1 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
3 PR rtl-optimization/58968
4 * lra-spills.c (return_regno_p): New function.
5 (lra_final_code_change): Use it.
7 2013-11-04 Joseph Myers <joseph@codesourcery.com>
9 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document
11 * target.def (float_exceptions_rounding_supported_p): New hook.
12 * targhooks.c (default_float_exceptions_rounding_supported_p): New
14 * targhooks.h (default_float_exceptions_rounding_supported_p):
16 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
18 * doc/tm.texi: Regenerate.
19 * config.gcc (powerpc*-*-linux*): Set extra_objs.
20 * config/rs6000/rs6000-linux.c: New file.
21 * config/rs6000/rs6000-protos.h
22 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
23 * config/rs6000/linux.h
24 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
25 * config/rs6000/linux64.h
26 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
27 * config/rs6000/t-linux (rs6000-linux.o): New rule.
28 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
30 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
32 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
33 Replace the define_insn_and_split with a define_insn and two
34 define_splits, with the split after reload re-permuting the source
35 register to its original value.
36 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
37 (*vsx_le_perm_store_v8hi): Likewise.
38 (*vsx_le_perm_store_v16qi): Likewise.
40 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
42 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
45 2013-11-04 Jakub Jelinek <jakub@redhat.com>
47 PR tree-optimization/58946
48 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
49 bbs with bbinfo[idx].op != NULL before all blocks with
50 bbinfo[idx].op == NULL.
52 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
54 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
55 (avr_dump_double_int_hex): Likewise.
56 (avr_log_vadump): Remove %D and %X handling.
57 * config/avr/avr.c (avr_double_int_push_digit): Delete.
58 (avr_map_op_t): Change map from double_int to unsigned int.
59 (avr_map_op): Update accordingly.
60 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
61 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
62 unsigned ints rather than double_ints.
64 2013-11-03 Marek Polacek <polacek@redhat.com>
66 Implement -fsanitize=vla-bound.
67 * opts.c (common_handle_option): Handle vla-bound.
68 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE):
70 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
71 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
73 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
75 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
78 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
80 * config/i386/constraints.md (Ts, Tv): New address constrains.
81 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
82 constraint for address_no_seg_operand.
83 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
84 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
85 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
86 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
87 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
88 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
89 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
90 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
91 constraint for vsib_address_operand.
93 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
95 * gcse.c (pre_delete): Remove references to regmove from comments.
96 * recog.c: (validate_replace_rtx_1): Likewise.
97 * config/rl78/rl78.c: Likewise.
98 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
99 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
100 * common/config/mmix/mmix-common.c: Likewise.
102 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
104 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
105 * gimplify.c (gimplify_compound_lval): Likewise.
106 * graphite-clast-to-gimple.c (gloog): Likewise.
107 * graphite-dependences.c (loop_is_parallel_p): Likewise.
108 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
109 (limit_scop); Likewise.
110 (build_scops): Likewise.
111 (dot_scop): Likewise.
112 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
113 (build_scop_drs): Likewise.
114 (insert_stmts): Likewise.
115 (insert_out_of_ssa_copy): Likewise.
116 (remove_phi): Likewise.
117 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
118 * hw-doloop.c (discover_loop): Likewise.
119 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
120 * tree-dfa.c (dump_enumerated_decls): Likewise.
121 * tree-if-conv.c (if_convertable_loop_p): Likewise.
122 * tree-inline.c (tree_function_versioning): Likewise.
123 * tree-loop-distribution.c (build_rdg): Likewise.
124 (rdg_flag_vertex_and_dependent): Likewise.
125 (distribute_loop): Likewise.
126 * tree-parloops.c (loop_parallel_p): Likewise.
127 (eliminate_local_variables): Likewise.
128 (separate_decls_in_region): Likewise.
129 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
130 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
131 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
132 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
133 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
134 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
135 (vectorizable_condition): Likewise.
137 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
139 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
140 * configure: Regenerate.
141 * config/i386/i386.md (*movdi_internal): Change
142 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
143 (*movdf_internal): Ditto.
144 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
145 * config/i386/sse.md (vec_concatv2di): Output interunit movq
146 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
148 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
150 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
153 2013-10-31 David Malcolm <dmalcolm@redhat.com>
155 Automated part of renaming of symtab_node_base to symtab_node.
157 Patch autogenerated by rename_symtab.py from
158 https://github.com/davidmalcolm/gcc-refactoring-scripts
159 revision 58bb219cc090b2f4516a9297d868c245495ee622
160 with ChangeLog entry fixed up by hand.
162 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
164 (cgraph_node_for_asm): Likewise.
165 * cgraph.h (symtab_node_base): Likewise.
166 (cgraph_node): Likewise.
167 (varpool_node): Likewise.
168 (is_a_helper <cgraph_node>::test): Likewise.
169 (is_a_helper <varpool_node>::test): Likewise.
170 (symtab_nodes): Likewise.
171 (symtab_register_node): Likewise.
172 (symtab_unregister_node): Likewise.
173 (symtab_remove_node): Likewise.
174 (symtab_get_node): Likewise.
175 (symtab_node_for_asm): Likewise.
176 (symtab_node_asm_name): Likewise.
177 (symtab_node_name): Likewise.
178 (symtab_insert_node_to_hashtable): Likewise.
179 (symtab_add_to_same_comdat_group): Likewise.
180 (symtab_dissolve_same_comdat_group_list): Likewise.
181 (dump_symtab_node): Likewise.
182 (debug_symtab_node): Likewise.
183 (dump_symtab_base): Likewise.
184 (verify_symtab_node): Likewise.
185 (verify_symtab_base): Likewise.
186 (symtab_used_from_object_file_p): Likewise.
187 (symtab_alias_ultimate_target): Likewise.
188 (symtab_resolve_alias): Likewise.
189 (fixup_same_cpp_alias_visibility): Likewise.
190 (symtab_for_node_and_aliases): Likewise.
191 (symtab_nonoverwritable_alias): Likewise.
192 (availability symtab_node_availability): Likewise.
193 (symtab_semantically_equivalent_p): Likewise.
194 (fixup_same_cpp_alias_visibility): Likewise.
195 (symtab_prevail_in_asm_name_hash): Likewise.
198 (varpool_first_variable): Likewise.
199 (varpool_next_variable): Likewise.
200 (varpool_first_static_initializer): Likewise.
201 (varpool_next_static_initializer): Likewise.
202 (varpool_first_defined_variable): Likewise.
203 (varpool_next_defined_variable): Likewise.
204 (cgraph_first_defined_function): Likewise.
205 (cgraph_next_defined_function): Likewise.
206 (cgraph_first_function): Likewise.
207 (cgraph_next_function): Likewise.
208 (cgraph_first_function_with_gimple_body): Likewise.
209 (cgraph_next_function_with_gimple_body): Likewise.
210 (symtab_alias_target): Likewise.
211 (symtab_real_symbol_p): Likewise.
212 (symtab_can_be_discarded): Likewise.
213 * cgraphbuild.c (mark_address): Likewise.
214 (mark_load): Likewise.
215 (mark_store): Likewise.
216 * cgraphunit.c (decide_is_symbol_needed): Likewise.
218 (enqueue_node): Likewise.
219 (referred_to_p): Likewise.
220 (cgraph_process_same_body_aliases): Likewise.
221 (analyze_functions): Likewise.
222 (handle_alias_pairs): Likewise.
223 (output_weakrefs): Likewise.
225 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
226 * ipa-inline-analysis.c (inline_write_summary): Likewise.
227 * ipa-prop.c (remove_described_reference): Likewise.
228 (try_decrement_rdesc_refcount): Likewise.
229 (ipa_edge_duplication_hook): Likewise.
230 * ipa-ref.c (ipa_record_reference): Likewise.
231 (ipa_maybe_record_reference): Likewise.
232 (ipa_clone_ref): Likewise.
233 (ipa_clone_references): Likewise.
234 (ipa_clone_referring): Likewise.
235 (ipa_find_reference): Likewise.
236 (ipa_remove_stmt_references): Likewise.
237 (ipa_clear_stmts_in_references): Likewise.
238 * ipa-ref.h (symtab_node_base): Likewise.
240 (ipa_record_reference): Likewise.
241 (ipa_maybe_record_reference): Likewise.
242 (ipa_clone_references): Likewise.
243 (ipa_clone_referring): Likewise.
244 (ipa_clone_ref): Likewise.
245 (ipa_find_reference): Likewise.
246 (ipa_remove_stmt_references): Likewise.
247 (ipa_clear_stmts_in_references): Likewise.
248 * ipa-reference.c (ipa_reference_write_optimization_summary):
250 * ipa.c (enqueue_node): Likewise.
251 (process_references): Likewise.
252 (walk_polymorphic_call_targets): Likewise.
253 (symtab_remove_unreachable_nodes): Likewise.
254 (address_taken_from_non_vtable_p): Likewise.
255 (comdat_can_be_unshared_p_1): Likewise.
256 (comdat_can_be_unshared_p): Likewise.
257 (can_replace_by_local_alias): Likewise.
258 (function_and_variable_visibility): Likewise.
259 * is-a.h: Likewise (within example in comment).
260 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
261 (lto_symtab_encoder_encode): Likewise.
262 (lto_symtab_encoder_delete_node): Likewise.
263 (lto_symtab_encoder_in_partition_p): Likewise.
264 (lto_set_symtab_encoder_in_partition): Likewise.
265 (output_refs): Likewise.
266 (compute_ltrans_boundary): Likewise.
267 (output_symtab): Likewise.
268 (input_node): Likewise.
269 (input_ref): Likewise.
270 (input_edge): Likewise.
271 (input_cgraph_1): Likewise.
272 (input_refs): Likewise.
273 (output_cgraph_opt_summary): Likewise.
274 (input_node_opt_summary): Likewise.
275 (input_cgraph_opt_section): Likewise.
276 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
278 * lto-streamer-out.c (lto_output): Likewise.
279 (output_symbol_p): Likewise.
280 (produce_symtab): Likewise.
281 * lto-streamer.h (lto_encoder_entry): Likewise.
282 (lto_free_function_in_decl_state_for_node): Likewise.
283 (lto_symtab_encoder_encode): Likewise.
284 (lto_symtab_encoder_delete_node): Likewise.
285 (lto_symtab_encoder_in_partition_p): Likewise.
286 (lto_set_symtab_encoder_in_partition): Likewise.
287 (lto_symtab_encoder_lookup): Likewise.
288 (lsei_node): Likewise.
289 (lto_symtab_encoder_deref): Likewise.
290 * symtab.c (symtab_hash): Likewise.
291 (assembler_name_hash): Likewise.
292 (symtab_nodes): Likewise.
293 (hash_node): Likewise.
295 (hash_node_by_assembler_name): Likewise.
296 (eq_assembler_name): Likewise.
297 (insert_to_assembler_name_hash): Likewise.
298 (unlink_from_assembler_name_hash): Likewise.
299 (symtab_prevail_in_asm_name_hash): Likewise.
300 (symtab_register_node): Likewise.
301 (symtab_insert_node_to_hashtable): Likewise.
302 (symtab_unregister_node): Likewise.
303 (symtab_get_node): Likewise.
304 (symtab_remove_node): Likewise.
305 (symtab_initialize_asm_name_hash): Likewise.
306 (symtab_node_for_asm): Likewise.
307 (symtab_add_to_same_comdat_group): Likewise.
308 (symtab_dissolve_same_comdat_group_list): Likewise.
309 (symtab_node_asm_name): Likewise.
310 (symtab_node_name): Likewise.
311 (dump_symtab_base): Likewise.
312 (dump_symtab_node): Likewise.
313 (dump_symtab): Likewise.
314 (debug_symtab_node): Likewise.
315 (verify_symtab_base): Likewise.
316 (verify_symtab_node): Likewise.
317 (verify_symtab): Likewise.
318 (symtab_used_from_object_file_p): Likewise.
319 (symtab_node_availability): Likewise.
320 (symtab_alias_ultimate_target): Likewise.
321 (fixup_same_cpp_alias_visibility): Likewise.
322 (symtab_resolve_alias): Likewise.
323 (symtab_for_node_and_aliases): Likewise.
324 (symtab_for_node_and_aliases): Likewise.
325 (symtab_nonoverwritable_alias_1): Likewise.
326 (symtab_nonoverwritable_alias): Likewise.
327 (symtab_semantically_equivalent_p): Likewise.
328 * value-prof.c (init_node_map): Likewise.
329 * varasm.c (find_decl): Likewise.
330 * varpool.c (varpool_node_for_asm): Likewise.
331 (varpool_remove_unreferenced_decls): Likewise.
333 2013-10-31 David Malcolm <dmalcolm@redhat.com>
335 Manual part of renaming of symtab_node_base to symtab_node.
337 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
338 clashes with the preferred name for the base class.
339 (const_symtab_node): Remove redundant typedef.
341 2013-10-31 Jakub Jelinek <jakub@redhat.com>
343 * optabs.c (expand_vec_perm): Avoid vector mode punning
345 * expmed.c (store_bit_field_1): Likewise.
346 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
347 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
348 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
349 * config/i386/i386.c (ix86_expand_vector_move_misalign,
350 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
351 ix86_expand_sse_unpack, ix86_expand_args_builtin,
352 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
353 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
354 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
355 expand_vec_perm_vpshufb2_vpermq,
356 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
357 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
358 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
359 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
362 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
364 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
365 Use gcc_unreachable for unhandled modes. Do not check results of
366 expand_simple_binop. If not expanded to target, move the result.
368 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
369 Shiva Chen <shiva0217@gmail.com>
371 * config.gcc (nds32*-*-*): Add nds32 target.
372 * config/nds32/nds32.c: New file.
373 * config/nds32/nds32.h: New file.
374 * config/nds32/nds32.md: New file.
375 * config/nds32/constants.md: New file.
376 * config/nds32/constraints.md: New file.
377 * config/nds32/iterators.md: New file.
378 * config/nds32/nds32-doubleword.md: New file.
379 * config/nds32/nds32-intrinsic.md: New file.
380 * config/nds32/nds32_intrinsic.h: New file.
381 * config/nds32/nds32-modes.def: New file.
382 * config/nds32/nds32-multiple.md: New file.
383 * config/nds32/nds32.opt: New file.
384 * config/nds32/nds32-opts.h: New file.
385 * config/nds32/nds32-protos.h: New file.
386 * config/nds32/nds32-peephole2.md: New file.
387 * config/nds32/pipelines.md: New file.
388 * config/nds32/predicates.md: New file.
389 * config/nds32/t-mlibs: New file.
390 * common/config/nds32: New directory and files.
392 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
393 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
394 * doc/install.texi (Cross-Compiler-Specific Options): Document
395 --with-nds32-lib for nds32 target.
396 * doc/extend.texi (Function Attributes, Target Builtins): Document
397 nds32 specific attributes.
399 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
401 * lra-constraints (process_alt_operands): Use the result
402 elimination register for operand when matching constraints.
404 2013-10-31 Jakub Jelinek <jakub@redhat.com>
406 * tree-vrp.c (maybe_set_nonzero_bits): New function.
407 (remove_range_assertions): Call it.
409 * tree.c (tree_ctz): New function.
410 * tree.h (tree_ctz): New prototype.
411 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
412 first argument from tree to const_tree.
413 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
414 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
415 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
417 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
418 * tree-vect-loop.c (vect_analyze_loop_operations,
419 vect_transform_loop): Don't force scalar loop for bound just because
420 number of iterations is unknown, only do it if it is not known to be
421 a multiple of vectorization_factor.
422 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
424 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
425 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
426 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
427 (set_nonzero_bits, get_nonzero_bits): New prototypes.
428 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
429 a default def isn't partially constant.
430 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
431 is known to be partially zero.
432 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
433 to see if a default def isn't partially constant.
434 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
435 creation of a range, if VR_RANGE, try to improve nonzero_bits from
437 (set_nonzero_bits, get_nonzero_bits): New functions.
439 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
440 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
441 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
442 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
443 uses but in the condition and ASSERT_EXPR and the other successor of
444 the predecessor bb is __builtin_unreachable (), set_range_info of the
445 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
447 2013-10-31 Martin Jambor <mjambor@suse.cz>
449 PR rtl-optimization/58934
451 2013-10-30 Martin Jambor <mjambor@suse.cz>
452 PR rtl-optimization/10474
453 * ira.c (find_moveable_pseudos): Do not calculate dominance info
455 (interesting_dest_for_shprep): New function.
456 (split_live_ranges_for_shrink_wrap): Likewise.
457 (ira): Calculate dominance info and df analysis. Call
458 split_live_ranges_for_shrink_wrap.
460 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
461 Yury Gribov <y.gribov@samsung.com>
464 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
465 shadow pointer to avoid clobbering the main one.
467 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
469 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
471 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
474 * ira-color.c (update_costs_from_copies): Add new parameter. Use
475 it for calling update_costs_from_allocno.
476 (assign_hard_reg): Call restore_costs_from_copies only for
477 !retry_p. Pass new argument to update_costs_from_copies.
478 (color_pass): Pass new argument to update_costs_from_copies.
479 (ira_mark_allocation_change): Ditto.
481 2013-10-30 Sharad Singhai <singhai@google.com>
484 * opts.c (common_handle_option): Remove deprecated option
485 -ftree-vectorizer-verbose.
486 * doc/invoke.texi (Debugging Options): Ditto.
487 * opts-global.c (handle_common_deferred_options): Ditto.
488 (dump_remap_tree_vectorizer_verbose): Delete.
489 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
491 2013-10-30 DJ Delorie <dj@redhat.com>
493 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
494 that take no arguments.
496 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
499 * reload1.c (update_eliminables_and_spill): New function, broken
501 (reload): Use it. Check for frame size change after frame size
502 alignment, and call update_eliminables_and_spill first if continue-ing.
504 2013-10-30 Cong Hou <congh@google.com>
507 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
508 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
509 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
511 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
513 PR rtl-optimization/58369
514 * reload1.c (compute_reload_subreg_offset): New function.
515 (choose_reload_regs): Use it to pass endian-correct
516 offset to subreg_regno_offset.
518 2013-10-30 Tobias Burnus <burnus@net-b.de>
521 * gcc/tree-cfg.c (replace_loop_annotate): Replace warning by
524 2013-10-30 Jason Merrill <jason@redhat.com>
526 * configure.ac (loose_warn): Add -Wno-format if
527 --disable-build-format-warnings.
529 2013-10-30 David Malcolm <dmalcolm@redhat.com>
531 * cgraphunit.c (analyze_functions): Split symtab_node declarations
532 onto multiple lines to make things easier for rename_symtab.py.
534 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
535 (symtab_semantically_equivalent_p): Likewise.
537 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
540 * lra.c (check_rtl): Remove address check before LRA work.
542 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
544 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
545 POINTER_PLUS_EXPR in the defining statement.
547 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
550 * tree-pass.h (make_pass_regmove): Remove.
551 * timevar.def (TV_REGMOVE): Remove.
552 * passes.def (pass_regmove): Remove.
553 * opts.c (default_options_table): Remove entry for regmove.
554 * doc/passes.texi: Remove regmove pass description.
555 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
557 (-fdump-rtl-regmove): Ditto.
558 * common.opt (foptimize-register-move, fregmove): Ignore.
559 * Makefile.in (OBJS): Remove regmove.o.
561 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
563 (struct ira_allocno) New member allocno_prefs.
564 (ALLOCNO_PREFS): New macro.
565 (ira_prefs, ira_prefs_num): New external vars.
566 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
567 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
568 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
570 (ira_add_allocno_copy_to_list): Remove prototype.
571 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
572 (ira_pref_iterator): New type.
573 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
574 (FOR_EACH_PREF): New macro.
575 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
576 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
578 (ira_setup_alts): New function.
579 (decrease_live_ranges_number): New function.
580 (ira): Call the above function.
581 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
582 (ira_create_allocno): Initialize allocno prefs.
583 (pref_pool, pref_vec): New static vars.
584 (initiate_prefs, find_allocno_pref, ira_create_pref): New
586 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
587 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
588 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
589 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
590 (ira_add_allocno_copy_to_list): Make static. Rename to
591 add_allocno_copy_to_list.
592 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
593 swap_allocno_copy_ends_if_necessary.
594 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
595 ira_remove_allocno_prefs.
596 (ira_flattening): Ditto.
597 (ira_build): Call initiate_prefs, print_prefs.
598 (ira_destroy): Call finish_prefs.
599 * ira-color.c (struct update_cost_record): New.
600 (struct allocno_color_data): Add new member update_cost_records.
601 (update_cost_record_pool): New static var.
602 (init_update_cost_records, get_update_cost_record): New functions.
603 (free_update_cost_record_list, finish_update_cost_records): Ditto.
604 (struct update_cost_queue_elem): Add member from.
605 (initiate_cost_update): Call init_update_cost_records.
606 (finish_cost_update): Call finish_update_cost_records.
607 (queue_update_cost, get_next_update_cost): Add new param from.
608 (Update_allocno_cost, update_costs_from_allocno): New functions.
609 (update_costs_from_prefs): Ditto.
610 (update_copy_costs): Rename to update_costs_from_copies.
611 (restore_costs_from_copies): New function.
612 (update_conflict_hard_regno_costs): Don't go back.
613 (assign_hard_reg): Call restore_costs_from_copies. Add printing
615 (pop_allocnos): Add priniting more debug info.
616 (color_allocnos): Remove prefs for conflicting hard regs.
617 Call update_costs_from_prefs.
618 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
619 (get_dup_num): Rename, modify, and move to ira.c
620 (process_regs_for_copy): Add prefs.
621 (add_insn_allocno_copies): Put src as first arg of
622 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
623 * ira-costs.c (record_reg_classes): Modify and move code into
624 record_operands_costs.
625 (find_costs_and_classes): Create prefs for the hard reg of small
627 (process_bb_node_for_hard_reg_moves): Add prefs.
629 2013-10-30 Richard Biener <rguenther@suse.de>
632 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
633 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
635 (pre_and_rev_post_order_compute): ... which now wraps it.
636 * graph.c (draw_cfg_nodes_no_loops): Use
637 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
640 2013-10-30 Christian Bruel <christian.bruel@st.com>
642 * gcc/config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
643 (sh_expand_cmpstr): Handle known align and schedule improvements.
644 * gcc/config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
645 * gcc/config/sh/sh.md (cmpstrnsi): New pattern.
647 2013-10-30 Martin Jambor <mjambor@suse.cz>
649 PR rtl-optimization/10474
650 * ira.c (find_moveable_pseudos): Do not calculate dominance info
652 (interesting_dest_for_shprep): New function.
653 (split_live_ranges_for_shrink_wrap): Likewise.
654 (ira): Calculate dominance info and df analysis. Call
655 split_live_ranges_for_shrink_wrap.
657 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
660 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
662 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
664 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
665 * tree.h (POINTER_BOUNDS_P): New.
666 (BOUNDED_TYPE_P): New.
668 (pointer_bounds_type_node): New.
669 * tree.c (build_common_tree_nodes): Initialize
670 pointer_bounds_type_node.
671 * gimple.h (gimple_call_get_nobnd_arg_index): New.
672 (gimple_call_num_nobnd_args): New.
673 (gimple_call_nobnd_arg): New.
674 (gimple_return_retbnd): New.
675 (gimple_return_set_retbnd): New
676 * gimple.c (gimple_build_return): Increase number of ops
677 for return statement.
678 (gimple_call_get_nobnd_arg_index): New.
679 * gimple-pretty-print.c (dump_gimple_return): Print second op.
681 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
683 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
684 with "chkp ctor" and "bnd_legacy" attributes.
685 * gimplify.c (gimplify_init_constructor): Avoid infinite
686 loop during gimplification of bounds initializer.
688 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
690 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
691 (handle_bnd_legacy): New.
692 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
693 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
696 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
698 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
699 (BT_FN_PTR_CONST_PTR): New.
700 (BT_FN_CONST_PTR_CONST_PTR): New.
701 (BT_FN_PTR_CONST_PTR_SIZE): New.
702 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
703 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
704 (BT_FN_VOID_CONST_PTR_SIZE): New.
705 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
706 * chkp-builtins.def: New.
707 * builtins.def: include chkp-builtins.def.
708 (DEF_CHKP_BUILTIN): New.
709 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
710 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
711 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
712 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
713 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
714 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
715 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
716 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
717 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
718 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
719 * common.opt (fcheck-pointer-bounds): New.
720 * toplev.c (process_options): Check Pointer Bounds Checker is
722 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
724 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
726 * target.def (builtin_chkp_function): New.
727 (chkp_bound_type): New.
728 (chkp_bound_mode): New.
729 (fn_abi_va_list_bounds_size): New.
730 (load_bounds_for_arg): New.
731 (store_bounds_for_arg): New.
732 * targhooks.h (default_load_bounds_for_arg): New.
733 (default_store_bounds_for_arg): New.
734 (default_fn_abi_va_list_bounds_size): New.
735 (default_chkp_bound_type): New.
736 (default_chkp_bound_mode): New.
737 (default_builtin_chkp_function): New.
738 * targhooks.c (default_load_bounds_for_arg): New.
739 (default_store_bounds_for_arg): New.
740 (default_fn_abi_va_list_bounds_size): New.
741 (default_chkp_bound_type): New.
742 (default_chkp_bound_mode); New.
743 (default_builtin_chkp_function): New.
744 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
745 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
746 (TARGET_STORE_BOUNDS_FOR_ARG): New.
747 (TARGET_BUILTIN_CHKP_FUNCTION): New.
748 (TARGET_CHKP_BOUND_TYPE): New.
749 (TARGET_CHKP_BOUND_MODE): New.
750 * doc/tm.texi: Regenerated.
751 * langhooks.h (lang_hooks): Add chkp_supported field.
752 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
753 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
755 2013-10-29 Andrew Pinski <apinski@cavium.com>
757 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
758 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
759 fails, combining the branches anyways.
760 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
761 increases the number of combinings.
762 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
764 2013-10-29 Mike Stump <mikestump@comcast.net>
766 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
767 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
768 (make_vector_mode): Increase namebuf to 16.
769 (emit_insn_modes_h): When processing BImode, don't
770 also match partial int modes.
771 (emit_class_narrowest_mode): Likewise.
773 * config/bfin/bfin-modes.def: Add precision to PDI.
774 * config/m32c/m32c-modes.def: Add precision to PSI.
775 * config/msp430/msp430-modes.def: Add precision to PSI.
776 * config/rs6000/rs6000-modes.def: Add precision to PTI.
777 * config/sh/sh-modes.def: Add precision to PSI and PDI.
779 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
782 * config/sh/sh.md (*addc): Rename existing variations to ...
783 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
784 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
785 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
786 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
788 2013-10-29 Teresa Johnson <tejohnson@google.com>
791 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
792 insanities when updating probabilities.
794 2013-10-29 David Malcolm <dmalcolm@redhat.com>
796 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
797 prettyprinter for cgraph_node to reflect the conversion of the
798 symtable types to a C++ class hierarchy: it now *is* a
799 symtab_node_base, rather than having one (named "symbol").
801 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
803 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
804 __cilkrts_pop_frame. If matched, then return true for built-in
806 (expand_builtin): Added BUILT_IN_CILK_DETACH and
807 BUILT_IN_CILK_POP_FRAME case.
808 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
809 (lhs_cilk_detect_spawn): Likewise.
810 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
811 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
812 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
813 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
814 (LANG_HOOKS_CILKPLUS): Likewise.
815 * tree.h (CILK_SPAWN_FN): Likewise.
816 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
817 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
818 (OBJS): Added cilk-common.o.
819 (BUILTINS_DEF): Added cilk-builtins.def.
820 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
821 (lhd_cilk_detect_spawn): Likewise.
822 * langhooks.h (lang_hooks_for_cilkplus): New struct.
823 (struct lang_hooks): Added new field called "cilkplus."
824 * cilk-common.c: New file.
826 * cilk-builtins.def: Likewise.
827 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
828 "__cilk" macro and set it to 200.
829 * function.h (struct function::cilk_frame_decl): New field.
830 (struct function::is_cilk_function): Likewise.
831 (struct function::calls_cilk_spawn): Likewise.
832 * gimplify.c (gimplify_call_expr): Added a check if the function call
833 being gimplified is a spawn detach point. If so, then add pop_frame
834 and detach function calls.
835 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
836 for gimplifying _Cilk_spawn and _Cilk_sync statements.
837 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
838 function. If so, added a _Cilk_sync and gimplified it.
839 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
840 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
841 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
843 (can_inline_edge_p): Prevent inling of spawnee function.
844 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
845 for functions that use Cilk keywords.
846 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
847 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
848 CILK_SYNC_STMT cases.
849 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
851 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
852 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
853 * passes.texi (Cilk Keywords): New section that describes the compiler
854 code changes for handling Cilk Keywords.
856 2013-10-29 David Malcolm <dmalcolm@redhat.com>
858 Patch autogenerated by refactor_symtab.py from
859 https://github.com/davidmalcolm/gcc-refactoring-scripts
860 revision 58bb219cc090b2f4516a9297d868c245495ee622
862 * asan.c (asan_finish_file): Update for conversion of symtab types to
863 a true class hierarchy.
864 * cfgexpand.c (estimated_stack_frame_size): Likewise.
865 * cgraph.c (cgraph_get_body): Likewise.
866 (cgraph_get_create_real_symbol_node): Likewise.
867 (verify_cgraph_node): Likewise.
868 (verify_edge_corresponds_to_fndecl): Likewise.
869 (verify_edge_count_and_frequency): Likewise.
870 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
871 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
872 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
873 (cgraph_node_cannot_return): Likewise.
874 (cgraph_set_pure_flag_1): Likewise.
875 (cgraph_set_const_flag_1): Likewise.
876 (cgraph_set_nothrow_flag_1): Likewise.
877 (cgraph_make_node_local_1): Likewise.
878 (cgraph_for_node_and_aliases): Likewise.
879 (cgraph_for_node_thunks_and_aliases): Likewise.
880 (cgraph_node_can_be_local_p): Likewise.
881 (cgraph_node_cannot_be_local_p_1): Likewise.
882 (cgraph_function_body_availability): Likewise.
883 (dump_cgraph_node): Likewise.
884 (cgraph_rtl_info): Likewise.
885 (cgraph_mark_address_taken_node): Likewise.
886 (cgraph_remove_node): Likewise.
887 (cgraph_release_function_body): Likewise.
888 (cgraph_update_edges_for_call_stmt_node): Likewise.
889 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
890 (cgraph_make_edge_direct): Likewise.
891 (cgraph_resolve_speculation): Likewise.
892 (cgraph_speculative_call_info): Likewise.
893 (cgraph_turn_edge_to_speculative): Likewise.
894 (cgraph_create_edge_1): Likewise.
895 (cgraph_set_call_stmt): Likewise.
896 (cgraph_node_for_asm): Likewise.
897 (cgraph_add_thunk): Likewise.
898 (cgraph_same_body_alias): Likewise.
899 (cgraph_create_function_alias): Likewise.
900 (cgraph_create_node): Likewise.
901 (cgraph_create_empty_node): Likewise.
902 (record_function_versions): Likewise.
903 (used_from_object_file_p): Likewise.
904 * cgraph.h (symtab_can_be_discarded): Likewise.
905 (symtab_real_symbol_p): Likewise.
906 (cgraph_mark_force_output_node): Likewise.
907 (cgraph_edge_recursive_p): Likewise.
908 (symtab_alias_target): Likewise.
909 (varpool_all_refs_explicit_p): Likewise.
910 (varpool_can_remove_if_no_refs): Likewise.
911 (cgraph_only_called_directly_or_aliased_p): Likewise.
912 (cgraph_next_function_with_gimple_body): Likewise.
913 (cgraph_first_function_with_gimple_body): Likewise.
914 (cgraph_function_with_gimple_body_p): Likewise.
915 (cgraph_next_function): Likewise.
916 (cgraph_first_function): Likewise.
917 (cgraph_next_defined_function): Likewise.
918 (cgraph_first_defined_function): Likewise.
919 (varpool_next_defined_variable): Likewise.
920 (varpool_first_defined_variable): Likewise.
921 (varpool_next_static_initializer): Likewise.
922 (varpool_first_static_initializer): Likewise.
923 (varpool_next_variable): Likewise.
924 (varpool_first_variable): Likewise.
925 (varpool_node_name): Likewise.
928 (is_a_helper <varpool_node>::test): Likewise.
929 (is_a_helper <cgraph_node>::test): Likewise.
930 (varpool_variable_node): Likewise.
931 (cgraph_function_or_thunk_node): Likewise.
932 (varpool_alias_target): Likewise.
933 (cgraph_alias_target): Likewise.
934 (cgraph_node_name): Likewise.
935 (varpool_node_asm_name): Likewise.
936 (cgraph_node_asm_name): Likewise.
937 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
938 (cgraph_rebuild_references): Likewise.
939 (rebuild_cgraph_edges): Likewise.
940 (record_eh_tables): Likewise.
941 (build_cgraph_edges): Likewise.
942 (mark_store): Likewise.
943 (mark_load): Likewise.
944 (mark_address): Likewise.
945 (record_type_list): Likewise.
946 (record_reference): Likewise.
947 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
948 (cgraph_materialize_clone): Likewise.
949 (cgraph_function_versioning): Likewise.
950 (cgraph_copy_node_for_versioning): Likewise.
951 (update_call_expr): Likewise.
952 (cgraph_find_replacement_node): Likewise.
953 (cgraph_create_virtual_clone): Likewise.
954 (cgraph_clone_node): Likewise.
955 * cgraphunit.c (compile): Likewise.
956 (output_weakrefs): Likewise.
957 (output_in_order): Likewise.
958 (expand_function): Likewise.
959 (assemble_thunks_and_aliases): Likewise.
960 (expand_thunk): Likewise.
961 (mark_functions_to_output): Likewise.
962 (handle_alias_pairs): Likewise.
963 (analyze_functions): Likewise.
964 (walk_polymorphic_call_targets): Likewise.
965 (varpool_finalize_decl): Likewise.
966 (process_function_and_variable_attributes): Likewise.
967 (cgraph_process_same_body_aliases): Likewise.
968 (analyze_function): Likewise.
969 (cgraph_add_new_function): Likewise.
970 (cgraph_finalize_function): Likewise.
971 (referred_to_p): Likewise.
972 (cgraph_reset_node): Likewise.
973 (cgraph_process_new_functions): Likewise.
974 (enqueue_node): Likewise.
975 (decide_is_symbol_needed): Likewise.
976 * coverage.c (coverage_compute_profile_id): Likewise.
977 * dbxout.c (dbxout_expand_expr): Likewise.
978 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
979 (reference_to_unused): Likewise.
980 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
981 * gimplify.c (unvisit_body): Likewise.
982 (unshare_body): Likewise.
983 * ipa-cp.c (ipcp_generate_summary): Likewise.
984 (ipcp_decision_stage): Likewise.
985 (identify_dead_nodes): Likewise.
986 (decide_whether_version_node): Likewise.
987 (decide_about_value): Likewise.
988 (perhaps_add_new_callers): Likewise.
989 (create_specialized_node): Likewise.
990 (update_profiling_info): Likewise.
991 (ipcp_propagate_stage): Likewise.
992 (estimate_local_effects): Likewise.
993 (good_cloning_opportunity_p): Likewise.
994 (devirtualization_time_bonus): Likewise.
995 (propagate_constants_accross_call): Likewise.
996 (initialize_node_lattices): Likewise.
997 (ipcp_cloning_candidate_p): Likewise.
998 (determine_versionability): Likewise.
999 (print_all_lattices): Likewise.
1000 (print_lattice): Likewise.
1001 (ipcp_discover_new_direct_edges): Likewise.
1002 * ipa-devirt.c (ipa_devirt): Likewise.
1003 (likely_target_p): Likewise.
1004 (update_type_inheritance_graph): Likewise.
1005 (possible_polymorphic_call_target_p): Likewise.
1006 (dump_possible_polymorphic_call_targets): Likewise.
1007 (devirt_variable_node_removal_hook): Likewise.
1008 (record_binfo): Likewise.
1009 (maybe_record_node): Likewise.
1010 (build_type_inheritance_graph): Likewise.
1011 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1012 (inline_generate_summary): Likewise.
1013 (inline_analyze_function): Likewise.
1014 (do_estimate_growth): Likewise.
1015 (simple_edge_hints): Likewise.
1016 (estimate_node_size_and_time): Likewise.
1017 (estimate_edge_devirt_benefit): Likewise.
1018 (compute_inline_parameters): Likewise.
1019 (estimate_function_body_sizes): Likewise.
1020 (compute_bb_predicates): Likewise.
1021 (initialize_inline_failed): Likewise.
1022 (dump_inline_summary): Likewise.
1023 (dump_inline_edge_summary): Likewise.
1024 * ipa-inline-transform.c (inline_transform): Likewise.
1025 (preserve_function_body_p): Likewise.
1026 (save_inline_function_body): Likewise.
1027 (inline_call): Likewise.
1028 (clone_inlined_nodes): Likewise.
1029 (can_remove_node_now_p): Likewise.
1030 (can_remove_node_now_p_1): Likewise.
1031 * ipa-inline.c (early_inliner): Likewise.
1032 (early_inline_small_functions): Likewise.
1033 (inline_always_inline_functions): Likewise.
1034 (ipa_inline): Likewise.
1035 (flatten_function): Likewise.
1036 (inline_small_functions): Likewise.
1037 (speculation_useful_p): Likewise.
1038 (recursive_inlining): Likewise.
1039 (update_caller_keys): Likewise.
1040 (reset_edge_caches): Likewise.
1041 (update_edge_key): Likewise.
1042 (edge_badness): Likewise.
1043 (relative_time_benefit): Likewise.
1044 (want_inline_self_recursive_call_p): Likewise.
1045 (want_inline_small_function_p): Likewise.
1046 (want_early_inline_function_p): Likewise.
1047 (num_calls): Likewise.
1048 (can_early_inline_edge_p): Likewise.
1049 (can_inline_edge_p): Likewise.
1050 (report_inline_failed_reason): Likewise.
1051 * ipa-profile.c (ipa_profile): Likewise.
1052 (ipa_propagate_frequency): Likewise.
1053 (ipa_propagate_frequency_1): Likewise.
1054 (ipa_profile_generate_summary): Likewise.
1055 * ipa-prop.c (ipcp_transform_function): Likewise.
1056 (read_replacements_section): Likewise.
1057 (ipa_prop_read_section): Likewise.
1058 (ipa_modify_call_arguments): Likewise.
1059 (ipa_print_node_params): Likewise.
1060 (propagate_controlled_uses): Likewise.
1061 (update_indirect_edges_after_inlining): Likewise.
1062 (remove_described_reference): Likewise.
1063 (ipa_make_edge_direct_to_target): Likewise.
1064 (ipa_analyze_node): Likewise.
1065 (ipa_analyze_params_uses): Likewise.
1066 (ipa_compute_jump_functions): Likewise.
1067 (ipa_get_callee_param_type): Likewise.
1068 (ipa_print_node_jump_functions): Likewise.
1069 (ipa_initialize_node_params): Likewise.
1070 (ipa_populate_param_decls): Likewise.
1071 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
1072 (write_agg_replacement_chain): Likewise.
1073 (ipa_write_node_info): Likewise.
1074 (ipa_edge_duplication_hook): Likewise.
1075 (try_decrement_rdesc_refcount): Likewise.
1076 * ipa-pure-const.c (propagate_nothrow): Likewise.
1077 (propagate_pure_const): Likewise.
1078 (pure_const_read_summary): Likewise.
1079 (pure_const_write_summary): Likewise.
1080 (analyze_function): Likewise.
1081 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
1082 (ipa_ref_referring_ref_list): Likewise.
1083 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
1084 (ipa_remove_stmt_references): Likewise.
1085 (ipa_find_reference): Likewise.
1086 (ipa_dump_referring): Likewise.
1087 (ipa_dump_references): Likewise.
1088 (ipa_record_reference): Likewise.
1089 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
1090 (ipa_reference_write_optimization_summary): Likewise.
1091 (write_node_summary_p): Likewise.
1092 (propagate): Likewise.
1093 (read_write_all_from_decl): Likewise.
1094 (generate_summary): Likewise.
1095 (analyze_function): Likewise.
1096 (propagate_bits): Likewise.
1097 (ipa_reference_get_not_written_global): Likewise.
1098 (ipa_reference_get_not_read_global): Likewise.
1099 * ipa-split.c (execute_split_functions): Likewise.
1100 (split_function): Likewise.
1101 * ipa-utils.c (ipa_merge_profiles): Likewise.
1102 (dump_cgraph_node_set): Likewise.
1103 (ipa_reverse_postorder): Likewise.
1104 (ipa_edge_within_scc): Likewise.
1105 (ipa_get_nodes_in_cycle): Likewise.
1106 (ipa_free_postorder_info): Likewise.
1107 (ipa_reduced_postorder): Likewise.
1108 (searchc): Likewise.
1109 (recursive_call_p): Likewise.
1110 * ipa.c (ipa_cdtor_merge): Likewise.
1111 (record_cdtor_fn): Likewise.
1112 (function_and_variable_visibility): Likewise.
1113 (varpool_externally_visible_p): Likewise.
1114 (cgraph_externally_visible_p): Likewise.
1115 (comdat_can_be_unshared_p): Likewise.
1116 (comdat_can_be_unshared_p_1): Likewise.
1117 (address_taken_from_non_vtable_p): Likewise.
1118 (ipa_discover_readonly_nonaddressable_vars): Likewise.
1119 (symtab_remove_unreachable_nodes): Likewise.
1120 (walk_polymorphic_call_targets): Likewise.
1121 (process_references): Likewise.
1122 (enqueue_node): Likewise.
1123 (has_addr_references_p): Likewise.
1124 (cgraph_non_local_node_p_1): Likewise.
1125 * is-a.h (varpool_analyze_node): Likewise.
1126 * lto-cgraph.c (input_symtab): Likewise.
1127 (merge_profile_summaries): Likewise.
1128 (input_cgraph_1): Likewise.
1129 (input_edge): Likewise.
1130 (input_varpool_node): Likewise.
1131 (input_node): Likewise.
1132 (input_overwrite_node): Likewise.
1133 (compute_ltrans_boundary): Likewise.
1134 (output_refs): Likewise.
1135 (lto_output_varpool_node): Likewise.
1136 (lto_output_node): Likewise.
1137 (reachable_from_other_partition_p): Likewise.
1138 (referenced_from_other_partition_p): Likewise.
1139 (lto_output_edge): Likewise.
1140 (output_node_opt_summary): Likewise.
1141 (add_node_to): Likewise.
1142 (reachable_from_this_partition_p): Likewise.
1143 (lto_set_symtab_encoder_in_partition): Likewise.
1144 (lto_symtab_encoder_in_partition_p): Likewise.
1145 (lto_set_symtab_encoder_encode_initializer): Likewise.
1146 (lto_symtab_encoder_encode_initializer_p): Likewise.
1147 (lto_set_symtab_encoder_encode_body): Likewise.
1148 (lto_symtab_encoder_encode_body_p): Likewise.
1149 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1151 * lto-streamer-in.c (lto_read_body): Likewise.
1152 (fixup_call_stmt_edges): Likewise.
1153 (fixup_call_stmt_edges_1): Likewise.
1154 * lto-streamer-out.c (produce_symtab): Likewise.
1155 (output_symbol_p): Likewise.
1156 (write_symbol): Likewise.
1157 (lto_output): Likewise.
1158 (copy_function): Likewise.
1159 (output_function): Likewise.
1160 * passes.c (function_called_by_processed_nodes_p): Likewise.
1161 (ipa_write_optimization_summaries): Likewise.
1162 (ipa_write_summaries): Likewise.
1163 (do_per_function_toporder): Likewise.
1164 (do_per_function): Likewise.
1165 (dump_passes): Likewise.
1166 * symtab.c (symtab_semantically_equivalent_p): Likewise.
1167 (symtab_nonoverwritable_alias): Likewise.
1168 (symtab_nonoverwritable_alias_1): Likewise.
1169 (symtab_for_node_and_aliases): Likewise.
1170 (symtab_resolve_alias): Likewise.
1171 (fixup_same_cpp_alias_visibility): Likewise.
1172 (symtab_alias_ultimate_target): Likewise.
1173 (symtab_used_from_object_file_p): Likewise.
1174 (verify_symtab_base): Likewise.
1175 (dump_symtab_base): Likewise.
1176 (symtab_node_name): Likewise.
1177 (symtab_node_asm_name): Likewise.
1178 (symtab_dissolve_same_comdat_group_list): Likewise.
1179 (symtab_add_to_same_comdat_group): Likewise.
1180 (symtab_unregister_node): Likewise.
1181 (symtab_insert_node_to_hashtable): Likewise.
1182 (symtab_register_node): Likewise.
1183 (unlink_from_assembler_name_hash): Likewise.
1184 (insert_to_assembler_name_hash): Likewise.
1185 (eq_assembler_name): Likewise.
1186 (hash_node_by_assembler_name): Likewise.
1187 (eq_node): Likewise.
1188 (hash_node): Likewise.
1189 * toplev.c (wrapup_global_declaration_2): Likewise.
1190 * trans-mem.c (ipa_tm_execute): Likewise.
1191 (ipa_tm_transform_clone): Likewise.
1192 (ipa_tm_transform_transaction): Likewise.
1193 (ipa_tm_transform_calls_redirect): Likewise.
1194 (ipa_tm_insert_gettmclone_call): Likewise.
1195 (ipa_tm_insert_irr_call): Likewise.
1196 (ipa_tm_create_version): Likewise.
1197 (ipa_tm_create_version_alias): Likewise.
1198 (ipa_tm_mark_forced_by_abi_node): Likewise.
1199 (ipa_tm_mark_force_output_node): Likewise.
1200 (ipa_tm_diagnose_tm_safe): Likewise.
1201 (ipa_tm_mayenterirr_function): Likewise.
1202 (ipa_tm_scan_irr_function): Likewise.
1203 (ipa_tm_note_irrevocable): Likewise.
1204 (ipa_tm_scan_calls_clone): Likewise.
1205 (get_cg_data): Likewise.
1206 * tree-eh.c (tree_could_trap_p): Likewise.
1207 * tree-emutls.c (ipa_lower_emutls): Likewise.
1208 (create_emultls_var): Likewise.
1209 (lower_emutls_function_body): Likewise.
1210 (gen_emutls_addr): Likewise.
1211 (emutls_decl): Likewise.
1212 (new_emutls_decl): Likewise.
1213 * tree-inline.c (tree_function_versioning): Likewise.
1214 (optimize_inline_calls): Likewise.
1215 (expand_call_inline): Likewise.
1216 (estimate_num_insns): Likewise.
1217 (copy_bb): Likewise.
1218 (delete_unreachable_blocks_update_callgraph): Likewise.
1219 * tree-nested.c (gimplify_all_functions): Likewise.
1220 (create_nesting_tree): Likewise.
1221 (check_for_nested_with_variably_modified): Likewise.
1222 * tree-pretty-print.c (dump_function_header): Likewise.
1223 * tree-profile.c (tree_profiling): Likewise.
1224 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
1225 (modify_function): Likewise.
1226 (convert_callers): Likewise.
1227 (convert_callers_for_node): Likewise.
1228 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
1229 (associate_varinfo_to_alias): Likewise.
1230 (create_variable_info_for): Likewise.
1231 (get_constraint_for_ssa_var): Likewise.
1232 * tree-vectorizer.c (increase_alignment): Likewise.
1233 * tree.c (find_decls_types_in_var): Likewise.
1234 (find_decls_types_in_node): Likewise.
1235 (free_lang_data_in_decl): Likewise.
1236 * value-prof.c (gimple_ic_transform): Likewise.
1237 (gimple_ic): Likewise.
1238 (check_ic_target): Likewise.
1239 (init_node_map): Likewise.
1240 * varasm.c (decl_binds_to_current_def_p): Likewise.
1241 (default_binds_local_p_1): Likewise.
1242 (dump_tm_clone_pairs): Likewise.
1243 (assemble_alias): Likewise.
1244 (find_decl): Likewise.
1245 (mark_decl_referenced): Likewise.
1246 * varpool.c (varpool_for_node_and_aliases): Likewise.
1247 (varpool_extra_name_alias): Likewise.
1248 (varpool_create_variable_alias): Likewise.
1249 (add_new_static_var): Likewise.
1250 (varpool_finalize_named_section_flags): Likewise.
1251 (varpool_remove_unreferenced_decls): Likewise.
1252 (enqueue_node): Likewise.
1253 (varpool_assemble_decl): Likewise.
1254 (assemble_aliases): Likewise.
1255 (varpool_analyze_node): Likewise.
1256 (cgraph_variable_initializer_availability): Likewise.
1257 (varpool_add_new_variable): Likewise.
1258 (ctor_for_folding): Likewise.
1259 (dump_varpool_node): Likewise.
1260 (varpool_remove_initializer): Likewise.
1261 (varpool_remove_node): Likewise.
1262 (varpool_node_for_decl): Likewise.
1263 (varpool_create_empty_node): Likewise.
1264 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
1265 (ix86_get_function_versions_dispatcher): Likewise.
1267 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1269 * cgraph.h (symtab_node_base): Convert to a class;
1270 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
1271 chain_next/prev from symtab_node_def.
1272 (cgraph_node): Inherit from symtab_node; add GTY option
1273 tag ("SYMTAB_FUNCTION").
1274 (varpool_node): Inherit from symtab_node; add GTY option
1275 tag ("SYMTAB_VARIABLE").
1276 (symtab_node_def): Remove.
1277 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
1278 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
1279 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
1280 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
1282 * ipa-ref.h (symtab_node_def): Drop.
1283 (symtab_node): Change underlying type from symtab_node_def to
1285 (const_symtab_node): Likwise.
1287 * is-a.h: Update examples in comment.
1289 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
1290 (assembler_name_hash): Likewise.
1292 2013-10-29 Martin Liska <marxin.liska@gmail.com>
1294 * doc/tree-ssa.texi (gimple_phi_result): Document.
1295 (gimple_phi_num_args, gimple_phi_arg): Likewise.
1296 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
1297 (PHI_RESULT, PHI_NUM_ARGS): Remove.
1298 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
1300 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1302 * expr.h: Revert change and include tree-core.h.
1303 * rtl.h: Revert change and don't include tree-core.h.
1305 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1307 * config/darwin.c: Include gimple.h.
1308 * config/i386/winnt.c: Likewise.
1310 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
1312 PR tree-optimization/19831
1313 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
1315 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1317 * tree-outof-ssa.h: Remove include files.
1318 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
1320 * tree-ssa-coalesce.c: Likewise.
1321 * cfgexpand.c: Likewise.
1322 * tree-ssa-ter.c: Likewise.
1323 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
1324 * lto-streamer.h: Likewise.
1325 * cgraphbuild.c: Add gimple.h to include list.
1326 * data-streamer-in.c: Likewise.
1327 * ipa-cp.c: Likewise.
1328 * tree-streamer.c: Likewise.
1329 * lto-compress.c: Likewise.
1330 * ipa-reference.c: Likewise.
1331 * data-streamer-out.c: Likewise.
1332 * lto-cgraph.c: Likewise.
1333 * cgraphclones.c: Likewise.
1334 * ipa-utils.c: Likewise.
1335 * data-streamer.c: Likewise.
1336 * ipa-split.c: Likewise.
1337 * lto-section-in.c: Likewise.
1338 * tree-streamer-out.c: Likewise.
1339 * ipa-prop.c: Likewise.
1340 * tree-streamer-in.c: Likewise.
1341 * symtab.c: Likewise.
1342 * opts-global.c: Likewise.
1343 * lto-opts.c: Likewise.
1344 * lto-section-out.c: Likewise.
1345 * lto-streamer.c: Likewise.
1346 * rtl.h: Add tree-core.h to include list.
1347 * expr.h: Remove tree-core.h from include list.
1348 * gimple.h: Likewise.
1349 * ipa-utils.h: Likewise.
1350 * streamer-hooks.h: Likewise.
1351 * streamer-hooks.c: Include input.h.
1353 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1355 * config/arm/arm.c (cortexa7_extra_costs): New table.
1356 (arm_cortex_a7_tune): New.
1357 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
1359 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
1361 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
1363 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1365 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
1366 to use autogenerated markers for a class-hierarchy, every class
1367 must have a GTY marker.
1368 * gengtype.h (struct type): Add linked list of subclasses to
1369 the "s" member of the union.
1370 (add_subclass): New decl.
1371 * gengtype-state.c (read_state_struct_type): Set up subclass
1373 * gengtype.c (get_ultimate_base_class): New.
1374 (add_subclass): New.
1375 (new_structure): Set up subclass linked list.
1376 (set_gc_used_type): Propagate usage information to subclasses.
1377 (output_mangled_typename): Use get_ultimate_base_class.
1378 (walk_subclasses): Use the subclass linked list, avoiding an
1379 O(N^2) when writing out all types.
1380 (walk_type): Issue an error if the base class is missing a tag,
1381 rather than generating bogus C code. Add a gcc_unreachable
1382 default case, in case people omit tags from concrete subclasses,
1383 or get the values wrong.
1384 (write_func_for_structure): Issue an error for subclasses for
1385 which the base doesn't have a "desc", since otherwise the
1386 autogenerated routines for the base would silently fail to visit
1387 any subclass fields.
1388 (write_root): Use get_ultimate_base_class, tweaking constness of
1389 tp to match that function's signature.
1391 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1393 * doc/gty.texi (GTY Options): Add note about inheritance to
1394 description of desc and tag.
1395 (Inheritance and GTY): New.
1397 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1399 * gengtype-parse.c (opts_have): Drop "static" so that
1400 we can use this from gengtype.c.
1401 * gengtype.c (set_gc_used_type): Mark any base class as used;
1402 update field traversal to visit inherited fields.
1403 (output_mangled_typename): Convert references to classes within
1404 an inheritance hierarchy to reference the ultimate base class,
1405 since only it will have gt_ functions.
1406 (get_string_option): New.
1407 (walk_subclasses): New.
1408 (walk_type): Treat GTY structs that have a "desc" as being the
1409 root of an inheritance hierarchy. Generate a switch on it
1410 within the marking function which walks all subclasses, adding
1411 cases for them via walk_subclasses. For subclasses, visit all
1412 fields of the type (including inherited ones).
1413 (write_func_for_structure): Don't write fns for subclasses, only
1414 for the ultimate base class within an inheritance hierarchy.
1415 Subclasses-marking will be handled by the base class marking functions.
1416 (write_types): Likewise.
1417 (write_local_func_for_structure): Likewise.
1418 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
1419 a "tag" option (and are thus concrete subclasses).
1420 (write_root): Use the marker function for the ultimate base class.
1421 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
1422 (opts_have): Add declaration.
1424 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
1426 * lra-spills.c (lra_final_code_change): Remove useless move insns
1427 originated from moves of pseudos.
1429 2013-10-28 Jeff Law <law@redhat.com>
1431 * lower-subreg.c (resolve_simple_move): Fix comment typo.
1433 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
1435 * df-scan.c (df_collection_rec): Adjust.
1436 (copy_defs): New constant.
1437 (copy_uses): Likewise.
1438 (copy_eq_uses): Likewise.
1439 (copy_mw): Likewise.
1440 (copy_all): Likewise.
1441 (df_insn_rescan): Adjust.
1442 (df_notes_rescan): Likewise.
1443 (df_swap_refs): Likewise.
1444 (df_sort_and_compress_refs): Likewise.
1445 (df_sort_and_compress_mws): Likewise.
1446 (df_install_refs): Likewise.
1447 (df_install_mws): Likewise.
1448 (df_refs_add_to_chains): Add flags parameter controlling which vectors
1450 (df_bb_refs_record): Adjust.
1451 (df_record_entry_block_defs): Likewise.
1452 (df_record_exit_block_defs): Likewise.
1453 (df_refs_verify): Likewise.
1454 (df_mws_verify): Likewise.
1455 (df_insn_refs_verify): Likewise.
1456 (df_bb_verify): Likewise.
1457 * ipa-pure-const.c (finish_state): Remove.
1458 (propagate): Adjust.
1459 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
1460 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
1461 var-tracking.c: Adjust.
1462 * vec.c (stack_vecs): Remove.
1463 (register_stack_vec): Likewise.
1464 (stack_vec_register_index): Likewise.
1465 (unregister_stack_vec): Likewise.
1466 * vec.h (struct va_stack): Remove.
1467 (struct vec<T, A, vl_ptr>): Specialize as
1468 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
1469 allocation strategy compatable with the vl_ptr layout.
1470 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
1471 specialization anyway.
1472 (class stack_vec): New class.
1473 (vec_stack_alloc): Remove.
1474 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
1476 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1477 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1478 Sergey Lega <sergey.s.lega@intel.com>
1479 Anna Tikhonova <anna.tikhonova@intel.com>
1480 Ilya Tocar <ilya.tocar@intel.com>
1481 Andrey Turetskiy <andrey.turetskiy@intel.com>
1482 Ilya Verbin <ilya.verbin@intel.com>
1483 Kirill Yukhin <kirill.yukhin@intel.com>
1484 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1486 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
1487 (*prefetch_avx512pf_<mode>): New.
1488 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
1489 (avx512f_maskcmp<mode>3): Ditto.
1490 (vashrv16si3): Ditto.
1492 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1493 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1494 Sergey Lega <sergey.s.lega@intel.com>
1495 Anna Tikhonova <anna.tikhonova@intel.com>
1496 Ilya Tocar <ilya.tocar@intel.com>
1497 Andrey Turetskiy <andrey.turetskiy@intel.com>
1498 Ilya Verbin <ilya.verbin@intel.com>
1499 Kirill Yukhin <kirill.yukhin@intel.com>
1500 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1502 * config/i386/i386.md (any_truncate): New.
1503 (trunsuffix): Ditto.
1504 * config/i386/predicates.md (const_8_to_9_operand): New.
1505 (const_10_to_11_operand): Ditto.
1506 (const_12_to_13_operand): Ditto.
1507 (const_14_to_15_operand): Ditto.
1508 (const_16_to_19_operand): Ditto.
1509 (const_20_to_23_operand): Ditto.
1510 (const_24_to_27_operand): Ditto.
1511 (const_28_to_31_operand): Ditto.
1512 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
1513 (cvtusi2<ssescalarmodesuffix>32): New.
1514 (cvtusi2<ssescalarmodesuffix>64): Ditto.
1515 (ufloatv16siv16sf2): Ditto.
1516 (avx512f_fix_notruncv16sfv16si): Ditto.
1517 (avx512f_ufix_notruncv16sfv16si): Ditto.
1518 (avx512f_vcvtss2usi): Ditto.
1519 (avx512f_vcvtss2usiq): Ditto.
1520 (avx512f_vcvttss2usi): Ditto.
1521 (avx512f_vcvttss2usiq): Ditto.
1522 (avx512f_vcvtsd2usi): Ditto.
1523 (avx512f_vcvtsd2usiq): Ditto.
1524 (avx512f_vcvttsd2usi): Ditto.
1525 (avx512f_vcvttsd2usiq): Ditto.
1526 (ufloatv8siv8df): Ditto.
1527 (avx512f_cvtdq2pd512_2): Ditto.
1528 (avx512f_cvtpd2dq512): Ditto.
1529 (avx512f_ufix_notruncv8dfv8si): Ditto.
1530 (avx512f_cvtpd2ps512): Ditto.
1531 (vec_unpacks_lo_v16sf): Ditto.
1532 (vec_unpacks_hi_v16sf): Ditto.
1533 (vec_unpacks_float_hi_v16si): Ditto.
1534 (vec_unpacks_float_lo_v16si): Ditto.
1535 (avx512f_unpckhps512): Ditto.
1536 (avx512f_unpcklps512): Ditto.
1537 (avx512f_movshdup512): Ditto.
1538 (avx512f_movsldup512): Ditto.
1539 (vec_extract_lo_v32hi): Ditto.
1540 (vec_extract_hi_v32hi): Ditto.
1541 (vec_extract_lo_v64qi): Ditto.
1542 (vec_extract_hi_v64qi): Ditto.
1543 (avx512f_unpckhpd512): Ditto.
1544 (avx512f_movddup512): Ditto.
1545 (avx512f_unpcklpd512): Ditto.
1546 (*avx512f_unpcklpd512): Ditto.
1547 (avx512f_shufps512_1): Ditto.
1548 (avx512f_shufpd512_1): Ditto.
1549 (avx512f_interleave_highv8di): Ditto.
1550 (avx512f_interleave_lowv8di): Ditto.
1551 (PMOV_DST_MODE): Ditto.
1552 (pmov_src_mode): Ditto.
1553 (pmov_src_lower): Ditto.
1555 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
1556 (*avx512f_<code>v8div16qi2): Ditto.
1557 (*avx512f_<code>v8div16qi2_store): Ditto.
1558 (vec_widen_umult_even_v16si): Ditto.
1559 (*vec_widen_umult_even_v16si): Ditto.
1560 (vec_widen_smult_even_v16si): Ditto.
1561 (*vec_widen_smult_even_v16si): Ditto.
1562 (avx512f_interleave_highv16si): Ditto.
1563 (avx512f_interleave_lowv16si): Ditto.
1564 (avx512f_<code>v16qiv16si2): Ditto.
1565 (avx512f_<code>v16hiv16si2): Ditto.
1566 (avx512f_<code>v8qiv8di2): Ditto.
1567 (avx512f_<code>v8hiv8di2): Ditto.
1568 (avx512f_<code>v8siv8di2): Ditto.
1569 (avx512cd_maskb_vec_dupv8di): Ditto.
1570 (avx512cd_maskw_vec_dupv16si): Ditto.
1571 (avx512f_vcvtph2ps512): Ditto.
1572 (avx512f_vcvtps2ph512): Ditto.
1573 (VEC_EXTRACT_MODE): Extened with wider modes.
1574 (VEC_PERM_AVX2): Ditto.
1575 (VEC_PERM_CONST): Ditto.
1577 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
1579 * config/arc/arc.c (arc_ccfsm_post_advance):
1580 Add comment about TYPE_RETURN.
1582 2013-10-28 Bin Cheng <bin.cheng@arm.com>
1584 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
1585 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
1586 (strip_offset): Convert offset to unsigned number.
1588 2013-10-27 Tom de Vries <tom@codesourcery.com>
1590 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
1591 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
1593 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
1595 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
1596 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
1597 (TARGET_BIG_ENDIAN): New macro.
1598 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
1599 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
1600 * config/sh/sh.c: Likewise.
1601 * config/sh/sh.h: Likewise.
1603 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
1605 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
1606 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
1608 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1610 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
1612 (cmpstr_t, cmpstrsi): Fix formatting.
1614 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1617 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
1618 addressing, do not use general_operand for memory operands.
1620 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
1623 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1624 * lra-spills.c (lra_final_code_change): Remove useless move insns.
1626 2013-10-26 Jeff Law <law@redhat.com>
1628 * predict.c (PRED_MUDFLAP): Remove.
1629 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
1631 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
1632 (OBJS): Remove tree-nomudflap.o
1633 (GTFILES): Remove tree-mudflap.c
1634 * builtins.c (expand_builtin_alloc): Remove mudflap support.
1635 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
1636 (mfwrap_spec, mflib_spec): Likewise.
1637 (cpp_unique_options, cc1_options, static_specs): Likewise.
1638 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
1639 * passes.def: Likewise.
1640 * toplev.c (compile_file, process_options): Likewise.
1641 * tree-inline.c (copy_tree_r): Likewise.
1642 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
1643 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
1644 (build_constant_desc, output_constant_def_contents): Likewise.
1645 (categorize_decl_for_section): Likewise.
1646 * tree-mudflap.c: Removed.
1647 * tree-mudflap.h: Removed.
1648 * tree-nomudflap.c: Removed.
1649 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
1650 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
1651 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
1652 * config/sol2.h (MFLIB_SPEC): Likewise.
1653 * doc/install.texi: Remove mudflap references.
1654 * doc/passes.texi: Similarly.
1655 * doc/sourcebuild.texi: Similarly.
1656 * doc/invoke.texi: Remove mudlfap related options.
