
21790 Change type to hash_table. Update dependent calls and types.
21791 (htab_t opt_info::insns_with_var_to_expand): Likewise.
21792
21793 * lto-streamer.h (struct string_slot): Move from data-streamer.h
21794 (struct string_slot_hasher): New.
21795 (htab_t output_block::string_hash_table):
21796 Change type to hash_table. Update dependent calls and types.
21797
21798 * lto-streamer-in.c (freeing_string_slot_hasher): New.
21799 (htab_t file_name_hash_table):
21800 Change type to hash_table. Update dependent calls and types.
21801
21802 * lto-streamer-out.c: Update output_block::string_hash_table dependent
21803 calls and types.
21804
21805 * lto-streamer.c (htab_t tree_htab):
21806 Change type to hash_table. Update dependent calls and types.
21807
21808 * omp-low.c: Include gimplify-ctx.h.
21809
21810 * passes.c (htab_t name_to_pass_map):
21811 Change type to hash_table. Update dependent calls and types.
21812 (pass_traverse): Rename to passes_pass_traverse.
21813
21814 * plugin.c (htab_t event_tab):
21815 Change type to hash_table. Update dependent calls and types.
21816
21817 * postreload-gcse.c (htab_t expr_table):
21818 Change type to hash_table. Update dependent calls and types.
21819 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
21820
21821 * sese.c (debug_rename_map_1): Make extern.
21822 (htab_t copy_bb_and_scalar_dependences::rename_map):
21823 Change type to hash_table. Update dependent calls and types.
21824
21825 * sese.h (extern debug_rename_map): Move to .c file.
21826
21827 * store-motion.c (htab_t store_motion_mems_table):
21828 Change type to hash_table. Update dependent calls and types.
21829
21830 * trans-mem.c (htab_t tm_new_mem_hash):
21831 Change type to hash_table. Update dependent calls and types.
21832
21833 * tree-browser.c (htab_t TB_up_ht):
21834 Change type to hash_table. Update dependent calls and types.
21835
21836 * tree-cfg.c (htab_t discriminator_per_locus):
21837 Change type to hash_table. Update dependent calls and types.
21838
21839 * tree-complex.c: Include tree-hasher.h
21840 (htab_t complex_variable_components):
21841 Change type to hash_table. Update dependent calls and types.
21842
21843 * tree-eh.c (htab_t finally_tree):
21844 Change type to hash_table. Update dependent calls and types.
21845
21846 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
21847 struct int_tree_hasher.
21848 (extern int_tree_map_eq): Likewise.
21849 (uid_decl_map_hash): Removed.
21850 (extern decl_tree_map_eq): Likewise.
21851
21852 * tree-hasher.h: New.
21853 (struct int_tree_hasher): New.
21854 (typedef int_tree_htab_type): New.
21855
21856 * tree-inline.c: Include gimplify-ctx.h.
21857
21858 * tree-mudflap.c: Include gimplify-ctx.h.
21859
21860 * tree-parloops.c: Include tree-hasher.h.
21861 (htab_t eliminate_local_variables_stmt::decl_address):
21862 Change type to hash_table. Update dependent calls and types.
21863 (htab_t separate_decls_in_region::decl_copies): Likewise.
21864
21865 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
21866 Change type to hash_table. Update dependent calls and types.
21867
21868 * tree-sra.c (candidates):
21869 Change type to hash_table. Update dependent calls and types.
21870
21871 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
21872 in tree-flow.h.
21873 (int_tree_map_hash): Likewise.
21874
21875 * tree-ssa-dom.c (htab_t avail_exprs):
21876 Change type to hash_table. Update dependent calls and types.
21877
21878 * tree-ssa-live.c (var_map_base_init::tree_to_index):
21879 Change type to hash_table. Update dependent calls and types.
21880
21881 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
21882 Change type to hash_table. Update dependent calls and types.
21883
21884 * tree-ssa-phiopt.c (seen_ssa_names):
21885 Change type to hash_table. Update dependent calls and types.
21886
21887 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
21888 Change type to hash_table. Update dependent calls and types.
21889
21890 * tree-ssa-uncprop.c (equiv):
21891 Change type to hash_table. Update dependent calls and types.
21892
21893 2013-04-25 Jakub Jelinek <jakub@redhat.com>
21894
21895 PR rtl-optimization/57003
21896 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
21897 call note_stores with kill_clobbered_value callback again after
21898 killing regs_invalidated_by_call.
21899
21900 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21901
21902 * config/aarch64/aarch64-simd.md
21903 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
21904 (aarch64_simd_bsl<mode>): Likewise.
21905 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
21906
21907 2013-04-25 Marek Polacek <polacek@redhat.com>
21908
21909 PR tree-optimization/57066
21910 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
21911
21912 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21913
21914 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
21915
21916 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21917
21918 * config/aarch64/aarch64-builtins.c
21919 (aarch64_fold_builtin): New.
21920 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
21921 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
21922 * config/aarch64/aarch64-simd-builtins.def (abs): New.
21923 * config/aarch64/arm_neon.h
21924 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
21925
21926 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21927 Tejas Belagod <tejas.belagod@arm.com>
21928
21929 * config/aarch64/aarch64-builtins.c
21930 (aarch64_gimple_fold_builtin): New.
21931 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
21932 * config/aarch64/aarch64-simd-builtins.def (addv): New.
21933 * config/aarch64/aarch64-simd.md (addpv4sf): New.
21934 (addvv4sf): Update.
21935 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
21936
21937 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21938
21939 * config/aarch64/aarch64.md
21940 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
21941
21942 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21943
21944 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
21945 (*ngcsi_uxtw): New pattern.
21946
21947 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21948 Julian Brown <julian@codesourcery.com>
21949
21950 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
21951 (TB_DREG): Add T_V4HF.
21952 (v4hf_UP): New macro.
21953 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21954 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21955 Handle initialisation of V4HF. Adjust initialisation of reinterpret
21956 built-ins.
21957 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21958 (arm_vector_mode_supported_p): Handle V4HF.
21959 (arm_mangle_map): Handle V4HFmode.
21960 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
21961 * config/arm/arm_neon_builtins.def: Add entries for
21962 vcvtv4hfv4sf, vcvtv4sfv4hf.
21963 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
21964 (neon_vcvtv4hfv4sf): Likewise.
21965 * config/arm/neon-gen.ml: Handle half-precision floating point
21966 features.
21967 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
21968 * config/arm/arm_neon.h: Regenerate.
21969 * config/arm/neon.ml (type elts): Add F16.
21970 (type vectype): Add T_float16x4, T_floatHF.
21971 (type vecmode): Add V4HF.
21972 (type features): Add Requires_FP_bit feature.
21973 (elt_width): Handle F16.
21974 (elt_class): Likewise.
21975 (elt_of_class_width): Likewise.
21976 (mode_of_elt): Refactor.
21977 (type_for_elt): Handle F16, fix error messages.
21978 (vectype_size): Handle T_float16x4.
21979 (vcvt_sh): New function.
21980 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
21981 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
21982 (string_of_mode): Handle V4HF.
21983 * doc/arm-neon-intrinsics.texi: Regenerate.
21984
21985 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21986
21987 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
21988 format specifier in 'X' case.
21989
21990 2013-04-25 Alan Modra <amodra@gmail.com>
21991
21992 PR target/57052
21993 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
21994 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
21995 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
21996 Repeat for many other rotate/shift and mask patterns using subregs.
21997 Name lshiftrt insns.
21998 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
21999 on WORDS_BIG_ENDIAN.
22000
22001 2013-04-25 Alan Modra <amodra@gmail.com>
22002
22003 * config.gcc: Support little-endian powerpc-linux targets.
22004 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
22005 (LINK_OS_LINUX_SPEC): Define.
22006 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
22007 Preserve MASK_LITTLE_ENDIAN.
22008 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
22009 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
22010 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
22011 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
22012 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
22013 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
22014 Correct fp word order for little-endian. Don't shift toc entries
22015 smaller than a word for little-endian.
22016 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
22017 (bswapdi2 splits): Correct low-part subreg for little-endian.
22018 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
22019 low/high where such is correct only for be.
22020 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
22021 little-endian for -mcall-aixdesc.
22022
22023 2013-04-25 Alan Modra <amodra@gmail.com>
22024
22025 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
22026 replace_equiv_address_nv.
22027
22028 2013-04-25 Alan Modra <amodra@gmail.com>
22029
22030 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
22031
22032 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22033
22034 Revert:
22035 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22036 * rtl.h (struct rtx_def): ...
22037
22038 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22039
22040 PR rtl-optimizations/57046
22041 * lra-constraints (split_reg): Set up lra_risky_transformations_p
22042 for multi-reg splits.
22043
22044 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
22045
22046 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
22047
22048 2013-04-24 Sterling Augustine <saugustine@google.com>
22049
22050 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
22051 (comp_dir_string, debug_str_dwo_section): New.
22052 (DEBUG_STR_DWO_SECTION): Rename to ...
22053 (DEBUG_DWO_STR_SECTION): ... this.
22054 (DEBUG_NORM_STR_SECTION): Delete.
22055 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
22056 (DEBUG_STR_DWO_SECTION_FLAGS): New.
22057 (find_AT_string): Move most logic to ...
22058 (find_AT_string_in_table): ... here. New.
22059 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
22060 add_skeleton_AT_string. Delete logic.
22061 (output_skeleton_debug_sections): Remove call to
22062 add_top_level_skeleton_die_attrs.
22063 (add_comp_dir_attribute): Move logic to comp_dir_string.
22064 (dwarf2out_init): Initialize debug_str_dwo_section.
22065 (output_indirect_string): Call find_string_form.
22066 (output_indirect_strings): Rewrite.
22067 (prune_unused_types): Empty skeleton_debug_str_hash.
22068 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
22069 (dwarf2out_finish): Call output_indirect_strings.
22070
22071 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22072
22073 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
22074
22075 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22076
22077 * rtl.h (struct rtx_def): Add comment for field jump.
22078 (LRA_SUBREG_P): New macro.
22079 * recog.c (register_operand): Check LRA_SUBREG_P.
22080 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22081 stack frame.
22082 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22083 (lra_final_code_change): Skip subreg change for operators.
22084 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22085 if there are no operand changes.
22086 * lra-constraints.c (curr_insn_set): New.
22087 (match_reload): Set LRA_SUBREG_P.
22088 (emit_spill_move): Ditto.
22089 (check_and_process_move): Use curr_insn_set. Process only single
22090 set insns. Don't initialize sec_mem_p and change_p.
22091 (simplify_operand_subreg): Use LRA_SUBREG_P.
22092 (reg_in_class_p): New function.
22093 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22094 of #ifdef. Add code to remove cycling.
22095 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22096 non-null disp. Reload inner instead of disp when base and index
22097 are null. Try to put lo_sum into register.
22098 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22099 (check_and_process_move): Move code for move cost check to
22100 simple_move_p. Remove equiv_substitution.
22101 (simple_move_p): New function.
22102 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22103 curr_insn_set. Call check_and_process_move only for single set
22104 insns. Use the new function. Move call of check_and_process_move
22105 after operand equiv substitution and address process.
22106
22107 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
22108
22109 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
22110 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
22111 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
22112
22113 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22114
22115 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
22116
22117 2013-04-24 Marek Polacek <polacek@redhat.com>
22118
22119 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
22120 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
22121 (select_loops_exit_conditions): Likewise.
22122 (number_of_iterations_for_all_loops): Likewise.
22123 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
22124 (scev_analysis): Likewise.
22125
22126 2013-04-02 Catherine Moore <clm@codesourcery.com>
22127 Chao-ying Fu <fu@mips.com>
22128
22129 * config/mips/micromips.md (jraddiusp): New pattern.
22130 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
22131 instruction if possible.
22132
22133 2013-04-24 Alan Modra <amodra@gmail.com>
22134
22135 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
22136
22137 2013-04-24 Julian Brown <julian@codesourcery.com>
22138 Chung-Lin Tang <cltang@codesourcery.com>
22139
22140 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
22141 dependency behavior in enumeration type DIE generation. Add TODO note
22142 to comments about future DW_FORM_sdata/udata re-work of related code.
22143
22144 2013-04-23 Lawrence Crowl <crowl@google.com>
22145
22146 * Makefile.in: Update as needed below.
22147
22148 * hash-table.h (class hash_table):
22149 Correct many methods with parameter types compare_type to the correct
22150 value_type. (Correct code was unlikely to notice the change.)
22151 (hash_table::elements_with_deleted) New.
22152 (class hashtable::iterator): New.
22153 (hashtable::begin()): New.
22154 (hashtable::end()): New.
22155 (FOR_EACH_HASH_TABLE_ELEMENT): New.
22156
22157 * statistics.c (statistics_hashes):
22158 Change type to hash_table. Update dependent calls and types.
22159
22160 * tree-into-ssa.c (var_infos):
22161 Change type to hash_table. Update dependent calls and types.
22162
22163 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
22164 Change type to hash_table. Update dependent calls and types.
22165
22166 * tree-ssa-loop-im.c (struct mem_ref.refs):
22167 Change type to hash_table. Update dependent calls and types.
22168
22169 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
22170 Change type to hash_table. Update dependent calls and types.
22171
22172 * tree-ssa-sccvn.c (vn_tables_s::nary):
22173 Change type to hash_table. Update dependent calls and types.
22174 (vn_tables_s::phis): Likewise.
22175 (vn_tables_s::references): Likewise.
22176
22177 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
22178 (vn_reference_eq): Update parameter and return types.
22179
22180 * tree-ssa-structalias.c (pointer_equiv_class_table):
22181 Change type to hash_table. Update dependent calls and types.
22182 (location_equiv_class_table): Likewise.
22183
22184 * tree-vect-data-refs.c: Consequential changes for making
22185 peeling a hash_table.
22186
22187 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
22188 (destroy_loop_vec_info): Dependent hash_table update.
22189
22190 * tree-vectorizer.h (peeling_htab):
22191 Change type to hash_table. Update dependent calls and types.
22192
22193 2013-04-23 Shiva Chen <shiva0217@gmail.com>
22194
22195 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
22196 to check the register content is equal or not.
22197 * lra-constraints.c (match_reload): Use lra_assign_reg_val
22198 to assign register content record.
22199 * lra-eliminations.c (update_reg_eliminate): Use
22200 lra_update_reg_val_offset to update register content offset.
22201 * lra-int.h (struct lra_reg): Add offset member.
22202 (lra_reg_val_equal_p): New static inline function.
22203 (lra_update_reg_val_offset): New static inline function.
22204 (lra_assign_reg_val): New static inline function.
22205 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
22206 to assign register content record.
22207 (initialize_lra_reg_info_element): Initial offset to zero.
22208
22209 2013-04-23 Catherine Moore <clm@codesourcery.com>
22210
22211 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
22212 operands. Record compression.
22213
22214 2013-04-23 Xinliang David Li <davidxl@google.com>
22215
22216 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
22217
22218 2013-04-23 Richard Biener <rguenther@suse.de>
22219
22220 PR middle-end/57036
22221 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
22222 parameter, only add abnormal goto edges from the copied body
22223 if the call could perform abnormal gotos.
22224 (copy_cfg_body): Adjust.
22225
22226 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
22227
22228 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
22229
22230 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
22231
22232 * coretypes.h (gimple_stmt_iterator): Add struct to make
22233 compatible with C.
22234
22235 2013-04-23 Richard Biener <rguenther@suse.de>
22236
22237 PR tree-optimization/57026
22238 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
22239 from SSA names occuring in abnormal PHI nodes.
22240
22241 2013-04-22 Andi Kleen <ak@linux.intel.com>
22242
22243 * lto/lto.c (print_lto_report_1): Fix LTO report names.
22244
22245 2013-04-22 Andi Kleen <ak@linux.intel.com>
22246
22247 * lto/lto.c (print_lto_report_1): Declare early.
22248 (read_cgraph_and_symbols): Call print_lto_report_1 early.
22249
22250 2013-04-22 Andi Kleen <ak@linux.intel.com>
22251
22252 * common.opt (-flto-report-wpa): Add.
22253 * doc/invoke.texi (-flto-report-wpa): Add.
22254 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
22255 (lto_main): dito.
22256
22257 2013-04-22 Xinliang David Li <davidxl@google.com>
22258
22259 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
22260 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
22261 * Makefile.in: New dependency
22262
22263 David Daney <ddaney.cavm@gmail.com>
22264
22265 * configure.ac (gcc_cv_as_micromips_support): Use the
22266 --fatal-warnings option.
22267 * configure: Regenerate.
22268
22269 2013-04-22 Marek Polacek <polacek@redhat.com>
22270
22271 PR sanitizer/56990
22272 * tsan.c (instrument_expr): Don't instrument expression
22273 in case its size is zero.
22274
22275 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
22276
22277 PR target/57032
22278 Revert:
22279 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22280
22281 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22282
22283 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22284
22285 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
22286 (gimple_stmt_iterator): New typedef.
22287 * gimple.h (gimple_stmt_iterator): Rename to...
22288 (gimple_stmt_iterator_d): ... This.
22289 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
22290 trees be valid for GIMPLE and GENERIC.
22291 (TARGET_GIMPLE_FOLD_BUILTIN): New.
22292 * gimple-fold.c (gimple_fold_call): Call target hook
22293 gimple_fold_builtin.
22294 * hooks.c (hook_bool_gsiptr_false): New.
22295 * hooks.h (hook_bool_gsiptr_false): New.
22296 * target.def (fold_stmt): New.
22297 * doc/tm.texi: Regenerate.
22298
22299 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
22300
22301 PR target/57018
22302 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
22303 a set sp if no stack realignment.
22304
22305 2013-04-22 Nick Clifton <nickc@redhat.com>
22306
22307 * config.gcc (tilegx-linux): Extend extra_objs rather than
22308 overwriting it.
22309 (tilepro-linux): Likewise.
22310
22311 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22312
22313 * config/aarch64/aarch64-builtins.c
22314 (CF): Remove.
22315 (CF0, CF1, CF2, CF3, CF4, CF10): New.
22316 (VAR<1-12>): Add MAP parameter.
22317 (BUILTIN_*): Likewise.
22318 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
22319 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
22320 (aarch64_ushl_n<mode>): Likewise.
22321 (aarch64_sshr_n<mode>): Likewise.
22322 (aarch64_ushr_n<mode>): Likewise.
22323 (aarch64_<maxmin><mode>): Likewise.
22324 (aarch64_sqrt<mode>): Likewise.
22325 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
22326 (vshr<q>_n_*): Likewise.
22327
22328 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22329
22330 * config/aarch64/aarch64-builtins.c
22331 (aarch64_simd_builtin_type_mode): Handle SF types.
22332 (sf_UP): Define.
22333 (BUILTIN_GPF): Define.
22334 (aarch64_init_simd_builtins): Handle SF types.
22335 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
22336 (frecps): Likewise.
22337 (frecpx): Likewise.
22338 * config/aarch64/aarch64-simd.md
22339 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
22340 (aarch64_frecpe<mode>): New.
22341 (aarch64_frecps<mode>): Likewise.
22342 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
22343 (v8type): Add frecp<esx>.
22344 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
22345 (aarch64_frecps<mode>): Likewise.
22346 * config/aarch64/iterators.md (FRECP): New.
22347 (frecp_suffix): Likewise.
22348 * config/aarch64/arm_neon.h
22349 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
22350
22351 2013-04-22 Christian Bruel <christian.bruel@st.com>
22352
22353 PR target/56995
22354 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
22355 (REG_CLASS_NAMES): Idem.
22356 (REG_CLASS_CONTENTS): Idem.
22357 (REGCLASS_HAS_FP_REG): Idem.
22358 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
22359 (sh_conditional_register_usage): Idem.
22360
22361 2013-04-21 Jeff Law <law@redhat.com>
22362
22363 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
22364 (ssa_forward_propagate_and_combine): Use it.
22365
22366 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22367
22368 * lra.c: Update the flow chart diagram.
22369
22370 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22371
22372 PR rtl-optimization/56847
22373 * lra-constraints.c (process_alt_operands): Discourage alternative
22374 with non-matche doffsettable memory constraint fro memory with
22375 known offset.
22376
22377 2013-04-19 Richard Biener <rguenther@suse.de>
22378
22379 PR tree-optimization/56982
22380 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
22381 function.
22382 * gimplify.c (gimplify_call_expr): Notice special calls.
22383 (gimplify_modify_expr): Likewise.
22384 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
22385 abnormal control flow receivers.
22386 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
22387 in the same way as cfun->has_nonlocal_labels.
22388 (gimple_purge_dead_abnormal_call_edges): Likewise.
22389 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
22390 receivers start a basic-block.
22391
22392 2013-04-19 Richard Biener <rguenther@suse.de>
22393
22394 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
22395 member ...
22396 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
22397 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
22398 (SLP_TREE_LOAD_PERMUTATION): Add.
22399 (vect_transform_slp_perm_load): Adjust prototype.
22400 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
22401 (vect_free_slp_instance): Likewise.
22402 (vect_create_new_slp_node): Likewise.
22403 (vect_supported_slp_permutation_p): Remove.
22404 (vect_slp_rearrange_stmts): Adjust.
22405 (vect_supported_load_permutation_p): Likewise. Inline
22406 vect_supported_slp_permutation_p here.
22407 (vect_analyze_slp_instance): Compute load permutations per
22408 slp node instead of per instance.
22409 (vect_get_slp_defs): Adjust.
22410 (vect_transform_slp_perm_load): Likewise.
22411 (vect_schedule_slp_instance): Remove redundant code.
22412 (vect_schedule_slp): Remove hack for PR56270, add it ...
22413 * tree-vect-stmts.c (vectorizable_load): ... here, do not
22414 CSE loads for SLP. Adjust.
22415
22416 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22417
22418 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
22419 spelling in two comments.
22420
22421 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22422
22423 PR target/56797
22424 * config/arm/arm.c (load_multiple_sequence): Require SP
22425 as base register for loads if SP is in the register list.
22426
22427 2013-04-19 Martin Jambor <mjambor@suse.cz>
22428
22429 PR tree-optimization/56718
22430 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
22431 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
22432 and made public. Adjusted all callers.
22433 (ipa_intraprocedural_devirtualization): New function.
22434 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
22435 (ipa_intraprocedural_devirtualization): Likewise.
22436 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
22437
22438 2013-04-19 Richard Biener <rguenther@suse.de>
22439
22440 PR tree-optimization/57000
22441 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
22442
22443 2013-04-19 Terry Guo <terry.guo@arm.com>
22444
22445 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
22446 Replace with ...
22447 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
22448 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
22449 (cortex_m4_fmacs): Use new reservations.
22450 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
22451
22452 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
22453
22454 PR rtl-optimization/56999
22455 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
22456 related code.
22457 (lra_coalesce): Remove split_origin_bitmap and related code.
22458 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
22459 ranges if necessary.
22460
22461 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
22462
22463 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
22464 New array.
22465 (ix86_expand_call): Remove clobbered_registers array and use
22466 x86_64_ms_sysv_extra_clobbered_registers instead.
22467 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
22468 Declare here.
22469 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
22470 predicate.
22471 * config/i386/i386.md (*call_rex64_ms_sysv): Use
22472 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
22473 (*call_value_rex64_ms_sysv): Ditto.
22474
22475 2013-04-18 Cary Coutant <ccoutant@google.com>
22476
22477 * dwarf2out.c (output_pubnames): Check die_perennial_p of
22478 parent instead of die_mark.
22479
22480 2013-04-18 Diego Novillo <dnovillo@google.com>
22481
22482 * gimple.c (create_gimple_tmp): New.
22483 (get_expr_type): New.
22484 (build_assign): New.
22485 (build_type_cast): New.
22486 * gimple.h (enum ssa_mode): Define.
22487 (gimple_seq_set_location): New.
22488 * asan.c (build_check_stmt): Change some gimple_build_* calls
22489 to use build_assign and build_type_cast.
22490
22491 2013-04-18 Richard Biener <rguenther@suse.de>
22492
22493 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
22494 handle negative step. Remove redundant checks.
22495 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
22496 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
22497 for negative step and grouped loads fail to vectorize.
22498
22499 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22500
22501 * emit-rtl.c (reset_insn_used_flags): New function.
22502 (reset_all_used_flags): Use it.
22503 (verify_insn_sharing): New function.
22504 (verify_rtl_sharing): Fix verification for SEQUENCEs.
22505
22506 2013-04-18 Jakub Jelinek <jakub@redhat.com>
22507
22508 PR tree-optimization/56984
22509 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
22510 and (x >> M) >= N don't register any assertion if N << M is the
22511 minimum value.
22512
22513 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22514
22515 * lower-subreg.c (resolve_simple_move): If called self-recursive,
22516 do not delete_insn insns that have not yet been emitted, only
22517 unlink them with remove_insn.
22518 * df-scan.c (df_insn_delete): Revert r197492.
22519
22520 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
22521
22522 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
22523 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
22524
22525 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22526
22527 * config/arm/arm.md (movsicc_insn): Convert define_insn into
22528 define_insn_and_split.
22529 (and_scc,ior_scc,negscc): Likewise.
22530 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
22531
22532 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22533
22534 * config/arm/arm.c (use_return_insn): Return 0 for targets that
22535 can benefit from using a sequence of LDRD instructions in epilogue
22536 instead of a single LDM instruction.
22537
22538 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
22539
22540 PR 45688
22541 * doc/extend.texi: Fix typo.
22542
22543 2013-04-17 Richard Biener <rguenther@suse.de>
22544
22545 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
22546 (vect_build_slp_tree): ... here.
22547 (vect_build_slp_tree_1): Compute which stmts of the SLP group
22548 match. Remove special-casing of mismatched complex loads.
22549 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
22550 re-try the match with swapped commutative operands.
22551 (vect_supported_load_permutation_p): Remove special-casing of
22552 mismatched complex loads.
22553 (vect_analyze_slp_instance): Adjust.
22554
22555 2013-04-17 Richard Biener <rguenther@suse.de>
22556
22557 PR rtl-optimization/56921
22558 * cfgloop.h (struct loop): Add simple_loop_desc member.
22559 (struct niter_desc): Mark with GTY(()).
22560 (simple_loop_desc): Do not use aux field but simple_loop_desc.
22561 * loop-iv.c (get_simple_loop_desc): Likewise.
22562 (free_simple_loop_desc): Likewise.
22563
22564 Revert
22565 2013-04-16 Richard Biener <rguenther@suse.de>
22566
22567 PR rtl-optimization/56921
22568 * loop-init.c (pass_rtl_move_loop_invariants): Add
22569 TODO_do_not_ggc_collect to todo_flags_finish.
22570 (pass_rtl_unswitch): Same.
22571 (pass_rtl_unroll_and_peel_loops): Same.
22572 (pass_rtl_doloop): Same.
22573
22574 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
22575
22576 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
22577 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
22578 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
22579 references.
22580 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
22581 * tree-streamer.c (record_common_node): Adjust reference in comment.
22582
22583 2013-04-17 Terry Guo <terry.guo@arm.com>
22584
22585 * config/arm/cortex-m4.md: Add a new bypass.
22586
22587 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22588
22589 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
22590 New pattern.
22591 (*subs_<optab><mode>_multp2): New pattern.
22592 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22593 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22594
22595 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22596
22597 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
22598 (*subs_mul_imm_<mode>): New pattern.
22599
22600 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
22601
22602 PR target/56948
22603 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
22604 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
22605 (vsx_movti_32bit): Change j->wa to O->wa.
22606
22607 2013-04-16 Richard Biener <rguenther@suse.de>
22608
22609 PR rtl-optimization/56921
22610 * loop-init.c (pass_rtl_move_loop_invariants): Add
22611 TODO_do_not_ggc_collect to todo_flags_finish.
22612 (pass_rtl_unswitch): Same.
22613 (pass_rtl_unroll_and_peel_loops): Same.
22614 (pass_rtl_doloop): Same.
22615
22616 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
22617
22618 * config/arm/arm.c (emit_multi_reg_push): New declaration
22619 for an existing function.
22620 (arm_emit_strd_push): New function.
22621 (arm_expand_prologue): Used here.
22622 (arm_emit_ldrd_pop): New function.
22623 (arm_expand_epilogue): Used here.
22624 (arm_get_frame_offsets): Update condition.
22625 (arm_emit_multi_reg_pop): Add a special case for load of a single
22626 register with writeback.
22627
22628 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
22629
22630 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
22631 description.
22632
22633 2013-04-16 Richard Biener <rguenther@suse.de>
22634
22635 PR tree-optimization/56756
22636 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
22637 (first_mem_ref_loc): New.
22638 (execute_sm): Place the load temporarily before a previous
22639 access instead of in the latch edge to ensure its SSA dependencies
22640 are defined at points dominating the load.
22641
22642 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
22643
22644 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
22645 correct fix by moving header and footer insn to the footer of
22646 the merged basic block. Clear BB_END of the merged-away block.
22647
22648 PR middle-end/43631
22649 * emit-rtl.c (make_note_raw): New function.
22650 (link_insn_into_chain): New static inline function.
22651 (add_insn): Use it.
22652 (add_insn_before, add_insn_after): Factor insn chain linking code...
22653 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
22654 using link_insn_into_chain.
22655 (note_outside_basic_block_p): New helper function for emit_note_after
22656 and emit_note_before.
22657 (emit_note_after): Use nobb variant of add_insn_after if the note
22658 should not be contained in a basic block.
22659 (emit_note_before): Use nobb variant of add_insn_before if the note
22660 should not be contained in a basic block.
22661 (emit_note_copy): Use make_note_raw.
22662 (emit_note): Likewise.
22663 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
22664 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
22665 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
22666 the moved barrier the tail of the basic block it follows.
22667 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
22668
22669 2013-04-15 Jakub Jelinek <jakub@redhat.com>
22670
22671 PR tree-optimization/56962
22672 * gimple-ssa-strength-reduction.c (record_increment): Only set
22673 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
22674 either rhs1 or rhs2 is equal to c->base_expr.
22675
22676 2013-04-15 Richard Biener <rguenther@suse.de>
22677
22678 PR tree-optimization/56933
22679 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
22680 member.
22681 (GROUP_READ_WRITE_DEPENDENCE): Remove.
22682 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
22683 * tree-vect-data-refs.c (vect_analyze_group_access): Move
22684 dependence check ...
22685 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
22686 ... here.
22687 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
22688 GROUP_READ_WRITE_DEPENDENCE.
22689
22690 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22691
22692 * emit-rtl.c (reset_all_used_flags): New function.
22693 (verify_rtl_sharing): Call reset_all_used_flags before and after
22694 performing the checks.
22695
22696 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22697
22698 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
22699 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
22700 * config/arm/constraints.md (De): New constraint.
22701 * config/arm/neon.md (anddi3_neon): Delete.
22702 (neon_vand<mode>): Expand to standard anddi3 pattern.
22703 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
22704 Move earlier in the file.
22705 (neon_inv_logic_op2): Likewise.
22706 (arm_anddi_operand_neon): New predicate.
22707
22708 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22709
22710 * configure.ac (gcc_cv_ld_as_needed): Set
22711 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
22712 Use -z ignore, -z record on *-*-solaris2*.
22713 (HAVE_LD_AS_NEEDED): Update comment.
22714 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
22715 * configure: Regenerate.
22716 * config.in: Regenerate.
22717 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
22718 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
22719 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
22720 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
22721 equivalents. Fix markup.
22722 * doc/tm.texi: Regenerate.
22723
22724 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
22725
22726 * config/i386/i386.opt: New option mstack-protector-guard=.
22727 * config/i386/i386-opts.h: Add enum stack_protector_guard.
22728 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
22729 TARGET_SSP_TLS_GUARD.
22730 * config/i386/i386.c (ix86_option_override_internal): Set
22731 ix86_stack_protector_guard.
22732 * config/i386/i386.md (stack_protect_set): Enable for
22733 TARGET_SSP_TLS_GUARD only.
22734 (stack_protect_set_<mode>): Ditto.
22735 (stack_protect_test): Ditto.
22736 (stack_protect_test_<mode>): Ditto.
22737 * doc/invoke.texi (i386 Option): Document.
22738
22739 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
22740
22741 PR target/56890
22742 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
22743 (S_MODES): Set H_MODE bit.
22744 (SF_MODES): Set only S_MODE and SF_MODE bits.
22745 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
22746 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
22747 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
22748 <MODE_FLOAT>: Likewise.
22749
22750 2013-04-15 Joey Ye <joey.ye@arm.com>
22751
22752 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
22753
22754 2013-04-15 Joey Ye <joey.ye@arm.com>
22755
22756 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
22757 for real far jump.
22758 (thumb_far_jump_used_p): Count instruction size and set
22759 far_jump_used.
22760
22761 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
22762
22763 * reorg.c (fill_simple_delay_slots): Reindent block of code.
22764 * resource.c (mark_target_live_regs): Reformat conditional block.
22765
22766 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22767
22768 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
22769 notes, they are emitted only just before final.
22770 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
22771
22772 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22773
22774 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
22775 * cfgrtl.c (delete_insn): Call it here instead.
22776 * lra-spills.c (lra_final_code_change): Use delete_insn.
22777 * haifa-sched.c (sched_remove_insn): Likewise.