1658 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1660 PR rtl-optimization/58759
1661 * lra-constraints.c (lra_constraints): Remove wrong condition to
1662 remove insn setting up an equivalent pseudo.
1664 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1666 * config/rs6000/rs6000-protos.h
1667 (rs6000_secondary_memory_needed_mode): New prototype.
1668 * config/rs6000/rs6000.c: Include ira.h.
1669 (TARGET_LRA_P): Redefine.
1670 (rs6000_legitimate_offset_address_p): Call
1671 legitimate_constant_pool_address_p in strict mode for LRA.
1672 (rs6000_legitimate_address_p): Ditto.
1673 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
1674 (rs6000_emit_move): Add LRA version of code to generate load/store
1676 (rs6000_secondary_memory_needed_mode): New.
1677 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
1678 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
1679 constants, memory, and FP registers.
1680 (rs6000_lra_p): New.
1681 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
1682 * config/rs6000/rs6000.opt (mlra): New option.
1683 * lra-spills.c (lra_final_code_change): Remove useless move insns.
1685 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
1687 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
1688 has_single_use () and not do the conversion if has_single_use ()
1689 returns false for the multiplication result.
1691 2013-10-25 David Malcolm <dmalcolm@redhat.com>
1693 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
1694 to "NODE", since this works on a "tree", not an
1696 (CONSTANT_CLASS_P): Likewise.
1699 (INDIRECT_REF_P): Likewise.
1700 (REFERENCE_CLASS_P): Likewise.
1701 (COMPARISON_CLASS_P): Likewise.
1702 (UNARY_CLASS_P): Likewise.
1703 (BINARY_CLASS_P): Likewise.
1704 (STATEMENT_CLASS_P): Likewise.
1705 (VL_EXP_CLASS_P): Likewise.
1706 (EXPRESSION_CLASS_P): Likewise.
1707 (IS_TYPE_OR_DECL_P): Likewise.
1709 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
1711 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
1712 ADDR_EXPR in the defining statement.
1714 2013-10-25 Richard Biener <rguenther@suse.de>
1716 PR tree-optimization/58626
1717 * tree-loop-distribution.c (enum rdg_dep_type): Remove
1718 anti_dd, output_dd and input_dd.
1719 (struct rdg_edge): Remove level and relation members.
1720 (RDGE_LEVEL, RDGE_RELATION): Remove.
1721 (dot_rdg_1): Adjust.
1722 (create_rdg_edge_for_ddr): Remove.
1723 (create_rdg_edges_for_scalar): Adjust.
1724 (create_edge_for_control_dependence): Likewise.
1725 (create_rdg_edges): Split into ...
1726 (create_rdg_flow_edges): ... this
1727 (create_rdg_cd_edges): ... and this.
1729 (build_rdg): Likewise, do not compute data dependences or
1731 (pg_add_dependence_edges): New function.
1733 (distribute_loop): First apply all non-dependence based
1734 partition mergings. Then compute dependences between partitions
1735 and merge and order partitions according to them.
1737 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
1739 PR rtl-optimization/58831
1740 * alias.c (init_alias_analysis): At the beginning of each iteration,
1741 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
1743 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
1745 * recog.c (search_ofs): New static variable moved from...
1746 (peep2_find_free_register): ...here.
1747 (peephole2_optimize): Initialize it.
1749 2013-10-25 Tobias Burnus <burnus@net-b.de>
1751 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
1753 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
1755 * config/i386/i386.h (TARGET_MPX): New define.
1756 (TARGET_MPX_P): Ditto.
1758 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
1760 * config/i386/constraints.md (B): New.
1763 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
1764 * config/i386/i386-modes.def (BND32): New.
1766 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
1767 * config/i386/i386.c (isa_opts): Add mmpx.
1768 (regclass_map): Add bound registers.
1769 (dbx_register_map): Likewise.
1770 (dbx64_register_map): Likewise.
1771 (svr4_dbx_register_map): Likewise.
1773 (ix86_option_override_internal): Support MPX ISA.
1774 (ix86_conditional_register_usage): Support bound registers.
1775 (print_reg): Likewise.
1776 (ix86_code_end): Add MPX bnd prefix.
1777 (output_set_got): Likewise.
1778 (ix86_output_call_insn): Likewise.
1779 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
1780 (ix86_print_operand_punct_valid_p): Likewise.
1781 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
1783 (ix86_class_likely_spilled_p): Add bound regs support.
1784 (ix86_hard_regno_mode_ok): Likewise.
1785 (x86_order_regs_for_local_alloc): Likewise.
1786 (ix86_bnd_prefixed_insn_p): New.
1787 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
1788 (FIXED_REGISTERS): Add bound registers.
1789 (CALL_USED_REGISTERS): Likewise.
1790 (REG_ALLOC_ORDER): Likewise.
1791 (HARD_REGNO_NREGS): Likewise.
1793 (VALID_BND_REG_MODE): New.
1794 (FIRST_BND_REG): New.
1795 (LAST_BND_REG): New.
1796 (reg_class): Add BND_REGS.
1797 (REG_CLASS_NAMES): Likewise.
1798 (REG_CLASS_CONTENTS): Likewise.
1800 (ANY_BND_REG_P): New.
1802 (HI_REGISTER_NAMES): Add bound registers.
1803 * config/i386/i386.md (UNSPEC_BNDMK): New.
1804 (UNSPEC_BNDMK_ADDR): New.
1805 (UNSPEC_BNDSTX): New.
1806 (UNSPEC_BNDLDX): New.
1807 (UNSPEC_BNDLDX_ADDR): New.
1808 (UNSPEC_BNDCL): New.
1809 (UNSPEC_BNDCU): New.
1810 (UNSPEC_BNDCN): New.
1811 (UNSPEC_MPX_FENCE): New.
1814 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
1815 (length_immediate): Likewise.
1816 (prefix_0f): Likewise.
1818 (prefix_rep): Check for bnd prefix.
1819 (length_nobnd): New.
1820 (length): Use length_nobnd if specified.
1825 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
1828 (simple_return_internal): Likewise.
1829 (simple_return_pop_internal): Likewise.
1830 (*indirect_jump): Add MPX bnd prefix.
1831 (*tablejump_1): Likewise.
1832 (simple_return_internal_long): Likewise.
1833 (simple_return_indirect_internal): Likewise.
1837 (*mov<mode>_internal_mpx): New.
1838 (<mode>_<bndcheck>): New.
1839 (*<mode>_<bndcheck>): New.
1844 * config/i386/predicates.md (lea_address_operand): Rename to...
1845 (address_no_seg_operand): ... this.
1846 (address_mpx_no_base_operand): New.
1847 (address_mpx_no_index_operand): New.
1848 (bnd_mem_operator): New.
1849 * config/i386/i386.opt (mmpx): New.
1850 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
1851 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
1853 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
1855 * mode-classes.def (MODE_POINTER_BOUNDS): New.
1856 * tree.def (POINTER_BOUNDS_TYPE): New.
1857 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
1858 (POINTER_BOUNDS_MODE): New.
1859 (make_pointer_bounds_mode): New.
1860 * machmode.h (POINTER_BOUNDS_MODE_P): New.
1861 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
1862 (layout_type): Support POINTER_BOUNDS_TYPE.
1863 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
1864 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
1865 (type_contains_placeholder_1): Likewise.
1866 * tree.h (POINTER_BOUNDS_TYPE_P): New.
1867 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
1868 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
1870 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
1872 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
1873 mode of address computation when calling memory_address_addr_space.
1875 2013-08-24 Richard Henderson <rth@twiddle.net>
1878 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
1879 instead of create_convert_operand_to.
1880 (maybe_emit_sync_lock_test_and_set): Likewise.
1881 (expand_atomic_compare_and_swap): Likewise.
1882 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
1884 2013-08-24 Sriraman Tallam <tmsriram@google.com>
1886 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
1887 Change param_is to use the struct and not the pointer to the struct.
1889 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
1891 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
1892 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
1893 Move to targhooks.c.
1894 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
1895 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
1896 Relocate from gimplify.c.
1897 * targhooks.h: Add 2 prototypes.
1898 * tree.h: Delete 2 prototypes.
1900 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
1902 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
1903 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
1905 2013-10-24 Cong Hou <congh@google.com>
1907 * convert.c (convert_to_real): Guard those unsafe math function
1908 convertions with flag_unsafe_math_optimizations. Handle sqrt()
1911 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
1914 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
1916 (cgraph_create_edge): Use the new argument.
1917 (cgraph_create_indirect_edge): Likewise.
1919 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
1921 * gcc/config/arc/arc.c (arc_ccfsm_post_advance): Also handle
1923 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
1924 changes statep->state.
1926 2013-10-24 Tobias Burnus <burnus@net-b.de>
1929 * tree-cfg.c (replace_loop_annotate): New function.
1930 (execute_build_cfg): Call it.
1931 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
1932 * internal-fn.c (expand_ANNOTATE): New function.
1933 * internal-fn.def (ANNOTATE): Define as new internal function.
1934 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
1935 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
1936 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
1937 * doc/extend.texi (Pragmas): Document #pragma ivdep.
1938 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
1940 2013-10-17 Ian Bolton <ian.bolton@arm.com>
1941 Marcus Shawcroft <marcus.shawcroft@arm.com>
1943 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
1944 Special case reload SP+C into none GENERAL_REGS.
1946 2013-10-24 Michael Matz <matz@suse.de>
1948 * gengtype.c (is_file_equal): Check that files will be same length.
1950 2013-10-25 Christian Bruel <christian.bruel@st.com>
1952 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
1953 * gcc/config/sh/t-sh (sh-mem.o): New rule.
1954 * gcc/config/sh/sh-mem.cc (expand_block_move): Moved here.
1955 (sh_expand_cmpstr): New function.
1956 * gcc/config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
1957 * gcc/config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
1958 * gcc/config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
1959 (rotlhi3_8): Rename.
1961 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1963 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
1964 * Makefile.in (CXXDEPMODE): Assign and change users.
1965 (CCDEPMODE): Delete.
1966 * configure: Regenerate.
1968 2013-10-23 David Malcolm <dmalcolm@redhat.com>
1970 * gengtype-parse.c (require_without_advance): New.
1971 (type): For GTY-marked types that are not GTY((user)), parse any
1972 base classes, requiring them to be single-inheritance, and not
1973 be templates. For non-GTY-marked types and GTY((user)),
1974 continue to skip over any C++ inheritance specification.
1975 * gengtype-state.c (state_writer::write_state_struct_type):
1976 Write base class of type (if any).
1977 (read_state_struct_type): Read base class of type (if any).
1978 * gengtype.c (new_structure): Add a "base_class" parameter.
1979 (create_optional_field_): Update for new parameter to new_structure.
1980 (adjust_field_rtx_def): Likewise.
1981 (adjust_field_tree_exp): Likewise.
1982 * gengtype.h (struct type): Add "base_class" field to the s
1984 (new_structure): Add "base" parameter.
1986 2013-10-23 Sriraman Tallam <tmsriram@google.com>
1989 * config/i386/i386.c (ix86_option_override_internal):
1990 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
1991 (ix86_valid_target_attribute_tree):
1992 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
1993 Change TARGET_SSE to TARGET_SSE_P (opts->...)
1995 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
1997 * tree-ssa-loop.h: Remove include files.
1998 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
1999 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
2000 * graphite-clast-to-gimple.c: Likewise.
2001 * graphite-scop-detection.c: Likewise.
2002 * graphite-sese-to-poly.c: Likewise.
2003 * ipa-inline-analysis.c: Likewise.
2004 * ipa-pure-const.c: Likewise.
2005 * loop-init.c: Likewise.
2006 * passes.c: Likewise.
2007 * predict.c: Likewise.
2008 * tree-cfg.c: Likewise.
2009 * tree-cfgcleanup.c: Likewise.
2010 * tree-chrec.c: Likewise.
2011 * tree-data-ref.c: Likewise.
2012 * tree-loop-distribution.c: Likewise.
2013 * tree-parloops.c: Likewise.
2014 * tree-predcom.c: Likewise.
2015 * tree-scalar-evolution.c: Likewise.
2016 * tree-ssa-address.c: Likewise.
2017 * tree-ssa.c: Likewise.
2018 * tree-ssa-dce.c: Likewise.
2019 * tree-ssa-loop.c: Likewise.
2020 * tree-ssa-loop-im.c: Likewise.
2021 * tree-ssa-loop-ivcanon.c: Likewise.
2022 * tree-ssa-loop-ivopts.c: Likewise.
2023 * tree-ssa-loop-manip.c: Likewise.
2024 * tree-ssa-loop-niter.c: Likewise.
2025 * tree-ssa-loop-prefetch.c: Likewise.
2026 * tree-ssa-loop-unswitch.c: Likewise.
2027 * tree-ssa-reassoc.c: Likewise.
2028 * tree-vect-data-refs.c: Likewise.
2029 * tree-vect-loop.c: Likewise.
2030 * tree-vect-loop-manip.c: Likewise.
2031 * tree-vectorizer.c: Likewise.
2032 * tree-vect-stmts.c: Likewise.
2033 * tree-vrp.c: Likewise.
2035 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2037 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
2039 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2041 PR tree-optimization/58775
2042 PR tree-optimization/58791
2043 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
2044 (insert_stmt_after): Rewritten, don't move the stmt, but really
2046 (get_stmt_uid_with_default): Remove.
2047 (build_and_add_sum): Use insert_stmt_after and
2048 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
2049 (update_range_test): Set uid on stmts added by
2050 force_gimple_operand_gsi. Don't immediately modify statements
2051 in inter-bb optimization, just update oe->op values.
2052 (optimize_range_tests): Return bool whether any changed have been made.
2053 (update_ops): New function.
2054 (struct inter_bb_range_test_entry): New type.
2055 (maybe_optimize_range_tests): Perform statement changes here.
2056 (not_dominated_by, appears_later_in_bb, get_def_stmt,
2057 ensure_ops_are_available): Remove.
2058 (find_insert_point): Rewritten.
2059 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
2060 return LHS of the (new resp. old) stmt. Don't call
2061 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
2062 instead of last, move new stmt at the right place.
2063 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
2064 (negate_value): Likewise. Set uids.
2065 (break_up_subtract_bb): Initialize uids.
2066 (reassociate_bb): Adjust rewrite_expr_tree caller.
2067 (do_reassoc): Don't call renumber_gimple_stmt_uids.
2069 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
2072 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
2073 TARGET_32BIT final condition.
2074 (mulsi3_internal2 and splitter): Same.
2076 2013-10-23 Jeff Law <law@redhat.com>
2078 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
2079 through joiner blocks with abnormal outgoing edges.
2081 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
2082 Add parameter JOINERS, to allow/disallow threading through joiner
2084 (thread_block): New. Call thread_block_1.
2085 (mark_threaded_blocks): Remove code to filter out certain cases
2086 of threading through joiner blocks.
2087 (thread_through_all_blocks): Document how we can have a dangling
2088 edge AUX field and clear it.
2090 2013-10-23 Ian Lance Taylor <iant@google.com>
2092 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
2093 (C++ Dialect Options): Likewise.
2094 (Optimize Options): Likewise.
2096 2013-10-23 Tom de Vries <tom@codesourcery.com>
2098 PR tree-optimization/58805
2099 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
2101 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2103 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
2104 sequence based on get_range_info returned range.
2106 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2108 * tree-ssa.h: Remove all #include's
2109 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2110 * alias.c: Move required includes from tree-ssa.h.
2112 * builtins.c: Likewise.
2113 * calls.c: Likewise.
2114 * cfgexpand.c: Likewise.
2115 * cfghooks.c: Likewise.
2116 * cfgloop.c: Likewise.
2117 * cfgloopmanip.c: Likewise.
2118 * cgraph.c: Likewise.
2119 * cgraphbuild.c: Likewise.
2120 * cgraphclones.c: Likewise.
2121 * cgraphunit.c: Likewise.
2123 * except.c: Likewise.
2125 * final.c: Likewise.
2126 * fold-const.c: Likewise.
2127 * ggc-page.c: Likewise.
2128 * gimple-builder.c: Likewise.
2129 * gimple-fold.c: Likewise.
2130 * gimple-iterator.c: Likewise.
2131 * gimple-low.c: Likewise.
2132 * gimple-pretty-print.c: Likewise.
2133 * gimple-ssa-strength-reduction.c: Likewise.
2134 * gimple-streamer-in.c: Likewise.
2135 * gimple-streamer-out.c: Likewise.
2136 * gimplify.c: Likewise.
2137 * graphite-blocking.c: Likewise.
2138 * graphite-clast-to-gimple.c: Likewise.
2139 * graphite-dependences.c: Likewise.
2140 * graphite-interchange.c: Likewise.
2141 * graphite-optimize-isl.c: Likewise.
2142 * graphite-poly.c: Likewise.
2143 * graphite-scop-detection.c: Likewise.
2144 * graphite-sese-to-poly.c: Likewise.
2145 * graphite.c: Likewise.
2146 * ipa-cp.c: Likewise.
2147 * ipa-inline-analysis.c: Likewise.
2148 * ipa-inline-transform.c: Likewise.
2149 * ipa-inline.c: Likewise.
2150 * ipa-prop.c: Likewise.
2151 * ipa-pure-const.c: Likewise.
2152 * ipa-reference.c: Likewise.
2153 * ipa-split.c: Likewise.
2154 * ipa-utils.c: Likewise.
2155 * loop-init.c: Likewise.
2156 * lto-cgraph.c: Likewise.
2157 * lto-section-in.c: Likewise.
2158 * lto-section-out.c: Likewise.
2159 * lto-streamer-in.c: Likewise.
2160 * lto-streamer-out.c: Likewise.
2161 * lto-streamer.c: Likewise.
2162 * omp-low.c: Likewise.
2163 * passes.c: Likewise.
2164 * predict.c: Likewise.
2165 * print-tree.c: Likewise.
2166 * profile.c: Likewise.
2168 * targhooks.c: Likewise.
2169 * tracer.c: Likewise.
2170 * trans-mem.c: Likewise.
2171 * tree-call-cdce.c: Likewise.
2172 * tree-cfg.c: Likewise.
2173 * tree-cfgcleanup.c: Likewise.
2174 * tree-chrec.c: Likewise.
2175 * tree-complex.c: Likewise.
2176 * tree-data-ref.c: Likewise.
2177 * tree-dfa.c: Likewise.
2178 * tree-eh.c: Likewise.
2179 * tree-emutls.c: Likewise.
2180 * tree-if-conv.c: Likewise.
2181 * tree-inline.c: Likewise.
2182 * tree-into-ssa.c: Likewise.
2183 * tree-loop-distribution.c: Likewise.
2184 * tree-mudflap.c: Likewise.
2185 * tree-nested.c: Likewise.
2186 * tree-nrv.c: Likewise.
2187 * tree-object-size.c: Likewise.
2188 * tree-outof-ssa.c: Likewise.
2189 * tree-parloops.c: Likewise.
2190 * tree-phinodes.c: Likewise.
2191 * tree-predcom.c: Likewise.
2192 * tree-pretty-print.c: Likewise.
2193 * tree-profile.c: Likewise.
2194 * tree-scalar-evolution.c: Likewise.
2195 * tree-sra.c: Likewise.
2196 * tree-ssa-address.c: Likewise.
2197 * tree-ssa-alias.c: Likewise.
2198 * tree-ssa-ccp.c: Likewise.
2199 * tree-ssa-coalesce.c: Likewise.
2200 * tree-ssa-copy.c: Likewise.
2201 * tree-ssa-copyrename.c: Likewise.
2202 * tree-ssa-dce.c: Likewise.
2203 * tree-ssa-dom.c: Likewise.
2204 * tree-ssa-dse.c: Likewise.
2205 * tree-ssa-forwprop.c: Likewise.
2206 * tree-ssa-ifcombine.c: Likewise.
2207 * tree-ssa-live.c: Likewise.
2208 * tree-ssa-loop-ch.c: Likewise.
2209 * tree-ssa-loop-im.c: Likewise.
2210 * tree-ssa-loop-ivcanon.c: Likewise.
2211 * tree-ssa-loop-ivopts.c: Likewise.
2212 * tree-ssa-loop-manip.c: Likewise.
2213 * tree-ssa-loop-niter.c: Likewise.
2214 * tree-ssa-loop-prefetch.c: Likewise.
2215 * tree-ssa-loop-unswitch.c: Likewise.
2216 * tree-ssa-loop.c: Likewise.
2217 * tree-ssa-math-opts.c: Likewise.
2218 * tree-ssa-operands.c: Likewise.
2219 * tree-ssa-phiopt.c: Likewise.
2220 * tree-ssa-phiprop.c: Likewise.
2221 * tree-ssa-pre.c: Likewise.
2222 * tree-ssa-propagate.c: Likewise.
2223 * tree-ssa-reassoc.c: Likewise.
2224 * tree-ssa-sccvn.c: Likewise.
2225 * tree-ssa-sink.c: Likewise.
2226 * tree-ssa-strlen.c: Likewise.
2227 * tree-ssa-structalias.c: Likewise.
2228 * tree-ssa-tail-merge.c: Likewise.
2229 * tree-ssa-ter.c: Likewise.
2230 * tree-ssa-threadedge.c: Likewise.
2231 * tree-ssa-threadupdate.c: Likewise.
2232 * tree-ssa-uncprop.c: Likewise.
2233 * tree-ssa-uninit.c: Likewise.
2234 * tree-ssa.c: Likewise.
2235 * tree-ssanames.c: Likewise.
2236 * tree-stdarg.c: Likewise.
2237 * tree-streamer-in.c: Likewise.
2238 * tree-switch-conversion.c: Likewise.
2239 * tree-tailcall.c: Likewise.
2240 * tree-vect-data-refs.c: Likewise.
2241 * tree-vect-generic.c: Likewise.
2242 * tree-vect-loop-manip.c: Likewise.
2243 * tree-vect-loop.c: Likewise.
2244 * tree-vect-patterns.c: Likewise.
2245 * tree-vect-slp.c: Likewise.
2246 * tree-vect-stmts.c: Likewise.
2247 * tree-vectorizer.c: Likewise.
2248 * tree-vrp.c: Likewise.
2251 * value-prof.c: Likewise.
2252 * var-tracking.c: Likewise.
2253 * varpool.c: Likewise.
2254 * vtable-verify.c: Likewise.
2256 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2258 * config/tilegx/tilegx.c: Include "tree.h".
2260 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2262 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
2263 the info, not after it.
2264 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
2265 dump_ssaname_info call.
2266 (pp_gimple_stmt_1): Adjust caller.
2267 (dump_phi_nodes): Likewise. Don't print "# " here.
2269 2013-10-22 Jan Hubicka <jh@suse.cz>
2271 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
2273 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
2274 * i386.c (expand_small_movmem_or_setmem): New function.
2275 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
2277 (alg_usable_p): Add support for value ranges; cleanup.
2278 (ix86_expand_set_or_movmem): Add support for misaligned moves.
2280 2013-10-22 Sterling Augustine <saugustine@google.com>
2282 * doc/invoke.texi: Document -ggnu-pubnames.
2283 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
2285 * dwarf2out.c: Include gdb/gdb-index.h.
2286 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
2287 debug_generate_pub_sections.
2288 (is_java, output_pubtables, output_pubname): New functions.
2289 (include_pubname_in_output): Handle debug_generate_pub_sections at
2291 (size_of_pubnames): Use new local space_for_flags based on
2292 debug_generate_pub_sections.
2293 (output_pubnames): Unify pubnames and pubtypes output logic.
2294 Genericize comments. Call output_pubname.
2295 (dwarf2out_finish): Move logic to output_pubnames and call it.
2297 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
2300 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
2301 Remove CCCmode handling.
2302 <case LTU>: Return 'c' suffix for CCCmode.
2303 <case GEU>: Return 'nc' suffix for CCCmode.
2304 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
2305 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
2306 (*sub<mode>3_cc_overflow): Ditto.
2307 (*subsi3_zext_cc_overflow): Ditto.
2309 2013-10-22 Steve Ellcey <sellcey@mips.com>
2311 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
2312 (AND (NOT OP1) (NOT OP2)).
2314 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
2316 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2317 meaning of merge-high and merge-low masks for little endian; avoid
2318 use of vector-pack masks for little endian for mismatched modes.
2320 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2322 * config/tilepro/tilepro.c: Include "tree.h".
2324 2013-10-22 Andreas Schwab <schwab@suse.de>
2326 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
2329 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
2331 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
2332 [Ed Smith-Rowland]): New entries.
2333 ([Stephen M. Webb]): Update.
2335 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
2337 * tree-ssa-ter.h: Remove duplicate copy of file contents.
2339 2013-10-21 Marek Polacek <polacek@redhat.com>
2342 * fold-const.c (fold_range_test): Return 0 if the type is not
2345 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
2347 * system.h: Move hwint.h include further down.
2348 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
2349 gcc_checking_asserts.
2350 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
2352 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2354 Fix volatile issues in optimize_bit_field_compare.
2355 * fold-const.c (optimize_bit_field_compare): Bail out if
2356 lvolatilep or rvolatilep.
2358 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2360 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
2361 and get_inner_reference returning different pmode for non-volatile
2362 bit-field members dependent on flag_strict_volatile_bitfields.
2363 * stor-layout.c (layout_decl): Remove special handling of
2364 flag_strict_volatile_bitfields.
2365 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
2366 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
2368 2013-10-21 Paulo Matos <pmatos@broadcom.com>
2370 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
2373 2013-10-21 Jeff Law <law@redhat.com>
2375 * tree-ssa-threadedge.c (thread_through_normal_block): New
2376 argument VISITED. Remove VISISTED as a local variable. When we
2377 have a threadable jump, verify the destination of the jump has not
2379 (thread_across_edge): Allocate VISITED bitmap once at function
2380 scope and use it throughout. Make sure to set appropriate bits in
2381 VISITED for E (start of jump thread path).
2382 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
2383 through a joiner if any edge on the path has a recorded jump thread.
2385 2013-10-21 Ian Lance Taylor <iant@google.com>
2387 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
2388 don't imply that attributes can solve all problems.
2389 (Directory Options): Fix typo.
2391 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2393 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
2394 extend and extend_add.
2396 2013-10-21 Richard Biener <rguenther@suse.de>
2398 PR tree-optimization/58794
2399 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
2400 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
2402 2013-10-21 Richard Biener <rguenther@suse.de>
2405 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
2406 to (T) X for sign-changing conversions (or no conversion).
2408 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
2410 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
2412 2013-10-20 Jan Hubicka <jh@suse.cz>
2414 * config/i386/i386-tune.def: Add comment; organize into categories
2416 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2418 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
2419 argument. Update function comment.
2420 (expand_set_or_movmem_via_rep): New function combining
2421 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
2422 (expand_movmem_via_rep_mov): Remove.
2423 expand_setmem_via_rep_stos): Remove.
2424 (expand_movmem_epilogue): Update calls correspondingly.
2425 (expand_setmem_epilogue_via_loop): Likewise.
2427 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
2428 (expand_set_or_movmem_prologue): New function combining
2429 expand_movmem_prologue and expand_setmem_prologue.
2430 (expand_movmem_prologue): Remove.
2431 (expand_setmem_prologue): Remove.
2432 (expand_set_or_movmem_constant_prologue): New function combining
2433 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
2434 (expand_constant_movmem_prologue): Remove.
2435 (expand_constant_setmem_prologue): Remove.
2436 (promote_duplicated_reg): Allow vector-const0 value.
2437 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
2438 and ix86_expand_setmem.
2439 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
2440 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
2442 2013-10-21 Diego Novillo <dnovillo@google.com>
2444 * asan.c: Include tree.h
2445 * bb-reorder.c: Likewise.
2446 * cfgcleanup.c: Likewise.
2447 * cfgloopmanip.c: Likewise.
2448 * data-streamer-in.c: Likewise.
2449 * data-streamer-out.c: Likewise.
2450 * data-streamer.c: Likewise.
2451 * dwarf2cfi.c: Likewise.
2452 * graphite-blocking.c: Likewise.
2453 * graphite-clast-to-gimple.c: Likewise.
2454 * graphite-dependences.c: Likewise.
2455 * graphite-interchange.c: Likewise.
2456 * graphite-optimize-isl.c: Likewise.
2457 * graphite-poly.c: Likewise.
2458 * graphite-scop-detection.c: Likewise.
2459 * graphite-sese-to-poly.c: Likewise.
2460 * graphite.c: Likewise.
2461 * ipa-devirt.c: Likewise.
2462 * ipa-profile.c: Likewise.
2465 * loop-init.c: Likewise.
2466 * loop-unroll.c: Likewise.
2467 * lower-subreg.c: Likewise.
2468 * lto/lto-object.c: Likewise.
2469 * recog.c: Likewise.
2470 * reginfo.c: Likewise.
2471 * tree-loop-distribution.c: Likewise.
2472 * tree-parloops.c: Likewise.
2473 * tree-ssa-strlen.c: Likewise.
2474 * tree-streamer.c: Likewise.
2475 * value-prof.c: Likewise.
2476 * target-globals.c: Likewise.
2477 * expr.h: Include tree-core.h instead of tree.h.
2478 * gimple.h: Likewise.
2479 * ipa-prop.h: Likewise.
2480 * ipa-utils.h: Likewise.
2481 * lto-streamer.h: Likewise.
2482 * streamer-hooks.h: Likewise.
2483 * ipa-reference.h: Include cgraph.h instead of tree.h.
2484 * cgraph.h: Include basic-block.h instead of tree.h.
2485 * tree-streamer.h: Do not include tree.h.
2486 * genattrtab.c (write_header): Generate inclusion of
2488 * genautomata.c (main): Likewise.
2489 * genemit.c: Likewise.
2490 * genopinit.c: Likewise.
2491 * genoutput.c (output_prologue): Likewise.
2492 * genpeep.c: Likewise.
2494 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2496 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
2498 (vec_unpacku_hi_v8hi): Likewise.
2499 (vec_unpacku_lo_v16qi): Likewise.
2500 (vec_unpacku_lo_v8hi): Likewise.
2502 2013-10-20 Jan Hubicka <jh@suse.cz>
2504 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
2505 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
2506 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
2508 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
2510 * config/mips/mips.h (ISA_HAS_WSBH): Define.
2511 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
2513 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
2515 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
2518 * system.h: Undef m_slot.
2520 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2522 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
2523 all elements for both endian flavors.
2525 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2528 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
2529 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
2530 and SI_REG for 64bit SYSV ABI targets.
2532 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2534 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
2535 to multi_reg_return. Clarify that we are skipping USEs of multiple
2536 return registers. Use bool type where appropriate.
2538 2013-10-18 Jan Hubicka <jh@suse.cz>
2540 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
2542 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
2543 (X86_TUNE_PUSH_MEMORY): Likewise.
2544 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
2545 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
2546 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
2548 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
2549 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
2551 (ix86_option_override_internal): Update to use tune features instead
2554 2013-10-18 Cong Hou <congh@google.com>
2556 PR tree-optimization/58508
2557 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
2558 statement that contains data refs with zero-step.
2560 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2562 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
2563 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
2564 * gimple-low.c (gimple_check_call_arg,
2565 gimple_check_call_matching_types): Move to cgraph.c.
2566 * gimple-low.h: Remove prototype.
2567 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
2568 Relocate from gimple-low.c.
2569 * cgraph.h: Add prototype. Don't include basic-block.h.
2570 * gimplify.c: Add gimple-low to include list.
2571 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
2572 * tree-eh.c: Add gimple-low to include list.
2573 * tree-nested.c: Likewise.
2574 * cfgexpand.c: Add tree-ssa-address.h to include list.
2576 * gimple-fold.c: Likewise.
2577 * gimple-ssa-strength-reduction.c: Likewise.
2578 * trans-mem.c: Likewise.
2579 * tree-mudflap.c: Likewise.
2580 * tree-ssa-loop-ivopts.c: Likewise.
2581 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
2582 (degenerate_phi_result): Move to tree-phinodes.c.
2583 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
2584 * tree-ssa-threadedge.c: Likewise.
2585 * tree-vrp.c: Likewise.
2586 * tree-phinodes.c (degenerate_phi_result): Relocate here.
2587 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
2588 * tree-phinodes.h (degenerate_phi_result): Add prototype.
2589 * tree-ssa-copy.c: Include tree-ssa-dom.h.
2590 * tree-ssa-forwprop.c: Likewise.
2591 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
2592 pass_data_cleanup_cfg_post_optimizing,
2593 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
2594 * tree-optimize.c: Delete File.
2595 * graphite.c: Include tree-cfgcleanup.h.
2596 * passes.c: Likewise.
2597 * tree-cfg.c: Likewise.
2598 * tree-profile.c: Likewise.
2599 * tree-ssa-dse.c: Likewise.
2600 * tree-ssa-loop-ivcanon.c: Likewise.
2601 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
2602 * tree-outof-ssa.c: Include sbitmap.h.
2603 * tree-ssa-live.c: Likewise.
2604 * tree-ssa-propagate.c: Likewise.
2605 * tree-ssa-structalias.c: Likewise.
2606 * tree-stdarg.c: Likewise.
2607 * Makefile.in (OBJS): Delete tree-optimize.o.
2608 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
2609 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
2610 * varasm.c: Include basic-block.h.
2611 * cfgloop.h: Include function.h instead of basic-block.h
2612 (bb_loop_depth): Move to cfgloop.c.
2613 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
2615 2013-10-18 Teresa Johnson <tejohnson@google.com>
2617 * predict.c (probably_never_executed): Compare frequency-based
2618 count to number of training runs.
2619 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
2621 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2623 * config/arm/arm.c (cortexa9_extra_costs): New table.
2624 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
2626 2013-10-18 Jeff Law <law@redhat.com>
2628 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
2630 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
2631 (register_jump_thread): Add "registered_jump_thread" debug
2633 * dbgcnt.def (registered_jump_thread): New debug counter.
2635 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2637 * config/rs6000/rs6000.c: Include cgraph.h.
2639 2013-10-18 Teresa Johnson <tejohnson@google.com>
2641 * tree-ssa-tail-merge.c (replace_block_by): Update edge
2642 weights during merging.
2644 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2646 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
2647 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
2648 * tree-ssa-operands.h: Remove prototype.
2649 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
2650 * gimple.c (virtual_operand_p): Relocate from gimple.c.
2651 * gimple.h: Add prototype.
2652 * gimple-ssa.h: Include tree-ssa-operands.h.
2653 * tree-dump.c: Add tree-cfg.h to include list.
2654 * tree-ssa-alias.c: Add ipa-reference.h to include list.
2655 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
2656 * config/i386/i386.c: Don't include tree-flow.h.
2657 * config/rs6000/rs6000.c: Likewise.
2659 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2661 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
2663 2013-10-18 Richard Biener <rguenther@suse.de>
2665 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
2666 or TYPE_UNSIGNED of integral types.
2667 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
2668 NULL_TREE for zero-precision integral types.
2670 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
2672 * config/aarch64/arm_neon.h
2673 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
2675 2013-10-17 Sriraman Tallam <tmsriram@google.com>
2678 * opth-gen.awk: Define target_flags_explicit.
2680 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2682 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
2683 fields to the reg_addr array that describes the valid addressing
2684 mode for any register, general purpose registers, floating point
2685 registers, and Altivec registers.
2686 (FIRST_RELOAD_REG_CLASS): Likewise.
2687 (LAST_RELOAD_REG_CLASS): Likewise.
2688 (struct reload_reg_map_type): Likewise.
2689 (reload_reg_map_type): Likewise.
2690 (RELOAD_REG_VALID): Likewise.
2691 (RELOAD_REG_MULTIPLE): Likewise.
2692 (RELOAD_REG_INDEXED): Likewise.
2693 (RELOAD_REG_OFFSET): Likewise.
2694 (RELOAD_REG_PRE_INCDEC): Likewise.
2695 (RELOAD_REG_PRE_MODIFY): Likewise.
2696 (reg_addr): Likewise.
2697 (mode_supports_pre_incdec_p): New helper functions to say whether
2698 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
2699 (mode_supports_pre_modify_p): Likewise.
2700 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
2701 print the valid address mode bits for each mode.
2702 (rs6000_debug_print_mode): Likewise.
2703 (rs6000_debug_reg_global): Likewise.
2704 (rs6000_setup_reg_addr_masks): New function to set up the address
2705 mask bits for each type.
2706 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
2707 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
2708 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
2709 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
2710 PRE_MODIFY are supported.
2711 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
2712 registers, instead of {src,dest}_av_p.
2713 (rs6000_print_options_internal): Tweak the debug output slightly.
2715 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
2717 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
2720 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
2722 * tree-flow.h (struct omp_region): Move to omp-low.c.
2723 Remove omp_ prototypes and variables.
2724 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
2725 (copy_var_decl): Relocate prototype from tree-flow.h.
2726 * gimple.c (copy_var_decl): Relocate from omp-low.c.
2727 * tree.h: Move prototype to omp-low.h.
2728 * omp-low.h: New File. Relocate prototypes here.
2729 * omp-low.c (struct omp_region): Make local here.
2730 (root_omp_region): Make static.
2731 (copy_var_decl) Move to gimple.c.
2732 (new_omp_region): Make static.
2733 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
2734 * tree-cfg.c: Include omp-low.h.
2735 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
2736 * gimplify.c: Include omp-low.h.
2737 * tree-parloops.c: Likewise.
2739 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
2741 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
2742 operands are in memory, prefer to force non-matched operand 1 to
2745 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2748 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
2749 restrict TImode addresses to single indirect registers if both
2750 -mquad-memory and -mvsx-timode are used.
2751 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
2752 we should emit load/store quad. Remove using %y for quad memory
2755 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
2756 constraints to allow load/store quad on machines where TImode is
2757 not allowed in VSX registers. Use 'n' instead of 'F' constraint
2758 for TImode to load integer constants.
2760 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2762 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
2764 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
2766 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
2767 handling of STACK_REG.
2769 2013-10-17 Richard Biener <rguenther@suse.de>
2771 PR tree-optimization/58143
2772 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
2774 (rewrite_to_defined_overflow): Likewise.
2775 (move_computations_dom_walker::before_dom): Rewrite stmts
2776 with undefined signed overflow that are not always executed
2777 into unsigned arithmetic.
2779 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
2782 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
2783 explicit isa flag to be an options variable, instead of using
2784 global_options_set. Remove define from rs6000.h.
2785 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
2787 * config/rs6000/rs6000.c (rs6000_option_override_internal):
2788 Initialize rs6000_isa_flags_explicit.
2789 (rs6000_function_specific_save): Add gcc_options* parameter, so
2790 that the powerpc builds after the 2013-10-15 changes.
2791 (rs6000_function_specific_restore): Likewise.
2793 2013-10-16 DJ Delorie <dj@redhat.com>
2795 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
2796 op is a REG before checking REGNO.
2797 (rl78_alloc_physical_registers): Verify pattern is a SET before
2800 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2802 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
2804 (vec_unpacks_lo_v4sf): Likewise.
2805 (vec_unpacks_float_hi_v4si): Likewise.
2806 (vec_unpacks_float_lo_v4si): Likewise.
2807 (vec_unpacku_float_hi_v4si): Likewise.
2808 (vec_unpacku_float_lo_v4si): Likewise.
2810 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2812 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
2813 (vsx_concat_v2sf): Likewise.
2815 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
2817 * config/aarch64/aarch64.md
2818 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
2820 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
2822 PR tree-optimization/58697
2823 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
2824 estimated_loop_iterations_int.
2825 (max_stmt_executions_int): Call get_max_loop_iterations_int.
2826 (get_max_loop_iterations_int): New. HWINT version of
2827 get_max_loop_iterations.
2828 * cfgloop.h: Add prototypes.
2829 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
2830 * loop-unroll.c (decide_peel_once_rolling): Call
2831 get_estimated_loop_iterations_int.
2832 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
2833 * tree-ssa-loop-niter.h: Tweak prototypes.
2835 2013-10-16 David Malcolm <dmalcolm@redhat.com>
2837 * gengtype-parse.c (struct_field_seq): Ignore access-control
2838 keywords ("public:" etc).
2840 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
2842 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
2843 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
2845 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
2847 * config/arm/arm.opt (mlra): New option.
2848 * config/arm/arm.c (arm_lra_p): New function.
2849 (TARGET_LRA_P): Define.
2851 2013-10-16 Paulo Matos <pmatos@broadcom.com>
2853 * tree-core.h (tree_code_name): Remove.
2854 * tree.h (get_tree_code_name): New prototype.
2855 * tree.c (tree_code_name): Make static.
2856 (get_tree_code_name): New function.
2857 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
2858 tree_class_check_failed, tree_range_check_failed,
2859 tree_not_class_check_failed, omp_clause_check_failed,
2860 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
2861 wrapper get_tree_code_name instead of calling tree_code_name directly.
2862 * tree-vrp.c (dump_asserts_for): Likewise.
2863 * tree-dump.c (dequeue_and_dump): Likewise.
2864 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
2865 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
2866 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
2867 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
2868 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
2869 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
2870 dump_gimple_omp_for): Likewise.
2871 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
2872 * tree-ssa-pre.c (print_pre_expr): Likewise.
2873 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
2874 * print-tree.c (print_node_brief, print_node): Likewise.
2875 * gimple.c (gimple_check_failed): Likewise.
2876 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
2877 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
2878 * config/mep/mep.c (mep_validate_vliw): Likewise.
2879 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
2880 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
2882 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
2884 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
2887 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
2889 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
2890 (MULTILIB_MATCHES): Add multilib for -march=v8.
2892 2013-10-15 Sriraman Tallam <tmsriram@google.com>
2895 * optc-save-gen.awk: Add extra parameter to the save and restore
2897 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
2898 * tree.c (build_optimization_node): New parameter. Add extra parameter
2899 to call to cl_optimization_save.
2900 (build_target_option_node): New parameter. Add extra parameter
2901 to call to cl_target_option_save.
2902 * tree.h (build_optimization_node): New parameter.
2903 (build_target_option_node): New parameter.
2904 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
2905 build_optimization_node and build_target_option_node.
2906 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
2907 (handle_pragma_push_options): Ditto.
2908 * toplev.c (process_options): Ditto.
2909 * opts.c (init_options_struct): Check for opts_set non-null.
2910 * target.def (target_option.save): New parameter.
2911 (target_option.restore): New parameter.
2912 * tm.texi: Generate.
2913 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
2914 (ix86_pragma_target_parse): Ditto.
2915 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
2917 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
2918 to build_optimization_node and build_target_option_node.
2919 (rs6000_valid_attribute_p): Ditto.
2920 (rs6000_pragma_target_parse): Ditto.
2921 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
2923 * config/i386/i386.h:
2924 TARGET_64BIT_P: New Macro
2925 TARGET_MMX_P: New Macro.
2926 TARGET_3DNOW_P: New Macro.
2927 TARGET_3DNOW_A_P: New Macro.
2928 TARGET_SSE_P: New Macro.
2929 TARGET_SSE2_P: New Macro.
2930 TARGET_SSE3_P: New Macro.
2931 TARGET_SSSE3_P: New Macro.
2932 TARGET_SSE4_1_P: New Macro.
2933 TARGET_SSE4_2_P: New Macro.
2934 TARGET_AVX_P: New Macro.
2935 TARGET_AVX2_P: New Macro.
2936 TARGET_AVX512F_P: New Macro.
2937 TARGET_AVX512PF_P: New Macro.
2938 TARGET_AVX512ER_P: New Macro.
2939 TARGET_AVX512CD_P: New Macro.
2940 TARGET_FMA_P: New Macro.
2941 TARGET_SSE4A_P: New Macro.
2942 TARGET_FMA4_P: New Macro.
2943 TARGET_XOP_P: New Macro.
2944 TARGET_LWP_P: New Macro.
2945 TARGET_ABM_P: New Macro.
2946 TARGET_BMI_P: New Macro.
2947 TARGET_BMI2_P: New Macro.
2948 TARGET_LZCNT_P: New Macro.
2949 TARGET_TBM_P: New Macro.
2950 TARGET_POPCNT_P: New Macro.
2951 TARGET_SAHF_P: New Macro.
2952 TARGET_MOVBE_P: New Macro.
2953 TARGET_CRC32_P: New Macro.
2954 TARGET_AES_P: New Macro.
2955 TARGET_PCLMUL_P: New Macro.
2956 TARGET_CMPXCHG16B_P: New Macro.
2957 TARGET_FSGSBASE_P: New Macro.
2958 TARGET_RDRND_P: New Macro.
2959 TARGET_F16C_P: New Macro.
2960 TARGET_RTM_P: New Macro.
2961 TARGET_HLE_P: New Macro.
2962 TARGET_RDSEED_P: New Macro.
2963 TARGET_PRFCHW_P: New Macro.
2964 TARGET_ADX_P: New Macro.
2965 TARGET_FXSR_P: New Macro.
2966 TARGET_XSAVE_P: New Macro.
2967 TARGET_XSAVEOPT_P: New Macro.
2968 TARGET_LP64_P: New Macro.
2969 TARGET_X32_P: New Macro.
2970 TARGET_FPMATH_DEFAULT_P: New Macro.
2971 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
2972 * config/i386/i386.c (ix86_option_override_internal): New parameters.
2974 Change ix86_tune_string to access opts->x_ix86_tune_string.
2975 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
2976 Change ix86_arch_string to access opts->x_ix86_arch_string.
2977 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
2978 Change ix86_pmode to access opts->x_ix86_pmode.
2979 Change ix86_abi to access opts->x_ix86_abi.
2980 Change ix86_cmodel to access opts->x_ix86_cmodel.
2981 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
2982 Change ix86_isa_flags_explicit to access
2983 opts->x_ix86_isa_flags_explicit.
2984 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
2985 Change ix86_regparm to access opts->x_ix86_regparm.
2986 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
2987 Change ix86_preferred_stack_boundary_arg to access
2988 opts->x_ix86_preferred_stack_boundary_arg.
2989 Change ix86_force_align_arg_pointer to access
2990 opts->x_ix86_force_align_arg_pointer.
2991 Change ix86_incoming_stack_boundar_arg to access
2992 opts->x_ix86_incoming_stack_boundar_arg.
2993 Change ix86_fpmath to access opts->x_ix86_fpmath.
2994 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
2995 Change ix86_recip_name to access opts->x_ix86_recip_name.
2996 Change ix86_stack_protector_guard to access
2997 opts->x_ix86_stack_protector_guard.
2998 Change ix86_tune_memcpy_strategy to access
2999 opts->x_ix86_tune_memcpy_strategy.
3000 Change ix86_tune_memset_strategy to access
3001 opts->x_ix86_tune_memset_strategy.
3002 Change global_options to access opts.
3003 Change global_options_set to access opts_set.
3004 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
3005 Change TARGET_MMX to TARGET_MMX_P (opts->...).
3006 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
3007 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
3008 Change TARGET_SSE to TARGET_SSE_P (opts->...).
3009 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
3010 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
3011 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
3012 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
3013 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
3014 Change TARGET_AVX to TARGET_AVX_P (opts->...).
3015 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
3016 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
3017 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
3018 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
3019 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
3020 Change TARGET_FMA to TARGET_FMA_P (opts->...).
3021 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
3022 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
3023 Change TARGET_XOP to TARGET_XOP_P (opts->...).
3024 Change TARGET_LWP to TARGET_LWP_P (opts->...).
3025 Change TARGET_ABM to TARGET_ABM_P (opts->...).
3026 Change TARGET_BMI to TARGET_BMI_P (opts->...).
3027 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
3028 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
3029 Change TARGET_TBM to TARGET_TBM_P (opts->...).
3030 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
3031 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
3032 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
3033 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
3034 Change TARGET_AES to TARGET_AES_P (opts->...).
3035 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
3036 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
3037 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
3038 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
3039 Change TARGET_F16C to TARGET_F16C_P (opts->...).
3040 Change TARGET_RTM to TARGET_RTM_P (opts->...).
3041 Change TARGET_HLE to TARGET_HLE_P (opts->...).
3042 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
3043 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
3044 Change TARGET_ADX to TARGET_ADX_P (opts->...).
3045 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
3046 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
3047 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
3048 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
3049 Change TARGET_X32 to TARGET_X32_P (opts->...).
3050 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
3051 Change TARGET_FLOAT_RETURNS_IN_80387 to
3052 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
3053 (ix86_function_specific_save): New parameter. Use opts-> fields
3054 to replace global fields.
3055 (ix86_function_specific_restore): Ditto.
3056 (ix86_valid_target_attribute_inner_p): New parameters.
3058 Fix call to ix86_handle_option and set_option.
3059 (ix86_valid_target_attribute_tree): New parameters.
3060 Change global_options to access opts.
3061 Change global_options_set to access opts_set.
3062 Fix call to ix86_valid_target_attribute_inner_p.
3063 Change ix86_tune_string to access opts->x_ix86_tune_string.
3064 Change ix86_arch_string to access opts->x_ix86_arch_string.
3065 Change ix86_fpmath to access opts->x_ix86_fpmath
3066 Fix call to ix86_option_override_internal.
3067 Fix call to ix86_add_new_builtins.
3068 Fix calls to build_optimization_node and build_target_option_node.
3069 (ix86_valid_target_attribute_p): Remove access to global_options.
3070 Use new gcc_options structure func_options.
3071 Fix call to ix86_valid_target_attribute_tree.
3072 Fix call to build_optimization_node.
3073 (get_builtin_code_for_version): Fix call to
3074 ix86_valid_target_attribute_tree.
3076 2013-10-15 David Malcolm <dmalcolm@redhat.com>
3078 * Makefile.in (PICFLAG): New.
3079 (enable_host_shared): New.
3080 (INTERNAL_CFLAGS): Use PICFLAG.
3081 (LIBIBERTY): Use pic build of libiberty.a if configured with
3082 --enable-host-shared.
3083 * configure.ac: Add --enable-host-shared, setting up new
3085 * configure: Regenerate.
3086 * doc/install.texi (--enable-shared): Add note contrasting it with ...
3087 (--enable-host-shared): New option.
3089 2013-10-15 Richard Biener <rguenther@suse.de>
3091 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
3092 for built-in functions.
3094 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
3096 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
3097 (optimize_range_tests_1): New function,
3098 extracted from optimize_range_tests.
3099 (optimize_range_tests_xor): Similarly.
3100 (optimize_range_tests_diff): New function.
3101 (optimize_range_tests): Use optimize_range_tests_1.
3103 2013-10-15 Cong Hou <congh@google.com>
3105 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
3106 requirement of the reduction pattern so that one operand of the
3107 reduction operation can come from outside of the loop.
3109 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3111 * config/arm/neon-schedgen.ml: Remove.
3112 * config/arm/cortex-a9-neon.md: Remove comment regarding
3115 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3117 * config/arm/types: Remove old neon types.
3119 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3121 * config/arm/cortex-a7.md
3122 (cortex_a7_neon_type): New.
3123 (cortex_a7_neon_mul): Update for new types.
3124 (cortex_a7_neon_mla): Likewise.
3125 (cortex_a7_neon): Likewise.
3127 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3129 * config/arm/cortex-a15-neon.md
3130 (cortex_a15_neon_type): New,
3132 (cortex_a15_neon_int_1): Remove.
3133 (cortex_a15_neon_int_2): Likewise.
3134 (cortex_a15_neon_int_3): Likewise.
3135 (cortex_a15_neon_int_4): Likewise.
3136 (cortex_a15_neon_int_5): Likewise.
3137 (cortex_a15_neon_vqneg_vqabs): Likewise.
3138 (cortex_a15_neon_vmov): Likewise.
3139 (cortex_a15_neon_vaba): Likewise.
3140 (cortex_a15_neon_vaba_qqq): Likewise.
3141 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3142 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3143 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3145 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3146 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3147 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
3148 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3149 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3150 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3151 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3152 (cortex_a15_neon_shift_1): Likewise.
3153 (cortex_a15_neon_shift_2): Likewise.
3154 (cortex_a15_neon_shift_3): Likewise.
3155 (cortex_a15_neon_vshl_ddd): Likewise.
3156 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3157 (cortex_a15_neon_vsra_vrsra): Likewise.
3158 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3159 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3160 (cortex_a15_neon_bp_3cycle): Likewise.
3161 (cortex_a15_neon_ldm_2): Likewise.
3162 (cortex_a15_neon_stm_2): Likewise.
3163 (cortex_a15_neon_mcr): Likewise.
3164 (cortex_a15_neon_mrc): Likewise.
3165 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3166 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3167 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3168 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3169 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3170 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3171 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3172 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3173 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3174 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3175 (cortex_a15_neon_bp_simple): Likewise.
3176 (cortex_a15_neon_bp_2cycle): Likewise.
3177 (cortex_a15_neon_bp_3cycle): Likewise.
3178 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3179 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3180 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3181 (cortex_a15_neon_vld2_4_regs): Likewise.
3182 (cortex_a15_neon_vld3_vld4): Likewise.
3183 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3184 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3185 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
3186 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
3187 (cortex_a15_neon_vst3_vst4): Rename to...
3188 (cortex_a15_neon_vst4): This, update for new attributes.
3189 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
3190 (cortex_a15_neon_vld3_vld4_lane): Likewise.
3191 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3192 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3193 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3194 (cortex_a15_neon_ldm_2): Likewise.
3195 (cortex_a15_neon_stm_2): Likewise.
3196 (cortex_a15_neon_mcr): Likewise.
3197 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3198 (cortex_a15_neon_mrc): Likewise.
3199 (cortex_a15_neon_mrrc): Likewise.
3201 (cortex_a15_neon_abd): New.
3202 (cortex_a15_neon_abd_q): Likewise.
3203 (cortex_a15_neon_aba): Likewise.
3204 (cortex_a15_neon_aba_q): Likewise.
3205 (cortex_a15_neon_acc): Likewise.
3206 (cortex_a15_neon_acc_q): Likewise.
3207 (cortex_a15_neon_arith_basic): Likewise.
3208 (cortex_a15_neon_arith_complex): Likewise.
3209 (cortex_a15_neon_multiply): Likewise.
3210 (cortex_a15_neon_multiply_q): Likewise.
3211 (cortex_a15_neon_mla): Likewise.
3212 (cortex_a15_neon_mla_q): Likewise.
3213 (cortex_a15_neon_sat_mla_long): Likewise.
3214 (cortex_a15_neon_shift_acc): Likewise.
3215 (cortex_a15_neon_shift_imm_basic): Likewise.
3216 (cortex_a15_neon_shift_imm_complex): Likewise.
3217 (cortex_a15_neon_shift_reg_basic): Likewise.
3218 (cortex_a15_neon_shift_reg_basic_q): Likewise.
3219 (cortex_a15_neon_shift_reg_complex): Likewise.
3220 (cortex_a15_neon_shift_reg_complex_q): Likewise.
3221 (cortex_a15_neon_fp_negabs): Likewise
3222 (cortex_a15_neon_fp_arith): Likewise
3223 (cortex_a15_neon_fp_arith_q): Likewise
3224 (cortex_a15_neon_fp_cvt_int): Likewise
3225 (cortex_a15_neon_fp_cvt_int_q): Likewise
3226 (cortex_a15_neon_fp_cvt_16): Likewise
3227 (cortex_a15_neon_fp_mul): Likewise
3228 (cortex_a15_neon_fp_mul_q): Likewise
3229 (cortex_a15_neon_fp_mla): Likewise
3230 (cortex_a15_neon_fp_mla_q): Likewise
3231 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
3232 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
3233 (cortex_a15_neon_bitops): Likewise.
3234 (cortex_a15_neon_bitops_q): Likewise.
3235 (cortex_a15_neon_from_gp): Likewise.
3236 (cortex_a15_neon_from_gp_q): Likewise.
3237 (cortex_a15_neon_tbl3_tbl4): Likewise.
3238 (cortex_a15_neon_zip_q): Likewise.
3239 (cortex_a15_neon_to_gp): Likewise.
3240 (cortex_a15_neon_load_a): Likewise.
3241 (cortex_a15_neon_load_b): Likewise.
3242 (cortex_a15_neon_load_c): Likewise.
3243 (cortex_a15_neon_load_d): Likewise.
3244 (cortex_a15_neon_load_e): Likewise.
3245 (cortex_a15_neon_load_f): Likewise.
3246 (cortex_a15_neon_store_a): Likewise.
3247 (cortex_a15_neon_store_b): Likewise.
3248 (cortex_a15_neon_store_c): Likewise.
3249 (cortex_a15_neon_store_d): Likewise.
3250 (cortex_a15_neon_store_e): Likewise.
3251 (cortex_a15_neon_store_f): Likewise.
3252 (cortex_a15_neon_store_g): Likewise.
3253 (cortex_a15_neon_store_h): Likewise.
3254 (cortex_a15_vfp_to_from_gp): Likewise.
3256 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3258 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
3260 (cortex_a9_neon_vshl_ddd): Remove.
3261 (cortex_a9_neon_vst3_vst4): Likewise.
3262 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3264 (cortex_a9_neon_bit_ops_q): New.
3266 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
3267 (cortex_a9_neon_int_2): Likewise.
3268 (cortex_a9_neon_int_3): Likewise.
3269 (cortex_a9_neon_int_4): Likewise.
3270 (cortex_a9_neon_int_5): Likewise.
3271 (cortex_a9_neon_vqneg_vqabs): Likewise.
3272 (cortex_a9_neon_vmov): Likewise.
3273 (cortex_a9_neon_vaba): Likewise.
3274 (cortex_a9_neon_vaba_qqq): Likewise.
3275 (cortex_a9_neon_shift_1): Likewise.
3276 (cortex_a9_neon_shift_2): Likewise.
3277 (cortex_a9_neon_shift_3): Likewise.
3278 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3279 (cortex_a9_neon_vsra_vrsra): Likewise.
3280 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3281 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3282 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3284 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3285 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3286 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3288 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3289 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3290 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3291 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3292 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3293 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3294 (cortex_a9_neon_fp_vsum): Likewise.
3295 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3296 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3297 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3298 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3299 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3300 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3301 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3302 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3303 (cortex_a9_neon_bp_simple): Likewise.
3304 (cortex_a9_neon_bp_2cycle): Likewise.
3305 (cortex_a9_neon_bp_3cycle): Likewise.
3306 (cortex_a9_neon_ldr): Likewise.
3307 (cortex_a9_neon_str): Likewise.
3308 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3309 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3310 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3311 (cortex_a9_neon_vld2_4_regs): Likewise.
3312 (cortex_a9_neon_vld3_vld4): Likewise.
3313 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3314 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3315 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3316 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3317 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3318 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3319 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3320 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3321 (cortex_a9_neon_mcr): Likewise.
3322 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3323 (cortex_a9_neon_mrc): Likewise.
3324 (cortex_a9_neon_mrrc): Likewise.
3326 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3328 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
3330 (cortex_a8_neon_vshl_ddd): Remove.
3331 (cortex_a8_neon_vst3_vst4): Likewise.
3332 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3334 (cortex_a8_neon_bit_ops_q): New.
3336 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
3337 (cortex_a8_neon_int_2): Likewise..
3338 (cortex_a8_neon_int_3): Likewise.
3339 (cortex_a8_neon_int_5): Likewise.
3340 (cortex_a8_neon_vqneg_vqabs): Likewise.
3341 (cortex_a8_neon_int_4): Likewise.
3342 (cortex_a8_neon_vaba): Likewise.
3343 (cortex_a8_neon_vaba_qqq): Likewise.
3344 (cortex_a8_neon_shift_1): Likewise.
3345 (cortex_a8_neon_shift_2): Likewise.
3346 (cortex_a8_neon_shift_3): Likewise.
3347 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3348 (cortex_a8_neon_vsra_vrsra): Likewise.
3349 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3350 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3351 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3353 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3354 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3355 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3357 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3358 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3359 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3360 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3361 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3362 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3363 (cortex_a8_neon_fp_vsum): Likewise.
3364 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3365 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3366 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3367 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3368 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3369 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3370 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3371 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3372 (cortex_a8_neon_bp_simple): Likewise.
3373 (cortex_a8_neon_bp_2cycle): Likewise.
3374 (cortex_a8_neon_bp_3cycle): Likewise.
3375 (cortex_a8_neon_ldr): Likewise.
3376 (cortex_a8_neon_str): Likewise.
3377 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3378 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3379 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3380 (cortex_a8_neon_vld2_4_regs): Likewise.
3381 (cortex_a8_neon_vld3_vld4): Likewise.
3382 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3383 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3384 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3385 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3386 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3387 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3388 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3389 (cortex_a8_neon_mcr): Likewise.
3390 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3391 (cortex_a8_neon_mrc): Likewise.
3392 (cortex_a8_neon_mrrc): Likewise.
3394 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3396 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
3398 * config/aarch64/aarch64-simd.md (neon_type): Remove.
3399 (aarch64_simd_dup<mode>): Add "type" attribute.
3400 (aarch64_dup_lane<mode>): Likewise.
3401 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
3402 (*aarch64_simd_mov<mode>): Likewise.
3403 (aarch64_simd_mov_from_<mode>low): Likewise.
3404 (aarch64_simd_mov_from_<mode>high): Likewise.
3405 (orn<mode>3): Likewise.
3406 (bic<mode>3): Likewise.
3407 (add<mode>3): Likewise.
3408 (sub<mode>3): Likewise.
3409 (mul<mode>3): Likewise.
3410 (*aarch64_mul3_elt<mode>): Likewise.
3411 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
3412 (*aarch64_mul3_elt_to_128df): Likewise.
3413 (*aarch64_mul3_elt_to_64v2df): Likewise.
3414 (neg<mode>2): Likewise.
3415 (abs<mode>2): Likewise.
3416 (abd<mode>_3): Likewise.
3417 (aba<mode>_3): Likewise.
3418 (fabd<mode>_3): Likewise.
3419 (*fabd_scalar<mode>3): Likewise.
3420 (and<mode>3): Likewise.
3421 (ior<mode>3): Likewise.
3422 (xor<mode>3): Likewise.
3423 (one_cmpl<mode>2): Likewise.
3424 (aarch64_simd_vec_set<mode>): Likewise.
3425 (aarch64_simd_lshr<mode>): Likewise.
3426 (aarch64_simd_ashr<mode>): Likewise.
3427 (aarch64_simd_imm_shl<mode>): Likewise.
3428 (aarch64_simd_reg_sshl<mode): Likewise.
3429 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
3430 (aarch64_simd_reg_shl<mode>_signed): Likewise.
3431 (aarch64_simd_vec_setv2di): Likewise.
3432 (aarch64_simd_vec_set<mode>): Likewise.
3433 (aarch64_mla<mode>): Likewise.
3434 (*aarch64_mla_elt<mode>): Likewise.
3435 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
3436 (aarch64_mls<mode>): Likewise.
3437 (*aarch64_mls_elt<mode>): Likewise.
3438 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
3439 (<su><maxmin><mode>3): Likewise.
3440 (move_lo_quad_<mode>): Likewise.
3441 (aarch64_simd_move_hi_quad_<mode>): Likewise.
3442 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
3443 (vec_pack_trunc_<mode>): Likewise.
3444 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
3445 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
3446 (*aarch64_<su>mlal_lo<mode>): Likewise.
3447 (*aarch64_<su>mlal_hi<mode>): Likewise.
3448 (*aarch64_<su>mlsl_lo<mode>): Likewise.
3449 (*aarch64_<su>mlsl_hi<mode>): Likewise.
3450 (*aarch64_<su>mlal<mode>): Likewise.
3451 (*aarch64_<su>mlsl<mode>): Likewise.
3452 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
3453 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
3454 (add<mode>3): Likewise.
3455 (sub<mode>3): Likewise.
3456 (mul<mode>3): Likewise.
3457 (div<mode>3): Likewise.
3458 (neg<mode>2): Likewise.
3459 (abs<mode>2): Likewise.
3460 (fma<mode>4): Likewise.
3461 (*aarch64_fma4_elt<mode>): Likewise.
3462 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
3463 (*aarch64_fma4_elt_to_128df): Likewise.
3464 (*aarch64_fma4_elt_to_64v2df): Likewise.
3465 (fnma<mode>4): Likewise.
3466 (*aarch64_fnma4_elt<mode>): Likewise.
3467 (*aarch64_fnma4_elt_<vswap_width_name><mode>
3468 (*aarch64_fnma4_elt_to_128df): Likewise.
3469 (*aarch64_fnma4_elt_to_64v2df): Likewise.
3470 (<frint_pattern><mode>2): Likewise.
3471 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
3472 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
3473 (vec_unpacks_lo_v4sf): Likewise.
3474 (aarch64_float_extend_lo_v2df): Likewise.
3475 (vec_unpacks_hi_v4sf): Likewise.
3476 (aarch64_float_truncate_lo_v2sf): Likewise.
3477 (aarch64_float_truncate_hi_v4sf): Likewise.
3478 (aarch64_vmls<mode>): Likewise.
3479 (<su><maxmin><mode>3): Likewise.
3480 (<maxmin_uns><mode>3): Likewise.
3481 (reduc_<sur>plus_<mode>): Likewise.
3482 (reduc_<sur>plus_v2di): Likewise.
3483 (reduc_<sur>plus_v2si): Likewise.
3484 (reduc_<sur>plus_<mode>): Likewise.
3485 (aarch64_addpv4sf): Likewise.
3486 (clz<mode>2): Likewise.
3487 (reduc_<maxmin_uns>_<mode>): Likewise.
3488 (reduc_<maxmin_uns>_v2di): Likewise.
3489 (reduc_<maxmin_uns>_v2si): Likewise.
3490 (reduc_<maxmin_uns>_<mode>): Likewise.
3491 (reduc_<maxmin_uns>_v4sf): Likewise.
3492 (aarch64_simd_bsl<mode>_internal): Likewise.
3493 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
3494 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
3495 (aarch64_get_lane<mode>): Likewise.
3496 (*aarch64_combinez<mode>): Likewise.
3497 (aarch64_combine<mode>): Likewise.
3498 (aarch64_simd_combine<mode>): Likewise.
3499 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
3500 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
3501 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
3502 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
3503 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
3504 (aarch64_<sur>h<addsub><mode>): Likewise.
3505 (aarch64_<sur><addsub>hn<mode>): Likewise.
3506 (aarch64_<sur><addsub>hn2<mode>): Likewise.
3507 (aarch64_pmul<mode>): Likewise.
3508 (aarch64_<su_optab><optab><mode>): Likewise.
3509 (aarch64_<sur>qadd<mode>): Likewise.
3510 (aarch64_sqmovun<mode>): Likewise.
3511 (aarch64_<sur>qmovn<mode>): Likewise.
3512 (aarch64_s<optab><mode>): Likewise.
3513 (aarch64_sq<r>dmulh<mode>): Likewise.
3514 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3515 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
3516 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3517 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
3518 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3519 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3520 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
3521 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
3522 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
3523 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
3524 (aarch64_sqdmull<mode>): Likewise.
3525 (aarch64_sqdmull_lane<mode>_internal): Likewise.
3526 (aarch64_sqdmull_n<mode>): Likewise.
3527 (aarch64_sqdmull2<mode>_internal): Likewise.
3528 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
3529 (aarch64_sqdmull2_n<mode>_internal): Likewise.
3530 (aarch64_<sur>shl<mode>): Likewise.
3531 (aarch64_<sur>q<r>shl<mode>
3532 (aarch64_<sur>shll_n<mode>): Likewise.
3533 (aarch64_<sur>shll2_n<mode>): Likewise.
3534 (aarch64_<sur>shr_n<mode>): Likewise.
3535 (aarch64_<sur>sra_n<mode>): Likewise.
3536 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
3537 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
3538 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
3539 (aarch64_cm<optab><mode>): Likewise.
3540 (aarch64_cm<optab>di): Likewise.
3541 (aarch64_cm<optab><mode>): Likewise.
3542 (aarch64_cm<optab>di): Likewise.
3543 (aarch64_cmtst<mode>): Likewise.
3544 (aarch64_cmtstdi): Likewise.
3545 (aarch64_cm<optab><mode>): Likewise.
3546 (*aarch64_fac<optab><mode>): Likewise.
3547 (aarch64_addp<mode>): Likewise.
3548 (aarch64_addpdi): Likewise.
3549 (sqrt<mode>2): Likewise.
3550 (vec_load_lanesoi<mode>): Likewise.
3551 (vec_store_lanesoi<mode>): Likewise.
3552 (vec_load_lanesci<mode>): Likewise.
3553 (vec_store_lanesci<mode>): Likewise.
3554 (vec_load_lanesxi<mode>): Likewise.
3555 (vec_store_lanesxi<mode>): Likewise.
3556 (*aarch64_mov<mode>): Likewise.
3557 (aarch64_ld2<mode>_dreg): Likewise.
3558 (aarch64_ld2<mode>_dreg): Likewise.
3559 (aarch64_ld3<mode>_dreg): Likewise.
3560 (aarch64_ld3<mode>_dreg): Likewise.
3561 (aarch64_ld4<mode>_dreg): Likewise.
3562 (aarch64_ld4<mode>_dreg): Likewise.
3563 (aarch64_tbl1<mode>): Likewise.
3564 (aarch64_tbl2v16qi): Likewise.
3565 (aarch64_combinev16qi): Likewise.
3566 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
3567 (aarch64_st2<mode>_dreg): Likewise.
3568 (aarch64_st2<mode>_dreg): Likewise.
3569 (aarch64_st3<mode>_dreg): Likewise.
3570 (aarch64_st3<mode>_dreg): Likewise.
3571 (aarch64_st4<mode>_dreg): Likewise.
3572 (aarch64_st4<mode>_dreg): Likewise.
3573 (*aarch64_simd_ld1r<mode>): Likewise.
3574 (aarch64_frecpe<mode>): Likewise.
3575 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
3576 (aarch64_frecps<mode>): Likewise.
3578 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3580 * config/arm/iterators.md (V_elem_ch): New.
3582 (VQH_type): Likewise.
3583 * config/arm/arm.md (is_neon_type): New.
3584 (conds): Use is_neon_type.
3585 (anddi3_insn): Update type attribute.
3586 (xordi3_insn): Likewise.
3587 (one_cmpldi2): Likewise.
3588 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
3589 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
3590 (*movmisalign<mode>_neon_store): Likewise.
3591 (*movmisalign<mode>_neon_load): Likewise.
3592 (vec_set<mode>_internal): Likewise.
3593 (vec_set<mode>_internal): Likewise.
3594 (vec_setv2di_internal): Likewise.
3595 (vec_extract<mode>): Likewise.
3596 (vec_extract<mode>): Likewise.
3597 (vec_extractv2di): Likewise.
3598 (*add<mode>3_neon): Likewise.
3599 (adddi3_neon): Likewise.
3600 (*sub<mode>3_neon): Likewise.
3601 (subdi3_neon): Likewise.
3602 (fma<VCVTF:mode>4): Likewise.
3603 (fma<VCVTF:mode>4_intrinsic): Likewise.
3604 (*fmsub<VCVTF:mode>4): Likewise.
3605 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3606 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3607 (ior<mode>3): Likewise.
3608 (and<mode>3): Likewise.
3609 (orn<mode>3_neon): Likewise.
3610 (orndi3_neon): Likewise.
3611 (bic<mode>3_neon): Likewise.
3612 (bicdi3_neon): Likewise.
3613 (xor<mode>3): Likewise.
3614 (one_cmpl<mode>2): Likewise.
3615 (abs<mode>2): Likewise.
3616 (neg<mode>2): Likewise.
3617 (negdi2_neon): Likewise.
3618 (*umin<mode>3_neon): Likewise.
3619 (*umax<mode>3_neon): Likewise.
3620 (*smin<mode>3_neon): Likewise.
3621 (*smax<mode>3_neon): Likewise.
3622 (vashl<mode>3): Likewise.
3623 (vashr<mode>3_imm): Likewise.
3624 (vlshr<mode>3_imm): Likewise.
3625 (ashl<mode>3_signed): Likewise.
3626 (ashl<mode>3_unsigned): Likewise.
3627 (neon_load_count): Likewise.
3628 (ashldi3_neon_noclobber): Likewise.
3629 (ashldi3_neon): Likewise.
3630 (signed_shift_di3_neon): Likewise.
3631 (unsigned_shift_di3_neon): Likewise.
3632 (ashrdi3_neon_imm_noclobber): Likewise.
3633 (lshrdi3_neon_imm_noclobber): Likewise.
3634 (<shift>di3_neon): Likewise.
3635 (widen_ssum<mode>3): Likewise.
3636 (widen_usum<mode>3): Likewise.
3637 (quad_halves_<code>v4si): Likewise.
3638 (quad_halves_<code>v4sf): Likewise.
3639 (quad_halves_<code>v8hi): Likewise.
3640 (quad_halves_<code>v16qi): Likewise.
3641 (reduc_splus_v2di): Likewise.
3642 (neon_vpadd_internal<mode>): Likewise.
3643 (neon_vpsmin<mode>): Likewise.
3644 (neon_vpsmax<mode>): Likewise.
3645 (neon_vpumin<mode>): Likewise.
3646 (neon_vpumax<mode>): Likewise.
3647 (*ss_add<mode>_neon): Likewise.
3648 (*us_add<mode>_neon): Likewise.
3649 (*ss_sub<mode>_neon): Likewise.
3650 (*us_sub<mode>_neon): Likewise.
3651 (neon_vadd<mode>_unspec): Likewise.
3652 (neon_vaddl<mode>): Likewise.
3653 (neon_vaddw<mode>): Likewise.
3654 (neon_vhadd<mode>): Likewise.
3655 (neon_vqadd<mode>): Likewise.
3656 (neon_vaddhn<mode>): Likewise.
3657 (neon_vmul<mode>): Likewise.
3658 (neon_vfms<VCVTF:mode>): Likewise.
3659 (neon_vmlal<mode>): Likewise.
3660 (neon_vmls<mode>): Likewise.
3661 (neon_vmlsl<mode>): Likewise.
3662 (neon_vqdmulh<mode>): Likewise.
3663 (neon_vqdmlal<mode>): Likewise.
3664 (neon_vqdmlsl<mode>): Likewise.
3665 (neon_vmull<mode>): Likewise.
3666 (neon_vqdmull<mode>): Likewise.
3667 (neon_vsub<mode>_unspec): Likewise.
3668 (neon_vsubl<mode>): Likewise.
3669 (neon_vsubw<mode>): Likewise.
3670 (neon_vqsub<mode>): Likewise.
3671 (neon_vhsub<mode>): Likewise.
3672 (neon_vsubhn<mode>): Likewise.
3673 (neon_vceq<mode>): Likewise.
3674 (neon_vcge<mode>): Likewise.
3675 (neon_vcgeu<mode>): Likewise.
3676 (neon_vcgt<mode>): Likewise.
3677 (neon_vcgtu<mode>): Likewise.
3678 (neon_vcle<mode>): Likewise.
3679 (neon_vclt<mode>): Likewise.
3680 (neon_vcage<mode>): Likewise.
3681 (neon_vcagt<mode>): Likewise.
3682 (neon_vtst<mode>): Likewise.
3683 (neon_vabd<mode>): Likewise.
3684 (neon_vabdl<mode>): Likewise.
3685 (neon_vaba<mode>): Likewise.
3686 (neon_vabal<mode>): Likewise.
3687 (neon_vmax<mode>): Likewise.
3688 (neon_vmin<mode>): Likewise.
3689 (neon_vpaddl<mode>): Likewise.
3690 (neon_vpadal<mode>): Likewise.
3691 (neon_vpmax<mode>): Likewise.
3692 (neon_vpmin<mode>): Likewise.
3693 (neon_vrecps<mode>): Likewise.
3694 (neon_vrsqrts<mode>): Likewise.
3695 (neon_vqabs<mode>): Likewise.
3696 (neon_vqneg<mode>): Likewise.
3697 (neon_vcls<mode>): Likewise.
3698 (clz<mode>2): Likewise.
3699 (popcount<mode>2): Likewise.
3700 (neon_vrecpe<mode>): Likewise.
3701 (neon_vrsqrte<mode>): Likewise.
3702 (neon_vget_lane<mode>_sext_internal): Likewise.
3703 (neon_vget_lane<mode>_zext_internal): Likewise.
3704 (neon_vdup_n<mode>): Likewise.
3705 (neon_vdup_n<mode>): Likewise.
3706 (neon_vdup_nv2di): Likewise.
3707 (neon_vdup_lane<mode>_interal): Likewise.
3708 (*neon_vswp<mode>): Likewise.
3709 (neon_vcombine<mode>): Likewise.
3710 (float<mode><V_cvtto>2): Likewise.
3711 (floatuns<mode><V_cvtto>2): Likewise.
3712 (fix_trunc<mode><V_cvtto>2): Likewise.
3713 (fixuns_trunc<mode><V_cvtto>2
3714 (neon_vcvt<mode>): Likewise.
3715 (neon_vcvt<mode>): Likewise.
3716 (neon_vcvtv4sfv4hf): Likewise.
3717 (neon_vcvtv4hfv4sf): Likewise.
3718 (neon_vcvt_n<mode>): Likewise.
3719 (neon_vcvt_n<mode>): Likewise.
3720 (neon_vmovn<mode>): Likewise.
3721 (neon_vqmovn<mode>): Likewise.
3722 (neon_vqmovun<mode>): Likewise.
3723 (neon_vmovl<mode>): Likewise.
3724 (neon_vmul_lane<mode>): Likewise.
3725 (neon_vmul_lane<mode>): Likewise.
3726 (neon_vmull_lane<mode>): Likewise.
3727 (neon_vqdmull_lane<mode>): Likewise.
3728 (neon_vqdmulh_lane<mode>): Likewise.
3729 (neon_vqdmulh_lane<mode>): Likewise.
3730 (neon_vmla_lane<mode>): Likewise.
3731 (neon_vmla_lane<mode>): Likewise.
3732 (neon_vmlal_lane<mode>): Likewise.
3733 (neon_vqdmlal_lane<mode>): Likewise.
3734 (neon_vmls_lane<mode>): Likewise.
3735 (neon_vmls_lane<mode>): Likewise.
3736 (neon_vmlsl_lane<mode>): Likewise.
3737 (neon_vqdmlsl_lane<mode>): Likewise.
3738 (neon_vext<mode>): Likewise.
3739 (neon_vrev64<mode>): Likewise.
3740 (neon_vrev32<mode>): Likewise.
3741 (neon_vrev16<mode>): Likewise.
3742 (neon_vbsl<mode>_internal): Likewise.
3743 (neon_vshl<mode>): Likewise.
3744 (neon_vqshl<mode>): Likewise.
3745 (neon_vshr_n<mode>): Likewise.
3746 (neon_vshrn_n<mode>): Likewise.
3747 (neon_vqshrn_n<mode>): Likewise.
3748 (neon_vqshrun_n<mode>): Likewise.
3749 (neon_vshl_n<mode>): Likewise.
3750 (neon_vqshl_n<mode>): Likewise.
3751 (neon_vqshlu_n<mode>): Likewise.
3752 (neon_vshll_n<mode>): Likewise.
3753 (neon_vsra_n<mode>): Likewise.
3754 (neon_vsri_n<mode>): Likewise.
3755 (neon_vsli_n<mode>): Likewise.
3756 (neon_vtbl1v8qi): Likewise.
3757 (neon_vtbl2v8qi): Likewise.
3758 (neon_vtbl3v8qi): Likewise.
3759 (neon_vtbl4v8qi): Likewise.
3760 (neon_vtbl1v16qi): Likewise.
3761 (neon_vtbl2v16qi): Likewise.
3762 (neon_vcombinev16qi): Likewise.
3763 (neon_vtbx1v8qi): Likewise.
3764 (neon_vtbx2v8qi): Likewise.
3765 (neon_vtbx3v8qi): Likewise.
3766 (neon_vtbx4v8qi): Likewise.
3767 (*neon_vtrn<mode>_insn): Likewise.
3768 (*neon_vzip<mode>_insn): Likewise.
3769 (*neon_vuzp<mode>_insn): Likewise.
3770 (neon_vld1<mode>): Likewise.
3771 (neon_vld1_lane<mode>): Likewise.
3772 (neon_vld1_lane<mode>): Likewise.
3773 (neon_vld1_dup<mode>): Likewise.
3774 (neon_vld1_dup<mode>): Likewise.
3775 (neon_vld1_dupv2di): Likewise.
3776 (neon_vst1<mode>): Likewise.
3777 (neon_vst1_lane<mode>): Likewise.
3778 (neon_vst1_lane<mode>): Likewise.
3779 (neon_vld2<mode>): Likewise.
3780 (neon_vld2<mode>): Likewise.
3781 (neon_vld2_lane<mode>): Likewise.
3782 (neon_vld2_lane<mode>): Likewise.
3783 (neon_vld2_dup<mode>): Likewise.
3784 (neon_vst2<mode>): Likewise.
3785 (neon_vst2<mode>): Likewise.
3786 (neon_vst2_lane<mode>): Likewise.
3787 (neon_vst2_lane<mode>): Likewise.
3788 (neon_vld3<mode>): Likewise.
3789 (neon_vld3qa<mode>): Likewise.
3790 (neon_vld3qb<mode>): Likewise.
3791 (neon_vld3_lane<mode>): Likewise.
3792 (neon_vld3_lane<mode>): Likewise.
3793 (neon_vld3_dup<mode>): Likewise.
3794 (neon_vst3<mode>): Likewise.
3795 (neon_vst3qa<mode>): Likewise.
3796 (neon_vst3qb<mode>): Likewise.
3797 (neon_vst3_lane<mode>): Likewise.
3798 (neon_vst3_lane<mode>): Likewise.
3799 (neon_vld4<mode>): Likewise.
3800 (neon_vld4qa<mode>): Likewise.
3801 (neon_vld4qb<mode>): Likewise.
3802 (neon_vld4_lane<mode>): Likewise.
3803 (neon_vld4_lane<mode>): Likewise.
3804 (neon_vld4_dup<mode>): Likewise.
3805 (neon_vst4<mode>): Likewise.
3806 (neon_vst4qa<mode>): Likewise.
3807 (neon_vst4qb<mode>): Likewise.
3808 (neon_vst4_lane<mode>): Likewise.
3809 (neon_vst4_lane<mode>): Likewise.
3810 (neon_vec_unpack<US>_lo_<mode>): Likewise.
3811 (neon_vec_unpack<US>_hi_<mode>): Likewise.
3812 (neon_vec_<US>mult_lo_<mode>): Likewise.
3813 (neon_vec_<US>mult_hi_<mode>): Likewise.
3814 (neon_vec_<US>shiftl_<mode>): Likewise.
3815 (neon_unpack<US>_<mode>): Likewise.
3816 (neon_vec_<US>mult_<mode>): Likewise.
3817 (vec_pack_trunc_<mode>): Likewise.
3818 (neon_vec_pack_trunc_<mode>): Likewise.
3819 (neon_vabd<mode>_2): Likewise.
3820 (neon_vabd<mode>_3): Likewise.
3822 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3824 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
3825 (load_pair): Update type attribute.
3826 (store_pair): Update type attribute.
3827 * config/aarch64/iterators.md (q): New.
3829 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3831 * config/arm/types.md: Add new types for Neon insns.
3833 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
3834 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3835 Sergey Lega <sergey.s.lega@intel.com>
3836 Anna Tikhonova <anna.tikhonova@intel.com>
3837 Ilya Tocar <ilya.tocar@intel.com>
3838 Andrey Turetskiy <andrey.turetskiy@intel.com>
3839 Ilya Verbin <ilya.verbin@intel.com>
3840 Kirill Yukhin <kirill.yukhin@intel.com>
3841 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3843 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
3844 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
3845 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
3847 (srcp14<mode>): Ditto.
3848 (rsqrt14<mode>): Ditto.
3849 (rsqrt14<mode>): Ditto.
3850 (avx512f_vmscalef<mode>): Ditto.
3851 (avx512f_scalef<mode>): Ditto.
3852 (avx512f_getexp<mode>): Ditto.
3853 (avx512f_sgetexp<mode>): Ditto.
3854 (avx512f_fixupimm<mode>): Ditto.
3855 (avx512f_sfixupimm<mode>): Ditto.
3856 (avx512f_rndscale<mode>): Ditto.
3857 (*avx512er_exp2<mode>): Ditto.
3858 (*avx512er_rcp28<mode>): Ditto.
3859 (avx512er_rsqrt28<mode>): Ditto.
3860 (avx512f_getmant<mode>): Ditto.
3861 (avx512f_getmant<mode>): Ditto.
3862 (avx512f_rndscale<mode>): Fix formatting.
3864 2013-10-15 Martin Jambor <mjambor@suse.cz>
3866 * ipa-utils.h (ipa_edge_within_scc): Declare.
3867 * ipa-cp.c (edge_within_scc): Moved...
3868 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
3870 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
3871 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3872 Sergey Lega <sergey.s.lega@intel.com>
3873 Anna Tikhonova <anna.tikhonova@intel.com>
3874 Ilya Tocar <ilya.tocar@intel.com>
3875 Andrey Turetskiy <andrey.turetskiy@intel.com>
3876 Ilya Verbin <ilya.verbin@intel.com>
3877 Kirill Yukhin <kirill.yukhin@intel.com>
3878 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3880 * config/i386/predicates.md (const_8_to_15_operand): New.
3881 (const_16_to_31_operand): Ditto.
3882 * config/i386/sse.md (V8FI): New.
3884 (reduc_splus_v8df): Ditto.
3885 (reduc_splus_v16sf): Ditto.
3886 (avx512f_vextract<shuffletype>32x4_1): Ditto.
3887 (vec_extract_hi_<mode>): Ditto.
3888 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
3889 (vec_set_lo_<mode>): Ditto.
3890 (vec_set_hi_<mode>): Ditto.
3891 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
3892 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
3893 (avx512f_pshufd_1): Ditto.
3894 (avx512f_broadcast<mode>): Ditto.
3895 (avx512f_broadcast<mode>): Ditto.
3896 (define_split): Split vec_extract_lo into move.
3897 (ssequartermode): Ditto.
3898 (ssedoublemode): Extened with wider modes.
3899 (vec_extract_lo_<mode>): Ditto.
3901 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
3902 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3903 Sergey Lega <sergey.s.lega@intel.com>
3904 Anna Tikhonova <anna.tikhonova@intel.com>
3905 Ilya Tocar <ilya.tocar@intel.com>
3906 Andrey Turetskiy <andrey.turetskiy@intel.com>
3907 Ilya Verbin <ilya.verbin@intel.com>
3908 Kirill Yukhin <kirill.yukhin@intel.com>
3909 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3911 * config/i386/predicates.md (register_or_constm1_operand): New.
3912 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
3913 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
3914 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
3915 UNSPEC_SCATTER_PREFETCH
3917 (avx512f_ucmp<mode>3): Ditto.
3918 (avx512f_vternlog<mode>): Ditto.
3919 (avx512f_align<mode>): Ditto.
3920 (<shift_insn><mode>3): Ditto.
3921 (avx512f_<rotate>v<mode>): Ditto.
3922 (avx512f_<rotate><mode>): Ditto.
3923 (avx512f_eq<mode>3): Ditto.
3924 (avx512f_eq<mode>3_1): Ditto.
3925 (avx512f_gt<mode>3): Ditto.
3926 (avx512f_testm<mode>3): Ditto.
3927 (avx512f_testnm<mode>3): Ditto.
3928 (avx512pf_gatherpf<mode>): Ditto.
3929 (*avx512pf_gatherpf<mode>_mask): Ditto.
3930 (*avx512pf_gatherpf<mode>): Ditto.
3931 (avx512pf_scatterpf<mode>): Ditto.
3932 (*avx512pf_scatterpf<mode>_mask): Ditto.
3933 (*avx512pf_scatterpf<mode>): Ditto.
3934 (avx512f_vec_dup_gpr<mode>): Ditto.
3935 (clz<mode>2): Ditto.
3936 (conflict<mode>): Ditto.
3937 (REDUC_SMINMAX_MODE): Extened with wider modes.
3938 (reduc_<code>_<mode>): Ditto.
3939 (vlshr<mode>3): Ditto.
3940 (vashl<mode>3): Ditto.
3942 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
3943 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3944 Sergey Lega <sergey.s.lega@intel.com>
3945 Anna Tikhonova <anna.tikhonova@intel.com>
3946 Ilya Tocar <ilya.tocar@intel.com>
3947 Andrey Turetskiy <andrey.turetskiy@intel.com>
3948 Ilya Verbin <ilya.verbin@intel.com>
3949 Kirill Yukhin <kirill.yukhin@intel.com>
3950 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3952 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
3955 (avx512fmaskmode): Ditto.
3956 (bcstscalarsuff): Ditto.
3957 (avx512f_blendm<mode>): Ditto.
3958 (cmp_imm_predicate): Ditto.
3959 (avx512f_cmp<mode>3): Ditto.
3960 (avx512f_vec_dup<mode>): Ditto.
3961 (avx512f_vec_dup_mem<mode>): Ditto.
3962 (avx512f_vpermi2var<mode>3): Ditto.
3963 (avx512f_vpermt2var<mode>3): Ditto.
3964 (vec_init<mode>): Ditto.
3965 (avx512f_gathersi<mode>): Ditto.
3966 (*avx512f_gathersi<mode>): Ditto.
3967 (*avx512f_gathersi<mode>_2): Ditto.
3968 (avx512f_gatherdi<mode>): Ditto.
3969 (*avx512f_gatherdi<mode>): Ditto.
3970 (*avx512f_gatherdi<mode>_2): Ditto.
3971 (avx512f_scattersi<mode>): Ditto.
3972 (*avx512f_scattersi<mode>): Ditto.
3973 (avx512f_scatterdi<mode>): Ditto.
3974 (*avx512f_scatterdi<mode>): Ditto.
3975 (sseintprefix): Extened with wider modes.
3976 (VEC_GATHER_IDXSI): Ditto.
3977 (VEC_GATHER_IDXDI): Ditto.
3978 (VEC_GATHER_SRCDI): Ditto.
3980 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
3981 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3983 * config/arm/t-aprofile: New file.
3984 * config.gcc: Handle --with-multilib-list option.
3986 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
3988 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
3991 2013-10-15 Richard Biener <rguenther@suse.de>
3993 * tree-loop-distribution.c (build_empty_rdg): Inline into
3995 (rdg_flag_vertex): Inline into single user.
3996 (rdg_flag_vertex_and_dependent): Likewise.
3997 (build_rdg_partition_for_vertex): Remove processed bitmap.
3998 (rdg_build_partitions): Simplify.
4000 2013-10-15 Richard Biener <rguenther@suse.de>
4002 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
4003 Restructure forwarding through conversions and copies to
4004 avoid performing copy-propagation the wrong way. Adjust
4005 recursion invocations.
4006 (forward_propagate_addr_expr): Add argument stating if we
4007 are recursing from a single-use.
4008 (ssa_forward_propagate_and_combine): Adjust.
4010 2013-10-14 David Malcolm <dmalcolm@redhat.com>
4012 * dumpfile.h (gcc::dump_manager): New class, to hold state
4013 relating to dumpfile management.
4014 (get_dump_file_name): Remove in favor of method of dump_manager.
4015 (dump_initialized_p): Likewise.
4016 (dump_start): Likewise.
4017 (dump_finish): Likewise.
4018 (dump_switch_p): Likewise.
4019 (dump_register): Likewise.
4020 (get_dump_file_info): Likewise.
4021 * context.c (gcc::context::context): Construct the dump_manager
4023 * context.h (gcc::context::get_dumps): New.
4024 (gcc::context::m_dumps): New.
4025 * coverage.c (coverage_init): Port to dump_manager API.
4026 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
4027 (extra_dump_files_in_use): Likewise.
4028 (extra_dump_files_alloced): Likewise.
4029 (gcc::dump_manager::dump_manager): New.
4030 (dump_register): Convert to...
4031 (gcc::dump_manager::dump_register): ...method, replacing
4032 function-static next_dump with m_next_dump field.
4033 (get_dump_file_info): Convert to...
4034 (gcc::dump_manager::get_dump_file_info): ...method.
4035 (get_dump_file_name): Convert to...
4036 (gcc::dump_manager::get_dump_file_name): ...method.
4037 (dump_start): Convert to...
4038 (gcc::dump_manager::dump_start): ...method.
4039 (dump_finish): Convert to...
4040 (gcc::dump_manager::dump_finish): ...method.
4041 (dump_begin): Replace body with...
4042 (gcc::dump_manager::dump_begin): ...new method.
4043 (dump_phase_enabled_p): Convert to...
4044 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4045 (dump_phase_enabled_p): Convert to...
4046 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4047 (dump_initialized_p): Convert to...
4048 (gcc::dump_manager::dump_initialized_p): ...method.
4049 (dump_flag_name): Replace body with...
4050 (gcc::dump_manager::dump_flag_name): ...new method.
4051 (dump_enable_all): Convert to...
4052 (gcc::dump_manager::dump_enable_all): ...new method.
4053 (opt_info_enable_passes): Convert to...
4054 (gcc::dump_manager::opt_info_enable_passes): ...new method.
4055 (dump_switch_p_1): Convert to...
4056 (gcc::dump_manager::dump_switch_p_1): ...new method.
4057 (dump_switch_p): Convert to...
4058 (gcc::dump_manager::dump_switch_p): ...new method.
4059 (opt_info_switch_p): Port to dump_manager API.
4060 (enable_rtl_dump_file): Likewise.
4061 * opts-global.c (handle_common_deferred_options): Port to new
4063 * passes.c (pass_manager::finish_optimization_passes): Likewise.
4064 (pass_manager::register_one_dump_file): Likewise.
4065 (pass_manager::register_pass): Likewise.
4066 (pass_init_dump_file): Likewise.
4067 (pass_fini_dump_file): Likewise.
4068 * statistics.c (statistics_early_init): Likewise.
4070 2013-10-14 Richard Biener <rguenther@suse.de>
4072 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
4073 iterative_hash_canonical_type, gimple_canonical_type_hash,
4074 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
4075 gimple_register_canonical_type, print_gimple_types_stats,
4076 free_gimple_type_tables): Move to lto/lto.c
4077 (gt-gimple.h): Do not include.
4078 * gimple.h (gimple_register_canonical_type,
4079 print_gimple_types_stats, free_gimple_type_tables): Remove.
4080 * Makefile.in (GTFILES): Remove gimple.c.
4082 2013-10-14 Travis Snoozy <quandary@remstate.com>
4085 * config/msp430/msp430.c (msp430_option_override): Correct thinko
4086 scanning for msp430x targets.
4088 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
4091 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
4092 (registers_ok_for_ldd_peep): Move around.
4093 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
4094 * config/sparc/sparc.md (widening peepholes): Use it.
4096 2013-10-14 Richard Biener <rguenther@suse.de>
4100 * gimple.c (iterative_hash_canonical_type): Make sure to
4101 record the hash into the correct hashtable slot.
4103 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
4105 PR rtl-optimization/58662
4106 * combine.c (try_combine): Take into account death nodes on I2 when
4107 splitting a PARALLEL of two independent SETs. Fix dump message.
4109 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
4112 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
4113 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
4114 * config/sh/t-sh (sh_treg_combine.o): New entry.
4115 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
4116 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
4117 (register_sh_passes): New function. Register sh_treg_combine pass.
4118 (sh_option_override): Invoke it.
4119 (sh_canonicalize_comparison): Handle op0_preserve_value.
4120 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
4121 opportunities. Canonicalize branch condition.
4122 (nott): Allow only if pseudos can be created for non-SH2A.
4124 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
4127 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
4128 (ix86_expand_movmem): Replace copy_addr_to_reg with
4129 ix86_copy_addr_to_reg.
4130 (ix86_expand_setmem): Likewise.
4132 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
4134 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
4135 provided by ix86_fp_compare_mode instead of CCFPUmode.
4137 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
4139 * config/aarch64/arm_neon.h
4140 (vtbx<1,3>_<psu>8): Fix register constriants.
4142 2013-10-11 Jeff Law <law@redhat.com>
4144 PR tree-optimization/58640
4145 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
4146 threading paths that cross over two loop entry points.
4148 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4150 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
4151 handle vector float as well.
4152 (*vsx_le_perm_load_v4si): Likewise.
4153 (*vsx_le_perm_store_v2di): Likewise.
4154 (*vsx_le_perm_store_v4si): Likewise.
4156 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4158 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
4159 directly to circumvent subtract from splat{31} workaround.
4160 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
4162 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
4163 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
4164 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
4165 (altivec_vperm_<mode>): Convert to define_insn_and_split to
4166 separate big and little endian logic.
4167 (*altivec_vperm_<mode>_internal): New define_insn.
4168 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
4169 separate big and little endian logic.
4170 (*altivec_vperm_<mode>_uns_internal): New define_insn.
4171 (vec_permv16qi): Add little endian logic.
4173 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
4175 * doc/extend.texi (returns_nonnull): Remove arguments.
4177 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4178 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4179 Sergey Lega <sergey.s.lega@intel.com>
4180 Anna Tikhonova <anna.tikhonova@intel.com>
4181 Ilya Tocar <ilya.tocar@intel.com>
4182 Andrey Turetskiy <andrey.turetskiy@intel.com>
4183 Ilya Verbin <ilya.verbin@intel.com>
4184 Kirill Yukhin <kirill.yukhin@intel.com>
4185 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4187 * config/i386/sse.md (VI48F_256_512): New.
4188 (avx2_permvar<mode>): Change to ...
4189 (<avx2_avx512f>_permvar<mode>): This.
4191 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4192 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4193 Sergey Lega <sergey.s.lega@intel.com>
4194 Anna Tikhonova <anna.tikhonova@intel.com>
4195 Ilya Tocar <ilya.tocar@intel.com>
4196 Andrey Turetskiy <andrey.turetskiy@intel.com>
4197 Ilya Verbin <ilya.verbin@intel.com>
4198 Kirill Yukhin <kirill.yukhin@intel.com>
4199 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4201 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
4202 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
4203 * config/i386/sse.md (VI4_AVX): New.
4204 (sf2simodelower): Ditto.
4205 (sse2_cvtps2dq): Change to ...
4206 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
4208 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4209 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4210 Sergey Lega <sergey.s.lega@intel.com>
4211 Anna Tikhonova <anna.tikhonova@intel.com>
4212 Ilya Tocar <ilya.tocar@intel.com>
4213 Andrey Turetskiy <andrey.turetskiy@intel.com>
4214 Ilya Verbin <ilya.verbin@intel.com>
4215 Kirill Yukhin <kirill.yukhin@intel.com>
4216 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4218 * config/i386/sse.md (V_512): New.
4220 (vcond<V_512:mode><VF_512:mode>): Ditto.
4221 (vcond<V_512:mode><VI_512:mode>): Ditto.
4222 (vcondu<V_512:mode><VI_512:mode>): Ditto.
4224 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4225 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4226 Sergey Lega <sergey.s.lega@intel.com>
4227 Anna Tikhonova <anna.tikhonova@intel.com>
4228 Ilya Tocar <ilya.tocar@intel.com>
4229 Andrey Turetskiy <andrey.turetskiy@intel.com>
4230 Ilya Verbin <ilya.verbin@intel.com>
4231 Kirill Yukhin <kirill.yukhin@intel.com>
4232 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4234 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
4235 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
4237 (fma<mode>4): Removed condition.
4238 (fms<mode>4): Ditto.
4239 (fnma<mode>4): Ditto.
4240 (fnms<mode>4): Ditto.
4241 (fma4i_fmadd_<mode>): Ditto.
4242 (*fma_fmadd_<mode>): Ditto.
4243 (*fma_fmsub_<mode>): Ditto.
4244 (*fma_fnmadd_<mode>): Ditto.
4245 (*fma_fnmsub_<mode>): Ditto.
4246 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
4247 (*fma_fmaddsub_<mode>): Ditto.
4248 (*fma_fmsubadd_<mode>): Ditto.
4249 (*fmai_fmadd_<mode>): Ditto.
4250 (*fmai_fmsub_<mode>): Ditto.
4251 (*fmai_fnmadd_<mode>): Ditto.
4252 (*fmai_fnmsub_<mode>): Ditto.
4254 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4255 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4256 Sergey Lega <sergey.s.lega@intel.com>
4257 Anna Tikhonova <anna.tikhonova@intel.com>
4258 Ilya Tocar <ilya.tocar@intel.com>
4259 Andrey Turetskiy <andrey.turetskiy@intel.com>
4260 Ilya Verbin <ilya.verbin@intel.com>
4261 Kirill Yukhin <kirill.yukhin@intel.com>
4262 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4264 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
4265 (VI124_256): Changed to ...
4266 (VI124_256_48_512): This.
4267 (ssepackmode): Extended with wider modes.
4268 (<code><mode>3): Changed iterator.
4269 (*avx2_<code><mode>3): Ditto.
4270 (vec_pack_trunc_<mode>): Ditto.
4272 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4273 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4274 Sergey Lega <sergey.s.lega@intel.com>
4275 Anna Tikhonova <anna.tikhonova@intel.com>
4276 Ilya Tocar <ilya.tocar@intel.com>
4277 Andrey Turetskiy <andrey.turetskiy@intel.com>
4278 Ilya Verbin <ilya.verbin@intel.com>
4279 Kirill Yukhin <kirill.yukhin@intel.com>
4280 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4282 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
4283 (VI8F_256_512): Ditto.
4284 (abs<mode>2): Changed iterator.
4285 (avx2_perm<mode>): Changed to ...
4286 (<avx2_avx512f>_perm<mode>): This.
4287 (avx2_perm<mode>_1): Changed to ...
4288 (<avx2_avx512f>_perm<mode>_1): This.
4290 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4291 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4292 Sergey Lega <sergey.s.lega@intel.com>
4293 Anna Tikhonova <anna.tikhonova@intel.com>
4294 Ilya Tocar <ilya.tocar@intel.com>
4295 Andrey Turetskiy <andrey.turetskiy@intel.com>
4296 Ilya Verbin <ilya.verbin@intel.com>
4297 Kirill Yukhin <kirill.yukhin@intel.com>
4298 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4300 * config/i386/sse.md (VI48_AVX512F): New.
4301 (VI48_AVX2): Changed to ...
4302 (VI48_AVX2_48_AVX512F): This.
4303 (avx2_ashrv<mode>): Changed to ...
4304 (<avx2_avx512f>_ashrv<mode>): This.
4305 (avx2_<shift_insn>v<mode>): Changed to ...
4306 (<avx2_avx512f>_<shift_insn>v<mode>): This.
4308 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4309 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4310 Sergey Lega <sergey.s.lega@intel.com>
4311 Anna Tikhonova <anna.tikhonova@intel.com>
4312 Ilya Tocar <ilya.tocar@intel.com>
4313 Andrey Turetskiy <andrey.turetskiy@intel.com>
4314 Ilya Verbin <ilya.verbin@intel.com>
4315 Kirill Yukhin <kirill.yukhin@intel.com>
4316 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4318 * config/i386/sse.md (VI4_AVX512F): New.
4319 (VI8_AVX2_AVX512F): Ditto.
4320 (mul<mode>3): Extended with wider modes.
4321 (*<sse4_1_avx2>_mul<mode>3): Ditto.
4322 (mul<mode>3): Ditto.
4323 (vec_widen_<s>mult_odd_<mode>): Ditto.
4325 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4326 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4327 Sergey Lega <sergey.s.lega@intel.com>
4328 Anna Tikhonova <anna.tikhonova@intel.com>
4329 Ilya Tocar <ilya.tocar@intel.com>
4330 Andrey Turetskiy <andrey.turetskiy@intel.com>
4331 Ilya Verbin <ilya.verbin@intel.com>
4332 Kirill Yukhin <kirill.yukhin@intel.com>
4333 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4335 * config/i386/sse.md (VI2_AVX512F): New.
4336 (VI124_AVX512F): Ditto.
4337 (sseunpackmode): Extended with wider modes.
4338 (sseunpackfltmode): Ditto.
4339 (vec_unpacks_float_hi_<mode>): Ditto.
4340 (vec_unpacks_float_lo_<mode>): Ditto.
4341 (vec_unpacku_float_hi_<mode>): Ditto.
4342 (vec_unpacku_float_lo_<mode>): Ditto.
4343 (vec_unpacks_lo_<mode>): Ditto.
4344 (vec_unpacks_hi_<mode>): Ditto.
4345 (vec_unpacku_lo_<mode>): Ditto.
4346 (vec_unpacku_hi_<mode>): Ditto.
4348 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4349 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4350 Sergey Lega <sergey.s.lega@intel.com>
4351 Anna Tikhonova <anna.tikhonova@intel.com>
4352 Ilya Tocar <ilya.tocar@intel.com>
4353 Andrey Turetskiy <andrey.turetskiy@intel.com>
4354 Ilya Verbin <ilya.verbin@intel.com>
4355 Kirill Yukhin <kirill.yukhin@intel.com>
4356 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4358 * config/i386/i386.md (multdiv): New.
4359 (multdiv_mnemonic): Ditto.
4360 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
4361 (<sse>_vm<multdiv_mnemonic><mode>3): This.
4362 (<sse>_vmdiv<mode>3): Removed.
4364 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4365 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4366 Sergey Lega <sergey.s.lega@intel.com>
4367 Anna Tikhonova <anna.tikhonova@intel.com>
4368 Ilya Tocar <ilya.tocar@intel.com>
4369 Andrey Turetskiy <andrey.turetskiy@intel.com>
4370 Ilya Verbin <ilya.verbin@intel.com>
4371 Kirill Yukhin <kirill.yukhin@intel.com>
4372 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4374 * config/i386/sse.md (V): Extended with wider modes.
4376 (ssehalfvecmode): Ditto.
4378 (ssepackfltmode): Ditto.
4379 (avx_vec_concat<mode>): Ditto.
4380 (V_256_512): New iterator.
4381 (VF2_512_256): Ditto.
4382 (si2dfmode): New attribute.
4383 (si2dfmodelower): Ditto.
4385 (concat_tg_mode): Ditto.
4386 (floatv4siv4df2): Changed to ...
4387 (float<si2dfmodelower><mode>2): This.
4388 (avx_cvtps2pd256): Changed to ...
4389 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
4390 (vec_pack_trunc_v4df): Changed to ...
4391 (vec_pack_trunc_<mode>): This.
4392 (avx_vpermil<mode>): Changed to ...
4393 (<sse2_avx_avx512f>_vpermil<mode>): This.
4394 (<fixsuffix>fix_truncv8dfv8si2): New.
4395 (vec_pack_sfix_trunc_v8df): Ditto.
4396 (avx512f_rndscale<mode>): Ditto.
4397 (avx512f_roundpd512): Ditto.
4398 (vec_pack_ufix_trunc_<mode>): Updated iterator.
4400 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4401 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4402 Sergey Lega <sergey.s.lega@intel.com>
4403 Anna Tikhonova <anna.tikhonova@intel.com>
4404 Ilya Tocar <ilya.tocar@intel.com>
4405 Andrey Turetskiy <andrey.turetskiy@intel.com>
4406 Ilya Verbin <ilya.verbin@intel.com>
4407 Kirill Yukhin <kirill.yukhin@intel.com>
4408 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4410 * config/i386/i386.md (any_fix): New iterator.
4411 (fixsuffix): New attribute.
4412 * config/i386/sse.md (VF1): Extened with wider modes.
4416 (sseintvecmodelower): Ditto.
4417 (ssescalarmode): Ditto.
4418 (ssescalarnum): Ditto.
4420 (ssexmmmode): Ditto.
4421 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
4422 (<sse>_rcp<mode>2): Change iterator.
4423 (rsqrt<mode>2): Ditto.
4424 (<sse>_rsqrt<mode>2): Ditto.
4425 (avx2_vec_dup<mode>): Ditto.
4426 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
4427 (round<mode>2_sfix): Ditto.
4428 (avx2_pbroadcast<mode>): Ditto.
4429 (*andnot<mode>3): Handle XI mode.
4430 (*<code><mode>3): Ditto.
4431 (AVXTOSSEMODE): Removed.
4432 (avx_vpermil<mode>): Changed to ...
4433 (<sse2_avx_avx512f>_vpermil<mode>): This.
4435 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4436 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4437 Sergey Lega <sergey.s.lega@intel.com>
4438 Anna Tikhonova <anna.tikhonova@intel.com>
4439 Ilya Tocar <ilya.tocar@intel.com>
4440 Andrey Turetskiy <andrey.turetskiy@intel.com>
4441 Ilya Verbin <ilya.verbin@intel.com>
4442 Kirill Yukhin <kirill.yukhin@intel.com>
4443 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4445 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
4446 (<sse>_comi): Ditto.
4447 (<sse>_ucomi): Ditto.
4448 (sse_cvtss2siq_2): Ditto.
4449 (sse2_cvtsd2si): Ditto.
4450 (sse2_cvtsd2siq): Ditto.
4451 (sse2_cvttsd2si): Ditto.
4452 (sse2_cvttsd2siq): Ditto.
4453 (<shift_insn><mode>3): Ditto.
4454 (sse2_cvtsi2sdq): Update constraint and prefix.
4455 (sse_cvtsi2ss): Update prefix.
4456 (sse_cvtsi2ssq): Ditto.
4458 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4460 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
4461 ignore internal calls.
4463 2013-10-11 Richard Biener <rguenther@suse.de>
4465 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
4466 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
4467 TREE_OVERFLOW is set.
4469 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
4471 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
4473 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
4474 * gimple.h: Likewise.
4476 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
4478 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
4481 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
4483 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
4484 * configure: Regenerate.
4486 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4488 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
4489 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
4490 formatting fixes, use pp_colon instead of pp_character (..., ':'),
4491 similarly pp_right_paren.
4492 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
4493 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
4494 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
4495 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
4497 (omp_declare_simd_clauses_equal,
4498 omp_remove_redundant_declare_simd_attrs): New functions.
4499 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
4500 (walk_tree_1): Handle new OpenMP 4.0 clauses.
4501 * tree.h (OMP_LOOP_CHECK): Define.
4502 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
4503 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
4504 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
4505 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
4506 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
4507 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
4508 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
4509 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
4510 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
4511 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
4512 OMP_CLAUSE_SIMDLEN_EXPR): Define.
4513 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
4514 (omp_remove_redundant_declare_simd_attrs): New prototype.
4515 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
4516 GIMPLE_OMP_TEAMS): New codes.
4517 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
4518 * omp-low.c (struct omp_context): Add cancel_label and cancellable
4520 (target_nesting_level): New variable.
4521 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
4522 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
4523 for collapse > 1 static schedule unless ordered.
4524 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
4525 (determine_parallel_type): Adjust get_ws_args_for caller.
4526 (install_var_field): Handle mask & 4 for double indirection.
4527 (scan_sharing_clauses): Ignore shared clause on teams construct.
4528 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
4529 (create_omp_child_function): If inside target or declare target
4530 constructs, set "omp declare target" attribute on the child function.
4531 (find_combined_for): New function.
4532 (scan_omp_parallel): Handle combined loops.
4533 (scan_omp_target, scan_omp_teams): New functions.
4534 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
4535 restrictions and set ctx->cancellable for cancellable constructs.
4536 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
4537 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
4538 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
4539 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
4540 (omp_clause_aligned_alignment): New function.
4541 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
4542 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
4543 on teams constructs. Handle user defined reductions and new
4545 (lower_reduction_clauses): Don't set placeholder to address of ref
4546 if it has already the right type.
4547 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
4548 (expand_parallel_call): Use the new non-_start suffixed builtins,
4549 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
4550 and GOMP_parallel_end after the call.
4551 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
4552 (expand_omp_for_init_counts): Handle combined loops.
4553 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
4555 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
4556 end of cancellable loops.
4557 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
4558 Likewise. Handle collapse > 1 loops.
4559 (expand_omp_simd): Handle combined loops.
4560 (expand_omp_for): Add inner_stmt argument, adjust callers of
4561 expand_omp_for* functions, use expand_omp_for_static*chunk even
4562 for collapse > 1 unless ordered.
4563 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
4564 of cancellable sections.
4565 (expand_omp_single): Remove need_barrier variable, just rely on
4566 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
4567 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
4568 (expand_omp_atomic_load, expand_omp_atomic_store,
4569 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
4570 (expand_omp_target): New function.
4571 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
4572 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
4573 (build_omp_regions_1): Immediately close region for
4574 GF_OMP_TARGET_KIND_UPDATE.
4575 (maybe_add_implicit_barrier_cancel): New function.
4576 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
4578 (lower_omp_single): Likewise. Add clobber after the barrier.
4579 (lower_omp_taskgroup): New function.
4580 (lower_omp_for): Handle combined loops. Adjust
4581 lower_rec_input_clauses caller. Handle cancellation.
4582 (lower_depend_clauses): New function.
4583 (lower_omp_taskreg): Lower depend clauses. Adjust
4584 lower_rec_input_clauses caller. Add clobber after the call. Handle
4586 (lower_omp_target, lower_omp_teams): New functions.
4587 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
4588 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
4589 and GOMP_cancellation_point calls.
4590 (lower_omp): Fold stmts inside of target region.
4591 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
4592 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4593 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
4594 (BT_FN_VOID_OMPFN_PTR_UINT,
4595 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
4596 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
4597 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
4598 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
4599 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
4600 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
4601 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
4602 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
4603 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
4604 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
4605 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
4606 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
4607 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
4608 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
4609 BUILT_IN_GOMP_PARALLEL_END.
4610 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
4611 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4612 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
4613 GF_OMP_FOR_KIND_DISTRIBUTE.
4614 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
4615 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
4616 (dump_gimple_omp_return): Print lhs if it has any.
4617 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
4618 gimple_omp_atomic_seq_cst_p.
4619 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
4620 and GIMPLE_OMP_TEAMS.
4621 * langhooks.c (lhd_omp_mappable_type): New function.
4622 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
4623 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
4625 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
4626 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
4627 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
4628 (struct gimplify_omp_ctx): Add combined_loop field.
4629 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
4630 on stmts inside of target region.
4631 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
4632 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
4633 ORT_TARGET and ORT_TARGET_DATA.
4634 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
4636 (omp_notice_threadprivate_variable): Complain about threadprivate
4637 variables in target region.
4638 (omp_notice_variable): Complain about vars with non-mappable type
4639 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
4640 (omp_check_private): Ignore ORT_TARGET* regions.
4641 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
4642 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
4643 (find_combined_omp_for): New function.
4644 (gimplify_omp_for): Handle gimplification of combined loops.
4645 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
4647 (gimplify_omp_target_update): New function.
4648 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
4649 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
4650 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
4651 (gimplify_body): If fndecl has "omp declare target" attribute, add
4652 implicit ORT_TARGET context around it.
4653 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
4654 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
4655 * tree-nested.c (convert_nonlocal_reference_stmt,
4656 convert_local_reference_stmt, convert_gimple_call): Handle
4657 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4658 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
4659 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
4660 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
4661 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
4662 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
4663 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
4664 BUILT_IN_GOMP_LOOP_END_CANCEL,
4665 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
4666 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
4667 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
4668 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
4669 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
4670 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
4671 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
4672 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
4673 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
4674 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
4675 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
4676 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
4677 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
4678 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
4679 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
4680 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4681 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
4682 gimple_build_omp_teams): New functions.
4683 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
4684 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
4685 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
4686 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4687 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
4688 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
4689 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
4690 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
4691 GF_OMP_ATOMIC_SEQ_CST): New.
4692 (gimple_build_omp_taskgroup, gimple_build_omp_target,
4693 gimple_build_omp_teams): New prototypes.
4694 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
4695 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4696 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
4697 GIMPLE_OMP_SINGLE as end of range.
4698 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
4699 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
4700 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
4701 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
4702 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
4703 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
4704 gimple_omp_target_kind, gimple_omp_target_set_kind,
4705 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
4706 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
4707 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
4708 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
4709 gimple_omp_teams_set_clauses): New inlines.
4710 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
4711 and GIMPLE_OMP_TASKGROUP.
4712 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
4713 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
4714 enum omp_clause_proc_bind_kind): New.
4715 (union omp_clause_subcode): Add depend_kind, map_kind and
4716 proc_bind_kind fields.
4717 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
4718 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4719 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
4720 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
4721 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
4723 2013-10-10 Teresa Johnson <tejohnson@google.com>
4725 * predict.c (tree_estimate_probability): Add new parameter
4726 for estimate_bb_frequencies.
4727 (estimate_bb_frequencies): Add new parameter to force estimation.
4728 (rebuild_frequencies): When max frequency in function is small,
4729 recompute counts from frequencies.
4730 * predict.h (estimate_bb_frequencies): New parameter.
4732 2013-10-10 David Malcolm <dmalcolm@redhat.com>
4734 * ipa-inline.c (ipa_inline): Fix leak of "order" when
4735 optimizations are disabled.
4737 2013-10-10 David Malcolm <dmalcolm@redhat.com>
4739 * coverage.c (coverage_finish): Fix leak of da_file_name.
4741 2013-10-10 Jan Hubicka <jh@suse.cz>
4743 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
4744 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
4747 2013-10-10 Jakub Jelinek <jakub@redhat.com>
4750 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
4751 if any labels are in FALLTHRU_BB, use a special label emitted
4752 immediately after the asm goto insn rather than label_rtx
4754 (expand_asm_stmt): Adjust caller.
4755 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
4756 edge if the last insn in predecessor is a jump with single successor,
4757 but it isn't simplejump_p.
4759 2013-10-10 Richard Biener <rguenther@suse.de>
4761 PR tree-optimization/58656
4762 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
4764 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4766 * gimplify.c: Include expr.h and tm_p.h for targets with special
4767 va-arg padding requirements.
4769 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4771 * tree-flow.h: Move some prototypes to gimple.h.
4772 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
4773 * gimple.h: Relocate some prototypes from tree-flow.h
4774 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
4776 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
4777 (build_va_arg_indirect_ref): Relocate and make static.
4778 (std_gimplify_va_arg_expr): Relocate here.
4779 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
4780 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
4782 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4784 * doc/md.texi: Document the mnemonic attribute.
4786 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4789 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
4790 (set_attr_alternative x ...) when searching for user defined
4793 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4795 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
4797 2013-10-09 Easwaran Raman <eraman@google.com>
4799 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
4800 * cfgexpand.c (defer_stack_allocation): ...use here
4801 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
4803 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
4805 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
4806 (operand_equal_for_value_replacement): New function, extracted from
4807 value_replacement and enhanced to catch more cases.
4808 (value_replacement): Use operand_equal_for_value_replacement.
4810 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4812 * loop-doloop.c (doloop_modify, doloop_optimize): Use
4813 get_max_loop_iterations.
4815 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4817 * config/arm/aarch-common.c (arm_early_load_addr_dep):
4818 Place comment above function.
4820 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4822 * tree-flow.h: Remove all remaining prototypes, enums and structs that
4823 are not related to tree-cfg.c.
4824 * tree-ssa-address.h: New file. Relocate prototypes.
4825 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
4826 (addr_for_mem_ref): New. Combine call to get_address_description and
4827 return addr_for_mem_ref.
4828 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
4829 * tree-ssa-live.h: Adjust prototypes.
4830 * passes.c: Include tree-ssa-live.h.
4831 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
4832 * graphite.c (graphite_transform_loops): Make static.
4833 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
4834 make_pass_graphite, pass_data_graphite_transforms,
4835 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
4836 * ipa-pure-const.c (warn_function_noreturn): Make static.
4837 (execute_warn_function_noreturn, gate_warn_function_noreturn,
4838 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
4839 Relocate from tree-cfg.c
4840 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
4842 (execute_warn_function_noreturn, gate_warn_function_noreturn,
4843 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
4844 Move to ipa-pure-const.c.
4845 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
4846 Relocate from tree-optimize.c.
4847 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
4848 make_pass_fixup_cfg): Move to tree-cfg.c.
4849 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
4850 Relocate some prototypes.
4851 * tree-data-ref.h (tree_check_data_deps) Add prototype.
4852 * tree-dump.c (dump_function_to_file): Remove prototype.
4853 Add tree-flow.h to the include file.
4854 * tree-dump.h: Remove prototype.
4855 * tree-parloops.h: New File. Add prototypes.
4856 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
4857 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
4858 from tree-ssa-loop.c.
4859 * tree-predcom.c (run_tree_predictive_commoning,
4860 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
4861 Relocate here from tree-ssa-loop.c.
4862 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
4863 ssa_name_values.release ().
4864 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
4865 (ssa_name_values): Relocate from tree-flow.h.
4866 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
4867 * tree-ssa-loop.c (run_tree_predictive_commoning,
4868 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
4869 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
4870 make_pass_graphite, pass_data_graphite_transforms,
4871 make_pass_graphite_transforms, gate_tree_parallelize_loops,
4872 tree_parallelize_loops, pass_data_parallelize_loops,
4873 make_pass_parallelize_loops): Move to other files.
4874 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
4876 * tree.h: Remove prototypes from tree-address.c.
4878 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4880 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
4881 (struct int_tree_map): Move to tree-hasher.h
4882 (SCALE, LABEL, PERCENT): Move to gimple.h
4883 * tree-flow-inline.h: Delete. Move functions to other files.
4884 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
4885 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
4886 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
4887 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
4888 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
4889 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
4890 * tree-hasher.h: Don't include tree-flow.h.
4891 (struct int_tree_map): Relocate from tree-flow.h.
4892 * tree-sra.c (contains_view_convert_expr_p): Relocate from
4893 tree-flow-inline.h and make static.
4894 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
4896 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
4897 tree-flow-inline.h and make static.
4898 * tree.h (is_global_var, may_be_aliased): Relocate from
4900 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
4901 * value-prof.c: No longer include tree-flow-inline.h.
4902 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
4904 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4906 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
4907 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
4908 (enum move_pos): Move to tree-ssa-loop-im.h
4909 * cfgloop.h: Move some prototypes.
4910 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
4911 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
4912 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
4913 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
4914 (loop_containing_stmt): Relocate from tree-flow-inline.h.
4915 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
4916 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
4917 (enum move_pos): Relocate here.
4918 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
4920 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
4921 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
4922 make_pass_lim): Relocate here from tree-ssa-loop.c.
4923 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
4925 (loop_edge_to_cancel, unloop_loops): Make static.
4926 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
4927 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
4928 (tree_complete_unroll, gate_tree_complete_unroll,
4929 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
4930 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
4931 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
4932 * tree-ssa-loop-ivopts.c: Remove local prototypes.
4933 (stmt_invariant_in_loop_p): Remove unused function.
4934 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
4935 * tree-ssa-loop-manip.h: New file. Add prototypes.
4936 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
4937 (gcov_type_to_double_int): Move to cfgloop.h.
4938 (double_int_cmp, bound_index,
4939 estimate_numbers_of_iterations_loop): Make static.
4940 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
4941 (max_loop_iterations): Factor out get_max_loop_iterations.
4942 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
4944 * tree-ssa-loop-niter.h: New file. Add prototypes.
4945 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
4946 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
4947 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
4948 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
4949 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
4950 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
4951 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
4952 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
4953 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
4954 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
4955 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
4956 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
4957 pass_data_complete_unroll, make_pass_complete_unroll,
4958 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
4959 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
4960 tree-ssa-loop-ivcanon.c.
4961 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
4962 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
4963 tree-ssa-loop-prefetch.c.
4964 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
4966 (get_lsm_tmp_name): Relocate and add suffix parameter.
4967 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
4968 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
4969 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
4970 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
4971 (get_estimated_loop_iterations): Factor out accessor from
4972 estimated_loop_iterations in tree-ssa-loop-niter.c.
4973 (get_max_loop_iterations): Factor out accessor from
4974 _max_loop_iterations in tree-ssa-niter.c.
4975 * loop-unroll.c (decide_unroll_constant_iterations,
4976 decide_unroll_runtime_iterations, decide_peel_simple,
4977 decide_unroll_stupid): Use new get_* accessors.
4979 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
4981 PR tree-optimization/20318
4982 * doc/extend.texi (returns_nonnull): New function attribute.
4983 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
4985 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
4986 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
4988 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
4991 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
4992 false if both components are bitfields.
4994 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
4996 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
4997 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
4998 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
4999 * config/aarch64/aarch64.h
5000 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
5001 * config/aarch64/aarch64-simd-builtins.def
5002 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
5004 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5006 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
5008 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5010 * config/aarch64/arm_neon.h (vdiv_f64): Added.
5012 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5014 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
5015 (vneg_f64): New intrinsic.
5016 (vneg_s8): Asm replaced with C.
5017 (vneg_s16): Likewise.
5018 (vneg_s32): Likewise.
5019 (vneg_s64): New intrinsic.
5020 (vnegq_f32): Asm replaced with C.
5021 (vnegq_f64): Likewise.
5022 (vnegq_s8): Likewise.
5023 (vnegq_s16): Likewise.
5024 (vnegq_s32): Likewise.
5025 (vnegq_s64): Likewise.
5027 2013-10-09 Renlin Li <Renlin.Li@arm.com>
5029 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
5031 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5033 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
5034 packed stack special handling.
5035 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
5036 back to fixed stack slots for FPRs saved due to stdarg.
5038 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5040 * config/s390/s390.c (s390_frame_info): Restructure function.
5042 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5044 * config/s390/s390.c (struct s390_frame_layout): New field
5046 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
5047 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
5048 regs_ever_clobbered to char*.
5049 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
5050 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
5051 (s390_register_info): Enable FPR save slots. Move/Copy some
5052 functionality into ...
5053 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
5054 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
5056 (s390_frame_info): Do gpr slot allocation here now. stdarg does
5057 not imply a stack frame.
5058 (s390_init_frame_layout): Remove variable clobbered_regs.
5059 (s390_update_register_info): Remove function.
5060 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
5061 cannot be used for register renaming.
5062 (s390_hard_regno_scratch_ok): New function.
5063 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
5064 (s390_initial_elimination_offset): Change offset calculation of
5065 the return address pointer.
5066 (save_gprs): Deal with only r6 being saved from the call-saved regs.
5067 (restore_gprs): Set frame related flag.
5068 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
5069 (s390_emit_prologue): Call s390_register_info instead of
5070 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
5071 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
5072 (s390_optimize_prologue): Call s390_optimize_register_info.
5073 Try to remove also FPR slot save/restore INSNs. Remove frame
5074 related flags from restore INSNs.
5076 2013-10-08 DJ Delorie <dj@redhat.com>
5078 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
5081 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
5082 avoid conflict with the MI use of %c.
5083 * config/rl78/rl78-real.md: change %c to %C throughout.
5084 * config/rl78/rl78-virt.md: Likewise.
5086 2013-10-08 Jan Hubicka <jh@suse.cz>
5088 * config/i386/i386.c (ix86_option_override_internal): Switch
5089 to SSE math for -ffast-math when target ISA supports SSE2.
5091 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
5093 * tree-flow.h: Remove some prototypes.
5094 * tree.h: Remove some protypes, add a couple.
5095 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
5096 using_eh_for_cleanups_p): Add interface routines for front ends.
5097 * tree-eh.h: New file. Add protoptyes.
5098 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
5099 (add_stmt_to_eh_lp_fn): Make static.
5100 (lower_try_finally): Use new using_eh_for_cleanups_p.
5101 * emit-rtl.c: Include tree-eh.h.
5102 * gimple.h: Include tree-eh.h.
5104 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
5106 PR tree-optimization/58480
5107 * tree-vrp.c (infer_nonnull_range): New function.
5108 (infer_value_range): Call infer_nonnull_range.
5110 2013-10-08 Dehao Chen <dehao@google.com>
5112 PR tree-optimization/58619
5113 * tree-inline.c (copy_phis_for_bb): Combine location data
5116 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5119 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
5120 RTX_FRAME_RELATED_P on INSN.
5122 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5124 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
5125 (altivec_expand_vec_perm_const): Call it.
5127 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5129 * config/rs6000/vector.md (mov<mode>): Emit permuted move
5130 sequences for LE VSX loads and stores at expand time.
5131 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
5133 * config/rs6000/rs6000.c (rs6000_const_vec): New.
5134 (rs6000_gen_le_vsx_permute): New.