22778 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
22779 returning to the nop pool.
22780 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
22781 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
22782
22783 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22784
22785 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
22786 * doc/tm.texi: Regenerated.
22787
22788 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
22789
22790 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
22791 QImode checks.
22792
22793 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22794
22795 * df-core.c (df_find_def): Compare register numbers.
22796 (df_find_use): Likewise.
22797
22798 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
22799
22800 PR target/56903
22801 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
22802 lra_in_progress for return.
22803
22804 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22805
22806 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
22807 define_insn into define_insn_and_split and emit movsicc patterns.
22808
22809 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22810
22811 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
22812
22813 2013-04-12 Richard Biener <rguenther@suse.de>
22814
22815 * tree-pass.h (TODO_do_not_ggc_collect): New.
22816 * passes.c (execute_one_ipa_transform_pass): Honor
22817 TODO_do_not_ggc_collect.
22818 (execute_one_pass): Likewise.
22819
22820 Revert
22821 2013-04-10 Richard Biener <rguenther@suse.de>
22822
22823 * passes.c (init_optimization_passes): Remove reload pass.
22824 * ira.c (do_reload): Merge into ...
22825 (ira): ... this.
22826 (rest_of_handle_reload): Remove.
22827 (pass_reload): Likewise.
22828 * config/i386/i386.c (ix86_option_override): Refer to ira instead
22829 of reload for vzeroupper pass placement.
22830
22831 2013-04-12 Jakub Jelinek <jakub@redhat.com>
22832
22833 PR tree-optimization/56918
22834 PR tree-optimization/56920
22835 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
22836 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
22837 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
22838 use wide_mul_with_sign method.
22839
22840 2013-04-12 Richard Biener <rguenther@suse.de>
22841
22842 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
22843 not be considered a gimple constant.
22844
22845 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
22846
22847 * fold-const.c (const_binop): Handle vector shifts by a scalar.
22848 (fold_binary_loc): Call const_binop also for mixed vector-scalar
22849 operations.
22850
22851 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
22852 Jakub Jelinek <jakub@redhat.com>
22853
22854 * opts.c: Include diagnostic-color.h.
22855 (common_handle_option): Handle OPT_fdiagnostics_color_.
22856 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
22857 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
22858 (diagnostic-color.o): New.
22859 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
22860 (diagnostic_color_rule): New enum.
22861 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
22862 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
22863 the location string.
22864 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
22865 either NULL, or color kind.
22866 * diagnostic-color.c: New file.
22867 * diagnostic-color.h: New file.
22868 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
22869 arguments.
22870 * doc/invoke.texi (-fdiagnostics-color): Document.
22871 * pretty-print.h (pp_show_color): Define.
22872 (struct pretty_print_info): Add show_color field.
22873 * diagnostic.c: Include diagnostic-color.h.
22874 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
22875 macros. Colorize error:, warning: etc. strings and also the location
22876 string.
22877 (diagnostic_show_locus): Colorize the caret line.
22878 * pretty-print.c: Include diagnostic-color.h.
22879 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
22880 inside of %< %> quotes or quoted through q format modifier.
22881
22882 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22883
22884 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
22885
22886 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22887
22888 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
22889 code in CC_NZ mode.
22890 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
22891 pattern.
22892
22893 2013-04-11 Marek Polacek <polacek@redhat.com>
22894
22895 PR tree-optimization/48184
22896 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
22897
22898 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
22899
22900 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
22901 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
22902 (skip_simple_arithmetic): Tidy up.
22903 * tree.h (skip_simple_constant_arithmetic): Declare.
22904
22905 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22906
22907 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
22908
22909 2013-04-11 Richard Biener <rguenther@suse.de>
22910
22911 * tree-vect-loop.c (get_initial_def_for_induction): Properly
22912 generate vector constants.
22913
22914 2013-04-11 Richard Biener <rguenther@suse.de>
22915
22916 PR tree-optimization/56878
22917 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
22918 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
22919 New function.
22920 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
22921 Prefer to align the DR with the most invariant base address.
22922
22923 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
22924
22925 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
22926 comment.
22927
22928 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
22929
22930 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
22931 floating-point vector comparisons against 0.
22932
22933 2013-04-11 Jakub Jelinek <jakub@redhat.com>
22934
22935 PR tree-optimization/56899
22936 * fold-const.c (extract_muldiv_1): Apply distributive law
22937 only if TYPE_OVERFLOW_WRAPS (ctype).
22938
22939 2013-04-11 Bin Cheng <bin.cheng@arm.com>
22940
22941 PR target/56124
22942 * ira-costs.c (scan_one_insn): Check whether the source rtx of
22943 loading has side effect.
22944
22945 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
22946
22947 * config/sparc/sparc.c: Include tree-pass.h.
22948 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
22949 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
22950 head of file. Change return type. Split off gate function.
22951 (sparc_gate_work_around_errata): New function.
22952 (pass_work_around_errata): New pass definition.
22953 (insert_pass_work_around_errata) New pass insert definition to
22954 insert pass_work_around_errata just after delayed-branch scheduling.
22955 (sparc_option_override): Insert the pass.
22956 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
22957
22958 2013-04-10 David S. Miller <davem@davemloft.net>
22959
22960 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
22961 or -mcpu=hypersparc.
22962
22963 * target.def (cstore_mode): New hook.
22964 * target.h: Include insn-codes.h
22965 * targhooks.c: Likewise.
22966 (default_cstore_mode): New function.
22967 * targhooks.h: Declare it.
22968 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
22969 * doc/tm.texi: Rebuild.
22970 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
22971 target hook, rather than inspecting the insn_data.
22972 * config/sparc/sparc.c (sparc_cstore_mode): New function.
22973 (TARGET_CSTORE_MODE): Redefine.
22974 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
22975 result patterns.
22976 * config/sparc/predicates.md (cstore_result_operand): New special
22977 predicate.
22978 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
22979 Use it for operand 0.
22980 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
22981 (*snesi_special): Likewise.
22982 (*snesi_zero): Likewise.
22983 (*seqsi_zero): Likewise.
22984 (*sltu_insn): Likewise.
22985 (*sgeu_insn): Likewise.
22986 (*seqdi_special): Make operand 0 and comparison operation be of
22987 DImode.
22988 (*snedi_special): Likewise.
22989 (*snedi_special_vis3): Likewise.
22990 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
22991 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
22992 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
22993 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
22994 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
22995 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
22996 (*sltu_extend_sp64): Likewise.
22997 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
22998 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
22999 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
23000 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
23001 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
23002
23003 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
23004
23005 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
23006 (aarch64_start_file): Use the new function.
23007
23008 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23009 Jason Merrill <jason@redhat.com>
23010
23011 * common.opt: Add -gdwarf.
23012 * opts.c (common_handle_option): Handle it.
23013 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
23014
23015 2013-04-10 Richard Biener <rguenther@suse.de>
23016
23017 * passes.c (execute_todo): Do not call ggc_collect conditional here.
23018 (execute_one_ipa_transform_pass): But unconditionally here.
23019 (execute_one_pass): And here.
23020 (init_optimization_passes): Remove reload pass.
23021 * tree-pass.h (TODO_ggc_collect): Remove.
23022 (pass_reload): Likewise.
23023 * ira.c (do_reload): Merge into ...
23024 (ira): ... this.
23025 (rest_of_handle_reload): Remove.
23026 (pass_reload): Likewise.
23027 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23028 of reload for vzeroupper pass placement.
23029 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
23030 and todo_flags_finish of all passes.
23031
23032 2013-04-10 Richard Biener <rguenther@suse.de>
23033
23034 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
23035 first_const_oprnd field, rename first_def_type to first_op_type.
23036 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
23037 (vect_get_and_check_slp_defs): Always use the type of the
23038 operand. Allow mixed vect_external_def, vect_constant_def types.
23039 (vect_get_constant_vectors): Handle mixed vect_external_def,
23040 vect_constant_def types.
23041
23042 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
23043
23044 PR tree-optimization/55524
23045 * tree-ssa-math-opts.c
23046 (convert_mult_to_fma): Don't use an fms construct
23047 when we don't have an fms operation, but fnma, and it looks
23048 likely that we'll be able to use the latter.
23049
23050 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
23051
23052 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
23053 function.
23054 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
23055 inline fail caused by overwritable functions.
23056
23057 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
23058
23059 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
23060 unnecessary bits in the constant power of two case.
23061
23062 2013-04-10 Richard Biener <rguenther@suse.de>
23063
23064 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
23065 broken code swapping operands.
23066 (vect_build_slp_tree): Do not compute load permutations here.
23067 (vect_analyze_slp_instance): Compute load permutations here,
23068 after building the SLP tree.
23069
23070 2013-04-09 Christian Bruel <christian.bruel@st.com>
23071
23072 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
23073 of next/prev_real_insn.
23074
23075 2013-04-09 Jan Hubicka <jh@suse.cz>
23076
23077 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
23078 Drop aliased parameter.
23079 (function_and_variable_visibility): Do not handle alias pairs.
23080 * cgraph.c (varpool_externally_visible_p): Update prototype.
23081 * varpool.c (varpool_add_new_variable): Update.
23082
23083 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23084
23085 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
23086
23087 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
23088
23089 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
23090
23091 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
23092
23093 2013-04-09 Marek Polacek <polacek@redhat.com>
23094
23095 PR tree-optimization/48762
23096 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
23097
23098 2013-04-09 Richard Biener <rguenther@suse.de>
23099
23100 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
23101 dealing with cost.
23102 (vect_build_slp_tree): Likewise.
23103 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
23104 calculating the cost of a SLP instance.
23105 (vect_analyze_slp_instance): Use it from here, after building
23106 the SLP tree.
23107
23108 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23109
23110 PR middle-end/56883
23111 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
23112 expand_omp_for_static_chunk): Use simple_p = true in
23113 force_gimple_operand_gsi calls when assigning to addressable decls.
23114
23115 2013-04-09 Jeff Law <law@redhat.com>
23116
23117 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
23118 when the boolean was created by converting a wider object which
23119 had a boolean range.
23120
23121 2013-04-09 Richard Biener <rguenther@suse.de>
23122
23123 * tree-vectorizer.h (slp_void_p): Remove.
23124 (slp_tree): Typedef before _slp_tree declaration.
23125 (struct _slp_tree): Use a vector of slp_tree as children.
23126 (vect_get_place_in_interleaving_chain): Remove.
23127 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
23128 Move ...
23129 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
23130 and make static.
23131 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
23132 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
23133 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
23134 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
23135 Use slp_node instead of slp_void_p and adjust.
23136
23137 2013-04-09 Richard Biener <rguenther@suse.de>
23138
23139 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
23140 work that is not necessary.
23141
23142 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23143
23144 PR tree-optimization/56854
23145 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
23146 forward into clobber stmts if it would change MEM_REF lhs into
23147 non-MEM_REF.
23148
23149 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
23150
23151 * tree.c (type_hash_lookup, type_hash_add): Make static.
23152 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
23153
23154 2013-04-09 Richard Biener <rguenther@suse.de>
23155
23156 * tree.h (unsave_expr_now): Remove.
23157 * tree-inline.c (mark_local_for_remap_r): Remove.
23158 (unsave_expr_1): Likewise.
23159 (unsave_r): Likewise.
23160 (unsave_expr_now): Likewise.
23161 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
23162 (propagate_tree_value): Likewise.
23163
23164 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
23165
23166 * doc/rtl.texi (sequence): Rewrite documentation to match the
23167 current use of SEQUENCE rtl objects.
23168 * rtl.def (SEQUENCE): Likewise.
23169
23170 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
23171 Update documentation.
23172 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
23173 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
23174
23175 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
23176
23177 2013-04-08 Teresa Johnson <tejohnson@google.com>
23178
23179 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
23180 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
23181 methods.
23182 (estimate_edge_size_and_time): Add comment to suggest using rounding
23183 methods.
23184 (estimate_node_size_and_time): Ditto.
23185 (remap_edge_change_prob): Use helper rounding divide methods.
23186 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
23187 (gimple_mod_pow2_value_transform): Ditto.
23188 (gimple_mod_subtract_transform): Ditto.
23189 (gimple_ic_transform): Ditto.
23190 (gimple_stringops_transform): Ditto.
23191 * stmt.c (conditional_probability): Ditto.
23192 (emit_case_dispatch_table): Ditto.
23193 * lto-cgraph.c (merge_profile_summaries): Ditto.
23194 * tree-optimize.c (execute_fixup_cfg): Ditto.
23195 * cfgcleanup.c (try_forward_edges): Ditto.
23196 * cfgloopmanip.c (scale_loop_profile): Ditto.
23197 (loopify): Ditto.
23198 (duplicate_loop_to_header_edge): Ditto.
23199 (lv_adjust_loop_entry_edge): Ditto.
23200 * tree-vect-loop.c (vect_transform_loop): Ditto.
23201 * profile.c (compute_branch_probabilities): Ditto.
23202 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
23203 * lto-streamer-in.c (input_cfg): Ditto.
23204 * gimple-streamer-in.c (input_bb): Ditto.
23205 * ipa-cp.c (update_profiling_info): Ditto.
23206 (update_specialized_profile): Ditto.
23207 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
23208 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
23209 rounding methods.
23210 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23211 (compute_trg_info): Ditto.
23212 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
23213 (purge_dead_edges): Ditto.
23214 * loop-unswitch.c (unswitch_loop): Ditto.
23215 * cgraphclones.c (cgraph_clone_edge): Ditto.
23216 (cgraph_clone_node): Ditto.
23217 * tree-inline.c (copy_bb): Ditto.
23218 (copy_edges_for_bb): Ditto.
23219 (initialize_cfun): Ditto.
23220 (copy_cfg_body): Ditto.
23221 (expand_call_inline): Ditto.
23222
23223 2013-04-08 Kai Tietz <ktietz@redhat.com>
23224
23225 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
23226 TARGET_CYGWIN64 by TARGET_64BIT.
23227
23228 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23229
23230 * config/epiphany/epiphany.md (GPR_1): New constant.
23231 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
23232 * config/epiphany/epiphany.c (gen_compare_reg):
23233 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
23234 is already in place.
23235 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
23236 Don't require being called during rtl expansion; If y operlaps r0,
23237 return 0.
23238 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
23239 (epiphany_expand_epilogue): Likewise.
23240
23241 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
23242 Don't use CC_FPmode for ORDERED / UNORDERED.
23243 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
23244
23245 * config/epiphany/constraints.md (CnL): New constraint.
23246 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
23247 * config/epiphany/predicates.md (add_operand): Allow 1024.
23248
23249 * config/epiphany/epiphany.md (logical_op): New code iterator.
23250 (op_mnc): New code attribute.
23251 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
23252 (mov_f+1, mov_f+2): New peephole2 patterns.
23253
23254 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
23255 (cstoresi4): Also allow re-use of zero result when doing a NE
23256 comparison to a non-zero operand.
23257 Use (clobber (scratch)) for first insn if the gpr output is not needed.
23258
23259 * config/epiphany/epiphany.md (<insn_opname>v2si3):
23260 Use gen_addsi3_i / gen_subsi3_i.
23261
23262 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23263
23264 PR c++/34949
23265 PR c++/50243
23266 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
23267 contain anything but clobbers, at most one __builtin_stack_restore,
23268 optionally debug stmts and final resx, and if it has at least one
23269 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
23270 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
23271 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
23272 which isn't defaut definition, remove them.
23273 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
23274 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
23275 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
23276 with MEM_REF LHS with SSA_NAME address.
23277
23278 2013-04-08 Jeff Law <law@redhat.com>
23279
23280 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
23281
23282 2013-04-08 Richard Biener <rguenther@suse.de>
23283
23284 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
23285 extra newline.
23286 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
23287 determined vector type.
23288 (vect_analyze_data_refs): Likewise.
23289 (vect_get_new_vect_var): Adjust.
23290 (vect_create_destination_var): Preserve SSA name versions.
23291 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
23292 not dump anything here.
23293
23294 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23295
23296 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
23297 Add member lr_slot_known.
23298 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
23299 if necessary.
23300 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
23301 Remove code that sets lr_slot_offset according to what a previous
23302 version of epiphany_emit_save_restore used to do.
23303 (epiphany_emit_save_restore): When doing an lr save or restore,
23304 set/verify lr_slot_known and lr_slot_offset.
23305
23306 2013-04-08 Xinyu Qi <xyqi@marvell.com>
23307
23308 PR target/54338
23309 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
23310 in ALL_REGS.
23311
23312 2013-04-08 Richard Biener <rguenther@suse.de>
23313
23314 * alias.c (find_base_term): Fix thinko in previous change.
23315
23316 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23317
23318 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
23319 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
23320 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
23321 if possible to compute val.
23322 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
23323 For QImode integers don't require anything about precision. Use
23324 const_with_all_bytes_same to find out if the constant doesn't have
23325 repeated bytes in it.
23326
23327 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23328
23329 * config/s390/s390.c (s390_expand_insv): Only accept insertions
23330 within mode size.
23331
23332 2013-04-08 Marek Polacek <polacek@redhat.com>
23333
23334 PR rtl-optimization/48182
23335 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
23336 value to 1.
23337
23338 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23339
23340 PR target/55487
23341 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
23342 nuses, make sure we have a label.
23343
23344 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23345
23346 PR target/56843
23347 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
23348 (rs6000_emit_swdiv_low_precision): Remove.
23349 (rs6000_emit_swdiv): Rewrite to handle between one and four
23350 iterations of Newton-Raphson generally; modify required number of
23351 iterations for some cases.
23352 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
23353
23354 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
23355
23356 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
23357 set-but-unused variable.
23358
23359 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
23360 basic blocks of released function bodies garbage-collectable.
23361
23362 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
23363 (struct rtl_opt_pass): Add TODO_df_finish.
23364
23365 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
23366
23367 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23368
23369 * config/arm/constraints.md (q): New constraint.
23370 * config/arm/ldrdstrd.md: New file.
23371 * config/arm/arm.md (ldrdstrd.md) New include.
23372 (arm_movdi): Use "q" instead of "r" constraint
23373 for double-word memory access.
23374 (movdf_soft_insn): Likewise.
23375 * config/arm/vfp.md (movdi_vfp): Likewise.
23376 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
23377 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
23378 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
23379 (mem_ok_for_ldrd_strd): Likewise.
23380 (output_move_double): Update assertion.
23381
23382 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23383
23384 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
23385
23386 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23387
23388 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
23389 define_insn_and_split.
23390 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
23391
23392 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23393
23394 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
23395 define_insn_and_split.
23396 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
23397 (shiftsi3_compare): New pattern.
23398 (rrx): New pattern.
23399 * config/arm/unspecs.md (UNSPEC_RRX): New.
23400
23401 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23402
23403 * config/arm/arm.md (negdi_extendsidi): New pattern.
23404 (negdi_zero_extendsidi): Likewise.
23405
23406 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23407
23408 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
23409 define_insn_and_split.
23410 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
23411 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
23412
23413 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23414
23415 * config/arm/arm.md (arm_subdi3): Convert define_insn into
23416 define_insn_and_split.
23417 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
23418 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
23419
23420 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23421
23422 * config/arm/arm.md (subsi3_carryin): New pattern.
23423 (subsi3_carryin_const): Likewise.
23424 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
23425 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
23426
23427 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23428
23429 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
23430
23431 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23432
23433 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
23434 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
23435
23436 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23437
23438 * config/arm/arm.c (arm_expand_builtin): Change fcode
23439 type to unsigned int.
23440
23441 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23442
23443 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
23444
23445 2013-04-04 Ian Lance Taylor <iant@google.com>
23446
23447 * doc/standards.texi (Standards): The Go frontend supports the Go 1
23448 language standard.
23449
23450 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
23451
23452 PR middle-end/56729
23453 * df-scan.c (df_insn_delete): Disable failing assert.
23454
23455 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23456
23457 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
23458 New function prototype.
23459 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
23460 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
23461 (arm_builtin_vectorized_function): New function.
23462
23463 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23464
23465 * config/arm/arm_neon_builtins.def: New file.
23466 * config/arm/arm.c (neon_builtin_data): Move contents to
23467 arm_neon_builtins.def.
23468 (enum arm_builtins): Include neon builtin definitions.
23469 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
23470 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
23471
23472 2013-04-04 Marek Polacek <polacek@redhat.com>
23473
23474 PR tree-optimization/48186
23475 * predict.c (maybe_hot_frequency_p): Return false if
23476 HOT_BB_FREQUENCY_FRACTION is 0.
23477 (cgraph_maybe_hot_edge_p): Likewise.
23478
23479 2013-04-04 Richard Biener <rguenther@suse.de>
23480
23481 PR tree-optimization/56826
23482 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
23483 more accurately.
23484
23485 2013-04-04 Richard Biener <rguenther@suse.de>
23486
23487 PR tree-optimization/56213
23488 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
23489 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
23490
23491 2013-04-04 Richard Biener <rguenther@suse.de>
23492
23493 PR tree-optimization/56837
23494 * tree-loop-distribution.c (classify_partition): For non-zero
23495 values require that the value has the same precision as its
23496 mode to be useful as memset value.
23497
23498 2013-04-03 Nick Clifton <nickc@redhat.com>
23499
23500 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
23501 (fmssf4): Use fmsf.s on E3V5 architectures.
23502 (fnmasf4): Use fnmaf.s on E3V5 architectures.
23503 (fnmssf4): Use fnmsf.s on E3V5 architectures.
23504
23505 2013-04-03 Jeff Law <law@redhat.com>
23506
23507 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
23508 (lra-eliminations.o): Likewise.
23509
23510 2013-04-03 Teresa Johnson <tejohnson@google.com>
23511
23512 * gcov-io.c (compute_working_sets): Moved most of body of old
23513 compute_working_sets here from profile.c.
23514 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
23515 (gcov_working_set_t): Moved typedef here from basic-block.h
23516 (compute_working_set): Declare.
23517 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
23518 (get_working_sets): Renamed from compute_working_set,
23519 replace most of body with call to new compute_working_sets.
23520 (get_exec_counts): Replace call to compute_working_sets
23521 to get_working_sets.
23522 * profile.h (get_working_sets): Renamed from compute_working_set.
23523 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
23524 to get_working_sets.
23525 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
23526 * gcov-dump.c (dump_working_sets): New function.
23527
23528 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
23529
23530 * hwint.c (sext_hwi, zext_hwi): New functions.
23531 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
23532 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
23533 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
23534 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
23535 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
23536 (sext_hwi, zext_hwi): New functions.
23537
23538 2013-04-03 Jeff Law <law@redhat.com>
23539
23540 PR tree-optimization/56799
23541 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
23542 back test for widening conversion erroneously dropped in prior change.
23543
23544 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23545
23546 PR target/56809
23547 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
23548 instead of next_real_insn.
23549
23550 2013-04-03 Marek Polacek <polacek@redhat.com>
23551
23552 PR sanitizer/55702
23553 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
23554
23555 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23556
23557 PR target/56809
23558 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
23559 next_real_insn.
23560 (thumb1_output_casesi): Likewise.
23561 (thumb2_output_casesi): Likewise.
23562
23563 2013-04-03 Richard Biener <rguenther@suse.de>
23564
23565 PR tree-optimization/56817
23566 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
23567 Split out ...
23568 (tree_unroll_loops_completely_1): ... new function to manually
23569 walk the loop tree, properly defering outer loops of unrolled
23570 loops to later iterations.
23571
23572 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23573
23574 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
23575 (vectorizable_load): Likewise.
23576 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
23577 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
23578
23579 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23580
23581 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
23582 BIT_FIELD_REF.
23583
23584 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
23585
23586 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
23587
23588 2013-04-03 Bin Cheng <bin.cheng@arm.com>
23589
23590 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
23591
23592 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23593
23594 PR tree-optimization/56790
23595 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
23596 folding.
23597
23598 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23599
23600 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
23601 Handle VEC_MERGE.
23602 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
23603 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
23604 equal arguments.
23605
23606 2013-04-03 Jakub Jelinek <jakub@redhat.com>
23607
23608 PR c/19449
23609 * tree.h (force_folding_builtin_constant_p): New decl.
23610 * builtins.c (force_folding_builtin_constant_p): New variable.
23611 (fold_builtin_constant_p): Fold immediately also if
23612 force_folding_builtin_constant_p.
23613
23614 2013-04-03 Richard Biener <rguenther@suse.de>
23615
23616 PR tree-optimization/56812
23617 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
23618 DRs of the same interleaving chain are independent.
23619
23620 2013-04-02 Jason Merrill <jason@redhat.com>
23621
23622 * gdbinit.in (pbb): Use debug fn.
23623
23624 2013-04-02 Lawrence Crowl <crowl@google.com>
23625
23626 * sese.h (struct ivtype_map_elt_s): Remove unused.
23627 (extern debug_ivtype_map): Remove unused.
23628 (extern eq_ivtype_map_elts): Remove unused.
23629 * sese.c (debug_ivtype_map): Removed unused.
23630 (debug_ivtype_map_1): Removed unused.
23631 (debug_ivtype_elt): Remove unused.
23632 (eq_ivtype_map_elts): Remove unused.
23633
23634
23635 2013-04-02 Kai Tietz <ktietz@redhat.com>
23636
23637 PR target/52790
23638 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
23639 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
23640 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
23641 function.
23642 (legitimize_pe_coff_symbol): Likewise.
23643 (is_imported_p): New helper-function.
23644 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
23645 for Windows x64 targets.
23646 (ix86_expand_prologue): Optimize for pe-coff targets.
23647 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
23648 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
23649 medium/large code-model.
23650 (legitimize_pic_address): Likewise.
23651 (legitimize_tls_address): Likewise.
23652 (ix86_expand_call): Likewise.
23653 (x86_output_mi_thunk): Likewise.
23654 (get_dllimport_decl): Add new beimport argument.
23655 (construct_plt_address): Don't assert for x64 pe-coff targets.
23656 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
23657 targets.
23658 (SYMBOL_FLAG_STUBVAR): New macro.
23659 (SYMBOL_REF_STUBVAR_P): Likewise.
23660 * config/i386/winnt.c (stub_list): New structure.
23661 (stub_head): New local variable.
23662 (i386_pe_record_stub): New function.
23663 (i386_pe_file_end): Emit refptr-stubs.
23664
23665 2013-04-02 Jakub Jelinek <jakub@redhat.com>
23666
23667 PR rtl-optimization/56745
23668 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
23669 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
23670
23671 PR c++/34949
23672 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
23673 and both of them are MEM_REFs, just compare first argument for
23674 equality and attempt to deal even with differing offsets.
23675
23676 PR c++/34949
23677 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
23678 of gimple_clobber_p to be MEM_REF.
23679 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
23680 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
23681 after gimplification.
23682 * asan.c (get_mem_ref_of_assignment): Don't instrument
23683 gimple_clobber_p stmts.
23684 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
23685 gimple_clobber_p stmt if they have MEM_REF lhs and
23686 are dead because of another gimple_clobber_p stmt.
23687 * tree-ssa-live.c (clear_unused_block_pointer): Treat
23688 gimple_clobber_p stmts like debug stmts.
23689 (remove_unused_locals): Remove clobbers with MEM_REF lhs
23690 that refer to unused VAR_DECLs or uninitialized values.
23691 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
23692 gimple_clobber_p stmts if they refer to removed parameters.
23693 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
23694 formatting.
23695
23696 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
23697
23698 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
23699 using SWI48 mode attribute.
23700
23701 2013-04-02 Wei Mi <wmi@google.com>
23702
23703 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
23704 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
23705 *<rotate_insn><mode>3_mask in i386.md.
23706
23707 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
23708
23709 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
23710
23711 2013-04-02 Richard Biener <rguenther@suse.de>
23712
23713 PR tree-optimization/56778
23714 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
23715 Runtime alias tests are not supported for gather loads.
23716 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
23717 stmts referenced from SSA operands before updating SSA form.
23718
23719 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
23720 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23721
23722 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
23723 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
23724 * config/arm/cortex-a53.md: New file.
23725 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
23726 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
23727 * config/arm/arm.c (arm_issue_rate): Likewise.
23728 * config/arm/arm-tune.md: Regenerate
23729 * config/arm/arm-tables.opt: Regenerate.
23730 * config/arm/arm-cores.def: Add cortex-a53.
23731
23732 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
23733
23734 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
23735 non-static link.
23736
23737 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23738
23739 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
23740 scalar load/store operations using B/H registers.
23741 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
23742
23743 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23744
23745 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
23746 scalar move.
23747 * config/aarch64/aarch64.c
23748 (aarch64_simd_scalar_immediate_valid_for_move): New.
23749 * config/aarch64/aarch64-protos.h
23750 (aarch64_simd_scalar_immediate_valid_for_move): New.
23751 * config/aarch64/constraints.md (Dh, Dq): New.
23752 * config/aarch64/iterators.md (hq): New.
23753
23754 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
23755
23756 * reorg.c (get_branch_condition): Deal with conditional returns.
23757 (fill_simple_delay_slots): Remove dead code dealing with jumps.
23758
23759 2013-04-01 Wei Mi <wmi@google.com>
23760
23761 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
23762 Truncate operand 2 using %b asm operand modifier.
23763 (*<shift_insn><mode>3_mask): Ditto.
23764 (*<rotate_insn><mode>3_mask): Ditto.
23765
23766 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
23767
23768 PR middle-end/56798
23769 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
23770
23771 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
23772
23773 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
23774 of next_real_insn.
23775 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
23776
23777 2013-03-30 Lawrence Crowl <crowl@google.com>
23778
23779 * dse.c (clear_alias_sets): Remove never set.
23780 (disqualified_clear_alias_sets): Remove never set.
23781 (clear_alias_mode_pool): Remove never set.
23782 (dse_step0): Remove condition that is never true.
23783 (canon_address): Remove condition that is never true.
23784 (dse_step7): Remove condition that is never true.
23785 (rest_of_handle_dse): Remove condition that is never true.
23786 (rest_of_handle_dse::did_global): Remove never read from above.
23787 (dse_step2_spill): Remove never called from above.
23788 (dse_step5_spill): Remove never called from above.
23789
23790 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
23791
23792 * doc/md.texi (Standard Names) <casesi>: Update documentation for
23793 JUMP_TABLE_DATA changes.
23794 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
23795 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
23796 (Insns) <jump_table_data>: New entry.
23797 * doc/tm.texi: Regenerate.
23798
23799 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
23800
23801 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
23802 for table jump at the end of a basic block using tablejump_p.
23803 * targhooks.c (default_invalid_within_doloop): Likewise.
23804 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
23805 target hook implementation that is identical to the default hook.
23806 (rs6000_invalid_within_doloop): Remove.
23807
23808 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
23809 unused variable from tablejump_p call.
23810
23811 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
23812 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
23813 (INSN_DELETED_P): Likewise.
23814 (emit_jump_table_data): New prototype.
23815 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
23816 after 4th as unused.
23817 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
23818 * sched-vis.c (print_insn): Likewise.
23819 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
23820 insn for compatibility with back ends that use next_active_insn to
23821 identify jump table data.
23822 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
23823 (remove_insn): Likewise.
23824 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
23825 to be emitted.
23826 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
23827 (emit_jump_table_data): New function.
23828
23829 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
23830 basic block, a JUMP_TABLE_DATA never is.
23831 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
23832 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
23833 off from code handling real insns.
23834 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
23835 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
23836 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
23837 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
23838 is not a NONDEBUG_INSN_P.
23839 * ira-costs.c (scan_one_insn): Likewise.
23840 * jump.c (mark_all_labels): Likewise.
23841 (mark_jump_label_1): Likewise.
23842 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
23843 * lra.c (get_insn_freq): Expect all insns reaching here to be in
23844 a basic block.
23845 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
23846 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
23847 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
23848 JUMP_TABLE_DATA_P insns.
23849 (calculate_elim_costs_all_insns): Likewise.
23850 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
23851 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
23852 (delete_output_reload): Code style fixups.
23853 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
23854 insn flags on this non-insn.
23855 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
23856 as scheduling barriers, for pre-change compatibility.
23857 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
23858 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
23859
23860 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
23861 redundant JUMP_TABLE_DATA_P test.
23862 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
23863 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
23864 (frv_for_each_packet): Likewise.
23865 * config/i386/i386.c (min_insn_size): Likewise.
23866 (ix86_avoid_jump_mispredicts): Likewise.
23867 * config/m32r/m32r.c (m32r_is_insn): Likewise.
23868 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
23869 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
23870 (mips16_insn_length): Robustify.
23871 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
23872 (mips16_split_long_branches): Likewise.
23873 * config/pa/pa.c (pa_combine_instructions): Likewise.
23874 * config/rs6000/rs6000.c (get_next_active_insn): Treat
23875 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
23876 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
23877 as contributing to pool range lengths.
23878 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
23879 Remove redundant JUMP_TABLE_DATA_P test.
23880 (sh_loop_align): Likewise.
23881 (split_branches): Likewise.
23882 (sh_insn_length_adjustment): Likewise.
23883 * config/spu/spu.c (get_branch_target): Likewise.
23884
23885 2013-03-29 Jan Hubicka <jh@suse.cz>
23886
23887 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
23888 gcov streaming; stream hot bb threshold to ltrans.