5135 (rs6000_gen_le_vsx_load): New.
5136 (rs6000_gen_le_vsx_store): New.
5137 (rs6000_gen_le_vsx_move): New.
5138 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
5139 (*vsx_le_perm_load_v4si): New.
5140 (*vsx_le_perm_load_v8hi): New.
5141 (*vsx_le_perm_load_v16qi): New.
5142 (*vsx_le_perm_store_v2di): New.
5143 (*vsx_le_perm_store_v4si): New.
5144 (*vsx_le_perm_store_v8hi): New.
5145 (*vsx_le_perm_store_v16qi): New.
5146 (*vsx_xxpermdi2_le_<mode>): New.
5147 (*vsx_xxpermdi4_le_<mode>): New.
5148 (*vsx_xxpermdi8_le_V8HI): New.
5149 (*vsx_xxpermdi16_le_V16QI): New.
5150 (*vsx_lxvd2x2_le_<mode>): New.
5151 (*vsx_lxvd2x4_le_<mode>): New.
5152 (*vsx_lxvd2x8_le_V8HI): New.
5153 (*vsx_lxvd2x16_le_V16QI): New.
5154 (*vsx_stxvd2x2_le_<mode>): New.
5155 (*vsx_stxvd2x4_le_<mode>): New.
5156 (*vsx_stxvd2x8_le_V8HI): New.
5157 (*vsx_stxvd2x16_le_V16QI): New.
5159 2013-10-07 Renlin Li <Renlin.Li@arm.com>
5161 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
5163 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5165 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
5166 loop to work also for 31bit ABI.
5167 Save the stack pointer for frame_size > 0.
5169 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5171 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
5172 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
5173 constraint letters from expanders.
5174 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
5175 retry count to general_operand.
5176 ("tabort"): Give operand 0 a mode.
5177 ("tabort_1"): Add mode and constraint letter for operand 0.
5178 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
5180 2013-10-04 Jeff Law <law@redhat.com>
5182 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
5184 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
5186 (thread_across_edge): Here. Call it.
5188 2013-10-04 Cary Coutant <ccoutant@google.com>
5190 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
5191 discarding a location list expression (or a piece of one).
5193 2013-10-03 Jan Hubicka <jh@suse.cz>
5195 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
5196 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
5197 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
5199 2013-10-03 Jan Hubicka <jh@suse.cz>
5201 * config/i386/i386.c (ix86_option_override_internal): Do not enable
5202 accumulate-outgoing-args when producing unwind info.
5204 2013-10-03 Wei Mi <wmi@google.com>
5206 * lra-constraints.c (insert_move_for_subreg): New function
5207 extracted from simplify_operand_subreg.
5208 (simplify_operand_subreg): Add reload for paradoxical subreg.
5210 2013-10-03 Rong Xu <xur@google.com>
5212 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
5213 the candidate of builtin_expect such that we should fix the
5214 size/time estimation.
5215 (estimate_function_body_sizes): Do the acutally size/time fix-up
5218 2013-10-03 Rong Xu <xur@google.com>
5220 * predict.c (tree_predict_by_opcode): Get the probability
5221 for builtin_expect from param builtin_expect_probability.
5222 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
5223 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
5224 * doc/invoke.texi: Add documentation for builtin-expect-probability.
5226 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
5229 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
5230 * calls.c (alloca_call_p): Use get_callee_fndecl.
5231 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
5232 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
5234 (vrp_visit_stmt): Remove duplicated code.
5236 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
5238 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
5239 ceildf2, btruncdf2, instead of vsx_* name.
5241 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
5242 iterators to only do V2DF and V4SF here. Move the DF code to
5243 rs6000.md where it is combined with SF mode. Replace <VSv> with
5244 just 'v' since only vector operations are handled with these insns
5245 after moving the DF support to rs6000.md.
5246 (vsx_sub<mode>3): Likewise.
5247 (vsx_mul<mode>3): Likewise.
5248 (vsx_div<mode>3): Likewise.
5249 (vsx_fre<mode>2): Likewise.
5250 (vsx_neg<mode>2): Likewise.
5251 (vsx_abs<mode>2): Likewise.
5252 (vsx_nabs<mode>2): Likewise.
5253 (vsx_smax<mode>3): Likewise.
5254 (vsx_smin<mode>3): Likewise.
5255 (vsx_sqrt<mode>2): Likewise.
5256 (vsx_rsqrte<mode>2): Likewise.
5257 (vsx_fms<mode>4): Likewise.
5258 (vsx_nfma<mode>4): Likewise.
5259 (vsx_copysign<mode>3): Likewise.
5260 (vsx_btrunc<mode>2): Likewise.
5261 (vsx_floor<mode>2): Likewise.
5262 (vsx_ceil<mode>2): Likewise.
5263 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
5264 (vsx_sminsf3): Likewise.
5265 (vsx_fmadf4): Likewise.
5266 (vsx_fmsdf4): Likewise.
5267 (vsx_nfmadf4): Likewise.
5268 (vsx_nfmsdf4): Likewise.
5269 (vsx_cmpdf_internal1): Likewise.
5271 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
5272 simpler to select whether a target has SPE or traditional floating
5273 point support in iterators.
5274 (TARGET_DF_SPE): Likewise.
5275 (TARGET_SF_FPR): Likewise.
5276 (TARGET_DF_FPR): Likewise.
5277 (TARGET_SF_INSN): Macros to say whether floating point support
5278 exists for a given operation for expanders.
5279 (TARGET_DF_INSN): Likewise.
5281 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
5282 combining of SF/DF mode operations, using both traditional and VSX
5290 (abs<mode>2): Combine SF/DF modes using traditional floating point
5291 instructions. Add support for using the upper DF registers with
5292 VSX support, and SF registers with power8-vector support. Update
5293 expanders for operations supported by both the SPE and traditional
5294 floating point units.
5295 (abs<mode>2_fpr): Likewise.
5296 (nabs<mode>2): Likewise.
5297 (nabs<mode>2_fpr): Likewise.
5298 (neg<mode>2): Likewise.
5299 (neg<mode>2_fpr): Likewise.
5300 (add<mode>3): Likewise.
5301 (add<mode>3_fpr): Likewise.
5302 (sub<mode>3): Likewise.
5303 (sub<mode>3_fpr): Likewise.
5304 (mul<mode>3): Likewise.
5305 (mul<mode>3_fpr): Likewise.
5306 (div<mode>3): Likewise.
5307 (div<mode>3_fpr): Likewise.
5308 (sqrt<mode>3): Likewise.
5309 (sqrt<mode>3_fpr): Likewise.
5310 (fre<Fs>): Likewise.
5311 (rsqrt<mode>2): Likewise.
5312 (cmp<mode>_fpr): Likewise.
5313 (smax<mode>3): Likewise.
5314 (smin<mode>3): Likewise.
5315 (smax<mode>3_vsx): Likewise.
5316 (smin<mode>3_vsx): Likewise.
5317 (negsf2): Delete SF operations that are merged with DF.
5324 (fmasf4_fpr): Likewise.
5325 (fmssf4_fpr): Likewise.
5326 (nfmasf4_fpr): Likewise.
5327 (nfmssf4_fpr): Likewise.
5328 (sqrtsf2): Likewise.
5329 (rsqrtsf_internal1): Likewise.
5330 (smaxsf3): Likewise.
5331 (sminsf3): Likewise.
5332 (cmpsf_internal1): Likewise.
5333 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
5334 (negdf2): Delete DF operations that are merged with SF.
5336 (nabsdf2): Likewise.
5342 (rsqrtdf_internal1): Likewise.
5343 (fmadf4_fpr): Likewise.
5344 (fmsdf4_fpr): Likewise.
5345 (nfmadf4_fpr): Likewise.
5346 (nfmsdf4_fpr): Likewise.
5347 (sqrtdf2): Likewise.
5348 (smaxdf3): Likewise.
5349 (smindf3): Likewise.
5350 (cmpdf_internal1): Likewise.
5351 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
5352 (btrunc<mode>2): Delete separate expander, and combine with the
5353 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
5354 (btrunc<mode>2_fpr): Likewise.
5355 (ceil<mode>2): Likewise.
5356 (ceil<mode>2_fpr): Likewise.
5357 (floor<mode>2): Likewise.
5358 (floor<mode>2_fpr): Likewise.
5359 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
5360 Add support for using the upper registers with VSX and
5361 power8-vector. Move insns to be closer to the define_expands. On
5362 VSX systems, prefer the traditional form of FMA over the VSX
5363 version, since the traditional form allows the target not to
5364 overlap with the inputs.
5365 (fms<mode>4_fpr): Likewise.
5366 (nfma<mode>4_fpr): Likewise.
5367 (nfms<mode>4_fpr): Likewise.
5369 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5370 Richard Earnshaw <richard.earnshaw@arm.com>
5372 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
5373 (struct mult_cost_table): Likewise.
5374 (struct mem_cost_table): Likewise.
5375 (struct fp_cost_table): Likewise.
5376 (struct vector_cost_table): Likewise.
5377 (cpu_cost_table): Likewise.
5378 * config/arm/arm.opt (mold-rts-costs): New option.
5379 (mnew-generic-costs): Likewise.
5380 * config/arm/arm.c (generic_extra_costs): New table.
5381 (cortexa15_extra_costs): Likewise.
5382 (arm_slowmul_tune): Use NULL as new costs.
5383 (arm_fastmul_tune): Likewise.
5384 (arm_strongarm_tune): Likewise.
5385 (arm_xscale_tune): Likewise.
5386 (arm_9e_tune): Likewise.
5387 (arm_v6t2_tune): Likewise.
5388 (arm_cortex_a5_tune): Likewise.
5389 (arm_cortex_a9_tune): Likewise.
5390 (arm_v6m_tune): Likewise.
5391 (arm_fa726te_tune): Likewise.
5392 (arm_cortex_a15_tune): Use cortex15_extra_costs.
5393 (arm_cortex_tune): Use generict_extra_costs.
5394 (shifter_op_p): New function.
5395 (arm_unspec_cost): Likewise.
5396 (LIBCALL_COST): Define.
5397 (arm_new_rtx_costs): New function.
5398 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
5399 table is available. Use old costs otherwise unless mnew-generic-costs
5401 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
5402 (cpu_cost_table): Declare.
5404 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
5407 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
5408 (*subs_mul_imm_<mode>)
5409 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
5410 (*sub_<shift>_<mode>)
5411 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
5412 Remove k constraint.
5414 2013-10-03 Ian Bolton <ian.bolton@arm.com>
5416 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
5418 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
5420 2013-10-02 Teresa Johnson <tejohnson@google.com>
5422 * predict.c (probably_never_executed): New function.
5423 (probably_never_executed_bb_p): Invoke probably_never_executed.
5424 (probably_never_executed_edge_p): Ditto.
5425 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
5426 Treat profile insanities conservatively.
5428 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
5430 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
5432 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
5434 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
5435 use it. Use smaller increase for scratch. Don't increase reject
5436 for early clobber scratch.
5437 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
5438 setting eliminated regs except setting fp from hfp.
5439 (lra_eliminate): Check lra_insn_recog_data on NULL.
5441 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5444 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
5445 setting -mvsx-timode by default until the underlying problem is fixed.
5446 (RS6000_CPU, power7 defaults): Likewise.
5448 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5450 * config/x-linux (host-linux.o): Remove header dependencies.
5451 Use $(COMPILE) and $(POSTCOMPILE).
5452 * config/t-linux-android (linux-android.o): Ditto.
5454 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5456 * Makefile.in (expmed.o-warn): Remove.
5458 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5460 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
5461 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
5463 2013-10-02 Teresa Johnson <tejohnson@google.com>
5465 * dojump.c (do_jump_1): Divide probability between
5466 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
5468 2013-10-02 Tom Tromey <tromey@redhat.com>
5470 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
5472 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5474 * tree-flow.h: Remove some prototypes.
5475 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
5476 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
5477 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
5478 mark_virtual_phi_result_for_renaming): Relocate here.
5479 * tree-into-ssa.h: Add prototypes.
5480 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
5481 single_pred_before_succ_order.
5482 (blocks_in_phiopt_order): Rename and move to cfganal.c.
5483 (nonfreeing_call_p) Move to gimple.c.
5484 * cfganal.c (single_pred_before_succ_order): Move and renamed from
5486 * basic-block.h (single_pred_before_succ_order): Add prototype.
5487 * gimple.c (nonfreeing_call_p): Relocate here.
5488 * gimple.h: Add prototype.
5489 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
5490 * tree-ssa-dom.h: New file. Relocate prototypes here.
5491 * tree-ssa.h: Include tree-ssa-dom.h.
5493 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5495 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
5496 Use $(COMPILE) and $(POSTCOMPILE).
5498 * config/alpha/x-alpha (driver-alpha.o): Ditto.
5500 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5502 * tree-flow.h: Remove some prototypes.
5503 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
5504 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5505 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
5506 propagate_tree_value*): Move from here to...
5507 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
5508 propagate_tree_value*): Relocate here.
5509 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5510 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
5511 * gimple-fold.c: Remove gimple-fold.h from include list.
5512 * tree-vrp.c: Remove gimple-fold.h from include list.
5513 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
5514 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
5515 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
5516 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
5517 * sese.c: Add tree-ssa-propagate.h to include list.
5519 2013-10-02 Richard Biener <rguenther@suse.de>
5521 * tree-loop-distribution.c: Include tree-vectorizer.h for
5523 (enum partition_kind): Remove PKIND_REDUCTION.
5524 (struct partition_s): Remove has_writes member, add reduction_p member.
5525 (partition_alloc): Adjust.
5526 (partition_builtin_p): Likewise.
5527 (partition_has_writes): Remove.
5528 (partition_reduction_p): New function.
5529 (partition_merge_into): Likewise.
5530 (generate_code_for_partition): Commonize builtin partition
5532 (rdg_cannot_recompute_vertex_p): Remove.
5533 (already_processed_vertex_p): Likewise.
5534 (rdg_flag_vertex): Do not set has_writes.
5535 (classify_partition): Adjust.
5536 (rdg_build_partitions): Do not set has_writes, treat all
5537 partitions as useful.
5538 (distribute_loop): Record number of library calls generated. Adjust.
5539 (tree_loop_distribution): Report number of loops and library
5540 calls generated as opt-info.
5542 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5544 * tree-flow.h: Include new .h files. Move prototypes.
5545 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
5546 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
5547 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
5548 * tree-pretty-print.h: Add prototypes from tree-flow.h.
5549 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
5550 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
5551 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
5552 * tree.h (get_ref_base_and_extent): Move prototype out.
5553 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
5555 * gimple-low.h: New File. Add prototypes from tree-flow.h.
5556 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
5557 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
5558 * tree-scalar-evolution.c: Include tree.h.
5559 * sese.c: Include tree.h.
5560 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
5561 * dwarf2out.c: Include tree-dfa.h.
5562 * tree-chrec.c: Include tree.h.
5563 * tree-data-ref.c: Include tree.h.
5565 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
5567 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
5570 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5572 * config/t-sol2 (sol2-c.o): Remove header dependencies.
5573 Use $(COMPILE) and $(POSTCOMPILE).
5574 (sol2-cxx.o): Likewise.
5575 (sol2-stubs.o): Likewise.
5577 * config/x-solaris (host-solaris.o): Likewise.
5579 * config/sparc/t-sparc (sparc.o): Remove.
5580 (sparc-c.o): Remove header dependencies.
5581 Use $(COMPILE) and $(POSTCOMPILE).
5582 * config/sparc/x-sparc: Likewise.
5584 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
5586 * config/arc/arc-opts.h: Add 2013 to Copyright years.
5587 * config/arc/arc700.md: Likewise.
5588 * config/arc/arc-modes.def: Likewise.
5589 * config/arc/arc-simd.h: Likewise.
5590 * config/arc/t-arc-uClibc: Likewise.
5591 * config/arc/t-arc-newlib: Likewise.
5593 2013-10-02 Renlin Li <renlin.li@arm.com>
5595 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
5597 (aarch64_expand_epilogue): Likewise.
5599 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5600 Yufeng Zhang <yufeng.zhang@arm.com>
5602 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
5604 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
5605 'base_in' represent a conversion and legal_cast_p_1 holds; set
5606 'base_in' with the returned value from get_unwidened.
5608 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5610 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
5611 plus_constant is not used.
5613 2013-10-01 Wei Mi <wmi@google.com>
5615 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
5616 * config/i386/i386.md: Add define_peephole2 to
5617 break partial reg stall for cvtss2sd/cvtsd2ss.
5619 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5621 * config/arc/arc.c (pass_arc_ifcvt::clone):
5622 Update for ctxt_ -> m_ctxt change.
5624 2013-10-01 Jeff Law <law@redhat.com>
5626 * tree-ssa-threadupdate.c (struct redirection_data): Delete
5627 outgoing_edge and intermediate_edge fields. Instead store the path.
5628 (redirection_data::hash): Hash on the last edge's destination index.
5629 (redirection_data::equal): Check the entire thread path.
5630 (lookup_redirectio_data): Corresponding changes.
5631 (create_edge_and_update_destination_phis): Likewise.
5632 (thread_single_edge): Likewise.
5634 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5635 Diego Novillo <dnovillo@google.com>
5637 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
5638 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
5639 (vld32wh_insn, vld32wl_insn): Delete commented-out old
5640 versions of these patterns.
5642 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
5643 (__builtin_arc_aligned): Likewise.
5645 * config/arc/arc.md: Expand adc_0 comment stating the intended
5646 purpose and why it isn't ready.
5647 Replace commented out call_value_via_label_mixed with a
5648 plain comment about bl_s.
5650 * config/arc/arc.c (stdio.h): Don't include directly.
5651 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
5652 Remove if (1) condition.
5653 (arc_encode_section_info): Fix comment.
5655 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5657 * config/arc/arc.c (arc_conditional_register_usage):
5658 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
5659 Also set reg_alloc_order for DMA config regs.
5661 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5662 Jeremy Bennett <jeremy.bennett@embecosm.com>
5664 * doc/install.texi (--with-cpu): Mention ARC.
5665 (arc-*-elf32): New paragraph.
5666 (arc-linux-uclibc): Likewise.
5667 * doc/md.texi (Machine Constraints): Add ARC part.
5668 * doc/invoke.texi: (menu): Add ARC Options.
5669 (Machine Dependent Options) <ARC Options>: Add synopsis.
5670 (node ARC Options): Add.
5671 * doc/extend.texi (long_call / short_call attribute): Add ARC.
5672 (ARC Built-in Functions): New section defining
5673 generic ARC built-in functions.
5674 (ARC SIMD Built-in Functions): New section defining SIMD specific
5676 (Declaring Attributes of Functions): Extended
5677 description of short_call and long_call attributes for ARC and
5678 added index entries.
5680 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
5681 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
5682 Joern Rennecke <joern.rennecke@embecosm.com>
5683 Muhammad Khurram Riaz <khurram.riaz@arc.com>
5684 Brendan Kehoe <brendan@zen.org>
5685 Michael Eager <eager@eagercon.com>
5686 Simon Cook <simon.cook@embecosm.com>
5687 Jeremy Bennett <jeremy.bennett@embecosm.com>
5689 * config/arc, common/config/arc: New directories.
5691 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5692 Brendan Kehoe <brendan@zen.org>
5693 Simon Cook <simon.cook@embecosm.com>
5695 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
5697 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
5699 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
5700 * tree-ssa-coalesce.h: New. Move prototype to here.
5701 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
5702 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
5703 (gimple_can_coalesce_p): Move to...
5704 * gimple.c (gimple_can_coalesce_p): Here.
5706 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
5708 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
5709 (dump_decl_set): Move to gimple.c.
5710 * gimple.h: Don't include tree-ssa-operands.h.
5711 (dump_decl_set): Add prototype.
5712 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
5713 Move to gimple-ssa.h.
5714 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
5715 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
5716 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
5717 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
5718 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
5720 (dump_decl_set): Relocate here.
5721 * gimple-ssa.h: New file.
5722 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
5723 Relocate from gimple.h.
5724 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
5725 * tree-ssa-operands.c (swap_ssa_operands): Rename from
5726 swap_tree_operands and remove non-ssa path.
5727 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
5728 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
5730 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
5731 vect_is_simple_reduction_1): Use swap_ssa_operands.
5732 * tree-flow.h: Move various prototypes to tree-phinodes.h.
5733 (enum need_phi_state): Move to tree-into-ssa.c.
5734 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
5735 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
5736 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
5737 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
5738 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
5739 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
5740 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
5741 num_imm_uses): Move to ssa-iterators.h.
5742 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
5743 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
5745 (op_iter_done, op_iter_next_def, op_iter_next_tree,
5746 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
5747 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
5748 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
5749 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
5750 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
5751 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
5752 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
5753 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
5754 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
5755 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
5756 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
5757 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
5758 (set_phi_nodes): Move to tree-phinodes.h.
5759 * tree-ssa-operands.h (enum ssa_op_iter_type,
5760 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
5761 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
5762 (dump_decl_set): Remove prototype.
5763 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
5764 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
5765 (set_phi_nodes): Relocate from tree-flow-inline.h.
5766 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
5768 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
5769 include list. Temporarily add gimple.h to include list.
5770 * ssa-iterators.h: New file.
5771 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
5772 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
5773 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
5774 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
5775 Relocate from tree-ssa-operands.h.
5776 (delink_imm_use, link_imm_use_to_list, link_imm_use,
5777 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
5778 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
5779 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
5780 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
5781 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
5782 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
5783 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
5784 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
5785 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
5786 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
5787 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
5788 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
5789 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
5790 Relocate from tree-flow-inline.h.
5791 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
5793 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
5796 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
5797 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
5798 Insert '\t' to output template.
5799 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
5800 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
5801 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
5802 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
5804 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
5806 * doc/install.texi (Host/target specific installation notes for GCC):
5807 Put @anchor before @heading.
5808 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
5809 Use @email for email addresses.
5811 2013-10-01 Jeff Law <law@redhat.com>
5813 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
5814 a vec. Only delete the path if we create one without successfully
5815 registering a jump thread.
5816 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
5818 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
5819 (paths): New vector of jump threading paths.
5820 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
5821 (THREAD_PATH): New accessor macro for the entire thread path.
5822 (lookup_redirection_data): Get intermediate and final outgoing edge
5823 from the thread path.
5824 (create_edge_and_update_destination_phis): Copy the threading path.
5825 (ssa_fix_duplicate_block_edges): Get edges and block types from the
5826 jump threading path.
5827 (ssa_redirect_edges): Get edges and block types from the jump threading
5828 path. Free the path vector.
5829 (thread_block): Get edges from the jump threading path. Look at the
5830 entire path to see if we thread to a loop exit. If we cancel a jump
5831 thread request, then free the path vector.
5832 (thread_single_edge): Get edges and block types from the jump threading
5833 path. Free the path vector.
5834 (thread_through_loop_header): Get edges and block types from the jump
5835 threading path. Free the path vector.
5836 (mark_threaded_blocks): Iterate over the vector of paths and store
5837 the path on the appropriate edge. Get edges and block types from the
5838 jump threading path.
5839 (mark_threaded_blocks): Get edges and block types from the jump
5840 threading path. Free the path vector.
5841 (thread_through_all_blocks): Use the vector of paths rather than
5842 a vector of 3-edge sets.
5843 (register_jump_thread): Accept pointer to a path vector rather
5844 than the path vector itself. Store the path vector for later use.
5847 2013-10-01 Jakub Jelinek <jakub@redhat.com>
5848 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5851 * config/s390/s390.c (s390_split_branches): Modify check for table
5853 (s390_chunkify_start): Rearrange table jump insn check in order to
5854 deal with compare and branch insns correctly.
5856 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
5860 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
5861 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
5862 define_insn_and_split.
5863 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
5864 (shiftsi3_compare): New pattern.
5866 * config/arm/unspecs.md (UNSPEC_RRX): New.
5868 2013-10-01 Alan Modra <amodra@gmail.com>
5870 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
5871 casing inout operands.
5873 2013-10-01 Richard Biener <rguenther@suse.de>
5875 PR tree-optimization/58553
5876 * tree-loop-distribution.c (struct partition_s): Add niter member.
5877 (classify_partition): Populate niter member for the partition
5878 and properly identify whether the relevant store happens before
5879 or after the loop exit.
5880 (generate_memset_builtin): Use niter member from the partition.
5881 (generate_memcpy_builtin): Likewise.
5883 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
5885 * vec.h (vec_prefix, vec): Prefix member names with "m_".
5886 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
5888 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
5890 * basic-block.h (edge_list): Prefix member names with "m_".
5891 * context.h (context): Likewise.
5892 * domwalk.h (dom_walker): Likewise.
5893 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
5894 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
5895 * hash-table.h (hash_table): Likewise.
5896 * machmode.h (bit_field_mode_iterator): Likewise.
5897 * pass_manager.h (pass_list): Likewise.
5898 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
5899 * tree-pass.h (pass_data): Likewise.
5900 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
5901 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
5902 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
5903 * asan.c (pass_data_asan): Update accordingly.
5904 * cfganal.c (control_dependences::find_control_dependence): Likewise.
5905 (control_dependences::control_dependences): Likewise.
5906 (control_dependences::~control_dependences): Likewise.
5907 (control_dependences::~control_dependences): Likewise.
5908 (control_dependences::get_edges_dependent_on): Likewise.
5909 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
5910 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
5911 * context.c (gcc::context::context): Likewise.
5912 * cprop.c (pass_rtl_cprop::clone): Likewise.
5913 * domwalk.c (dom_walker::walk): Likewise.
5914 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
5915 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
5916 * mode-switching.c (pass_mode_switching::clone): Likewise.
5917 * passes.c (opt_pass::opt_pass): Likewise.
5918 (pass_manager::pass_manager): Likewise.
5919 * predict.c (pass_strip_predict_hints::clone): Likewise.
5920 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
5921 (pass_split_all_insns::clone): Likewise.
5922 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
5924 (bit_field_mode_iterator::next_mode): Likewise.
5925 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
5926 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
5927 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
5928 * tree-complex.c (pass_lower_complex::clone): Likewise.
5929 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
5930 * tree-object-size.c (pass_object_sizes::clone): Likewise.
5931 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
5932 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
5933 (pass_fold_builtins::clone): Likewise.
5934 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
5935 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
5936 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
5937 (pass_cd_dce::clone): Likewise.
5938 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
5939 (pass_phi_only_cprop::clone): Likewise.
5940 * tree-ssa-dse.c (pass_dse::clone): Likewise.
5941 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
5942 * tree-ssa-loop.c (pass_lim::clone): Likewise.
5943 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
5944 * tree-ssa-pre.c (pass_fre::clone): Likewise.
5945 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
5946 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
5947 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
5948 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
5949 * tree-vrp.c (pass_vrp::clone): Likewise.
5950 * tsan.c (pass_tsan::clone): Likewise.
5952 2013-09-30 Jakub Jelinek <jakub@redhat.com>
5955 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
5956 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
5959 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
5960 optimization, punt if sign_bit_p looked through any zero extension.
5962 2013-09-30 Teresa Johnson <tejohnson@google.com>
5964 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
5965 Update redirected out edge count in joiner case.
5966 (ssa_redirect_edges): Common the joiner and non-joiner cases
5967 so that joiner case gets profile updates.
5969 2013-09-30 Richard Biener <rguenther@suse.de>
5971 PR tree-optimization/58554
5972 * tree-loop-distribution.c (classify_partition): Require
5973 unconditionally executed stores for memcpy and memset recognition.
5974 (tree_loop_distribution): Calculate dominance info.
5976 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
5978 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
5979 (NO_PROFILE_COUNTERS): Likewise.
5980 (PROFILE_HOOK): Likewise.
5981 (FUNCTION_PROFILER): Likewise.
5982 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
5984 2013-09-30 Iain Sandoe <iain@codesourcery.com>
5986 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
5987 calls and defines in TARGET_MACHO conditional.
5988 (load_macho_picbase_di): Likewise.
5989 (reload_macho_picbase): Likewise.
5990 (reload_macho_picbase_si): Likewise.
5991 (reload_macho_picbase_di): Likewise.
5992 (nonlocal_goto_receiver): Likewise.
5994 2013-09-30 Nick Clifton <nickc@redhat.com>
5996 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
5997 that use the MSP430X ISA.
5998 (msp430_option_override): Scan -mmcu command line option for any
5999 MCU name that supports the MSP430X ISA.
6000 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
6001 -mmcu options which enable the MSP430X ISA.
6003 2013-09-30 Richard Biener <rguenther@suse.de>
6006 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
6007 before looking for setjmp-like calls.
6009 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6012 * config/darwin-protos.h (machopic_get_function_picbase): New.
6013 * config/darwin.c (machopic_get_function_picbase): New.
6014 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
6015 label for a new func. (load_macho_picbase_di): Likewise.
6016 (reload_macho_picbase): New expand.
6017 (reload_macho_picbase_si): New insn.
6018 (reload_macho_picbase_di): New insn.
6019 (nonlocal_goto_receiver): New define and split.
6020 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
6021 (unspecv enum): Add UNSPECV_NLGR.
6023 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6025 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
6026 that altivec registers are correctly sized on Darwin.
6028 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6030 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
6031 darwin-driver.o): Use COMPILE and POSTCOMPILE.
6032 * config/x-darwin (host-darwin.o): Likewise.
6033 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
6034 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
6035 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
6037 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
6039 * doc/invoke.texi: Fix usage of @tie{} command.
6041 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
6043 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
6045 2013-09-28 Jan Hubicka <jh@suse.cz>
6047 * config/i386/x86-tune.def: Add documentation for each of the options;
6050 2013-09-28 Jan Hubicka <jh@suse.cz>
6052 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
6054 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
6055 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
6056 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
6057 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
6058 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
6060 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
6062 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
6063 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
6064 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
6065 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
6066 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
6067 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
6068 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
6069 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
6070 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
6071 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
6072 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
6073 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
6074 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
6075 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
6076 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
6077 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
6078 gimple.h, godump.c, graphite-clast-to-gimple.c,
6079 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
6080 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
6081 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
6082 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6083 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
6084 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
6085 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
6086 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
6087 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
6088 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
6089 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
6090 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
6091 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
6092 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
6093 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
6094 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
6095 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
6096 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
6097 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
6098 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
6099 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
6100 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
6101 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
6102 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
6103 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6104 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6105 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
6106 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
6107 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
6108 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6109 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6110 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
6111 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
6112 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6113 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6114 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
6115 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
6116 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
6117 whitespace before "(".
6119 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
6121 * expr.h (extract_bit_field): Remove packedp parameter.
6122 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
6123 from forward declaration.
6124 (store_split_bit_field): Remove packedp arg from calls to
6125 extract_fixed_bit_field.
6126 (extract_bit_field_1): Remove packedp parameter and packedp
6127 argument from recursive calls and calls to extract_fixed_bit_field.
6128 (extract_bit_field): Remove packedp parameter and corresponding
6129 arg to extract_bit_field_1.
6130 (extract_fixed_bit_field): Remove packedp parameter. Remove code
6132 (extract_split_bit_field): Remove packedp arg from call to
6133 extract_fixed_bit_field.
6134 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
6135 (copy_blkmode_from_reg): Likewise.
6136 (copy_blkmode_to_reg): Likewise.
6137 (read_complex_part): Likewise.
6138 (store_field): Likewise.
6139 (expand_expr_real_1): Likewise.
6140 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
6141 to extract_bit_field.
6142 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
6143 call to extract_bit_field.
6144 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
6145 call to extract_bit_field.
6146 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
6147 and special packedp behavior from -fstrict-volatile-bitfields
6150 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6152 * lra-eliminations.c (init_elim_table): Guard value_p.
6154 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6156 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
6157 DFmode, DImode, and SFmode in the upper VSX registers based on the
6158 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
6159 if -mpower8-vector. Combine -mvsx-timode handling with the rest
6160 of the VSX register handling.
6162 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
6164 (zero_extendsidi2_lfiwzx): Add support for loading into the
6165 Altivec registers with -mpower8-vector. Use wu/wv constraints to
6166 only do VSX memory options on Altivec registers.
6167 (extendsidi2_lfiwax): Likewise.
6168 (extendsfdf2_fpr): Likewise.
6169 (mov<mode>_hardfloat, SF/SD modes): Likewise.
6170 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
6171 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
6172 (movdi_internal64): Likewise.
6174 2013-09-27 Xinliang David Li <davidxl@google.com>
6176 * opts.c (finish_options): Adjust parameters
6177 according to vect cost model.
6178 (common_handle_option): Set dynamic vect cost
6180 targhooks.c (default_add_stmt_cost): Compute stmt cost
6182 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
6183 Use helper function.
6184 * tree-vectorizer.h (unlimited_cost_model): New function.
6185 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
6186 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
6188 (vect_enhance_data_refs_alignment): Ditto.
6189 * flag-types.h: New enum.
6190 * common/config/i386/i386-common.c (ix86_option_init_struct):
6191 No need to initialize vect_cost_model flag.
6192 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
6195 2013-09-27 Diego Novillo <dnovillo@google.com>
6197 * gimple.h (enum ssa_mode): Remove.
6199 2013-09-27 Paulo Matos <pmatos@broadcom.com>
6201 * cfgloop.h (number_of_loops): Fix typo in check for null.
6203 2013-09-27 Jakub Jelinek <jakub@redhat.com>
6206 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
6207 are children of outermost saved_cfun's loop, and set it up to
6208 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
6209 if loop != loop0 and SESE region contains bbs that belong to loop0.
6211 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6213 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
6214 (binary_scale_code_p, get_base_term, get_index_term): New functions.
6215 (set_address_segment, set_address_base, set_address_index)
6216 (set_address_disp): Accept the argument unconditionally.
6217 (baseness): Remove must_be_base_p and must_be_index_p checks.
6218 (decompose_normal_address): Classify as much as possible in the
6221 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6223 * cse.c (count_reg_usage): Handle INT_LIST.
6224 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
6225 * reginfo.c (reg_scan_mark_refs): Likewise.
6226 * reload1.c (eliminate_regs_1): Likewise.
6228 2013-09-27 Iain Sandoe <iain@codesourcery.com>
6231 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
6234 2013-09-27 Richard Biener <rguenther@suse.de>
6236 PR tree-optimization/58459
6237 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
6238 restriction not propagating into loops.
6240 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
6242 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
6243 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
6244 * doc/tree-ssa.texi (Walking use-def chains): Delete.
6246 2013-09-26 Richard Biener <rguenther@suse.de>
6248 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
6250 2013-09-26 Richard Biener <rguenther@suse.de>
6252 * alias.h (component_uses_parent_alias_set): Rename to ...
6253 (component_uses_parent_alias_set_from): ... this.
6254 * alias.c (component_uses_parent_alias_set): Rename to ...
6255 (component_uses_parent_alias_set_from): ... this and return
6257 (reference_alias_ptr_type_1): Use the result from
6258 component_uses_parent_alias_set_from instead of stripping
6259 components one at a time.
6260 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
6262 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6264 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
6265 Move prototypes to...
6266 * tree-ssa-ter.h: New File. Move prototypes here.
6267 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
6268 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
6269 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
6270 from is_replaceable_p.
6271 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
6272 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
6273 (process_replaceable): Use ter_is_replaceable_p.
6274 (find_replaceable_in_bb): Use ter_is_replaceable_p.
6275 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
6276 newly refactored ssa_is_replaceable_p.
6277 * cfgexpand.c: Include tree-outof-ssa.h.
6278 * ssaexpand.h: Delete.
6280 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6282 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
6283 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
6285 (create_gimple_tmp): Delete.
6286 (get_expr_type, build_assign, build_type_cast): Move to...
6287 * gimple-builder.c: New File.
6288 (get_expr_type): Relocate from gimple.c.
6289 (build_assign, build_type_cast): Change to only create ssanames.
6290 * gimple.h: Move prototypes to...
6291 * gimple-builder.h: New File. Here.
6292 * tree-ssa.h: And here.
6293 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
6294 count_uses_and_derefs): Relocate from gimple.c.
6295 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
6296 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
6297 * tree-ssa-math-opts (execute_cse_reciprocals): Use
6298 gimple_replace_ssa_lhs.
6299 * asan.c: Include gimple-builder.h.
6300 * Makefile.in: Add gimple-builder.o.
6302 2013-09-26 Richard Biener <rguenther@suse.de>
6304 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
6305 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
6306 (loe_visit_block): Use gcc_checking_assert.
6307 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
6308 gimple_assign_ssa_name_copy_p.
6309 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
6312 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
6314 * config/rs6000/t-rs6000 (rs6000.o): Remove.
6315 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
6317 2013-09-26 Richard Biener <rguenther@suse.de>
6319 PR tree-optimization/58539
6320 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
6321 the fact that debug statements are not taking part in loop-closed
6324 2013-09-26 Nick Clifton <nickc@redhat.com>
6326 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
6327 time warning message.
6328 (msp430_print_operand_raw): Delete unused letter parameter.
6329 (TARGET_PRINT_OPERAND_ADDRESS): Define.
6330 (msp430_print_operand_address): New function.
6331 (msp430_print_operand): Move address printing code from here to
6333 * config/msp430/msp430.md (movsipsi2): Add comment in generated
6335 (zero_extendpsisi2): Likewise.
6336 (extendpsisi2): New pattern.
6337 (andneghi3): New pattern.
6339 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
6341 * config/aarch64/aarch64.opt (mlra): New option.
6342 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
6343 (TARGET_LRA_P): Define.
6345 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
6347 * expr.c (expand_assignment): Remove obsolete comment.
6349 2013-09-25 Jeff Law <law@redhat.com>
6351 * tree-flow.h (thread_through_all_blocks): Prototype moved into
6352 tree-ssa-threadupdate.h.
6353 (register_jump_thread): Similarly.
6354 * tree-ssa-threadupdate.h: New header file.
6355 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
6356 * tree-vrp.c: Likewise.
6357 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
6358 (thread_around_empty_blocks): Change type of path vector argument to
6359 an edge,type pair from just an edge. Initialize both elements when
6360 appending to a jump threading path. Tweak references to elements
6362 (thread_across_edge): Similarly. Release memory for the elements
6364 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
6365 (dump_jump_thread_path): New function broken out from
6366 register_jump_thread.
6367 (register_jump_thread): Use dump_jump_thread_path. Change type of
6368 path vector entries. Search the path for NULL edges and dump
6369 the path if one is found. Tweak the conversion of path to 3-edge
6370 form to use the block copy type information embedded in the path.
6372 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6374 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
6376 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6377 Vladimir Makarov <vmakarov@redhat.com>
6379 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
6380 from the least significant bit.
6381 (strip_address_mutations): Add bitfield operations handling.
6382 (must_be_index_p): Add shifting and rotate operations handling.
6383 (set_address_base): Use must_be_base_p predicate.
6384 (set_address_index): Use must_be_index_p predicate.
6386 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
6387 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6388 Sergey Lega <sergey.s.lega@intel.com>
6389 Anna Tikhonova <anna.tikhonova@intel.com>
6390 Ilya Tocar <ilya.tocar@intel.com>
6391 Andrey Turetskiy <andrey.turetskiy@intel.com>
6392 Ilya Verbin <ilya.verbin@intel.com>
6393 Kirill Yukhin <kirill.yukhin@intel.com>
6394 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6396 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
6398 (ix86_expand_vector_move_misalign): Support new unaligned load and
6399 stores and use new names.
6400 (CODE_FOR_sse2_storedqu): Rename to ...
6401 (CODE_FOR_sse2_storedquv16qi): ... this.
6402 (CODE_FOR_sse2_loaddqu): Rename to ...
6403 (CODE_FOR_sse2_loaddquv16qi): ... this.
6404 (CODE_FOR_avx_loaddqu256): Rename to ...
6405 (CODE_FOR_avx_loaddquv32qi): ... this.
6406 (CODE_FOR_avx_storedqu256): Rename to ...
6407 (CODE_FOR_avx_storedquv32qi): ... this.
6408 * config/i386/i386.md (fpint_logic): New.
6409 * config/i386/sse.md (VMOVE): Extend for AVX512.
6413 (VI_UNALIGNED_LOADSTORE): Ditto.
6414 (sse2_avx_avx512f): Ditto.
6415 (sse2_avx2): Extend for AVX512.
6416 (sse4_1_avx2): Ditto.
6417 (avx2_avx512f): New.
6418 (sse): Extend for AVX512.
6421 (avxsizesuffix): Ditto.
6422 (sseintvecmode): Ditto.
6424 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
6425 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
6426 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6427 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
6428 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6429 (<sse2_avx_avx512f>_storedqu<mode): ... this.
6430 (<sse>_movnt<mode>): Replace constraint "x" with "v".
6431 (STORENT_MODE): Extend for AVX512.
6432 (*absneg<mode>2): Replace constraint "x" with "v".
6433 (*mul<mode>3): Ditto.
6434 (*ieee_smin<mode>3): Ditto.
6435 (*ieee_smax<mode>3): Ditto.
6436 (avx_cmp<mode>3): Replace VF with VF_128_256.
6437 (*<sse>_maskcmp<mode>3_comm): Ditto.
6438 (<sse>_maskcmp<mode>3): Ditto.
6439 (<sse>_andnot<mode>3): Extend for AVX512.
6440 (<code><mode>3, anylogic): Replace VF with VF_128_256.
6441 (<code><mode>3, fpint_logic): New.
6442 (*<code><mode>3): Extend for AVX512.
6443 (avx512flogicsuff): New.
6444 (avx512f_<logic><mode>): Ditto.
6445 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
6447 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
6448 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
6449 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
6450 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
6451 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
6452 (xop_vpermil2<mode>3): Ditto.
6453 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
6454 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
6455 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
6456 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
6458 2013-09-25 Tom Tromey <tromey@redhat.com>
6460 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
6461 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
6462 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
6463 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
6464 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
6465 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
6466 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
6467 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
6468 (GRAPHITE_HTAB_H): Remove.
6470 2013-09-25 Tom Tromey <tromey@redhat.com>
6472 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
6474 2013-09-25 Tom Tromey <tromey@redhat.com>
6476 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
6478 2013-09-25 Tom Tromey <tromey@redhat.com>
6480 * config/i386/t-i386 (i386.o): Remove.
6481 (i386-c.o): Use COMPILE and POSTCOMPILE.
6483 2013-09-25 Tom Tromey <tromey@redhat.com>
6485 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
6487 2013-09-25 Tom Tromey <tromey@redhat.com>
6489 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
6490 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
6491 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
6492 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
6493 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
6494 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
6495 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
6496 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
6497 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
6498 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
6499 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
6500 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
6501 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
6502 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
6503 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
6504 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
6505 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
6506 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
6507 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
6508 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
6509 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
6510 (tree-ssa-pre.o, tree-ssa-sccvn.o)
6511 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
6512 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
6513 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
6514 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
6515 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
6516 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
6517 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
6518 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
6519 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
6520 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
6521 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
6522 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
6523 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
6524 (graphite-clast-to-gimple.o, graphite-dependences.o)
6525 (graphite-interchange.o, graphite-poly.o)
6526 (graphite-scop-detection.o, graphite-sese-to-poly.o)
6527 (graphite-optimize-isl.o, tree-vect-loop.o)
6528 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
6529 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
6530 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
6531 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
6532 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
6533 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
6534 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
6535 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
6536 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
6537 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
6538 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
6539 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
6540 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
6541 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
6542 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
6543 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
6544 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
6545 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
6546 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
6547 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
6548 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
6549 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
6550 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
6551 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
6552 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
6553 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
6554 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
6555 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
6556 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
6557 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
6558 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
6559 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
6560 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
6561 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
6562 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
6563 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
6564 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
6565 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
6566 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
6567 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
6568 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
6569 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
6570 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
6571 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
6572 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
6573 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
6574 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
6575 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
6576 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
6577 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
6578 (gcov-dump.o): Remove.
6579 (default-c.o): Use COMPILE and POSTCOMPILE.
6580 (CFLAGS-gcc.o): New variable.
6581 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
6583 2013-09-25 Tom Tromey <tromey@redhat.com>
6585 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
6586 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
6587 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
6588 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
6589 (c-family/c-pragma.o, c-family/c-pretty-print.o)
6590 (c-family/c-semantics.o, c-family/c-ada-spec.o)
6591 (c-family/array-notation-common.o, c-family/stub-objc.o)
6592 (c-family/c-ubsan.o): Remove.
6594 2013-09-25 Tom Tromey <tromey@redhat.com>
6596 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
6598 2013-09-25 Tom Tromey <tromey@redhat.com>
6600 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
6601 to add -DENABLE_SHARED_LIBGCC.
6602 (gcc.o): Don't use subshell.
6604 2013-09-25 Tom Tromey <tromey@redhat.com>
6606 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
6607 * configure.ac: Don't invoke AM_PROG_CC_C_O.
6608 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
6609 * configure, config.in: Rebuild.
6611 2013-09-25 Tom Tromey <tromey@redhat.com>
6613 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
6614 (COMPILE, POSTCOMPILE): New variables.
6615 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
6616 (DEPFILES): New variable.
6617 Include ".Po" files.
6618 * configure.ac: Add checks for dependency checking.
6619 * configure, aclocal.m4: Regenerate.
6621 2013-09-25 Tom Tromey <tromey@redhat.com>
6623 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
6624 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
6626 2013-09-25 Tom Tromey <tromey@redhat.com>
6628 * Makefile.in (generated_files): Add options.h,
6629 target-hooks-def.h, insn-opinit.h,
6630 common/common-target-hooks-def.h, pass-instances.def,
6631 c-family/c-target-hooks-def.h.
6633 2013-09-25 Jeff Law <law@redhat.com>
6635 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
6636 than foo[foo.length () - 1] to access last member in a vec.
6637 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
6639 2013-09-25 Richard Biener <rguenther@suse.de>
6642 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
6644 2013-09-25 Jan Hubicka <jh@suse.cz>
6646 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
6649 2013-09-25 Marek Polacek <polacek@redhat.com>
6652 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
6653 when determining the type name.
6655 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
6657 * config/sh/sh.md: Fix formatting.
6659 2013-09-24 Xinliang David Li <davidxl@google.com>
6661 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
6662 max peel iterations parameter.
6663 * param.def: New parameter.
6664 * doc/invoke.texi: Document New parameter.
6666 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
6668 * gimple-pretty-print.c: Various whitespace tweaks.
6669 * tree-core.h: Likewise.
6670 * tree-pretty-print.c: Likewise.
6671 * tree-ssa-alias.c: Likewise.
6672 * tree-ssa-copy.c: Likewise.
6673 * tree-ssanames.c: Likewise.
6674 * tree-ssanames.h: Likewise.
6675 * tree-vrp.c: Likewise.
6677 2013-09-24 Alan Modra <amodra@gmail.com>
6681 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
6682 for output operands that disallow regs. Don't use EXPAND_WRITE on
6685 2013-09-24 Richard Biener <rguenther@suse.de>
6688 * tree.c (reference_alias_ptr_type): Move ...
6689 * alias.c (reference_alias_ptr_type): ... here and implement
6690 in terms of the new reference_alias_ptr_type_1.
6691 (ref_all_alias_ptr_type_p): New helper.
6692 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
6693 use ref_all_alias_ptr_type_p.
6694 (get_deref_alias_set): Add flag_strict_aliasing check here.
6695 (reference_alias_ptr_type_1): New function, split out from ...
6696 (get_alias_set): ... here.
6697 (alias_ptr_types_compatible_p): New function.
6698 * alias.h (reference_alias_ptr_type): Declare.
6699 (alias_ptr_types_compatible_p): Likewise.
6700 * tree.h (reference_alias_ptr_type): Remove.
6701 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
6702 to compare MEM_REF alias types.
6704 2013-09-24 Richard Biener <rguenther@suse.de>
6706 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
6708 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6710 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
6711 reload helper function arrays into a single array reg_addr.
6712 (reload_fpr_gpr): Likewise.
6713 (reload_gpr_vsx): Likewise.
6714 (reload_vsx_gpr): Likewise.
6715 (struct rs6000_reg_addr): Likewise.
6716 (reg_addr): Likewise.
6717 (rs6000_debug_reg_global): Change rs6000_vector_reload,
6718 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
6719 (rs6000_init_hard_regno_mode_ok): Likewise.
6720 (rs6000_secondary_reload_direct_move): Likewise.
6721 (rs6000_secondary_reload): Likewise.
6723 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
6724 constraints: wu, ww, and wy. Repurpose wv constraint added during
6725 power8 changes. Put wg constraint in alphabetical order.
6727 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
6728 for future work to add ISA 2.07 VSX single precision support.
6729 (-mvsx-scalar-double): Change default from -1 to 1, update
6730 documentation comment.
6731 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
6732 (-mupper-regs-df): New debug switch to control whether DF values
6733 can go in the traditional Altivec registers.
6734 (-mupper-regs-sf): New debug switch to control whether SF values
6735 can go in the traditional Altivec registers.
6737 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
6739 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
6740 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
6741 Add new constraints, wu/ww/wy. Repurpose wv constraint.
6742 (rs6000_debug_legitimate_address_p): Print if we are running
6743 before, during, or after reload.
6744 (rs6000_secondary_reload): Add a comment.
6745 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
6747 * config/rs6000/constraints.md (wa constraint): Sort w<x>
6748 constraints. Update documentation string.
6749 (wd constraint): Likewise.
6750 (wf constraint): Likewise.
6751 (wg constraint): Likewise.
6752 (wn constraint): Likewise.
6753 (ws constraint): Likewise.
6754 (wt constraint): Likewise.
6755 (wx constraint): Likewise.
6756 (wz constraint): Likewise.
6757 (wu constraint): New constraint for ISA 2.07 SFmode scalar
6759 (ww constraint): Likewise.
6760 (wy constraint): Likewise.
6761 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
6762 in the previous submissions.
6763 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
6765 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
6767 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
6768 (REG_BR_PROB): Say that the probability is stored in an int_list.
6769 * reg-notes.def: Update commentary to mention INT_LIST.
6770 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
6771 (INT_LIST): New rtx.
6772 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
6773 * rtlanal.c (int_reg_note_p): New function.
6774 (alloc_reg_note): Assert that the note does not have an int argument.
6775 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
6776 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
6777 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
6778 rather than an INSN_LIST. Handle INT_LIST.
6779 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
6780 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
6781 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
6782 Manipulate them as ints rather than rtxes.
6783 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
6784 * regmove.c (copy_src_to_dest): Likewise.
6785 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
6787 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
6788 into the cases that need it.
6789 * config/arm/arm.c (arm_unwind_emit): Likewise.
6791 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
6792 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
6793 * loop-doloop.c (add_test, doloop_modify): Likewise.
6794 * loop-unswitch.c (compare_and_jump_seq): Likewise.
6795 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
6796 * predict.c (combine_predictions_for_insn): Likewise.
6797 * print-rtl.c (print_rtx): Handle INT_LIST.
6798 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
6799 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
6800 * config/arm/arm.c (emit_unlikely_jump): Likewise.
6801 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
6802 (ix86_split_fp_branch, predict_jump): Likewise.
6803 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
6804 * config/sh/sh.c (expand_cbranchsi4): Likewise.
6805 * config/spu/spu.c (ea_load_store_inline): Likewise.
6807 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
6808 value of a REG_BR_PROB note.
6809 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
6810 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
6811 * emit-rtl.c (try_split): Likewise.
6812 * predict.c (br_prob_note_reliable_p): Likewise.
6813 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
6814 * reorg.c (mostly_true_jump): Likewise.
6815 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
6816 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
6817 * config/i386/i386.c (ix86_print_operand): Likewise.
6818 * config/ia64/ia64.c (ia64_print_operand): Likewise.
6819 * config/mmix/mmix.c (mmix_print_operand): Likewise.
6820 * config/rs6000/rs6000.c (output_cbranch): Likewise.
6821 * config/s390/s390.c (s390_expand_tbegin): Likewise.
6822 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
6823 * config/sparc/sparc.c (output_cbranch): Likewise.
6824 * config/spu/spu.c (get_branch_target): Likewise.
6825 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
6826 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
6828 2013-09-23 Jan Hubicka <jh@suse.cz>
6830 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
6832 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
6833 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
6835 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
6836 (gimple_fold_call): Dump inconsistent devirtualizations; add
6837 sanity check for type based devirtualizations.
6838 * ipa-prop.c: Include ipa-utils.h
6839 (ipa_intraprocedural_devirtualization): Add sanity check.
6840 (try_make_edge_direct_virtual_call): Likewise.
6842 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
6844 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
6845 assignment statements.
6847 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
6849 * gimple-pretty-print.c (dump_ssaname_info): New function.
6850 (dump_gimple_phi): Call it.
6851 (pp_gimple_stmt_1): Likewise.
6852 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
6853 (range_info_def): Declare.
6854 * tree-pretty-print.c (pp_double_int): New function.
6855 (dump_generic_node): Call it.
6856 * tree-pretty-print.h (pp_double_int): Declare.
6857 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
6858 * tree-ssanames.h (range_info_def): New structure.
6859 (value_range_type): Move definition here.
6860 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
6862 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
6864 (set_range_info): New function.
6865 (get_range_info): Likewise.
6866 (duplicate_ssa_name_range_info): Likewise.
6867 (duplicate_ssa_name_fn): Check pointer type and call
6868 duplicate_ssa_name_range_info.
6869 * tree-ssa-copy.c (fini_copy_prop): Likewise.
6870 * tree-vrp.c (value_range_type): Remove definition, now in
6872 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
6873 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
6874 (SSA_NAME_RANGE_INFO): New macro.
6876 2013-09-23 Richard Biener <rguenther@suse.de>
6878 PR tree-optimization/58464
6879 * tree-ssa-pre.c (phi_trans_lookup): Remove.
6880 (phi_trans_add): Change to add conditionally on being not present.
6881 (phi_translate_1): Remove recursion detection here.
6882 (phi_translate): Pre-seed the cache with NULL to catch
6883 recursion here in a more generic way.
6884 (bitmap_find_leader): Adjust comment.
6885 (get_representative_for): Dump value-numbers.
6886 (create_expression_by_pieces): Likewise.
6887 (insert_into_preds_of_block): Likewise.
6889 2013-09-23 Christian Bruel <christian.bruel@st.com>
6892 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
6893 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
6895 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
6898 2013-09-20 Renlin Li <renlin.li@arm.com>
6900 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
6902 (aarch64_expand_epilogue): Likewise.
6903 (aarch64_legitimize_reload_address): Likewise.
6905 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
6907 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
6908 NULL_TREE before pushing them onto the vector. Likewise for labels.
6910 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
6912 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
6913 comparison operators when -fno-trapping-math is in effect.
6914 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
6915 comparison operators in TFmode and assert that unsupported operators
6917 (ia64_print_operand): Likewise.
6919 2013-09-21 Jan Hubicka <jh@suse.cz>
6921 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
6922 (sse_typeless_stores): Enable for core
6923 (sse_load0_by_pxor): Likewise.
6924 (four_jump_limit): Disable for core.
6925 (pad_returns): Likewise.
6926 (avoid_vector_decode): Likewise.
6927 (fuse_cmp_and_branch): Enable for cores.
6928 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
6930 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
6933 * config/pa/pa.c (pa_option_override): Disable auto increment and
6934 decrement instructions until reload is completed.
6936 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
6937 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
6938 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
6940 2013-09-20 DJ Delorie <dj@redhat.com>
6941 Nick Clifton <nickc@redhat.com>
6943 * config/rl78/rl78.c: Various whitespace and comment tweaks.
6944 (need_to_save): Save bank 0 on interrupts.
6945 (characterize_address): Strip far address wrappers.
6946 (rl78_as_legitimate_address): Likewise.
6947 (transcode_memory_rtx): Likewise.
6948 (rl78_peep_movhi_p): Disable this peephole after devirt.
6949 (rl78_propogate_register_origins): Forget all origins when a
6951 * config/rl78/rl78-virt.md: Various whitespace tweaks.
6952 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
6954 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
6955 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
6956 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
6958 * config/rl78/constraints.md: Various whitespace and paren tweaks.
6960 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
6962 * config/pa/pa.md: In "scc" insn patterns, change output template to
6963 handle const0_rtx in reg_or_0_operand operands.
6965 2013-09-20 Martin Husemann <martin@NetBSD.org>
6968 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
6969 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
6971 2013-09-20 Richard Biener <rguenther@suse.de>
6974 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
6975 remembering SSA name version and block index.
6976 (new_scev_info_str): Adjust.
6977 (hash_scev_info): Likewise. Also hash the block index.
6978 (eq_scev_info): Adjust.
6979 (find_var_scev_info): Likewise.
6980 (struct instantiate_cache_entry): Remove.
6981 (struct instantiate_cache_type): Use a htab to map name, block
6983 (instantiate_cache_type::~instantiate_cache_type): Adjust.
6984 (get_instantiated_value_entry): Likewise.
6985 (hash_idx_scev_info, eq_idx_scev_info): New functions.
6986 (instantiate_scev_name): Adjust.
6988 2013-09-20 Jeff Law <law@redhat.com>
6990 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
6992 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
6994 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
6995 Call aarch64_simd_expand_args to update op[argc].
6997 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
6999 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
7002 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7004 * gengtype.c (file_rules): Added rule for *.cc files.
7005 (get_output_file_with_visibility): Give fatal message when no
7008 2013-09-20 Renlin Li <renlin.li@arm.com>
7010 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
7011 (aarch64_expand_epilogue): Likewise.
7012 (aarch64_legitimize_reload_address): Likewise.
7014 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
7017 * expr.c (expand_assignment): Remove misalignp code path.
7019 2013-09-20 Marek Polacek <polacek@redhat.com>
7022 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
7023 TYPE_PRECISION. Add asserts.
7025 2013-09-20 Richard Biener <rguenther@suse.de>
7027 PR tree-optimization/58453
7028 * tree-loop-distribution.c (distribute_loop): Apply the cost
7029 model for -ftree-loop-distribute-patterns, too.
7031 2013-09-20 Richard Biener <rguenther@suse.de>
7034 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
7035 make type comparison less strict.
7037 2013-09-20 Alan Modra <amodra@gmail.com>
7039 * configure: Regenerate.
7040 * aclocal.m4: Regenerate.
7042 2013-09-20 Marek Polacek <polacek@redhat.com>
7045 * doc/extend.texi: Document that attribute used is meant to be used
7046 on variables with static storage duration.
7048 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7050 PR tree-optimization/58472
7051 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
7052 simd_lane_access set inv_p = false.
7053 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
7054 the simduid magic VAR_DECL.
7056 2013-09-19 Jan Hubicka <jh@suse.cz>
7058 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
7060 2013-09-17 Jeff Law <law@redhat.com>
7062 * tree-ssa-dom.c (record_temporary_equivalences): New function
7063 split out of dom_opt_dom_walker::after_dom_children.
7064 (dom_opt_dom_walker::thread_across_edge): Move common code
7065 in here from dom_opt_dom_walker::after_dom_children.
7066 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
7068 2013-09-19 Jan Hubicka <jh@suse.cz>
7070 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
7071 (TARGET_GENERIC): Use PROCESOR_GENERIC
7072 (enum processor_type): Unify generic32 and 64.
7073 * i386.md (cpu): Likewise.
7074 * x86-tune.def (use_leave): Enable for generic32.
7075 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
7076 * athlon.md: Change generic64 to generic in all occurences.
7077 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
7078 (ix86_target_macros_internal): Likewise.
7079 * driver-i386.c (host_detect_local_cpu): Likewise.
7080 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
7082 (generic_memcpy, generic_memset, generic_cost): This one.
7083 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
7084 (m_GENERIC32, m_GENERIC64): Remove.
7085 (m_GENERIC): Turn into one flag.
7086 (processor_target): Unify generic tunnings.
7087 (ix86_option_override_internal): Replace generic32/64 by generic.
7088 (ix86_issue_rate): Likewise.
7089 (ix86_adjust_cost): Likewise.
7091 2013-09-19 Jan Hubicka <jh@suse.cz>
7093 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
7094 of speculative flag.
7096 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7098 * omp-low.c (expand_omp_sections): Always pass len - 1 to
7099 GOMP_sections_start, even if !exit_reachable.
7101 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7103 * lra-constraints.c (need_for_all_save_p): Use macro
7104 HARD_REGNO_CALL_PART_CLOBBERED.
7105 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
7106 set up pseudo conflict hard regs.
7108 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7111 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
7114 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7116 PR rtl-optimization/58438
7117 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
7118 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
7119 from previous subpasses.
7121 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
7123 * arm.c (arm_get_frame_offsets): Validate architecture supports
7124 LDRD/STRD before accepting the tuning preference.
7125 (arm_expand_prologue): Likewise.
7126 (arm_expand_epilogue): Likewise.
7128 2013-09-18 Richard Biener <rguenther@suse.de>
7130 PR tree-optimization/58417
7131 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
7132 have chrecs with symbols defined in the loop as operands.
7133 (chrec_fold_multiply): Likewise.
7134 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
7135 parameters before folding binary operations.
7136 (struct instantiate_cache_entry_hasher): Remove.
7137 (struct instantiate_cache_type): Use a pointer-map.
7138 (instantiate_cache_type::instantiate_cache_type): New function.
7139 (instantiate_cache_type::get): Likewise.
7140 (instantiate_cache_type::set): Likewise.
7141 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7142 (get_instantiated_value_entry): Likewise.
7143 (global_cache): New global.
7144 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
7145 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
7146 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
7147 (instantiate_scev_name): Adjust.
7148 (instantiate_scev): Construct global instead of local cache.
7149 (resolve_mixers): Likewise.
7151 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
7152 Paolo Carlini <paolo.carlini@oracle.com>
7155 * doc/implement-cxx.texi: Fix references to the C++ standards.
7157 2013-09-18 Jakub Jelinek <jakub@redhat.com>
7159 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
7160 * tree-vect-data-refs.c (vect_analyze_data_refs): For
7161 simd_lane_access drs, update also DR_ALIGNED_TO.
7163 2013-09-18 Marek Polacek <polacek@redhat.com>
7166 * doc/extend.texi: Document no_sanitize_undefined attribute.
7167 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
7168 no_sanitize_undefined attribute.
7170 2013-09-18 Nick Clifton <nickc@redhat.com>
7172 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
7173 (ASM_DECLARE_FUNCTION_NAME): Define.
7175 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
7177 * compare-elim.c (find_comparison_dom_walker): New class
7178 (find_comparisons_in_bb): Rename to
7179 find_comparison_dom_walker::before_dom_children
7180 (find_comparisons): Adjust
7181 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
7183 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
7184 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
7185 (init_walk_dominator_tree): Remove declaration.
7186 (fini_walk_dominator_tree): Remove declaration.
7187 * fwprop.c (single_def_use_dom_walker): New class
7188 (single_def_use_enter_block): Convert to
7189 single_def_use_dom_walker::before_dom_children.
7190 (single_def_use_leave_block): Convert to
7191 single_def_use_dom_walker::after_dom_children.
7192 (build_single_def_use_links): Adjust.
7193 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
7195 (find_candidates_in_block): Convert to
7196 find_candidates_dom_walker::before_dom_children.
7197 (execute_strength_reduction): Adjust.
7198 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
7199 (sese_dom_walker): New class.
7200 (sese_dom_walker::sese_dom_walker): New constructor.
7201 (sese_dom_walker::~sese_dom_walker): New destructor.
7202 (build_sese_conditions_before): Convert to
7203 sese_dom_walker::before_dom_children.
7204 (build_sese_conditions_after): Convert to
7205 sese_dom_walker::after_dom_children.
7206 (build_poly_scop): Adjust
7207 * tree-into-ssa.c (rewrite_dom_walker): New class
7208 (rewrite_enter_block): Convert to
7209 rewrite_dom_walker::before_dom_children.
7210 (rewrite_leave_block): Convert to
7211 rewrite_dom_walker::after_dom_children.
7212 (rewrite_update_dom_walker): New class.
7213 (rewrite_update_enter_block): Convert to
7214 rewrite_update_dom_walker::before_dom_children.
7215 (rewrite_update_leave_block): Convert to
7216 rewrite_update_dom_walker::after_dom_children.
7217 (rewrite_blocks, rewrite_into_ssa): Adjust.
7218 (mark_def_dom_walker): New class.
7219 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
7220 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
7221 (mark_def_sites_blocks): Convert to
7222 mark_def_dom_walker::before_dom_children.
7223 (mark_def_site_blocks): Remove.
7224 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
7225 (tree_ssa_dominator_optimize): Adjust.
7226 (dom_thread_across_edge): Convert to method
7227 dom_opt_dom_walker::thread_across_edge.
7228 (dom_opt_enter_block): Convert to member function
7229 dom_opt_dom_walker::before_dom_children.
7230 (dom_opt_leave_block): Convert to member function
7231 dom_opt_dom_walker::after_dom_children.
7232 * tree-ssa-dse.c (dse_dom_walker): New class.
7233 (dse_enter_block): Convert to member function
7234 dse_dom_walker::before_dom_children.
7235 (tree_ssa_dse): Adjust.
7236 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
7237 (determine_invariantness_stmt): Convert to method
7238 invariantness_dom_walker::before_dom_children.
7239 (determine_invariantness): Remove
7240 (move_computations_dom_walker): New class.
7241 (move_computations_stmt): Convert to method
7242 move_computations_dom_walker::before_dom_children.
7243 (move_computations, tree_ssa_lim): Adjust.
7244 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
7245 (nt_init_block): Convert to method
7246 notrappping_dom_walker::before_dom_children.
7247 (nt_fini_block): Convert to method
7248 method nontrapping_dom_walker::after_dom_children.
7249 (get_non_trapping): Adjust.
7250 * tree-ssa-pre.c (eliminate_dom_walker): New class.
7251 (eliminate_bb): Convert to method
7252 eliminate_dom_walker::before_dom_children.
7253 (eliminate_leave_block): Convert to method
7254 eliminate_dom_walker::after_dom_children.
7255 (eliminate): Adjust.
7256 * tree-ssa-strlen.c (strlen_dom_walker): New class.
7257 (strlen_enter_block): Convert to method
7258 strlen_dom_walker::before_dom_children.
7259 (strlen_leave_block): Convert to method
7260 method strlen_dom_walker::after_dom_children.
7261 (tree_ssa_strlen): Adjust.
7262 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
7263 (tree_ssa_uncprop): Adjust.
7264 (uncprop_leave_block): Convert to method
7265 uncprop_dom_walker::after_dom_children.
7266 (uncprop_leave_block): Convert to method
7267 uncprop_dom_walker::before_dom_children.
7269 2013-09-18 Bin Cheng <bin.cheng@arm.com>
7271 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
7272 branch in same basic block. Check both src and dest of the move insn.
7274 2013-09-17 Nick Clifton <nickc@redhat.com>
7276 * config/rl78/rl78-real.md (bf): New pattern.
7278 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
7279 (rl78_print_operand): Do not put a # before a %B.
7280 * config/rl78/rl78.opt: Tweak doc strings.
7282 2013-09-17 DJ Delorie <dj@redhat.com>
7284 * config/rl78/constraints.md (Wcv): Allow up to $r31.
7285 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
7286 (rl78_option_override): Likewise, if -mallregs.
7287 (is_virtual_register): Likewise.
7288 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
7289 (REGNO_OK_FOR_BASE_P): Likewise.
7290 * config/rl78/rl78.opt (-mallregs): New.
7292 2013-09-17 Nick Clifton <nickc@redhat.com>
7294 * config/rl78/rl78.c (need_to_save): Change return type to bool.
7295 For interrupt functions: save all call clobbered registers if the
7296 interrupt handler is not a leaf function.
7297 (rl78_expand_prologue): Always recompute the frame information.
7298 For interrupt functions: only select bank 0 if one of the bank 0
7299 registers is going to be psuhed.
7301 2013-09-17 DJ Delorie <dj@redhat.com>
7303 * config/rl78/constraints.md: For each W* constraint, rename to C*
7304 and create a W* constraint that checks for an optional ES: prefix
7306 * config/rl78/rl78.md (UNS_ES_ADDR): New.
7307 (es_addr): New. Used to wrap far addresses.
7308 * config/rl78/rl78-protos.h (rl78_es_addr): New.
7309 (rl78_es_base): New.
7310 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
7311 wrapped far addresses.
7312 (rl78_print_operand_1): Unwrap far addresses before processing.
7313 (rl78_lo16): Wrap far addresses in unspecs.
7314 (rl78_es_addr): New.
7315 (rl78_es_base): New.
7316 (insn_ok_now): Check for not-yet-wrapped far addresses.
7317 (transcode_memory_rtx): Properly re-wrap far addresses.
7319 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
7321 * config/sparc/t-rtems: Add leon3 multilibs.
7323 2013-09-17 Cong Hou <congh@google.com>
7325 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
7326 when checking the dot production pattern. The type of rhs operand
7327 of multiply is now checked correctly.
7329 2013-09-17 Jeff Law <law@redhat.com>
7331 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
7332 edge implied equivalences into successor phis.
7333 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
7334 here from tree-ssa-threadedge.c.
7335 (mark_threaded_blocks): When threading through a joiner, if both
7336 successors of the joiner's clone reach the same block, verify the
7337 PHI arguments are equal. If not, cancel the jump threading request.
7338 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
7339 tree-ssa-threadupdate.c
7340 (thread_across_edge): Don't check PHI argument equality when
7341 threading through joiner block here.
7343 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
7345 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
7346 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
7347 (warn_uninit, warn_uninitialized_vars,
7348 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7349 Move to tree-ssa-uninit.c.
7350 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
7351 (has_undefined_value_p): New. Pass dependant parts of
7352 ssa_undefined_value_p.
7353 (uninit_undefined_value_p): Use has_undefined_value_p.
7354 (warn_uninit, warn_uninitialized_vars,
7355 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7356 Move from tree-ssa.c.
7357 * tree-ssa.h: Adjust prototypes.
7359 2013-09-17 Jan Hubicka <jh@suse.cz>
7362 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
7363 * ipa-inline.c (can_inline_edge_p): Do not downgrade
7364 FUNCTION_NOT_OPTIMIZED.
7365 * ipa-inline-analysis.c (compute_inline_parameters): Function
7366 not optimized is not inlinable unless it is alwaysinline.
7367 (inline_analyze_function): Force calls in not optimized
7368 function not inlinable.
7370 2013-09-17 Jan Hubicka <jh@suse.cz>
7373 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
7375 * ipa.c (function_and_variable_visibility): Likewise.
7376 * ipa-profile.c (ipa_profile): Likewise.
7378 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
7381 * cgraph.c (cgraph_function_body_availability): Check for ifunc
7382 attribute, and don't inline the resolver in this case.
7384 2013-09-17 Teresa Johnson <tejohnson@google.com>
7386 * coverage.c (get_coverage_counts): Add missing newline.
7388 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7390 PR tree-optimization/58088
7391 * fold-const.c (mask_with_trailing_zeros): New function.
7392 (fold_binary_loc): Make sure we don't recurse infinitely
7393 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
7394 Use mask_with_trailing_zeros where appropriate.
7396 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
7398 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
7399 of 'prev' var to get better distance estimation.
7401 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
7403 * tree-inline.h (struct copy_body_data): Add transform_parameter.
7404 * tree-inline.c (is_parameter_of): New predicate.
7405 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
7406 a parameter has been remapped.
7407 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
7408 (optimize_inline_calls): Initialize transform_parameter.
7409 (copy_gimple_seq_and_replace_locals): Likewise.
7410 (tree_function_versioning): Likewise.
7411 (maybe_inline_call_in_expr): Likewise.
7413 2013-09-17 Nick Clifton <nickc@redhat.com>
7415 * config/msp430/msp430-protos.h: Add prototypes for new functions.
7416 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
7418 (is_attr_func): New function.
7419 (msp430_is_interrupt_func): New function.
7420 (is_naked_func): New function.
7421 (is_reentrant_func): New function.
7422 (is_critical_func): New function.
7423 (msp430_start_function): Add annotations for function attributes.
7424 (msp430_attr): New function.
7425 (msp430_attribute_table): New.
7426 (msp430_function_section): New function.
7427 (TARGET_ASM_FUNCTION_SECTION): Define.
7428 (msp430_builtin): New enum.
7429 (msp430_init_builtins): New function.
7430 (msp430_builtin_devl): New function.
7431 (msp430_expand_builtin): New function.
7432 (TARGET_INIT_BUILTINS): Define.
7433 (TARGET_EXPAND_BUILTINS): Define.
7434 (TARGET_BUILTIN_DECL): Define.
7435 (msp430_expand_prologue): Add support for naked, interrupt,
7436 critical and reentrant functions.
7437 (msp430_expand_epilogue): Likewise.
7438 (msp430_print_operand): Handle 'O' character.
7439 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
7441 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
7442 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
7443 (pushm): Use a 'n' rather than an 'i' constraint.
7444 (msp_return): Add generation of the interrupt return instruction.
7445 (disable_interrupts): New pattern.
7446 (enable_interrupts): New pattern.
7447 (push_intr_state): New pattern.
7448 (pop_intr_state): New pattern.
7449 (bic_SR): New pattern.
7450 (bis_SR): New pattern.
7451 * doc/extend.texi: Document MSP430 function attributes and builtin
7454 2013-09-17 Richard Biener <rguenther@suse.de>
7456 PR tree-optimization/58432
7457 * tree-loop-distribution.c (tree_loop_distribution): Also
7458 scan PHIs for outside loop uses and seed a partition from them.
7460 2013-09-17 Bin Cheng <bin.cheng@arm.com>
7462 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
7463 (restructure_reference): Call backtrace_base_for_ref.
7465 2013-09-17 Alan Modra <amodra@gmail.com>
7468 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
7471 2013-09-16 DJ Delorie <dj@redhat.com>
7473 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
7474 vregs location for RL78/G10.
7475 (rl78_expand_prologue): Avoid SEL on G10.
7476 (rl78_expand_epilogue): Likewise.
7477 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
7478 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
7479 __RL78_G10__ when appropriate.
7480 (ASM_SPEC): Pass -mg10 along to the assembler.
7481 * config/rl78/rl78.md (sel_rb): Disable for G10.
7482 * config/rl78/rl78.opt: Add -mg10 option.
7483 * config/rl78/t-rl78: Add -mg10 multilib.
7485 2013-09-16 Xinliang David Li <davidxl@google.com>
7487 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
7488 * omp-low.c (omp_max_vf): Ditto.
7489 (expand_omp_simd): Ditto.
7490 * tree-vectorizer.c (vectorize_loops): Ditto.
7491 (gate_vect_slp): Ditto.
7492 (gate_increase_alignment): Ditto.
7493 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
7494 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
7495 (gate_tree_vectorize): Name change.
7496 (tree_vectorize): Ditto.
7497 (pass_vectorize::gate): Call new function.
7498 (pass_vectorize::execute): Ditto.
7499 * opts.c: O3 default setting change.
7500 (finish_options): Check new flag.
7501 * doc/invoke.texi: Document new flags.
7502 * common.opt: New flags.
7504 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
7506 * doc/tm.texi.in (Cond Exec Macros): Remove node.
7507 (Condition Code): Don't reference it.
7508 * doc/tm.texi: Regenerate.
7510 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
7513 * lra-constraints.c (undo_optional_reloads): Consider all optional
7514 reload even if it did not get a hard reg.
7516 2013-09-16 Teresa Johnson <tejohnson@google.com>
7518 * dumpfile.c (dump_loc): Remove newline emission.
7519 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
7520 emission to dump_printf_loc calls where missing.
7521 (vect_mark_for_runtime_alias_test): Ditto.
7522 (vect_analyze_data_ref_dependence): Ditto.
7523 (vect_analyze_data_ref_dependences): Ditto.
7524 (vect_slp_analyze_data_ref_dependence): Ditto.
7525 (vect_slp_analyze_data_ref_dependences): Ditto.
7526 (vect_compute_data_ref_alignment): Ditto.
7527 (vect_update_misalignment_for_peel): Ditto.
7528 (vect_verify_datarefs_alignment): Ditto.
7529 (vector_alignment_reachable_p): Ditto.
7530 (vect_get_data_access_cost): Ditto.
7531 (vect_enhance_data_refs_alignment): Ditto.
7532 (vect_find_same_alignment_drs): Ditto.
7533 (vect_analyze_data_refs_alignment): Ditto.
7534 (vect_analyze_group_access): Ditto.
7535 (vect_analyze_data_ref_access): Ditto.
7536 (vect_analyze_data_ref_accesses): Ditto.
7537 (vect_prune_runtime_alias_test_list): Ditto.
7538 (vect_analyze_data_refs): Ditto.
7539 (vect_create_addr_base_for_vector_ref): Ditto.
7540 (vect_create_data_ref_ptr): Ditto.
7541 (vect_grouped_store_supported): Ditto.
7542 (vect_grouped_load_supported): Ditto.
7543 * value-prof.c (check_counter): Ditto.
7544 (check_ic_target): Ditto.
7545 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
7546 (vect_recog_widen_mult_pattern): Ditto.
7547 (vect_recog_widen_sum_pattern): Ditto.
7548 (vect_recog_over_widening_pattern): Ditto.
7549 (vect_recog_widen_shift_pattern): Ditto.
7550 (vect_recog_rotate_pattern): Ditto.
7551 (vect_recog_vector_vector_shift_pattern): Ditto.
7552 (vect_recog_divmod_pattern): Ditto.
7553 (vect_recog_mixed_size_cond_pattern): Ditto.
7554 (vect_recog_bool_pattern): Ditto.
7555 (vect_pattern_recog_1): Ditto.
7556 (vect_pattern_recog): Ditto.
7557 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
7558 (vect_is_simple_iv_evolution): Ditto.
7559 (vect_analyze_scalar_cycles_1): Ditto.
7560 (vect_get_loop_niters): Ditto.
7561 (vect_analyze_loop_1): Ditto.
7562 (vect_analyze_loop_form): Ditto.
7563 (vect_analyze_loop_operations): Ditto.
7564 (vect_analyze_loop_2): Ditto.
7565 (vect_analyze_loop): Ditto.
7566 (report_vect_op): Ditto.
7567 (vect_is_slp_reduction): Ditto.
7568 (vect_is_simple_reduction_1): Ditto.
7569 (vect_get_known_peeling_cost): Ditto.
7570 (vect_estimate_min_profitable_iters): Ditto.
7571 (vect_model_reduction_cost): Ditto.
7572 (vect_model_induction_cost): Ditto.
7573 (get_initial_def_for_induction): Ditto.
7574 (vect_create_epilog_for_reduction): Ditto.
7575 (vectorizable_reduction): Ditto.
7576 (vectorizable_induction): Ditto.
7577 (vectorizable_live_operation): Ditto.
7578 (vect_loop_kill_debug_uses): Ditto.
7579 (vect_transform_loop): Ditto.
7580 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
7581 (vect_stmt_relevant_p): Ditto.
7582 (process_use): Ditto.
7583 (vect_mark_stmts_to_be_vectorized): Ditto.
7584 (vect_model_simple_cost): Ditto.
7585 (vect_model_promotion_demotion_cost): Ditto.
7586 (vect_model_store_cost): Ditto.
7587 (vect_get_store_cost): Ditto.
7588 (vect_model_load_cost): Ditto.
7589 (vect_get_load_cost): Ditto.
7590 (vect_init_vector_1): Ditto.
7591 (vect_get_vec_def_for_operand): Ditto.
7592 (vect_finish_stmt_generation): Ditto.
7593 (vectorizable_call): Ditto.
7594 (vectorizable_conversion): Ditto.
7595 (vectorizable_assignment): Ditto.
7596 (vectorizable_shift): Ditto.
7597 (vectorizable_operation): Ditto.
7598 (vectorizable_store): Ditto.
7599 (vectorizable_load): Ditto.
7600 (vectorizable_condition): Ditto.
7601 (vect_analyze_stmt): Ditto.
7602 (vect_transform_stmt): Ditto.
7603 (vect_is_simple_use): Ditto.
7604 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
7605 (vect_can_advance_ivs_p): Ditto.
7606 (vect_update_ivs_after_vectorizer): Ditto.
7607 (vect_do_peeling_for_loop_bound): Ditto.
7608 (vect_gen_niters_for_prolog_loop): Ditto.
7609 (vect_update_inits_of_drs): Ditto.
7610 (vect_create_cond_for_alias_checks): Ditto.
7611 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
7612 (vect_build_slp_tree_1): Ditto.
7613 (vect_supported_load_permutation_p): Ditto.
7614 (vect_analyze_slp_instance): Ditto.
7615 (vect_analyze_slp): Ditto.
7616 (vect_make_slp_decision): Ditto.
7617 (vect_detect_hybrid_slp): Ditto.
7618 (vect_bb_vectorization_profitable_p): Ditto.
7619 (vect_slp_analyze_bb_1): Ditto.
7620 (vect_update_slp_costs_according_to_vf): Ditto.
7621 (vect_get_mask_element): Ditto.
7622 (vect_transform_slp_perm_load): Ditto.
7623 (vect_schedule_slp_instance): Ditto.
7624 (vect_schedule_slp): Ditto.
7625 (vect_slp_transform_bb): Ditto.
7626 * profile.c (read_profile_edge_counts): Ditto.
7627 (compute_branch_probabilities): Ditto.
7628 * coverage.c (get_coverage_counts): Ditto.
7630 2013-09-16 Diego Novillo <dnovillo@google.com>
7632 * tree-core.h: Add missing comment lines from refactoring of tree.h.
7634 2013-09-16 Jan Hubicka <jh@suse.cz>
7636 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
7637 abstract functions; for static functions check the presence of body.
7639 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7641 * config/aarch64/aarch64-simd-builtins.def (fma): New.
7642 * config/aarch64/aarch64-simd.md
7643 (aarch64_mla_elt<mode>): New.
7644 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7645 (aarch64_mls_elt<mode>): Likewise.
7646 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7647 (aarch64_fma4_elt<mode>): Likewise.
7648 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7649 (aarch64_fma4_elt_to_128v2df): Likewise.
7650 (aarch64_fma4_elt_to_64df): Likewise.
7651 (fnma<mode>4): Likewise.
7652 (aarch64_fnma4_elt<mode>): Likewise.
7653 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7654 (aarch64_fnma4_elt_to_128v2df): Likewise.
7655 (aarch64_fnma4_elt_to_64df): Likewise.
7656 * config/aarch64/iterators.md (VDQSF): New.
7657 * config/aarch64/arm_neon.h
7658 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
7659 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
7661 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7663 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
7664 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
7665 (aarch64_mul3_elt_to_128df): Likewise.
7666 (aarch64_mul3_elt_to_64v2df): Likewise.
7667 * config/aarch64/iterators.md (VEL): Also handle DFmode.
7669 (VMUL_CHANGE_NLANES) Likewise.
7672 * config/aarch64/arm_neon.h
7673 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
7675 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7677 * config/aarch64/arm_neon.h
7678 (vcvtx_high_f32_f64): Fix parameters.
7680 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7681 Uros Bizjak <ubizjak@gmail.com>
7683 * config/alpha.c: Include tree-ssanames.h.
7685 2013-09-16 Richard Biener <rguenther@suse.de>
7687 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
7688 (dot_rdg_1): Handle control_dd.
7689 (create_edge_for_control_dependence): New function.
7690 (create_rdg_edges): Add control dependences if asked for.
7691 (build_rdg): Likewise.
7692 (generate_loops_for_partition): If there are not necessary
7693 control stmts remove all their dependencies.
7694 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
7695 (distribute_loop): Pass on control dependences.
7696 (tree_loop_distribution): Compute control dependences and remove
7697 restriction on number of loop nodes.
7699 2013-09-16 Jakub Jelinek <jakub@redhat.com>
7701 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
7704 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
7706 * cse.c (try_const_anchors): Punt on CC modes.
7708 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7710 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
7712 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
7715 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
7718 2013-09-14 Iain Sandoe <iain@codesourcery.com>
7721 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
7723 (darwin_objc1_section): Likewise.
7724 (darwin_file_end): Emit Image Info section when required.
7726 2013-09-14 Jan Hubicka <jh@suse.cz>
7728 * tree-into-ssa.c (gate_into_ssa): New.
7729 (pass_data_build_ssa): Use it.
7730 * cgraph.h (expand_thunk): Update prototype.
7731 * cgraphunit.c (analyze_function): Expand thunks early.
7732 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
7733 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
7734 set lowered flag; do not add new function.
7735 (assemble_thunks_and_aliases): Update.
7736 * tree-ssa.c (gate_init_datastructures): New gate.
7737 (pass_data_init_datastructures): Use it.
7739 2013-09-14 Iain Sandoe <iain@codesourcery.com>
7742 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
7743 xmm register set described in the psABI.
7745 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
7747 * dwarf2out.c (should_emit_struct_debug): Add check
7748 for type_decl variable is not NULL.
7750 2013-09-13 Jacek Caban <jacek@codeweavers.com>
7752 * config.gcc: Use new winnt-c.c target hooks
7753 * config/t-winnt: New file
7754 * config/winnt-c.c: New file
7755 * doc/tm.texi.in: Document new hook
7756 * doc/tm.texi: Regenerated
7758 2013-09-13 Jan Hubicka <jh@suse.cz>
7761 * ipa-inline.c (check_callers): New function.
7762 (check_caller_edge): Remove.
7763 (want_inline_function_to_all_callers_p): Also permit alises that are
7765 (inline_to_all_callers): Terminate the walk when devirtualization
7766 introduce new calls.
7768 2013-09-13 Jan Hubicka <jh@suse.cz>
7770 * ipa-inline-analysis.c (struct growth_data): Add node.
7771 (do_estimate_growth_1): Fix detection of recursion.
7773 2013-09-13 Jakub Jelinek <jakub@redhat.com>
7775 PR tree-optimization/58392
7776 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
7777 to avoid shadowing of outer loop variable. If
7778 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
7779 replace_by_duplicate_decl simduid of loops that have it set and
7780 set dest_cfun->has_simduid_loops and/or
7781 dest_cfun->has_force_vect_loops.
7782 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
7783 instead of maybe_lookup_decl.
7784 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
7785 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
7786 call. Copy over force_vect and copy and remap simduid. Set
7787 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
7788 (copy_cfg_body): Remove blocks_to_copy argument. Use
7789 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
7790 caller. Don't set cfun->has_simduid_loops and/or
7791 cfun->has_force_vect_loops here.
7792 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
7794 (expand_call_inline, tree_function_versioning): Adjust copy_body
7797 2013-09-13 Martin Jambor <mjambor@suse.cz>
7800 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
7801 the assert if the edge was a speculative one.
7803 2013-09-13 Richard Biener <rguenther@suse.de>
7805 * tree-data-ref.h (known_dependences_p): Move here ...
7806 * tree-loop-distribution.c (known_dependences_p): ... from here.
7807 (dump_rdg_component, debug_rdg_component): Remove.
7809 (generate_loops_for_partition): Use gimple_uid instead of
7810 relying on matching stmt visit order.
7811 (rdg_build_partitions): Take starting stmt vector.
7812 (ldist_gen): Merge into ...
7813 (distribute_loop): ... this function. Do not compute starting
7815 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
7817 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7819 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
7820 Set type attribute correctly. Set predicable_short_it attribute.
7821 (cmpsi_shiftsi): Remove %? from output template.
7823 2013-09-13 Richard Biener <rguenther@suse.de>
7825 * tree-loop-distribution.c (struct rdg_component,
7826 rdg_defs_used_in_other_loops_p, free_rdg_components,
7827 rdg_build_components): Remove.
7828 (stmts_from_loop): Do not record virtual PHIs.
7829 (generate_loops_for_partition): Skip virtual PHIs.
7830 (build_rdg_partition_for_component): Rename to ...
7831 (build_rdg_partition_for_vertex): ... this and adjust.
7832 (rdg_build_partitions): Take a vector of starting vertices
7833 instead of components. Remove unnecessary leftover handling.
7834 (ldist_gen): Do not build components or record other stores.
7835 (distribute_loop): Do not distribute loops containing stmts
7838 2013-09-13 Christian Bruel <christian.bruel@st.com>
7841 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
7843 2013-09-13 Kai Tietz <ktietz@redhat.com>
7845 * config.gcc: Separate cases for mingw and cygwin targets,
7846 and add 64-bit cygwin target case.
7848 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
7849 dll-export inline-functions.
7850 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
7853 2013-09-13 Jeff Law <law@redhat.com>
7857 2013-09-06 Jeff Law <law@redhat.com>
7859 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
7860 edge implied equivalences into successor phis.
7862 2013-09-12 DJ Delorie <dj@redhat.com>
7864 * config/rl78/rl78-virt.md: Change from | to \; for asm line
7867 2013-09-12 Brooks Moses <bmoses@google.com>
7870 * Makefile.in: Do not install driver binaries in $(target)/bin.
7872 2013-09-12 DJ Delorie <dj@redhat.com>
7874 * config/rl78/rl78.opt (mrelax): New.
7875 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
7876 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
7878 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
7880 (rl78_expand_epilogue): Likewise.
7882 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
7885 * lra-eliminations.c (remove_reg_equal_offset_note): New.
7886 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
7887 pointer elimination with using remove_reg_equal_offset_note.
7889 2013-09-12 DJ Delorie <dj@redhat.com>
7891 * config/msp430/: New port.
7892 * config.gcc (msp430): Added.
7893 * doc/invoke.texi: Document MSP430 options.
7894 * doc/install.texi: Document msp430-elf
7895 * doc/md.texi: Document msp430-elf
7896 * doc/contrib.texi: Document msp430-elf
7898 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
7901 2013-09-12 Martin Jambor <mjambor@suse.cz>
7904 * ipa-prop.c (remove_described_reference): Give up if the edge in the
7905 reference descriptor is NULL.
7906 (ipa_edge_removal_hook): If owning a reference descriptor, set its
7909 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
7911 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
7912 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
7913 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
7914 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
7915 make_temp_ssa_name): move to tree-ssanames.h
7916 * tree-ssa-alias.h: Move prototype.
7917 * tree-ssa.h: Include tree-ssanames.h.
7918 * tree-ssanames.c (FREE_SSANAMES): Move to here.
7919 * tree-ssanames.h: New. Move items from tree-flow*.h
7920 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
7922 2013-09-12 Richard Biener <rguenther@suse.de>
7924 PR tree-optimization/58404
7925 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
7926 propagate non-invariant addresses into dereferences wrapped
7927 in component references.
7929 2013-09-12 Richard Biener <rguenther@suse.de>
7931 PR tree-optimization/58402
7932 * passes.def: Move pass_late_warn_uninitialized later.
7934 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
7936 * tree-ssa.h: New. Move content from tree-flow.h and
7938 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
7939 Move prototypes belonging to tree-ssa.c.
7940 * tree-flow-inline.h (redirect_edge_var_map_def,
7941 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
7943 * gimple.h: Adjust prototypes.
7944 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
7946 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
7947 * tree.h: Move prototype to tree-ssa.h.
7948 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
7949 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
7950 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
7951 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
7952 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
7953 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
7954 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
7955 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
7956 graphite-blocking.c, graphite-clast-to-gimple.c,
7957 graphite-dependences.c, graphite-interchange.c,
7958 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
7959 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
7960 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
7961 ipa-reference.c, ipa-split.c, ipa-utils.c,
7962 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
7963 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
7964 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
7965 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
7966 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
7967 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
7968 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
7969 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
7970 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
7971 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
7972 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
7973 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
7974 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
7975 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
7976 value-prof.c, var-tracking.c,
7977 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
7979 2013-09-12 Richard Biener <rguenther@suse.de>
7981 PR tree-optimization/58396
7982 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
7983 (build_rdg): Take a loop-nest parameter, fix memleaks.
7984 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
7986 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
7988 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
7991 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
7993 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
7994 The FMA instruction names should have a 'v' prefix.
7996 2013-09-12 Richard Biener <rguenther@suse.de>
7998 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
7999 (dot_rdg): Use popen instead of system in optional code.
8000 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
8001 (already_processed_vertex_p): Adjust.
8002 (has_anti_or_output_dependence, predecessor_has_mem_write,
8003 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
8004 rdg_flag_uses): Remove.
8005 (rdg_flag_vertex): Simplify.
8006 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
8008 (build_rdg_partition_for_component): Process the first vertex
8009 of a component only.
8010 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
8012 2013-09-12 Alan Modra <amodra@gmail.com>
8014 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
8016 2013-09-11 DJ Delorie <dj@redhat.com>
8017 Nick Clifton <nickc@redhat.com>
8019 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
8020 (rl78_stack_based_mem): New.
8021 * config/rl78/constraints.md (Iv08): New.
8030 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
8032 (movsi): Change from expand to insn-and-split.
8033 (ashrsi3): Clobber AX.
8037 * config/rl78/rl78.md (CC_REG): Fix.
8038 (addsi3): Allow memory and immediate operands.
8039 (addsi3_internal): Split into...
8040 (addsi3_internal_virt): ...new, and ...
8041 (addsi3_internal_real): ...new.
8043 (subsi3_internal_virt): New.
8044 (subsi3_internal_real): New.
8045 (mulsi3): Add memory operand.
8046 (mulsi3_rl78): Likewise.
8047 (mulsi3_g13): Likewise.
8048 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
8049 (cbranchqi4_real): Add more constraint options.
8050 (cbranchhi4_real): Expand pattern.
8051 (cbranchhi4_real_signed): New.
8052 (cbranchhi4_real_inverted): New.
8053 (cbranchsi4_real_lt): New.
8054 (cbranchsi4_real_ge): New.
8055 (cbranchsi4_real_signed): New.
8056 (cbranchsi4_real): New.
8058 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
8060 (lshrsi3_virt): Likewise.
8061 (ashlsi3_virt): Likewise.
8062 (cbranchqi4_virt_signed): New.
8063 (cbranchhi4_virt_signed): New.
8064 (cbranchsi4_virt): New.
8065 * config/rl78/rl78.c: Whitespace fixes throughout.
8066 (move_elim_pass): New.
8067 (pass_data_rl78_move_elim): New.
8068 (pass_rl78_move_elim): New.
8069 (make_pass_rl78_move_elim): New.
8070 (rl78_devirt_info): Run devirt earlier.
8071 (rl78_move_elim_info): New.
8072 (rl78_asm_file_start): Register it.
8073 (rl78_split_movsi): New.
8074 (rl78_as_legitimate_address): Allow virtual base registers when
8076 (rl78_addr_space_convert): Remove spurious debug stuff.
8077 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
8078 (rl78_print_operand): More cases for not printing '#'.
8079 (rl78_expand_compare): Remove most of the logic.
8080 (content_memory): New.
8081 (clear_content_memory): New.
8082 (get_content_index): New.
8083 (get_content_name): New.
8084 (display_content_memory): New.
8085 (update_content): New.
8086 (record_content): New.
8087 (already_contains): New.
8088 (insn_ok_now): Re-recog insns with virtual registers.
8089 (add_postponed_content_update): New.
8090 (process_postponed_content_update): New.
8091 (gen_and_emit_move): New.
8092 (transcode_memory_rtx): Record new location content.
8093 Use gen_and_emit_move.
8094 (force_into_acc): New.
8095 (move_to_acc): Use gen_and_emit_move.
8096 (move_from_acc): Likewise.
8097 (move_acc_to_reg): Likewise.
8098 (move_to_x): Likewise.
8099 (move_to_hl): Likewise.
8100 (move_to_de): Likewise.
8101 (rl78_alloc_physical_registers_op1): Record location content.
8102 (has_constraint): New.
8103 (rl78_alloc_physical_registers_op2): Record location content.
8105 (rl78_alloc_physical_registers_ro1): Likewise.
8106 (rl78_alloc_physical_registers_cmp): Likewise.
8107 (rl78_alloc_physical_registers_umul): Likewise.
8108 (rl78_alloc_address_registers_macax): New.
8109 (rl78_alloc_physical_registers): Initialize and set location
8110 content memory as needed.
8111 (rl78_reorg): Make sure split2 is called.
8112 (rl78_rtx_costs): New.
8114 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
8116 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
8117 for (not (neg ...)) and (neg (not ...)) cases.
8119 2013-09-11 Richard Biener <rguenther@suse.de>
8122 * passes.def: Split critical edges before late uninit warning passes.
8123 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
8125 2013-09-11 Jakub Jelinek <jakub@redhat.com>
8127 PR tree-optimization/58385
8128 * fold-const.c (build_range_check): If both low and high are NULL,
8129 use omit_one_operand_loc to preserve exp side-effects.
8131 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8133 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
8135 2013-09-11 Richard Biener <rguenther@suse.de>
8137 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
8138 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
8139 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
8140 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
8141 stmts_from_loop, known_dependences_p, build_empty_rdg,
8142 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
8143 * tree-loop-distribution.c: ... here.
8144 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
8145 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
8146 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
8147 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
8148 * tree-loop-distribution.c: ... here.
8149 * tree-loop-distribution.c: Include gimple-pretty-print.h.
8150 (struct partition_s): Add loops member.
8151 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
8152 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
8153 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
8155 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8156 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8157 Sergey Lega <sergey.s.lega@intel.com>
8158 Anna Tikhonova <anna.tikhonova@intel.com>
8159 Ilya Tocar <ilya.tocar@intel.com>
8160 Andrey Turetskiy <andrey.turetskiy@intel.com>
8161 Ilya Verbin <ilya.verbin@intel.com>
8162 Kirill Yukhin <kirill.yukhin@intel.com>
8163 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8165 * config/i386/constraints.md (k): New.
8167 * config/i386/i386.c (const regclass_map): Add new mask registers.
8168 (dbx_register_map): Ditto.
8169 (dbx64_register_map): Ditto.
8170 (svr4_dbx_register_map): Ditto.
8171 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
8173 (ix86_preferred_reload_class): Disable constants for mask registers.
8174 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
8175 (ix86_hard_regno_mode_ok): Support new mask registers.
8176 (x86_order_regs_for_local_alloc): Ditto.
8177 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
8178 (FIXED_REGISTERS): Add new mask registers.
8179 (CALL_USED_REGISTERS): Ditto.
8180 (REG_ALLOC_ORDER): Ditto.
8181 (VALID_MASK_REG_MODE): New.
8182 (FIRST_MASK_REG): Ditto.
8183 (LAST_MASK_REG): Ditto.
8184 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
8185 (MAYBE_MASK_CLASS_P): New.
8186 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
8187 (REG_CLASS_CONTENTS): Ditto.
8188 (MASK_REGNO_P): New.
8189 (ANY_MASK_REG_P): Ditto.
8190 (HI_REGISTER_NAMES): Add new mask registers.
8191 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
8192 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
8194 (attribute "type"): Add mskmov, msklog.
8195 (attribute "length_immediate"): Support them.
8196 (attribute "memory"): Ditto.
8197 (attribute "prefix_0f"): Ditto.
8198 (*movhi_internal): Support new mask registers.
8199 (*movqi_internal): Ditto.
8200 (define_split): Split out clobber pattern is a logic
8201 insn on mask registers.
8202 (*k<logic><mode>): New.
8203 (*andhi_1): Extend to support mask regs.
8204 (*andqi_1): Extend to support mask regs.
8206 (define_split): Split and-not to and and not if operands
8208 (*<code><mode>_1): Separate HI mode to new pattern...
8209 (*<code>hi_1): This.
8210 (*<code>qi_1): Extend to support mask regs.
8212 (kortestzhi): Ditto.
8213 (kortestchi): Ditto.
8215 (*one_cmpl<mode>2_1): Remove HImode and handle it...
8216 (*one_cmplhi2_1): ...Here, now with mask registers support.
8217 (*one_cmplqi2_1): Support new mask registers.
8218 (HI/QImode arithmetics splitter): Don't split if mask registers
8220 (HI/QImode not splitter): Ditto.
8221 * config/i386/predicated.md (mask_reg_operand): New.
8222 (general_reg_operand): Ditto.
8224 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8226 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
8227 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
8229 2013-09-10 Jeff Law <law@redhat.com>
8231 PR tree-optimization/58380
8232 * tree-ssa-threadupdate.c (thread_block): Recognize another case
8233 of threading through a buried loop header.
8235 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
8236 return value for single successor case.
8238 2013-09-10 Jan Hubicka <jh@suse.cz>
8240 * ipa-devirt.c (ipa_devirt): Enable with LTO.
8242 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
8245 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
8246 support conditional execution.
8247 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
8249 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
8251 * lra.c (lra): Clear lra_optional_reload_pseudos before every
8253 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
8254 Check destination too to check move insn.
8255 (undo_optional_reloads): Add check that the original peudo did not
8256 changed its allocation and the optional reload was inherited on last
8257 inheritance pass. Break loop after deciding to keep optional reload.
8258 (lra_undo_inheritance): Add check that inherited pseudo still in
8261 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
8263 * config/aarch64/aarch64.md (generic_sched): New.
8264 * config/aarch64/aarch64-generic.md (load): Make conditional
8265 on generic_sched attribute.
8266 (nonload): Likewise.
8268 2013-09-10 Jan Hubicka <jh@suse.cz>
8270 * lto-cgraph.c: Include ipa-utils.h.
8271 (compute_ltrans_boundary): Also add possible targets into the boundary.
8273 2013-09-10 Jan Hubicka <jh@suse.cz>
8275 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
8276 VAR_DECL of vtable rather than full expression.
8278 2013-09-10 Jan Hubicka <jh@suse.cz>
8279 Paolo Carlini <paolo.carlini@oracle.com>
8281 * cgraphunit.c (analyze_functions): Save input_location, set it
8282 to UNKNOWN_LOCATION and restore it at the end.
8284 2013-09-10 Martin Jambor <mjambor@suse.cz>
8286 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
8287 represented by a thunk.
8289 2013-09-10 Jeff Law <law@redhat.com>
8291 PR tree-optimization/58343
8292 * tree-ssa-threadupdate.c (thread_block): Identify and disable
8293 jump threading requests through loop headers buried in the middle
8294 of a jump threading path.
8296 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
8297 in return value/type.
8299 2013-09-10 Jakub Jelinek <jakub@redhat.com>
8301 PR rtl-optimization/58365
8302 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
8303 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
8306 2013-09-10 Richard Biener <rguenther@suse.de>
8308 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
8309 * tree-data-ref.c (create_rdg_vertices): Collect all data
8310 references, signal failure to the caller, use data-ref API.
8311 (build_rdg): Compute data references only once. Maintain lifetime
8312 of data references and data dependences from within RDG.
8313 (free_rdg): Free dependence relations.
8314 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
8315 inventing extra dependences.
8316 (distribute_loop): Update for RDG API changes.
8318 2013-09-10 Kai Tietz <ktietz@redhat.com>
8320 * doc/invoke.texi (fms-extensions): Document changed
8321 behavior for ms-abi targets.
8322 * config/i386/i386.c (ix86_option_override_internal):
8323 Set default value of option -fms-extension for ms-abi targets.
8325 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
8327 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
8329 2013-09-10 Alan Modra <amodra@gmail.com>
8332 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
8334 2013-09-10 Alan Modra <amodra@gmail.com>
8336 * config/rs6000/predicates.md (add_cint_operand): New.
8337 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
8338 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
8339 using add_cint_operand.
8340 (largetoc_high_plus_aix): Likewise.
8342 2013-09-09 Jakub Jelinek <jakub@redhat.com>
8344 PR tree-optimization/58364
8345 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
8346 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
8347 the current range can't be an unconditional true or false.
8349 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
8351 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
8353 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
8355 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
8357 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
8360 * doc/invoke.texi (-Wdelete-incomplete): Document it.
8362 2013-09-09 Ian Bolton <ian.bolton@arm.com>
8364 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
8365 NO_REGS for immediate that can't be moved directly into FP_REGS.
8367 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8369 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
8370 comparison with negated operand.
8371 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
8374 2013-09-09 Richard Biener <rguenther@suse.de>
8377 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
8378 of a subloop record all its block as affecting loop-closed SSA form.
8380 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8382 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
8383 of an rtx/bitpos pair.
8384 (store_fixed_bit_field): Update accordingly.
8386 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8388 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
8390 * builtins.c (expand_errno_check): Likewise.
8391 * dwarf2cfi.c (init_return_column_size): Likewise.
8392 * except.c (sjlj_mark_call_sites): Likewise.
8393 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
8394 * lra-constraints.c (emit_inc): Likewise.
8395 * ree.c (combine_set_extension): Likewise.
8396 * regmove.c (fixup_match_2): Likewise.
8397 * reload1.c (inc_for_reload): Likewise.
8399 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8401 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
8402 (force_to_mode, simplify_shift_const_1, simplify_comparison):
8403 Use gen_int_mode with the mode of the associated simplify_* call.
8404 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
8405 * expmed.c (expand_shift_1): Likewise.
8406 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8407 * loop-iv.c (iv_number_of_iterations): Likewise.
8408 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8409 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
8411 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8413 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
8414 of the associated expand_* call.
8415 (asan_emit_stack_protection): Likewise.
8416 * builtins.c (round_trampoline_addr): Likewise.
8417 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
8418 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
8419 (emit_store_flag): Likewise.
8420 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
8421 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
8423 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8424 * ifcvt.c (noce_try_store_flag_constants): Likewise.
8425 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8426 * modulo-sched.c (generate_prolog_epilog): Likewise.
8427 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
8428 (expand_ctz, expand_ffs, expand_unop): Likewise.
8430 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8432 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
8433 of the associated gen_rtx_* call.
8434 * caller-save.c (init_caller_save): Likewise.
8435 * combine.c (find_split_point, make_extraction): Likewise.
8436 (make_compound_operation): Likewise.
8437 * dwarf2out.c (mem_loc_descriptor): Likewise.
8438 * explow.c (plus_constant, probe_stack_range): Likewise.
8439 * expmed.c (expand_mult_const): Likewise.
8440 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
8441 * reload1.c (init_reload): Likewise.
8442 * valtrack.c (cleanup_auto_inc_dec): Likewise.
8443 * var-tracking.c (adjust_mems): Likewise.
8444 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
8445 rather than gen_rtx_fmt_ee.
8447 2013-09-09 Jan Hubicka <jh@suse.cz>
8450 * value-prof.c (gimple_ic): Copy also abnormal edges.
8452 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8454 * asan.c (asan_shadow_cst): Use gen_int_mode.
8456 2013-09-08 Jan Hubicka <jh@suse.cz>
8458 * ipa-profile.c: Add toplevel comment.
8459 (ipa_propagate_frequency_1): Be more conservative when profile is read.
8460 (contains_hot_call_p): New function.
8461 (ipa_propagate_frequency): Set frequencies based on counts when
8463 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
8464 profile; do not tamper with profile after inlining if it is read.
8466 2013-09-08 Jan Hubicka <jh@suse.cz>
8468 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
8471 2013-09-08 Jan Hubicka <jh@suse.cz>
8473 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
8476 2013-09-08 Jeff Law <law@redhat.com>
8479 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
8482 2013-09-08 Andi Kleen <ak@linux.intel.com>
8484 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
8486 2013-09-08 Jan Hubicka <jh@suse.cz>
8488 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
8490 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
8492 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
8493 for non-debug insns.
8494 * lra.c (new_insn_reg): Take the containing insn as a parameter.
8495 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
8496 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
8499 2013-09-08 Jan Hubicka <jh@suse.cz>
8501 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
8502 targets and devirtualize to BUILT_IN_UNREACHABLE.
8503 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
8504 * ipa.c (walk_polymorphic_call_targets): New function.
8505 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
8506 functions; use the new timevar.
8507 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
8508 was removed from the program.
8509 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
8510 not consider it in the walk when its vtable is dead.
8511 (possible_polymorphic_call_targets_1): Pass anonymous flag to
8513 (devirt_variable_node_removal_hook): New function.
8514 (possible_polymorphic_call_targets): Also register
8515 devirt_variable_node_removal_hook.
8516 (ipa_devirt): Do not do non-speculative devirtualization.
8517 (gate_ipa_devirt): One execute if devirtualizing speculatively.
8519 2013-09-08 Jan Hubicka <jh@suse.cz>
8521 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
8522 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
8523 varpool_remove_variable_insertion_hook): Declare.
8524 * varpool.c (varpool_node_hook_list): New structure.
8525 (first_varpool_node_removal_hook,
8526 first_varpool_variable_insertion_hook): New variables.
8527 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
8528 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
8529 varpool_remove_variable_insertion_hook,
8530 varpool_call_variable_insertion_hooks): New functions.
8531 (varpool_remove_node): Use it.
8533 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
8536 * diagnostic.c (diagnostic_build_prefix): When s.file is
8537 "<built-in>" don't output line and column numbers.
8539 2013-09-06 Jan Hubicka <jh@suse.cz>
8541 * cgraphunit.c (expand_thunk): Get body before touching arguments.
8542 * lto-streamer-out.c: Stream thunks, too.
8543 * lto-streamer-in.c (input_function): Pop cfun here
8544 (lto_read_body): Instead of here.
8546 2013-09-06 Caroline Tice <cmtice@google.com>
8548 * doc/install.texi: Add documentation for the --enable-vtable-verify
8549 and the --disable-libvtv configure options.
8551 2013-09-06 Jeff Law <law@redhat.com>
8553 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8554 edge implied equivalences into successor phis.
8556 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
8558 * resource.c (mark_referenced_resources): Handle COND_EXEC.
8560 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
8562 * resource.c (mark_target_live_regs): Compute resources taking
8563 into account if a call is predicated or not.
8565 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
8567 * toplev.c (output_stack_usage): Be prepared for suffixes created by
8568 the compiler in the function names.
8570 2013-09-06 Jan Hubicka <jh@suse.cz>
8573 * ipa-inline.c (has_caller_p): New function.
8574 (want_inline_function_to_all_callers_p): Use it.
8575 (sum_callers, inline_to_all_callers): Break out from ...
8576 (ipa_inline): ... here.
8578 2013-09-06 Jan Hubicka <jh@suse.cz>
8580 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
8581 only when AVX is enabled.
8583 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8585 * config/aarch64/aarch64.md
8586 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
8587 is fpsimd_<load/store>2.
8588 (load_pair<mode>): Likewise.
8589 (store_pair<mode>): Likewise.
8591 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8593 * config/arm/types.md (type): Add "mrs" type.
8594 * config/aarch64/aarch64.md
8595 (aarch64_load_tp_hard): Make type "mrs".
8597 (load_tp_hard): Make type "mrs".
8598 * config/arm/cortex-a15.md: Update with new attributes.
8599 * config/arm/cortex-a5.md: Update with new attributes.
8600 * config/arm/cortex-a53.md: Update with new attributes.
8601 * config/arm/cortex-a7.md: Update with new attributes.
8602 * config/arm/cortex-a8.md: Update with new attributes.
8603 * config/arm/cortex-a9.md: Update with new attributes.
8604 * config/arm/cortex-m4.md: Update with new attributes.
8605 * config/arm/cortex-r4.md: Update with new attributes.
8606 * config/arm/fa526.md: Update with new attributes.
8607 * config/arm/fa606te.md: Update with new attributes.
8608 * config/arm/fa626te.md: Update with new attributes.
8609 * config/arm/fa726te.md: Update with new attributes.
8611 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8613 * config/aarch64/aarch64.md
8614 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
8615 (*movtf_aarch64): Likewise.
8617 (thumb1_movdi_insn): Use "multiple" for type where more than one
8618 instruction is used for a move.
8619 (*arm32_movhf): Likewise.
8620 (*thumb_movdf_insn): Likewise.
8622 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8624 * config/arm/types.md (type): Rename fcpys to fmov.
8626 (*arm_movsi_vfp): Rename type fcpys as fmov.
8627 (*thumb2_movsi_vfp): Likewise
8628 (*movhf_vfp_neon): Likewise
8629 (*movhf_vfp): Likewise
8630 (*movsf_vfp): Likewise
8631 (*thumb2_movsf_vfp): Likewise
8632 (*movsfcc_vfp): Likewise
8633 (*thumb2_movsfcc_vfp): Likewise
8634 * config/aarch64/aarch64-simd.md
8635 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
8636 * config/aarch64/aarch64.md
8637 (*movsi_aarch64): Replace type mov_reg with fmovs.
8638 (*movdi_aarch64): Likewise
8639 (*movsf_aarch64): Likewise
8640 (*movdf_aarch64): Likewise
8642 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
8643 * config/arm/iwmmxt.md
8644 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
8645 * config/arm/arm1020e.md: Update with new attributes.
8646 * config/arm/cortex-a15-neon.md: Update with new attributes.
8647 * config/arm/cortex-a5.md: Update with new attributes.
8648 * config/arm/cortex-a53.md: Update with new attributes.
8649 * config/arm/cortex-a7.md: Update with new attributes.
8650 * config/arm/cortex-a8-neon.md: Update with new attributes.
8651 * config/arm/cortex-a9.md: Update with new attributes.
8652 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8653 * config/arm/cortex-r4f.md: Update with new attributes.
8654 * config/arm/marvell-pj4.md: Update with new attributes.
8655 * config/arm/vfp11.md: Update with new attributes.
8657 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8659 * config/aarch64/aarch64.md
8660 (*madd<mode>): Fix type attribute.
8661 (*maddsi_uxtw): Likewise.
8662 (*msub<mode>): Likewise.
8663 (*msubsi_uxtw): Likewise.
8664 (<su_optab>maddsidi4): Likewise.
8665 (<su_optab>msubsidi4): Likewise.
8667 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8669 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
8670 * config/arm/arm.md (core_cycles): Remove fdiv.
8671 * config/arm/vfp.md:
8672 (*sqrtsf2_vfp): Update for attribute changes.
8673 (*sqrtdf2_vfp): Likewise.
8674 * config/aarch64/aarch64.md:
8675 (sqrt<mode>2): Update for attribute changes.
8676 * config/arm/arm1020e.md: Update with new attributes.
8677 * config/arm/cortex-a15-neon.md: Update with new attributes.
8678 * config/arm/cortex-a5.md: Update with new attributes.
8679 * config/arm/cortex-a53.md: Update with new attributes.
8680 * config/arm/cortex-a7.md: Update with new attributes.
8681 * config/arm/cortex-a8-neon.md: Update with new attributes.
8682 * config/arm/cortex-a9.md: Update with new attributes.
8683 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8684 * config/arm/cortex-r4f.md: Update with new attributes.
8685 * config/arm/marvell-pj4.md: Update with new attributes.
8686 * config/arm/vfp11.md: Update with new attributes.
8688 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8690 * config/arm/types.md
8691 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
8692 * config/aarch64/aarch64.md
8693 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
8695 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
8696 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
8697 (float<GPI:mode><GPF:mode>2): Likewise.
8699 (*truncsisf2_vfp): Update with new attributes.
8700 (*truncsidf2_vfp): Likewise.
8701 (fixuns_truncsfsi2): Likewise.
8702 (fixuns_truncdfsi2): Likewise.
8703 (*floatsisf2_vfp): Likewise.
8704 (*floatsidf2_vfp): Likewise.
8705 (floatunssisf2): Likewise.
8706 (floatunssidf2): Likewise.
8707 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
8708 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
8709 * config/arm/arm1020e.md: Update with new attributes.
8710 * config/arm/cortex-a15-neon.md: Update with new attributes.
8711 * config/arm/cortex-a5.md: Update with new attributes.
8712 * config/arm/cortex-a53.md: Update with new attributes.
8713 * config/arm/cortex-a7.md: Update with new attributes.
8714 * config/arm/cortex-a8-neon.md: Update with new attributes.
8715 * config/arm/cortex-a9.md: Update with new attributes.
8716 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8717 * config/arm/cortex-r4f.md: Update with new attributes.
8718 * config/arm/marvell-pj4.md: Update with new attributes.
8719 * config/arm/vfp11.md: Update with new attributes.
8721 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8723 * config/aarch64/arm_neon.h
8724 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
8725 (vqtbx<1,2,3,4><q>_s8): Likewise.
8727 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8729 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
8730 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
8731 (add<mode>3): Add type attribute.
8732 (add<mode>3): Likewise.
8733 (usadd<mode>3): Likewise.
8734 (ssadd<mode>3): Likewise.
8735 (sub<mode>3): Likewise.
8736 (sub<mode>3): Likewise.
8737 (ussub<mode>3): Likewise.
8738 (sssub<mode>3): Likewise.
8739 (ssmulsa3): Likewise.
8740 (usmulusa3): Likewise.
8741 (arm_usatsihi): Likewise.
8743 (*movdi_vfp): Add types for all instructions.
8744 (*movdi_vfp_cortexa8): Likewise.
8745 (*movhf_vfp_neon): Likewise.
8746 (*movhf_vfp): Likewise.
8747 (*movdf_vfp): Likewise.
8748 (*thumb2_movdf_vfp): Likewise.
8749 (*thumb2_movdfcc_vfp): Likewise.
8750 * config/arm/arm.md: Add type attribute to all insn patterns.
8751 (*thumb1_adddi3): Add type attribute.
8752 (*arm_adddi3): Likewise.
8753 (*adddi_sesidi_di): Likewise.
8754 (*adddi_zesidi_di): Likewise.
8755 (*thumb1_addsi3): Likewise.
8756 (addsi3_compare0): Likewise.
8757 (*addsi3_compare0_scratch): Likewise.
8758 (*compare_negsi_si): Likewise.
8759 (cmpsi2_addneg): Likewise.
8760 (*addsi3_carryin_<optab>): Likewise.
8761 (*addsi3_carryin_alt2_<optab>): Likewise.
8762 (*addsi3_carryin_clobercc_<optab>): Likewise.
8763 (*subsi3_carryin): Likewise.
8764 (*subsi3_carryin_const): Likewise.
8765 (*subsi3_carryin_compare): Likewise.
8766 (*subsi3_carryin_compare_const): Likewise.
8767 (*arm_subdi3): Likewise.
8768 (*thumb_subdi3): Likewise.
8769 (*subdi_di_zesidi): Likewise.
8770 (*subdi_di_sesidi): Likewise.
8771 (*subdi_zesidi_di): Likewise.
8772 (*subdi_sesidi_di): Likewise.
8773 (*subdi_zesidi_ze): Likewise.
8774 (thumb1_subsi3_insn): Likewise.
8775 (*arm_subsi3_insn): Likewise.
8776 (*anddi3_insn): Likewise.
8777 (*anddi_zesidi_di): Likewise.
8778 (*anddi_sesdi_di): Likewise.
8779 (*ne_zeroextracts): Likewise.
8780 (*ne_zeroextracts): Likewise.
8781 (*ite_ne_zeroextr): Likewise.
8782 (*ite_ne_zeroextr): Likewise.
8783 (*anddi_notdi_di): Likewise.
8784 (*anddi_notzesidi): Likewise.
8785 (*anddi_notsesidi): Likewise.
8786 (andsi_notsi_si): Likewise.
8787 (thumb1_bicsi3): Likewise.
8788 (*iordi3_insn): Likewise.
8789 (*iordi_zesidi_di): Likewise.
8790 (*iordi_sesidi_di): Likewise.
8791 (*thumb1_iorsi3_insn): Likewise.
8792 (*xordi3_insn): Likewise.
8793 (*xordi_zesidi_di): Likewise.
8794 (*xordi_sesidi_di): Likewise.
8795 (*arm_xorsi3): Likewise.
8796 (*andsi_iorsi3_no): Likewise.
8797 (*smax_0): Likewise.
8798 (*smax_m1): Likewise.
8799 (*arm_smax_insn): Likewise.
8800 (*smin_0): Likewise.
8801 (*arm_smin_insn): Likewise.
8802 (*arm_umaxsi3): Likewise.
8803 (*arm_uminsi3): Likewise.
8804 (*minmax_arithsi): Likewise.
8805 (*minmax_arithsi_): Likewise.
8806 (*satsi_<SAT:code>): Likewise.
8807 (arm_ashldi3_1bit): Likewise.
8808 (arm_ashrdi3_1bit): Likewise.
8809 (arm_lshrdi3_1bit): Likewise.
8810 (*arm_negdi2): Likewise.
8811 (*thumb1_negdi2): Likewise.
8812 (*arm_negsi2): Likewise.
8813 (*thumb1_negsi2): Likewise.
8814 (*negdi_extendsid): Likewise.
8815 (*negdi_zero_extend): Likewise.
8816 (*arm_abssi2): Likewise.
8817 (*thumb1_abssi2): Likewise.
8818 (*arm_neg_abssi2): Likewise.
8819 (*thumb1_neg_abss): Likewise.
8820 (one_cmpldi2): Likewise.
8821 (extend<mode>di2): Likewise.
8822 (*compareqi_eq0): Likewise.
8823 (*arm_extendhisi2addsi): Likewise.
8824 (*arm_movdi): Likewise.
8825 (*thumb1_movdi_insn): Likewise.
8826 (*arm_movt): Likewise.
8827 (*thumb1_movsi_insn): Likewise.
8828 (pic_add_dot_plus_four): Likewise.
8829 (pic_add_dot_plus_eight): Likewise.
8830 (tls_load_dot_plus_eight): Likewise.
8831 (*thumb1_movhi_insn): Likewise.
8832 (*thumb1_movsf_insn): Likewise.
8833 (*movdf_soft_insn): Likewise.
8834 (*thumb_movdf_insn): Likewise.
8835 (cbranchsi4_insn): Likewise.
8836 (cbranchsi4_scratch): Likewise.
8837 (*negated_cbranchsi4): Likewise.
8838 (*tbit_cbranch): Likewise.
8839 (*tlobits_cbranch): Likewise.
8840 (*tstsi3_cbranch): Likewise.
8841 (*cbranchne_decr1): Likewise.
8842 (*addsi3_cbranch): Likewise.
8843 (*addsi3_cbranch_scratch): Likewise.
8844 (*arm_cmpdi_insn): Likewise.
8845 (*arm_cmpdi_unsig): Likewise.
8846 (*arm_cmpdi_zero): Likewise.
8847 (*thumb_cmpdi_zero): Likewise.
8848 (*deleted_compare): Likewise.
8849 (*mov_scc): Likewise.
8850 (*mov_negscc): Likewise.
8851 (*mov_notscc): Likewise.
8852 (*cstoresi_eq0_thumb1_insn): Likewise.
8853 (cstoresi_nltu_thumb1): Likewise.
8854 (cstoresi_ltu_thu): Likewise.
8855 (thumb1_addsi3_addgeu): Likewise.
8856 (*arm_jump): Likewise.
8857 (*thumb_jump): Likewise.
8858 (*check_arch2): Likewise.
8859 (arm_casesi_internal): Likewise.
8860 (thumb1_casesi_dispatch): Likewise.
8861 (*arm_indirect_jump): Likewise.
8862 (*thumb1_indirect_jump): Likewise.
8864 (*and_scc): Likewise.
8865 (*ior_scc): Likewise.
8866 (*compare_scc): Likewise.
8867 (*cond_move): Likewise.
8868 (*cond_arith): Likewise.
8869 (*cond_sub): Likewise.
8870 (*cmp_ite0): Likewise.
8871 (*cmp_ite1): Likewise.
8872 (*cmp_and): Likewise.
8873 (*cmp_ior): Likewise.
8874 (*ior_scc_scc): Likewise.
8875 (*ior_scc_scc_cmp): Likewise.
8876 (*and_scc_scc): Likewise.
8877 (*and_scc_scc_cmp): Likewise.
8878 (*and_scc_scc_nod): Likewise.
8879 (*negscc): Likewise.
8880 (movcond_addsi): Likewise.
8881 (movcond): Likewise.
8882 (*ifcompare_plus_move): Likewise.
8883 (*if_plus_move): Likewise.
8884 (*ifcompare_move_plus): Likewise.
8885 (*if_move_plus): Likewise.
8886 (*ifcompare_arith_arith): Likewise.
8887 (*if_arith_arith): Likewise.
8888 (*ifcompare_arith_move): Likewise.
8889 (*if_arith_move): Likewise.
8890 (*ifcompare_move_arith): Likewise.
8891 (*if_move_arith): Likewise.
8892 (*ifcompare_move_not): Likewise.
8893 (*if_move_not): Likewise.
8894 (*ifcompare_not_move): Likewise.
8895 (*if_not_move): Likewise.
8896 (*ifcompare_shift_move): Likewise.
8897 (*if_shift_move): Likewise.
8898 (*ifcompare_move_shift): Likewise.
8899 (*if_move_shift): Likewise.
8900 (*ifcompare_shift_shift): Likewise.
8901 (*ifcompare_not_arith): Likewise.
8902 (*ifcompare_arith_not): Likewise.
8903 (*if_arith_not): Likewise.
8904 (*ifcompare_neg_move): Likewise.
8905 (*if_neg_move): Likewise.
8906 (*ifcompare_move_neg): Likewise.
8907 (*if_move_neg): Likewise.
8908 (prologue_thumb1_interwork): Likewise.
8909 (*cond_move_not): Likewise.
8910 (*sign_extract_onebit): Likewise.
8911 (*not_signextract_onebit): Likewise.
8912 (stack_tie): Likewise.
8913 (align_4): Likewise.
8914 (align_8): Likewise.
8915 (consttable_end): Likewise.
8916 (consttable_1): Likewise.
8917 (consttable_2): Likewise.
8918 (consttable_4): Likewise.
8919 (consttable_8): Likewise.
8920 (consttable_16): Likewise.
8921 (*thumb1_tablejump): Likewise.
8922 (prefetch): Likewise.
8923 (force_register_use): Likewise.
8924 (thumb_eh_return): Likewise.
8925 (load_tp_hard): Likewise.
8926 (load_tp_soft): Likewise.
8927 (tlscall): Likewise.
8928 (*arm_movtas_ze): Likewise.
8929 (*arm_rev): Likewise.
8930 (*arm_revsh): Likewise.
8931 (*arm_rev16): Likewise.
8932 * config/arm/thumb2.md
8933 (*thumb2_smaxsi3): Likewise.
8934 (*thumb2_sminsi3): Likewise.
8935 (*thumb32_umaxsi3): Likewise.
8936 (*thumb2_uminsi3): Likewise.
8937 (*thumb2_negdi2): Likewise.
8938 (*thumb2_abssi2): Likewise.
8939 (*thumb2_neg_abss): Likewise.
8940 (*thumb2_movsi_insn): Likewise.
8941 (tls_load_dot_plus_four): Likewise.
8942 (*thumb2_movhi_insn): Likewise.
8943 (*thumb2_mov_scc): Likewise.
8944 (*thumb2_mov_negs): Likewise.
8945 (*thumb2_mov_negs): Likewise.
8946 (*thumb2_mov_nots): Likewise.
8947 (*thumb2_mov_nots): Likewise.
8948 (*thumb2_movsicc_): Likewise.
8949 (*thumb2_movsfcc_soft_insn): Likewise.
8950 (*thumb2_indirect_jump): Likewise.
8951 (*thumb2_and_scc): Likewise.
8952 (*thumb2_ior_scc): Likewise.
8953 (*thumb2_ior_scc_strict_it): Likewise.
8954 (*thumb2_cond_move): Likewise.
8955 (*thumb2_cond_arith): Likewise.
8956 (*thumb2_cond_ari): Likewise.
8957 (*thumb2_cond_sub): Likewise.
8958 (*thumb2_negscc): Likewise.
8959 (*thumb2_movcond): Likewise.
8960 (thumb2_casesi_internal): Likewise.
8961 (thumb2_casesi_internal_pic): Likewise.
8962 (*thumb2_alusi3_short): Likewise.
8963 (*thumb2_mov<mode>_shortim): Likewise.
8964 (*thumb2_addsi_short): Likewise.
8965 (*thumb2_subsi_short): Likewise.
8966 (thumb2_addsi3_compare0): Likewise.
8967 (*thumb2_cbz): Likewise.
8968 (*thumb2_cbnz): Likewise.
8969 (*thumb2_one_cmplsi2_short): Likewise.
8970 (*thumb2_negsi2_short): Likewise.
8971 (*orsi_notsi_si): Likewise.
8972 * config/arm/arm1020e.md: Update with new attributes.
8973 * config/arm/arm1026ejs.md: Update with new attributes.
8974 * config/arm/arm1136jfs.md: Update with new attributes.