23889 * predict.c (get_hot_bb_threshold): Break out from ....
23890 (maybe_hot_count_p): ... here.
23891 (set_hot_bb_threshold): New function.
23892 * lto-section-in.c (lto_section_name): Add profile.
23893 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
23894 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
23895 and data-streamer.h
23896 (histogram_entry): New structure.
23897 (histogram, histogram_pool): New global vars.
23898 (histogram_hash): New structure.
23899 (histogram_hash::hash): New method.
23900 (histogram_hash::equal): Likewise.
23901 (account_time_size): New function.
23902 (cmp_counts): New function.
23903 (dump_histogram): New function.
23904 (ipa_profile_generate_summary): New function.
23905 (ipa_profile_write_summary): New function.
23906 (ipa_profile_read_summary): New function.
23907 (ipa_profile): Decide on threshold.
23908 (pass_ipa_profile): Add ipa_profile_write_summary and
23909 ipa_profile_read_summary.
23910 * Makefile.in (ipa.o): Update dependencies.
23911 * lto-streamer.h (LTO_section_ipa_profile): New section.
23912
23913 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
23914
23915 * tree.h (VAR_P): New.
23916
23917 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
23918
23919 PR lto/56777
23920 * doc/invoke.texi ([-fwhole-program]): Fix typo.
23921
23922 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
23923
23924 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
23925 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
23926 (control_flow_insn_p): Likewise.
23927 * cfgrtl.c (duplicate_insn_chain): Likewise.
23928 * final.c (get_attr_length_1): Likewise.
23929 (shorten_branches): Likewise.
23930 (final_scan_insn): Likewise.
23931 * function.c (instantiate_virtual_regs): Likewise.
23932 * gcse.c (insert_insn_end_basic_block): Likewise.
23933 * ira-costs.c (scan_one_insn): Likewise.
23934 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
23935 * lra.c (check_rtl): Likewise.
23936 * reload1.c (elimination_costs_in_insn): Likewise.
23937 * reorg.c (follow_jumps): Likewise.
23938
23939 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
23940 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
23941 (thumb_far_jump_used_p): Likewise.
23942 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
23943 (workaround_speculation): Likewise.
23944 (add_sched_insns_for_speculation): Likewise.
23945 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
23946 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
23947 (frv_for_each_packet): Likewise.
23948 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
23949 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
23950 (final_emit_insn_group_barriers): Likewise.
23951 * config/m32r/m32r.c (m32r_is_insn): Likewise.
23952 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
23953 (mips16_insn_length): Likewise.
23954 * config/pa/pa.c (pa_reorg): Likewise.
23955 (pa_combine_instructions): Likewise.
23956 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
23957 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
23958 (sh_reorg): Likewise.
23959 (split_branches): Likewise.
23960 * config/spu/spu.c (get_branch_target): Likewise.
23961
23962 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
23963 JUMP_TABLE_DATA_P.
23964
23965 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
23966
23967 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
23968 Fix declaration name.
23969
23970 2013-03-28 Lawrence Crowl <crowl@google.com>
23971
23972 * graphds.h (struct graph.indicies): Remove unused.
23973 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
23974 (SCOP_ORIGINAL_PDDRS): Remove unused.
23975 * sese.h (extern insert_loop_close_phis): Removed unused.
23976 (extern insert_guard_phis): Removed unused.
23977 (extern ivtype_map_elt_info): Removed unused.
23978 (new_ivtype_map_elt): Removed unused.
23979 * sese.c (ivtype_map_elt_info): Removed unused.
23980
23981 2013-03-28 Lawrence Crowl <crowl@google.com>
23982
23983 * Makefile.in: Add several missing include dependences.
23984 (DUMPFILE_H): New.
23985 (test-dump.o): New. This object is not added to any executable,
23986 but is present for ad-hoc testing.
23987 * bitmap.c
23988 (debug (const bitmap_head_def &)): New.
23989 (debug (const bitmap_head_def *)): New.
23990 * bitmap.h
23991 (extern debug (const bitmap_head_def &)): New.
23992 (extern debug (const bitmap_head_def *)): New.
23993 * cfg.c
23994 (debug (edge_def &)): New.
23995 (debug (edge_def *)): New.
23996 * cfghooks.c
23997 (debug (basic_block_def &)): New.
23998 (debug (basic_block_def *)): New.
23999 * dumpfile.h
24000 (dump_node (const_tree, int, FILE *)): Correct source file.
24001 * dwarf2out.c
24002 (debug (die_struct &)): New.
24003 (debug (die_struct *)): New.
24004 * dwarf2out.h
24005 (extern debug (die_struct &)): New.
24006 (extern debug (die_struct *)): New.
24007 * gimple-pretty-print.c
24008 (debug (gimple_statement_d &)): New.
24009 (debug (gimple_statement_d *)): New.
24010 * gimple-pretty-print.h
24011 (extern debug (gimple_statement_d &)): New.
24012 (extern debug (gimple_statement_d *)): New.
24013 * ira-build.c
24014 (debug (ira_allocno_copy &)): New.
24015 (debug (ira_allocno_copy *)): New.
24016 (debug (ira_allocno &)): New.
24017 (debug (ira_allocno *)): New.
24018 * ira-int.h
24019 (extern debug (ira_allocno_copy &)): New.
24020 (extern debug (ira_allocno_copy *)): New.
24021 (extern debug (ira_allocno &)): New.
24022 (extern debug (ira_allocno *)): New.
24023 * ira-lives.c
24024 (debug (live_range &)): New.
24025 (debug (live_range *)): New.
24026 * lra-int.h
24027 (debug (lra_live_range &)): New.
24028 (debug (lra_live_range *)): New.
24029 * lra-lives.c
24030 (debug (lra_live_range &)): New.
24031 (debug (lra_live_range *)): New.
24032 * omega.c
24033 (debug (omega_pb_d &)): New.
24034 (debug (omega_pb_d *)): New.
24035 * omega.h
24036 (extern debug (omega_pb_d &)): New.
24037 (extern debug (omega_pb_d *)): New.
24038 * print-rtl.c
24039 (debug (const rtx_def &)): New.
24040 (debug (const rtx_def *)): New.
24041 * print-tree.c
24042 (debug_tree (tree): Move within file.
24043 (debug_raw (const tree_node &)): New.
24044 (debug_raw (const tree_node *)): New.
24045 (dump_tree_via_hooks (const tree_node *, int)): New.
24046 (debug (const tree_node &)): New.
24047 (debug (const tree_node *)): New.
24048 (debug_verbose (const tree_node &)): New.
24049 (debug_verbose (const tree_node *)): New.
24050 (debug_head (const tree_node &)): New.
24051 (debug_head (const tree_node *)): New.
24052 (debug_body (const tree_node &)): New.
24053 (debug_body (const tree_node *)): New.
24054 (debug_vec_tree (tree): Move and reimplement in terms of dump.
24055 (debug (vec<tree, va_gc> &)): New.
24056 (debug (vec<tree, va_gc> *)): New.
24057 * rtl.h
24058 (extern debug (const rtx_def &)): New.
24059 (extern debug (const rtx_def *)): New.
24060 * sbitmap.c
24061 (debug_raw (simple_bitmap_def &)): New.
24062 (debug_raw (simple_bitmap_def *)): New.
24063 (debug (simple_bitmap_def &)): New.
24064 (debug (simple_bitmap_def *)): New.
24065 * sbitmap.h
24066 (extern debug (simple_bitmap_def &)): New.
24067 (extern debug (simple_bitmap_def *)): New.
24068 (extern debug_raw (simple_bitmap_def &)): New.
24069 (extern debug_raw (simple_bitmap_def *)): New.
24070 * sel-sched-dump.c
24071 (debug (vinsn_def &)): New.
24072 (debug (vinsn_def *)): New.
24073 (debug_verbose (vinsn_def &)): New.
24074 (debug_verbose (vinsn_def *)): New.
24075 (debug (expr_def &)): New.
24076 (debug (expr_def *)): New.
24077 (debug_verbose (expr_def &)): New.
24078 (debug_verbose (expr_def *)): New.
24079 (debug (vec<rtx> &)): New.
24080 (debug (vec<rtx> *)): New.
24081 * sel-sched-dump.h
24082 (extern debug (vinsn_def &)): New.
24083 (extern debug (vinsn_def *)): New.
24084 (extern debug_verbose (vinsn_def &)): New.
24085 (extern debug_verbose (vinsn_def *)): New.
24086 (extern debug (expr_def &)): New.
24087 (extern debug (expr_def *)): New.
24088 (extern debug_verbose (expr_def &)): New.
24089 (extern debug_verbose (expr_def *)): New.
24090 (extern debug (vec<rtx> &)): New.
24091 (extern debug (vec<rtx> *)): New.
24092 * sel-sched-ir.h
24093 (_list_iter_cond_expr): Make inline instead of static.
24094 * sreal.c
24095 (debug (sreal &)): New.
24096 (debug (sreal *)): New.
24097 * sreal.h
24098 (extern debug (sreal &)): New.
24099 (extern debug (sreal *)): New.
24100 * tree.h
24101 (extern debug_raw (const tree_node &)): New.
24102 (extern debug_raw (const tree_node *)): New.
24103 (extern debug (const tree_node &)): New.
24104 (extern debug (const tree_node *)): New.
24105 (extern debug_verbose (const tree_node &)): New.
24106 (extern debug_verbose (const tree_node *)): New.
24107 (extern debug_head (const tree_node &)): New.
24108 (extern debug_head (const tree_node *)): New.
24109 (extern debug_body (const tree_node &)): New.
24110 (extern debug_body (const tree_node *)): New.
24111 (extern debug (vec<tree, va_gc> &)): New.
24112 (extern debug (vec<tree, va_gc> *)): New.
24113 * tree-cfg.c
24114 (debug (struct loop &)): New.
24115 (debug (struct loop *)): New.
24116 (debug_verbose (struct loop &)): New.
24117 (debug_verbose (struct loop *)): New.
24118 * tree-dump.c: Add header dependence.
24119 * tree-flow.h
24120 (extern debug (struct loop &)): New.
24121 (extern debug (struct loop *)): New.
24122 (extern debug_verbose (struct loop &)): New.
24123 (extern debug_verbose (struct loop *)): New.
24124 * tree-data-ref.c
24125 (debug (data_reference &)): New.
24126 (debug (data_reference *)): New.
24127 (debug (vec<data_reference_p> &)): New.
24128 (debug (vec<data_reference_p> *)): New.
24129 (debug (vec<ddr_p> &)): New.
24130 (debug (vec<ddr_p> *)): New.
24131 * tree-data-ref.h
24132 (extern debug (data_reference &)): New.
24133 (extern debug (data_reference *)): New.
24134 (extern debug (vec<data_reference_p> &)): New.
24135 (extern debug (vec<data_reference_p> *)): New.
24136 (extern debug (vec<ddr_p> &)): New.
24137 (extern debug (vec<ddr_p> *)): New.
24138 * tree-ssa-alias.c
24139 (debug (pt_solution &)): New.
24140 (debug (pt_solution *)): New.
24141 * tree-ssa-alias.h
24142 (extern debug (pt_solution &)): New.
24143 (extern debug (pt_solution *)): New.
24144 * tree-ssa-alias.c
24145 (debug (_var_map &)): New.
24146 (debug (_var_map *)): New.
24147 (debug (tree_live_info_d &)): New.
24148 (debug (tree_live_info_d *)): New.
24149 * tree-ssa-alias.h
24150 (extern debug (_var_map &)): New.
24151 (extern debug (_var_map *)): New.
24152 (extern debug (tree_live_info_d &)): New.
24153 (extern debug (tree_live_info_d *)): New.
24154
24155 2013-03-28 Jan Hubicka <jh@suse.cz>
24156
24157 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
24158
24159 2013-03-28 Ian Bolton <ian.bolton@arm.com>
24160
24161 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
24162 record only when desired or required.
24163
24164 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
24165
24166 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
24167 *vec_extractv2di_1_rex64. Use x64 isa attribute.
24168
24169 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24170
24171 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
24172 (*andsi3_compare0_uxtw): New pattern.
24173 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
24174 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
24175
24176 2013-03-28 Jan Hubicka <jh@suse.cz>
24177
24178 * data-streamer-in.c (streamer_read_gcov_count): New function.
24179 * gimple-streamer-out.c: Include value-prof.h.
24180 (output_gimple_stmt): Output histogram.
24181 (output_bb): Use streamer_write_gcov_count.
24182 * value-prof.c: Include data-streamer.h
24183 (dump_histogram_value): Add HIST_TYPE_MAX.
24184 (stream_out_histogram_value): New function.
24185 (stream_in_histogram_value): New function.
24186 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
24187 (stream_out_histogram_value, stream_in_histogram_value): Declare.
24188 * data-streamer-out.c (streamer_write_gcov_count): New function.
24189 (streamer_write_gcov_count_stream): New function.
24190 * lto-cgraph.c (lto_output_edge): Update counter streaming.
24191 (lto_output_node): Likewise.
24192 (input_node, input_edge): Likewise.
24193 * lto-streamer-out.c (output_cfg): Update streaming.
24194 * lto-streamer-in.c (input_cfg): Likewise.
24195 * data-streamer.h (streamer_write_gcov_count,
24196 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
24197 * gimple-streamer-in.c: Include value-prof.h
24198 (input_gimple_stmt): Input histograms.
24199 (input_bb): Update profile streaming.
24200
24201 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24202
24203 * genmodes.c (emit_max_int): New function.
24204 (emit_insn_modes_h): Added call to emit_max_function.
24205 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
24206 Added doc.
24207 * machmode.def: Fixed comment.
24208
24209 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24210
24211 * combine.c (try_combine): Removed useless assert.
24212 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
24213
24214 2013-03-28 Marek Polacek <polacek@redhat.com>
24215 Richard Biener <rguenther@suse.de>
24216
24217 PR tree-optimization/56695
24218 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
24219 build signed result of a vector comparison.
24220 * tree-cfg.c (verify_gimple_comparison): Check that a result
24221 of a vector comparison has signed type.
24222
24223 2013-03-28 Richard Biener <rguenther@suse.de>
24224
24225 PR tree-optimization/37021
24226 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
24227 do not restrict gaps between groups.
24228 * tree-vect-stmts.c (vectorizable_load): Properly account for
24229 a gap between groups.
24230
24231 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
24232
24233 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
24234 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
24235 is not enabled.
24236
24237 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
24238
24239 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
24240 * doc/extend.texi (Named Address Spaces): Ditto.
24241 (Variable Attributes): Ditto.
24242
24243 2013-03-27 Kai Tietz <ktietz@redhat.com>
24244
24245 * config.build: Add support for cygwin x64 target.
24246 * config.gcc: Likewise.
24247 * config.host: Likewise.
24248 * configure.ac: Likewise
24249 * configure: Regenerated.
24250
24251 2013-03-27 Kai Tietz <ktietz@redhat.com>
24252
24253 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
24254 * config/i386/t-cygwin-w64: New file.
24255 * config/i386/cygwin-w64.h: New file.
24256 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
24257 and add support for x64-cygwin target.
24258 (CPP_SPEC): Likewise.
24259 (CXX_WRAP_SPEC_LIST): Undefine before define.
24260 (LIBGCJ_SONAME): Use 15 as version.
24261
24262 2013-03-27 Richard Biener <rguenther@suse.de>
24263
24264 PR tree-optimization/56716
24265 * tree-ssa-structalias.c (perform_var_substitution): Adjust
24266 dumping for ref nodes.
24267
24268 2013-03-27 Martin Jambor <mjambor@suse.cz>
24269
24270 PR tree-optimization/55334
24271 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
24272 restricted pointers to arrays.
24273
24274 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
24275
24276 * Makefile.in (.SUFFIXES): Add .cc.
24277 (.c.o): Apply same recipe for implicit rule .cc.o.
24278
24279 2013-03-27 Richard Biener <rguenther@suse.de>
24280
24281 PR tree-optimization/37021
24282 * tree-vect-data-refs.c (vect_check_strided_load): Allow
24283 REALPART/IMAGPART_EXPRs around the supported refs.
24284 * tree-ssa-structalias.c (find_func_aliases): Assume that
24285 floating-point values are not used to transfer pointers.
24286
24287 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
24288
24289 * target.def (TARGET_HAS_IFUNC_P): New target hook.
24290 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
24291 * doc/tm.texi: Regenerate.
24292 * targhooks.h (default_has_ifunc_p): New.
24293 * targhooks.c (default_has_ifunc_p): Ditto.
24294 * config/linux-protos.h: New file.
24295 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
24296 hook for linux which disables support of indirect functions in android.
24297 * config/linux-android.c: New file.
24298 * config/t-linux-android.c: Ditto.
24299 * config.gcc: Added new object file linux-android.o.
24300 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
24301 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
24302 * varasm.c (do_assemble_alias): Likewise.
24303 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
24304 doesn't support indirect functions.
24305 * configure: Regenerate.
24306
24307 2013-03-27 Bin Cheng <bin.cheng@arm.com>
24308
24309 PR target/56102
24310 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
24311 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
24312 mult-word mode.
24313
24314 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24315
24316 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
24317
24318 2013-03-27 Terry Guo <terry.guo@arm.com>
24319
24320 * config/arm/arm-cores.def: Added core cortex-r7.
24321 * config/arm/arm-tune.md: Regenerated.
24322 * config/arm/arm-tables.opt: Regenerated.
24323 * doc/invoke.texi: Added entry for core cortex-r7.
24324
24325 2013-03-27 Walter Lee <walt@tilera.com>
24326
24327 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
24328 double-decrement of next_scratch_regno.
24329
24330 2013-03-27 Walter Lee <walt@tilera.com>
24331
24332 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
24333 input operands.
24334 (insn_v1mulus): Ditto.
24335 (insn_v2muls): Ditto.
24336
24337 2013-03-27 Walter Lee <walt@tilera.com>
24338
24339 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
24340 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
24341
24342 2013-03-27 Walter Lee <walt@tilera.com>
24343
24344 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
24345 (*sibcall_value): Ditto.
24346
24347 2013-03-27 Walter Lee <walt@tilera.com>
24348
24349 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
24350 (insn_mnz_v8qi): ... this ...
24351 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
24352 vector equivalent.
24353 (insn_v<n>mnz): Replaced by ...
24354 (insn_v1mnz): ... this ...
24355 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
24356 equivalent.
24357 (insn_mz_<mode>): Replaced by ...
24358 (insn_mz_v8qi): ... this ...
24359 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
24360 vector equivalent.
24361 (insn_v<n>mz): Replaced by ...
24362 (insn_v1mz): ... this ...
24363 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
24364 equivalent.
24365
24366 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
24367
24368 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
24369
24370 2013-03-26 Roland McGrath <mcgrathr@google.com>
24371
24372 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
24373 than fprintf with a non-constant, non-format string.
24374
24375 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24376
24377 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
24378 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
24379 operand 0 predicate.
24380 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
24381 attribute. Use general_x64nomem_operand as operand 1 predicate.
24382 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
24383 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24384 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
24385 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24386 (mov<mode>_insv_1): Remove expander. Merge insn with
24387 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
24388 Use general_x64nomem_operand as operand 1 predicate.
24389 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
24390 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
24391 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24392 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
24393 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
24394 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
24395 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24396 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
24397 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
24398 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
24399 (general_x64nomem_operand): Ditto.
24400
24401 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
24402
24403 * config/rtems.opt: Add -pthread option.
24404
24405 2013-03-26 Richard Biener <rguenther@suse.de>
24406
24407 * alias.c (find_base_term): Avoid redundant and not used recursion.
24408 (base_alias_check): Get the initial base term from the caller.
24409 (true_dependence_1): Compute and pass base terms to base_alias_check.
24410 (write_dependence_p): Likewise.
24411 (may_alias_p): Likewise.
24412
24413 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
24414
24415 * config/aarch64/aarch64.c (aarch64_classify_address): Support
24416 PC-relative load in SI modes and above only.
24417
24418 2013-03-26 Xinyu Qi <xyqi@marvell.com>
24419
24420 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
24421 * config/arm/iwmmxt.md (WCGR0): Update.
24422 (WCGR1, WCGR2, WCGR3): Likewise.
24423
24424 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24425
24426 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
24427 Use x64 and nox64 isa attributes.
24428
24429 2013-03-26 Richard Biener <rguenther@suse.de>
24430
24431 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
24432 alignment computations and rely on get_object_alignment_1
24433 for the !TYPE_P case.
24434 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
24435
24436 2013-03-26 Walter Lee <walt@tilera.com>
24437
24438 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
24439 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
24440
24441 2013-03-25 Jeff Law <law@redhat.com>
24442
24443 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
24444 check for INTEGRAL_TYPE_P that was missing due to checking in
24445 wrong version of prior patch.
24446
24447 2013-03-25 Walter Lee <walt@tilera.com>
24448
24449 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
24450 TILEGX_INSN_SHUFFLEBYTES1.
24451 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
24452 shufflebytes1.
24453 (tilegx_builtins): Ditto.
24454 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
24455
24456 2013-03-25 Walter Lee <walt@tilera.com>
24457
24458 * config/tilegx/tilegx.md (floatsisf2): New pattern.
24459 (floatunssisf2): New pattern.
24460 (floatsidf2): New pattern.
24461 (floatunssidf2): New pattern.
24462
24463 2013-03-25 Walter Lee <walt@tilera.com>
24464
24465 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
24466 tests for constraint J, K, N, P.
24467
24468 2013-03-25 Walter Lee <walt@tilera.com>
24469
24470 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
24471 Use indirect/pcrel encoding.
24472 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
24473 Ditto.
24474
24475 2013-03-25 Steve Ellcey <sellcey@mips.com>
24476
24477 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
24478 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
24479 * config/mips/mips.c (mips_option_override): Set IMADD default.
24480 * config/mips/mips.h (PTF_AVOID_IMADD): New.
24481 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
24482 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
24483 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
24484 * doc/invoke.texi (-mimadd/-mno-imadd): New.
24485
24486 2013-03-25 Jeff Law <law@redhat.com>
24487
24488 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
24489 slightly to avoid creating and folding useless trees. Simplify
24490 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
24491
24492 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
24493
24494 * config/i386/i386.md (*zero_extendsidi2): Merge with
24495 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
24496 * config/i386/predicates.md (x86_64_zext_operand): Rename from
24497 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
24498 targets. Clarify comment.
24499
24500 2013-03-25 Martin Jambor <mjambor@suse.cz>
24501
24502 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
24503 pass-through jump functions differently.
24504 (ipa_read_jump_function): Likewise. Also use setter functions to set
24505 up jump functions.
24506
24507 2013-03-25 Martin Jambor <mjambor@suse.cz>
24508
24509 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
24510 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
24511 process it.
24512 (ipa_get_indirect_edge_target): New function.
24513 (devirtualization_time_bonus): New parameter known_aggs, pass it to
24514 ipa_get_indirect_edge_target. Update all callers.
24515 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
24516 ipa_get_indirect_edge_target_1 instead of calling
24517 ipa_get_indirect_edge_target.
24518 (create_specialized_node): Pass aggvlas to
24519 ipcp_discover_new_direct_edges.
24520
24521 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24522
24523 * config/arm/arm.md (f_sels, f_seld): New types.
24524 (*cmov<mode>): New pattern.
24525 * config/arm/predicates.md (arm_vsel_comparison_operator): New
24526 predicate.
24527
24528 2013-03-25 Kai Tietz <ktietz@redhat.com>
24529
24530 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
24531 POSIX-printf for mingw-hosted builds.
24532
24533 2013-03-25 Richard Biener <rguenther@suse.de>
24534
24535 PR middle-end/56694
24536 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
24537 must-not-throw stmt location.
24538
24539 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24540
24541 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
24542 Emit load-acquire versions when acq is true.
24543 (arm_emit_store_exclusive): Add rel parameter.
24544 Emit store-release versions when rel is true.
24545 (arm_split_compare_and_swap): Use acquire-release instructions
24546 instead.
24547 of barriers when appropriate.
24548 (arm_split_atomic_op): Likewise.
24549 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
24550 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
24551 (VUNSPEC_SLX): Likewise.
24552 (VUNSPEC_LDA): Likewise.
24553 (VUNSPEC_STL): Likewise.
24554 * config/arm/sync.md (atomic_load<mode>): New pattern.
24555 (atomic_store<mode>): Likewise.
24556 (arm_load_acquire_exclusive<mode>): Likewise.
24557 (arm_load_acquire_exclusivesi): Likewise.
24558 (arm_load_acquire_exclusivedi): Likewise.
24559 (arm_store_release_exclusive<mode>): Likewise.
24560
24561 2013-03-25 Catherine Moore <clm@codesourcery.com>
24562
24563 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
24564 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
24565 * config/mip/predicates.md (lwsp_swsp_operand,
24566 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
24567 sb16_operand, db4_operand, db7_operand, ib3_operand,
24568 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
24569 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
24570 andi16_operand): New predicates.
24571 * config/mips/mips.md (compression): New attribute.
24572 (enabled): New attribute.
24573 (length): Consider compression in computing length.
24574 (shift_compression): New code attribute.
24575 (*add<mode>3): New operands. Record compression.
24576 (sub<mode>3): Likewise.
24577 (one_cmpl<mode>2): Likewise.
24578 (*and<mode>3): Likewise.
24579 (*ior<mode>3): Likewise.
24580 (unnamed pattern for xor): Likewise.
24581 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
24582 (*<optab><mode>3): Likewise.
24583 (*mov<mode>_internal: Likewise.
24584 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
24585 (mips_unsigned_immediate_p): New.
24586 (umips_lwsp_swsp_address_p): New.
24587 (m16_based_address_p): New.
24588 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
24589 (mips_unsigned_immediate_p): New prototype.
24590 (lwsp_swsp_address_p): New prototype.
24591 (m16_based_address_p): New prototype.
24592 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
24593 (mips_signed_immediate_p): New function.
24594 (m16_based_address_p): New function.
24595 (lwsp_swsp_address_p): New function.
24596 (mips_print_operand_punctuation): Recognize short delay slot insns
24597 for microMIPS.add<mode>3"
24598
24599 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24600
24601 PR target/56720
24602 * config/arm/iterators.md (v_cmp_result): New mode attribute.
24603 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
24604
24605 2013-03-25 Richard Biener <rguenther@suse.de>
24606
24607 PR tree-optimization/56689
24608 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
24609 any edge.
24610
24611 2013-03-25 Richard Biener <rguenther@suse.de>
24612
24613 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
24614 of bitmap.
24615 (memory_references): Likewise.
24616 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
24617 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
24618 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
24619 (gather_mem_refs_in_loops): Fold into ...
24620 (analyze_memory_references): ... this. Move initialization
24621 to tree_ssa_lim_initialize.
24622 (fill_always_executed_in): Rename to ...
24623 (fill_always_executed_in_1): ... this.
24624 (fill_always_executed_in): Move contains_call computation to
24625 this new function from ...
24626 (tree_ssa_lim_initialize): ... here.
24627 (tree_ssa_lim): Call fill_always_executed_in.
24628
24629 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24630
24631 * postreload.c (reload_combine): Fix code detecting returns.
24632
24633 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24634
24635 * function.c (emit_use_return_register_into_block): On cc0 targets,
24636 do not emit the sequence between cc0 setter and user.
24637
24638 2013-03-25 Kai Tietz <ktietz@redhat.com>
24639
24640 * config/i386/predicates.md (local_symbolic_operand): Interpret
24641 dll-imported symbols as none-local.
24642
24643 2013-03-25 Richard Biener <rguenther@suse.de>
24644
24645 * tree-ssa-loop-im.c (struct depend): Remove.
24646 (struct lim_aux_data): Make depends a vec of gimples.
24647 (free_lim_aux_data): Adjust.
24648 (add_dependency): Likewise.
24649 (set_level): Likewise.
24650
24651 2013-03-25 Richard Biener <rguenther@suse.de>
24652
24653 PR middle-end/56434
24654 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
24655 the pointer returned by calls with ECF_MALLOC set.
24656
24657 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24658
24659 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
24660
24661 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24662
24663 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
24664 using MMXMODE mode iterator.
24665 (*move<mode>_internal): Merge with *movv2sf_internal and
24666 *movv2sf_internal_rex64 using MMXMODE mode iterator.
24667
24668 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24669
24670 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
24671 (record_last_mem_set_info): Likewise.
24672
24673 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
24674 of XNEWVEC followed by memset.
24675 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
24676
24677 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24678
24679 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
24680 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
24681 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
24682 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
24683 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
24684 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
24685 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
24686 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
24687 BARRIER_P instead of GET_CODE.
24688
24689 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24690
24691 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
24692 inaccuracy in the probing code.
24693
24694 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
24695 (ctrapdi4): Likewise.
24696
24697 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24698
24699 * calls.c (expand_call): Add missing guard to code handling return
24700 of non-BLKmode structures in MSB.
24701 * function.c (expand_function_end): Likewise.
24702
24703 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24704
24705 * combine.c (try_combine): Adjust comment. Do not add the set of
24706 insn #0 if the destination indirectly is set or dies in insn #2.
24707 Tidy up code to distribute a new note.
24708
24709 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24710
24711 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
24712 also for alternatives 16 and 17.
24713
24714 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24715
24716 * config/i386/sse.md (*mov<mode>_internal): Merge with
24717 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
24718 Emit insn template depending on type attribute. Use
24719 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24720 movd instead of movq mnemonic for interunit moves. Rewrite mode
24721 attribute calculation. Remove unit attribute calculation.
24722 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
24723 Set prefix_data16 attribute for DImode ssemov types.
24724 Use Ym instead of y for SSE-MMX conversion alternatives.
24725 Reorder operand constraints.
24726
24727 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
24728
24729 * df.h (df_insn_delete): Adjust prototype.
24730 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
24731 and let it decide whether mark the basic block dirty.
24732 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
24733 * df-scan.c (df_insn_info_delete): New helper function, split
24734 off from df_insn_delete.
24735 (df_scan_free_bb_info): Use it.
24736 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
24737 Likewise.
24738 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
24739 that the insn is actually an insn and it has a non-NULL basic block.
24740 Do not mark basic block dirty if only deleting a DEBUG_INSN.
24741
24742 2013-03-22 Richard Biener <rguenther@suse.de>
24743
24744 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
24745 dep_ref members.
24746 (mem_ref_alloc): Do not allocate them.
24747 (refs_independent_p): Do not query or maintain a cache.
24748
24749 2013-03-22 Richard Biener <rguenther@suse.de>
24750
24751 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
24752 (gather_mem_refs_in_loops): Do not compute it.
24753 (analyze_memory_references): Do not allocate it.
24754 (tree_ssa_lim_finalize): Do not free it.
24755 (for_all_locs_in_loop): Do not query all_refs_in_loop.
24756
24757 2013-03-22 Richard Biener <rguenther@suse.de>
24758
24759 * is-a.h (as_a): Use gcc_checking_assert.
24760
24761 2013-03-22 Ian Bolton <ian.bolton@arm.com>
24762
24763 * config/aarch64/aarch64.c (aarch64_print_operand): New
24764 format specifier for printing a constant in hex.
24765 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
24766 format specifier for printing second operand.
24767
24768 2013-03-22 Richard Biener <rguenther@suse.de>
24769
24770 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
24771 bitmaps.
24772 (gather_mem_refs_in_loops): Perform store accumulation here.
24773 (create_vop_ref_mapping_loop): Remove.
24774 (create_vop_ref_mapping): Likewise.
24775 (analyze_memory_references): Initialize refs_stored_in_loop.
24776 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
24777 (record_indep_loop): Remove.
24778 (record_dep_loop): New function.
24779 (ref_indep_loop_p_1): Adjust to only walk over references
24780 in the loop, not its subloops.
24781 (ref_indep_loop_p): Rename to ...
24782 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
24783 maintaining a more fine-grained cache.
24784 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
24785 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
24786
24787 2013-03-22 Richard Biener <rguenther@suse.de>
24788
24789 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
24790 (struct mem_ref): Make accesses_in_loop a vec of a vec of
24791 aggregate mem_ref_loc.
24792 (free_mem_ref_locs): Inline into ...
24793 (memref_free): ... this and adjust.
24794 (mem_ref_alloc): Adjust.
24795 (mem_ref_locs_alloc): Remove.
24796 (record_mem_ref_loc): Adjust.
24797 (get_all_locs_in_loop): Rewrite into ...
24798 (for_all_locs_in_loop): ... this iterator.
24799 (rewrite_mem_ref_loc): New functor.
24800 (rewrite_mem_refs): Use for_all_locs_in_loop.
24801 (sm_set_flag_if_changed): New functor.
24802 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
24803 (ref_always_accessed): New functor.
24804 (ref_always_accessed_p): Use for_all_locs_in_loop.
24805
24806 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24807
24808 * tree-pass.h (PROP_gimple_lvec): New.
24809 * passes.c (dump_properties): Handle PROP_gimple_lvec.
24810 (init_optimization_passes): Move pass_lower_vector.
24811 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
24812 PROP_gimple_lvec.
24813 (pass_lower_vector): Provide PROP_gimple_lvec.
24814 (pass_lower_vector_ssa): Likewise.
24815 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
24816
24817 2013-03-21 Mark Wielaard <mjw@redhat.com>
24818
24819 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
24820
24821 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
24822
24823 * config/i386/i386.md (*movdi_internal): Disparage slightly
24824 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
24825 conversion alternatives.
24826
24827 2013-03-21 Jakub Jelinek <jakub@redhat.com>
24828
24829 PR middle-end/48087
24830 * diagnostic.def (DK_WERROR): New kind.
24831 * diagnostic.h (werrorcount): Define.
24832 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
24833 promoted to DK_ERROR, increment DK_WERROR counter instead of
24834 DK_ERROR counter.
24835 * toplev.c (toplev_main): Call print_ignored_options even if
24836 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
24837 even if just werrorcount is non-zero.
24838
24839 PR debug/55608
24840 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
24841 on failure.
24842 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
24843 (string_cst_pool_decl): New function.
24844 (optimize_one_addr_into_implicit_ptr): New function.
24845 (resolve_addr_in_expr): Optimize DWARF location expression
24846 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
24847 which doesn't live in memory, but has DW_AT_location or
24848 DW_AT_const_value, or refers to a string literal, into
24849 DW_OP_GNU_implicit_pointer.
24850 (optimize_location_into_implicit_ptr): New function.
24851 (resolve_addr): If removing DW_AT_location of a variable because
24852 it was DW_OP_addr of address of the variable, but the variable doesn't
24853 live in memory, try to emit const value attribute for the initializer.
24854
24855 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24856
24857 * tree.h (VECTOR_TYPE_P): New macro.
24858 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
24859 TYPE_MODE): Use it.
24860 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
24861 VEC_COND_EXPR cannot be lvalues.
24862 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
24863
24864 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24865
24866 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
24867 Restrict the transformation to equal modes.
24868
24869 2013-03-21 Richard Biener <rguenther@suse.de>
24870
24871 PR tree-optimization/39326
24872 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
24873 (MEM_ANALYZABLE): Adjust.
24874 (record_mem_ref_loc): Move bitmap ops ...
24875 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
24876 unanalyzable refs, do not record locations for it.
24877 (analyze_memory_references): Allocate ref zero as shared
24878 unanalyzable ref.
24879 (refs_independent_p): Do not test for unanalyzed mems here.
24880 (ref_indep_loop_p_1): Special-case disambiguation against
24881 the unanalyzed ref.
24882 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
24883
24884 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
24885
24886 * config/arm/arm-protos.h (tune_params): Add
24887 prefer_neon_for_64bits field.
24888 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
24889 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
24890 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
24891 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
24892 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
24893 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
24894 (arm_option_override): Handle -mneon-for-64bits new option.
24895 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
24896 (prefer_neon_for_64bits): Declare new variable.
24897 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
24898 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
24899 (arch_enabled): Handle new arch types. Remove support for onlya8
24900 and nota8.
24901 (one_cmpldi2): Use new arch names.
24902 (zero_extend<mode>di2, extend<mode>di2): Ditto.
24903 * config/arm/arm.opt (mneon-for-64bits): Add option.
24904 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
24905 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
24906 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
24907 of onlya8.
24908 * doc/invoke.texi (-mneon-for-64bits): Document.
24909
24910 2013-03-21 Richard Biener <rguenther@suse.de>
24911
24912 PR tree-optimization/39326
24913 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
24914 (sort_bbs_in_loop_postorder_cmp): New function.
24915 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
24916
24917 2013-03-21 Richard Biener <rguenther@suse.de>
24918
24919 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
24920 (vect_insert_into_interleaving_chain): Likewise.
24921 (vect_drs_dependent_in_basic_block): Inline ...
24922 (vect_slp_analyze_data_ref_dependence): ... here. New function,
24923 split out from ...
24924 (vect_analyze_data_ref_dependence): ... here. Simplify.
24925 (vect_check_interleaving): Simplify.
24926 (vect_analyze_data_ref_dependences): Likewise. Split out ...
24927 (vect_slp_analyze_data_ref_dependences): ... this new function.
24928 (dr_group_sort_cmp): New function.
24929 (vect_analyze_data_ref_accesses): Compute data-reference groups
24930 here instead of in vect_analyze_data_ref_dependence. Use
24931 a more efficient algorithm.
24932 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
24933 vect_slp_analyze_data_ref_dependences. Call
24934 vect_analyze_data_ref_accesses earlier.
24935 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
24936 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
24937 (vect_slp_analyze_data_ref_dependences): New prototype.
24938
24939 2013-03-21 Richard Biener <rguenther@suse.de>
24940
24941 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
24942 ref is stored in the loop.
24943 (find_refs_for_sm): Walk only over all stores.
24944 (store_motion_loop): Allocate from lim_bitmap_obstack.
24945 (store_motion): Likewise.
24946
24947 2013-03-21 Richard Biener <rguenther@suse.de>
24948
24949 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
24950 Update virtual SSA form.
24951
24952 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
24953
24954 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
24955 * configure: Regenerate.
24956 * config.in: Regenerate.
24957 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
24958 if !HAVE_LD_EH_FRAME_CIEV3.
24959
24960 2013-03-21 Richard Biener <rguenther@suse.de>
24961
24962 * tree-cfg.c (verify_expr_no_block): New function.
24963 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
24964 nor DECL_VALUE_EXPR have locations with associated blocks.
24965 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
24966 (clear_unused_block_pointer): Remove code dealing with
24967 blocks in DECL_DEBUG_EXPR locations.
24968
24969 2013-03-21 Richard Biener <rguenther@suse.de>
24970
24971 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
24972 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
24973 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
24974 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
24975 instead of DECL_DEBUG_EXPR_IS_FROM.
24976 * gimplify.c (gimplify_modify_expr): Likewise.
24977 * tree-cfg.c (verify_expr_location_1): Likewise.
24978 * tree-complex.c (create_one_component_var): Likewise.
24979 * tree-sra.c (create_access_replacement): Likewise.
24980 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
24981 (clear_unused_block_pointer): Likewise.
24982 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
24983 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
24984 * var-tracking.c (var_debug_decl): Likewise.
24985 (track_expr_p): Likewise.
24986 * tree-inline.c (add_local_variables): Likewise. Set
24987 DECL_HAS_DEBUG_EXPR_P after copying it.
24988 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
24989 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
24990
24991 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
24992
24993 PR bootstrap/56656
24994 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
24995 * configure: Regenerate.
24996 * config.in: Regenerate.
24997 * config/i386/i386.md (*movdf_internal): Use
24998 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24999 movd instead of movq mnemonic for interunit moves.
25000 (*movdi_internal): Ditto.
25001
25002 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25003
25004 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
25005 (abd<mode>_3): New pattern.
25006 (aba<mode>_3): New pattern.
25007 (fabd<mode>_3): New pattern.
25008
25009 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25010
25011 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
25012 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
25013 occurrence of REGISTER_PREFIX as its empty string.
25014
25015 2013-03-20 Jeff Law <law@redhat.com>
25016
25017 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
25018 addititional equivalences for equality comparisons between an SSA_NAME
25019 and a constant where the SSA_NAME was set from a widening conversion.
25020
25021 2013-03-20 Walter Lee <walt@tilera.com>
25022
25023 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
25024
25025 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25026
25027 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
25028 depending on type attribute.
25029 (*movti_internal): Ditto.
25030 (*movtf_internal): Ditto.
25031 (*movxf_internal): Ditto.
25032 (*movdf_internal): Ditto.
25033 (*movsf_internal): Ditto.
25034
25035 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25036
25037 * config/i386/i386.md (*movti_internal): Set prefix attribute to
25038 maybe_vex for sselog1 and ssemov types.
25039 (*movdi_internal): Reorder operand constraints.
25040 (*movsi_internal): Ditto. Set prefix attribute to
25041 maybe_vex for sselog1 and ssemov types.
25042 (*movtf_internal): Set prefix attribute to maybe_vex
25043 for sselog1 and ssemov types.
25044 (*movdf_internal): Ditto. Set prefix_data16 attribute for
25045 DImode ssemov types. Reorder operand constraints.
25046 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
25047 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
25048 attribute for SImode ssemov types. Reorder operand constraints.
25049
25050 2013-03-20 Martin Jambor <mjambor@suse.cz>
25051
25052 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
25053 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
25054
25055 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
25056
25057 * config/rs6000/predicates.md (indexed_address, update_address_mem
25058 update_indexed_address_mem): New predicates.
25059 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
25060 attribute for load/store instructions.
25061 * config/rs6000/dfp.md (movsd_store): Likewise.
25062 (movsd_load): Likewise.
25063 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
25064 (unnamed HI->DI extend define_insn): Likewise.
25065 (unnamed SI->DI extend define_insn): Likewise.
25066 (unnamed QI->SI extend define_insn): Likewise.
25067 (unnamed QI->HI extend define_insn): Likewise.
25068 (unnamed HI->SI extend define_insn): Likewise.
25069 (unnamed HI->SI extend define_insn): Likewise.
25070 (extendsfdf2_fpr): Likewise.
25071 (movsi_internal1): Likewise.
25072 (movsi_internal1_single): Likewise.
25073 (movhi_internal): Likewise.
25074 (movqi_internal): Likewise.
25075 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
25076 attribute for load/store instructions.
25077 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
25078 instructions.
25079 (mov<mode>_softfloat): Likewise.
25080 (mov<mode>_hardfloat32): Likewise.
25081 (mov<mode>_hardfloat64): Likewise.
25082 (mov<mode>_softfloat64): Likewise.
25083 (movdi_internal32): Likewise.
25084 (movdi_internal64): Likewise.
25085 (probe_stack_<mode>): Likewise.
25086
25087 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
25088
25089 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
25090 floating point, and decimal floating point to reload iterator.
25091
25092 * config/rs6000/constraints.md (wl constraint): New constraints to
25093 return FLOAT_REGS if certain options are used to reduce the number
25094 of separate patterns that exist in the file.
25095 (wx constraint): Likewise.
25096 (wz constraint): Likewise.
25097
25098 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25099 -mdebug=reg, print wg, wl, wx, and wz constraints.
25100 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25101 Initialize the reload functions for 64-bit binary/decimal floating
25102 point types.
25103 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25104 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25105 create the buffer on the stack to overcome not having a 32-bit
25106 load and store.
25107 (rs6000_emit_move): Likewise.
25108 (rs6000_secondary_memory_needed_rtx): Likewise.
25109 (rs6000_alloc_sdmode_stack_slot): Likewise.
25110 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25111 via xxlxor, just like DFmode 0.0.
25112
25113 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
25114 define as 1 if we are running on a power7 or newer.
25115 (enum r6000_reg_class_enum): Add new constraints.
25116
25117 * config/rs6000/dfp.md (movsd): Delete, combine with binary
25118 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25119 with other moves by using conditional constraits (wg). Use LFIWZX
25120 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
25121 (movsd splitter): Likewise.
25122 (movsd_hardfloat): Likewise.
25123 (movsd_softfloat): Likewise.
25124
25125 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
25126 binary and decimal floating point moves.
25127 (fmove_ok): New attributes to combine binary and decimal floating
25128 point moves, and to combine power6x (mfpgpr) moves along normal
25129 floating moves.
25130 (real_value_to_target): Likewise.
25131 (f32_lr): Likewise.
25132 (f32_lm): Likewise.
25133 (f32_li): Likewise.
25134 (f32_sr): Likewise.
25135 (f32_sm): Likewise.
25136 (f32_si): Likewise.
25137 (movsf): Combine binary and decimal floating point moves. Combine
25138 power6x (mfpgpr) moves with other moves by using conditional
25139 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
25140 (mov<mode> for SFmode/SDmode); Likewise.
25141 (SFmode/SDmode splitters): Likewise.
25142 (movsf_hardfloat): Likewise.
25143 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
25144 (movsf_softfloat): Likewise.
25145 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
25146
25147 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
25148 wx and wz constraints.
25149
25150 * config/rs6000/constraints.md (wg constraint): New constraint to
25151 return FLOAT_REGS if -mmfpgpr (power6x) was used.
25152
25153 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
25154 constraint.
25155
25156 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25157 -mdebug=reg, print wg, wl, wx, and wz constraints.
25158 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25159 Initialize the reload functions for 64-bit binary/decimal floating
25160 point types.
25161 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25162 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25163 create the buffer on the stack to overcome not having a 32-bit
25164 load and store.
25165 (rs6000_emit_move): Likewise.
25166 (rs6000_secondary_memory_needed_rtx): Likewise.
25167 (rs6000_alloc_sdmode_stack_slot): Likewise.
25168 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25169 via xxlxor, just like DFmode 0.0.
25170
25171 * config/rs6000/dfp.md (movdd): Delete, combine with binary
25172 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25173 with other moves by using conditional constraits (wg). Use LFIWZX
25174 and STFIWX for loading SDmode on power7.
25175 (movdd splitters): Likewise.
25176 (movdd_hardfloat32): Likewise.
25177 (movdd_softfloat32): Likewise.
25178 (movdd_hardfloat64_mfpgpr): Likewise.
25179 (movdd_hardfloat64): Likewise.
25180 (movdd_softfloat64): Likewise.
25181
25182 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
25183 64-bit binary and decimal floating point moves.
25184 (FMOVE64X): Likewise.
25185 (movdf): Combine 64-bit binary and decimal floating point moves.
25186 Combine power6x (mfpgpr) moves with other moves by using
25187 conditional constraits (wg).
25188 (mov<mode> for DFmode/DDmode): Likewise.
25189 (DFmode/DDmode splitters): Likewise.
25190 (movdf_hardfloat32): Likewise.
25191 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
25192 (movdf_softfloat32): Likewise.
25193 (movdf_hardfloat64_mfpgpr): Likewise.
25194 (movdf_hardfloat64): Likewise.
25195 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
25196 (movdf_softfloat64): Likewise.
25197 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
25198 (reload_<mode>_load): Move to later in the file so they aren't in
25199 the middle of the floating point move insns.
25200 (reload_<mode>_store): Likewise.
25201
25202 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
25203 constraint.
25204
25205 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
25206 constraint if -mdebug=reg.
25207 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
25208 Enable using dd reload support if needed.
25209
25210 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
25211 binary and decimal floating point moves in rs6000.md.
25212 (movtd_internal): Likewise.
25213
25214 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
25215 decimal floating point moves.
25216 (movtf): Likewise.
25217 (movtf_internal): Likewise.
25218 (mov<mode>_internal, TDmode/TFmode): Likewise.
25219 (movtf_softfloat): Likewise.
25220 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
25221
25222 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
25223 movdi_internal64, using wg constraint for move direct operations.
25224 (movdi_internal64): Likewise.
25225
25226 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
25227 MODES_TIEABLE_P for selected modes. Print the numerical value of
25228 the various virtual registers. Use GPR/FPR first/last values,
25229 instead of hard coding the register numbers. Print which modes
25230 have reload functions registered.
25231 (rs6000_option_override_internal): If -mdebug=reg, trace the options
25232 settings before/after setting cpu, target and subtarget settings.
25233 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
25234 and for secondary reload failures in rs6000_secondary_reload_inner.
25235 (rs6000_secondary_reload_fail): Likewise.
25236 (rs6000_secondary_reload_inner): Likewise.
25237
25238 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
25239 macros for first/last GPR and FPR registers.
25240 (LAST_GPR_REGNO): Likewise.
25241 (FIRST_FPR_REGNO): Likewise.
25242 (LAST_FPR_REGNO): Likewise.
25243
25244 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
25245 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
25246 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
25247 (vcond<mode><mode>): Likewise.
25248 (vcondu<mode><mode>): Likewise.
25249 (vector_gtu<mode>): Likewise.
25250 (vector_gte<mode>): Likewise.
25251 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
25252 to prevent the compiler from converting DImode operations to TImode.
25253 (ior<mode>3): Likewise.
25254 (and<mode>3): Likewise.
25255 (one_cmpl<mode>2): Likewise.
25256 (nor<mode>3): Likewise.
25257 (andc<mode>3): Likewise.
25258
25259 * config/rs6000/constraints.md (wt constraint): New constraint
25260 that returns VSX_REGS if TImode is allowed in VSX registers.
25261
25262 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
25263 constant under VSX.
25264
25265 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
25266 similar to TImode, but it is restricted to being in the GPRs.
25267
25268 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
25269 TImode to occupy a single VSX register.
25270
25271 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
25272 -mvsx-timode for power7/power8.
25273 (power7 cpu): Likewise.
25274 (power8 cpu): Likewise.
25275
25276 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
25277 sure that TFmode/TDmode take up two registers if they are ever
25278 allowed in the upper VSX registers.
25279 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
25280 registers.
25281 (rs6000_init_hard_regno_mode_ok): Likewise.
25282 (rs6000_debug_reg_global): Add debugging for PTImode and wt
25283 constraint. Print if LRA is turned on.
25284 (rs6000_option_override_internal): Give an error if -mvsx-timode
25285 and VSX is not enabled.
25286 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
25287 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
25288 to reg+offset addressing. Use PTImode when checking offset
25289 addresses for validity.
25290 (reg_offset_addressing_ok_p): Likewise.
25291 (rs6000_legitimate_offset_address_p): Likewise.
25292 (rs6000_legitimize_address): Likewise.
25293 (rs6000_legitimize_reload_address): Likewise.
25294 (rs6000_legitimate_address_p): Likewise.
25295 (rs6000_eliminate_indexed_memrefs): Likewise.
25296 (rs6000_emit_move): Likewise.
25297 (rs6000_secondary_reload): Likewise.
25298 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
25299 reloads to fpr registers to continue to use reg+offset addressing,
25300 but 64-bit reloads to altivec registers need reg+reg addressing.
25301 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
25302 it. Treat LO_SUM like a PLUS operation.
25303 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
25304 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
25305 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
25306 registers to share a register with a smaller sized type, since VSX
25307 puts scalars in the upper 64-bits.
25308 (print_operand): Add support for PTImode.
25309 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
25310 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
25311 registers, but don't have arithmetic support.
25312 (rs6000_memory_move_cost): Add test for VSX.
25313 (rs6000_opt_masks): Add -mvsx-timode.
25314
25315 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
25316 for TImode.
25317 (VSs): Likewise.
25318 (VSr): Use wt constraint for TImode.
25319 (VSv): Drop TImode support.
25320 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
25321 (vsx_movti_64bit): Likewise.
25322 (vsx_movti_32bit): Likewise.
25323 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
25324 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
25325 one '?' on the appropriate output constraint. Do not allow TImode
25326 logical operations on 32-bit systems.
25327 (vsx_ior<mode>3): Likewise.
25328 (vsx_xor<mode>3): Likewise.
25329 (vsx_one_cmpl<mode>2): Likewise.
25330 (vsx_nor<mode>3): Likewise.
25331 (vsx_andc<mode>3): Likewise.
25332 (vsx_concat_<mode>): Likewise.
25333 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
25334
25335 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
25336 OPTION_MASK_VSX_TIMODE.
25337 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
25338 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
25339
25340 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
25341 (TI2 iterator): New iterator for TImode, PTImode.
25342 (wd mode attribute): Add values for vector types.
25343 (movti_string): Replace TI move operations with operations for TImode
25344 and PTImode. Add support for TImode being allowed in VSX registers.
25345 (mov<mode>_string, TImode/PTImode): Likewise.
25346 (movti_ppc64): Likewise.
25347 (mov<mode>_ppc64, TImode/PTImode): Likewise.
25348 (TI mode splitters): Likewise.
25349
25350 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
25351 constraint.
25352
25353 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
25354
25355 PR tree-optimization/56355
25356 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
25357 Also handle integers with undefined overflow.
25358
25359 2013-03-20 Catherine Moore <clm@codesourcery.com>
25360 Maciej W. Rozycki <macro@codesourcery.com>
25361 Tom de Vries <tom@codesourcery.com>
25362 Nathan Sidwell <nathan@codesourcery.com>
25363 Iain Sandoe <iain@codesourcery.com>
25364 Nathan Froyd <froydnj@codesourcery.com>
25365 Chao-ying Fu <fu@mips.com>
25366
25367 * doc/extend.texi: (micromips, nomicromips, nocompression):
25368 Document new function attributes.
25369 * doc/invoke.texi (minterlink-compressed, mmicromips,
25370 m14k, m14ke, m14kec): Document new options.
25371 (minterlink-mips16): Update documentation.
25372 * doc/md.texi (ZC, ZD): Document new constraints.
25373 * configure.ac (gcc_cv_as_micromips): Check if linker
25374 supports the .set micromips directive.
25375 * configure: Regenerate.
25376 * config.in: Regenerate.
25377 * config/mips/mips-tables.opt: Regenerate.
25378 * config/mips/micromips.md: New file.
25379 * constraints.md (ZC, ZD): New constraints.
25380 * config/mips/predicates.md (movep_src_register): New predicate.
25381 (movep_src_operand): New predicate.
25382 (non_volatile_mem_operand): New predicate.
25383 * config/mips/mips.md (multimem): New type.
25384 (length): Differentiate between 17-bit and 18-bit branch offsets.
25385 (MOVEP1, MOVEP2): New mode iterator.
25386 (mov_<load>l): Use ZC constraint.
25387 (mov_<load>r): Likewise.
25388 (mov_<store>l): Likewise.
25389 (mov_<store>r): Likewise.
25390 (*branch_equality<mode>_inverted): Add microMIPS support.
25391 (*branch_equality<mode>): Likewise.
25392 (*jump_absolute): Likewise.
25393 (indirect_jump_<mode>): Likewise.
25394 (tablejump_<mode>): Likewise.
25395 (<optab>_internal): Likewise.
25396 (sibcall_internal): Likewise.
25397 (sibcall_value_internal): Likewise.
25398 (prefetch): Use constraint ZD.
25399 * config/mips/mips.opt (minterlink-compressed): New option.
25400 (minterlink-mips16): Now an alias for minterlink-compressed.
25401 (mmicromips): New option.
25402 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
25403 (compare_and_swap_12): Likewise.
25404 (sync_add<mode>): Likewise.
25405 (sync_<optab>_12): Likewise.
25406 (sync_old_<optab>_12): Likewise.
25407 (sync_new_<optab>_12): Likewise.
25408 (sync_nand_12): Likewise.
25409 (sync_old_nand_12): Likewise.
25410 (sync_new_nand_12): Likewise.
25411 (sync_sub<mode>): Likewise.
25412 (sync_old_add<mode>): Likewise.
25413 (sync_old_sub<mode>): Likewise.
25414 (sync_new_add<mode>): Likewise.
25415 (sync_new_sub<mode>): Likewise.
25416 (sync_<optab><mode>): Likewise.
25417 (sync_old_<optab><mode>): Likewise.
25418 (sync_new_<optab><mode>): Likewise.
25419 (sync_nand<mode>): Likewise.
25420 (sync_old_nand<mode>): Likewise.
25421 (sync_new_nand<mode>): Likewise.
25422 (sync_lock_test_and_set<mode>): Likewise.
25423 (test_and_set_12): Likewise.
25424 (atomic_compare_and_swap<mode>): Likewise.
25425 (atomic_exchange<mode>_llsc): Likewise.
25426 (atomic_fetch_add<mode>_llsc): Likewise.
25427 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
25428 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
25429 (umips_save_restore_pattern_p): Likewise.
25430 (umips_load_store_pair_p): Likewise.
25431 (umips_output_load_store_pair): Likewise.
25432 (umips_movep_target_p): Likewise.
25433 (umips_12bit_offset_address_p): Likewise.
25434 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
25435 (mips_base_mips16): Rename this...
25436 (mips_base_compression_flags): ...to this. Update all uses.
25437 (mips_attribute_table): Add micromips, nomicromips and nocompression.
25438 (mips_mips16_decl_p): Delete.
25439 (mips_nomips16_decl_p): Delete.
25440 (mips_get_compress_on_flags): New function.
25441 (mips_get_compress_off_flags): New function.
25442 (mips_get_compress_mode): New function.
25443 (mips_get_compress_on_name): New function.
25444 (mips_get_compress_off_name): New function.
25445 (mips_insert_attributes): Support multiple compression types.
25446 (mips_merge_decl_attributes): Likewise.
25447 (umips_12bit_offset_address_p): New function.
25448 (mips_start_function_definition): Emit .set micromips directive.
25449 (mips_call_may_need_jalx_p): New function.
25450 (mips_function_ok_for_sibcall): Add microMIPS support.
25451 (mips_print_operand_punctuation): Support short delay slots and
25452 compact jumps.
25453 (umips_swm_mask, umips_swm_encoding): New.
25454 (umips_build_save_restore): New function.
25455 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
25456 (was_mips16_p): Remove.
25457 (old_compression_mode): New.
25458 (mips_set_compression_mode): New function.
25459 (mips_set_current_function): Add microMIPS support.
25460 (mips_option_override): Likewise.
25461 (umips_save_restore_pattern_p): New function.
25462 (umips_output_save_restore): New function.
25463 (umips_load_store_pair_p_1): New function.
25464 (umips_load_store_pair_p): New function.
25465 (umips_output_load_store_pair_1): New function.
25466 (umips_output_load_store_pair): New function.
25467 (umips_movep_target_p) New function.
25468 (mips_prepare_pch_save): Add microMIPS support.
25469 * config/mips/mips.h (TARGET_COMPRESSION): New.
25470 (TARGET_CPU_CPP_BUILTINS): Update macro
25471 to use new compression flags and to support microMIPS.
25472 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
25473 (MIPS_ARCH_FLOAT_SPEC): Likewise.
25474 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
25475 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
25476 (ASM_SPEC): Support mmicromips and mno-micromips.
25477 (M16STORE_REG_P): New macro.
25478 (MIPS_CALL): Support TARGET_MICROMIPS.
25479 (MICROMIPS_J): New macro.
25480 (mips_base_mips16): Rename this...
25481 (mips_base_compression_flags): ...to this.
25482 (UMIPS_12BIT_OFFSET_P): New macro.
25483 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
25484 (MULTILIB_DIRNAMES): Likewise.
25485 2013-03-20 Richard Biener <rguenther@suse.de>
25486
25487 PR tree-optimization/56661
25488 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
25489 the result does not have to be distinct.
25490
25491 2013-03-20 Richard Biener <rguenther@suse.de>
25492
25493 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
25494 remap_gimple_op_r.
25495
25496 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
25497 Steven Bosscher <steven@gcc.gnu.org>
25498
25499 PR rtl-optimization/56605
25500 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
25501
25502 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25503
25504 PR bootstrap/56656
25505 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
25506 that require movd instead of movq.
25507
25508 2013-03-20 Richard Biener <rguenther@suse.de>
25509
25510 * tree-ssa-structalias.c (struct variable_info): Add pointer
25511 to the first field of an aggregate with sub-vars. Make
25512 this and the pointer to the next subfield its ID.
25513 (vi_next): New function.
25514 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
25515 storedanything_id, integer_id): Increment by one.
25516 (new_var_info, get_call_vi, lookup_call_clobber_vi,
25517 get_call_clobber_vi): Adjust.
25518 (solution_set_expand): Simplify and speedup.
25519 (solution_set_add): Inline into ...
25520 (set_union_with_increment): ... this. Adjust accordingly.
25521 (do_sd_constraint): Likewise.
25522 (do_ds_constraint): Likewise.
25523 (do_complex_constraint): Simplify.
25524 (build_pred_graph): Adjust.
25525 (solve_graph): Likewise. Simplify and speedup.
25526 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
25527 get_constraint_for_component_ref, get_constraint_for_1,
25528 first_vi_for_offset, first_or_preceding_vi_for_offset,
25529 create_function_info_for, create_variable_info_for_1,
25530 create_variable_info_for, intra_create_variable_infos): Adjust.
25531 (init_base_vars): Push NULL for ID zero.
25532 (compute_points_to_sets): Adjust.
25533
25534 2013-03-20 Richard Biener <rguenther@suse.de>
25535
25536 * cfgloop.c (verify_loop_structure): Streamline and avoid
25537 ICEing on corrupt loop tree.
25538 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
25539 loop tree.
25540
25541 2013-03-20 Richard Biener <rguenther@suse.de>
25542
25543 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
25544 check whether an SSA update is needed.
25545
25546 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
25547
25548 * config/mips/constraints.md (T): Rename to...
25549 (Yf): ...this.
25550 (U): Rename to...
25551 (Yd): ...this.
25552 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
25553 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
25554
25555 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25556
25557 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
25558 (*subsi3_carryin_uxtw): Likewise.
25559
25560 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25561
25562 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
25563 (*rorsi3_insn_uxtw): Likewise.
25564
25565 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25566
25567 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
25568 (*extrsi5_insn_uxtw): Likewise.
25569
25570 2013-03-19 Richard Biener <rguenther@suse.de>
25571
25572 PR tree-optimization/56273
25573 * passes.c (init_optimization_passes): Move second VRP after DOM.
25574
25575 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
25576
25577 * config/i386/i386.md (*movti_internal): Merge from
25578 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
25579 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
25580 nox64 isa attributes.
25581
25582 2013-03-18 Richard Biener <rguenther@suse.de>
25583
25584 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
25585 (unite): Likewise.
25586 (merge_node_constraints): Likewise.
25587 (build_succ_graph): Likewise.
25588 (valid_graph_edge): Inline into single caller.
25589 (unify_nodes): Likewise. Use bitmap_set_bit return value
25590 and cache varinfo.
25591 (scc_visit): Fix formatting and variable use.
25592 (do_sd_constraint): Use gcc_checking_assert.
25593 (do_ds_constraint): Likewise.
25594 (do_complex_constraint): Likewise.
25595 (condense_visit): Likewise. Cleanup.
25596 (dump_pred_graph): New function.
25597 (perform_var_substitution): Dump the pred-graph before
25598 variable substitution.
25599 (find_equivalent_node): Use gcc_checking_assert.
25600 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
25601
25602 2013-03-18 Richard Biener <rguenther@suse.de>
25603
25604 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
25605 Remove cond_expr_stmt_list argument and do not gimplify the
25606 built expression.
25607 (vect_loop_versioning): Adjust.
25608 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
25609 Cleanup to use less temporaries.
25610 (vect_create_data_ref_ptr): Cleanup.
25611
25612 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25613
25614 PR tree-optimization/56635
25615 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
25616 require types_compatible_p types.
25617
25618 2013-03-18 Nick Clifton <nickc@redhat.com>
25619
25620 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
25621 spurious backslash.
25622
25623 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
25624 Add missing line to comment describing function.
25625
25626 2013-03-18 Richard Biener <rguenther@suse.de>
25627
25628 PR tree-optimization/56210
25629 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
25630 Handle string / character search functions.
25631 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
25632
25633 2013-03-18 Richard Biener <rguenther@suse.de>
25634
25635 PR middle-end/56483
25636 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
25637 and implement properly.
25638 * gimple.h (gimple_cond_single_var_p): Remove.
25639
25640 2013-03-18 Richard Biener <rguenther@suse.de>
25641
25642 * tree-data-ref.h (find_data_references_in_loop): Declare.
25643 * tree-data-ref.c (get_references_in_stmt): Use a stack
25644 vector pre-allocated in the callers.
25645 (find_data_references_in_stmt): Adjust.
25646 (graphite_find_data_references_in_stmt): Likewise.
25647 (create_rdg_vertices): Likewise.
25648 (find_data_references_in_loop): Export.
25649 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
25650 Compute dependences here...
25651 (vect_analyze_data_refs): ...not here. When we encounter
25652 a non-vectorizable data reference in basic-block vectorization
25653 truncate the data reference vector. Do not bother to
25654 fixup data-dependence information for gather loads.
25655 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
25656 of data references, as reported.
25657
25658 2013-03-18 Richard Biener <rguenther@suse.de>
25659
25660 PR tree-optimization/3713
25661 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
25662 has_constants and expr.
25663 (stmt_has_constants): Properly valueize SSA names when deciding
25664 whether the stmt has constants.
25665
25666 2013-03-18 Richard Biener <rguenther@suse.de>
25667
25668 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
25669 whole function when there is nothing to do.
25670 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
25671 * tree-vectorizer.c (vectorize_loops): Update virtual and
25672 loop-closed SSA once.
25673 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
25674
25675 2013-03-18 Richard Biener <rguenther@suse.de>
25676
25677 PR middle-end/56113
25678 * domwalk.c (bb_postorder): New global static.
25679 (cmp_bb_postorder): New function.
25680 (walk_dominator_tree): Replace scheme imposing an order for
25681 visiting dominator sons by one sorting them at the time they
25682 are pushed on the stack.
25683
25684 2013-03-18 Richard Biener <rguenther@suse.de>
25685
25686 PR tree-optimization/39326
25687 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
25688 (struct mem_ref): Replace mem member with ao_ref typed member.
25689 (MEM_ANALYZABLE): Adjust.
25690 (memref_eq): Likewise.
25691 (mem_ref_alloc): Likewise.
25692 (gather_mem_refs_stmt): Likewise.
25693 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
25694 (execute_sm_if_changed_flag_set): Adjust.
25695 (execute_sm): Likewise.
25696 (ref_always_accessed_p): Likewise.
25697 (refs_independent_p): Likewise.
25698 (can_sm_ref_p): Likewise.
25699
25700 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25701
25702 PR c/56566
25703 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
25704 return 1 even for !unsignedp.