8975 * config/arm/arm926ejs.md: Update with new attributes.
8976 * config/arm/cortex-a15.md: Update with new attributes.
8977 * config/arm/cortex-a5.md: Update with new attributes.
8978 * config/arm/cortex-a53.md: Update with new attributes.
8979 * config/arm/cortex-a7.md: Update with new attributes.
8980 * config/arm/cortex-a8.md: Update with new attributes.
8981 * config/arm/cortex-a9.md: Update with new attributes.
8982 * config/arm/cortex-m4.md: Update with new attributes.
8983 * config/arm/cortex-r4.md: Update with new attributes.
8984 * config/arm/fa526.md: Update with new attributes.
8985 * config/arm/fa606te.md: Update with new attributes.
8986 * config/arm/fa626te.md: Update with new attributes.
8987 * config/arm/fa726te.md: Update with new attributes.
8989 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8991 * config/aarch64/aarch64-simd.md
8992 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
8993 <vwx> iterator to ensure correct register choice.
8994 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
8995 (aarch64_sqdmull_n<mode>): Likewise.
8996 (aarch64_sqdmull2_n<mode>_internal): Likewise.
8997 * config/aarch64/arm_neon.h
8998 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
8999 (vml<as><q>_n_<su>16): Likewise.
9000 (vml<as>l_high_lane<q>_<su>16): Likewise.
9001 (vml<as>l_high_n_<su>16): Likewise.
9002 (vml<as>l_lane<q>_<su>16): Likewise.
9003 (vml<as>l_n_<su>16): Likewise.
9004 (vmul<q>_lane<q>_<su>16): Likewise.
9005 (vmul<q>_n_<su>16): Likewise.
9006 (vmull_lane<q>_<su>16): Likewise.
9007 (vmull_n_<su>16): Likewise.
9008 (vmull_high_lane<q>_<su>16): Likewise.
9009 (vmull_high_n_<su>16): Likewise.
9010 (vqrdmulh<q>_n_s16): Likewise.
9012 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
9014 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
9015 have the correct lane parameter.
9017 2013-09-06 Richard Biener <rguenther@suse.de>
9019 * cfganal.c (control_dependences::~control_dependences):
9020 Properly free all of the vector.
9022 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
9025 * config/i386/i386.c (ix86_conditional_register_usage):
9026 Proper initialize extended SSE registers.
9028 2013-09-06 Jan Hubicka <jh@suse.cz>
9030 PR tree-optimization/58311
9031 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
9033 2013-09-06 Jan Hubicka <jh@suse.cz>
9035 * Makefile.in (tree-sra.o): Update dependencies.
9036 * tree-sra.c: Include ipa-utils.h
9037 (scan_function): Use recursive_call_p.
9038 (has_caller_p): New function.
9039 (cgraph_for_node_and_aliases): Count also callers of aliases.
9041 2013-09-06 Jan Hubicka <jh@suse.cz>
9044 * cgraph.h (symtab_semantically_equivalent_p): Declare.
9045 * tree-tailcall.c: Include ipa-utils.h.
9046 (find_tail_calls): Use it.
9047 * ipa-pure-const.c (check_call): Likewise.
9048 * ipa-utils.c (recursive_call_p): New function.
9049 * ipa-utils.h (recursive_call_p): Dclare.
9050 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
9051 (symtab_semantically_equivalent_p): New function.
9052 * Makefile.in (tree-tailcall.o): Update dependencies.
9054 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9056 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
9059 2013-09-06 Richard Biener <rguenther@suse.de>
9061 * lto-streamer.h (lto_global_var_decls): Remove.
9062 * Makefile.in (OBJS): Remove lto-symtab.o.
9063 (lto-symtab.o): Remove.
9064 (GTFILES): Remove lto-symtab.c
9065 * lto-symtab.c: Move to lto/
9067 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9069 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
9070 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
9071 (UNSPEC_FPINT_RINT): New constant definitions.
9072 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
9073 definition with 2 attributes.
9074 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
9075 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
9078 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9080 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
9081 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
9082 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
9083 attribute to "z196_alone".
9084 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
9087 2013-09-06 Richard Biener <rguenther@suse.de>
9089 * basic-block.h (class control_dependences): New.
9090 * tree-ssa-dce.c (control_dependence_map): Remove.
9092 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
9093 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
9094 find_pdom, find_control_dependence, find_all_control_dependences):
9096 (mark_control_dependent_edges_necessary,
9097 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
9098 tree_dce_done, perform_tree_ssa_dce): Adjust.
9099 * cfganal.c (set_control_dependence_map_bit,
9100 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
9101 find_all_control_dependences): Move from tree-ssa-dce.c and
9102 implement as methods of control_dependences class.
9103 (control_dependences::control_dependences): New.
9104 (control_dependences::~control_dependences): Likewise.
9105 (control_dependences::get_edges_dependent_on): Likewise.
9106 (control_dependences::get_edge): Likewise.
9108 2013-09-04 Jan Hubicka <jh@suse.cz>
9110 * tree.c (types_same_for_odr): Drop overactive check.
9111 * ipa-devirt.c (hash_type_name): Likewise.
9113 2013-09-04 Jan Hubicka <jh@suse.cz>
9115 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
9116 (analyze_functions): ... here.
9118 2013-09-04 Jan Hubicka <jh@suse.cz>
9121 * cgraphunit.c (analyze_functions): Clear AUX fields
9122 after processing; initialize assembler name has.
9124 2013-09-05 Jeff Law <law@redhat.com>
9126 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
9127 from thread_around_empty_block. Record threading path into PATH.
9128 Recurse if threading through the initial block is successful.
9129 (thread_across_edge): Corresponding changes to slightly simplify.
9131 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9133 * config/aarch64/aarch64.md
9134 (type): Remove frecpe, frecps, frecpx.
9135 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
9136 fix to be a TARGET_SIMD instruction.
9137 (aarch64_frecps): Remove.
9138 * config/aarch64/aarch64-simd.md
9139 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
9140 (aarch64_frecps<mode>): Handle all float/vector of float modes.
9142 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9143 Sofiane Naci <sofiane.naci@arm.com>
9145 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
9146 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
9147 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
9148 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
9149 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
9150 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
9151 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
9152 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
9153 into "clz, "rbit". Rename "shift" to "shift_imm".
9154 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
9155 changes. Update for attribute changes all occurrences of arlo_* and
9157 * config/arm/arm-fixed.md: Update for attribute changes
9158 all occurrences of arlo_* types.
9159 * config/arm/thumb2.md: Update for attribute changes all occurrences
9161 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
9162 (cortexa7_older_only): Likewise.
9163 (cortexa7_younger): Likewise.
9164 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9165 (1020alu_shift_op): Likewise.
9166 (1020alu_shift_reg_op): Likewise.
9167 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9168 (alu_shift_op): Likewise.
9169 (alu_shift_reg_op): Likewise.
9170 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9171 (11_alu_shift_op): Likewise.
9172 (11_alu_shift_reg_op): Likewise.
9173 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9174 (9_alu_shift_reg_op): Likewise.
9175 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
9177 (cortex_a15_alu_shift): Likewise.
9178 (cortex_a15_alu_shift_reg): Likewise.
9179 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
9181 (cortex_a5_alu_shift): Likewise.
9182 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
9184 (cortex_a53_alu_shift): Likewise.
9185 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
9187 (cortex_a7_alu_reg): Likewise.
9188 (cortex_a7_alu_shift): Likewise.
9189 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
9191 (cortex_a8_alu_shift): Likewise.
9192 (cortex_a8_alu_shift_reg): Likewise.
9193 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9194 (cortex_a9_dp_shift): Likewise.
9195 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
9197 * config/arm/cortex-r4.md
9198 (cortex_r4_alu): Update for attribute changes.
9199 (cortex_r4_mov): Likewise.
9200 (cortex_r4_alu_shift_reg): Likewise.
9201 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9202 (526_alu_shift_op): Likewise.
9203 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9204 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9205 (626te_alu_shift_op): Likewise.
9206 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
9207 (726te_alu_shift_op): Likewise.
9208 (726te_alu_shift_reg_op): Likewise.
9209 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9210 (mp626_alu_shift_op): Likewise.
9211 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
9212 (pj4_alu_conds): Likewise.
9213 (pj4_shift): Likewise.
9214 (pj4_shift_conds): Likewise.
9215 (pj4_alu_shift): Likewise.
9216 (pj4_alu_shift_conds): Likewise.
9217 * config/aarch64/aarch64.md: Update for attribute change
9218 all occurrences of arlo_* and shift* types.
9220 2013-09-05 Mike Stump <mikestump@comcast.net>
9222 * tree.h: Move documentation for tree_function_decl to tree-core.h
9223 with the declaration.
9225 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
9228 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
9229 looking for widest mode.
9231 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
9233 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
9235 2013-09-05 Richard Biener <rguenther@suse.de>
9237 PR tree-optimization/58137
9238 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
9239 Do not create vectors of pointers.
9240 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
9241 types for the components of the vector initializer.
9242 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
9243 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
9245 2013-09-05 Martin Jambor <mjambor@suse.cz>
9247 * ipa-prop.c (remove_described_reference): Accept missing references,
9248 return false if that hppens, otherwise return true.
9249 (cgraph_node_for_jfunc): New function.
9250 (try_decrement_rdesc_refcount): Likewise.
9251 (try_make_edge_direct_simple_call): Use them.
9252 (ipa_edge_removal_hook): Remove references from rdescs.
9253 (ipa_edge_duplication_hook): Clone rdescs and their references
9254 when the new edge has the same caller as the old one.
9255 * cgraph.c (cgraph_resolve_speculation): Remove speculative
9256 reference before removing any edges.
9258 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
9260 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
9262 * thumb2.md (thumb2_storewb_parisi): New pattern.
9264 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
9266 * config/aarch64/aarch64-option-extensions.def: Add
9267 AARCH64_OPT_EXTENSION of 'crc'.
9268 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
9269 (AARCH64_ISA_CRC): Ditto.
9270 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
9271 description of the CRC extension.
9273 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
9275 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
9276 * config/rs6000/linux.h: Ditto.
9277 * alpha/linux.h: Ditto.
9278 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
9279 no_c99_libc_has_function.
9280 * config/c6x/uclinux-elf.h: Ditto.
9281 * config/lm32/uclinux-elf.h: Ditto.
9282 * config/m68k/uclinux.h: Ditto.
9283 * config/moxie/uclinux.h: Ditto.
9284 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
9285 (crisv32-*-linux*, cris-*-linux*): Ditto.
9286 * config/bfin/bfin.c: Include "tm_p.h".
9288 2013-09-05 Richard Biener <rguenther@suse.de>
9290 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
9291 check for a definition without a basic-block.
9293 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9294 Sofiane Naci <sofiane.naci@arm.com>
9296 * config/aarch64/aarch64.md
9297 (*movti_aarch64): Rename r_2_f and f_2_r.
9298 (*movsf_aarch64): Likewise.
9299 (*movdf_aarch64): Likewise.
9300 (*movtf_aarch64): Likewise.
9301 (aarch64_movdi_<mode>low): Likewise.
9302 (aarch64_movdi_<mode>high): Likewise.
9303 (aarch64_mov<mode>high_di): Likewise.
9304 (aarch64_mov<mode>low_di): Likewise.
9305 (aarch64_movtilow_tilow): Likewise.
9306 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
9307 values to config/arm/types.md
9308 (attribute "conds"): Update for attribute change.
9309 (anddi3_insn): Likewise.
9310 (iordi3_insn): Likewise.
9311 (xordi3_insn): Likewise.
9312 (one_cmpldi2): Likewise.
9313 * config/arm/types.md (type): Add Neon types.
9314 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
9315 use "type" attribute.
9316 (movmisalign<mode>_neon_store): Likewise.
9317 (movmisalign<mode>_neon_load): Likewise.
9318 (vec_set<mode>_internal): Likewise.
9319 (vec_setv2di_internal): Likewise.
9320 (vec_extract<mode>): Likewise.
9321 (vec_extractv2di): Likewise.
9322 (add<mode>3_neon): Likewise.
9323 (adddi3_neon): Likewise.
9324 (sub<mode>3_neon): Likewise.
9325 (subdi3_neon): Likewise.
9326 (mul<mode>3_neon): Likewise.
9327 (mul<mode>3add<mode>_neon): Likewise.
9328 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
9329 (fma<VCVTF:mode>4)): Likewise.
9330 (fma<VCVTF:mode>4_intrinsic): Likewise.
9331 (fmsub<VCVTF:mode>4)): Likewise.
9332 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
9333 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
9334 (ior<mode>3): Likewise.
9335 (and<mode>3): Likewise.
9336 (anddi3_neon): Likewise.
9337 (orn<mode>3_neon): Likewise.
9338 (orndi3_neon): Likewise.
9339 (bic<mode>3_neon): Likewise.
9340 (bicdi3_neon): Likewise.
9341 (xor<mode>3): Likewise.
9342 (one_cmpl<mode>2): Likewise.
9343 (abs<mode>2): Likewise.
9344 (neg<mode>2): Likewise.
9345 (umin<mode>3_neon): Likewise.
9346 (umax<mode>3_neon): Likewise.
9347 (smin<mode>3_neon): Likewise.
9348 (smax<mode>3_neon): Likewise.
9349 (vashl<mode>3): Likewise.
9350 (vashr<mode>3_imm): Likewise.
9351 (vlshr<mode>3_imm): Likewise.
9352 (ashl<mode>3_signed): Likewise.
9353 (ashl<mode>3_unsigned): Likewise.
9354 (neon_load_count): Likewise.
9355 (ashldi3_neon_noclobber): Likewise.
9356 (signed_shift_di3_neon): Likewise.
9357 (unsigned_shift_di3_neon): Likewise.
9358 (ashrdi3_neon_imm_noclobber): Likewise.
9359 (lshrdi3_neon_imm_noclobber): Likewise.
9360 (widen_ssum<mode>3): Likewise.
9361 (widen_usum<mode>3): Likewise.
9362 (quad_halves_<code>v4si): Likewise.
9363 (quad_halves_<code>v4sf): Likewise.
9364 (quad_halves_<code>v8hi): Likewise.
9365 (quad_halves_<code>v16qi): Likewise.
9366 (reduc_splus_v2di): Likewise.
9367 (neon_vpadd_internal<mode>): Likewise.
9368 (neon_vpsmin<mode>): Likewise.
9369 (neon_vpsmax<mode>): Likewise.
9370 (neon_vpumin<mode>): Likewise.
9371 (neon_vpumax<mode>): Likewise.
9372 (ss_add<mode>_neon): Likewise.
9373 (us_add<mode>_neon): Likewise.
9374 (ss_sub<mode>_neon): Likewise.
9375 (us_sub<mode>_neon): Likewise.
9376 (neon_vadd<mode>_unspec): Likewise.
9377 (neon_vaddl<mode>): Likewise.
9378 (neon_vaddw<mode>): Likewise.
9379 (neon_vhadd<mode>): Likewise.
9380 (neon_vqadd<mode>): Likewise.
9381 (neon_vaddhn<mode>): Likewise.
9382 (neon_vmul<mode>): Likewise.
9383 (neon_vmla<mode>): Likewise.
9384 (neon_vmlal<mode>): Likewise.
9385 (neon_vmls<mode>): Likewise.
9386 (neon_vmlsl<mode>): Likewise.
9387 (neon_vqdmulh<mode>): Likewise.
9388 (neon_vqdmlal<mode>): Likewise.
9389 (neon_vqdmlsl<mode>): Likewise.
9390 (neon_vmull<mode>): Likewise.
9391 (neon_vqdmull<mode>): Likewise.
9392 (neon_vsub<mode>_unspec): Likewise.
9393 (neon_vsubl<mode>): Likewise.
9394 (neon_vsubw<mode>): Likewise.
9395 (neon_vqsub<mode>): Likewise.
9396 (neon_vhsub<mode>): Likewise.
9397 (neon_vsubhn<mode>): Likewise.
9398 (neon_vceq<mode>): Likewise.
9399 (neon_vcge<mode>): Likewise.
9400 (neon_vcgeu<mode>): Likewise.
9401 (neon_vcgt<mode>): Likewise.
9402 (neon_vcgtu<mode>): Likewise.
9403 (neon_vcle<mode>): Likewise.
9404 (neon_vclt<mode>): Likewise.
9405 (neon_vcage<mode>): Likewise.
9406 (neon_vcagt<mode>): Likewise.
9407 (neon_vtst<mode>): Likewise.
9408 (neon_vabd<mode>): Likewise.
9409 (neon_vabdl<mode>): Likewise.
9410 (neon_vaba<mode>): Likewise.
9411 (neon_vabal<mode>): Likewise.
9412 (neon_vmax<mode>): Likewise.
9413 (neon_vmin<mode>): Likewise.
9414 (neon_vpaddl<mode>): Likewise.
9415 (neon_vpadal<mode>): Likewise.
9416 (neon_vpmax<mode>): Likewise.
9417 (neon_vpmin<mode>): Likewise.
9418 (neon_vrecps<mode>): Likewise.
9419 (neon_vrsqrts<mode>): Likewise.
9420 (neon_vqabs<mode>): Likewise.
9421 (neon_vqneg<mode>): Likewise.
9422 (neon_vcls<mode>): Likewise.
9423 (clz<mode>2): Likewise.
9424 (popcount<mode>2): Likewise.
9425 (neon_vrecpe): Likewise.
9426 (neon_vrsqrte): Likewise.
9427 (neon_vget_lane<mode>_sext_internal): Likewise.
9428 (neon_vget_lane<mode>_zext_internal): Likewise.
9429 (neon_vdup_n<mode>): Likewise.
9430 (neon_vdup_nv2di): Likewise.
9431 (neon_vdpu_lane<mode>_internal): Likewise.
9432 (neon_vswp<mode>): Likewise.
9433 (float<mode><V_cvtto>2): Likewise.
9434 (floatuns<mode><V_cvtto>2): Likewise.
9435 (fix_trunc<mode><V_cvtto>)2): Likewise
9436 (fixuns_trunc<mode><V_cvtto)2): Likewise.
9437 (neon_vcvt<mode>): Likewise.
9438 (neon_vcvtv4sfv4hf): Likewise.
9439 (neon_vcvtv4hfv4sf): Likewise.
9440 (neon_vcvt_n<mode>): Likewise.
9441 (neon_vmovn<mode>): Likewise.
9442 (neon_vqmovn<mode>): Likewise.
9443 (neon_vqmovun<mode>): Likewise.
9444 (neon_vmovl<mode>): Likewise.
9445 (neon_vmul_lane<mode>): Likewise.
9446 (neon_vmull_lane<mode>): Likewise.
9447 (neon_vqdmull_lane<mode>): Likewise.
9448 (neon_vqdmulh_lane<mode>): Likewise.
9449 (neon_vmla_lane<mode>): Likewise.
9450 (neon_vmlal_lane<mode>): Likewise.
9451 (neon_vqdmlal_lane<mode>): Likewise.
9452 (neon_vmls_lane<mode>): Likewise.
9453 (neon_vmlsl_lane<mode>): Likewise.
9454 (neon_vqdmlsl_lane<mode>): Likewise.
9455 (neon_vext<mode>): Likewise.
9456 (neon_vrev64<mode>): Likewise.
9457 (neon_vrev32<mode>): Likewise.
9458 (neon_vrev16<mode>): Likewise.
9459 (neon_vbsl<mode>_internal): Likewise.
9460 (neon_vshl<mode>): Likewise.
9461 (neon_vqshl<mode>): Likewise.
9462 (neon_vshr_n<mode>): Likewise.
9463 (neon_vshrn_n<mode>): Likewise.
9464 (neon_vqshrn_n<mode>): Likewise.
9465 (neon_vqshrun_n<mode>): Likewise.
9466 (neon_vshl_n<mode>): Likewise.
9467 (neon_vqshl_n<mode>): Likewise.
9468 (neon_vqshlu_n<mode>): Likewise.
9469 (neon_vshll_n<mode>): Likewise.
9470 (neon_vsra_n<mode>): Likewise.
9471 (neon_vsri_n<mode>): Likewise.
9472 (neon_vsli_n<mode>): Likewise.
9473 (neon_vtbl1v8qi): Likewise.
9474 (neon_vtbl2v8qi): Likewise.
9475 (neon_vtbl3v8qi): Likewise.
9476 (neon_vtbl4v8qi): Likewise.
9477 (neon_vtbx1v8qi): Likewise.
9478 (neon_vtbx2v8qi): Likewise.
9479 (neon_vtbx3v8qi): Likewise.
9480 (neon_vtbx4v8qi): Likewise.
9481 (neon_vtrn<mode>_internal): Likewise.
9482 (neon_vzip<mode>_internal): Likewise.
9483 (neon_vuzp<mode>_internal): Likewise.
9484 (neon_vld1<mode>): Likewise.
9485 (neon_vld1_lane<mode>): Likewise.
9486 (neon_vld1_dup<mode>): Likewise.
9487 (neon_vld1_dupv2di): Likewise.
9488 (neon_vst1<mode>): Likewise.
9489 (neon_vst1_lane<mode>): Likewise.
9490 (neon_vld2<mode>): Likewise.
9491 (neon_vld2_lane<mode>): Likewise.
9492 (neon_vld2_dup<mode>): Likewise.
9493 (neon_vst2<mode>): Likewise.
9494 (neon_vst2_lane<mode>): Likewise.
9495 (neon_vld3<mode>): Likewise.
9496 (neon_vld3qa<mode>): Likewise.
9497 (neon_vld3qb<mode>): Likewise.
9498 (neon_vld3_lane<mode>): Likewise.
9499 (neon_vld3_dup<mode>): Likewise.
9500 (neon_vst3<mode>): Likewise.
9501 (neon_vst3qa<mode>): Likewise.
9502 (neon_vst3qb<mode>): Likewise.
9503 (neon_vst3_lane<mode>): Likewise.
9504 (neon_vld4<mode>): Likewise.
9505 (neon_vld4qa<mode>): Likewise.
9506 (neon_vld4qb<mode>): Likewise.
9507 (neon_vld4_lane<mode>): Likewise.
9508 (neon_vld4_dup<mode>): Likewise.
9509 (neon_vst4<mode>): Likewise.
9510 (neon_vst4qa<mode>): Likewise.
9511 (neon_vst4qb<mode>): Likewise.
9512 (neon_vst4_lane<mode>): Likewise.
9513 (neon_vec_unpack<US>_lo_<mode>): Likewise.
9514 (neon_vec_unpack<US>_hi_<mode>): Likewise.
9515 (neon_vec_<US>mult_lo_<mode>): Likewise.
9516 (neon_vec_<US>mult_hi_<mode>): Likewise.
9517 (neon_vec_<US>shiftl_<mode>): Likewise.
9518 (neon_unpack<US>_<mode>): Likewise.
9519 (neon_vec_<US>mult_<mode>): Likewise.
9520 (vec_pack_trunc_<mode>): Likewise.
9521 (neon_vec_pack_trunk_<mode>): Likewise.
9522 (neon_vabd<mode>_2): Likewise.
9523 (neon_vabd<mode>_3): Likewise.
9524 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
9525 (thumb2_movsi_vfp): Likewise.
9526 (movdi_vfp): Likewise.
9527 (movdi_vfp_cortexa8): Likewise.
9528 (movhf_vfp_neon): Likewise.
9529 (movhf_vfp): Likewiwse.
9530 (movsf_vfp): Likewiwse.
9531 (thumb2_movsf_vfp): Likewiwse.
9532 (movdf_vfp): Likewise.
9533 (thumb2_movdf_vfp): Likewise.
9534 (movsfcc_vfp): Likewise.
9535 (thumb2_movsfcc_vfp): Likewise.
9536 (movdfcc_vfp): Likewise.
9537 (thumb2_movdfcc_vfp): Likewise.
9538 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
9539 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
9540 (v10_v2c): Likewise.
9541 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
9543 (cortex_a15_neon_int_2): Likewise.
9544 (cortex_a15_neon_int_3): Likewise.
9545 (cortex_a15_neon_int_4): Likewise.
9546 (cortex_a15_neon_int_5): Likewise.
9547 (cortex_a15_neon_vqneg_vqabs): Likewise.
9548 (cortex_a15_neon_vmov): Likewise.
9549 (cortex_a15_neon_vaba): Likewise.
9550 (cortex_a15_neon_vaba_qqq): Likewise.
9551 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9552 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9553 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
9554 scalar_64_32_long_scalar): Likewise.
9555 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9556 (cortex_a15_neon_mla_qqq_8_16): Likewise.
9557 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9558 lotype_qdd_64_32_long): Likewise.
9559 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9560 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9561 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
9562 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9563 (cortex_a15_neon_shift_1): Likewise.
9564 (cortex_a15_neon_shift_2): Likewise.
9565 (cortex_a15_neon_shift_3): Likewise.
9566 (cortex_a15_neon_vshl_ddd): Likewise.
9567 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9568 (cortex_a15_neon_vsra_vrsra): Likewise.
9569 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
9570 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
9571 (cortex_a15_neon_fp_vmul_ddd): Likewise.
9572 (cortex_a15_neon_fp_vmul_qqd): Likewise.
9573 (cortex_a15_neon_fp_vmla_ddd): Likewise.
9574 (cortex_a15_neon_fp_vmla_qqq): Likewise.
9575 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
9576 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
9577 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9578 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9579 (cortex_a15_neon_bp_simple): Likewise.
9580 (cortex_a15_neon_bp_2cycle): Likewise.
9581 (cortex_a15_neon_bp_3cycle): Likewise.
9582 (cortex_a15_neon_vld1_1_2_regs): Likewise.
9583 (cortex_a15_neon_vld1_3_4_regs): Likewise.
9584 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9585 (cortex_a15_neon_vld2_4_regs): Likewise.
9586 (cortex_a15_neon_vld3_vld4): Likewise.
9587 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9588 (cortex_a15_neon_vst1_3_4_regs): Likewise.
9589 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
9590 (cortex_a15_neon_vst3_vst4): Likewise.
9591 (cortex_a15_neon_vld1_vld2_lane): Likewise.
9592 (cortex_a15_neon_vld3_vld4_lane" 10
9593 (cortex_a15_neon_vst1_vst2_lane): Likewise.
9594 (cortex_a15_neon_vst3_vst4_lane): Likewise.
9595 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
9596 (cortex_a15_neon_ldm_2): Likewise.0
9597 (cortex_a15_neon_stm_2): Likewise.
9598 (cortex_a15_neon_mcr): Likewise.
9599 (cortex_a15_neon_mcr_2_mcrr): Likewise.
9600 (cortex_a15_neon_mrc): Likewise.
9601 (cortex_a15_neon_mrrc): Likewise.
9602 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9604 (cortex_a15_alu_shift): Likewise.
9605 (cortex_a15_alu_shift_reg): Likewise.
9606 (cortex_a15_mult32): Likewise.
9607 (cortex_a15_mult64): Likewise.
9608 (cortex_a15_block): Likewise.
9609 (cortex_a15_branch): Likewise.
9610 (cortex_a15_load1): Likewise.
9611 (cortex_a15_load3): Likewise.
9612 (cortex_a15_store1): Likewise.
9613 (cortex_a15_store3): Likewise.
9614 (cortex_a15_call): Likewise.
9615 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
9616 (cortex_a5_f2r): Likewise.
9617 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
9619 (cortex_a53_f2r): Likewise.
9620 * config/arm/cortex-a7.md
9621 (cortex_a7_branch): Update for attribute change.
9622 (cortex_a7_call): Likewise.
9623 (cortex_a7_alu_imm): Likewise.
9624 (cortex_a7_alu_reg): Likewise.
9625 (cortex_a7_alu_shift): Likewise.
9626 (cortex_a7_mul): Likewise.
9627 (cortex_a7_load1): Likewise.
9628 (cortex_a7_store1): Likewise.
9629 (cortex_a7_load2): Likewise.
9630 (cortex_a7_store2): Likewise.
9631 (cortex_a7_load3): Likewise.
9632 (cortex_a7_store3): Likewise.
9633 (cortex_a7_load4): Likewise.
9634 (cortex_a7_store4): Likewise.
9635 (cortex_a7_fpalu): Likewise.
9636 (cortex_a7_fconst): Likewise.
9637 (cortex_a7_fpmuls): Likewise.
9638 (cortex_a7_neon_mul): Likewise.
9639 (cortex_a7_fpmacs): Likewise.
9640 (cortex_a7_neon_mla: Likewise.
9641 (cortex_a7_fpmuld: Likewise.
9642 (cortex_a7_fpmacd: Likewise.
9643 (cortex_a7_fpfmad: Likewise.
9644 (cortex_a7_fdivs: Likewise.
9645 (cortex_a7_fdivd: Likewise.
9646 (cortex_a7_r2f: Likewise.
9647 (cortex_a7_f2r: Likewise.
9648 (cortex_a7_f_flags: Likewise.
9649 (cortex_a7_f_loads: Likewise.
9650 (cortex_a7_f_loadd: Likewise.
9651 (cortex_a7_f_stores: Likewise.
9652 (cortex_a7_f_stored: Likewise.
9653 (cortex_a7_neon): Likewise.
9654 * config/arm/cortex-a8-neon.md
9655 (cortex_a8_neon_mrc): Update for attribute change.
9656 (cortex_a8_neon_mrrc): Likewise.
9657 (cortex_a8_neon_int_1): Likewise.
9658 (cortex_a8_neon_int_2): Likewise.
9659 (cortex_a8_neon_int_3): Likewise.
9660 (cortex_a8_neon_int_4): Likewise.
9661 (cortex_a8_neon_int_5): Likewise.
9662 (cortex_a8_neon_vqneg_vqabs): Likewise.
9663 (cortex_a8_neon_vmov): Likewise.
9664 (cortex_a8_neon_vaba): Likewise.
9665 (cortex_a8_neon_vaba_qqq): Likewise.
9666 (cortex_a8_neon_vsma): Likewise.
9667 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9668 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9669 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
9670 long_scalar): Likewise.
9671 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9672 (cortex_a8_neon_mla_qqq_8_16): Likewise.
9673 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9674 long_scalar_qdd_64_32_long): Likewise.
9675 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9676 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9677 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
9678 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9679 (cortex_a8_neon_shift_1): Likewise.
9680 (cortex_a8_neon_shift_2): Likewise.
9681 (cortex_a8_neon_shift_3): Likewise.
9682 (cortex_a8_neon_vshl_ddd): Likewise.
9683 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9684 (cortex_a8_neon_vsra_vrsra): Likewise.
9685 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
9686 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
9687 (cortex_a8_neon_fp_vsum): Likewise.
9688 (cortex_a8_neon_fp_vmul_ddd): Likewise.
9689 (cortex_a8_neon_fp_vmul_qqd): Likewise.
9690 (cortex_a8_neon_fp_vmla_ddd): Likewise.
9691 (cortex_a8_neon_fp_vmla_qqq): Likewise.
9692 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
9693 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
9694 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9695 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9696 (cortex_a8_neon_bp_simple): Likewise.
9697 (cortex_a8_neon_bp_2cycle): Likewise.
9698 (cortex_a8_neon_bp_3cycle): Likewise.
9699 (cortex_a8_neon_ldr): Likewise.
9700 (cortex_a8_neon_str): Likewise.
9701 (cortex_a8_neon_vld1_1_2_regs): Likewise.
9702 (cortex_a8_neon_vld1_3_4_regs): Likewise.
9703 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9704 (cortex_a8_neon_vld2_4_regs): Likewise.
9705 (cortex_a8_neon_vld3_vld4): Likewise.
9706 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9707 (cortex_a8_neon_vst1_3_4_regs): Likewise.
9708 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
9709 (cortex_a8_neon_vst3_vst4): Likewise.
9710 (cortex_a8_neon_vld1_vld2_lane): Likewise.
9711 (cortex_a8_neon_vld3_vld4_lane): Likewise.
9712 (cortex_a8_neon_vst1_vst2_lane): Likewise.
9713 (cortex_a8_neon_vst3_vst4_lane): Likewise.
9714 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
9715 (cortex_a8_neon_mcr): Likewise.
9716 (cortex_a8_neon_mcr_2_mcrr): Likewise.
9717 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
9718 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
9720 (ca9_neon_mrrc): Likewise.
9721 (cortex_a9_neon_int_1): Likewise.
9722 (cortex_a9_neon_int_2): Likewise.
9723 (cortex_a9_neon_int_3): Likewise.
9724 (cortex_a9_neon_int_4): Likewise.
9725 (cortex_a9_neon_int_5): Likewise.
9726 (cortex_a9_neon_vqneg_vqabs): Likewise.
9727 (cortex_a9_neon_vmov): Likewise.
9728 (cortex_a9_neon_vaba): Likewise.
9729 (cortex_a9_neon_vaba_qqq): Likewise.
9730 (cortex_a9_neon_vsma): Likewise.
9731 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9732 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9733 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
9734 long_scalar): Likewise.
9735 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9736 (cortex_a9_neon_mla_qqq_8_16): Likewise.
9737 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9738 long_scalar_qdd_64_32_long): Likewise.
9739 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9740 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9741 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
9742 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9743 (cortex_a9_neon_shift_1): Likewise.
9744 (cortex_a9_neon_shift_2): Likewise.
9745 (cortex_a9_neon_shift_3): Likewise.
9746 (cortex_a9_neon_vshl_ddd): Likewise.
9747 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9748 (cortex_a9_neon_vsra_vrsra): Likewise.
9749 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
9750 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
9751 (cortex_a9_neon_fp_vsum): Likewise.
9752 (cortex_a9_neon_fp_vmul_ddd): Likewise.
9753 (cortex_a9_neon_fp_vmul_qqd): Likewise.
9754 (cortex_a9_neon_fp_vmla_ddd): Likewise.
9755 (cortex_a9_neon_fp_vmla_qqq): Likewise.
9756 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
9757 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
9758 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9759 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9760 (cortex_a9_neon_bp_simple): Likewise.
9761 (cortex_a9_neon_bp_2cycle): Likewise.
9762 (cortex_a9_neon_bp_3cycle): Likewise.
9763 (cortex_a9_neon_ldr): Likewise.
9764 (cortex_a9_neon_str): Likewise.
9765 (cortex_a9_neon_vld1_1_2_regs): Likewise.
9766 (cortex_a9_neon_vld1_3_4_regs): Likewise.
9767 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9768 (cortex_a9_neon_vld2_4_regs): Likewise.
9769 (cortex_a9_neon_vld3_vld4): Likewise.
9770 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9771 (cortex_a9_neon_vst1_3_4_regs): Likewise.
9772 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
9773 (cortex_a9_neon_vst3_vst4): Likewise.
9774 (cortex_a9_neon_vld1_vld2_lane): Likewise.
9775 (cortex_a9_neon_vld3_vld4_lane): Likewise.
9776 (cortex_a9_neon_vst1_vst2_lane): Likewise.
9777 (cortex_a9_neon_vst3_vst4_lane): Likewise.
9778 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
9779 (cortex_a9_neon_mcr): Likewise.
9780 (cortex_a9_neon_mcr_2_mcrr): Likewise.
9781 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
9782 (cortex_a9_fps): Likewise.
9783 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
9785 (cortex_m4_fmuls): Likewise.
9786 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
9788 (cortex_r4_mrc): Likewise.
9789 * config/arm/iterators.md: Update comment referring to neon_type.
9790 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
9791 (iwmmxt_movsi_insn): Likewise.
9792 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
9794 (pj4_core_to_vfp): Likewise.
9795 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
9797 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
9798 (vfp_fstore): Likewise.
9799 * doc/md.texi: Change references to neon_type to refer to type.
9801 2013-09-04 Dodji Seketeli <dodji@redhat.com>
9803 * tree.h (DECL_BUILT_IN): Fix typo in comment.
9805 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
9807 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
9810 2013-09-04 Easwaran Raman <eraman@google.com>
9813 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
9814 (build_and_add_sum): Use it.
9815 (appears_later_in_bb): Simplify code.
9817 2013-09-04 Teresa Johnson <tejohnson@google.com>
9819 * dumpfile.c (dump_finish): Don't close stderr/stdout.
9821 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
9823 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
9825 2013-09-04 Jan Hubicka <jh@suse.cz>
9827 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
9828 * ipa-devirt.c: Include diganostic.h
9829 (odr_type_d): Add types and types_set.
9830 (hash_type_name): Work for types with vtables during LTO.
9831 (odr_hasher::remove): Fix comment; destroy types_set.
9832 (add_type_duplicate): New function,
9833 (get_odr_type): Use it.
9834 (dump_type_inheritance_graph): Dump type duplicates.
9835 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
9837 * tree.c (types_same_for_odr): Give exact answers on types with
9840 2013-09-04 Dodji Seketeli <dodji@redhat.com>
9842 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
9843 explaining their differences.
9845 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
9847 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
9849 2013-09-03 Jeff Law <law@redhat.com>
9851 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
9852 when not threading through a joiner block. Pass joiner/no joiner
9853 state to register_jump_thread.
9854 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
9855 state from argument rather than implying on path length.
9856 Dump the entire jump thread path into debugging dump.
9857 * tree-flow.h (register_jump_thread): Update prototype.
9859 2013-08-29 Xinliang David Li <davidxl@google.com>
9861 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
9862 Remove a trivial gcc_assert.
9864 2013-08-29 Xinliang David Li <davidxl@google.com>
9866 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
9867 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
9868 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
9869 Delay base decl alignment adjustment.
9870 * tree-vectorizer.c (vect_destroy_datarefs): New function.
9871 * tree-vectorizer.h: New data structure.
9872 (set_dr_misalignment): New function.
9873 (dr_misalignment): Ditto.
9874 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
9875 (vectorizable_load): Ditto.
9876 (ensure_base_align): New function.
9877 (vectorize_loops): Add dbg_cnt support.
9878 (execute_vect_slp): Ditto.
9879 * dbgcnt.def: New debug counter.
9880 * Makefile: New dependency.
9882 2013-09-03 Meador Inge <meadori@codesourcery.com>
9886 2013-08-30 Meador Inge <meadori@codesourcery.com>
9888 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
9890 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
9892 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
9893 function descriptor.
9895 2013-09-03 Richard Biener <rguenther@suse.de>
9897 * tree-affine.c (add_elt_to_tree): Fix association issue,
9898 avoid useless converts and make sure to always return a
9899 properly typed result.
9901 2013-09-03 Richard Biener <rguenther@suse.de>
9904 * fold-const.c (negate_expr_p): Fix division case.
9905 (negate_expr): Likewise.
9907 2013-09-03 Richard Biener <rguenther@suse.de>
9910 * tree-streamer-out.c: Include tm.h.
9911 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
9913 2013-09-03 Jan Hubicka <jh@suse.cz>
9915 * tree-profile.c (tree_profiling): Cleanup CFG when done.
9917 2013-09-03 Alan Modra <amodra@gmail.com>
9919 * config.gcc (powerpc*-*-linux*): Add support for little-endian
9920 multilibs to big-endian target and vice versa.
9921 * config/rs6000/t-linux64: Use := assignment on all vars.
9922 (MULTILIB_EXTRA_OPTS): Remove fPIC.
9923 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
9924 * config/rs6000/t-linux64le: New file.
9925 * config/rs6000/t-linux64bele: New file.
9926 * config/rs6000/t-linux64lebe: New file.
9928 2013-09-02 Jan Hubicka <jh@suse.cz>
9930 * ipa-inline-transform.c (inline_transform): Do not
9931 optimize_inline_calls when not optimizing.
9933 2013-09-02 Jan Hubicka <jh@suse.cz>
9935 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
9936 duplicated nodes for assembler names.
9937 * symtab.c (symtab_unregister_node): Do not attempt to unlink
9938 hard registers from assembler name hash.
9940 2013-09-02 Jan Hubicka <jh@suse.cz>
9942 * ipa-split.c (execute_split_functions): Split externally visible
9943 functions called once.
9945 2013-09-02 Martin Jambor <mjambor@suse.cz>
9948 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
9949 linked list. When finding the correct duplicate, also consider also
9950 the caller in additon to its inlined_to node.
9952 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
9954 * config/aarch64/aarch64-simd-builtins.def
9955 (dup_lane_scalar): Remove.
9956 * config/aarch64/aarch64-simd.md
9957 (aarch64_simd_dup): Add 'w->w' alternative.
9958 (aarch64_dup_lane<mode>): Allow for VALL.
9959 (aarch64_dup_lane_scalar<mode>): Remove.
9960 (aarch64_dup_lane_<vswap_width_name><mode>): New.
9961 (aarch64_get_lane_signed<mode>): Add w->w altenative.
9962 (aarch64_get_lane_unsigned<mode>): Likewise.
9963 (aarch64_get_lane<mode>): Likewise.
9964 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
9965 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
9966 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
9967 (VCON): Change container of V2SF.
9968 (vswap_width_name): Likewise.
9969 * config/aarch64/arm_neon.h
9970 (__aarch64_vdup_lane_any): New.
9971 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
9972 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
9973 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
9975 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
9978 * expr.c (emit_move_complex): Do not move complex FP values as parts if
9979 the source or the destination is a single hard register.
9981 2013-09-02 Richard Biener <rguenther@suse.de>
9984 * tree-scalar-evolution.c (instantiate_scev_name): Allow
9987 2013-09-02 Richard Biener <rguenther@suse.de>
9989 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
9990 arithmetic to sizetype.
9992 2013-09-02 Bin Cheng <bin.cheng@arm.com>
9994 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
9995 Find auto-increment use both before and after candidate.
9997 2013-09-02 Marek Polacek <polacek@redhat.com>
9999 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
10001 2013-09-01 Jan Hubicka <jh@suse.cz>
10003 * Makefile.in: Add ipa-profile.o
10004 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
10005 * cgraph.c (struct cgraph_propagate_frequency_data,
10006 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
10007 ipa-profile.c; replace cgraph_ by ipa_ prefix.
10008 * cgraph.h (cgraph_propagate_frequency): Remove.
10009 * ipa-inline-analysis.c: Include ipa-utils.h;
10010 drop duplicated cfgloop.h.
10011 (inline_update_callee_summaries): Update.
10012 * ipa-profile.c: New file.
10013 * ipa-utils.h (ipa_propagate_frequency): Declare.
10014 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
10015 data-streamer.h, value-prof.h.
10016 (symtab_remove_unreachable_nodes): Update profile.
10017 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
10018 account_time_size, cmp_counts, dump_histogram,
10019 ipa_profile_generate_summary, ipa_profile_write_summary,
10020 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
10021 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
10022 Move to ipa-profile.c.
10024 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
10026 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
10028 2013-09-01 Jan Hubicka <jh@suse.cz>
10030 * common.opt (fdevirtualize-speculatively): New function.
10031 * invoke.texi (fdevirtualize-speculatively): Document.
10032 * ipa-devirt.c: Include ipa-inline.h
10033 (likely_target_p): New function.
10034 (ipa_devirt): New function.
10035 (gate_ipa_devirt): New function.
10036 (pass_data_ipa_devirt): New static var.
10037 (pass_ipa_devirt): Likewise.
10038 (make_pass_ipa_devirt): New function.
10039 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
10040 (common_handle_option): Disable devirtualization when
10041 value range profiling is available.
10042 * passes.def (pass_ipa_devirt): Add.
10043 * timever.def (TV_IPA_DEVIRT): New timevar.
10044 * tree-pass.h (make_pass_ipa_devirt):
10046 2013-09-01 Iain Sandoe <iain@codesourcery.com>
10048 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
10049 include sanitize(undefined).
10051 2013-08-31 Diego Novillo <dnovillo@google.com>
10053 * Makefile.in (TREE_CORE_H): Define.
10055 (GTFILES): Add tree-core.h.
10056 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
10058 * tree-core.h: New file.
10059 Move all data structures, enum, typedefs, global
10060 declarations and constants from ...
10061 * tree.h: ... here.
10063 2013-08-31 Jan Hubicka <jh@suse.cz>
10065 * bulitins.c (expand_builtin): Do not early exit for gcov
10066 instrumented functions.
10068 2013-08-31 Marek Polacek <polacek@redhat.com>
10070 * ubsan.c: Include tm_p.h.
10072 2013-08-31 Jan Hubicka <jh@suse.cz>
10074 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
10077 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
10078 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
10079 * tree-cfg.c (verify_gimple_label): ... here.
10080 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
10081 (ipa_merge_profiles): New function.
10082 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
10083 (lto_input_function_body): Likewise.
10084 * ipa-utils.h (ipa_merge_profiles): Declare.
10085 * lto-streamer.h (lto_input_function_body): Update prototype.
10086 (emit_label_in_global_context_p): Remove.
10087 * lto-symtab.c: Include ipa-utils.h
10088 (lto_cgraph_replace_node): Use ipa_merge_profiles.
10090 2013-08-31 Jan Hubicka <jh@suse.cz>
10092 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
10094 2013-08-31 Jan Hubicka <jh@suse.cz>
10096 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
10098 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
10100 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
10101 "cmp" RTX before signed_comparison_operator check to account
10102 for "code" changes.
10104 2013-08-30 Jan Hubicka <jh@suse.cz>
10106 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
10107 (detect_type_change_1): Rename to ...
10108 (detect_type_change): ... this one; early return on non-polymorphic
10110 (detect_type_change_ssa): Add comp_type parameter; update
10111 use of detect_type_change.
10112 (compute_complex_assign_jump_func): Add param_type parameter;
10113 update use of detect_type_change_ssa.
10114 (compute_complex_ancestor_jump_func): Likewise.
10115 (ipa_get_callee_param_type): New function.
10116 (ipa_compute_jump_functions_for_edge): Compute parameter type;
10117 update calls to the jump function computation functions.
10119 2013-08-30 Teresa Johnson <tejohnson@google.com>
10120 Steven Bosscher <steven@gcc.gnu.org>
10122 * cfgrtl.c (fixup_new_cold_bb): New routine.
10123 (commit_edge_insertions): Invoke fixup_partitions.
10124 (find_partition_fixes): New routine.
10125 (fixup_partitions): Ditto.
10126 (verify_hot_cold_block_grouping): Update comments.
10127 (rtl_verify_edges): Invoke find_partition_fixes.
10128 (rtl_verify_bb_pointers): Update comments.
10129 (rtl_verify_bb_layout): Ditto.
10130 * basic-block.h (probably_never_executed_edge_p): Declare.
10131 (fixup_partitions): Ditto.
10132 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
10133 * bb-reorder.c (sanitize_hot_paths): New function.
10134 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
10135 sanitize_hot_paths.
10136 * predict.c (probably_never_executed_edge_p): New routine.
10137 * cfg.c (check_bb_profile): Add partition insanity warnings.
10139 2013-08-30 Meador Inge <meadori@codesourcery.com>
10141 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10143 2013-08-30 Marek Polacek <polacek@redhat.com>
10145 * Makefile.in (ubsan.o): Add.
10146 (c-family/c-ubsan.o): Add.
10147 (builtins.o): Add ubsan.h dependency.
10148 * ubsan.h: New file.
10149 * ubsan.c: New file.
10150 * common.opt: Add -fsanitize=undefined option.
10151 (flag_sanitize): Add variable.
10152 (fsanitize=): Add option. Add Driver.
10153 (fsanitize=thread): Remove option.
10154 (fsanitize=address): Likewise.
10155 (static-libubsan): New option.
10156 * doc/invoke.texi: Document the new flag and -static-libubsan.
10157 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
10158 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
10159 * builtin-attrs.def (ATTR_COLD): Define.
10160 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10161 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
10162 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
10163 * flag-types.h (sanitize_code): New enum.
10164 * opts.c (common_handle_option): Parse command line arguments
10165 of -fsanitize=. Add -fsanitize=unreachable option.
10166 * varasm.c (get_variable_section): Adjust.
10167 (assemble_noswitch_variable): Likewise.
10168 (assemble_variable): Likewise.
10169 (output_constant_def_contents): Likewise.
10170 (categorize_decl_for_section): Likewise.
10171 (place_block_symbol): Likewise.
10172 (output_object_block): Likewise.
10173 * builtins.def: Likewise.
10174 * toplev.c (compile_file): Likewise.
10175 (process_options): Likewise.
10176 * cppbuiltin.c: Likewise.
10177 * tsan.c (tsan_pass): Likewise.
10178 (tsan_gate): Likewise.
10179 (tsan_gate_O0): Likewise.
10180 * cfgexpand.c (partition_stack_vars): Likewise.
10181 (expand_stack_vars): Likewise.
10182 (defer_stack_allocation): Likewise.
10183 (expand_used_vars): Likewise.
10184 * cfgcleanup.c (old_insns_match_p): Likewise.
10185 * asan.c (asan_finish_file): Likewise.
10186 (asan_instrument): Likewise.
10187 (gate_asan): Likewise.
10188 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
10189 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10190 (asan_global_struct): Use pointer_sized_int_node instead
10191 calling build_nonstandard_integer_type.
10192 (initialize_sanitizer_builtins): Likewise.
10193 (asan_finish_file): Likewise.
10194 * gcc.c: Document %{%:function(args):X}.
10195 (static_spec_functions): Add sanitize.
10196 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
10197 store funcval != NULL there.
10198 (do_spec_1): Adjust handle_spec_function caller.
10199 (handle_braces): Allow %:function(args) as condition.
10200 (sanitize_spec_function): New function.
10201 (ADD_STATIC_LIBUBSAN_LIBS): Define.
10202 (LIBUBSAN_SPEC): Likewise.
10203 (LIBUBSAN_EARLY_SPEC): Likewise.
10204 (SANITIZER_SPEC): Handle libubsan.
10205 (SANITIZER_EARLY_SPEC): Likewise.
10206 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
10207 instead of fsanitize=address.
10208 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
10209 instead of fsanitize=address*.
10210 * builtins.c: Include ubsan.h.
10211 (fold_builtin_0): Instrument __builtin_unreachable.
10212 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
10213 instead of flag_asan.
10214 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
10215 (pointer_sized_int_node): Define.
10216 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
10218 2013-08-30 Mike Stump <mikestump@comcast.net>
10220 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
10223 2013-08-29 Jan Hubicka <jh@suse.cz>
10225 * cgraph.c (cgraph_function_body_availability): Handle weakref
10227 * passes.def: Remove pass_fixup_cfg.
10228 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
10229 track when we need to remove functions.
10230 (gate_ipa_inline): Execute inlining always; add comment why.
10231 (pass_data_ipa_inline): Remove TODO_remove_functions.
10232 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
10233 do not produce summaries.
10234 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
10235 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
10236 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
10237 constant pool and vtable.
10239 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
10241 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
10242 New arm_neon.h's internal macros to specify 64-bit constants.
10243 Avoid using stdint.h's macros.
10245 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
10247 * recog.c (verify_changes): Verify that changes[i].old is non-zero
10248 before applying REG_P.
10250 2013-08-30 Jakub Jelinek <jakub@redhat.com>
10252 PR tree-optimization/58277
10253 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
10254 after seeing too many stmts with vdef in between dombb and current
10255 bb, invalidate everything.
10257 2013-08-30 Richard Biener <rguenther@suse.de>
10259 * fold-const.c (fold_single_bit_test): Fix overflow test.
10261 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
10263 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
10264 and which can live in a register, always retrieve the value on entry.
10265 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
10266 passed by invisible reference specially.
10267 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
10268 (vt_add_function_parameter): Correctly deal with a parameter passed by
10269 invisible reference.
10271 2013-08-30 Jan Hubicka <jh@suse.cz>
10273 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
10275 2013-08-30 Richard Biener <rguenther@suse.de>
10277 PR tree-optimization/58228
10278 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
10279 allow invariant loads in nested loop vectorization.
10281 2013-08-30 Richard Biener <rguenther@suse.de>
10283 PR tree-optimization/58223
10284 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
10285 (has_anti_or_output_dependence): ... this and adjust to also
10286 look for output dependences.
10287 (mark_nodes_having_upstream_mem_writes): Adjust.
10288 (rdg_flag_uses): Likewise.
10290 2013-08-30 Richard Biener <rguenther@suse.de>
10292 PR tree-optimization/58010
10293 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
10294 assert that we have a loop-closed PHI.
10296 2013-08-29 Jan Hubicka <jh@suse.cz>
10298 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
10299 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
10300 * lto-section-in.c (lto_free_function_in_decl_state): New function.
10301 (lto_free_function_in_decl_state_for_node): New function.
10303 2013-08-29 Xinliang David Li <davidxl@google.com>
10305 * loop-unroll.c (report_unroll_peel): Minor message change.
10306 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
10307 Emit alignment peeling message with default -fopt-info.
10308 (vect_loop_versioning): Emit loop version info message.
10309 * tree-vectorizer.c (vectorize_loops): Minor message change.
10310 (execute_vect_slp): Ditto.
10312 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
10314 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
10315 of the clone from the DECL_NAME of the original function.
10317 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
10319 * passes.c (register_pass): Add overload.
10320 * tree-pass.h (register_pass): Forward declare it. Add comment.
10322 2013-08-29 Jan Hubicka <jh@suse.cz>
10324 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
10325 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
10326 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
10327 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10328 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
10329 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
10330 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10331 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
10333 2013-08-29 Teresa Johnson <tejohnson@google.com>
10335 * dumpfile.c (dump_loc): Output column number.
10336 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
10337 * doc/invoke.texi: Document optall -fopt-info flag.
10338 * profile.c (read_profile_edge_counts): Use new dump framework.
10339 (compute_branch_probabilities): Ditto.
10340 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
10341 when pass not in any opt group.
10342 * pass_manager.h (pass_manager::get_pass_profile): New method.
10343 * value-prof.c (check_counter): Use new dump framework.
10344 (check_ic_target): Ditto.
10345 * coverage.c (get_coverage_counts): Ditto.
10346 (coverage_init): Setup new dump framework.
10348 2013-08-29 Richard Biener <rguenther@suse.de>
10350 PR tree-optimization/58246
10351 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
10352 handle the dominance check inside a basic-block.
10354 2013-08-29 Richard Biener <rguenther@suse.de>
10356 PR middle-end/57287
10357 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
10358 of default defs that appear in abnormal PHI nodes.
10360 2013-08-29 Richard Biener <rguenther@suse.de>
10362 PR tree-optimization/57685
10363 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
10364 single-use operands to avoid exponential complexity.
10366 2013-08-28 Dehao Chen <dehao@google.com>
10368 * ipa-inline.c (edge_badness): Fix integer underflow.
10370 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
10372 * gtm-builtins.def (_ITM_free): Declare leaf.
10374 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10377 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
10378 (*tls_local_dynamic_base_64_largepic): Likewise.
10379 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
10380 Remove predicate from call operand.
10381 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
10382 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
10384 2013-08-28 Jeff Law <law@redhat.com>
10386 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
10387 checks for the number of predecessors and successors allowed.
10388 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
10389 which require copying a joiner block if there is a request which
10390 is a subpath that requires no joiner block copying.
10392 2013-08-28 Jan Hubicka <jh@suse.cz>
10394 * lto-streamer-out.c (DFS_write_tree_body): Drop
10395 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
10396 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
10397 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
10398 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
10399 Do not read DECL_ERROR_ISSUED.
10400 (unpack_ts_decl_with_vis_value_fields): Do not read
10402 (lto_input_ts_binfo_tree_pointers): Do not read
10403 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
10404 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
10405 write DECL_ERROR_ISSUED..
10406 (pack_ts_decl_with_vis_value_fields): Do not write
10408 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
10409 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
10410 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
10411 * tree.h (tree_decl_common): Update comment.
10412 (DECL_ERROR_ISSUED): Remove.
10414 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10416 PR middle-end/58257
10417 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
10419 2013-08-28 Jan Hubicka <jh@suse.cz>
10421 * builtins.def (free): Declare leaf.
10423 2013-08-27 David Malcolm <dmalcolm@redhat.com>
10425 * gdbhooks.py: New.
10426 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
10427 * configure: Regenerate.
10429 2013-08-27 Martin Jambor <mjambor@suse.cz>
10431 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
10432 (ipa_ancestor_jf_data): Likewise.
10433 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
10434 (ipa_get_jf_pass_through_type_preserved): New function.
10435 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
10436 (ipa_get_jf_ancestor_type_preserved): New function.
10437 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
10438 (ipa_get_jf_ancestor_result): Likewise.
10439 (propagate_vals_accross_pass_through): Use
10440 ipa_get_jf_pass_through_result to do all the value mappings.
10441 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
10442 type_preserved flag.
10443 (ipa_set_jf_cst_copy): New function.
10444 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
10445 (ipa_set_jf_arith_pass_through): Likewise.
10446 (ipa_set_ancestor_jf): Likewise.
10447 (compute_complex_assign_jump_func): Set type_preserved instead of
10449 (ipa_compute_jump_functions_for_edge): Likewise.
10450 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
10451 (update_jump_functions_after_inlining): Update type_preserved.
10452 Explicitely create jump functions when combining one with pass_through.
10453 (ipa_write_jump_function): Stream the type_preserved flags.
10454 (ipa_read_jump_function): Likewise.
10456 2013-08-27 Jakub Jelinek <jakub@redhat.com>
10457 Aldy Hernandez <aldyh@redhat.com>
10459 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
10460 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
10461 * function.h (struct function): Add has_force_vect_loops and
10463 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
10464 * gimple.c (gimple_build_omp_critical): Add KIND argument and
10466 * gimple.def: Update CLAUSES comments.
10467 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
10468 (gimple_build_omp_for): Add argument to prototype.
10469 (gimple_omp_for_kind): New.
10470 (gimple_omp_for_set_kind): New.
10471 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
10472 GOVD_DATA_SHARE_CLASS.
10473 (enum omp_region_type): Add ORT_SIMD.
10474 (gimple_add_tmp_var): Handle ORT_SIMD.
10475 (gimplify_var_or_parm_decl): Same.
10476 (is_gimple_stmt): Same.
10477 (omp_firstprivatize_variable): Same.
10478 (omp_add_variable): Only use splay_tree_insert if lookup failed.
10479 (omp_notice_variable): Handle ORT_SIMD.
10480 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
10481 (omp_check_private): Handle ORT_SIMD.
10482 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10483 OMP_CLAUSE_SAFELEN.
10484 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
10485 Handle OMP_CLAUSE_LASTPRIVATE.
10486 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10487 OMP_CLAUSE_SAFELEN.
10488 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
10489 (gimplify_expr): Handle OMP_SIMD.
10490 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
10491 (expand_GOMP_SIMD_VF): New.
10492 (expand_GOMP_SIMD_LAST_LANE): New.
10493 * internal-fn.def (GOMP_SIMD_LANE): New.
10494 (GOMP_SIMD_VF): New.
10495 (GOMP_SIMD_LAST_LANE): New.
10496 * omp-low.c: Include target.h.
10497 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
10498 OMP_CLAUSE_SAFELEN.
10499 (check_omp_nesting_restrictions): Same.
10501 (lower_rec_simd_input_clauses): New.
10502 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
10504 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
10505 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
10506 (expand_omp_build_assign): New.
10507 (expand_omp_for_init_counts): New.
10508 (expand_omp_for_init_vars): New.
10509 (extract_omp_for_update_vars): New.
10510 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
10511 and rewrite accordingly.
10512 (expand_omp_simd): New.
10513 (expand_omp_for): Use expand_omp_simd.
10514 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
10515 (lower_omp_for): Do not lower the body.
10516 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
10517 in their own loops.
10518 * tree-flow.h (find_omp_clause): Remove prototype.
10519 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
10520 forcing vectorization of the loop, or if flag_tree_vectorize.
10521 (gate_tree_if_conversion): Similarly.
10522 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
10523 gimple_build_omp_for.
10524 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
10525 * tree-parloops (create_parallel_loop): Pass kind argument to
10526 gimple_build_omp_for.
10527 * tree-pretty-print.c (dump_omp_clause): Add cases for
10528 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
10529 OMP_CLAUSE__SIMDUID_.
10530 (dump_generic_node): Handle OMP_SIMD.
10531 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
10532 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
10533 unroll OMP_SIMD loops here.
10534 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
10535 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
10537 (vect_analyze_data_refs): Handle simd loops.
10538 * tree-vect-loop.c (vectorizable_live_operation): Handle
10540 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
10541 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
10542 (vectorizable_load): Same.
10543 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
10544 (struct simduid_to_vf): New.
10545 (simduid_to_vf::hash): New.
10546 (simduid_to-vf::equal): New.
10547 (struct simd_array_to_simduid): New.
10548 (simd_array_to_simduid::hash): New.
10549 (simd_array_to_simduid::equal): New.
10550 (adjust_simduid_builtins): New.
10551 (struct note_simd_array_uses_struct): New.
10552 (note_simd_array_uses_cb): New.
10553 (note_simd_array_uses): New.
10554 (vectorize_loops): Handle simd hints and adjust simd builtins
10556 * tree-vectorizer.h (struct _stmt_vec_info): Add
10557 simd_lane_access_p field.
10558 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
10559 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
10560 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
10561 (omp_clause_code_name): Same.
10562 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
10563 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
10564 * tree.def (OMP_SIMD): New entry.
10565 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
10566 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
10567 (OMP_CLAUSE_DECL): Adjust range for new clauses.
10568 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
10569 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
10570 (OMP_CLAUSE_LINEAR_STEP): New.
10571 (OMP_CLAUSE_SAFELEN_EXPR): New.
10572 (OMP_CLAUSE__SIMDUID__DECL): New.
10573 (find_omp_clause): New prototype.
10575 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
10577 * config/i386/driver-i386.c (host_detect_local_cpu): Update
10578 Haswell processor detection.
10580 2013-08-27 Christian Widmer <shadow@umbrox.de>
10583 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
10584 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
10585 AVX2 capable processors.
10587 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
10589 * config/aarch64/arm_neon.h: Replace all inline asm implementations
10590 of vget_low_* with implementations in terms of other intrinsics.
10592 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
10594 PR middle-end/57219
10595 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
10596 values to -1, 0 and 1.
10598 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
10600 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
10601 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
10602 (<optab><mode>3_insn): Remove.
10603 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
10604 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
10605 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
10606 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
10607 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
10608 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
10609 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
10610 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
10611 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
10612 (ror<mode>3_insn): Likewise.
10613 * config/aarch64/predicates.md (aarch64_simd_register): New.
10615 2013-08-27 Richard Biener <rguenther@suse.de>
10617 PR tree-optimization/57521
10618 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
10619 one edge is non-critical.
10620 (find_phi_replacement_condition): Make sure to use a non-critical
10621 edge. Cleanup and remove old bug workarounds.
10622 (bb_postdominates_preds): Remove.
10623 (if_convertible_loop_p_1): Do not compute post-dominators.
10624 (combine_blocks): Do not free post-dominators.
10625 (main_tree_if_conversion): Likewise.
10626 (pass_data_if_conversion): Add TODO_verify_ssa.
10628 2013-08-27 DJ Delorie <dj@redhat.com>
10630 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
10632 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
10634 * function.c (assign_parm_find_data_types): Set passed_mode and
10635 nominal_mode to the TYPE_MODE of nominal_type for the built
10636 pointer type in case of the struct-pass-by-reference.
10638 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10640 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
10641 (UINT16_TYPE): Change default to "unsigned int".
10643 * config/avr/avr.opt (mfract-convert-truncate): New option.
10644 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
10645 is set, round negative fractional integers according to n1169
10646 when converting to integer types.
10648 2013-08-26 Jan Hubicka <jh@suse.cz>
10650 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
10651 methods can not be called indirectly when their address is not taken.
10653 2013-08-26 Jan Hubicka <jh@suse.cz>
10655 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
10658 2013-08-26 Jan Hubicka <jh@suse.cz>
10660 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
10663 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10665 * reload.c (find_valid_class): Allow classes that do not include
10666 FIRST_PSEUDO_REGISTER - 1.
10668 2013-08-26 Jan Hubicka <jh@suse.cz>
10670 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
10671 fix edge count/frequency when speculation failed; fix type check
10672 for the direct call.
10674 2013-08-26 Jan Hubicka <jh@suse.cz>
10676 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
10678 2013-08-26 Jan Hubicka <jh@suse.cz>
10680 * ipa-inline-transform.c (inline_transform): Be ready for basic block
10681 to be changed by edge redirection.