25705
25706 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25707
25708 * config/i386/i386.md (isa): Add x64 and nox64.
25709 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
25710 (*pushtf): Enable *roF alternative for x64 isa only.
25711 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
25712 mode attribute of integer alternatives to DImode for TARGET_64BIT.
25713 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
25714 (*movtf_internal): Merge from *movtf_internal_rex64 and
25715 *movtf_internal_sse. Use x64 and nox64 isa attributes.
25716 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
25717 nox64 isa attributes.
25718 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
25719 nox64 isa attributes.
25720 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
25721
25722 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25723
25724 * config/alpha/alpha.c (TARGET_LRA_P): New define.
25725
25726 2013-03-17 Jakub Jelinek <jakub@redhat.com>
25727
25728 PR target/56640
25729 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
25730 class names. Remove trailing comma after "ALL_REGS".
25731
25732 2013-03-16 Jan Hubicka <jh@suse.cz>
25733
25734 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
25735 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
25736 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
25737 of cgraph_get_create_node.
25738 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
25739
25740 2013-03-16 Jason Merrill <jason@redhat.com>
25741
25742 PR debug/49090
25743 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
25744 with DW_AT_default_value.
25745
25746 2013-03-16 Jakub Jelinek <jakub@redhat.com>
25747
25748 * BASE-VER: Set to 4.9.0.
25749
25750 2013-03-14 Andi Kleen <ak@linux.intel.com>
25751
25752 PR target/56619
25753 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
25754 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
25755 Document _x* TSX intrinsics.
25756
25757 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
25758 David Holsgrove <david.holsgrove@xilinx.com>
25759
25760 * configure.ac: Add MicroBlaze TLS support detection.
25761 * configure: Regenerate.
25762 * config/microblaze/microblaze-protos.h
25763 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
25764 symbol_mentioned_p, label_mentioned_p): Add prototypes.
25765 * config/microblaze/microblaze.c (microblaze_address_type): Add
25766 ADDRESS_TLS and tls_reloc address types.
25767 (microblaze_address_info): Add tls_reloc.
25768 (TARGET_HAVE_TLS): Define.
25769 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
25770 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
25771 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
25772 load_tls_operand, microblaze_call_tls_get_addr,
25773 microblaze_legitimize_tls_address): New functions.
25774 (microblaze_classify_unspec): Handle UNSPEC_TLS.
25775 (get_base_reg): Use microblaze_tls_symbol_p.
25776 (microblaze_classify_address): Handle TLS.
25777 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
25778 label_mentioned_p and microblaze_tls_referenced_p.
25779 (microblaze_legitimize_address): Handle TLS.
25780 (microblaze_address_insns): Handle ADDRESS_TLS.
25781 (pic_address_needs_scratch): Handle TLS.
25782 (print_operand_address): Handle TLS.
25783 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
25784 (microblaze_expand_move): Handle TLS.
25785 (microblaze_legitimate_constant_p): Check
25786 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
25787 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
25788 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
25789 (PIC_OFFSET_TABLE_REGNUM): Set.
25790 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
25791 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
25792 (addsi3, movsi_internal2, movdf_internal): Update constraints
25793 * config/microblaze/predicates.md (arith_plus_operand): Define
25794 (move_operand): Redefine as move_src_operand,
25795 check microblaze_tls_referenced_p.
25796
25797 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25798
25799 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
25800 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
25801
25802 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25803
25804 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
25805 CC mode for AND.
25806
25807 2013-03-14 Jakub Jelinek <jakub@redhat.com>
25808
25809 PR tree-optimization/53265
25810 * common.opt (Waggressive-loop-optimizations): New option.
25811 * tree-ssa-loop-niter.c: Include tree-pass.h.
25812 (do_warn_aggressive_loop_optimizations): New function.
25813 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
25814 if number_of_latch_executions returned constant.
25815 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
25816 early. If number_of_latch_executions returned constant, set
25817 nb_iterations_upper_bound back to it.
25818 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
25819 field.
25820 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
25821 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
25822
25823 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
25824 (MULTILIB_OSDIRNAMES): Set.
25825 * genmultilib: If defaultosdirname doesn't start with :: , set
25826 defaultosdirname2 instead, clear it and emit two . multilib_raw
25827 entries instead of just one.
25828
25829 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
25830
25831 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
25832 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
25833 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
25834 (SUBTARGET_OVERRIDE_OPTIONS): New.
25835
25836 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
25837
25838 PR target/49880
25839 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
25840 (musermode): Convert to Var(TARGET_USERMODE).
25841 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
25842 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
25843 * config/sh/sh.c (sh_option_override): Use
25844 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
25845 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
25846 condition.
25847 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
25848 TARGET_SH4.
25849 (udivsi3_i4_single, divsi3_i4_single): Use
25850 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
25851
25852 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
25853
25854 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
25855 default setting.
25856
25857 2013-03-13 Richard Biener <rguenther@suse.de>
25858
25859 PR tree-optimization/56608
25860 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
25861 calls when vectorizing basic-blocks.
25862
25863 2013-03-13 Jakub Jelinek <jakub@redhat.com>
25864
25865 PR plugins/45078
25866 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
25867 tm_file.
25868
25869 2013-03-12 Jakub Jelinek <jakub@redhat.com>
25870
25871 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
25872
25873 2013-03-11 Jan Hubicka <jh@suse.cz>
25874
25875 PR lto/56557
25876 * lto-streamer-out.c (output_symbol_p): Skip references from
25877 constructors of external variables.
25878
25879 2013-03-11 Jan Hubicka <jh@suse.cz>
25880
25881 PR middle-end/56571
25882 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
25883 from pseudos.
25884 * emit-rtl.c (verify_rtx_sharing): Likewise.
25885 (copy_insn_1): Likewise.
25886 * rtl.c (copy_rtx): Likewise.
25887
25888 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
25889
25890 PR target/56591
25891 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
25892 output_operand_lossage message.
25893
25894 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
25895
25896 PR target/56470
25897 * arm.c (shift_op): Validate RTL pattern on the fly.
25898 (arm_print_operand, case 'S'): Don't use shift_operator to validate
25899 the RTL.
25900
25901 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25902
25903 PR target/56347
25904 * config/pa/pa.md (call_value): Check for calls to powf and direct to
25905 new call patterns that clobber %fr12.
25906 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
25907 split and postreload patterns.
25908 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
25909 registers %fr12 and %fr12R as call used.
25910
25911 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
25912
25913 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
25914 (canon_address, record_store, replace_read, check_mem_read_rtx,
25915 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
25916 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
25917 rest_of_handle_dse): Likewise.
25918
25919 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
25920
25921 PR middle-end/56524
25922 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
25923 Add base_optabs.
25924 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
25925 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
25926 (save_optabs_if_changed): Replace with...
25927 (init_tree_optimization_optabs): ...this.
25928 * optabs.c (save_optabs_if_changed): Rename to...
25929 (init_tree_optimization_optabs): ...this. Take the optimization node
25930 as argument. Do nothing if the base optabs are already correct.
25931 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
25932 to recompute optabs.
25933 * function.h (function): Remove optabs field.
25934 * function.c (invoke_set_current_function_hook): Call
25935 init_tree_optimization_optabs. Use the result to initialize
25936 this_fn_optabs.
25937
25938 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
25939
25940 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
25941 if GTMA_HAS_NO_INSTRUMENTATION.
25942 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
25943 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
25944 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
25945 * gimple-pretty-print.c (dump_gimple_transaction): Handle
25946 GTMA_HAS_NO_INSTRUMENTATION.
25947
25948 2013-03-08 Jakub Jelinek <jakub@redhat.com>
25949
25950 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
25951 libasan_preinit.o.
25952
25953 2013-03-08 Marek Polacek <polacek@redhat.com>
25954 Jakub Jelinek <jakub@redhat.com>
25955
25956 PR tree-optimization/56478
25957 * predict.c (is_comparison_with_loop_invariant_p): Change the
25958 type of loop_step to tree.
25959 (predict_loops): Adjust.
25960 (predict_iv_comparison): Perform the computations on double_ints.
25961
25962 2013-03-08 Richard Biener <rguenther@suse.de>
25963
25964 PR tree-optimization/56570
25965 * tree-cfg.c (verify_expr_location_1): Verify locations for
25966 DECL_DEBUG_EXPR.
25967 * tree-sra.c (create_access_replacement): Strip locations
25968 from DECL_DEBUG_EXPRs.
25969
25970 2013-03-08 Richard Biener <rguenther@suse.de>
25971
25972 * tree-inline.c (expand_call_inline): Do not associate
25973 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
25974 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
25975
25976 2013-03-08 Richard Biener <rguenther@suse.de>
25977
25978 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
25979 or block changes with -Og. Fix for location / block encoding
25980 changes and PHI arguments with locations.
25981
25982 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
25983
25984 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
25985 for all counters.
25986 (struct output_info): Likewise.
25987 (register_overhead): Remove bad gcc_assert.
25988 (bitmap_find_bit): If there is only a single bitmap element, do not
25989 count a miss as a search.
25990 (print_statistics): Update for counter type changes.
25991 (dump_bitmap_statistics): Likewise. Print headers such that they
25992 are properly lined up with the printed counters.
25993
25994 2013-03-07 Jakub Jelinek <jakub@redhat.com>
25995
25996 PR tree-optimization/56559
25997 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
25998 check that it has only a single use.
25999
26000 2013-03-07 Richard Biener <rguenther@suse.de>
26001
26002 * doc/invoke.texi (fwhole-program): Discourage use in combination
26003 with -flto.
26004
26005 2013-03-06 Jakub Jelinek <jakub@redhat.com>
26006
26007 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
26008
26009 PR tree-optimization/56539
26010 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
26011 instead of GSI_CONTINUE_LINKING as last argument to
26012 force_gimple_operand_gsi. Adjust function comment.
26013
26014 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
26015 aarch64-cores.def.
26016
26017 PR middle-end/56548
26018 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
26019 promoted mode, convert the result back to the original mode.
26020
26021 2013-03-06 Richard Biener <rguenther@suse.de>
26022
26023 PR middle-end/56294
26024 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
26025 (insert_updated_phi_nodes_compare_uids): New function.
26026 (update_ssa): Sort symbols_to_rename after UID before
26027 traversing it to insert PHI nodes.
26028
26029 2013-03-06 Richard Biener <rguenther@suse.de>
26030
26031 PR middle-end/50494
26032 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
26033 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
26034
26035 Revert
26036 2013-02-13 Richard Biener <rguenther@suse.de>
26037
26038 PR lto/50494
26039 * varasm.c (output_constant_def_1): Get the decl representing
26040 the constant as argument.
26041 (output_constant_def): Wrap output_constant_def_1.
26042 (make_decl_rtl): Use output_constant_def_1 with the decl
26043 representing the constant.
26044 (build_constant_desc): Optionally re-use a decl already
26045 representing the constant.
26046 (tree_output_constant_def): Adjust.
26047
26048 2013-03-06 Joey Ye <joey.ye@arm.com>
26049
26050 PR lto/50293
26051 * gcc.c (convert_white_space): New function.
26052 (main): Handles white space in function name.
26053
26054 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
26055
26056 PR target/56529
26057 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
26058 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
26059 to SH_DIV_CALL_TABLE for TARGET_SH2.
26060 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
26061 list.
26062 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
26063 call-table options.
26064
26065 2013-03-05 Sterling Augustine <saugustine@google.com>
26066 Cary Coutant <ccoutant@google.com>
26067
26068 PR debug/55364
26069 * dwarf2out.c (resolve_addr): Don't call
26070 remove_loc_list_addr_table_entries a second time for the same
26071 expression.
26072
26073 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26074
26075 PR debug/56510
26076 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
26077 (avoid_complex_debug_insns): New function.
26078 (expand_debug_locations): Call it.
26079
26080 PR rtl-optimization/56484
26081 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
26082 lifetimes of hard registers on small register class machines.
26083
26084 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26085
26086 * config/microblaze/microblaze-protos.h: Rename
26087 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
26088 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
26089 fast_interrupt.
26090 (microblaze_fast_interrupt_function_p): New function.
26091 (microblaze_is_interrupt_handler): Rename to
26092 microblaze_is_interrupt_variant and add fast_interrupt check.
26093 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
26094 (save_restore_insns): Likewise.
26095 (compute_frame_size): Likewise.
26096 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
26097 (microblaze_globalize_label): Likewise.
26098 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
26099 * config/microblaze/microblaze.md: Use wrapper
26100 microblaze_is_interrupt_variant.
26101
26102 2013-03-05 Kai Tietz <ktietz@redhat.com>
26103
26104 * sdbout.c (sdbout_one_type): Switch to current function's section
26105 supporting cold/hot.
26106
26107 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26108
26109 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
26110 -mxl-reorder.
26111
26112 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26113
26114 PR middle-end/56461
26115 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
26116 if VALGRIND_GET_VBITS is defined, temporarily make object
26117 memory all defined, and restore previous valgrind addressability
26118 and definability afterwards. Free this_object at the end.
26119
26120 PR middle-end/56461
26121 * lra.c (lra): Call lra_clear_live_ranges if live_p,
26122 right before calling lra_create_live_ranges, also call it
26123 when clearing live_p. Only call lra_clear_live_ranges
26124 at the end if live_p.
26125
26126 PR middle-end/56461
26127 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
26128
26129 2013-03-05 Richard Biener <rguenther@suse.de>
26130
26131 PR tree-optimization/56521
26132 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
26133 value-id.
26134
26135 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
26136
26137 PR c++/55135
26138 * except.h (remove_unreachable_eh_regions): New prototype.
26139 * except.c (remove_eh_handler_splicer): New function, split out
26140 of remove_eh_handler.
26141 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
26142 warning about running it on many EH regions one at a time.
26143 (remove_unreachable_eh_regions_worker): New function, walk the
26144 EH tree in depth-first order and remove non-marked regions.
26145 (remove_unreachable_eh_regions): New function.
26146 * tree-eh.c (mark_reachable_handlers): New function, split out
26147 from remove_unreachable_handlers.
26148 (remove_unreachable_handlers): Use mark_reachable_handlers and
26149 remove_unreachable_eh_regions.
26150 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
26151 and remove_unreachable_eh_regions.
26152
26153 2013-03-05 Richard Biener <rguenther@suse.de>
26154
26155 PR middle-end/56525
26156 * loop-init.c (fix_loop_structure): Remove loops in two stages,
26157 not freeing them until the end.
26158
26159 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26160
26161 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
26162
26163 2013-03-05 Richard Biener <rguenther@suse.de>
26164
26165 PR tree-optimization/56270
26166 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
26167 of loads after scheduling an SLP instance.
26168
26169 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26170
26171 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
26172 tic6x.exp.
26173 (check_gcc_parallelize): Run guality.exp as a separate job from
26174 vect.exp with unsorted.exp and $(dg_target_exps) separately from
26175 struct-layout-1.exp with stackalign.exp.
26176
26177 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
26178
26179 PR middle-end/56461
26180 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
26181 load_index sbitmap even if some bit in it isn't set.
26182
26183 PR middle-end/56461
26184 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
26185 (discover_iteration_bound_by_body_walk): Change queues to
26186 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
26187 spelling in comment. Call safe_push on queues[bound_index] directly.
26188 Release queues[queue_index] in every iteration unconditionally.
26189 Release bounds vector.
26190
26191 PR middle-end/56461
26192 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
26193 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
26194 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
26195 inner_phis vector.
26196
26197 2013-03-05 Richard Biener <rguenther@suse.de>
26198
26199 PR lto/56515
26200 * tree-inline.c (remap_blocks_to_null): New function.
26201 (expand_call_inline): When expanding a call stmt without
26202 an associated block inline remap all callee blocks to NULL.
26203
26204 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26205
26206 PR rtl-optimization/56494
26207 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
26208 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
26209 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
26210
26211 PR middle-end/56461
26212 * sel-sched-ir.c (free_sched_pools): Release
26213 succs_info_pool.stack[succs_info_pool.max_top] vectors too
26214 if succs_info_pool.max_top isn't -1.
26215
26216 PR bootstrap/56509
26217 * opts.c (opts_obstack, opts_concat): Moved to...
26218 * opts-common.c (opts_obstack, opts_concat): ... here.
26219
26220 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26221
26222 PR middle-end/56461
26223 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
26224
26225 2013-03-04 Martin Jambor <mjambor@suse.cz>
26226
26227 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
26228 all appropriate places.
26229
26230 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
26231
26232 PR tree-optimization/56424
26233 * ipa-split.c (split_function): Do not set the RSO flag if result is
26234 not by reference and its type is a register type.
26235
26236 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
26237
26238 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
26239 (microblaze_legitimate_pic_operand): Likewise
26240 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
26241 new function microblaze_legitimate_pic_operand
26242 * config/microblaze/microblaze-protos.h
26243 (microblaze_legitimate_pic_operand): Declare.
26244
26245 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26246
26247 * config/microblaze/predicates.md (call_insn_simple_operand):
26248 New predicate for supported rtx code types.
26249 * config/microblaze/microblaze.md (call_internal1): Use
26250 call_insn_simple_operand predicate.
26251
26252 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26253
26254 PR middle-end/56461
26255 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
26256 partitions.ordered_remove.
26257
26258 PR middle-end/56461
26259 * tree-vect-stmts.c (vectorizable_conversion): Don't call
26260 vec_oprnds0.create (1) for modifier == NONE.
26261
26262 PR middle-end/56461
26263 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
26264 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
26265 vec_oprnds1 right before pushing anything to it for
26266 scalar_shift_arg.
26267
26268 PR middle-end/56461
26269 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
26270 set nbbs to 0 instead of having separate code path.
26271 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
26272 instead of false as last argument if returning NULL.
26273
26274 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
26275
26276 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
26277 the attribute is now called "target" instead of "option".
26278 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
26279 * doc/tm.texi.in (Target Attributes): Likewise document the correct
26280 attribute/pragma name for TARGET_OPTION_VALID_P and
26281 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
26282 * doc/tm.texi: Regenerated.
26283
26284 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
26285
26286 * config/microblaze/microblaze.c:
26287 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
26288 * config/microblaze/microblaze.h: Add -mxl-reorder to
26289 DRIVER_SELF_SPECS.
26290 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
26291 instructions emitted if TARGET_REORDER.
26292 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
26293 or 0 for -m/-mno case, but initialises as 2 to detect default use case
26294 separately.
26295
26296 2013-03-01 Xinliang David Li <davidxl@google.com>
26297
26298 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
26299 walk length.
26300
26301 2013-03-01 Jakub Jelinek <jakub@redhat.com>
26302
26303 PR middle-end/56461
26304 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
26305 vector even when returning true. Fix up function comment formatting.
26306
26307 PR middle-end/56461
26308 * ira-build.c (ira_loop_nodes_count): New variable.
26309 (create_loop_tree_nodes): Initialize it.
26310 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
26311
26312 PR middle-end/56461
26313 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
26314 method on dr_chain and result_chain.
26315 * tree-vect-stmts.c (vectorizable_store): Only call
26316 result_chain.create if j == 0.
26317
26318 PR middle-end/56461
26319 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
26320 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
26321 before overwriting it.
26322
26323 2013-03-01 Tobias Burnus <burnus@net-b.de>
26324
26325 * doc/extended.texi (C Extensions): Change order in @menu
26326 to match @node.
26327 (Other MIPS Built-in Functions): Move last MIPS entry before
26328 "picoChip Built-in Functions".
26329 (SH Built-in Functions): Move after RX Built-in Functions.
26330 * doc/gcc.texi (Introduction): Change order in @menu
26331 to match @node.
26332 * doc/md.texi (Constraints): Ditto.
26333 * gty.texi (Type Information): Ditto.
26334 (User-provided marking routines for template types): Make
26335 subsection.
26336 * doc/invoke.texi (AArch64 Options): Move before
26337 "Adapteva Epiphany Options".
26338
26339 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
26340 Jakub Jelinek <jakub@redhat.com>
26341
26342 PR sanitizer/56454
26343 * asan.c (gate_asan): Lookup no_sanitize_address instead of
26344 no_address_safety_analysis attribute.
26345 * doc/extend.texi (no_address_safety_attribute): Rename to
26346 no_sanitize_address attribute, mention no_address_safety_analysis
26347 attribute as deprecated alias.
26348
26349 2013-02-28 Jakub Jelinek <jakub@redhat.com>
26350
26351 PR middle-end/56461
26352 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
26353 type to vec<vec<tree> > *.
26354 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
26355 to be vec<tree> instead of vec<tree> *, set vec_defs
26356 to vNULL and call vec_defs.create (number_of_vects), adjust other
26357 uses of vec_defs.
26358 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
26359 vectorizable_condition): Adjust vect_get_slp_defs callers.
26360
26361 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26362
26363 * config/aarch64/aarch64.c
26364 (aarch64_float_const_representable): Remove unused variable.
26365
26366 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26367
26368 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
26369
26370 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26371
26372 * config/aarch64/aarch64-builtins.c
26373 (aarch64_init_simd_builtins): Make static.
26374
26375 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26376
26377 * config/aarch64/aarch64.c
26378 (aarch64_simd_make_constant): Make static.
26379
26380 2013-02-28 Martin Jambor <mjambor@suse.cz>
26381
26382 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
26383 with no initialization to the RHS of debug statements.
26384
26385 2013-02-28 Martin Jambor <mjambor@suse.cz>
26386
26387 PR tree-optimization/56294
26388 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
26389 Adjust dumping.
26390 (get_access_replacement): Do not call create_access_replacement.
26391 Assert a replacement exists.
26392 (get_repl_default_def_ssa_name): Create the replacement declaration
26393 itself.
26394
26395 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26396
26397 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
26398 final_end_function.
26399
26400 2013-02-28 Marek Polacek <polacek@redhat.com>
26401
26402 PR rtl-optimization/56466
26403 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
26404 if we're changing a loop.
26405 (peel_loops_completely): Likewise.
26406
26407 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
26408
26409 PR c++/55813
26410 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
26411
26412 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26413
26414 PR target/56445
26415 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
26416 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
26417 INTX_FTYPE_FX, FX_FTYPE_INTX.
26418 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
26419
26420 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26421
26422 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
26423 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
26424 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
26425 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
26426 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
26427 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
26428 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
26429 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
26430 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
26431 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
26432 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
26433 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
26434 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
26435 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
26436 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
26437 (avrxmega6): Increase max flash segments from 5 to 6.
26438 * config/avr/t-multilib: Regenerate.
26439 * config/avr/avr-tables.opt: Regenerate.
26440 * doc/avr-mmcu.texi: Regenerate.
26441
26442 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26443
26444 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
26445 (avr_device_to_arch): Rename to avr_device_to_ld.
26446 (avr_device_to_as): New prototype.
26447 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
26448 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
26449 * config/avr/driver-avr.c (avr_device_to_as): New.
26450 (avr_device_to_arch): Rename to avr_device_to_ld.
26451
26452 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26453
26454 PR middle-end/56461
26455 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
26456 method on dr_chain and result_chain.
26457
26458 PR middle-end/56461
26459 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
26460 pointer_set_destroy on not_executed_last_iteration.
26461
26462 PR middle-end/56461
26463 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
26464
26465 PR middle-end/56461
26466 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
26467 FOR_EACH_DEFINED_FUNCTION when freeing state.
26468
26469 PR middle-end/56461
26470 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
26471 pool_free.
26472 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
26473 overwriting it.
26474
26475 PR middle-end/56461
26476 * ipa-cp.c (decide_whether_version_node): Call vec_free on
26477 known_aggs[i].items and release known_aggs vector.
26478
26479 PR middle-end/56461
26480 * ipa-reference.c (propagate): Free node_info even for alias nodes.
26481
26482 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26483
26484 * config/microblaze/microblaze.c (microblaze_emit_compare):
26485 Use xor for EQ/NE comparisions.
26486 * config/microblaze/microblaze.md (cstoresf4): Add constraints
26487 (cbranchsf4): Adjust operator to comparison_operator.
26488
26489 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26490
26491 PR middle-end/56461
26492 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
26493 vector.
26494 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
26495 vec_safe_push, always update *slot.
26496 (redirect_edge_var_map_clear): Use vec_free.
26497 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
26498 (free_var_map_entry): Use vec_free.
26499 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
26500 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
26501
26502 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
26503
26504 PR middle-end/45472
26505 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
26506 when the may_trap_p bit of the exprs being merged differs.
26507 Reorder tests for speculativeness in the logical and operator.
26508
26509 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26510
26511 * incpath.c (add_standard_paths): Use reconcat instead of concat
26512 where appropriate and avoid leaking memory.
26513
26514 * opts.h: Include obstack.h.
26515 (opts_concat): New prototype.
26516 (opts_obstack): New declaration.
26517 * opts.c (opts_concat): New function.
26518 (opts_obstack): New variable.
26519 (init_options_struct): Call gcc_init_obstack on opts_obstack.
26520 (finish_options): Use opts_concat instead of concat
26521 and XOBNEWVEC instead of XNEWVEC.
26522 * opts-common.c (generate_canonical_option, decode_cmdline_option,
26523 generate_option): Likewise.
26524 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
26525 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
26526
26527 PR target/56455
26528 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
26529 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
26530
26531 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26532
26533 PR middle-end/56461
26534 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
26535
26536 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26537
26538 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
26539 (arm_block_move_unaligned_straight): Likewise.
26540 (arm_adjust_block_mem): Likewise.
26541
26542 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26543
26544 PR target/48901
26545 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
26546 temp, cond and label.
26547 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
26548
26549 PR target/52500
26550 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
26551 * config/c6x/c6x.h (dbx_register_map): Update declaration.
26552
26553 PR target/52501
26554 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
26555 of prologue/epilogue functions.
26556
26557 PR target/52550
26558 * config/tilegx/tilegx.c (tilegx_expand_prologue):
26559 Remove unused variable cfa_offset.
26560 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
26561
26562 PR target/54639
26563 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
26564 type promotion to unsigned.
26565
26566 PR target/54640
26567 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
26568 for HOST_WIDE_INT of 32 bit / same size as int.
26569 (arm_block_move_unaligned_straight): Likewise.
26570 (arm_adjust_block_mem): Likewise.
26571
26572 PR target/54662
26573 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
26574 ALL_CFLAGS.
26575
26576 2013-02-26 Marek Polacek <polacek@redhat.com>
26577
26578 PR tree-optimization/56426
26579 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
26580
26581 2013-02-26 Richard Biener <rguenther@suse.de>
26582
26583 PR target/56444
26584 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
26585 unused variable loops.
26586
26587 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26588
26589 PR tree-optimization/56448
26590 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
26591 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
26592 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
26593 later operands of the references, or even first operand for
26594 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
26595
26596 PR tree-optimization/56443
26597 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
26598 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
26599 to type_for_mode langhook.
26600
26601 2013-02-25 Matt Turner <mattst88@gmail.com>
26602
26603 * doc/invoke.texi: Document r4700.
26604
26605 2013-02-25 Richard Biener <rguenther@suse.de>
26606
26607 PR tree-optimization/56175
26608 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
26609 split out from ...
26610 (simplify_bitwise_binary): ... here. Also guard the conversion
26611 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
26612
26613 2013-02-25 Catherine Moore <clm@codesourcery.com>
26614
26615 Revert:
26616 2013-02-24 Catherine Moore <clm@codesourcery.com>
26617 Maciej W. Rozycki <macro@codesourcery.com>
26618 Tom de Vries <tom@codesourcery.com>
26619 Nathan Sidwell <nathan@codesourcery.com>
26620 Iain Sandoe <iain@codesourcery.com>
26621 Nathan Froyd <froydnj@codesourcery.com>
26622 Chao-ying Fu <fu@mips.com>
26623
26624 * doc/extend.texi: (micromips, nomicromips, nocompression):
26625 Document new function attributes.
26626 * doc/invoke.texi (minterlink-compressed, mmicromips,
26627 m14k, m14ke, m14kec): Document new options.
26628 (minterlink-mips16): Update documentation.
26629 * doc/md.texi (ZC, ZD): Document new constraints.
26630 * configure.ac (gcc_cv_as_micromips): Check if linker
26631 supports the .set micromips directive.
26632 * configure: Regenerate.
26633 * config.in: Regenerate.
26634 * config/mips/mips-tables.opt: Regenerate.
26635 * config/mips/micromips.md: New file.
26636 * constraints.md (ZC, AD): New constraints.
26637 * config/mips/predicates.md (movep_src_register): New predicate.
26638 (movep_src_operand): New predicate.
26639 (non_volatile_mem_operand): New predicate.
26640 * config/mips/mips.md (multimem): New type.
26641 (length): Differentiate between 17-bit and 18-bit branch offsets.
26642 (MOVEP1, MOVEP2): New mode iterator.
26643 (mov_<load>l): Use ZC constraint.
26644 (mov_<load>r): Likewise.
26645 (mov_<store>l): Likewise.
26646 (mov_<store>r): Likewise.
26647 (*branch_equality<mode>_inverted): Add microMIPS support.
26648 (*branch_equality<mode>): Likewise.
26649 (*jump_absolute): Likewise.
26650 (indirect_jump_<mode>): Likewise.
26651 (tablejump_<mode>): Likewise.
26652 (<optab>_internal): Likewise.
26653 (sibcall_internal): Likewise.
26654 (sibcall_value_internal): Likewise.
26655 (prefetch): Use constraint ZD.
26656 * config/mips/mips.opt (minterlink-compressed): New option.
26657 (minterlink-mips16): Now an alias for minterlink-compressed.
26658 (mmicromips): New option.
26659 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26660 (compare_and_swap_12): Likewise.
26661 (sync_add<mode>): Likewise.
26662 (sync_<optab>_12): Likewise.
26663 (sync_old_<optab>_12): Likewise.
26664 (sync_new_<optab>_12): Likewise.
26665 (sync_nand_12): Likewise.
26666 (sync_old_nand_12): Likewise.
26667 (sync_new_nand_12): Likewise.
26668 (sync_sub<mode>): Likewise.
26669 (sync_old_add<mode>): Likewise.
26670 (sync_old_sub<mode>): Likewise.
26671 (sync_new_add<mode>): Likewise.
26672 (sync_new_sub<mode>): Likewise.
26673 (sync_<optab><mode>): Likewise.
26674 (sync_old_<optab><mode>): Likewise.
26675 (sync_new_<optab><mode>): Likewise.
26676 (sync_nand<mode>): Likewise.
26677 (sync_old_nand<mode>): Likewise.
26678 (sync_new_nand<mode>): Likewise.
26679 (sync_lock_test_and_set<mode>): Likewise.
26680 (test_and_set_12): Likewise.
26681 (atomic_compare_and_swap<mode>): Likewise.
26682 (atomic_exchange<mode>_llsc): Likewise.
26683 (atomic_fetch_add<mode>_llsc): Likewise.
26684 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26685 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26686 (umips_save_restore_pattern_p): Likewise.
26687 (umips_load_store_pair_p): Likewise.
26688 (umips_output_load_store_pair): Likewise.
26689 (umips_movep_target_p): Likewise.
26690 (umips_12bit_offset_address_p): Likewise.
26691 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26692 (mips_base_mips16): Rename this...
26693 (mips_base_compression_flags): ...to this. Update all uses.
26694 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26695 (mips_mips16_decl_p): Delete.
26696 (mips_nomips16_decl_p): Delete.
26697 (mips_get_compress_on_flags): New function.
26698 (mips_get_compress_off_flags): New function.
26699 (mips_get_compress_mode): New function.
26700 (mips_get_compress_on_name): New function.
26701 (mips_get_compress_off_name): New function.
26702 (mips_insert_attributes): Support multiple compression types.
26703 (mips_merge_decl_attributes): Likewise.
26704 (umips_12bit_offset_address_p): New function.
26705 (mips_start_function_definition): Emit .set micromips directive.
26706 (mips_call_may_need_jalx_p): New function.
26707 (mips_function_ok_for_sibcall): Add microMIPS support.
26708 (mips_print_operand_punctuation): Support short delay slots and
26709 compact jumps.
26710 (umips_swm_mask, umips_swm_encoding): New.
26711 (umips_build_save_restore): New function.
26712 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26713 (was_mips16_p): Remove.
26714 (old_compression_mode): New.
26715 (mips_set_compression_mode): New function.
26716 (mips_set_current_function): Add microMIPS support.
26717 (mips_option_override): Likewise.
26718 (umips_save_restore_pattern_p): New function.
26719 (umips_output_save_restore): New function.
26720 (umips_load_store_pair_p_1): New function.
26721 (umips_load_store_pair_p): New function.
26722 (umips_output_load_store_pair_1): New function.
26723 (umips_output_load_store_pair): New function.
26724 (umips_movep_target_p) New function.
26725 (mips_prepare_pch_save): Add microMIPS support.
26726 * config/mips/mips.h (TARGET_COMPRESSION): New.
26727 (TARGET_CPU_CPP_BUILTINS): Update macro
26728 to use new compression flags and to support microMIPS.
26729 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26730 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26731 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26732 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26733 (ASM_SPEC): Support mmicromips and mno-micromips.
26734 (M16STORE_REG_P): New macro.
26735 (MIPS_CALL): Support TARGET_MICROMIPS.
26736 (MICROMIPS_J): New macro.
26737 (mips_base_mips16): Rename this...
26738 (mips_base_compression_flags): ...to this.
26739 (UMIPS_12BIT_OFFSET_P): New macro.
26740 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26741 (MULTILIB_DIRNAMES): Likewise.
26742
26743 2013-02-25 Tom de Vries <tom@codesourcery.com>
26744
26745 PR rtl-optimization/56131
26746 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
26747 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
26748 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
26749
26750 2013-02-25 Tobias Burnus <burnus@net-b.de>
26751
26752 * doc/invoke.texi (-fsanitize=): Move from optimization
26753 to debugging options.
26754
26755 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26756
26757 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
26758
26759 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26760 Alexander Monakov <amonakov@ispras.ru>
26761
26762 PR middle-end/56077
26763 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
26764 flush pending lists also on non-jumps. Adjust comment.
26765
26766 2013-02-24 Catherine Moore <clm@codesourcery.com>
26767 Maciej W. Rozycki <macro@codesourcery.com>
26768 Tom de Vries <tom@codesourcery.com>
26769 Nathan Sidwell <nathan@codesourcery.com>
26770 Iain Sandoe <iain@codesourcery.com>
26771 Nathan Froyd <froydnj@codesourcery.com>
26772 Chao-ying Fu <fu@mips.com>
26773
26774 * doc/extend.texi: (micromips, nomicromips, nocompression):
26775 Document new function attributes.
26776 * doc/invoke.texi (minterlink-compressed, mmicromips,
26777 m14k, m14ke, m14kec): Document new options.
26778 (minterlink-mips16): Update documentation.
26779 * doc/md.texi (ZC, ZD): Document new constraints.
26780 * configure.ac (gcc_cv_as_micromips): Check if linker
26781 supports the .set micromips directive.
26782 * configure: Regenerate.
26783 * config.in: Regenerate.
26784 * config/mips/mips-tables.opt: Regenerate.
26785 * config/mips/micromips.md: New file.
26786 * constraints.md (ZC, AD): New constraints.
26787 * config/mips/predicates.md (movep_src_register): New predicate.
26788 (movep_src_operand): New predicate.
26789 (non_volatile_mem_operand): New predicate.
26790 * config/mips/mips.md (multimem): New type.
26791 (length): Differentiate between 17-bit and 18-bit branch offsets.
26792 (MOVEP1, MOVEP2): New mode iterator.
26793 (mov_<load>l): Use ZC constraint.
26794 (mov_<load>r): Likewise.
26795 (mov_<store>l): Likewise.
26796 (mov_<store>r): Likewise.
26797 (*branch_equality<mode>_inverted): Add microMIPS support.
26798 (*branch_equality<mode>): Likewise.
26799 (*jump_absolute): Likewise.
26800 (indirect_jump_<mode>): Likewise.
26801 (tablejump_<mode>): Likewise.
26802 (<optab>_internal): Likewise.
26803 (sibcall_internal): Likewise.
26804 (sibcall_value_internal): Likewise.
26805 (prefetch): Use constraint ZD.
26806 * config/mips/mips.opt (minterlink-compressed): New option.
26807 (minterlink-mips16): Now an alias for minterlink-compressed.
26808 (mmicromips): New option.
26809 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26810 (compare_and_swap_12): Likewise.
26811 (sync_add<mode>): Likewise.
26812 (sync_<optab>_12): Likewise.
26813 (sync_old_<optab>_12): Likewise.
26814 (sync_new_<optab>_12): Likewise.
26815 (sync_nand_12): Likewise.
26816 (sync_old_nand_12): Likewise.
26817 (sync_new_nand_12): Likewise.
26818 (sync_sub<mode>): Likewise.
26819 (sync_old_add<mode>): Likewise.
26820 (sync_old_sub<mode>): Likewise.
26821 (sync_new_add<mode>): Likewise.
26822 (sync_new_sub<mode>): Likewise.
26823 (sync_<optab><mode>): Likewise.
26824 (sync_old_<optab><mode>): Likewise.
26825 (sync_new_<optab><mode>): Likewise.
26826 (sync_nand<mode>): Likewise.
26827 (sync_old_nand<mode>): Likewise.
26828 (sync_new_nand<mode>): Likewise.
26829 (sync_lock_test_and_set<mode>): Likewise.
26830 (test_and_set_12): Likewise.
26831 (atomic_compare_and_swap<mode>): Likewise.
26832 (atomic_exchange<mode>_llsc): Likewise.
26833 (atomic_fetch_add<mode>_llsc): Likewise.
26834 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26835 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26836 (umips_save_restore_pattern_p): Likewise.
26837 (umips_load_store_pair_p): Likewise.
26838 (umips_output_load_store_pair): Likewise.
26839 (umips_movep_target_p): Likewise.
26840 (umips_12bit_offset_address_p): Likewise.
26841 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26842 (mips_base_mips16): Rename this...
26843 (mips_base_compression_flags): ...to this. Update all uses.
26844 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26845 (mips_mips16_decl_p): Delete.
26846 (mips_nomips16_decl_p): Delete.
26847 (mips_get_compress_on_flags): New function.
26848 (mips_get_compress_off_flags): New function.
26849 (mips_get_compress_mode): New function.
26850 (mips_get_compress_on_name): New function.
26851 (mips_get_compress_off_name): New function.
26852 (mips_insert_attributes): Support multiple compression types.
26853 (mips_merge_decl_attributes): Likewise.
26854 (umips_12bit_offset_address_p): New function.
26855 (mips_start_function_definition): Emit .set micromips directive.
26856 (mips_call_may_need_jalx_p): New function.
26857 (mips_function_ok_for_sibcall): Add microMIPS support.
26858 (mips_print_operand_punctuation): Support short delay slots and
26859 compact jumps.
26860 (umips_swm_mask, umips_swm_encoding): New.
26861 (umips_build_save_restore): New function.
26862 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26863 (was_mips16_p): Remove.
26864 (old_compression_mode): New.
26865 (mips_set_compression_mode): New function.
26866 (mips_set_current_function): Add microMIPS support.
26867 (mips_option_override): Likewise.
26868 (umips_save_restore_pattern_p): New function.
26869 (umips_output_save_restore): New function.
26870 (umips_load_store_pair_p_1): New function.
26871 (umips_load_store_pair_p): New function.
26872 (umips_output_load_store_pair_1): New function.
26873 (umips_output_load_store_pair): New function.
26874 (umips_movep_target_p) New function.
26875 (mips_prepare_pch_save): Add microMIPS support.
26876 * config/mips/mips.h (TARGET_COMPRESSION): New.
26877 (TARGET_CPU_CPP_BUILTINS): Update macro
26878 to use new compression flags and to support microMIPS.
26879 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26880 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26881 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26882 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26883 (ASM_SPEC): Support mmicromips and mno-micromips.
26884 (M16STORE_REG_P): New macro.
26885 (MIPS_CALL): Support TARGET_MICROMIPS.
26886 (MICROMIPS_J): New macro.
26887 (mips_base_mips16): Rename this...
26888 (mips_base_compression_flags): ...to this.
26889 (UMIPS_12BIT_OFFSET_P): New macro.
26890 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26891 (MULTILIB_DIRNAMES): Likewise.
26892
26893 2013-02-24 Jakub Jelinek <jakub@redhat.com>
26894
26895 PR target/52555
26896 * target-globals.c (save_target_globals): For init_reg_sets and
26897 target_reinit remporarily set this_fn_optabs to this_target_optabs.
26898
26899 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
26900
26901 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
26902 * config/aarch64/t-aarch64
26903 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
26904
26905 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
26906
26907 PR inline-asm/56148
26908 * lra-constraints.c (process_alt_operands): Reload operand
26909 conflicting with earlier clobber only if no more other conflicting
26910 operands.
26911
26912 2013-02-22 Jakub Jelinek <jakub@redhat.com>
26913
26914 PR sanitizer/56393
26915 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
26916 if not linking a shared library.
26917
26918 2013-02-22 Seth LaForge <sethml@google.com>
26919
26920 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
26921
26922 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
26923
26924 * config/arm/arm.md (split for extendsidi): Update condition.
26925 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
26926 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
26927 (qhs_zextenddi_cstr): Likewise.
26928
26929 2013-02-21 Jakub Jelinek <jakub@redhat.com>
26930
26931 PR middle-end/56420
26932 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
26933 avoid signed wrapping.
26934 (expand_mult): Handle properly multiplication by
26935 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
26936 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
26937 in the compiler if coeff is HOST_WIDE_INT_MIN.
26938 (expand_divmod): Don't make ext_op1 static, change it's type to
26939 uhwi. Avoid undefined behavior in -INTVAL (op1).
26940
26941 PR rtl-optimization/50339
26942 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
26943 field.
26944 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
26945 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
26946 into splitting_ashiftrt field.
26947 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
26948 ASHIFTRT.
26949 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
26950 choices.
26951
26952 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
26953
26954 PR middle-end/56108
26955 * trans-mem.c (execute_tm_mark): Do not expand transactions that
26956 are sure to go irrevocable.
26957
26958 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
26959
26960 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
26961 scalars are valid operands.
26962
26963 2013-02-21 Martin Jambor <mjambor@suse.cz>
26964
26965 PR tree-optimization/56310
26966 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
26967 only matching indices and non-negative final offsets.
26968 (intersect_aggregates_with_edge): Pass src_idx to
26969 agg_replacements_to_vector. Pass src_idx insstead of index to
26970 intersect_with_agg_replacements.
26971
26972 2013-02-21 Martin Jambor <mjambor@suse.cz>
26973
26974 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
26975 instead of hard-wired defaults.
26976
26977 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
26978
26979 * doc/invoke.texi (MIPS Options): Update documentation of the
26980 floating-point multiply-accumulate instruction restrictions.
26981
26982 2013-02-21 Kostya Serebryany <kcc@google.com>
26983
26984 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
26985 asan_shadow_offset on x86_64 linux.
26986
26987 2013-02-21 Richard Biener <rguenther@suse.de>
26988
26989 PR tree-optimization/56415
26990 Revert
26991 2013-02-11 Richard Biener <rguenther@suse.de>
26992
26993 PR tree-optimization/56273
26994 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
26995 first VRP run.
26996
26997 2013-02-21 Jakub Jelinek <jakub@redhat.com>
26998
26999 PR bootstrap/56258
27000 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
27001 instead of @itemx.
27002
27003 PR inline-asm/56405
27004 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
27005 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
27006
27007 2013-02-20 Jan Hubicka <jh@suse.cz>
27008
27009 PR tree-optimization/56265
27010 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
27011 when target is referenced for first time.
27012
27013 2013-02-20 Richard Biener <rguenther@suse.de>
27014
27015 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
27016 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
27017 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
27018 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
27019 not return anything.
27020 (rename_ssa_copies): Do not remove unused locals.
27021 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
27022 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
27023 * passes.c (execute_function_todo): Do not schedule unused locals
27024 removal if cleanup_tree_cfg did something.
27025 * tree-ssa-live.c (remove_unused_locals): Dump statistics
27026 about the number of removed locals.
27027
27028 2013-02-20 Richard Biener <rguenther@suse.de>
27029
27030 PR tree-optimization/56398
27031 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
27032
27033 2013-02-20 Martin Jambor <mjambor@suse.cz>
27034
27035 PR tree-optimization/55334
27036 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
27037 restricted pointers to arrays.
27038
27039 2013-02-20 Richard Biener <rguenther@suse.de>
27040 Jakub Jelinek <jakub@redhat.com>
27041
27042 PR tree-optimization/56396
27043 * tree-ssa-ccp.c (n_const_val): New static variable.
27044 (get_value): Return NULL for SSA names we don't have a lattice
27045 entry for.
27046 (ccp_initialize): Initialize n_const_val.
27047 * tree-ssa-copy.c (n_copy_of): New static variable.
27048 (init_copy_prop): Initialize n_copy_of.
27049 (get_value): Return NULL_TREE for SSA names we don't have a
27050 lattice entry for.
27051
27052 2013-02-20 Martin Jambor <mjambor@suse.cz>
27053
27054 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
27055
27056 2013-02-20 Richard Biener <rguenther@suse.de>
27057
27058 * genpreds.c (write_lookup_constraint): Do not compare first
27059 letter of the constraint again.
27060
27061 2013-02-20 Richard Biener <rguenther@suse.de>
27062
27063 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
27064 and ceil_log2.
27065 (get_use_iv_cost): Terminate hashtable walk when coming across
27066 an empty entry.
27067
27068 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
27069
27070 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
27071 reassociation for avx2 targets.
27072
27073 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27074
27075 * config/microblaze/microblaze.c: microblaze_has_clz = 0
27076 Add version check for v8.10.a to enable microblaze_has_clz
27077 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
27078 version and TARGET_PATTERN_COMPARE check
27079 * config/microblaze/microblaze.md: New clzsi2 instruction
27080
27081 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27082
27083 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
27084 function before branching.
27085
27086 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27087
27088 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
27089 DUMP_INSN_RTX_UID.
27090 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
27091
27092 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27093
27094 PR middle-end/55889
27095 * sel-sched.c: Include ira.h.
27096 (implicit_clobber_conflict_p): New function.
27097 (moveup_expr): Use it.
27098 * Makefile.in (sel-sched.o): Depend on ira.h.
27099
27100 2013-02-19 Richard Biener <rguenther@suse.de>
27101
27102 PR tree-optimization/56384
27103 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
27104 (vn_hash_type): Split out from ...
27105 (vn_hash_constant_with_type): ... here.
27106 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
27107 (vn_phi_eq): Compare types from vn_phi_s structure.
27108 (vn_phi_lookup): Populate vn_phi_s type.
27109 (vn_phi_insert): Likewise.
27110
27111 2013-02-19 Jakub Jelinek <jakub@redhat.com>
27112
27113 PR tree-optimization/56350
27114 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
27115 if haven't found reduction or nested cycle operand, rather than
27116 asserting we must find it.
27117
27118 PR tree-optimization/56381
27119 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
27120 to fold_build3.
27121
27122 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
27123 Jakub Jelinek <jakub@redhat.com>
27124
27125 PR target/52555
27126 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
27127 (swap_optab_enable): Same.
27128 (init_all_optabs): Use argument instead of global.
27129 * tree.h (struct tree_optimization_option): New field target_optabs.
27130 * expr.h (init_all_optabs): Add argument to prototype.
27131 (TREE_OPTIMIZATION_OPTABS): New.
27132 (save_optabs_if_changed): Protoize.
27133 * optabs.h: Declare this_fn_optabs.
27134 * optabs.c (save_optabs_if_changed): New.
27135 Declare this_fn_optabs.
27136 (init_optabs): Add argument to init_all_optabs() call.
27137 * function.c (invoke_set_current_function_hook): Handle per
27138 function optabs.
27139 * function.h (struct function): New field optabs.
27140 * config/mips/mips.c (mips_set_mips16_mode): Handle when
27141 optimization_current_node has changed.
27142 * target-globals.h (save_target_globals_default_opts): Protoize.
27143 * target-globals.c (save_target_globals_default_opts): New.
27144
27145 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27146
27147 PR target/56347
27148 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
27149 registers %fr12 and %fr12R as call used.
27150
27151 PR target/56214
27152 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
27153 and HImode, require all displacements to be an integer multiple of
27154 their mode size.
27155 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
27156 only allow QImode and HImode when reload is in progress and strict is
27157 true. Likewise for symbolic addresses. Use base14_operand to check
27158 displacements in REG+BASE addresses.
27159
27160 2013-02-18 Richard Biener <rguenther@suse.de>
27161
27162 PR tree-optimization/56366
27163 * tree-vect-loop.c (get_initial_def_for_induction): Properly
27164 handle sign-conversion of outer-loop initial induction value.
27165
27166 2013-02-18 Richard Biener <rguenther@suse.de>
27167
27168 PR middle-end/56349
27169 * cfghooks.c (merge_blocks): If we merge a latch into another
27170 block adjust references to it.
27171 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
27172 (verify_loop_structure): Verify that a recorded latch is in fact
27173 a latch.
27174
27175 2013-02-18 Richard Biener <rguenther@suse.de>
27176
27177 PR tree-optimization/56321
27178 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
27179 order SSA name release and virtual operand unlinking.
27180
27181 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27182
27183 * config/microblaze/microblaze.md (save_stack_block): Define.
27184 (restore_stack_block): Likewise.
27185
27186 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27187
27188 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
27189 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
27190 * config/microblaze/microblaze.c (microblaze_option_override):
27191 Bail out early for PIC modes when target does not support PIC.
27192
27193 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27194
27195 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
27196 Replace with a microblaze version.
27197 (microblaze_trampoline_init): Adapt for microblaze.
27198 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
27199 microblaze.
27200
27201 2013-02-16 Jakub Jelinek <jakub@redhat.com>
27202 Dodji Seketeli <dodji@redhat.com>
27203
27204 PR asan/56330
27205 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
27206 (instrument_mem_region_access): Do not forget to always put
27207 instrumentation of the of 'base' and 'base + len' in a "if (len !=
27208 0) statement, even for cases where either 'base' or 'base + len'
27209 are not instrumented -- because they have been previously
27210 instrumented. Simplify the logic by putting all the statements
27211 instrument 'base + len' inside a sequence, and then insert that
27212 sequence right before the current insertion point. Then, to
27213 instrument 'base + len', just get an iterator on that statement.
27214 And do not forget to update the pointer to iterator the function
27215 received as argument.
27216
27217 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
27218
27219 PR rtl-optimization/56348
27220 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
27221
27222 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
27223
27224 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
27225 (clean_graph_dump_file): Pass base to start_graph_dump.
27226
27227 2013-02-14 Richard Henderson <rth@redhat.com>
27228
27229 PR target/55941
27230 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
27231
27232 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
27233
27234 * collect2-aix.h: Define F_LOADONLY.
27235
27236 2013-02-14 Richard Biener <rguenther@suse.de>
27237
27238 PR lto/50494
27239 * varasm.c (output_constant_def_1): Get the decl representing
27240 the constant as argument.
27241 (output_constant_def): Wrap output_constant_def_1.
27242 (make_decl_rtl): Use output_constant_def_1 with the decl
27243 representing the constant.
27244 (build_constant_desc): Optionally re-use a decl already
27245 representing the constant.
27246 (tree_output_constant_def): Adjust.
27247
27248 2013-02-14 Dodji Seketeli <dodji@redhat.com>
27249
27250 Fix an asan crash
27251 * asan.c (instrument_builtin_call): Really put the length of the
27252 second source argument into src1_len.
27253
27254 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27255
27256 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
27257 argument. If it is false, don't create edge from then_bb to
27258 fallthru_bb.
27259 (insert_if_then_before_iter): Pass true to it.
27260 (build_check_stmt): Pass false to it.
27261 (transform_statements): Flush hash table only on extended basic
27262 block boundaries, rather than at the beginning of every bb.
27263 Don't flush hash table on nonfreeing_call_p calls.
27264 * tree-flow.h (nonfreeing_call_p): New prototype.
27265 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
27266
27267 2013-02-13 David S. Miller <davem@davemloft.net>
27268
27269 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
27270
27271 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
27272
27273 PR target/56184
27274 * ira.c (max_regno_before_ira): Move from ...
27275 (ira): ... here.
27276 (fix_reg_equiv_init): Use max_regno_before_ira instead of
27277 vec_safe_length.
27278
27279 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27280
27281 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
27282
27283 2013-02-13 Richard Biener <rguenther@suse.de>
27284
27285 PR lto/56295
27286 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
27287 globals in MEM_REFs.
27288
27289 2013-02-13 Richard Biener <rguenther@suse.de>
27290
27291 * loop-init.c (loop_optimizer_init): Clear loop state when
27292 re-initializing preserved loops.
27293 * loop-unswitch.c (unswitch_single_loop): Return whether
27294 we unswitched the loop. Do not verify loop state here.
27295 (unswitch_loops): When we unswitched a loop discover new loops.
27296
27297 2013-02-13 Kostya Serebryany <kcc@google.com>
27298
27299 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
27300 on x86_64 linux.
27301 * sanitizer.def: Rename __asan_init to __asan_init_v1.
27302
27303 2013-02-12 Dodji Seketeli <dodji@redhat.com>
27304
27305 Avoid instrumenting duplicated memory access in the same basic block
27306 * Makefile.in (asan.o): Add new dependency on hash-table.h
27307 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
27308 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
27309 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
27310 (free_mem_ref_resources, has_mem_ref_been_instrumented)
27311 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
27312 (get_mem_ref_of_assignment): New functions.
27313 (get_mem_refs_of_builtin_call): Extract from
27314 instrument_builtin_call and tweak a little bit to make it fit with
27315 the new signature.
27316 (instrument_builtin_call): Use the new
27317 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
27318 of is_gimple_builtin_call.
27319 (instrument_derefs, instrument_mem_region_access): Insert the
27320 instrumented memory reference into the hash table.
27321 (maybe_instrument_assignment): Renamed instrument_assignment into
27322 this, and change it to advance the iterator when instrumentation
27323 actually happened and return true in that case. This makes it
27324 homogeneous with maybe_instrument_assignment, and thus give a
27325 chance to callers to be more 'regular'.
27326 (transform_statements): Clear the memory reference hash table
27327 whenever we enter a new BB, when we cross a function call, or when
27328 we are done transforming statements. Use
27329 maybe_instrument_assignment instead of instrumentation. No more
27330 need to special case maybe_instrument_assignment and advance the
27331 iterator after calling it; it's now handled just like
27332 maybe_instrument_call. Update comment.
27333
27334 2013-02-13 Richard Biener <rguenther@suse.de>
27335
27336 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
27337 Fix loop discovery code.
27338
27339 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
27340
27341 PR inline-asm/56148
27342 * lra-constraints.c (process_alt_operands): Match early clobber
27343 operand with itself. Check conflicts with earlyclobber only if
27344 the operand is not reloaded. Prefer to reload conflicting operand
27345 if earlyclobber and matching operands are the same.
27346
27347 2013-02-12 Richard Biener <rguenther@suse.de>
27348
27349 PR lto/56297
27350 * lto-streamer-out.c (write_symbol): Do not output symbols
27351 for hard register variables.
27352
27353 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
27354
27355 PR target/54222
27356 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
27357 (umulsidi3_insn, mulsidi3_insn): New insns.
27358
27359 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
27360
27361 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
27362 (struct tune_params): Add vec_costs field.
27363 * config/arm/arm.c (arm_builtin_vectorization_cost)
27364 (arm_add_stmt_cost): New functions.
27365 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
27366 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
27367 (arm_default_vec_cost): New struct of type cpu_vec_costs.
27368 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
27369 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
27370 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
27371 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
27372
27373 2013-02-12 Richard Biener <rguenther@suse.de>
27374
27375 PR lto/56295
27376 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
27377 decls again if possible.
27378
27379 2013-02-12 Richard Biener <rguenther@suse.de>
27380
27381 PR middle-end/56288
27382 * tree-ssa.c (verify_ssa_name): Fix check, move
27383 SSA_NAME_IN_FREE_LIST check up.
27384
27385 2013-02-12 Jakub Jelinek <jakub@redhat.com>
27386 Steven Bosscher <steven@gcc.gnu.org>
27387
27388 PR rtl-optimization/56151
27389 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
27390 equal to op0 or op1, and last_insn pattern is CODE operation
27391 with MEM dest and one of the operands matches that MEM.
27392
27393 2013-02-11 Sriraman Tallam <tmsriram@google.com>
27394
27395 * doc/extend.texi: Document Function Multiversioning and "default"
27396 parameter string to target attribute.
27397 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
27398 target attribute parameter is "default".
27399 (ix86_compare_version_priority): Remove checks for target attribute.
27400 (ix86_mangle_function_version_assembler_name): Change error to sorry.
27401 Remove check for target attribute equal to NULL. Add assert.
27402 (ix86_generate_version_dispatcher_body): Change error to sorry.
27403
27404 2013-02-11 Iain Sandoe <iain@codesourcery.com>
27405 Jack Howarth <howarth@bromo.med.uc.edu>
27406 Patrick Marlier <patrick.marlier@gmail.com>
27407
27408 PR libitm/55693
27409 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
27410 define ENDFILE_SPEC as TM_DESTRUCTOR.
27411 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
27412
27413 2013-02-11 Alexander Potapenko <glider@google.com>
27414 Jack Howarth <howarth@bromo.med.uc.edu>
27415 Jakub Jelinek <jakub@redhat.com>
27416
27417 PR sanitizer/55617
27418 * config/darwin.c (cdtor_record): Rename ctor_record.
27419 (sort_cdtor_records): Rename sort_ctor_records.
27420 (finalize_dtors): New routine to sort destructors by
27421 priority before use in assemble_integer.
27422 (machopic_asm_out_destructor): Use finalize_dtors if needed.
27423
27424 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
27425
27426 PR rtl-optimization/56275
27427 * simplify-rtx.c (avoid_constant_pool_reference): Check that
27428 offset is non-negative and less than cmode size before
27429 calling simplify_subreg.
27430
27431 2013-02-11 Richard Biener <rguenther@suse.de>
27432
27433 PR tree-optimization/56264
27434 * cfgloop.h (fix_loop_structure): Adjust prototype.
27435 * loop-init.c (fix_loop_structure): Return the number of
27436 newly discovered loops.
27437 * tree-cfgcleanup.c (repair_loop_structures): When new loops
27438 are discovered, do a full loop-closed SSA rewrite.
27439
27440 2013-02-11 Richard Biener <rguenther@suse.de>
27441
27442 PR tree-optimization/56273
27443 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27444 first VRP run.
27445 (check_array_ref): Fix missing newline in dumps.
27446 (search_for_addr_array): Likewise.
27447
27448 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
27449
27450 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
27451
27452 2013-02-09 Jakub Jelinek <jakub@redhat.com>
27453
27454 PR target/56256
27455 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
27456
27457 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
27458
27459 PR rtl-optimization/56246
27460 * lra-constraints.c (simplify_operand_subreg): Try to reuse
27461 reload pseudo.
27462 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
27463 constraints are satisfied.
27464
27465 2013-02-08 Jeff Law <law@redhat.com>
27466
27467 PR debug/53948
27468 * emit-rtl.c (reg_is_parm_p): New function.
27469 * regs.h (reg_is_parm_p): New prototype.
27470 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
27471 callee-clobbered registers.
27472
27473 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
27474
27475 PR target/56043
27476 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
27477 If there is no implicit builtin declaration, just return NULL.
27478
27479 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27480
27481 * config/i386/sse.md (FMAMODEM): New mode iterator.
27482 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
27483 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
27484
27485 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27486
27487 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
27488 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
27489 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
27490
27491 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27492
27493 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
27494 (microblaze*-*-elf): Likewise.
27495 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
27496 LINK_SPEC.
27497 * config/microblaze/microblaze-c.c: Add builtin defines for
27498 _LITTLE_ENDIAN and _BIG_ENDIAN.
27499 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
27500 add to TARGET_DEFAULT flags.
27501 Expand ASM_SPEC and LINK_SPEC.
27502 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
27503 * config/microblaze/microblaze.md: Update extendsidi2 and
27504 movdi_internal instructions to use low-order / high-order reg
27505 print_operands.
27506 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
27507 options and inversemask / mask of LITTLE_ENDIAN.
27508 * config/microblaze/t-microblaze: Expand multilib options to
27509 include mlittle-endian (le) and update exceptions patterns.
27510
27511 2013-02-08 Jakub Jelinek <jakub@redhat.com>
27512
27513 PR rtl-optimization/56195
27514 * lra-constraints.c (get_reload_reg): Don't reuse regs
27515 if they have smaller mode than requested, if they have
27516 wider mode than requested, try to return a SUBREG.
27517
27518 PR tree-optimization/56250
27519 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
27520 if type is unsigned and code isn't MULT_EXPR.
27521
27522 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27523
27524 PR tree-optimization/56064
27525 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
27526 bits according to mode.
27527 * fixed-value.h (fixed_from_double_int)
27528 (const_fixed_from_double_int): Adjust comments.
27529
27530 2013-02-08 Richard Biener <rguenther@suse.de>
27531
27532 PR lto/56231
27533 * lto-streamer.h (struct data_in): Remove current_file, current_line
27534 and current_col members.
27535 * lto-streamer-out.c (lto_output_location): Stream changed bits
27536 en-block for efficiency.
27537 * lto-streamer-in.c (clear_line_info): Remove.
27538 (lto_input_location): Cache current file, line and column
27539 globally via local statics. Read changed bits en-block.
27540 (input_function): Do not call clear_line_info.
27541 (lto_read_body): Likewise.
27542 (lto_input_toplevel_asms): Likewise.
27543
27544 2013-02-08 Michael Matz <matz@suse.de>
27545
27546 PR tree-optimization/52448
27547 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
27548 (nt_call_phase): New static.
27549 (add_or_mark_expr): Only mark accesses with newer phase than any
27550 call seen.
27551 (nonfreeing_call_p): New.
27552 (nt_init_block): Update nt_call_phase, mark blocks as visited.
27553 (nt_fini_block): Keep blocks marked as visited.
27554 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
27555
27556 2013-02-08 Richard Biener <rguenther@suse.de>
27557
27558 * ira.c (ira): Free broken dominator information.
27559
27560 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27561
27562 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
27563
27564 2013-02-08 Marek Polacek <polacek@redhat.com>
27565
27566 * cfgloop.c (verify_loop_structure): Add more checking of headers.
27567
27568 2013-02-08 Richard Biener <rguenther@suse.de>
27569
27570 PR middle-end/56181
27571 * cfgloop.h (flow_loops_find): Adjust.
27572 (bb_loop_header_p): Declare.
27573 * cfgloop.c (bb_loop_header_p): New function split out from ...
27574 (flow_loops_find): ... here. Adjust function signature,
27575 support incremental loop structure update.
27576 (verify_loop_structure): Cleanup. Verify a loop is a loop.
27577 * cfgloopmanip.c (fix_loop_structure): Move ...
27578 * loop-init.c (fix_loop_structure): ... here.
27579 (apply_loop_flags): Split out from ...
27580 (loop_optimizer_init): ... here.
27581 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
27582 in incremental mode, only remove dead loops here.
27583
27584 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27585
27586 PR target/54222
27587 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
27588 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
27589 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
27590 (*round<mode>3.libgcc): New insns for fixed-modes.
27591 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
27592 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
27593 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
27594 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
27595 implementations. Define to __builtin_avr_absFX,
27596 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
27597 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
27598 __builtin_avr_countlsFX, respectively.
27599 * config/avr/avr-c.c (target.h): Include it.
27600 (enum avr_builtin_id): New enum.
27601 (avr_resolve_overloaded_builtin): New static function.
27602 (avr_register_target_pragmas): Use it to set
27603 targetm.resolve_overloaded_builtin.
27604 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
27605 tree nodes used by DEF_BUILTIN.
27606 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
27607 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
27608 <AVR_BUILTIN_xxBITS>: Same.
27609
27610 2013-02-08 Richard Biener <rguenther@suse.de>
27611
27612 * cfgloop.c (verify_loop_structure): Properly handle
27613 a loop exiting to another loop header.
27614 * ira-int.h (ira_loops): Remove.
27615 * ira.c (ira_loops): Remove.
27616 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
27617 (do_reload): Use loop_optimizer_finalize.
27618 * ira-build.c (create_loop_tree_nodes): Use get_loops and
27619 number_of_loops to access the loop tree.
27620 (more_one_region_p): Likewise.
27621 (finish_loop_tree_nodes): Likewise.
27622 (rebuild_regno_allocno_maps): Likewise.
27623 (mark_loops_for_removal): Likewise.
27624 (mark_all_loops_for_removal): Likewise.
27625 (remove_unnecessary_regions): Likewise.
27626 (ira_build): Likewise.
27627 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
27628
27629 2013-02-08 Richard Biener <rguenther@suse.de>
27630
27631 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
27632 * ipa-pure-const.c (analyze_function): Avoid calling
27633 mark_irreducible_loops twice.
27634 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
27635
27636 2013-02-07 David S. Miller <davem@davemloft.net>
27637
27638 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
27639 on 'reg'.
27640 * var-tracking.c (vt_add_function_parameter): Test the presence of
27641 HAVE_window_save properly and do not remap argument registers when
27642 we have a leaf function.
27643
27644 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
27645
27646 PR bootstrap/56227
27647 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
27648 instead of "ll".
27649 * config/i386/i386.c (ix86_print_operand): Ditto.
27650
27651 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27652
27653 * lra-constraints.c (process_alt_operands): Fix recently added comment.
27654
27655 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27656
27657 PR rtl-optimization/56225
27658 * lra-constraints.c (process_alt_operands): Check that reload hard
27659 reg can hold value for strict_low_part.
27660
27661 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27662
27663 PR debug/56154
27664 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
27665 dwarf2out_end_function.
27666 (in_first_function_p, maybe_at_text_label_p,
27667 first_loclabel_num_not_at_text_label): New variables.
27668 (dwarf2out_var_location): In the first function find out
27669 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
27670 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
27671 functions.
27672
27673 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
27674
27675 PR rtl-optimization/56178
27676 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
27677 SUBREG of a register. Tidy up related block of code.
27678 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
27679 note if the source is a register or a SUBREG of a register.