10683 2013-08-26 Jan Hubicka <jh@suse.cz>
10685 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
10686 formating; add sanity check.
10687 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
10688 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
10689 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
10690 (resolve_noninline_speculation): Update callee keys, too.
10692 2013-08-26 Jan Hubicka <jh@suse.cz>
10694 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
10695 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
10697 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10699 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
10702 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
10704 * config/i386/i386.c (ix86_debug_options): Remove prototype.
10705 (x86_64_elf_select_section): Ditto.
10706 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
10708 (ix86_pass_by_reference): Ditto.
10709 (output_set_got): Ditto.
10710 (ix86_unary_operator_ok): Ditto.
10711 (ix86_expand_builtin): Ditto.
10713 2013-08-23 Jan Hubicka <jh@suse.cz>
10715 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
10717 2013-08-23 Jan Hubicka <jh@suse.cz>
10719 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
10720 (tree_decl_with_vis): Add FINAL field.
10722 2013-08-23 Jeff Law <law@redhat.com>
10724 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
10725 the debugging dump when the expression is fully redundant.
10727 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
10729 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
10730 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
10731 * pretty-print.c (pp_formatted_text_data): Likewise.
10732 (pp_write_text_to_stream): Likewise.
10733 (pp_write_text_as_dot_label_to_stream): Likewise.
10734 (pp_append_r): Likewise.
10735 (pp_format): Likewise.
10736 (pp_flush): Likewise.
10737 (pp_clear_output_area): Likewise.
10738 (pp_append_text): Likewise.
10739 (pp_formatted_text): Likewise.
10740 (pp_remaining_character_count_for_line): Likewise.
10741 (pp_newline): Likewise.
10742 (pp_character): Likewise.
10743 (output_buffer::~output_buffer): Define.
10744 (pretty_printer::~pretty_printer): Destruct output buffer.
10745 * pretty-print.h (output_buffer::~output_buffer): Declare.
10746 (pretty_printer::~pretty_printer): Declare virtual.
10748 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
10751 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
10752 HOST_WIDE_INT_M1U): New macros.
10753 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
10754 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
10755 unsigned -1 for lshift.
10756 * cse.c (cse_insn): Likewise.
10757 * double-int.c (rshift_double, lshift_double): Likewise.
10758 * builtins.c (fold_builtin_bitop): Likewise.
10759 * combine.c (force_to_mode): Likewise.
10760 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
10761 * simplify-rtx.c (simplify_const_unary_operation,
10762 simplify_const_binary_operation): Likewise.
10763 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
10764 check_va_list_escapes): Likewise.
10765 * rtlanal.c (nonzero_bits1): Likewise.
10766 * expmed.c (expand_smod_pow2): Likewise.
10767 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
10769 2013-08-23 Jan Hubicka <jh@suse.cz>
10771 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
10772 as having address taken.
10774 2013-08-23 Jan Hubicka <jh@suse.cz>
10776 * ipa-utils.h (method_class_type): Declare.
10777 * ipa-devirt.c (method_class_type): Export.
10779 * cgraphunit.c (analyze_functions): Do basic devirtualization;
10780 do not walk base classes of anonymous types.
10782 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
10784 PR rtl-optimization/58220
10785 PR regression/58221
10786 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
10787 handle SEQUENCE insns properly.
10789 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
10791 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
10793 (pp_newline_and_indent): Likewise.
10794 (pp_separate_with): Likewise.
10795 * pretty-print.c (pp_newline_and_flush): Define.
10796 (pp_newline_and_indent): Likewise.
10797 (pp_separate_with): Likewise.
10799 2013-08-23 Jakub Jelinek <jakub@redhat.com>
10802 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
10803 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
10805 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
10807 * config/i386/predicates.md (ext_sse_reg_operand): New.
10808 * config/i386/i386.md (*movti_internal): Use
10809 predicate to determine if EVEX is needed.
10810 (*movsi_internal): Ditto.
10811 (*movdf_internal): Ditto.
10812 (*movsf_internal): Ditto.
10813 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
10815 2013-08-23 Jakub Jelinek <jakub@redhat.com>
10817 PR tree-optimization/58209
10818 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
10819 (find_tail_calls): Give up for pointer result types if m is non-NULL.
10820 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
10821 emit POINTER_PLUS_EXPR.
10822 (create_tailcall_accumulator): For pointer result type accumulate in
10825 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
10827 * configure.ac: Add backslashes missing from the last change.
10828 * configure: Regenerate.
10830 2013-08-22 Jan Hubicka <jh@susue.cz>
10832 * ipa.c (function_and_variable_visibility): First remember function
10833 was global and then make it local.
10835 2013-08-22 Julian Brown <julian@codesourcery.com>
10837 * configure.ac: Add aarch64 to list of arches which use "nop" in
10839 * configure: Regenerate.
10841 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10843 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
10844 gnu_libc_has_function.
10845 * config/s390/tpf.h: Likewise.
10847 2013-08-22 Jan Hubicka <jh@susue.cz>
10849 * timevar.c (validate_phases): Add cast.
10851 2013-08-22 Jan Hubicka <jh@susue.cz>
10853 * timevar.c (validate_phases): Use size_t for memory.
10854 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
10856 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
10858 * pretty-print.h (output_buffer::output_buffer): Declare.
10859 (pretty_printer::pretty_printer): Likewise.
10860 (pp_construct): Remove.
10861 * pretty-print.c (output_buffer::output_buffer): Define.
10862 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
10863 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
10864 (print_gimple_expr): Likewise.
10865 (print_gimple_seq): Likewise.
10866 (gimple_dump_bb): Likewise.
10867 * sched-vis.c (dump_value_slim): Likewise.
10868 (dump_insn_slim): Likewise.
10869 (dump_rtl_slim): Likewise.
10870 (str_pattern_slim): Likewise.
10871 * tree-mudflap.c (mf_varname_tree): Likewise.
10872 * graph.c (print_graph_cfg): Likewise.
10873 (start_graph_dump): Likewise.
10874 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
10876 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
10877 pretty printer initialization.
10878 * coretypes.h (diagnostic_context): Remove superflous type alias
10880 (pretty_printer): Likewise. Declare directly as a class.
10881 (pretty_print_info): Remove declaration as class.
10882 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
10883 and pp_clear_output_area.
10884 (asan_add_global): Likewise.
10886 2013-08-22 Jan Hubicka <jh@suse.cz>
10888 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
10889 * ipa-utils.h (update_type_inheritance_graph): Declare.
10890 (possible_polymorphic_call_target_p): Declare.
10891 (possible_polymorphic_call_target_p): New.
10892 * ipa-devirt.c: Update toplevel comments.
10893 (cached_polymorphic_call_targets): Move up.
10894 (odr_type_d): Move ID down.
10895 (polymorphic_type_binfo_p): Update comment.
10896 (odr_hasher::remove): Likewise;
10897 (get_odr_type): Set anonymous_namespace.
10898 (dump_odr_type): Dump it.
10899 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
10900 (maybe_record_node): Record node in cached_polymorphic_call_targets.
10901 (record_binfo): Add comment.
10902 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
10904 (devirt_node_removal_hook): Do not iCE when cache is freed.
10905 (possible_polymorphic_call_target_p): New predicate.
10906 (update_type_inheritance_graph): New function.
10908 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
10909 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
10910 Sergey Lega <sergey.s.lega@intel.com>
10911 Anna Tikhonova <anna.tikhonova@intel.com>
10912 Ilya Tocar <ilya.tocar@intel.com>
10913 Andrey Turetskiy <andrey.turetskiy@intel.com>
10914 Ilya Verbin <ilya.verbin@intel.com>
10915 Kirill Yukhin <kirill.yukhin@intel.com>
10916 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
10918 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
10919 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
10920 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
10921 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
10922 (OPTION_MASK_ISA_AVX2_UNSET): Update.
10923 (OPTION_MASK_ISA_AVX512F_UNSET): New.
10924 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
10925 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
10926 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
10927 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
10928 OPT_mavx512pf, OPT_mavx512er cases.
10929 * config/i386/constraints.md (v): New constraint.
10930 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
10931 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
10932 (bit_AVX512CD): New.
10933 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
10934 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
10935 * config/i386/i386-c.c (ix86_target_macros_internal):
10936 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
10938 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
10939 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
10940 * config/i386/i386.c (regclass_map, dbx_register_map)
10941 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
10942 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
10943 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
10944 -mavx512pf options.
10945 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
10946 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
10947 -mavx512pf options. Fix formatting.
10948 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
10949 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
10950 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
10951 -mavx512cd, -mavx512pf options.
10952 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
10953 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
10954 (ix86_preferred_output_reload_class): Replace SSE_REGS with
10956 (ix86_hard_regno_mode_ok): Support 512-bit registers.
10957 (ix86_set_reg_reg_cost): Ditto.
10958 (x86_order_regs_for_local_alloc): Ditto.
10959 (MAX_VECT_LEN): Extend to 64-byte.
10960 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
10961 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
10962 (TARGET_AVX512ER, TARGET_AVX512CD): New.
10963 (BIGGEST_ALIGNMENT): Extend to 512-bits.
10964 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
10965 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
10966 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
10967 (SSE_REG_MODE_P): Support new modes.
10968 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
10969 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
10970 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
10971 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
10972 (REG_CLASS_CONTENTS): Add new registers.
10973 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
10974 (EXT_REX_SSE_REGNO_P): New.
10975 (HI_REGISTER_NAMES): Add new registers.
10976 * config/i386/i386.md: Define constants for new registers.
10977 (mode): Add new 512-bit modes.
10978 (prefix): Support evex prefix.
10979 (isa): Support avx512f, noavx512f, fma_avx512f.
10980 (ssemodesuffix): Add new 512-bit modes.
10982 (*movxi_internal_avx512f): Ditto.
10983 (*movdi_internal): Replace constraint "x" with the new constraint "v".
10985 (*movsi_internal): Likewise.
10986 (*movdf_internal): Likewise.
10987 (*movsf_internal): Likewise.
10988 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
10989 (<code><mode>3): Likewise.
10990 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
10992 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
10993 with the new constraint "v".
10994 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
10996 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
10997 with the new constraint "v".
10998 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
10999 (<sse2>_storedqu<avxsizesuffix>): Likewise.
11000 (*<plusminus_insn><mode>3): Likewise.
11001 (<sse>_vm<plusminus_insn><mode>3): Likewise.
11002 (*mul<mode>3): Likewise.
11003 (<sse>_vmmul<mode>3): Likewise.
11004 (<sse>_div<mode>3): Likewise.
11005 (<sse>_vmdiv<mode>3): Likewise.
11006 (<sse>_sqrt<mode>2): Likewise.
11007 (<sse>_vmsqrt<mode>2): Likewise.
11008 (*<code><mode>3_finite): Likewise.
11009 (*<code><mode>3) <smaxmin>: Likewise.
11010 (<sse>_vm<code><mode>3): Likewise.
11011 (*<code><mode>3) <any_logic>: Likewise.
11012 (*fma_fmadd_<mode>): Likewise.
11013 (*fma_fmsub_<mode>): Likewise.
11014 (*fma_fnmadd_<mode>): Likewise.
11015 (*fma_fnmsub_<mode>): Likewise.
11016 (*fma_fmaddsub_<mode>): Likewise.
11017 (*fma_fmsubadd_<mode>): Likewise.
11018 (*fmai_fmadd_<mode>): Likewise.
11019 (*fmai_fmsub_<mode>): Likewise.
11020 (*fmai_fnmadd_<mode>): Likewise.
11021 (*fmai_fnmsub_<mode>): Likewise.
11022 (sse_cvtsi2ss): Likewise.
11023 (sse_cvtsi2ssq): Likewise.
11024 (sse_cvtss2si): Likewise.
11025 (sse_cvtss2si_2): Likewise.
11026 (sse_cvtss2siq): Likewise.
11027 (sse_cvtss2siq_2): Likewise.
11028 (sse_cvttss2si): Likewise.
11029 (sse_cvtss2siq_2): Likewise.
11030 (float<sseintvecmodelower><mode>2): Likewise.
11031 (sse2_cvtsd2si_2): Likewise.
11032 (sse2_cvtsd2siq_2): Likewise.
11033 (*<plusminus_insn><mode>3): Likewise.
11034 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
11035 (*<sse4_1_avx2>_mul<mode>3): Likewise.
11036 (ashr<mode>3): Likewise.
11037 (<shift_insn><mode>3): Likewise.
11038 (avx2_<code><mode>3): Likewise.
11039 (*avx2_<code><mode>3): Likewise.
11040 (*andnot<mode>3): Likewise.
11041 (*<code><mode>3) <any_logic>: Likewise.
11042 (abs<mode>2): Likewise.
11043 (avx2_permvar<mode>): Likewise.
11044 (avx2_perm<mode>_1): Likewise.
11045 (*avx_vpermilp<mode>): Likewise.
11046 (avx_vpermilvar<mode>3): Likewise.
11047 (avx2_ashrv<mode>): Likewise.
11048 (avx2_<shift_insn>v<mode>): Likewise.
11049 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
11051 * doc/rtl.texi: Document XImode.
11053 2013-08-21 Jeff Law <law@redhat.com>
11055 * tree-flow.h (register_jump_thread): Pass vector of edges
11056 instead of each important edge.
11057 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
11058 thread path into a vector and pass that to register_jump_thread.
11059 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
11060 passed in edge vector to the current 3-edge form.
11063 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11065 * dce.c (fini_dce): Call df_analyze again just in case
11066 delete_unmarked_insns removed anything.
11068 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
11070 * reload.h (struct reg_equivs): Rename to ..
11071 (struct reg_equivs_s): .. this.
11073 2013-08-20 Martin Liska <marxin.liska@gmail.com>
11075 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
11077 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11079 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
11081 2013-08-21 Jeff Law <law@redhat.com>
11083 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
11084 simplify assignments too. If the RHS collapses to a singleton
11085 range, then return the value for the range.
11087 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
11089 * config/i386/sse.md (V16): Rename to...
11091 (mov<mode>): Update iterator name.
11092 (*mov<mode>_internal): Ditto.
11093 (push<mode>1): Ditto.
11094 (movmisalign<mode>): Ditto.
11096 2013-08-20 Jan Hubicka <jh@suse.cz>
11099 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
11100 entry for direct edges.
11101 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
11103 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11105 Revert my last two changes, r201865 and r201864:
11108 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11110 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11111 instances can own GC refs.
11113 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11114 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11115 (gcc::context::gt_pch_nx): Likewise.
11116 (gcc::context::gt_pch_nx): Likewise.
11117 * ggc.h (gt_ggc_mx <T>): New.
11118 (gt_pch_nx_with_op <T>): New.
11119 (gt_pch_nx <T>): New.
11120 * passes.c (opt_pass::gt_ggc_mx): New.
11121 (opt_pass::gt_pch_nx): New.
11122 (opt_pass::gt_pch_nx_with_op): New.
11123 (pass_manager::gt_ggc_mx): New.
11124 (pass_manager::gt_pch_nx): New.
11125 (pass_manager::gt_pch_nx_with_op): New.
11126 (pass_manager::operator new): Use
11127 ggc_internal_cleared_alloc_stat rather than xcalloc.
11128 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11129 (pass_manager::gt_ggc_mx): New.
11130 (pass_manager::gt_pch_nx): New.
11131 (pass_manager::gt_pch_nx_with_op): New.
11132 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11133 (opt_pass::operator new): New.
11134 (opt_pass::gt_ggc_mx): New.
11135 (opt_pass::gt_pch_nx): New.
11136 (opt_pass::gt_pch_nx_with_op): New.
11139 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11141 * Makefile.in (GTFILES): Add context.h.
11142 * context.c (gcc::context::operator new): New.
11143 (gcc::context::gt_ggc_mx): New.
11144 (gcc::context::gt_pch_nx): New.
11145 (gcc::context::gt_pch_nx): New.
11146 * context.h (gcc::context): Add GTY((user)) marking.
11147 (gcc::context::operator new): New.
11148 (gcc::context::gt_ggc_mx): New.
11149 (gcc::context::gt_pch_nx): New.
11150 (gcc::context::gt_pch_nx): New.
11151 (g): Add GTY marking.
11152 (gt_ggc_mx (gcc::context *)): New.
11153 (gt_pch_nx (gcc::context *)): New.
11154 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11155 void *cookie)): New.
11156 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11158 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11160 * dce.c (fini_dce): Call df_analyze again just in case
11161 delete_unmarked_insns removed anything.
11163 2013-08-20 Teresa Johnson <tejohnson@google.com>
11165 PR rtl-optimizations/57451
11166 * final.c (reemit_insn_block_notes): Prevent lexical blocks
11167 from crossing split section boundaries.
11169 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11171 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
11172 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
11173 with MULTLIB_DEFAULTS.
11175 2013-08-20 Nick Clifton <nickc@redhat.com>
11177 * target.def (narrow_volatile_bitfield): Note that the default
11178 value is false, not !TARGET_STRICT_ALIGN.
11179 * doc/tm.texi: Regenerate.
11181 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
11183 Fix LIB_SPEC for systems without libpthread.
11185 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
11186 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
11188 * config/i386/linux-common.h: Likewise.
11189 * config/mips/linux-common.h: Likewise.
11191 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11193 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
11196 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11198 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11199 instances can own GC refs.
11201 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11202 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11203 (gcc::context::gt_pch_nx): Likewise.
11204 (gcc::context::gt_pch_nx): Likewise.
11205 * ggc.h (gt_ggc_mx <T>): New.
11206 (gt_pch_nx_with_op <T>): New.
11207 (gt_pch_nx <T>): New.
11208 * passes.c (opt_pass::gt_ggc_mx): New.
11209 (opt_pass::gt_pch_nx): New.
11210 (opt_pass::gt_pch_nx_with_op): New.
11211 (pass_manager::gt_ggc_mx): New.
11212 (pass_manager::gt_pch_nx): New.
11213 (pass_manager::gt_pch_nx_with_op): New.
11214 (pass_manager::operator new): Use
11215 ggc_internal_cleared_alloc_stat rather than xcalloc.
11216 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11217 (pass_manager::gt_ggc_mx): New.
11218 (pass_manager::gt_pch_nx): New.
11219 (pass_manager::gt_pch_nx_with_op): New.
11220 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11221 (opt_pass::operator new): New.
11222 (opt_pass::gt_ggc_mx): New.
11223 (opt_pass::gt_pch_nx): New.
11224 (opt_pass::gt_pch_nx_with_op): New.
11226 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11228 * Makefile.in (GTFILES): Add context.h.
11229 * context.c (gcc::context::operator new): New.
11230 (gcc::context::gt_ggc_mx): New.
11231 (gcc::context::gt_pch_nx): New.
11232 (gcc::context::gt_pch_nx): New.
11233 * context.h (gcc::context): Add GTY((user)) marking.
11234 (gcc::context::operator new): New.
11235 (gcc::context::gt_ggc_mx): New.
11236 (gcc::context::gt_pch_nx): New.
11237 (gcc::context::gt_pch_nx): New.
11238 (g): Add GTY marking.
11239 (gt_ggc_mx (gcc::context *)): New.
11240 (gt_pch_nx (gcc::context *)): New.
11241 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11242 void *cookie)): New.
11243 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11245 2013-08-20 Alan Modra <amodra@gmail.com>
11248 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
11249 (rs6000_emit_epilogue): Likewise.
11251 2013-08-19 Dehao Chen <dehao@google.com>
11253 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
11255 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
11256 Jakub Jelinek <jakub@redhat.com>
11258 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
11259 (BUILT_IN_FABSD64): Likewise.
11260 (BUILT_IN_FABSD128): Likewise.
11261 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
11262 (fold_builtin_1): Likewise.
11263 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
11264 destination and source operands.
11265 (*abstd2_fpr): Likewise.
11266 (*nabstd2_fpr): Likewise.
11268 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
11270 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
11273 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
11275 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
11278 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
11280 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
11281 * builtins.c (default_libc_has_function): New.
11282 (gnu_libc_has_function): Ditto.
11283 (no_c99_libc_has_function): Ditto.
11284 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
11285 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
11286 (fold_builtin_sincos): Likewise.
11287 (fold_builtin_cexp): Likewise.
11288 * builtins.def (DEF_C94_BUILTIN): Likewise.
11289 (DEF_C99_BUILTIN): Likewise.
11290 (DEF_C99_C90RES_BUILTIN): Likewise.
11291 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
11292 definitions to using this define.
11293 * config/darwin-protos.h (darwin_libc_has_function): New.
11294 * config/darwin.c (darwin_libc_has_function): Ditto.
11295 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
11296 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
11297 * config/darwin.h: Ditto.
11298 * config/elfos.h: Ditto.
11299 * config/freebsd.h: Ditto.
11300 * config/i386/cygming.h: Ditto.
11301 * config/i386/djgpp.h: Ditto.
11302 * config/i386/i386-interix.h: Ditto.
11303 * config/microblaze/microblaze.h: Ditto.
11304 * config/mmix/mmix.h: Ditto.
11305 * config/gnu-user.h: Ditto.
11306 * config/ia64/hpux.h: Ditto.
11307 * config/pa/pa-hpux.h: Ditto.
11308 * config/pdp11/pdp11.h: Ditto.
11309 * config/picochip/picochip.h: Ditto.
11310 * config/linux.h: Ditto.
11311 * config/netbsd.h: Ditto.
11312 * config/openbsd.h: Ditto.
11313 * config/rs6000/aix43.h: Ditto.
11314 * config/rs6000/aix51.h: Ditto.
11315 * config/rs6000/aix52.h: Ditto.
11316 * config/rs6000/aix53.h: Ditto.
11317 * config/rs6000/aix61.h: Ditto.
11318 * config/rs6000/darwin.h: Ditto.
11319 * config/rs6000/linux.h: Ditto.
11320 * config/rs6000/linux64.h: Ditto.
11321 * config/s390/tpf.h: Ditto.
11322 * config/sol2-10.h: Ditto.
11323 * config/sol2.h: Ditto.
11324 * config/vms/vms.h: Ditto.
11325 * config/vxworks.h: Ditto.
11326 * config/linux-android.c (linux_android_libc_has_function):
11327 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
11328 * config/linux-protos.h (linux_android_libc_has_function):
11330 * config/i386/i386.c (ix86_libc_has_function): New.
11331 * config/i386/i386-protos.h
11332 (ix86_libc_has_function): New declaration.
11333 * config/i386/i386.md
11334 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
11335 ("isinf<mode>2): Likewise.
11336 * convert.c (convert_to_integer): Using new target hook
11337 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
11338 TARGET_C99_FUNCTIONS.
11339 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
11340 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
11341 * coretypes.h (function_class): New enum for different
11342 classes of functions.
11343 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
11344 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
11345 (TARGET_HAS_SINCOS): Likewise.
11346 (TARGET_LIBC_HAS_FUNCTION): New.
11347 * doc/tm.texi: Regenerated.
11348 * targhooks.h (default_libc_has_function): New declaration.
11349 (no_c99_libc_has_function): Ditto.
11350 (gnu_libc_has_function): Ditto.
11351 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
11352 and TARGET_HAS_SINCOS.
11354 2013-08-18 Jan Hubicka <jh@suse.cz>
11356 * Makeifle-in (ipa-devirt.o): New.
11357 (GTFILES): Add ipa-utils.h and ipa-devirt.c
11358 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
11359 (analyze_functions): Look into possible targets of polymorphic call.
11360 * dumpfile.c (dump_files): Add type-inheritance dump.
11361 * dumpfile.h (TDI_inheritance): New.
11362 * ipa-devirt.c: New file.
11363 * ipa-utils.h (odr_type_d): Forward declare.
11364 (odr_type): New type.
11365 (build_type_inheritance_graph): Declare.
11366 (possible_polymorphic_call_targets): Declare and introduce inline
11367 variant when only edge is pased.
11368 (dump_possible_polymorphic_call_targets): Likewise.
11369 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
11370 * tree.c (type_in_anonymous_namespace_p): Break out from ...
11371 (types_same_for_odr): ... here.
11372 * tree.h (type_in_anonymous_namespace_p): Declare.
11374 2013-08-18 Jakub Jelinek <jakub@redhat.com>
11376 PR tree-optimization/58006
11377 * tree-parloops.c (take_address_of): Don't ICE if get_name
11379 (eliminate_local_variables_stmt): Remove clobber stmts.
11381 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
11383 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
11384 error message is issued for an alias to undefined symbol.
11386 2013-08-18 Jan Hubicka <jh@suse.cz>
11388 * cgraph.c (cgraph_create_indirect_edge): Discover
11389 polymorphic calls and record basic info into indirect_info.
11390 * gimple-fold.c (gimple_fold_call): When doing BINFO based
11391 devirtualization, ignore objc function calls.
11392 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
11393 call with no parm index info.
11394 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
11395 * tree.c (virtual_method_call_p): New function.
11396 * tree.h (virtual_method_call_p): Declare.
11398 2013-08-16 Jan Hubicka <jh@suse.cz>
11400 PR middle-end/58179
11401 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
11403 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
11405 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
11408 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11410 * gengtype.c (type_for_name): Add special-case support for
11411 locating types within the "gcc::" namespace.
11412 (open_base_files): Emit a "using namespace gcc" directive.
11414 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11417 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
11418 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
11420 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
11421 array instead of each individual operand as a separate argument.
11422 (emit_fusion_gpr_load): Likewise.
11423 (expand_fusion_gpr_load): Add new function declaration.
11425 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
11426 signature to have the operands passed as an array, instead of as
11427 separate arguments. Allow ZERO_EXTEND to be in the memory
11428 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
11429 depend on the register live/dead flags when peepholes are run.
11430 (expand_fusion_gpr_load): New function to be called from the
11431 peephole2 pass, to change the register that addis sets to be the
11433 (emit_fusion_gpr_load): Change the calling signature to have the
11434 operands passed as an array, instead of as separate arguments.
11435 Allow ZERO_EXTEND to be in the memory address, and also
11436 SIGN_EXTEND if -mpower8-fusion-sign.
11438 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
11439 unspec enumeration.
11440 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
11441 adjust the register addis loads up in the peephole2 pass. Do not
11442 depend on the register live/dead state when the peephole pass is done.
11444 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11446 * gengtype.c (create_user_defined_type): Ensure that the kind
11447 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
11448 declaration is seen before the GTY((user)) marking.
11450 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
11453 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
11455 2013-08-16 Jan Hubicka <jh@suse.cz>
11457 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
11458 arugment expected_type.
11459 (gimple_fold_call): Use it.
11460 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
11461 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
11462 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
11463 (try_make_edge_direct_virtual_call): Likewise.
11464 * tree.c (obj_type_ref_class): New.
11465 * tree.h (obj_type_ref_class): Use it.
11467 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
11469 * sched-vis.c (rtl_slim_pp_initialized): Remove.
11470 (rtl_slim_pp): Likewise.
11471 (init_rtl_slim_pretty_print): Likewise.
11472 (dump_value_slim): Don't call it. Use local pretty printer.
11473 (dump_insn_slim): Likewise.
11474 (dump_rtl_slim): Likewise.
11475 (str_pattern_slim): Likewise.
11476 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
11479 2013-08-16 Jakub Jelinek <jakub@redhat.com>
11481 PR tree-optimization/58164
11482 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
11483 walk gimple_goto_dest of GIMPLE_GOTO.
11485 PR tree-optimization/58165
11486 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
11487 bi_call must be the last stmt in a bb, don't split_block, instead
11488 use fallthru edge from it and give up if there is none.
11489 Release conds vector when returning early.
11491 2013-08-14 Xinliang David Li <davidxl@google.com>
11493 * config/i386/i386.c (ix86_option_override_internal):
11494 Remove unused variable and field.
11496 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11499 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
11500 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
11501 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
11502 and Linux, correct BLKmode alignment when 128-bit alignment is
11503 required and compatibility flag is not set.
11504 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
11505 for zero-size arguments when compatibility flag is not set.
11507 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11509 PR tree-optimization/58145
11510 * tree-sra.c (build_ref_for_offset): If prev_base has
11511 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
11513 2013-08-14 Xinliang David Li <davidxl@google.com>
11515 * config/i386/i386.c (ix86_option_override_internal):
11516 Fix uninitialized variable error.
11518 2013-08-14 Xinliang David Li <davidxl@google.com>
11520 * config/i386/i386.opt: Define two new options.
11521 * config/i386/x86-tune.def: Add arch selector field in macros.
11522 * config/i386/i386.h: Adjust macro definition.
11523 * config/i386/i386.c (ix86_option_override_internal):
11525 (parse_mtune_ctrl_str): New function.
11526 (set_ix86_tune_features): New function.
11527 (ix86_function_specific_restore): Call the new helper function.
11529 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
11531 PR rtl-optimization/57662
11532 * sel-sched.c (code_motion_process_successors): When the current insn
11533 is removed after the recursive traversal, break from the loop.
11534 Add comments and debug printouts.
11536 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11537 Alexandre Oliva <aoliva@redhat.com>
11540 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
11541 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
11542 there also UNSPEC_PLTOFF.
11544 2013-08-14 Marek Polacek <polacek@redhat.com>
11546 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
11547 NUM_CONDITIONS bit positions.
11549 2013-08-13 Cary Coutant <ccoutant@google.com>
11551 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
11552 (attr_checksum): Hash vector contents instead of pointer.
11553 (attr_checksum_ordered): Likewise.
11555 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
11557 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
11558 when Pmode != word_mode. Add length_address attribute.
11559 (sse3_monitor_<mode>): Merge from sse3_monitor and
11560 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
11561 Pmode != word_mode. Update insn length attribute.
11562 * config/i386/i386.c (ix86_option_override_internal): Update
11563 ix86_gen_monitor selection for merged sse3_monitor insn.
11565 2013-08-13 Julian Brown <julian@codesourcery.com>
11567 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
11568 perform invalid legitimization on greater-than-word-size modes for
11569 TARGET_E500_DOUBLE.
11571 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
11573 * ira.c (setup_class_translate_array): Use aclass instead of cl
11574 for classes not fully covered by allocno classes.
11576 2013-08-13 Jakub Jelinek <jakub@redhat.com>
11578 PR tree-optimization/57661
11579 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
11580 * tree-inline.c (tree_function_versioning): Initialize it.
11581 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
11582 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
11583 that is not being copied.
11586 * asan.c (instrument_strlen_call): Fix typo in comment.
11587 Use char * type even for the lhs of POINTER_PLUS_EXPR.
11589 2013-08-13 Steve Ellcey <sellcey@mips.com>
11591 * config/mips/mips.md (prefetch): Use lw instead of ld on
11592 loongson in 32bit mode.
11594 2013-08-13 Nick Clifton <nickc@redhat.com>
11596 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
11598 2013-08-13 Jan Hubicka <jh@suse.cz>
11600 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
11601 introduced edge; fix typo in sanity check.
11602 (cgraph_resolve_speculation): Export; improve diagnostic.
11603 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
11604 speculation at type mismatch.
11605 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
11606 (cgraph_resolve_speculation): Declare.
11607 (symtab_can_be_discarded): New function.
11608 * value-prof.c (gimple_ic_transform): Remove actual transform code.
11609 * ipa-inline-transform.c (speculation_removed): New global var.
11610 (clone_inlined_nodes): See if speculation can be removed.
11611 (inline_call): If speculations was removed, we growths may not match.
11612 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
11613 (speculation_useful_p): New function.
11614 (resolve_noninline_speculation): New function.
11615 (inline_small_functions): Resolve useless speculations.
11616 * ipa-inline.h (speculation_useful_p): Declare
11617 * ipa.c (can_replace_by_local_alias): Simplify.
11618 (ipa_profile): Produce speculative calls in non-lto, too;
11619 add simple cost model; produce local aliases.
11621 2013-08-13 David Malcolm <dmalcolm@redhat.com>
11623 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
11626 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
11628 * config/i386/i386.c (ix86_function_versions): Use error + inform.
11630 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
11632 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
11633 iterator instead of X87MODEF.
11635 2013-08-12 Perez Read <netfirewall@gmail.com>
11638 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
11639 operand 0 for intel asm alternative.
11640 (*movabs<mode>_2): Ditto for operand 1.
11642 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
11644 * config/aarch64/arm_none.h
11645 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
11647 2013-08-12 Nick Clifton <nickc@redhat.com>
11649 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
11651 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
11653 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
11654 expand for QI/HImode operand to produce more effictive code for
11655 unsigned char(short) --> float(double) conversion.
11657 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
11659 * doc/invoke.texi: Mention that -ftls-model does not force the final
11662 2013-08-12 Marek Polacek <polacek@redhat.com>
11663 Marc Glisse <marc.glisse@inria.fr>
11665 PR tree-optimization/57980
11666 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
11667 when creating -1 constant.
11669 2013-08-10 Jan Hubicka <jh@suse.cz>
11671 Workaround binutils PR14342.
11672 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
11673 (gimple_init_edge_profiler): Likewise.
11674 (gimple_gen_ic_func_profiler): Likewise.
11676 2013-08-09 Jan Hubicka <jh@suse.cz>
11678 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
11680 2013-08-09 Xinliang David Li <davidxl@google.com>
11682 * config/i386/stringop.def: New file.
11683 * config/i386/stringop.opt: New file.
11684 * config/i386/i386-opts.h: Include stringopt.def.
11685 * config/i386/i386.opt: Include stringopt.opt.
11686 * config/i386/i386.c (ix86_option_override_internal):
11687 Override default size based stringop inline strategies with options.
11688 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
11691 2013-08-09 Jan Hubicka <jh@suse.cz>
11693 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
11695 2013-08-09 Jan Hubicka <jh@suse.cz>
11697 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
11699 (dump_cgraph_node): Dump profile-id.
11700 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
11701 and common_target_probability.
11702 * lto-cgraph.c (lto_output_edge): Stream common targets.
11703 (lto_output_node): Stream profile ids.
11704 (input_node): Stream profile ids.
11705 (input_edge): Stream common targets.
11706 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
11707 * ipa.c: Include value-prof.h
11708 (ipa_profile_generate_summary): Turn indirect call statement histograms
11709 into common targets.
11710 (ipa_profile): Turn common targets into speculative edges.
11712 2013-08-09 Jan Hubicka <jh@suse.cz>
11714 * cgraph.h (cgraph_node): Add profile_id.
11715 * value-prof.c (cgraph_node_map): Turn into pointer_map.
11716 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
11717 (del_node_map): Update.
11718 (find_func_by_funcdef_no): Replace by ...
11719 (find_func_by_profile_id): ... this one.
11720 (gimple_ic_transform): Do not remove useful histograms when
11721 speculation is not done; dump info when indirect call removal
11723 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
11724 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
11725 (__gcov_indirect_call_profiler_v2): .. this one.
11726 * profile.h (init_node_map): Update.
11727 * coverage.c (coverage_compute_profile_id): New function.
11728 * coverage.h (coverage_compute_profile_id): Declare.
11729 * tree-profile.c (init_ic_make_global_vars): Make
11730 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
11731 (gimple_init_edge_profiler): Update prototype of
11732 __gcov_indirect_call_profiler.
11733 (gimple_gen_ic_func_profiler): Simplify.
11734 (tree_profiling): Use init_node_map
11736 2013-08-09 Jan Hubicka <jh@suse.cz>
11738 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
11739 non-speculative refs.
11740 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
11741 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
11742 (cgraph_set_call_stmt): Likewise.
11743 (cgraph_create_edge_1): Fix release checking compilatoin;
11744 clear lto_stmt_uid.
11745 (cgraph_free_edge): Free indirect info.
11746 (cgraph_turn_edge_to_speculative): New function.
11747 (cgraph_speculative_call_info): New function.
11748 (cgraph_make_edge_direct): Return direct edge; handle speculation.
11749 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
11750 (dump_cgraph_node): Dump speculation.
11751 (verify_edge_count_and_frequency): Accept speculative edges.
11752 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
11753 (verify_cgraph_node): Handle speculation.
11754 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
11755 (cgraph_set_call_stmt): Update prototype.
11756 (cgraph_make_edge_direct): Update prototype.
11757 (cgraph_speculative_call_info): Declare.
11758 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
11759 to change; update call of ipa_find_references.
11760 * ipa-ref.c (ipa_record_reference): Fix return value; clear
11761 lto_stmt_uid and speculative flags.
11762 (ipa_dump_references): Dump speculation.
11763 (ipa_clone_references): Clone speculative flag.
11764 (ipa_clone_referring): Likewise.
11765 (ipa_clone_ref): New function.
11766 (ipa_find_reference): Look into lto_stmt_uids
11767 (ipa_clear_stmts_in_references): Do not clear speculative calls.
11768 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
11769 (ipa_find_reference): Update declaration.
11770 (ipa_clone_ref): Declare.
11771 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
11772 stream speculative flag.
11773 (lto_output_ref): Stream statements uids and speculation.
11774 (input_ref): Likewise.
11775 (input_edge): Stream speuclation.
11776 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
11777 (cgraph_set_call_stmt_including_clones): Handle speculation.
11778 * ipa-inline.c (heap_edge_removal_hook): New function.
11779 (inline_small_functions): Register it.
11780 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
11781 also initialize refs.
11782 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
11784 (try_make_edge_direct_simple_call): Likewise.
11785 (try_make_edge_direct_simple_call): Likewise.
11786 (update_indirect_edges_after_inlining): Likewise.
11787 (remove_described_reference): Look proper lto_stmt_uid.
11788 (propagate_controlled_uses): Likewise.
11789 (propagate_controlled_uses): Liekwise.
11790 * tree-inline.c (copy_bb): Copy speculative edges.
11791 (redirect_all_calls): New function.
11792 (copy_cfg_body): Do redirection after loop info is updated.
11793 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
11795 2013-08-09 Jan Hubicka <jh@suse.cz>
11797 * lto-streamer-out.c (output_function): Renumber PHIs.
11798 * lto-streamer-in.c (input_function): Likewise.
11800 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
11802 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
11803 (get_lane_unsigned): Likewise.
11804 (dup_lane_scalar): Likewise.
11805 (get_lane): enable for VALL.
11806 * config/aarch64/aarch64-simd.md
11807 (aarch64_dup_lane_scalar<mode>): Remove.
11808 (aarch64_get_lane_signed<mode>): Likewise.
11809 (aarch64_get_lane_unsigned<mode>): Likewise.
11810 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
11811 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
11812 (aarch64_get_lane<mode>): Enable for all vector modes.
11813 (aarch64_get_lanedi): Remove misleading constraints.
11814 * config/aarch64/arm_neon.h
11815 (__aarch64_vget_lane_any): Define.
11816 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
11817 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
11818 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
11819 * config/aarch64/iterators.md (VDQQH): New.
11820 (VDQQHS): Likewise.
11821 (vwcore): Likewise.
11823 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
11825 * configure.ac: Add GAS check for LEON instructions on SPARC.
11826 * configure: Regenerate.
11827 * config.in: Likewise.
11828 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
11830 * config/sparc/sparc.opt (LEON, LEON3): New masks.
11831 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
11833 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
11834 (AS_LEON_FLAG): New macro.
11835 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
11836 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
11837 Deal with LEON and LEON3 for the memory model.
11838 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
11839 (atomic_compare_and_swap<mode>_1): Likewise.
11840 (*atomic_compare_and_swap<mode>_1): Likewise.
11842 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11844 * config/arm/neon.md (vcond): Fix floating-point vector
11845 comparisons against 0.
11847 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
11849 * lra-constraints.c (emit_spill_move): Remove assert.
11850 (process_alt_operands): Add more debugging
11851 output. Increase reject for spilling into memory. Decrease
11852 reject for reloading scratch.
11853 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
11855 2013-08-08 Steve Ellcey <sellcey@mips.com>
11857 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
11858 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
11859 micromips incompatible. Add nan2008.
11860 (MULTILIB_DIRNAMES): Add nan2008.
11861 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
11862 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
11863 and micromips incompatible. Add nan2008.
11864 (MULTILIB_DIRNAMES): Add nan2008.
11865 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
11867 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
11869 PR rtl-optimization/58079
11870 * combine.c (combine_simplify_rtx): Avoid using SUBST if
11871 simplify_comparison has widened a comparison with an integer.
11873 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11875 * config/arm/neon.md (movmisalign<mode>): Disable when we
11876 don't allow unaligned accesses.
11877 (*movmisalign<mode>_neon_store): Likewise.
11878 (*movmisalign<mode>_neon_load): Likewise.
11879 (*movmisalign<mode>_neon_store): Likewise.
11880 (*movmisalign<mode>_neon_load): Likewise.
11882 2013-08-08 Jan Hubicka <jh@suse.cz>
11884 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
11885 (make_pass_rebuild_cgraph_edges): Also clear references.
11886 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
11887 * ipa-inline-transform.c (inline_transform): Remove all references
11889 * cgraphunit.c (expand_function): Remove all references after
11891 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
11892 (ipa_find_reference): Rewrite to iterator.
11893 (remove_stmt_references): Likewise.
11894 (ipa_clear_stmts_in_references): New function.
11895 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
11896 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
11898 * ipa-split.c (split_function): Remove references in split function.
11900 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
11903 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
11904 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
11906 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
11909 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
11910 suggested mode for the assignment isn't compatible with the
11911 registers required.
11913 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
11916 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
11918 2013-08-07 Xinliang David Li <davidxl@google.com>
11920 * config/i386/i386.opt: New option -mtune-ctrl=.
11921 * config/i386/x86-tune.def: New file.
11922 * config/i386/i386.h: include x86-tune.def.
11923 * config/i386/i386.c (ix86_option_override_internal):
11924 Parsing -mtune-ctrl= option and set tune features.
11926 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
11929 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
11930 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
11933 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
11935 * rtl.h (update_alignments): Declare.
11936 * final.c (grow_label_align): New function extracted from...
11937 (shorten_branches): ...here. Call it.
11938 (update_alignments): New function.
11939 * reorg.c (sibling_labels): New variable.
11940 (get_label_before): Add SIBLING parameter. If it is non-zero, push
11941 the new label along with it onto the sibling_labels vector.
11942 (fill_simple_delay_slots): Adjust call to get_label_before.
11943 (fill_slots_from_thread): Likewise.
11944 (relax_delay_slots): Likewise.
11945 (make_return_insns): Likewise.
11946 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
11948 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
11950 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
11951 document its semantics.
11952 (diagnostic_report_diagnostic): Adjust accordingly.
11954 2013-08-07 David Malcolm <dmalcolm@redhat.com>
11956 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
11957 (sparc_option_override): ...and port to new C++ pass API.
11958 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
11960 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
11962 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
11964 2013-08-06 Caroline Tice <cmtice@google.com>
11966 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
11967 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
11968 * tree-pass.h: Add pass_vtable_verify.
11969 * varasm.c (assemble_variable): Add code to properly set the comdat
11970 section and name for the .vtable_map_vars section.
11971 (assemble_vtyv_preinit_initializer): New function.
11972 (default_sectin_type_flags): Make sure .vtable_map_vars section has
11974 * output.h: Add function decl for assemble_vtv_preinit_initializer.
11975 * vtable-verify.c: New file.
11976 * vtable-verify.h: New file.
11977 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
11978 initialiation levels.
11979 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
11980 * passes.def: Insert pass_vtable_verify.
11981 * aclocal.m4: Reorder includes.
11982 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
11983 -fvtv-counts options.
11984 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
11985 as appropriate, if -fvtable-verify=... is used.
11986 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
11987 -fvtable-verify=... is used.
11988 * Makefile.in (OBJS): Add vtable-verify.o to list.
11989 (vtable-verify.o): Add new build rule.
11990 (GTFILES): Add vtable-verify.c to list.
11991 * common.opt (fvtable-verify=): New flag.
11992 (vtv_priority): Values for fvtable-verify= flag.
11993 (fvtv-counts): New flag.
11994 (fvtv-debug): New flag.
11995 * tree.h (save_vtable_map_decl): New extern function decl.
11997 2013-08-07 David Malcolm <dmalcolm@redhat.com>
11999 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
12000 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
12001 (pass_data_rl78_devirt): ...new pass_data instance and...
12002 (make_pass_rl78_devirt): ...new function.
12003 (rl78_asm_file_start): Port pass registration to new C++ API.
12005 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12007 * coretypes.h (rtl_opt_pass): Add.
12008 (gcc::context): Add.
12009 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
12010 (epiphany_init): Port to new C++ pass API.
12011 (epiphany_optimize_mode_switching): Likewise.
12012 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
12013 (pass_manager::get_pass_mode_switching): New.
12014 (pass_manager::get_pass_peephole2): New.
12015 * mode-switching.c (pass_mode_switching): Add clone method.
12016 * recog.c (pass_peephole2): Add clone method.
12017 (pass_split_all_insns): Add clone method.
12019 2013-08-06 David Malcolm <dmalcolm@redhat.com>
12021 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
12022 (mips_option_override): ...here, porting to new C++ API for passes.
12024 2013-08-06 Jan Hubicka <jh@suse.cz>
12026 * cgraph.c (cgraph_get_body): New function based on lto.c
12028 * cgraph.h (cgraph_get_body): Declare.
12029 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
12031 * cgraphunit.c (expand_function): Get body prior expanding.
12032 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
12033 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
12035 * passes.c (do_per_function_toporder): Get body.
12036 * tree-inline.c (expand_call_inline): Get body prior inlining it.
12037 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
12039 2013-08-06 Martin Jambor <mjambor@suse.cz>
12042 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
12043 is not re-finalized. Rename second parameter to no_collect.
12045 2013-08-06 Martin Jambor <mjambor@suse.cz>
12047 PR middle-end/58041
12048 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
12049 MEM_REF has proper alignment information.
12051 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
12054 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
12056 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
12057 rtx (*) (rtx, ...) with insn_gen_fn.
12058 * genoutput.c (output_insn_data): Cast gen_? function pointers to
12059 insn_gen_fn::stored_funcptr. Add initializer braces.
12061 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12063 Rewrite how instances of passes are cloned to remove assumptions
12064 about their sizes (thus allowing pass subclasses to have
12065 additional data fields, albeit non-GC-managed ones at this point).
12067 * passes.c (make_pass_instance): Now that passes have clone
12068 methods, rewrite this function to eliminate XNEW and memcpy
12069 calls that used hardcoded sizes. Since this function no longer
12070 creates pass instances, rename it to...
12071 (add_pass_instance): ...this. Document the old way that passes were
12072 numbered and flagged, and rework this function to continue using it.
12073 (next_pass_1): Add an initial_pass argument for use by
12075 (position_pass): When adding multiple instances of a pass, use
12076 the pass's clone method, rather than relying on the XNEW/memcpy
12077 within the former make_pass_instance (now add_pass_instance).
12078 (pass_manager::pass_manager): When invoking next_pass_1, also supply
12079 the initial instance of the current pass within the pass manager.
12081 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12083 This is the automated part of the conversion of passes from C
12084 structs to C++ classes.
12086 Patch autogenerated by refactor_passes.py from
12087 https://github.com/davidmalcolm/gcc-refactoring-scripts
12088 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
12090 * asan.c (pass_asan): Convert from a global struct to a subclass of
12091 gimple_opt_pass along with...
12092 (pass_data_asan): ...new pass_data instance and...
12093 (make_pass_asan): ...new function.
12094 (pass_asan_O0): Convert from a global struct to a subclass of
12095 gimple_opt_pass along with...
12096 (pass_data_asan_O0): ...new pass_data instance and...
12097 (make_pass_asan_O0): ...new function.
12098 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
12099 subclass of rtl_opt_pass along with...
12100 (pass_data_inc_dec): ...new pass_data instance and...
12101 (make_pass_inc_dec): ...new function.
12102 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
12103 a subclass of rtl_opt_pass along with...
12104 (pass_data_reorder_blocks): ...new pass_data instance and...
12105 (make_pass_reorder_blocks): ...new function.
12106 (pass_duplicate_computed_gotos): Convert from a global struct to a
12107 subclass of rtl_opt_pass along with...
12108 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
12109 (make_pass_duplicate_computed_gotos): ...new function.
12110 (pass_partition_blocks): Convert from a global struct to a subclass of
12111 rtl_opt_pass along with...
12112 (pass_data_partition_blocks): ...new pass_data instance and...
12113 (make_pass_partition_blocks): ...new function.
12114 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
12115 struct to a subclass of rtl_opt_pass along with...
12116 (pass_data_branch_target_load_optimize1): ...new pass_data instance
12118 (make_pass_branch_target_load_optimize1): ...new function.
12119 (pass_branch_target_load_optimize2): Convert from a global struct to a
12120 subclass of rtl_opt_pass along with...
12121 (pass_data_branch_target_load_optimize2): ...new pass_data instance
12123 (make_pass_branch_target_load_optimize2): ...new function.
12124 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
12125 of rtl_opt_pass along with...
12126 (pass_data_jump): ...new pass_data instance and...
12127 (make_pass_jump): ...new function.
12128 (pass_jump2): Convert from a global struct to a subclass of
12129 rtl_opt_pass along with...
12130 (pass_data_jump2): ...new pass_data instance and...
12131 (make_pass_jump2): ...new function.
12132 * cfgexpand.c (pass_expand): Convert from a global struct to a
12133 subclass of rtl_opt_pass along with...
12134 (pass_data_expand): ...new pass_data instance and...
12135 (make_pass_expand): ...new function.
12136 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
12137 of rtl_opt_pass along with...
12138 (pass_data_free_cfg): ...new pass_data instance and...
12139 (make_pass_free_cfg): ...new function.
12140 (pass_into_cfg_layout_mode): Convert from a global struct to a
12141 subclass of rtl_opt_pass along with...
12142 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
12143 (make_pass_into_cfg_layout_mode): ...new function.
12144 (pass_outof_cfg_layout_mode): Convert from a global struct to a
12145 subclass of rtl_opt_pass along with...
12146 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
12147 (make_pass_outof_cfg_layout_mode): ...new function.
12148 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
12149 struct to a subclass of gimple_opt_pass along with...
12150 (pass_data_build_cgraph_edges): ...new pass_data instance and...
12151 (make_pass_build_cgraph_edges): ...new function.
12152 (pass_rebuild_cgraph_edges): Convert from a global struct to a
12153 subclass of gimple_opt_pass along with...
12154 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
12155 (make_pass_rebuild_cgraph_edges): ...new function.
12156 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
12157 subclass of gimple_opt_pass along with...
12158 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
12160 (make_pass_remove_cgraph_callee_edges): ...new function.
12161 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
12162 struct to a subclass of rtl_opt_pass along with...
12163 (pass_data_stack_adjustments): ...new pass_data instance and...
12164 (make_pass_stack_adjustments): ...new function.
12165 * combine.c (pass_combine): Convert from a global struct to a subclass
12166 of rtl_opt_pass along with...
12167 (pass_data_combine): ...new pass_data instance and...
12168 (make_pass_combine): ...new function.
12169 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
12170 global struct to a subclass of rtl_opt_pass along with...
12171 (pass_data_compare_elim_after_reload): ...new pass_data instance
12173 (make_pass_compare_elim_after_reload): ...new function.
12174 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
12175 of rtl_opt_pass along with...
12176 (pass_data_rtl_cprop): ...new pass_data instance and...
12177 (make_pass_rtl_cprop): ...new function.
12178 * cse.c (pass_cse): Convert from a global struct to a subclass of
12179 rtl_opt_pass along with...
12180 (pass_data_cse): ...new pass_data instance and...
12181 (make_pass_cse): ...new function.
12182 (pass_cse2): Convert from a global struct to a subclass of
12183 rtl_opt_pass along with...
12184 (pass_data_cse2): ...new pass_data instance and...
12185 (make_pass_cse2): ...new function.
12186 (pass_cse_after_global_opts): Convert from a global struct to a
12187 subclass of rtl_opt_pass along with...
12188 (pass_data_cse_after_global_opts): ...new pass_data instance and...
12189 (make_pass_cse_after_global_opts): ...new function.
12190 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
12191 of rtl_opt_pass along with...
12192 (pass_data_ud_rtl_dce): ...new pass_data instance and...
12193 (make_pass_ud_rtl_dce): ...new function.
12194 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
12195 rtl_opt_pass along with...
12196 (pass_data_fast_rtl_dce): ...new pass_data instance and...
12197 (make_pass_fast_rtl_dce): ...new function.
12198 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
12199 a subclass of rtl_opt_pass along with...
12200 (pass_data_df_initialize_opt): ...new pass_data instance and...
12201 (make_pass_df_initialize_opt): ...new function.
12202 (pass_df_initialize_no_opt): Convert from a global struct to a
12203 subclass of rtl_opt_pass along with...
12204 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
12205 (make_pass_df_initialize_no_opt): ...new function.
12206 (pass_df_finish): Convert from a global struct to a subclass of
12207 rtl_opt_pass along with...
12208 (pass_data_df_finish): ...new pass_data instance and...
12209 (make_pass_df_finish): ...new function.
12210 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
12211 rtl_opt_pass along with...
12212 (pass_data_rtl_dse1): ...new pass_data instance and...
12213 (make_pass_rtl_dse1): ...new function.
12214 (pass_rtl_dse2): Convert from a global struct to a subclass of
12215 rtl_opt_pass along with...
12216 (pass_data_rtl_dse2): ...new pass_data instance and...
12217 (make_pass_rtl_dse2): ...new function.
12218 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
12219 subclass of rtl_opt_pass along with...
12220 (pass_data_dwarf2_frame): ...new pass_data instance and...
12221 (make_pass_dwarf2_frame): ...new function.
12222 * except.c (pass_set_nothrow_function_flags): Convert from a global
12223 struct to a subclass of rtl_opt_pass along with...
12224 (pass_data_set_nothrow_function_flags): ...new pass_data instance
12226 (make_pass_set_nothrow_function_flags): ...new function.
12227 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
12228 subclass of rtl_opt_pass along with...
12229 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
12231 (make_pass_convert_to_eh_region_ranges): ...new function.
12232 * final.c (pass_compute_alignments): Convert from a global struct to a
12233 subclass of rtl_opt_pass along with...
12234 (pass_data_compute_alignments): ...new pass_data instance and...
12235 (make_pass_compute_alignments): ...new function.
12236 (pass_final): Convert from a global struct to a subclass of
12237 rtl_opt_pass along with...
12238 (pass_data_final): ...new pass_data instance and...
12239 (make_pass_final): ...new function.
12240 (pass_shorten_branches): Convert from a global struct to a subclass of
12241 rtl_opt_pass along with...
12242 (pass_data_shorten_branches): ...new pass_data instance and...
12243 (make_pass_shorten_branches): ...new function.
12244 (pass_clean_state): Convert from a global struct to a subclass of
12245 rtl_opt_pass along with...
12246 (pass_data_clean_state): ...new pass_data instance and...
12247 (make_pass_clean_state): ...new function.
12248 * function.c (pass_instantiate_virtual_regs): Convert from a global
12249 struct to a subclass of rtl_opt_pass along with...
12250 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
12251 (make_pass_instantiate_virtual_regs): ...new function.
12252 (pass_leaf_regs): Convert from a global struct to a subclass of
12253 rtl_opt_pass along with...
12254 (pass_data_leaf_regs): ...new pass_data instance and...
12255 (make_pass_leaf_regs): ...new function.
12256 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
12257 subclass of rtl_opt_pass along with...
12258 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
12260 (make_pass_thread_prologue_and_epilogue): ...new function.
12261 (pass_match_asm_constraints): Convert from a global struct to a
12262 subclass of rtl_opt_pass along with...
12263 (pass_data_match_asm_constraints): ...new pass_data instance and...
12264 (make_pass_match_asm_constraints): ...new function.
12265 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
12266 subclass of rtl_opt_pass along with...
12267 (pass_data_rtl_fwprop): ...new pass_data instance and...
12268 (make_pass_rtl_fwprop): ...new function.
12269 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
12270 rtl_opt_pass along with...
12271 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
12272 (make_pass_rtl_fwprop_addr): ...new function.
12273 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
12274 rtl_opt_pass along with...
12275 (pass_data_rtl_pre): ...new pass_data instance and...
12276 (make_pass_rtl_pre): ...new function.
12277 (pass_rtl_hoist): Convert from a global struct to a subclass of
12278 rtl_opt_pass along with...
12279 (pass_data_rtl_hoist): ...new pass_data instance and...
12280 (make_pass_rtl_hoist): ...new function.
12281 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
12282 subclass of gimple_opt_pass along with...
12283 (pass_data_lower_cf): ...new pass_data instance and...
12284 (make_pass_lower_cf): ...new function.
12285 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
12286 from a global struct to a subclass of gimple_opt_pass along with...
12287 (pass_data_strength_reduction): ...new pass_data instance and...
12288 (make_pass_strength_reduction): ...new function.
12289 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
12290 of rtl_opt_pass along with...
12291 (pass_data_rtl_ifcvt): ...new pass_data instance and...
12292 (make_pass_rtl_ifcvt): ...new function.
12293 (pass_if_after_combine): Convert from a global struct to a subclass of
12294 rtl_opt_pass along with...
12295 (pass_data_if_after_combine): ...new pass_data instance and...
12296 (make_pass_if_after_combine): ...new function.
12297 (pass_if_after_reload): Convert from a global struct to a subclass of
12298 rtl_opt_pass along with...
12299 (pass_data_if_after_reload): ...new pass_data instance and...
12300 (make_pass_if_after_reload): ...new function.
12301 * init-regs.c (pass_initialize_regs): Convert from a global struct to
12302 a subclass of rtl_opt_pass along with...
12303 (pass_data_initialize_regs): ...new pass_data instance and...
12304 (make_pass_initialize_regs): ...new function.
12305 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
12306 of ipa_opt_pass_d along with...
12307 (pass_data_ipa_cp): ...new pass_data instance and...
12308 (make_pass_ipa_cp): ...new function.
12309 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
12310 global struct to a subclass of gimple_opt_pass along with...
12311 (pass_data_inline_parameters): ...new pass_data instance and...
12312 (make_pass_inline_parameters): ...new function.
12313 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
12314 subclass of gimple_opt_pass along with...
12315 (pass_data_early_inline): ...new pass_data instance and...
12316 (make_pass_early_inline): ...new function.
12317 (pass_ipa_inline): Convert from a global struct to a subclass of
12318 ipa_opt_pass_d along with...
12319 (pass_data_ipa_inline): ...new pass_data instance and...
12320 (make_pass_ipa_inline): ...new function.
12321 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
12322 struct to a subclass of gimple_opt_pass along with...
12323 (pass_data_local_pure_const): ...new pass_data instance and...
12324 (make_pass_local_pure_const): ...new function.
12325 (pass_ipa_pure_const): Convert from a global struct to a subclass of
12326 ipa_opt_pass_d along with...
12327 (pass_data_ipa_pure_const): ...new pass_data instance and...
12328 (make_pass_ipa_pure_const): ...new function.
12329 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
12330 to a subclass of ipa_opt_pass_d along with...
12331 (pass_data_ipa_reference): ...new pass_data instance and...
12332 (make_pass_ipa_reference): ...new function.
12333 * ipa-split.c (pass_split_functions): Convert from a global struct to
12334 a subclass of gimple_opt_pass along with...
12335 (pass_data_split_functions): ...new pass_data instance and...
12336 (make_pass_split_functions): ...new function.
12337 (pass_feedback_split_functions): Convert from a global struct to a
12338 subclass of gimple_opt_pass along with...
12339 (pass_data_feedback_split_functions): ...new pass_data instance and...
12340 (make_pass_feedback_split_functions): ...new function.
12341 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
12342 global struct to a subclass of simple_ipa_opt_pass along with...
12343 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
12345 (make_pass_ipa_function_and_variable_visibility): ...new function.
12346 (pass_ipa_free_inline_summary): Convert from a global struct to a
12347 subclass of simple_ipa_opt_pass along with...
12348 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
12349 (make_pass_ipa_free_inline_summary): ...new function.
12350 (pass_ipa_whole_program_visibility): Convert from a global struct to a
12351 subclass of ipa_opt_pass_d along with...
12352 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
12354 (make_pass_ipa_whole_program_visibility): ...new function.
12355 (pass_ipa_profile): Convert from a global struct to a subclass of
12356 ipa_opt_pass_d along with...
12357 (pass_data_ipa_profile): ...new pass_data instance and...
12358 (make_pass_ipa_profile): ...new function.
12359 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
12360 ipa_opt_pass_d along with...
12361 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
12362 (make_pass_ipa_cdtor_merge): ...new function.
12363 * ira.c (pass_ira): Convert from a global struct to a subclass of
12364 rtl_opt_pass along with...
12365 (pass_data_ira): ...new pass_data instance and...
12366 (make_pass_ira): ...new function.
12367 (pass_reload): Convert from a global struct to a subclass of
12368 rtl_opt_pass along with...
12369 (pass_data_reload): ...new pass_data instance and...
12370 (make_pass_reload): ...new function.
12371 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
12372 subclass of rtl_opt_pass along with...
12373 (pass_data_cleanup_barriers): ...new pass_data instance and...
12374 (make_pass_cleanup_barriers): ...new function.
12375 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
12376 of rtl_opt_pass along with...
12377 (pass_data_loop2): ...new pass_data instance and...
12378 (make_pass_loop2): ...new function.
12379 (pass_rtl_loop_init): Convert from a global struct to a subclass of
12380 rtl_opt_pass along with...
12381 (pass_data_rtl_loop_init): ...new pass_data instance and...
12382 (make_pass_rtl_loop_init): ...new function.
12383 (pass_rtl_loop_done): Convert from a global struct to a subclass of
12384 rtl_opt_pass along with...
12385 (pass_data_rtl_loop_done): ...new pass_data instance and...
12386 (make_pass_rtl_loop_done): ...new function.
12387 (pass_rtl_move_loop_invariants): Convert from a global struct to a
12388 subclass of rtl_opt_pass along with...
12389 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
12390 (make_pass_rtl_move_loop_invariants): ...new function.
12391 (pass_rtl_unswitch): Convert from a global struct to a subclass of
12392 rtl_opt_pass along with...
12393 (pass_data_rtl_unswitch): ...new pass_data instance and...
12394 (make_pass_rtl_unswitch): ...new function.
12395 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
12396 subclass of rtl_opt_pass along with...
12397 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
12399 (make_pass_rtl_unroll_and_peel_loops): ...new function.
12400 (pass_rtl_doloop): Convert from a global struct to a subclass of
12401 rtl_opt_pass along with...
12402 (pass_data_rtl_doloop): ...new pass_data instance and...
12403 (make_pass_rtl_doloop): ...new function.
12404 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
12405 a subclass of rtl_opt_pass along with...
12406 (pass_data_lower_subreg): ...new pass_data instance and...
12407 (make_pass_lower_subreg): ...new function.
12408 (pass_lower_subreg2): Convert from a global struct to a subclass of
12409 rtl_opt_pass along with...
12410 (pass_data_lower_subreg2): ...new pass_data instance and...
12411 (make_pass_lower_subreg2): ...new function.
12412 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
12413 struct to a subclass of ipa_opt_pass_d along with...
12414 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
12415 (make_pass_ipa_lto_gimple_out): ...new function.
12416 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
12417 of ipa_opt_pass_d along with...
12418 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
12419 (make_pass_ipa_lto_finish_out): ...new function.
12420 * mode-switching.c (pass_mode_switching): Convert from a global struct
12421 to a subclass of rtl_opt_pass along with...
12422 (pass_data_mode_switching): ...new pass_data instance and...
12423 (make_pass_mode_switching): ...new function.
12424 * modulo-sched.c (pass_sms): Convert from a global struct to a
12425 subclass of rtl_opt_pass along with...
12426 (pass_data_sms): ...new pass_data instance and...
12427 (make_pass_sms): ...new function.
12428 * omp-low.c (pass_expand_omp): Convert from a global struct to a
12429 subclass of gimple_opt_pass along with...
12430 (pass_data_expand_omp): ...new pass_data instance and...
12431 (make_pass_expand_omp): ...new function.
12432 (pass_lower_omp): Convert from a global struct to a subclass of
12433 gimple_opt_pass along with...
12434 (pass_data_lower_omp): ...new pass_data instance and...
12435 (make_pass_lower_omp): ...new function.