27680
27681 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27682
27683 PR target/56228
27684 * config/rs6000/rs6000.md (ptrm): New mode attr.
27685 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
27686 call_value_indirect_aix<pttrsize>,
27687 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
27688 m in constraints.
27689
27690 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
27691
27692 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
27693 if -bnortl. Convert to strcmp and strncmp.
27694
27695 2013-02-07 Alan Modra <amodra@gmail.com>
27696
27697 PR target/54009
27698 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
27699 addresses won't wrap when offsetting.
27700 (rs6000_secondary_reload): Provide secondary reloads needed for
27701 wrapping LO_SUM addresses.
27702
27703 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
27704
27705 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
27706 MACH, just __MACH__.
27707
27708 2013-02-06 Richard Biener <rguenther@suse.de>
27709
27710 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
27711 instead of calling fix_loop_structure.
27712
27713 2013-02-06 Jakub Jelinek <jakub@redhat.com>
27714
27715 PR middle-end/56217
27716 * omp-low.c (use_pointer_for_field): Return false if
27717 lower_send_shared_vars doesn't generate any copy-out code.
27718
27719 2013-02-06 Tom de Vries <tom@codesourcery.com>
27720
27721 PR rtl-optimization/56131
27722 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
27723 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
27724 of the label is NULL. Add comment.
27725
27726 2013-02-05 Jakub Jelinek <jakub@redhat.com>
27727
27728 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
27729
27730 PR sanitizer/55374
27731 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
27732 (STATIC_LIBTSAN_LIBS): Likewise.
27733 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
27734 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
27735 is defined, don't add anything else beyond that.
27736 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
27737 (LINK_COMMAND_SPEC): Use them.
27738
27739 PR tree-optimization/56205
27740 * tree-stdarg.c (check_all_va_list_escapes): Return true if
27741 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
27742 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
27743
27744 2013-02-05 Richard Biener <rguenther@suse.de>
27745
27746 PR tree-optimization/53342
27747 PR tree-optimization/53185
27748 * tree-vectorizer.h (vect_check_strided_load): Remove.
27749 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
27750 not disallow peeling for vectorized strided loads.
27751 (vect_check_strided_load): Make static and simplify.
27752 (vect_analyze_data_refs): Adjust.
27753 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
27754 correctly when vectorizing strided loads.
27755
27756 2013-02-05 Richard Biener <rguenther@suse.de>
27757
27758 * doc/install.texi: Refer to ISL, not PPL.
27759
27760 2013-02-05 Jan Hubicka <jh@suse.cz>
27761
27762 PR tree-optimization/55789
27763 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
27764
27765 2013-02-05 Jan Hubicka <jh@suse.cz>
27766
27767 PR tree-optimization/55789
27768 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
27769 the dead call anyway.
27770
27771 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
27772
27773 PR sanitizer/55374
27774 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
27775
27776 2013-02-04 Alexander Potapenko <glider@google.com>
27777 Jack Howarth <howarth@bromo.med.uc.edu>
27778 Jakub Jelinek <jakub@redhat.com>
27779
27780 PR sanitizer/55617
27781 * config/darwin.c (sort_ctor_records): Stabilized qsort
27782 on constructor priority by using original position.
27783 (finalize_ctors): New routine to sort constructors by
27784 priority before use in assemble_integer.
27785 (machopic_asm_out_constructor): Use finalize_ctors if needed.
27786
27787 2013-02-04 Jakub Jelinek <jakub@redhat.com>
27788
27789 PR libstdc++/54314
27790 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
27791 about visibility on artificial decls.
27792 * config/sol2.c (solaris_assemble_visibility): Likewise.
27793
27794 2013-02-04 Kai Tietz <ktietz@redhat.com>
27795
27796 PR target/56186
27797 * config/i386/i386.c (function_value_ms_64): Add additional valtype
27798 argument and improve checking of return-argument types for 16-byte
27799 modes.
27800 (ix86_function_value_1): Add additional valtype argument on call
27801 of function_value_64.
27802 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
27803 handling infunction_value_64 function.
27804
27805 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
27806
27807 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
27808
27809 2013-02-04 Richard Biener <rguenther@suse.de>
27810
27811 PR tree-optimization/56188
27812 * tree-ssa-structalias.c (label_visit): Consider case with
27813 initially non-empty points-to set.
27814 (perform_var_substitution): Dump node mapping and clean up.
27815
27816 2013-02-04 Richard Guenther <rguenther@suse.de>
27817
27818 PR lto/56168
27819 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
27820 node prevail as last resort.
27821 (lto_symtab_merge_decls): Remove guard on LTRANS here.
27822 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
27823
27824 2013-02-04 Richard Biener <rguenther@suse.de>
27825
27826 PR tree-optimization/56113
27827 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
27828 Merge into ...
27829 (equiv_class_lookup_or_add): ... this.
27830 (label_visit): Adjust and fix error in previous patch.
27831 (perform_var_substitution): Adjust.
27832
27833 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
27834
27835 * config/sh/divtab.c: Fix formatting and comments throughout the file.
27836 * config/sh/sh4-300.md: Likewise.
27837 * config/sh/sh4a.md: Likewise.
27838 * config/sh/constraints.md: Likewise.
27839 * config/sh/sh.md: Likewise.
27840 * config/sh/netbsd-elf.h: Likewise.
27841 * config/sh/predicates.md: Likewise.
27842 * config/sh/sh-protos.h: Likewise.
27843 * config/sh/ushmedia.h: Likewise.
27844 * config/sh/linux.h: Likewise.
27845 * config/sh/sh.c: Likewise.
27846 * config/sh/superh.h: Likewise.
27847 * config/sh/elf.h: Likewise.
27848 * config/sh/sh4.md: Likewise.
27849 * config/sh/sh.h: Likewise.
27850
27851 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27852
27853 * config/pa/constraints.md: Adjust unused letters. Change "T"
27854 constraint to match_test floating_point_store_memory_operand().
27855 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
27856 (base14_operand): New.
27857 (floating_point_store_memory_operand): New.
27858 (integer_store_memory_operand): Revise to use base14_operand and
27859 reg_plus_base_memory_operand.
27860 (move_dest_operand): Allow symbolic_memory_operands.
27861 (symbolic_memory_operand): Check for LO_SOM.
27862 (symbolic_operand): Change default case to break.
27863 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
27864 CONST_DOUBLE values to be reloaded by putting them into memory when
27865 the destination is a floating point register.
27866 (movdf): Remove code to handle CONST_DOUBLE.
27867 (movsf): Likewise.
27868 (reload_indf_r1): New.
27869 (reload_insf_r1): New.
27870 Consistently use "Q" and "T" constraints with integer and floating
27871 point move instructions, respectively.
27872 (movdi): Remove FAIL.
27873 Change predicate for source operand unamed DImode move from
27874 general_operand to move_src_operand.
27875 (umulsidi3): Change predicate for destination operand to
27876 register_operand.
27877 Likewise for similar unamed patterns.
27878 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
27879 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
27880 (hppa_legitimize_address): Simplify mask calculation.
27881 (pa_emit_move_sequence): Revised handling of secondary reloads from
27882 REG+D addresses for floating point loads and stores. Directly handle
27883 loading CONST0_RTX (mode) to a floating point register.
27884 (pa_secondary_reload): Handle reloading DF and SFmode constant values
27885 to floating point registers. Don't restrict secondary reloads to
27886 floating point registers to integer modes. Revise some comments and
27887 cleanup some code.
27888 (TARGET_LEGITIMATE_ADDRESS_P): Define.
27889 (pa_legitimate_address_p): New.
27890 (pa_legitimize_reload_address): New.
27891 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
27892 (STRICT_REG_OK_FOR_BASE_P): New.
27893 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
27894 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
27895
27896 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
27897 Andrew Dixie <andrewd@gentrack.com>
27898
27899 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
27900 flag set.
27901
27902 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
27903
27904 * expmed.c (extract_bit_field_1): Pass the full width of the
27905 structure to get_best_reg_extraction_insn.
27906
27907 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
27908
27909 PR target/54601
27910 * configure.ac (use_cxa_atexit): Add AIX.
27911 * configure: Regenerate.
27912
27913 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
27914
27915 2013-02-01 Jakub Jelinek <jakub@redhat.com>
27916
27917 PR debug/54793
27918 * final.c (need_profile_function): New variable.
27919 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
27920 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
27921 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
27922 notes, targetm.asm_out.function_prologue doesn't emit anything,
27923 HAVE_prologue and profiler should be emitted before prologue,
27924 set need_profile_function instead of emitting it.
27925 (final_scan_insn): If need_profile_function, emit
27926 profile_function on the first NOTE_INSN_BASIC_BLOCK or
27927 NOTE_INSN_FUNCTION_BEG note.
27928
27929 2013-02-01 Richard Henderson <rth@redhat.com>
27930
27931 * config/rs6000/rs6000.md (smulditi3): New.
27932 (umulditi3): New.
27933
27934 * config/alpha/alpha.md (umulditi3): New.
27935
27936 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
27937
27938 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
27939 (ASM_OUTPUT_ALIGNED_LOCAL): New.
27940
27941 2013-02-01 Richard Biener <rguenther@suse.de>
27942
27943 PR tree-optimization/56113
27944 * tree-ssa-structalias.c (label_visit): Reduce work for
27945 single-predecessor nodes.
27946
27947 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
27948
27949 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
27950 range isn't testing for zero.
27951
27952 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
27953
27954 PR middle-end/56113
27955 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
27956
27957 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
27958 Nick Clifton <nickc@redhat.com>
27959
27960 * config/v850/constraints.md (Q): Define as a memory constraint.
27961 * config/v850/predicates.md (label_ref_operand): New predicate.
27962 (e3v5_shift_operand): New predicate.
27963 (ior_operator): New predicate.
27964 * config/v850/t-v850: Add e3v5 multilib.
27965 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
27966 (v850_gen_movdi): Prototype.
27967 * config/v850/v850.c: Add support for e3v5 architecture.
27968 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
27969 TARGET_V850E_UP.
27970 (construct_save_jarl): Add e3v5 long JARL support.
27971 (v850_adjust_insn_length): New function. Adjust length of call
27972 insns when using e3v5 instructions.
27973 (v850_gen_movdi): New function: Generate instructions to move a
27974 DImode value.
27975 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
27976 (CPP_SPEC): Define __v850e3v5__ as appropriate.
27977 (TARGET_USE_FPU): Enable for e3v5.
27978 (CONST_OK_FOR_W): New macro.
27979 (ADJUST_INSN_LENGTH): Define.
27980 * config/v850/v850.md (UNSPEC_LOOP): Define.
27981 (attr cpu): Add v850e3v5.
27982 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
27983 (movdi): New pattern.
27984 (movdi_internal): New pattern.
27985 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
27986 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
27987 (cstoresf4): Likewise.
27988 (cstoredf4): Likewise.
27989 (insv): New pattern.
27990 (rotlso3_a): New pattern.
27991 (rotlsi3_b): New pattern
27992 (rotlsi3_v850e3v5): New pattern.
27993 (doloop_begin): New pattern.
27994 (fix_loop_counter): New pattern.
27995 (doloop_end): New pattern.
27996 (branch_normal): Add e3v5 long branch support.
27997 (branch_invert): Likewise.
27998 (branch_z_normal): Likewise.
27999 (branch_z_invert): Likewise.
28000 (branch_nz_normal): Likewise.
28001 (branch_nz_invert): Likewise.
28002 (call_internal_short): Add e3v5 register-indirect JARL support.
28003 (call_internal_long): Likewise.
28004 (call_value_internal_short): Likewise.
28005 (call_value_internal_long): Likewise.
28006 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
28007 (mloop): New option.
28008 * config.gcc: Add support for configuring v840e3v5 target.
28009 * doc/invoke.texi: Document new v850 specific command line options.
28010
28011 2013-01-31 Paul Koning <ni1d@arrl.net>
28012
28013 PR debug/55059
28014 PR debug/54508
28015 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
28016 children if parent is a class.
28017 (prune_unused_types_prune): Don't add DW_AT_declaration.
28018
28019 2013-01-31 Richard Biener <rguenther@suse.de>
28020
28021 PR tree-optimization/56157
28022 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
28023 match up operand with SLP child.
28024
28025 2013-01-31 Jason Merrill <jason@redhat.com>
28026
28027 PR debug/54410
28028 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
28029 parameters the first time.
28030 (gen_scheduled_generic_parms_dies): Check completeness here.
28031
28032 2013-01-31 Richard Biener <rguenther@suse.de>
28033
28034 PR middle-end/53073
28035 * common.opt (faggressive-loop-optimizations): New flag,
28036 enabled by default.
28037 * doc/invoke.texi (faggressive-loop-optimizations): Document.
28038 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
28039 infer_loop_bounds_from_undefined by it.
28040
28041 2013-01-31 Richard Biener <rguenther@suse.de>
28042
28043 PR tree-optimization/56150
28044 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
28045 visit virtual operands.
28046 (find_uses_to_rename_bb): Likewise.
28047
28048 2013-01-31 Richard Biener <rguenther@suse.de>
28049
28050 PR tree-optimization/56150
28051 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
28052 mixed store non-store stmts.
28053
28054 2013-01-30 Jakub Jelinek <jakub@redhat.com>
28055
28056 PR sanitizer/55374
28057 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
28058 LIBASAN_EARLY_SPEC is defined.
28059 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
28060 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
28061 before %o.
28062 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
28063
28064 PR c++/55742
28065 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
28066 invalid args instead of ICEing on it.
28067 (ix86_valid_target_attribute_tree): Return error_mark_node if
28068 ix86_valid_target_attribute_inner_p failed.
28069 (ix86_valid_target_attribute_p): Return false only if
28070 ix86_valid_target_attribute_tree returned error_mark_node. Allow
28071 target("default") attribute.
28072 (sorted_attr_string): Change argument from const char * to tree,
28073 merge in all target attribute arguments rather than just one.
28074 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
28075 instead of free. Avoid using strcat.
28076 (ix86_mangle_function_version_assembler_name): Mangle
28077 target("default") as if no target attribute is present. Adjust
28078 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
28079 instead of xmalloc and XDELETEVEC instead of free.
28080 (ix86_function_versions): Don't return true if one of the decls
28081 doesn't have target attribute. If they don't and one of the decls
28082 is DECL_FUNCTION_VERSIONED, report an error. Adjust
28083 sorted_attr_string caller. Use XDELETEVEC instead of free.
28084 (ix86_supports_function_versions): Remove.
28085 (make_name): Fix up formatting.
28086 (make_dispatcher_decl): Remove resolver_name and its initialization.
28087 Avoid leaking memory.
28088 (is_function_default_version): Return true if there is
28089 target("default") attribute rather than no target attribute at all.
28090 (make_resolver_func): Avoid leaking memory.
28091 (ix86_generate_version_dispatcher_body): Likewise.
28092 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
28093 * target.def (supports_function_versions): Remove.
28094 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
28095 * doc/tm.texi: Regenerated.
28096
28097 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
28098
28099 PR rtl-optimization/56144
28100 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
28101 for values with side effects.
28102
28103 2013-01-30 Richard Biener <rguenther@suse.de>
28104
28105 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
28106 (sparseset_pop): Likewise.
28107 * cfganal.c (compute_idf): Likewise. Increase work-stack size
28108 to be able to use quick_push in the worker loop.
28109
28110 2013-01-30 Marek Polacek <polacek@redhat.com>
28111
28112 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
28113
28114 2013-01-30 Richard Biener <rguenther@suse.de>
28115
28116 PR lto/56147
28117 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
28118
28119 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
28120
28121 PR tree-optimization/56064
28122 * fixed-value.c (fixed_from_double_int): New function.
28123 * fixed-value.h (fixed_from_double_int): New prototype.
28124 (const_fixed_from_double_int): New static inline function.
28125 * fold-const.c (native_interpret_fixed): New static function.
28126 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
28127 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
28128 (native_encode_fixed): New static function.
28129 (native_encode_expr) <FIXED_CST>: Use it.
28130 (native_interpret_int): Move double_int worker code to...
28131 * double-int.c (double_int::from_buffer): ...this new static method.
28132 * double-int.h (double_int::from_buffer): Prototype it.
28133
28134 2013-01-30 Richard Biener <rguenther@suse.de>
28135
28136 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
28137 New pointer-map and obstack.
28138 (init_alias_vars): Allocate pointer-map and obstack.
28139 (delete_points_to_sets): Free them.
28140 (find_what_var_points_to): Cache result.
28141 (find_what_p_points_to): Adjust for changed interface of
28142 find_what_var_points_to.
28143 (compute_points_to_sets): Likewise.
28144 (ipa_pta_execute): Likewise.
28145
28146 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
28147
28148 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
28149 * configure: Regenerate.
28150 * config.in: Regenerate.
28151 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
28152 #nobits/#progbits if supported.
28153
28154 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
28155
28156 PR target/56121
28157 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
28158 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
28159 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
28160
28161 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28162
28163 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
28164 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
28165
28166 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28167
28168 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
28169 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
28170
28171 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28172
28173 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
28174 declaration.
28175 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
28176 * config/arm/cortex-a7.md: New bypasses using
28177 arm_mac_accumulator_is_result.
28178
28179 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28180
28181 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
28182 (cortex_a7_neon_mla): Likewise.
28183 (cortex_a7_fpfmad): New reservation.
28184 (cortex_a7_fpmacs): Use ffmas and update required units.
28185 (cortex_a7_fpmuld): Update required units and latency.
28186 (cortex_a7_fpmacd): Likewise.
28187 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
28188 (cortex_a7_neon). Likewise.
28189 (bypass) Update participating units.
28190
28191 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28192
28193 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
28194 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
28195 from fmac to ffma.
28196 * config/arm/vfp11.md (vfp_farith): Use ffmas.
28197 (vfp_fmul): Use ffmad.
28198 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
28199 (cortex_r4_fmacd): Use ffmad.
28200 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
28201 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
28202 (cortex_a9_fmacd): Use ffmad.
28203 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
28204 (cortex_a8_vfp_macd): Use ffmad.
28205 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
28206 (cortex_a5_fpmacd): Use ffmad.
28207 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
28208 (cortex_a15_vfp_macd): Use ffmad.
28209 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
28210
28211 2013-01-29 Jason Merrill <jason@redhat.com>
28212
28213 PR libstdc++/54314
28214 * varasm.c (default_assemble_visibility): Don't warn about
28215 visibility on artificial decls.
28216
28217 2013-01-29 Richard Biener <rguenther@suse.de>
28218
28219 PR tree-optimization/56113
28220 * tree-ssa-structalias.c (equiv_class_lookup): Also return
28221 the bitmap leader.
28222 (label_visit): Free duplicate bitmaps and record the leader instead.
28223 (perform_var_substitution): Adjust.
28224
28225 2013-01-29 Richard Biener <rguenther@suse.de>
28226
28227 PR tree-optimization/55270
28228 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
28229 the CFG, schedule loops for fixup.
28230
28231 2013-01-29 Nick Clifton <nickc@redhat.com>
28232
28233 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
28234 SP_REG.
28235
28236 2013-01-28 Leif Ekblad <leif@rdos.net>
28237
28238 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
28239 * config/i386/i386.h (TARGET_RDOS): New macro.
28240 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
28241 * config/i386/i386.c (ix86_option_override_internal): For 64bit
28242 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
28243 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
28244 DEFAULT_LARGE_SECTION_THRESHOLD.
28245 * config/i386/i386.md (R14_REG, R15_REG): New constants.
28246 * config/i386/rdos.h: New file.
28247 * config/i386/rdos64.h: New file.
28248
28249 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
28250
28251 PR other/54814
28252 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
28253 TEST_HARD_REG_BIT.
28254
28255 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28256
28257 PR rtl-optimization/56117
28258 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
28259 call cselib_lookup_from_insn on the MEM before calling
28260 add_insn_mem_dependence.
28261
28262 2013-01-28 Richard Biener <rguenther@suse.de>
28263
28264 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
28265 to a stmt that didn't have one.
28266 (copy_phis_for_bb): Likewise for PHI arguments.
28267 (copy_debug_stmt): Likewise for debug stmts.
28268
28269 2013-01-28 Richard Biener <rguenther@suse.de>
28270
28271 PR tree-optimization/56034
28272 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
28273 (partition_builtin_p): Adjust.
28274 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
28275 it is the last partition.
28276 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
28277 up the vertex for the definition.
28278 (classify_partition): Classify whether a partition is a
28279 PKIND_REDUCTION, thus has uses outside of the loop.
28280 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
28281 Merge all PKIND_REDUCTION partitions into the last partition.
28282 (tree_loop_distribution): Seed partitions from reductions as well.
28283
28284 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28285
28286 PR tree-optimization/56125
28287 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
28288 pow(x,c) into sqrt(x) * powi(x, n/2) or
28289 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
28290 optimizing for size.
28291 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
28292 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
28293 integer.
28294
28295 PR tree-optimization/56094
28296 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
28297 to UNKNOWN_LOCATION while gimplifying expr.
28298
28299 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
28300
28301 PR target/56114
28302 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
28303 operand 0 in movabs insn template for -masm=intel asm alternative.
28304 (*movabs<mode>_2): Ditto for operand 1.
28305
28306 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
28307
28308 PR target/54663
28309 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
28310 of microblaze-c.o
28311
28312 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28313
28314 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
28315 tm_file.
28316
28317 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
28318
28319 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
28320 Undef to avoid warning.
28321
28322 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28323
28324 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
28325 * configure: Regenerate.
28326
28327 2013-01-25 Jakub Jelinek <jakub@redhat.com>
28328
28329 PR tree-optimization/56098
28330 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
28331 for stmts with volatile ops.
28332 (cond_store_replacement): Don't optimize if assign has volatile ops.
28333 (cond_if_else_store_replacement_1): Don't optimize if either
28334 then_assign or else_assign have volatile ops.
28335 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
28336 volatile ops.
28337
28338 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28339
28340 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
28341
28342 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28343
28344 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
28345 missing ':' in asm example.
28346
28347 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
28348
28349 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
28350 entries into lane and laneq entries.
28351 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
28352 Remove AdvSIMD scalar modes.
28353 (aarch64_sq<r>dmulh_laneq<mode>): New.
28354 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
28355 modes.
28356 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
28357 builtin implementations to relfect changes in RTL in aarch64-simd.md.
28358 * config/aarch64/iterators.md (VCOND): New.
28359 (VCONQ): New.
28360
28361 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28362
28363 PR target/54222
28364 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
28365 Add NULL LIBNAME argument to existing definitions.
28366 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
28367 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
28368 * config/avr/avr.c (DEF_BUILTIN): Same.
28369 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
28370 (avr_expand_builtin): Expand to a vanilla call if a libgcc
28371 implementation is available (DECL_ASSEMBLER_NAME is set).
28372 (avr_fold_absfx): New static function.
28373 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
28374 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
28375 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
28376 AVR_BUILTIN_ABSLLK.
28377 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
28378 (abshk, absk, abslk, absllk): Provide as static inline functions.
28379
28380 2013-01-25 Marek Polacek <polacek@redhat.com>
28381
28382 PR tree-optimization/56035
28383 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
28384
28385 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
28386
28387 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
28388 (*movtf_internal_rex64): Add (!o,C) alternative
28389 (*movxf_internal_rex64): Ditto.
28390 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
28391
28392 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
28393
28394 * doc/invoke.texi: fix typo.
28395 * doc/objc.texi: fix typo.
28396
28397 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
28398
28399 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
28400 for the first two alternatives.
28401
28402 2013-01-24 Diego Novillo <dnovillo@google.com>
28403
28404 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
28405 (ggc-zone.o): Remove.
28406 * configure.ac: Remove option --with-gc.
28407 * configure: Re-generate.
28408 * doc/install.texi: Remove documentation for --with-gc.
28409 * gengtype.c (write_enum_defn): Remove. Update all users.
28410 (write_Types_process_field): Remove generation of gt_e_* argument.
28411 (output_type_enum): Remove. Update all users.
28412 (write_enum_defn): Remove. Update all users.
28413 (enum alloc_zone): Remove. Update all users.
28414 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
28415 * ggc-common.c (ggc_splay_alloc): Remove first argument.
28416 Update all callers.
28417 (struct ptr_data): Remove field TYPE. Update all users.
28418 (gt_pch_note_object): Remove argument TYPE. Update all users.
28419 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
28420 Update all users.
28421 * ggc-none.c (ggc_alloc_typed_stat): Remove.
28422 (struct alloc_zone): Remove.
28423 (ggc_internal_alloc_zone_stat): Remove.
28424 (ggc_internal_cleared_alloc_zone_stat): Remove.
28425 * ggc-page.c (ggc_alloc_typed_stat): Remove.
28426 (ggc_pch_count_object): Remove last argument. Update all users.
28427 (ggc_pch_alloc_object): Remove last argument. Update all users.
28428 (struct alloc_zone): Remove.
28429 * ggc-zone.c: Remove.
28430 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
28431 (struct alloc_zone): Remove.
28432 (ggc_alloc_typed_stat): Remove.
28433 (ggc_alloc_typed): Remove.
28434 (ggc_splay_alloc): Remove first argument.
28435 (rtl_zone): Remove. Update all users.
28436 (tree_zone): Remove. Update all users.
28437 (tree_id_zone): Remove. Update all users.
28438 (ggc_internal_zone_alloc_stat): Remove. Update all users.
28439 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
28440 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
28441 * tree-ssanames.c: Remove references to zone allocator in comments.
28442
28443 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
28444
28445 * config/avr/avr.c (avr_out_fract): Make register numbers that
28446 might be outside of source operand signed.
28447
28448 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
28449
28450 * config/i386/constraints.md (Yf): New constraint.
28451 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
28452 of f constraint to conditionaly disable x87 register preferences.
28453 (*movdf_internal): Ditto.
28454 (*movsf_internal): Ditto.
28455
28456 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
28457
28458 PR inline-asm/55934
28459 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
28460 that have operands with impossible constraints.
28461 Add a FIXME for a speed-up opportunity.
28462 * lra-constraints.c (process_alt_operands): Verify that a class
28463 selected from constraints on asms is valid for the operand mode.
28464 (curr_insn_transform): Remove incorrect comment.
28465
28466 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
28467
28468 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
28469 TOC operand is a valid symbol ref in the constant pool.
28470
28471 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28472
28473 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
28474
28475 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
28476
28477 PR target/54222
28478 * config/avr/stdfix.h: New file.
28479 * t-avr (stdfix-gcc.h): New rule to build it.
28480 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
28481
28482 2013-01-23 Kostya Serebryany <kcc@google.com>
28483
28484 * config/darwin.h: remove dependency on
28485 CoreFoundation (asan on Mac OS).
28486
28487 2013-01-23 Jakub Jelinek <jakub@redhat.com>
28488
28489 PR target/49069
28490 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
28491 instead of cmpdi_operand for first comparison operand.
28492 Don't assert that comparison operands aren't both constants.
28493
28494 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28495
28496 * doc/install.texi (Downloading the Source): Update references to
28497 downloading separate components.
28498
28499 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28500
28501 * doc/extend.texi (__int128): Improve grammar.
28502
28503 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
28504
28505 PR target/56028
28506 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
28507 alternative to (o,r).
28508 (*movdi_internal_rex64): Remove (!o,n) alternative.
28509 (DImode immediate->memory splitter): Remove.
28510 (DImode immediate->memory peephole2): Remove.
28511 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
28512 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
28513 alternative to (!o,*r).
28514 (*movtf_internal_sse): New pattern.
28515 (*movxf_internal_rex64): New pattern.
28516 (*movxf_internal): Disable for TARGET_64BIT.
28517 (*movdf_internal_rex64): Remove (!o,F) alternative.
28518
28519 2013-01-22 Jakub Jelinek <jakub@redhat.com>
28520
28521 PR middle-end/56074
28522 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
28523 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
28524 * tree-vect-loop-manip.c (find_loop_location): Also ignore
28525 stmt locations where LOCATION_LOCUS of the stmt location is
28526 UNKNOWN_LOCATION or BUILTINS_LOCATION.
28527
28528 PR target/55686
28529 * config/i386/i386.md (UNSPEC_STOS): New.
28530 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
28531 *strsetqi_1): Add UNSPEC_STOS.
28532
28533 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
28534
28535 PR c++/56067
28536 * doc/invoke.texi: Remove left over -Wsynth example.
28537
28538 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28539
28540 PR tree-optimization/56051
28541 * fold-const.c (fold_binary_loc): Don't fold
28542 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
28543 a narrowing conversion, or widening conversion from signed
28544 to unsigned.
28545
28546 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28547
28548 PR rtl-optimization/56023
28549 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
28550 dependent on debug instruction.
28551
28552 2013-01-21 Martin Jambor <mjambor@suse.cz>
28553
28554 PR middle-end/56022
28555 * function.c (allocate_struct_function): Call
28556 invoke_set_current_function_hook earlier.
28557
28558 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28559
28560 * reload1.c (init_reload): Only initialize reload_obstack
28561 during the first call.
28562
28563 2013-01-21 Marek Polacek <polacek@redhat.com>
28564
28565 * cfgloop.c (verify_loop_structure): Fix up grammar.
28566
28567 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
28568
28569 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
28570 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
28571
28572 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28573
28574 PR target/56058
28575 * config/arm/marvell-pj4.md: Update copyright year.
28576 Fix up use of alu to alu_reg and simple_alu_imm.
28577
28578 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28579
28580 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
28581
28582 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
28583
28584 PR target/55433
28585 * lra-constraints.c (curr_insn_transform): Don't reuse original
28586 insn for secondary memory move when memory mode should be different.
28587
28588 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28589
28590 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
28591 atomic_storedi_1): New patterns.
28592
28593 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
28594
28595 btver2 pipeline descriptions.
28596 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
28597 descriptions.
28598 * config/i386/i386.md (btver2_decode): New type attributes.
28599 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
28600 type attributes.
28601 * config/i386/btver2.md: New file describing btver2 pipelines.
28602
28603 2013-01-19 Andrew Pinski <apinski@cavium.com>
28604
28605 PR tree-optimization/52631
28606 * tree-ssa-sccvn (visit_use): Before looking up the original
28607 statement, try looking up the simplified expression.
28608
28609 2013-01-19 Anthony Green <green@moxielogic.com>
28610
28611 * config/moxie/moxie.c (moxie_expand_prologue): Set
28612 current_function_static_stack_size.
28613
28614 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28615
28616 PR tree-optimization/56029
28617 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
28618 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
28619
28620 2013-01-18 Sharad Singhai <singhai@google.com>
28621
28622 PR tree-optimization/55995
28623 * dumpfile.c (dump_loc): Print location only if available.
28624 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
28625
28626 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
28627
28628 PR target/55433
28629 * lra-constraints.c (curr_insn_transform): Reuse original insn for
28630 secondary memory move.
28631 (inherit_reload_reg): Use rclass instead of cl for
28632 check_secondary_memory_needed_p.
28633
28634 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28635
28636 PR middle-end/56015
28637 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
28638 the case where writing real complex part of target modifies op1.
28639
28640 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
28641
28642 * config/aarch64/aarch64-simd.md
28643 (aarch64_vcond_internal<mode>): Handle unordered cases.
28644 * config/aarch64/iterators.md (v_cmp_result): New.
28645
28646 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
28647 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28648
28649 * config/arm/marvell-pj4.md: New file.
28650 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
28651 * config/arm/arm.md (generic_sched): Add marvell_pj4.
28652 (generic_vfp): Likewise.
28653 * config/arm/arm-cores.def: Add marvell-pj4.
28654 * config/arm/arm-tune.md: Regenerate.
28655 * config/arm/arm-tables.opt: Regenerate.
28656 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
28657 * doc/invoke.texi: Document marvell-pj4.
28658
28659 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
28660
28661 * config/aarch64/arm_neon.h: Map scalar types to standard types.
28662
28663 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28664
28665 PR debug/54114
28666 PR debug/54402
28667 PR debug/49888
28668 * var-tracking.c (negative_power_of_two_p): New.
28669 (global_get_addr_cache, local_get_addr_cache): New.
28670 (get_addr_from_global_cache, get_addr_from_local_cache): New.
28671 (vt_canonicalize_addr): Rewrite using the above. Adjust the
28672 heading comment.
28673 (vt_stack_offset_p): Remove.
28674 (vt_canon_true_dep): Always canonicalize loc's address.
28675 (clobber_overlapping_mems): Make sure we have a MEM.
28676 (local_get_addr_clear_given_value): New.
28677 (val_reset): Clear local cached entries.
28678 (compute_bb_dataflow): Create and release the local cache.
28679 Disable duplicate MEMs clobbering.
28680 (emit_notes_in_bb): Clobber MEMs likewise.
28681 (vt_emit_notes): Create and release the local cache.
28682 (vt_initialize, vt_finalize): Create and release the global
28683 cache, respectively.
28684 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
28685
28686 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28687
28688 PR libmudflap/53359
28689 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
28690 not found in the symtab.
28691
28692 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28693
28694 PR debug/56006
28695 PR rtl-optimization/55547
28696 PR rtl-optimization/53827
28697 PR debug/53671
28698 PR debug/49888
28699 * alias.c (offset_overlap_p): New, factored out of...
28700 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
28701 the conservative special case for symbolic constants. Don't
28702 adjust zero sizes on alignment.
28703
28704 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
28705
28706 PR rtl-optimization/52573
28707 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
28708 REG_UNUSED for the same register.
28709
28710 2013-01-17 Richard Biener <rguenther@suse.de>
28711 Marek Polacek <polacek@redhat.com>
28712
28713 PR rtl-optimization/55833
28714 * loop-unswitch.c (unswitch_loops): Move loop verification...
28715 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
28716 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
28717 Set it to true when we're removing a loop from hierarchy tree in
28718 an irreducible region.
28719 (fix_bb_placements): Adjust caller.
28720 (fix_loop_placements): Likewise.
28721
28722 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
28723
28724 * config/avr/builtins.def (DEF_BUILTIN): Factor out
28725 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
28726 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
28727 Remove ID. Adjust comments.
28728 * config/avr/avr-c.c (avr_builtin_name): Remove.
28729 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
28730 * config/avr/avr.c (avr_tolower): New static function.
28731 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
28732 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
28733 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
28734 default expansion.
28735
28736 2013-01-17 Jan Hubicka <jh@suse.cz>
28737
28738 PR tree-optimization/55273
28739 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
28740
28741 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
28742
28743 PR target/55981
28744 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
28745 store through atomic_store<mode>_1.
28746 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
28747
28748 2013-01-17 Martin Jambor <mjambor@suse.cz>
28749
28750 PR tree-optimizations/55264
28751 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
28752 for virtual methods.
28753 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
28754 virtual methods before inlining is over.
28755 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
28756 virtual functions.
28757 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
28758 non-virtual.
28759
28760 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
28761
28762 PR rtl-optimization/56005
28763 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
28764 pending reads for prefetch.
28765
28766 2013-01-16 Ian Bolton <ian.bolton@arm.com>
28767
28768 * config/aarch64/aarch64.md
28769 (*cstoresi_neg_uxtw): New pattern.
28770 (*cmovsi_insn_uxtw): New pattern.
28771 (*<optab>si3_uxtw): New pattern.
28772 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
28773 (*<optab>si3_insn_uxtw): New pattern.
28774 (*bswapsi2_uxtw): New pattern.
28775
28776 2013-01-16 Richard Biener <rguenther@suse.de>
28777
28778 * tree-inline.c (tree_function_versioning): Remove set but
28779 never used variable.
28780
28781 2013-01-16 Richard Biener <rguenther@suse.de>
28782
28783 PR tree-optimization/55964
28784 * tree-flow.h (rename_variables_in_loop): Remove.
28785 (rename_variables_in_bb): Likewise.
28786 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
28787 (copy_loop_before): Adjust and delete update-ssa status.
28788 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
28789 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
28790 (rename_variables_in_loop): Remove.
28791 (slpeel_update_phis_for_duplicate_loop): Likewise.
28792 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
28793 use available cfg machinery instead of duplicating it.
28794 Update PHI nodes and perform poor-mans SSA update here.
28795 (slpeel_tree_peel_loop_to_edge): Adjust.
28796
28797 2013-01-16 Richard Biener <rguenther@suse.de>
28798
28799 PR tree-optimization/54767
28800 PR tree-optimization/53465
28801 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
28802 (vrp_visit_phi_node): For PHI arguments coming via backedges
28803 drop all symbolical range information.
28804 (execute_vrp): Compute backedges.
28805
28806 2013-01-16 Richard Biener <rguenther@suse.de>
28807
28808 * doc/install.texi: Update CLooG and ISL requirements to
28809 0.18.0 and 0.11.1.
28810
28811 2013-01-16 Christian Bruel <christian.bruel@st.com>
28812
28813 PR target/55301
28814 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
28815 (broken_move): Handle UNSPECV_SP_SWITCH_B.
28816 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
28817
28818 2013-01-16 DJ Delorie <dj@redhat.com>
28819
28820 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
28821 (UNSPECV_SP_SWITCH_E): New.
28822 (sp_switch_1): Change to an unspec.
28823 (sp_switch_2): Change to an unspec. Don't use post-inc when we
28824 replace $r15.
28825
28826 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
28827
28828 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
28829 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
28830 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
28831 (expand_mem_thread_fence): Ditto.
28832 (expand_mem_signal_fence): Ditto.
28833 (expand_atomic_load): Ditto.
28834 (expand_atomic_store): Ditto.
28835
28836 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
28837
28838 PR rtl-optimization/55547
28839 PR rtl-optimization/53827
28840 PR debug/53671
28841 PR debug/49888
28842 * alias.c (memrefs_conflict_p): Set sizes to negative after
28843 AND adjustments.
28844
28845 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28846
28847 PR target/55940
28848 * function.c (thread_prologue_and_epilogue_insns): Always
28849 add crtl->drap_reg to set_up_by_prologue.set, even if
28850 stack_realign_drap is false.
28851
28852 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
28853
28854 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
28855 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
28856 *call): Fix indention.
28857
28858 2013-01-15 Tom de Vries <tom@codesourcery.com>
28859
28860 PR target/55876
28861 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
28862 Update comment.
28863
28864 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
28865
28866 PR rtl-optimization/55153
28867 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
28868
28869 2013-01-15 Martin Jambor <mjambor@suse.cz>
28870
28871 PR tree-optimization/55920
28872 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
28873 accesses as grp_to_be_debug_replaced.
28874
28875 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28876
28877 PR tree-optimization/55920
28878 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
28879 there is non-useless type conversion needed from debug rhs to lhs,
28880 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
28881
28882 2013-01-15 Joseph Myers <joseph@codesourcery.com>
28883 Mikael Pettersson <mikpe@it.uu.se>
28884
28885 PR target/43961
28886 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
28887 Thumb.
28888 (ASM_OUTPUT_CASE_LABEL): Remove.
28889 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
28890 * final.c (shorten_branches): Update alignment of labels before
28891 jump tables if CASE_VECTOR_SHORTEN_MODE.
28892
28893 2013-01-15 Richard Biener <rguenther@suse.de>
28894
28895 PR bootstrap/55961
28896 * system.h: Do not include gmp.h for building host tools.
28897
28898 2013-01-15 Richard Biener <rguenther@suse.de>
28899
28900 PR middle-end/55882
28901 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
28902 account for bitpos when computing alignment.
28903
28904 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
28905
28906 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
28907 (ix86_target_macros_internal): Likewise.
28908
28909 * config/i386/i386.c (m_CORE2I7): Removed.
28910 (m_CORE_HASWELL): New macro.
28911 (m_CORE_ALL): Likewise.
28912 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
28913 (initial_ix86_arch_features): Likewise.
28914 (processor_target_table): Initializations for Core avx2.
28915 (cpu_names): New names "core-avx2".
28916 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
28917 PROCESSOR_CORE_HASWELL.
28918 (ix86_issue_rate): New case.
28919 (ia32_multipass_dfa_lookahead): Likewise.
28920 (ix86_sched_init_global): Likewise.
28921
28922 * config/i386/i386.h (TARGET_HASWELL): New macro.
28923 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
28924 (processor_type): New PROCESSOR_HASWELL.
28925
28926 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28927
28928 PR tree-optimization/55955
28929 * tree-vect-loop.c (vectorizable_reduction): Give up early on
28930 *SHIFT_EXPR and *ROTATE_EXPR codes.
28931
28932 PR tree-optimization/48766
28933 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
28934 -ftrapv disable -fwrapv.
28935
28936 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28937
28938 PR target/55974
28939 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
28940 etc. to 1 and not to __flash.
28941 Use LL suffix for __INT24_MAX__ with -mint8.
28942 Use ULL suffix for __UINT24_MAX__ with -mint8.
28943
28944 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28945
28946 * config/avr/avr-arch.h
28947 (struct base_arch_s): Use typedef avr_arch_t instead.
28948 (struct arch_info_s): Use typedef avr_arch_info_t instead.
28949 (struct mcu_type_s): Use typedef avr_mcu_t instead.
28950 * config/avr/avr.c: Same.
28951 * config/avr/avr-devices.c: Same.
28952 * config/avr/driver-avr.c: Same.
28953 * config/avr/gen-avr-mmcu-texi.c: Same.
28954 * config/avr/avr-mcus.def: Adjust comment.
28955
28956 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
28957
28958 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
28959 * config/aarch64/iterators.md (VALLDI): New.
28960
28961 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
28962 Andi Kleen <ak@linux.intel.com>
28963
28964 PR target/55948
28965 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
28966 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
28967 memmodel flag.
28968
28969 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28970
28971 * config/avr/avr-stdint.h: Remove trailing blanks.
28972 * config/avr/avr-log.h: Same.
28973 * config/avr/avr-arch.h: Same.
28974 * config/avr/avr-devices.c: Same.
28975 * config/avr/avr-dimode.md: Same.
28976 * config/avr/predicates.md: Same.
28977 * config/avr/avr-c.c: Same. And fix typo.
28978
28979 * config/avr/avr-protos.h: Same. And:
28980 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
28981 (init_cumulative_args): Rename to avr_init_cumulative_args.
28982 (expand_prologue): Rename to avr_expand_prologue.
28983 (expand_epilogue): Rename to avr_expand_epilogue.
28984 (adjust_insn_length): Rename to avr_adjust_insn_length.
28985 (notice_update_cc): Rename to avr_notice_update_cc.
28986 (final_prescan_insn): Rename to avr_final_prescan_insn.
28987 * config/avr/avr.c: Same.
28988 * config/avr/avr.h: Same.
28989 * config/avr/avr.md: Remove trailing blanks.
28990 (prologue): Use avr_expand_prologue.
28991 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
28992
28993 2013-01-14 Richard Biener <rguenther@suse.de>
28994
28995 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
28996 verify_location, collect_subblocks): New functions.
28997 (verify_gimple_in_cfg): Verify that locations only reference
28998 BLOCKs in the functions BLOCK tree.
28999
29000 2013-01-14 Richard Biener <rguenther@suse.de>
29001
29002 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
29003 PHI argument.
29004 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
29005 unshare reference.
29006 (insert_out_of_ssa_copy_on_edge): Likewise.
29007 (rewrite_close_phi_out_of_ssa): Likewise.
29008 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
29009 debug expressions.
29010 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
29011 propagated constants.
29012 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
29013 can not be shared.
29014
29015 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29016
29017 * config/avr/avr-modes.def: Add GPL copyright notice.
29018
29019 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
29020
29021 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
29022 MEMMODEL_MASK to determine memory model.
29023 (atomic_store<mode>): Ditto from operands[2].
29024 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
29025
29026 2013-01-13 Jakub Jelinek <jakub@redhat.com>
29027
29028 PR fortran/55935
29029 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
29030 (fold_gimple_assign): Don't call unshare_expr here.
29031 (fold_ctor_reference): Call unshare_expr.
29032
29033 2013-01-13 Terry Guo <terry.guo@arm.com>
29034
29035 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
29036 * doc/fragments.texi: Document MULTILIB_REUSE.
29037 * gcc.c (multilib_reuse): New internal spec.
29038 (set_multilib_dir): Also search multilib from multilib_reuse.
29039 * genmultilib (tmpmultilib3): Refactor code.
29040 (tmpmultilib4): Ditto.
29041 (multilib_reuse): New multilib argument.
29042
29043 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
29044
29045 * Makefile.in: Update copyright.
29046
29047 2013-01-12 Tom de Vries <tom@codesourcery.com>
29048
29049 PR middle-end/55890
29050 * calls.c (expand_call): Check if arg_nr is valid.
29051
29052 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
29053
29054 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
29055 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
29056 documentation. Add missing '__' in front of
29057 __builtin_ia32_packssdw256.
29058
29059 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29060
29061 PR target/55719
29062 * config/s390/s390.c (s390_preferred_reload_class): Do not return
29063 NO_REGS for larl operands.
29064 (s390_reload_larl_operand): Use s390_load_address instead of
29065 emit_move_insn.
29066
29067 2013-01-11 Richard Biener <rguenther@suse.de>
29068
29069 * tree-cfg.c (verify_node_sharing_1): Split out from ...
29070 (verify_node_sharing): ... here.
29071 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
29072
29073 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
29074
29075 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
29076 Substitute TREECHECKING.
29077 * configure: Regenerate.
29078 * Makefile.in (TREECHECKING): New.
29079
29080 2013-01-11 Richard Guenther <rguenther@suse.de>
29081
29082 PR tree-optimization/44061
29083 * tree-vrp.c (extract_range_basic): Compute zero as
29084 value-range for __builtin_constant_p of function parameters.
29085
29086 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
29087
29088 Update copyright years.
29089
29090 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
29091
29092 PR rtl-optimization/55672
29093 * lra-eliminations.c (mark_not_eliminable): Permit addition with
29094 const to be eliminable.
29095
29096 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
29097
29098 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
29099 * configure: Regenerate.
29100
29101 2013-01-10 Richard Biener <rguenther@suse.de>
29102
29103 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
29104
29105 2013-01-10 Richard Biener <rguenther@suse.de>
29106
29107 PR bootstrap/55792
29108 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
29109 locations for virtual PHI arguments.
29110 (rewrite_update_phi_arguments): Likewise.
29111
29112 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
29113
29114 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
29115 on to assembler.
29116
29117 2013-01-10 Jakub Jelinek <jakub@redhat.com>
29118
29119 PR tree-optimization/55921
29120 * tree-complex.c (expand_complex_asm): New function.
29121 (expand_complex_operations_1): Call it for GIMPLE_ASM.
29122
29123 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29124
29125 PR target/55718
29126 * config/s390/s390.c (s390_symref_operand_p)
29127 (s390_loadrelative_operand_p): Merge the two functions.
29128 (s390_check_qrst_address, print_operand_address): Add parameters
29129 to s390_loadrelative_operand_p invokation.
29130 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
29131 (s390_reload_larl_operand, s390_secondary_reload): Use
29132 s390_loadrelative_operand_p instead of s390_symref_operand_p.
29133 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
29134
29135 2013-01-09 Mike Stump <mikestump@comcast.net>
29136
29137 * dse.c (record_store): Remove unnecessary assert.
29138
29139 2013-01-09 Jan Hubicka <jh@suse.cz>
29140
29141 PR tree-optimization/55569
29142 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
29143 * cfgloop.h (scale_loop_profile): Likewise.
29144
29145 2013-01-09 Jan Hubicka <jh@suse.cz>
29146
29147 PR lto/45375
29148 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
29149 functions.
29150 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
29151
29152 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
29153
29154 PR middle-end/55114
29155 * expr.h (maybe_emit_group_store): Declare.
29156 * expr.c (maybe_emit_group_store): New function.
29157 * builtins.c (expand_builtin_int_roundingfn): Call it.
29158 (expand_builtin_int_roundingfn_2): Likewise.
29159
29160 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
29161
29162 PR rtl-optimization/55829
29163 * lra-constraints.c (match_reload): Add code for absent output.
29164 (curr_insn_transform): Add code for reloads of matched inputs
29165 without output.
29166
29167 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
29168
29169 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
29170 attribute of movddup insn to DF.
29171 (*vec_interleave_lowv2df): Ditto.
29172 (vec_dupv2df): Ditto.
29173
29174 2013-01-09 Jan Hubicka <jh@suse.cz>
29175
29176 PR tree-optimiation/55875
29177 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
29178 EVERY_ITERATION parameter.
29179 (number_of_iterations_exit): Check if exit is executed every iteration.
29180 (idx_infer_loop_bounds): Similarly here.
29181 (n_of_executions_at_most): Simplify
29182 to only test for cases where statement is dominated by the
29183 particular bound; handle correctly the "postdominance" test.
29184 (scev_probably_wraps_p): Use max loop iterations info
29185 as a global bound first.
29186
29187 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
29188 Nick Clifton <nickc@redhat.com>
29189
29190 * config/v850/v850.md (cbranchsf4): New pattern.
29191 (cstoresf4): New pattern.
29192 (cbranchdf4): New pattern.
29193 (cstoredf4): New pattern.
29194 (movsicc): Disallow floating point comparisons.
29195 (cmpsf_le_insn): Fix order of operators.
29196 (cmpsf_lt_insn): Likewise.
29197 (cmpsf_eq_insn): Likewise.
29198 (cmpdf_le_insn): Likewise.
29199 (cmpdf_lt_insn): Likewise.
29200 (cmpdf_eq_insn): Likewise.
29201 (cmpsf_ge_insn): Use LE comparison.
29202 (cmpdf_ge_insn): Likewise.
29203 (cmpsf_gt_insn): Use LT comparison.
29204 (cmpdf_gt_insn): Likewise.
29205 (cmpsf_ne_insn): Delete pattern.
29206 (cmpdf_ne_insn): Delete pattern.
29207 * config/v850/v850.c (v850_gen_float_compare): Use
29208 gen_cmpdf_eq_insn for NE comparison.
29209 (v850_float_z_comparison_operator)
29210 (v850_float_nz_comparison_operator): Move from here ...
29211 * config/v850/predicates.md: ... to here. Move GT and GE
29212 comparisons into v850_float_z_comparison_operator.
29213 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
29214 Delete prototype.
29215 (v850_float_nz_comparison_operator): Likewise.
29216
29217 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29218
29219 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
29220 with calls to gen_insvsi/gen_insvdi.
29221
29222 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29223
29224 * config/i386/i386.c (initial_ix86_tune_features): Set up
29225 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
29226
29227 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
29228 Jakub Jelinek <jakub@redhat.com>
29229
29230 PR tree-optimization/48189
29231 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
29232 If nitercst is 0, don't predict the exit edge.
29233
29234 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29235
29236 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
29237 in asm_fprintf with reg_names.
29238 (aarch64_print_operand_address): Likewise.
29239 (aarch64_return_addr): Likewise.
29240 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
29241
29242 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29243
29244 * config/pa/pa.h (VAL_U6_BITS_P): Define.
29245 (INT_U6_BITS): Likewise.
29246 * config/pa/predicates.md (uint6_operand): New predicate.
29247 (shift5_operand, shift6_operand): Likewise.
29248 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
29249 arith32_operand.
29250 (lshrdi3): Use shift6_operand.
29251 (shrpsi4, shrpdi4): New insn patterns.
29252 (extzv): Delete expander.
29253 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
29254 predicates in unamed zero extract patterns. Tighten common constraint.
29255 (extv): Delete expander.
29256 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
29257 predicates in unamed sign extract patterns. Tighten common constraint.
29258 (insv): Delete expander.
29259 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
29260 predicates in unamed insert patterns. Tighten common constraint.
29261 Change uint32_operand predicate to uint6_operand predicate in unamed
29262 DImode pattern to insert constant values of type 1...1xxxx.
29263
29264 2013-01-04 Jan Hubicka <jh@suse.cz>
29265
29266 PR tree-optimization/55823
29267 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
29268 issue.
29269
29270 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29271 Uros Bizjak <ubizjak@gmail.com>
29272
29273 PR rtl-optimization/55845
29274 * df-problems.c (can_move_insns_across): Stop scanning at
29275 volatile_insn_p source instruction or give up if
29276 across_from .. across_to range contains any volatile_insn_p
29277 instructions.
29278
29279 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29280
29281 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
29282 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
29283 Declare.
29284 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
29285 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
29286
29287 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29288
29289 PR fortran/55341
29290 * asan.c (asan_clear_shadow): New function.
29291 (asan_emit_stack_protection): Use it.
29292
29293 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29294
29295 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
29296 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
29297 with tab instead of space.
29298
29299 2013-01-08 Nick Clifton <nickc@redhat.com>
29300
29301 * config/rl78/rl78.c (rl78_expand_prologue): Always select
29302 register bank 0 at the start of an interrupt handler.
29303 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
29304 MDBH registers.
29305
29306 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29307
29308 * config/aarch64/aarch64-simd.md
29309 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
29310 (aarch64_simd_bsl): Likewise.
29311 (aarch64_vcond_internal<mode>): Likewise.
29312 (vcond<mode><mode>): Likewise.
29313 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
29314 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
29315
29316 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29317
29318 * config/aarch64/aarch64-builtins.c
29319 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
29320
29321 2013-01-08 Martin Jambor <mjambor@suse.cz>
29322
29323 PR debug/55579
29324 * tree-sra.c (analyze_access_subtree): Return true also after
29325 potentially creating a debug-only replacement.
29326
29327 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29328
29329 PR middle-end/55890
29330 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
29331
29332 PR tree-optimization/54120
29333 * tree-vrp.c (range_fits_type_p): Don't allow
29334 src_precision < precision from signed vr to unsigned_p
29335 if vr->min or vr->max is negative.
29336 (simplify_float_conversion_using_ranges): Test can_float_p
29337 against CODE_FOR_nothing.
29338
29339 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29340 Richard Biener <rguenther@suse.de>
29341
29342 PR middle-end/55851
29343 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
29344 types instead of just INTEGER_TYPE types.
29345
29346 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
29347
29348 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
29349 TF_SIZE): Define.
29350
29351 2013-01-07 Steve Ellcey <sellcey@mips.com>
29352
29353 PR target/42661
29354 * config/mips/mips.opt: Change mad to mmad to match documentation.
29355
29356 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29357
29358 PR target/55897
29359 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
29360 .progmemx.data now.
29361
29362 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29363
29364 PR target/55897
29365 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
29366 (avr_addrspace_t): Add .section_name field.
29367 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
29368 array size.
29369 (avr_addrspace): Same. Initialize .section_name. Remove last
29370 NULL entry. Put __memx into .progmemx.data.
29371 (progmem_section_prefix): Remove.
29372 (avr_asm_init_sections): No need to initialize progmem_section.
29373 (avr_asm_named_section): Use avr_addrspace[].section_name to get
29374 section name prefix.
29375 (avr_asm_select_section): Ditto. And use get_unnamed_section to
29376 retrieve the progmem section.
29377 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
29378 boundary to run over avr_addrspace[].
29379 (avr_register_target_pragmas): Ditto.
29380
29381 2013-01-06 Jakub Jelinek <jakub@redhat.com>
29382
29383 * varasm.c (output_constant_def_contents): For asan_protect_global
29384 protected strings, adjust DECL_ALIGN if needed, before testing for
29385 anchored symbols.
29386 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
29387 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
29388 normal decls.
29389 (output_object_block): For asan protected decls, emit asan padding
29390 after their contents.
29391 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
29392 (asan_finish_file): Test it here instead.
29393
29394 2013-01-07 Nick Clifton <nickc@redhat.com>
29395 Matthias Klose <doko@debian.org>
29396 Doug Kwan <dougkwan@google.com>
29397 H.J. Lu <hongjiu.lu@intel.com>
29398
29399 PR driver/55470
29400 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
29401
29402 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
29403
29404 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
29405
29406 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
29407
29408 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
29409
29410 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29411
29412 PR target/54461
29413 * doc/install.texi (Cross-Compiler-Specific Options): Document
29414 --with-avrlibc.
29415
29416 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
29417
29418 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
29419 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
29420 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
29421 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
29422 vqmovun_high_s64): Fix source operand number and update copyright.
29423
29424 2013-01-07 Richard Biener <rguenther@suse.de>
29425
29426 PR middle-end/55890
29427 * gimple.h (gimple_call_builtin_p): New overload.
29428 * gimple.c (validate_call): New function.
29429 (gimple_call_builtin_p): Likewise.
29430 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
29431 Use gimple_call_builtin_p.
29432 (find_func_clobbers): Likewise.
29433 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
29434 (strlen_optimize_stmt): Likewise.
29435
29436 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29437
29438 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
29439 (vld1q_dup_*): Likewise.
29440 (vld1_*): Likewise.
29441 (vld1q_*): Likewise.
29442 (vld1_lane_*): Likewise.
29443 (vld1q_lane_*): Likewise.
29444
29445 2013-01-07 Richard Biener <rguenther@suse.de>
29446
29447 * lto-streamer.h (LTO_minor_version): Bump to 2.
29448
29449 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29450
29451 * config/aarch64/aarch64-protos.h
29452 (aarch64_const_double_zero_rtx_p): Rename to...
29453 (aarch64_float_const_zero_rtx_p): ...this.
29454 (aarch64_float_const_representable_p): New.
29455 (aarch64_output_simd_mov_immediate): Likewise.
29456 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
29457 move immediate case.
29458 * config/aarch64/aarch64.c
29459 (aarch64_const_double_zero_rtx_p): Rename to...
29460 (aarch64_float_const_zero_rtx_p): ...this.
29461 (aarch64_print_operand): Allow printing of new constants.
29462 (aarch64_valid_floating_const): New.
29463 (aarch64_legitimate_constant_p): Check for valid floating-point
29464 constants.
29465 (aarch64_simd_valid_immediate): Likewise.
29466 (aarch64_vect_float_const_representable_p): New.
29467 (aarch64_float_const_representable_p): Likewise.
29468 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
29469 (aarch64_output_simd_mov_immediate): New.
29470 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
29471 (*movdf_aarch64): Likewise.
29472 * config/aarch64/constraints.md (Ufc): New.
29473 (Y): call aarch64_float_const_zero_rtx.
29474 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
29475
29476 2013-01-07 Richard Biener <rguenther@suse.de>
29477
29478 PR tree-optimization/55888
29479 PR tree-optimization/55862
29480 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
29481 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
29482 not if it is contained therein.
29483
29484 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29485
29486 * config/avr/t-avr: Typo.
29487
29488 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29489
29490 PR55243
29491 * config/avr/t-avr: Don't automatically rebuild
29492 $(srcdir)/config/avr/t-multilib
29493 $(srcdir)/config/avr/avr-tables.opt
29494 $(srcdir)/doc/avr-mmcu.texi
29495 (avr-mcus): New phony target to build them on request.
29496 (s-avr-mlib, s-avr-mmcu-texi): Remove.
29497 * avr/avr-mcus.def: Adjust comments.
29498
29499 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
29500
29501 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
29502
29503 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29504
29505 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
29506
29507 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29508
29509 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
29510
29511 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
29512
29513 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
29514 to generate profiling.
29515 * config/rs6000/aix64.h (LIB_SPEC): Same.
29516
29517 2013-01-04 Andrew Pinski <apinski@cavium.com>
29518
29519 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
29520 New function.
29521 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
29522
29523 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
29524
29525 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
29526 unconditionally.
29527 (ix86_expand_move): Ditto.
29528 (ix86_zero_extend_to_Pmode): Ditto.
29529 (ix86_expand_call): Ditto.
29530 (ix86_expand_special_args_builtin): Ditto.
29531 (ix86_expand_builtin): Ditto.
29532
29533 2013-01-04 Richard Biener <rguenther@suse.de>
29534
29535 PR tree-optimization/55862
29536 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
29537 translating them through PHI nodes.
29538
29539 2013-01-04 Martin Jambor <mjambor@suse.cz>
29540
29541 PR tree-optimization/55755
29542 * tree-sra.c (sra_modify_assign): Do not check that an access has no
29543 children when trying to avoid producing a VIEW_CONVERT_EXPR.
29544
29545 2013-01-04 Marek Polacek <polacek@redhat.com>
29546
29547 PR middle-end/55859
29548 * opts.c (default_options_optimization): Clarify error message.
29549
29550 2013-01-04 Richard Biener <rguenther@suse.de>
29551
29552 PR middle-end/55863
29553 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
29554 reassociation.
29555
29556 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29557
29558 PR target/53789
29559 * config/pa/pa.md (movsi): Revert previous change.
29560 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
29561 references.
29562
29563 2013-01-03 Richard Henderson <rth@redhat.com>
29564
29565 * config/i386/i386.c (ix86_expand_move): Always assign to op1
29566 after eliminating TLS symbols.
29567
29568 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
29569
29570 PR bootstrap/50167
29571 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
29572 * graphite-poly.c (debug_gmp_value): Likewise.
29573
29574 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
29575
29576 PR target/55712
29577 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
29578 selected code model, define __code_mode_small__, __code_model_medium__,
29579 __code_model_large__, __code_model_32__ or __code_model_kernel__.
29580 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
29581 xchg temporary register with %k. Declare temporary register as
29582 early clobbered.
29583 [__x86_64__]: For medium and large code models, preserve %rbx register.
29584
29585 2013-01-03 Richard Biener <rguenther@suse.de>
29586
29587 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
29588 (dump_subscript): Adjust.
29589 (finalize_ddr_dependent): Do not dump redundant info.
29590 (analyze_siv_subscript): Adjust.
29591 (subscript_dependence_tester): Likewise.
29592 (compute_affine_dependence): Likewise.
29593
29594 2013-01-03 Richard Biener <rguenther@suse.de>
29595
29596 Revert
29597 2013-01-03 Richard Biener <rguenther@suse.de>
29598
29599 PR tree-optimization/55857
29600 * tree-vect-stmts.c (vectorizable_load): Do not setup
29601 re-alignment for invariant loads.
29602
29603 2013-01-02 Richard Biener <rguenther@suse.de>
29604
29605 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29606 invariant load do not generate a vector load from the scalar location.
29607
29608 2013-01-03 Richard Biener <rguenther@suse.de>
29609
29610 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
29611 for not vectorizing.
29612 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
29613 not build INDIRECT_REFs, call get_name once only.
29614 (vect_create_data_ref_ptr): Likewise. Dump base object kind
29615 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
29616
29617 2013-01-03 Richard Biener <rguenther@suse.de>
29618
29619 PR tree-optimization/55857
29620 * tree-vect-stmts.c (vectorizable_load): Do not setup
29621 re-alignment for invariant loads.
29622
29623 2013-01-03 Richard Biener <rguenther@suse.de>
29624
29625 PR lto/55848
29626 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
29627 prefer a built-in decl.
29628
29629 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29630
29631 * gcc.c (process_command): Update copyright notice dates.
29632 * gcov.c (print_version): Likewise.
29633 * gcov-dump.c (print_version): Likewise.
29634
29635 PR rtl-optimization/55838
29636 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
29637 iv0.step, iv1.step and step.
29638
29639 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29640 Marc Glisse <marc.glisse@inria.fr>
29641
29642 PR tree-optimization/55832
29643 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
29644 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
29645 integer_{one,zero}_node.
29646
29647 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29648
29649 PR debug/54402
29650 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
29651 * var-tracking.c (reverse_op): Don't add reverse ops to
29652 VALUEs that have already
29653 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
29654
29655 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
29656
29657 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
29658
29659 2013-01-02 Teresa Johnson <tejohnson@google.com>
29660
29661 * dumpfile.c (dump_loc): Print filename with location.
29662 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
29663 new location_t parameter to emit complete unroll message with
29664 new dump framework.
29665 (canonicalize_loop_induction_variables): Compute loops location
29666 and pass to try_unroll_loop_completely.
29667 * loop-unroll.c (report_unroll_peel): New function.
29668 (peel_loops_completely): Use new dump format with location
29669 for main dumpfile message, and invoke report_unroll_peel on success.
29670 (decide_unrolling_and_peeling): Ditto.
29671 (decide_peel_once_rolling): Remove old dumpfile message subsumed
29672 by report_unroll_peel.
29673 (decide_peel_completely): Ditto.
29674 (decide_unroll_constant_iterations): Ditto.
29675 (decide_unroll_runtime_iterations): Ditto.
29676 (decide_peel_simple): Ditto.
29677 (decide_unroll_stupid): Ditto.
29678 * cfgloop.c (get_loop_location): New function.
29679 * cfgloop.h (get_loop_location): Declare.
29680
29681 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29682
29683 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
29684 NULL.
29685
29686 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29687
29688 PR middle-end/55198
29689 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
29690 BLKmode objects when EXPAND_MEMORY is specified.
29691
29692 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29693
29694 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
29695 in loop predicate.
29696 (fold_builtin_cpu): Do not share cpu model decls across statements.
29697
29698 2013-01-02 Jason Merrill <jason@redhat.com>
29699
29700 PR c++/55804
29701 * tree.c (build_array_type_1): Revert earlier change.
29702
29703 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
29704
29705 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
29706 "cortex-a57".
29707 * config/aarch64/aarch64-tune.md: Re-generate.
29708
29709 2013-01-02 Richard Biener <rguenther@suse.de>
29710
29711 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29712 invariant load do not generate a vector load from the scalar location.
29713
29714 2013-01-02 Richard Biener <rguenther@suse.de>
29715
29716 PR bootstrap/55784
29717 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
29718 * configure: Regenerate.
29719
29720 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29721
29722 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
29723 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
29724 (expand_builtin_int_roundingfn_2): Keep the original target around
29725 for the fallback case.
29726
29727 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29728
29729 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
29730 to be clear for sign changes.
29731
29732 2013-01-01 Jan Hubicka <jh@suse.cz>
29733
29734 * ipa-inline-analysis.c: Fix formatting.
29735
29736 2013-01-01 Jakub Jelinek <jakub@redhat.com>
29737
29738 PR tree-optimization/55831
29739 * tree-vect-loop.c (get_initial_def_for_induction): Use
29740 gsi_after_labels instead of gsi_start_bb.
29741 \f
29742 Copyright (C) 2013 Free Software Foundation, Inc.
29743
29744 Copying and distribution of this file, with or without modification,
29745 are permitted in any medium without royalty provided the copyright
29746 notice and this notice are preserved.