12436 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
12437 of gimple_opt_pass along with...
12438 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
12439 (make_pass_diagnose_omp_blocks): ...new function.
12440 * passes.c (pass_early_local_passes): Convert from a global struct to
12441 a subclass of simple_ipa_opt_pass along with...
12442 (pass_data_early_local_passes): ...new pass_data instance and...
12443 (make_pass_early_local_passes): ...new function.
12444 (pass_all_early_optimizations): Convert from a global struct to a
12445 subclass of gimple_opt_pass along with...
12446 (pass_data_all_early_optimizations): ...new pass_data instance and...
12447 (make_pass_all_early_optimizations): ...new function.
12448 (pass_all_optimizations): Convert from a global struct to a subclass
12449 of gimple_opt_pass along with...
12450 (pass_data_all_optimizations): ...new pass_data instance and...
12451 (make_pass_all_optimizations): ...new function.
12452 (pass_all_optimizations_g): Convert from a global struct to a subclass
12453 of gimple_opt_pass along with...
12454 (pass_data_all_optimizations_g): ...new pass_data instance and...
12455 (make_pass_all_optimizations_g): ...new function.
12456 (pass_rest_of_compilation): Convert from a global struct to a subclass
12457 of rtl_opt_pass along with...
12458 (pass_data_rest_of_compilation): ...new pass_data instance and...
12459 (make_pass_rest_of_compilation): ...new function.
12460 (pass_postreload): Convert from a global struct to a subclass of
12461 rtl_opt_pass along with...
12462 (pass_data_postreload): ...new pass_data instance and...
12463 (make_pass_postreload): ...new function.
12464 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
12465 subclass of rtl_opt_pass along with...
12466 (pass_data_gcse2): ...new pass_data instance and...
12467 (make_pass_gcse2): ...new function.
12468 * postreload.c (pass_postreload_cse): Convert from a global struct to
12469 a subclass of rtl_opt_pass along with...
12470 (pass_data_postreload_cse): ...new pass_data instance and...
12471 (make_pass_postreload_cse): ...new function.
12472 * predict.c (pass_profile): Convert from a global struct to a subclass
12473 of gimple_opt_pass along with...
12474 (pass_data_profile): ...new pass_data instance and...
12475 (make_pass_profile): ...new function.
12476 (pass_strip_predict_hints): Convert from a global struct to a subclass
12477 of gimple_opt_pass along with...
12478 (pass_data_strip_predict_hints): ...new pass_data instance and...
12479 (make_pass_strip_predict_hints): ...new function.
12480 * recog.c (pass_peephole2): Convert from a global struct to a subclass
12481 of rtl_opt_pass along with...
12482 (pass_data_peephole2): ...new pass_data instance and...
12483 (make_pass_peephole2): ...new function.
12484 (pass_split_all_insns): Convert from a global struct to a subclass of
12485 rtl_opt_pass along with...
12486 (pass_data_split_all_insns): ...new pass_data instance and...
12487 (make_pass_split_all_insns): ...new function.
12488 (pass_split_after_reload): Convert from a global struct to a subclass
12489 of rtl_opt_pass along with...
12490 (pass_data_split_after_reload): ...new pass_data instance and...
12491 (make_pass_split_after_reload): ...new function.
12492 (pass_split_before_regstack): Convert from a global struct to a
12493 subclass of rtl_opt_pass along with...
12494 (pass_data_split_before_regstack): ...new pass_data instance and...
12495 (make_pass_split_before_regstack): ...new function.
12496 (pass_split_before_sched2): Convert from a global struct to a subclass
12497 of rtl_opt_pass along with...
12498 (pass_data_split_before_sched2): ...new pass_data instance and...
12499 (make_pass_split_before_sched2): ...new function.
12500 (pass_split_for_shorten_branches): Convert from a global struct to a
12501 subclass of rtl_opt_pass along with...
12502 (pass_data_split_for_shorten_branches): ...new pass_data instance
12504 (make_pass_split_for_shorten_branches): ...new function.
12505 * ree.c (pass_ree): Convert from a global struct to a subclass of
12506 rtl_opt_pass along with...
12507 (pass_data_ree): ...new pass_data instance and...
12508 (make_pass_ree): ...new function.
12509 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
12510 subclass of rtl_opt_pass along with...
12511 (pass_data_stack_regs): ...new pass_data instance and...
12512 (make_pass_stack_regs): ...new function.
12513 (pass_stack_regs_run): Convert from a global struct to a subclass of
12514 rtl_opt_pass along with...
12515 (pass_data_stack_regs_run): ...new pass_data instance and...
12516 (make_pass_stack_regs_run): ...new function.
12517 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
12518 subclass of rtl_opt_pass along with...
12519 (pass_data_cprop_hardreg): ...new pass_data instance and...
12520 (make_pass_cprop_hardreg): ...new function.
12521 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
12522 subclass of rtl_opt_pass along with...
12523 (pass_data_reginfo_init): ...new pass_data instance and...
12524 (make_pass_reginfo_init): ...new function.
12525 * regmove.c (pass_regmove): Convert from a global struct to a subclass
12526 of rtl_opt_pass along with...
12527 (pass_data_regmove): ...new pass_data instance and...
12528 (make_pass_regmove): ...new function.
12529 * regrename.c (pass_regrename): Convert from a global struct to a
12530 subclass of rtl_opt_pass along with...
12531 (pass_data_regrename): ...new pass_data instance and...
12532 (make_pass_regrename): ...new function.
12533 * reorg.c (pass_delay_slots): Convert from a global struct to a
12534 subclass of rtl_opt_pass along with...
12535 (pass_data_delay_slots): ...new pass_data instance and...
12536 (make_pass_delay_slots): ...new function.
12537 (pass_machine_reorg): Convert from a global struct to a subclass of
12538 rtl_opt_pass along with...
12539 (pass_data_machine_reorg): ...new pass_data instance and...
12540 (make_pass_machine_reorg): ...new function.
12541 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
12542 of rtl_opt_pass along with...
12543 (pass_data_sched): ...new pass_data instance and...
12544 (make_pass_sched): ...new function.
12545 (pass_sched2): Convert from a global struct to a subclass of
12546 rtl_opt_pass along with...
12547 (pass_data_sched2): ...new pass_data instance and...
12548 (make_pass_sched2): ...new function.
12549 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
12550 to a subclass of rtl_opt_pass along with...
12551 (pass_data_stack_ptr_mod): ...new pass_data instance and...
12552 (make_pass_stack_ptr_mod): ...new function.
12553 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
12554 to a subclass of rtl_opt_pass along with...
12555 (pass_data_rtl_store_motion): ...new pass_data instance and...
12556 (make_pass_rtl_store_motion): ...new function.
12557 * tracer.c (pass_tracer): Convert from a global struct to a subclass
12558 of gimple_opt_pass along with...
12559 (pass_data_tracer): ...new pass_data instance and...
12560 (make_pass_tracer): ...new function.
12561 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
12562 to a subclass of gimple_opt_pass along with...
12563 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
12564 (make_pass_diagnose_tm_blocks): ...new function.
12565 (pass_lower_tm): Convert from a global struct to a subclass of
12566 gimple_opt_pass along with...
12567 (pass_data_lower_tm): ...new pass_data instance and...
12568 (make_pass_lower_tm): ...new function.
12569 (pass_tm_init): Convert from a global struct to a subclass of
12570 gimple_opt_pass along with...
12571 (pass_data_tm_init): ...new pass_data instance and...
12572 (make_pass_tm_init): ...new function.
12573 (pass_tm_mark): Convert from a global struct to a subclass of
12574 gimple_opt_pass along with...
12575 (pass_data_tm_mark): ...new pass_data instance and...
12576 (make_pass_tm_mark): ...new function.
12577 (pass_tm_edges): Convert from a global struct to a subclass of
12578 gimple_opt_pass along with...
12579 (pass_data_tm_edges): ...new pass_data instance and...
12580 (make_pass_tm_edges): ...new function.
12581 (pass_tm_memopt): Convert from a global struct to a subclass of
12582 gimple_opt_pass along with...
12583 (pass_data_tm_memopt): ...new pass_data instance and...
12584 (make_pass_tm_memopt): ...new function.
12585 (pass_ipa_tm): Convert from a global struct to a subclass of
12586 simple_ipa_opt_pass along with...
12587 (pass_data_ipa_tm): ...new pass_data instance and...
12588 (make_pass_ipa_tm): ...new function.
12589 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
12590 subclass of gimple_opt_pass along with...
12591 (pass_data_call_cdce): ...new pass_data instance and...
12592 (make_pass_call_cdce): ...new function.
12593 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
12594 subclass of gimple_opt_pass along with...
12595 (pass_data_build_cfg): ...new pass_data instance and...
12596 (make_pass_build_cfg): ...new function.
12597 (pass_split_crit_edges): Convert from a global struct to a subclass of
12598 gimple_opt_pass along with...
12599 (pass_data_split_crit_edges): ...new pass_data instance and...
12600 (make_pass_split_crit_edges): ...new function.
12601 (pass_warn_function_return): Convert from a global struct to a
12602 subclass of gimple_opt_pass along with...
12603 (pass_data_warn_function_return): ...new pass_data instance and...
12604 (make_pass_warn_function_return): ...new function.
12605 (pass_warn_function_noreturn): Convert from a global struct to a
12606 subclass of gimple_opt_pass along with...
12607 (pass_data_warn_function_noreturn): ...new pass_data instance and...
12608 (make_pass_warn_function_noreturn): ...new function.
12609 (pass_warn_unused_result): Convert from a global struct to a subclass
12610 of gimple_opt_pass along with...
12611 (pass_data_warn_unused_result): ...new pass_data instance and...
12612 (make_pass_warn_unused_result): ...new function.
12613 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
12614 a subclass of gimple_opt_pass along with...
12615 (pass_data_merge_phi): ...new pass_data instance and...
12616 (make_pass_merge_phi): ...new function.
12617 * tree-complex.c (pass_lower_complex): Convert from a global struct to
12618 a subclass of gimple_opt_pass along with...
12619 (pass_data_lower_complex): ...new pass_data instance and...
12620 (make_pass_lower_complex): ...new function.
12621 (pass_lower_complex_O0): Convert from a global struct to a subclass of
12622 gimple_opt_pass along with...
12623 (pass_data_lower_complex_O0): ...new pass_data instance and...
12624 (make_pass_lower_complex_O0): ...new function.
12625 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
12626 subclass of gimple_opt_pass along with...
12627 (pass_data_lower_eh): ...new pass_data instance and...
12628 (make_pass_lower_eh): ...new function.
12629 (pass_refactor_eh): Convert from a global struct to a subclass of
12630 gimple_opt_pass along with...
12631 (pass_data_refactor_eh): ...new pass_data instance and...
12632 (make_pass_refactor_eh): ...new function.
12633 (pass_lower_resx): Convert from a global struct to a subclass of
12634 gimple_opt_pass along with...
12635 (pass_data_lower_resx): ...new pass_data instance and...
12636 (make_pass_lower_resx): ...new function.
12637 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
12638 of gimple_opt_pass along with...
12639 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
12640 (make_pass_lower_eh_dispatch): ...new function.
12641 (pass_cleanup_eh): Convert from a global struct to a subclass of
12642 gimple_opt_pass along with...
12643 (pass_data_cleanup_eh): ...new pass_data instance and...
12644 (make_pass_cleanup_eh): ...new function.
12645 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
12646 to a subclass of simple_ipa_opt_pass along with...
12647 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
12648 (make_pass_ipa_lower_emutls): ...new function.
12649 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
12650 a subclass of gimple_opt_pass along with...
12651 (pass_data_if_conversion): ...new pass_data instance and...
12652 (make_pass_if_conversion): ...new function.
12653 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
12654 subclass of gimple_opt_pass along with...
12655 (pass_data_build_ssa): ...new pass_data instance and...
12656 (make_pass_build_ssa): ...new function.
12657 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
12658 global struct to a subclass of gimple_opt_pass along with...
12659 (pass_data_loop_distribution): ...new pass_data instance and...
12660 (make_pass_loop_distribution): ...new function.
12661 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
12662 subclass of gimple_opt_pass along with...
12663 (pass_data_mudflap_1): ...new pass_data instance and...
12664 (make_pass_mudflap_1): ...new function.
12665 (pass_mudflap_2): Convert from a global struct to a subclass of
12666 gimple_opt_pass along with...
12667 (pass_data_mudflap_2): ...new pass_data instance and...
12668 (make_pass_mudflap_2): ...new function.
12669 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
12670 subclass of gimple_opt_pass along with...
12671 (pass_data_mudflap_1): ...new pass_data instance and...
12672 (make_pass_mudflap_1): ...new function.
12673 (pass_mudflap_2): Convert from a global struct to a subclass of
12674 gimple_opt_pass along with...
12675 (pass_data_mudflap_2): ...new pass_data instance and...
12676 (make_pass_mudflap_2): ...new function.
12677 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
12678 gimple_opt_pass along with...
12679 (pass_data_nrv): ...new pass_data instance and...
12680 (make_pass_nrv): ...new function.
12681 (pass_return_slot): Convert from a global struct to a subclass of
12682 gimple_opt_pass along with...
12683 (pass_data_return_slot): ...new pass_data instance and...
12684 (make_pass_return_slot): ...new function.
12685 * tree-object-size.c (pass_object_sizes): Convert from a global struct
12686 to a subclass of gimple_opt_pass along with...
12687 (pass_data_object_sizes): ...new pass_data instance and...
12688 (make_pass_object_sizes): ...new function.
12689 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
12690 global struct to a subclass of gimple_opt_pass along with...
12691 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
12693 (make_pass_cleanup_cfg_post_optimizing): ...new function.
12694 (pass_fixup_cfg): Convert from a global struct to a subclass of
12695 gimple_opt_pass along with...
12696 (pass_data_fixup_cfg): ...new pass_data instance and...
12697 (make_pass_fixup_cfg): ...new function.
12698 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
12699 (make_pass_mudflap_1): ...new function.
12700 (pass_mudflap_2): Replace declaration with that of...
12701 (make_pass_mudflap_2): ...new function.
12702 (pass_asan): Replace declaration with that of...
12703 (make_pass_asan): ...new function.
12704 (pass_asan_O0): Replace declaration with that of...
12705 (make_pass_asan_O0): ...new function.
12706 (pass_tsan): Replace declaration with that of...
12707 (make_pass_tsan): ...new function.
12708 (pass_tsan_O0): Replace declaration with that of...
12709 (make_pass_tsan_O0): ...new function.
12710 (pass_lower_cf): Replace declaration with that of...
12711 (make_pass_lower_cf): ...new function.
12712 (pass_refactor_eh): Replace declaration with that of...
12713 (make_pass_refactor_eh): ...new function.
12714 (pass_lower_eh): Replace declaration with that of...
12715 (make_pass_lower_eh): ...new function.
12716 (pass_lower_eh_dispatch): Replace declaration with that of...
12717 (make_pass_lower_eh_dispatch): ...new function.
12718 (pass_lower_resx): Replace declaration with that of...
12719 (make_pass_lower_resx): ...new function.
12720 (pass_build_cfg): Replace declaration with that of...
12721 (make_pass_build_cfg): ...new function.
12722 (pass_early_tree_profile): Replace declaration with that of...
12723 (make_pass_early_tree_profile): ...new function.
12724 (pass_cleanup_eh): Replace declaration with that of...
12725 (make_pass_cleanup_eh): ...new function.
12726 (pass_sra): Replace declaration with that of...
12727 (make_pass_sra): ...new function.
12728 (pass_sra_early): Replace declaration with that of...
12729 (make_pass_sra_early): ...new function.
12730 (pass_early_ipa_sra): Replace declaration with that of...
12731 (make_pass_early_ipa_sra): ...new function.
12732 (pass_tail_recursion): Replace declaration with that of...
12733 (make_pass_tail_recursion): ...new function.
12734 (pass_tail_calls): Replace declaration with that of...
12735 (make_pass_tail_calls): ...new function.
12736 (pass_tree_loop): Replace declaration with that of...
12737 (make_pass_tree_loop): ...new function.
12738 (pass_tree_loop_init): Replace declaration with that of...
12739 (make_pass_tree_loop_init): ...new function.
12740 (pass_lim): Replace declaration with that of...
12741 (make_pass_lim): ...new function.
12742 (pass_tree_unswitch): Replace declaration with that of...
12743 (make_pass_tree_unswitch): ...new function.
12744 (pass_predcom): Replace declaration with that of...
12745 (make_pass_predcom): ...new function.
12746 (pass_iv_canon): Replace declaration with that of...
12747 (make_pass_iv_canon): ...new function.
12748 (pass_scev_cprop): Replace declaration with that of...
12749 (make_pass_scev_cprop): ...new function.
12750 (pass_empty_loop): Replace declaration with that of...
12751 (make_pass_empty_loop): ...new function.
12752 (pass_record_bounds): Replace declaration with that of...
12753 (make_pass_record_bounds): ...new function.
12754 (pass_graphite): Replace declaration with that of...
12755 (make_pass_graphite): ...new function.
12756 (pass_graphite_transforms): Replace declaration with that of...
12757 (make_pass_graphite_transforms): ...new function.
12758 (pass_if_conversion): Replace declaration with that of...
12759 (make_pass_if_conversion): ...new function.
12760 (pass_loop_distribution): Replace declaration with that of...
12761 (make_pass_loop_distribution): ...new function.
12762 (pass_vectorize): Replace declaration with that of...
12763 (make_pass_vectorize): ...new function.
12764 (pass_slp_vectorize): Replace declaration with that of...
12765 (make_pass_slp_vectorize): ...new function.
12766 (pass_complete_unroll): Replace declaration with that of...
12767 (make_pass_complete_unroll): ...new function.
12768 (pass_complete_unrolli): Replace declaration with that of...
12769 (make_pass_complete_unrolli): ...new function.
12770 (pass_parallelize_loops): Replace declaration with that of...
12771 (make_pass_parallelize_loops): ...new function.
12772 (pass_loop_prefetch): Replace declaration with that of...
12773 (make_pass_loop_prefetch): ...new function.
12774 (pass_iv_optimize): Replace declaration with that of...
12775 (make_pass_iv_optimize): ...new function.
12776 (pass_tree_loop_done): Replace declaration with that of...
12777 (make_pass_tree_loop_done): ...new function.
12778 (pass_ch): Replace declaration with that of...
12779 (make_pass_ch): ...new function.
12780 (pass_ccp): Replace declaration with that of...
12781 (make_pass_ccp): ...new function.
12782 (pass_phi_only_cprop): Replace declaration with that of...
12783 (make_pass_phi_only_cprop): ...new function.
12784 (pass_build_ssa): Replace declaration with that of...
12785 (make_pass_build_ssa): ...new function.
12786 (pass_build_alias): Replace declaration with that of...
12787 (make_pass_build_alias): ...new function.
12788 (pass_build_ealias): Replace declaration with that of...
12789 (make_pass_build_ealias): ...new function.
12790 (pass_dominator): Replace declaration with that of...
12791 (make_pass_dominator): ...new function.
12792 (pass_dce): Replace declaration with that of...
12793 (make_pass_dce): ...new function.
12794 (pass_dce_loop): Replace declaration with that of...
12795 (make_pass_dce_loop): ...new function.
12796 (pass_cd_dce): Replace declaration with that of...
12797 (make_pass_cd_dce): ...new function.
12798 (pass_call_cdce): Replace declaration with that of...
12799 (make_pass_call_cdce): ...new function.
12800 (pass_merge_phi): Replace declaration with that of...
12801 (make_pass_merge_phi): ...new function.
12802 (pass_split_crit_edges): Replace declaration with that of...
12803 (make_pass_split_crit_edges): ...new function.
12804 (pass_pre): Replace declaration with that of...
12805 (make_pass_pre): ...new function.
12806 (pass_profile): Replace declaration with that of...
12807 (make_pass_profile): ...new function.
12808 (pass_strip_predict_hints): Replace declaration with that of...
12809 (make_pass_strip_predict_hints): ...new function.
12810 (pass_lower_complex_O0): Replace declaration with that of...
12811 (make_pass_lower_complex_O0): ...new function.
12812 (pass_lower_complex): Replace declaration with that of...
12813 (make_pass_lower_complex): ...new function.
12814 (pass_lower_vector): Replace declaration with that of...
12815 (make_pass_lower_vector): ...new function.
12816 (pass_lower_vector_ssa): Replace declaration with that of...
12817 (make_pass_lower_vector_ssa): ...new function.
12818 (pass_lower_omp): Replace declaration with that of...
12819 (make_pass_lower_omp): ...new function.
12820 (pass_diagnose_omp_blocks): Replace declaration with that of...
12821 (make_pass_diagnose_omp_blocks): ...new function.
12822 (pass_expand_omp): Replace declaration with that of...
12823 (make_pass_expand_omp): ...new function.
12824 (pass_expand_omp_ssa): Replace declaration with that of...
12825 (make_pass_expand_omp_ssa): ...new function.
12826 (pass_object_sizes): Replace declaration with that of...
12827 (make_pass_object_sizes): ...new function.
12828 (pass_strlen): Replace declaration with that of...
12829 (make_pass_strlen): ...new function.
12830 (pass_fold_builtins): Replace declaration with that of...
12831 (make_pass_fold_builtins): ...new function.
12832 (pass_stdarg): Replace declaration with that of...
12833 (make_pass_stdarg): ...new function.
12834 (pass_early_warn_uninitialized): Replace declaration with that of...
12835 (make_pass_early_warn_uninitialized): ...new function.
12836 (pass_late_warn_uninitialized): Replace declaration with that of...
12837 (make_pass_late_warn_uninitialized): ...new function.
12838 (pass_cse_reciprocals): Replace declaration with that of...
12839 (make_pass_cse_reciprocals): ...new function.
12840 (pass_cse_sincos): Replace declaration with that of...
12841 (make_pass_cse_sincos): ...new function.
12842 (pass_optimize_bswap): Replace declaration with that of...
12843 (make_pass_optimize_bswap): ...new function.
12844 (pass_optimize_widening_mul): Replace declaration with that of...
12845 (make_pass_optimize_widening_mul): ...new function.
12846 (pass_warn_function_return): Replace declaration with that of...
12847 (make_pass_warn_function_return): ...new function.
12848 (pass_warn_function_noreturn): Replace declaration with that of...
12849 (make_pass_warn_function_noreturn): ...new function.
12850 (pass_cselim): Replace declaration with that of...
12851 (make_pass_cselim): ...new function.
12852 (pass_phiopt): Replace declaration with that of...
12853 (make_pass_phiopt): ...new function.
12854 (pass_forwprop): Replace declaration with that of...
12855 (make_pass_forwprop): ...new function.
12856 (pass_phiprop): Replace declaration with that of...
12857 (make_pass_phiprop): ...new function.
12858 (pass_tree_ifcombine): Replace declaration with that of...
12859 (make_pass_tree_ifcombine): ...new function.
12860 (pass_dse): Replace declaration with that of...
12861 (make_pass_dse): ...new function.
12862 (pass_nrv): Replace declaration with that of...
12863 (make_pass_nrv): ...new function.
12864 (pass_rename_ssa_copies): Replace declaration with that of...
12865 (make_pass_rename_ssa_copies): ...new function.
12866 (pass_sink_code): Replace declaration with that of...
12867 (make_pass_sink_code): ...new function.
12868 (pass_fre): Replace declaration with that of...
12869 (make_pass_fre): ...new function.
12870 (pass_check_data_deps): Replace declaration with that of...
12871 (make_pass_check_data_deps): ...new function.
12872 (pass_copy_prop): Replace declaration with that of...
12873 (make_pass_copy_prop): ...new function.
12874 (pass_vrp): Replace declaration with that of...
12875 (make_pass_vrp): ...new function.
12876 (pass_uncprop): Replace declaration with that of...
12877 (make_pass_uncprop): ...new function.
12878 (pass_return_slot): Replace declaration with that of...
12879 (make_pass_return_slot): ...new function.
12880 (pass_reassoc): Replace declaration with that of...
12881 (make_pass_reassoc): ...new function.
12882 (pass_rebuild_cgraph_edges): Replace declaration with that of...
12883 (make_pass_rebuild_cgraph_edges): ...new function.
12884 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
12885 (make_pass_remove_cgraph_callee_edges): ...new function.
12886 (pass_build_cgraph_edges): Replace declaration with that of...
12887 (make_pass_build_cgraph_edges): ...new function.
12888 (pass_local_pure_const): Replace declaration with that of...
12889 (make_pass_local_pure_const): ...new function.
12890 (pass_tracer): Replace declaration with that of...
12891 (make_pass_tracer): ...new function.
12892 (pass_warn_unused_result): Replace declaration with that of...
12893 (make_pass_warn_unused_result): ...new function.
12894 (pass_diagnose_tm_blocks): Replace declaration with that of...
12895 (make_pass_diagnose_tm_blocks): ...new function.
12896 (pass_lower_tm): Replace declaration with that of...
12897 (make_pass_lower_tm): ...new function.
12898 (pass_tm_init): Replace declaration with that of...
12899 (make_pass_tm_init): ...new function.
12900 (pass_tm_mark): Replace declaration with that of...
12901 (make_pass_tm_mark): ...new function.
12902 (pass_tm_memopt): Replace declaration with that of...
12903 (make_pass_tm_memopt): ...new function.
12904 (pass_tm_edges): Replace declaration with that of...
12905 (make_pass_tm_edges): ...new function.
12906 (pass_split_functions): Replace declaration with that of...
12907 (make_pass_split_functions): ...new function.
12908 (pass_feedback_split_functions): Replace declaration with that of...
12909 (make_pass_feedback_split_functions): ...new function.
12910 (pass_strength_reduction): Replace declaration with that of...
12911 (make_pass_strength_reduction): ...new function.
12912 (pass_ipa_lower_emutls): Replace declaration with that of...
12913 (make_pass_ipa_lower_emutls): ...new function.
12914 (pass_ipa_function_and_variable_visibility): Replace declaration with
12916 (make_pass_ipa_function_and_variable_visibility): ...new function.
12917 (pass_ipa_tree_profile): Replace declaration with that of...
12918 (make_pass_ipa_tree_profile): ...new function.
12919 (pass_early_local_passes): Replace declaration with that of...
12920 (make_pass_early_local_passes): ...new function.
12921 (pass_ipa_whole_program_visibility): Replace declaration with that
12923 (make_pass_ipa_whole_program_visibility): ...new function.
12924 (pass_ipa_lto_gimple_out): Replace declaration with that of...
12925 (make_pass_ipa_lto_gimple_out): ...new function.
12926 (pass_ipa_increase_alignment): Replace declaration with that of...
12927 (make_pass_ipa_increase_alignment): ...new function.
12928 (pass_ipa_inline): Replace declaration with that of...
12929 (make_pass_ipa_inline): ...new function.
12930 (pass_ipa_free_lang_data): Replace declaration with that of...
12931 (make_pass_ipa_free_lang_data): ...new function.
12932 (pass_ipa_free_inline_summary): Replace declaration with that of...
12933 (make_pass_ipa_free_inline_summary): ...new function.
12934 (pass_ipa_cp): Replace declaration with that of...
12935 (make_pass_ipa_cp): ...new function.
12936 (pass_ipa_reference): Replace declaration with that of...
12937 (make_pass_ipa_reference): ...new function.
12938 (pass_ipa_pure_const): Replace declaration with that of...
12939 (make_pass_ipa_pure_const): ...new function.
12940 (pass_ipa_pta): Replace declaration with that of...
12941 (make_pass_ipa_pta): ...new function.
12942 (pass_ipa_lto_finish_out): Replace declaration with that of...
12943 (make_pass_ipa_lto_finish_out): ...new function.
12944 (pass_ipa_tm): Replace declaration with that of...
12945 (make_pass_ipa_tm): ...new function.
12946 (pass_ipa_profile): Replace declaration with that of...
12947 (make_pass_ipa_profile): ...new function.
12948 (pass_ipa_cdtor_merge): Replace declaration with that of...
12949 (make_pass_ipa_cdtor_merge): ...new function.
12950 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
12952 (make_pass_cleanup_cfg_post_optimizing): ...new function.
12953 (pass_init_datastructures): Replace declaration with that of...
12954 (make_pass_init_datastructures): ...new function.
12955 (pass_fixup_cfg): Replace declaration with that of...
12956 (make_pass_fixup_cfg): ...new function.
12957 (pass_expand): Replace declaration with that of...
12958 (make_pass_expand): ...new function.
12959 (pass_instantiate_virtual_regs): Replace declaration with that of...
12960 (make_pass_instantiate_virtual_regs): ...new function.
12961 (pass_rtl_fwprop): Replace declaration with that of...
12962 (make_pass_rtl_fwprop): ...new function.
12963 (pass_rtl_fwprop_addr): Replace declaration with that of...
12964 (make_pass_rtl_fwprop_addr): ...new function.
12965 (pass_jump): Replace declaration with that of...
12966 (make_pass_jump): ...new function.
12967 (pass_jump2): Replace declaration with that of...
12968 (make_pass_jump2): ...new function.
12969 (pass_lower_subreg): Replace declaration with that of...
12970 (make_pass_lower_subreg): ...new function.
12971 (pass_cse): Replace declaration with that of...
12972 (make_pass_cse): ...new function.
12973 (pass_fast_rtl_dce): Replace declaration with that of...
12974 (make_pass_fast_rtl_dce): ...new function.
12975 (pass_ud_rtl_dce): Replace declaration with that of...
12976 (make_pass_ud_rtl_dce): ...new function.
12977 (pass_rtl_dce): Replace declaration with that of...
12978 (make_pass_rtl_dce): ...new function.
12979 (pass_rtl_dse1): Replace declaration with that of...
12980 (make_pass_rtl_dse1): ...new function.
12981 (pass_rtl_dse2): Replace declaration with that of...
12982 (make_pass_rtl_dse2): ...new function.
12983 (pass_rtl_dse3): Replace declaration with that of...
12984 (make_pass_rtl_dse3): ...new function.
12985 (pass_rtl_cprop): Replace declaration with that of...
12986 (make_pass_rtl_cprop): ...new function.
12987 (pass_rtl_pre): Replace declaration with that of...
12988 (make_pass_rtl_pre): ...new function.
12989 (pass_rtl_hoist): Replace declaration with that of...
12990 (make_pass_rtl_hoist): ...new function.
12991 (pass_rtl_store_motion): Replace declaration with that of...
12992 (make_pass_rtl_store_motion): ...new function.
12993 (pass_cse_after_global_opts): Replace declaration with that of...
12994 (make_pass_cse_after_global_opts): ...new function.
12995 (pass_rtl_ifcvt): Replace declaration with that of...
12996 (make_pass_rtl_ifcvt): ...new function.
12997 (pass_into_cfg_layout_mode): Replace declaration with that of...
12998 (make_pass_into_cfg_layout_mode): ...new function.
12999 (pass_outof_cfg_layout_mode): Replace declaration with that of...
13000 (make_pass_outof_cfg_layout_mode): ...new function.
13001 (pass_loop2): Replace declaration with that of...
13002 (make_pass_loop2): ...new function.
13003 (pass_rtl_loop_init): Replace declaration with that of...
13004 (make_pass_rtl_loop_init): ...new function.
13005 (pass_rtl_move_loop_invariants): Replace declaration with that of...
13006 (make_pass_rtl_move_loop_invariants): ...new function.
13007 (pass_rtl_unswitch): Replace declaration with that of...
13008 (make_pass_rtl_unswitch): ...new function.
13009 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
13010 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13011 (pass_rtl_doloop): Replace declaration with that of...
13012 (make_pass_rtl_doloop): ...new function.
13013 (pass_rtl_loop_done): Replace declaration with that of...
13014 (make_pass_rtl_loop_done): ...new function.
13015 (pass_web): Replace declaration with that of...
13016 (make_pass_web): ...new function.
13017 (pass_cse2): Replace declaration with that of...
13018 (make_pass_cse2): ...new function.
13019 (pass_df_initialize_opt): Replace declaration with that of...
13020 (make_pass_df_initialize_opt): ...new function.
13021 (pass_df_initialize_no_opt): Replace declaration with that of...
13022 (make_pass_df_initialize_no_opt): ...new function.
13023 (pass_reginfo_init): Replace declaration with that of...
13024 (make_pass_reginfo_init): ...new function.
13025 (pass_inc_dec): Replace declaration with that of...
13026 (make_pass_inc_dec): ...new function.
13027 (pass_stack_ptr_mod): Replace declaration with that of...
13028 (make_pass_stack_ptr_mod): ...new function.
13029 (pass_initialize_regs): Replace declaration with that of...
13030 (make_pass_initialize_regs): ...new function.
13031 (pass_combine): Replace declaration with that of...
13032 (make_pass_combine): ...new function.
13033 (pass_if_after_combine): Replace declaration with that of...
13034 (make_pass_if_after_combine): ...new function.
13035 (pass_ree): Replace declaration with that of...
13036 (make_pass_ree): ...new function.
13037 (pass_partition_blocks): Replace declaration with that of...
13038 (make_pass_partition_blocks): ...new function.
13039 (pass_match_asm_constraints): Replace declaration with that of...
13040 (make_pass_match_asm_constraints): ...new function.
13041 (pass_regmove): Replace declaration with that of...
13042 (make_pass_regmove): ...new function.
13043 (pass_split_all_insns): Replace declaration with that of...
13044 (make_pass_split_all_insns): ...new function.
13045 (pass_fast_rtl_byte_dce): Replace declaration with that of...
13046 (make_pass_fast_rtl_byte_dce): ...new function.
13047 (pass_lower_subreg2): Replace declaration with that of...
13048 (make_pass_lower_subreg2): ...new function.
13049 (pass_mode_switching): Replace declaration with that of...
13050 (make_pass_mode_switching): ...new function.
13051 (pass_sms): Replace declaration with that of...
13052 (make_pass_sms): ...new function.
13053 (pass_sched): Replace declaration with that of...
13054 (make_pass_sched): ...new function.
13055 (pass_ira): Replace declaration with that of...
13056 (make_pass_ira): ...new function.
13057 (pass_reload): Replace declaration with that of...
13058 (make_pass_reload): ...new function.
13059 (pass_clean_state): Replace declaration with that of...
13060 (make_pass_clean_state): ...new function.
13061 (pass_branch_prob): Replace declaration with that of...
13062 (make_pass_branch_prob): ...new function.
13063 (pass_value_profile_transformations): Replace declaration with that
13065 (make_pass_value_profile_transformations): ...new function.
13066 (pass_postreload_cse): Replace declaration with that of...
13067 (make_pass_postreload_cse): ...new function.
13068 (pass_gcse2): Replace declaration with that of...
13069 (make_pass_gcse2): ...new function.
13070 (pass_split_after_reload): Replace declaration with that of...
13071 (make_pass_split_after_reload): ...new function.
13072 (pass_branch_target_load_optimize1): Replace declaration with that
13074 (make_pass_branch_target_load_optimize1): ...new function.
13075 (pass_thread_prologue_and_epilogue): Replace declaration with that
13077 (make_pass_thread_prologue_and_epilogue): ...new function.
13078 (pass_stack_adjustments): Replace declaration with that of...
13079 (make_pass_stack_adjustments): ...new function.
13080 (pass_peephole2): Replace declaration with that of...
13081 (make_pass_peephole2): ...new function.
13082 (pass_if_after_reload): Replace declaration with that of...
13083 (make_pass_if_after_reload): ...new function.
13084 (pass_regrename): Replace declaration with that of...
13085 (make_pass_regrename): ...new function.
13086 (pass_cprop_hardreg): Replace declaration with that of...
13087 (make_pass_cprop_hardreg): ...new function.
13088 (pass_reorder_blocks): Replace declaration with that of...
13089 (make_pass_reorder_blocks): ...new function.
13090 (pass_branch_target_load_optimize2): Replace declaration with that
13092 (make_pass_branch_target_load_optimize2): ...new function.
13093 (pass_leaf_regs): Replace declaration with that of...
13094 (make_pass_leaf_regs): ...new function.
13095 (pass_split_before_sched2): Replace declaration with that of...
13096 (make_pass_split_before_sched2): ...new function.
13097 (pass_compare_elim_after_reload): Replace declaration with that of...
13098 (make_pass_compare_elim_after_reload): ...new function.
13099 (pass_sched2): Replace declaration with that of...
13100 (make_pass_sched2): ...new function.
13101 (pass_stack_regs): Replace declaration with that of...
13102 (make_pass_stack_regs): ...new function.
13103 (pass_stack_regs_run): Replace declaration with that of...
13104 (make_pass_stack_regs_run): ...new function.
13105 (pass_df_finish): Replace declaration with that of...
13106 (make_pass_df_finish): ...new function.
13107 (pass_compute_alignments): Replace declaration with that of...
13108 (make_pass_compute_alignments): ...new function.
13109 (pass_duplicate_computed_gotos): Replace declaration with that of...
13110 (make_pass_duplicate_computed_gotos): ...new function.
13111 (pass_variable_tracking): Replace declaration with that of...
13112 (make_pass_variable_tracking): ...new function.
13113 (pass_free_cfg): Replace declaration with that of...
13114 (make_pass_free_cfg): ...new function.
13115 (pass_machine_reorg): Replace declaration with that of...
13116 (make_pass_machine_reorg): ...new function.
13117 (pass_cleanup_barriers): Replace declaration with that of...
13118 (make_pass_cleanup_barriers): ...new function.
13119 (pass_delay_slots): Replace declaration with that of...
13120 (make_pass_delay_slots): ...new function.
13121 (pass_split_for_shorten_branches): Replace declaration with that of...
13122 (make_pass_split_for_shorten_branches): ...new function.
13123 (pass_split_before_regstack): Replace declaration with that of...
13124 (make_pass_split_before_regstack): ...new function.
13125 (pass_convert_to_eh_region_ranges): Replace declaration with that
13127 (make_pass_convert_to_eh_region_ranges): ...new function.
13128 (pass_shorten_branches): Replace declaration with that of...
13129 (make_pass_shorten_branches): ...new function.
13130 (pass_set_nothrow_function_flags): Replace declaration with that of...
13131 (make_pass_set_nothrow_function_flags): ...new function.
13132 (pass_dwarf2_frame): Replace declaration with that of...
13133 (make_pass_dwarf2_frame): ...new function.
13134 (pass_final): Replace declaration with that of...
13135 (make_pass_final): ...new function.
13136 (pass_rtl_seqabstr): Replace declaration with that of...
13137 (make_pass_rtl_seqabstr): ...new function.
13138 (pass_release_ssa_names): Replace declaration with that of...
13139 (make_pass_release_ssa_names): ...new function.
13140 (pass_early_inline): Replace declaration with that of...
13141 (make_pass_early_inline): ...new function.
13142 (pass_inline_parameters): Replace declaration with that of...
13143 (make_pass_inline_parameters): ...new function.
13144 (pass_update_address_taken): Replace declaration with that of...
13145 (make_pass_update_address_taken): ...new function.
13146 (pass_convert_switch): Replace declaration with that of...
13147 (make_pass_convert_switch): ...new function.
13148 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
13149 to a subclass of simple_ipa_opt_pass along with...
13150 (pass_data_ipa_tree_profile): ...new pass_data instance and...
13151 (make_pass_ipa_tree_profile): ...new function.
13152 * tree-sra.c (pass_sra_early): Convert from a global struct to a
13153 subclass of gimple_opt_pass along with...
13154 (pass_data_sra_early): ...new pass_data instance and...
13155 (make_pass_sra_early): ...new function.
13156 (pass_sra): Convert from a global struct to a subclass of
13157 gimple_opt_pass along with...
13158 (pass_data_sra): ...new pass_data instance and...
13159 (make_pass_sra): ...new function.
13160 (pass_early_ipa_sra): Convert from a global struct to a subclass of
13161 gimple_opt_pass along with...
13162 (pass_data_early_ipa_sra): ...new pass_data instance and...
13163 (make_pass_early_ipa_sra): ...new function.
13164 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
13165 subclass of gimple_opt_pass along with...
13166 (pass_data_ccp): ...new pass_data instance and...
13167 (make_pass_ccp): ...new function.
13168 (pass_fold_builtins): Convert from a global struct to a subclass of
13169 gimple_opt_pass along with...
13170 (pass_data_fold_builtins): ...new pass_data instance and...
13171 (make_pass_fold_builtins): ...new function.
13172 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
13173 subclass of gimple_opt_pass along with...
13174 (pass_data_copy_prop): ...new pass_data instance and...
13175 (make_pass_copy_prop): ...new function.
13176 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
13177 global struct to a subclass of gimple_opt_pass along with...
13178 (pass_data_rename_ssa_copies): ...new pass_data instance and...
13179 (make_pass_rename_ssa_copies): ...new function.
13180 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
13181 subclass of gimple_opt_pass along with...
13182 (pass_data_dce): ...new pass_data instance and...
13183 (make_pass_dce): ...new function.
13184 (pass_dce_loop): Convert from a global struct to a subclass of
13185 gimple_opt_pass along with...
13186 (pass_data_dce_loop): ...new pass_data instance and...
13187 (make_pass_dce_loop): ...new function.
13188 (pass_cd_dce): Convert from a global struct to a subclass of
13189 gimple_opt_pass along with...
13190 (pass_data_cd_dce): ...new pass_data instance and...
13191 (make_pass_cd_dce): ...new function.
13192 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
13193 subclass of gimple_opt_pass along with...
13194 (pass_data_dominator): ...new pass_data instance and...
13195 (make_pass_dominator): ...new function.
13196 (pass_phi_only_cprop): Convert from a global struct to a subclass of
13197 gimple_opt_pass along with...
13198 (pass_data_phi_only_cprop): ...new pass_data instance and...
13199 (make_pass_phi_only_cprop): ...new function.
13200 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
13201 subclass of gimple_opt_pass along with...
13202 (pass_data_dse): ...new pass_data instance and...
13203 (make_pass_dse): ...new function.
13204 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
13205 a subclass of gimple_opt_pass along with...
13206 (pass_data_forwprop): ...new pass_data instance and...
13207 (make_pass_forwprop): ...new function.
13208 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
13209 struct to a subclass of gimple_opt_pass along with...
13210 (pass_data_tree_ifcombine): ...new pass_data instance and...
13211 (make_pass_tree_ifcombine): ...new function.
13212 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
13213 subclass of gimple_opt_pass along with...
13214 (pass_data_ch): ...new pass_data instance and...
13215 (make_pass_ch): ...new function.
13216 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
13217 subclass of gimple_opt_pass along with...
13218 (pass_data_tree_loop): ...new pass_data instance and...
13219 (make_pass_tree_loop): ...new function.
13220 (pass_tree_loop_init): Convert from a global struct to a subclass of
13221 gimple_opt_pass along with...
13222 (pass_data_tree_loop_init): ...new pass_data instance and...
13223 (make_pass_tree_loop_init): ...new function.
13224 (pass_lim): Convert from a global struct to a subclass of
13225 gimple_opt_pass along with...
13226 (pass_data_lim): ...new pass_data instance and...
13227 (make_pass_lim): ...new function.
13228 (pass_tree_unswitch): Convert from a global struct to a subclass of
13229 gimple_opt_pass along with...
13230 (pass_data_tree_unswitch): ...new pass_data instance and...
13231 (make_pass_tree_unswitch): ...new function.
13232 (pass_predcom): Convert from a global struct to a subclass of
13233 gimple_opt_pass along with...
13234 (pass_data_predcom): ...new pass_data instance and...
13235 (make_pass_predcom): ...new function.
13236 (pass_vectorize): Convert from a global struct to a subclass of
13237 gimple_opt_pass along with...
13238 (pass_data_vectorize): ...new pass_data instance and...
13239 (make_pass_vectorize): ...new function.
13240 (pass_graphite): Convert from a global struct to a subclass of
13241 gimple_opt_pass along with...
13242 (pass_data_graphite): ...new pass_data instance and...
13243 (make_pass_graphite): ...new function.
13244 (pass_graphite_transforms): Convert from a global struct to a subclass
13245 of gimple_opt_pass along with...
13246 (pass_data_graphite_transforms): ...new pass_data instance and...
13247 (make_pass_graphite_transforms): ...new function.
13248 (pass_check_data_deps): Convert from a global struct to a subclass of
13249 gimple_opt_pass along with...
13250 (pass_data_check_data_deps): ...new pass_data instance and...
13251 (make_pass_check_data_deps): ...new function.
13252 (pass_iv_canon): Convert from a global struct to a subclass of
13253 gimple_opt_pass along with...
13254 (pass_data_iv_canon): ...new pass_data instance and...
13255 (make_pass_iv_canon): ...new function.
13256 (pass_scev_cprop): Convert from a global struct to a subclass of
13257 gimple_opt_pass along with...
13258 (pass_data_scev_cprop): ...new pass_data instance and...
13259 (make_pass_scev_cprop): ...new function.
13260 (pass_record_bounds): Convert from a global struct to a subclass of
13261 gimple_opt_pass along with...
13262 (pass_data_record_bounds): ...new pass_data instance and...
13263 (make_pass_record_bounds): ...new function.
13264 (pass_complete_unroll): Convert from a global struct to a subclass of
13265 gimple_opt_pass along with...
13266 (pass_data_complete_unroll): ...new pass_data instance and...
13267 (make_pass_complete_unroll): ...new function.
13268 (pass_complete_unrolli): Convert from a global struct to a subclass of
13269 gimple_opt_pass along with...
13270 (pass_data_complete_unrolli): ...new pass_data instance and...
13271 (make_pass_complete_unrolli): ...new function.
13272 (pass_parallelize_loops): Convert from a global struct to a subclass
13273 of gimple_opt_pass along with...
13274 (pass_data_parallelize_loops): ...new pass_data instance and...
13275 (make_pass_parallelize_loops): ...new function.
13276 (pass_loop_prefetch): Convert from a global struct to a subclass of
13277 gimple_opt_pass along with...
13278 (pass_data_loop_prefetch): ...new pass_data instance and...
13279 (make_pass_loop_prefetch): ...new function.
13280 (pass_iv_optimize): Convert from a global struct to a subclass of
13281 gimple_opt_pass along with...
13282 (pass_data_iv_optimize): ...new pass_data instance and...
13283 (make_pass_iv_optimize): ...new function.
13284 (pass_tree_loop_done): Convert from a global struct to a subclass of
13285 gimple_opt_pass along with...
13286 (pass_data_tree_loop_done): ...new pass_data instance and...
13287 (make_pass_tree_loop_done): ...new function.
13288 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
13289 struct to a subclass of gimple_opt_pass along with...
13290 (pass_data_cse_reciprocals): ...new pass_data instance and...
13291 (make_pass_cse_reciprocals): ...new function.
13292 (pass_cse_sincos): Convert from a global struct to a subclass of
13293 gimple_opt_pass along with...
13294 (pass_data_cse_sincos): ...new pass_data instance and...
13295 (make_pass_cse_sincos): ...new function.
13296 (pass_optimize_bswap): Convert from a global struct to a subclass of
13297 gimple_opt_pass along with...
13298 (pass_data_optimize_bswap): ...new pass_data instance and...
13299 (make_pass_optimize_bswap): ...new function.
13300 (pass_optimize_widening_mul): Convert from a global struct to a
13301 subclass of gimple_opt_pass along with...
13302 (pass_data_optimize_widening_mul): ...new pass_data instance and...
13303 (make_pass_optimize_widening_mul): ...new function.
13304 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
13305 subclass of gimple_opt_pass along with...
13306 (pass_data_phiopt): ...new pass_data instance and...
13307 (make_pass_phiopt): ...new function.
13308 (pass_cselim): Convert from a global struct to a subclass of
13309 gimple_opt_pass along with...
13310 (pass_data_cselim): ...new pass_data instance and...
13311 (make_pass_cselim): ...new function.
13312 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
13313 subclass of gimple_opt_pass along with...
13314 (pass_data_phiprop): ...new pass_data instance and...
13315 (make_pass_phiprop): ...new function.
13316 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
13317 subclass of gimple_opt_pass along with...
13318 (pass_data_pre): ...new pass_data instance and...
13319 (make_pass_pre): ...new function.
13320 (pass_fre): Convert from a global struct to a subclass of
13321 gimple_opt_pass along with...
13322 (pass_data_fre): ...new pass_data instance and...
13323 (make_pass_fre): ...new function.
13324 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
13325 subclass of gimple_opt_pass along with...
13326 (pass_data_reassoc): ...new pass_data instance and...
13327 (make_pass_reassoc): ...new function.
13328 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
13329 subclass of gimple_opt_pass along with...
13330 (pass_data_sink_code): ...new pass_data instance and...
13331 (make_pass_sink_code): ...new function.
13332 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
13333 subclass of gimple_opt_pass along with...
13334 (pass_data_strlen): ...new pass_data instance and...
13335 (make_pass_strlen): ...new function.
13336 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
13337 struct to a subclass of gimple_opt_pass along with...
13338 (pass_data_build_alias): ...new pass_data instance and...
13339 (make_pass_build_alias): ...new function.
13340 (pass_build_ealias): Convert from a global struct to a subclass of
13341 gimple_opt_pass along with...
13342 (pass_data_build_ealias): ...new pass_data instance and...
13343 (make_pass_build_ealias): ...new function.
13344 (pass_ipa_pta): Convert from a global struct to a subclass of
13345 simple_ipa_opt_pass along with...
13346 (pass_data_ipa_pta): ...new pass_data instance and...
13347 (make_pass_ipa_pta): ...new function.
13348 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
13349 subclass of gimple_opt_pass along with...
13350 (pass_data_uncprop): ...new pass_data instance and...
13351 (make_pass_uncprop): ...new function.
13352 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
13353 global struct to a subclass of gimple_opt_pass along with...
13354 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
13355 (make_pass_late_warn_uninitialized): ...new function.
13356 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
13357 to a subclass of gimple_opt_pass along with...
13358 (pass_data_init_datastructures): ...new pass_data instance and...
13359 (make_pass_init_datastructures): ...new function.
13360 (pass_early_warn_uninitialized): Convert from a global struct to a
13361 subclass of gimple_opt_pass along with...
13362 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
13363 (make_pass_early_warn_uninitialized): ...new function.
13364 (pass_update_address_taken): Convert from a global struct to a
13365 subclass of gimple_opt_pass along with...
13366 (pass_data_update_address_taken): ...new pass_data instance and...
13367 (make_pass_update_address_taken): ...new function.
13368 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
13369 struct to a subclass of gimple_opt_pass along with...
13370 (pass_data_release_ssa_names): ...new pass_data instance and...
13371 (make_pass_release_ssa_names): ...new function.
13372 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
13373 subclass of gimple_opt_pass along with...
13374 (pass_data_stdarg): ...new pass_data instance and...
13375 (make_pass_stdarg): ...new function.
13376 * tree-switch-conversion.c (pass_convert_switch): Convert from a
13377 global struct to a subclass of gimple_opt_pass along with...
13378 (pass_data_convert_switch): ...new pass_data instance and...
13379 (make_pass_convert_switch): ...new function.
13380 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
13381 to a subclass of gimple_opt_pass along with...
13382 (pass_data_tail_recursion): ...new pass_data instance and...
13383 (make_pass_tail_recursion): ...new function.
13384 (pass_tail_calls): Convert from a global struct to a subclass of
13385 gimple_opt_pass along with...
13386 (pass_data_tail_calls): ...new pass_data instance and...
13387 (make_pass_tail_calls): ...new function.
13388 * tree-vect-generic.c (pass_lower_vector): Convert from a global
13389 struct to a subclass of gimple_opt_pass along with...
13390 (pass_data_lower_vector): ...new pass_data instance and...
13391 (make_pass_lower_vector): ...new function.
13392 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
13393 gimple_opt_pass along with...
13394 (pass_data_lower_vector_ssa): ...new pass_data instance and...
13395 (make_pass_lower_vector_ssa): ...new function.
13396 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
13397 to a subclass of gimple_opt_pass along with...
13398 (pass_data_slp_vectorize): ...new pass_data instance and...
13399 (make_pass_slp_vectorize): ...new function.
13400 (pass_ipa_increase_alignment): Convert from a global struct to a
13401 subclass of simple_ipa_opt_pass along with...
13402 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
13403 (make_pass_ipa_increase_alignment): ...new function.
13404 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
13405 gimple_opt_pass along with...
13406 (pass_data_vrp): ...new pass_data instance and...
13407 (make_pass_vrp): ...new function.
13408 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
13409 subclass of simple_ipa_opt_pass along with...
13410 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
13411 (make_pass_ipa_free_lang_data): ...new function.
13412 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
13413 gimple_opt_pass along with...
13414 (pass_data_tsan): ...new pass_data instance and...
13415 (make_pass_tsan): ...new function.
13416 (pass_tsan_O0): Convert from a global struct to a subclass of
13417 gimple_opt_pass along with...
13418 (pass_data_tsan_O0): ...new pass_data instance and...
13419 (make_pass_tsan_O0): ...new function.
13420 * var-tracking.c (pass_variable_tracking): Convert from a global
13421 struct to a subclass of rtl_opt_pass along with...
13422 (pass_data_variable_tracking): ...new pass_data instance and...
13423 (make_pass_variable_tracking): ...new function.
13424 * web.c (pass_web): Convert from a global struct to a subclass of
13425 rtl_opt_pass along with...
13426 (pass_data_web): ...new pass_data instance and...
13427 (make_pass_web): ...new function.
13428 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
13429 declaration with that of...
13430 (make_pass_mode_switch_use): ...new function.
13431 (pass_resolve_sw_modes): Replace declaration with that of...
13432 (make_pass_resolve_sw_modes): ...new function.
13433 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
13434 from a global struct to a subclass of rtl_opt_pass along with...
13435 (pass_data_mode_switch_use): ...new pass_data instance and...
13436 (make_pass_mode_switch_use): ...new function.
13437 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
13438 from a global struct to a subclass of rtl_opt_pass along with...
13439 (pass_data_resolve_sw_modes): ...new pass_data instance and...
13440 (make_pass_resolve_sw_modes): ...new function.
13441 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
13442 struct to a subclass of rtl_opt_pass along with...
13443 (pass_data_insert_vzeroupper): ...new pass_data instance and...
13444 (make_pass_insert_vzeroupper): ...new function.
13445 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
13446 global struct to a subclass of rtl_opt_pass along with...
13447 (pass_data_work_around_errata): ...new pass_data instance and...
13448 (make_pass_work_around_errata): ...new function.
13449 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
13450 struct to a subclass of rtl_opt_pass along with...
13451 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
13452 (make_pass_mips_machine_reorg2): ...new function.
13454 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13456 * passes.c (pass_manager::operator new): New.
13458 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13460 Handwritten part of conversion of passes to C++ classes.
13462 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
13463 (toplev.o): Add dep on PASS_MANAGER_H.
13464 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
13465 of early local pases to reflect this moving from a global to a
13466 member of gcc::pass_manager.
13467 (cgraph_add_new_function): Likewise.
13468 * lto-cgraph.c (lto_output_node): Update for conversion of
13469 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
13470 * passes.c (opt_pass::clone): New.
13471 (opt_pass::gate): New.
13472 (opt_pass::execute): New.
13473 (opt_pass::opt_pass): New.
13474 (pass_manager::execute_early_local_passes): New.
13475 (pass_manager::execute_pass_mode_switching): new.
13476 (finish_optimization_passes): Convert to...
13477 (pass_manager::finish_optimization_passes): ...this.
13478 (finish_optimization_passes): Update for conversion of passes to
13480 (register_dump_files_1): Use has_gate since we cannot portably
13481 check a vtable entry against NULL.
13482 (dump_one_pass): Likewise.
13483 (ipa_write_summaries_2): Likewise.
13484 (ipa_write_optimization_summaries_1): Likewise.
13485 (ipa_read_summaries_1): Likewise.
13486 (ipa_read_optimization_summaries_1): Likewise.
13487 (execute_ipa_stmt_fixups): Likewise.
13488 (pass_manager::pass_manager): Rewrite pass-creation, invoking
13489 pass-creation functions rather than wiring up globals, and
13490 storing the results in fields of pass_manager generated using
13491 pass-instances.def.
13492 (pass_manager::dump_profile_report): Update for conversion of
13493 passes to C++ classes.
13494 (pass_manager::execute_ipa_summary_passes): Likewise.
13495 (execute_one_ipa_transform_pass): Likewise.
13496 (execute_one_pass): Use has_gate and has_execute since we cannot
13497 portably check a vtable entry against NULL.
13498 * pass_manager.h (pass_manager::finish_optimization_passes): New.
13499 (pass_manager): Use pass-instances.def to add fields for the
13500 various pass instances.
13501 * toplev.c (finalize): Update for move of
13502 finish_optimization_passes to a method of gcc::pass_manager.
13503 * toplev.h (finish_optimization_passes): Move to method of class
13505 * tree-pass.h (struct pass_data): New.
13506 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
13507 (opt_pass::gate): Convert to virtual function.
13508 (opt_pass::~opt_pass): New.
13509 (opt_pass::clone): New.
13510 (opt_pass::execute): Convert to virtual function.
13511 (opt_pass::opt_pass): New.
13512 (opt_pass::ctxt_): new.
13513 (gimple_opt_pass): Convert to subclass of opt_pass.
13514 (gimple_opt_pass::gimple_opt_pass): New.
13515 (rtl_opt_pass): Convert to subclass of opt_pass.
13516 (rtl_opt_pass::rtl_opt_pass): New.
13517 (ipa_opt_pass_d): Convert to subclass of opt_pass.
13518 (ipa_opt_pass_d::ipa_opt_pass_d): New.
13519 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
13520 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
13521 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
13522 invocation of pass_mode_switching to reflect this moving from a
13523 global to a member of gcc::pass_manager.
13524 (ix86_option_override): Rework how pass_insert_vzeroupper is
13525 added to the pass_manager to reflect autogenerated changes.
13526 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
13528 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
13530 PR rtl-optimization/57708
13531 * recog.c (peep2_find_free_register): Validate all regs in a
13534 2013-08-05 Jan Hubicka <jh@suse.cz>
13537 * cgraph.c (verify_cgraph_node): Accept local flags from other
13539 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
13540 (function_and_variable_visibility): Likewise.
13541 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
13543 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13545 * graph.c (init_graph_slim_pretty_print): Remove.
13546 (print_graph_cfg): Do not call it. Use local pretty printer.
13547 (start_graph_dump): Likewise.
13549 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13551 * gimple-pretty-print.c (buffer): Remove.
13552 (initialized): Likewise.
13553 (maybe_init_pretty_print): Likewise.
13554 (print_gimple_stmt): Do not call it. Use non-static local
13555 pretty_printer variable.
13556 (print_gimple_expr): Likewise.
13557 (print_gimple_seq): Likewise.
13558 (gimple_dump_bb): Likewise.
13560 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13562 * asan.c (asan_pp): Remove.
13563 (asan_pp_initialized): Likewise.
13564 (asan_pp_initialize): Likewise.
13565 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
13566 (asan_emit_stack_protection): Tidy. Use local pretty printer.
13567 (asan_add_global): Likewise.
13569 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
13571 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
13572 * diagnostic.h (diagnostic_flush_buffer): Adjust.
13573 * pretty-print.c (pp_formatted_text_data): Likewise.
13574 (pp_indent): Rename from pp_base_indent.
13575 (pp_format): Rename from pp_base_format.
13576 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
13577 (pp_format_verbatim): Rename from pp_base_format_verbatim.
13578 (pp_flush): Rename from pp_base_flush.
13579 (pp_set_line_maximum_length): Rename from
13580 pp_base_set_line_maximum_length.
13581 (pp_clear_output_area): Rename from pp_base_clear_output_area.
13582 (pp_set_prefix): Rename from pp_base_set_prefix.
13583 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
13584 (pp_emit_prefix): Rename from pp_base_emit_prefix.
13585 (pp_append_text): Rename from pp_base_append_text.
13586 (pp_formatted_text): Rename from pp_base_formatted_text.
13587 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
13588 (pp_remaining_character_count_for_line): Rename from
13589 pp_base_remaining_character_count_for_line.
13590 (pp_newline): Rename from pp_base_newline.
13591 (pp_character): Rename from pp_base_character.
13592 (pp_string): Rename from pp_base_string.
13593 (pp_maybe_space): Rename from pp_base_maybe_space.
13594 * asan.c (asan_pp_string): Adjust.
13595 (asan_emit_stack_protection): Likewise.
13596 (asan_add_global): Likewise.
13597 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
13598 * tree-mudflap.c (mf_varname_tree): Likewise.
13599 * tree-pretty-print.c (pp_tree_identifier): Rename from
13600 pp_base_tree_identifier.
13601 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
13602 Declare as function.
13604 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13606 * pretty-print.h (pp_bar_bar): New.
13607 (pp_ampersand_ampersand): Likewise.
13608 (pp_less_equal): Likewise.
13609 (pp_greater_equal): Likewise.
13610 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
13611 printer functions instead of pp_string or operators and punctuators.
13612 (dump_gimple_call): Likewise.
13613 (dump_gimple_omp_for): Likewise.
13614 (dump_gimple_transaction): Likewise.
13615 (dump_gimple_phi): Likewise.
13616 (pp_gimple_stmt_1): Likewise.
13617 * sched-vis.c (print_insn): Likewise.
13618 * tree-mudflap.c (mf_varname_tree): Likewise.
13619 * tree-pretty-print.c (dump_block_node): Likewise.
13620 (dump_generic_node): Likewise.
13622 2013-08-02 Jan Hubicka <jh@suse.cz>
13624 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
13626 * lto-streamer-out.c (tree_is_indexable): Results decls and
13627 parm decls are not indexable.
13628 (DFS_write_tree_body): Do not follow args and results.
13629 (hash_tree): Likewise.
13630 (output_functions): Rearrange so struct function is needed
13631 only when real body is output; be able to also ouptut abstract
13632 functions; output DECL_ARGUMENTS and DECL_RESULT.
13633 (lto_output): When not in WPA, ale store abstract functions.
13634 (write_symbol): Do not care about RESULT_DECL.
13635 (output_symbol_p): Handle correctly sbtract decls.
13636 * lto-streamer-in.c (input_function): Rearrange so struct
13637 function can be NULL at entry; allow streaming of
13638 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
13639 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
13640 sanity check during LTO.
13641 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
13642 RESULT_DECl and DECL_ARGUMENTS.
13643 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
13646 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13648 * pretty-print.h (pp_underscore): New.
13650 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
13651 printer functions instead of pp_character.
13652 (dump_binary_rhs): Likewise.
13653 (dump_ternary_rhs): Likewise.
13654 (dump_gimple_call_args): Likewise.
13655 (pp_points_to_solution): Likewise.
13656 (dump_gimple_call): Likewise.
13657 (dump_gimple_switch): Likewise.
13658 (dump_gimple_cond): Likewise.
13659 (dump_gimple_bind): Likewise.
13660 (dump_gimple_try): Likewise.
13661 (dump_gimple_omp_for): Likewise.
13662 (dump_gimple_omp_continue): Likewise.
13663 (dump_gimple_omp_single): Likewise.
13664 (dump_gimple_omp_sections): Likewise.
13665 (dump_gimple_omp_block): Likewise.
13666 (dump_gimple_omp_critical): Likewise.
13667 (dump_gimple_transaction): Likewise.
13668 (dump_gimple_asm): Likewise.
13669 (dump_gimple_phi): Likewise.
13670 (dump_gimple_omp_parallel): Likewise.
13671 (dump_gimple_omp_task): Likewise.
13672 (dump_gimple_omp_atomic_load): Likewise.
13673 (dump_gimple_omp_atomic_store): Likewise.
13674 (dump_gimple_mem_ops): Likewise.
13675 (pp_gimple_stmt_1): Likewise.
13676 (pp_cfg_jump): Likewise.
13677 (dump_implicit_edges): Likewise.
13678 (gimple_dump_bb_for_graph): Likewise.
13679 * graph.c (draw_cfg_node): Likewise.
13680 * langhooks.c (lhd_print_error_function): Likewise.
13681 * sched-vis.c (print_exp): Likewise.
13682 (print_value): Likewise.
13683 (print_pattern): Likewise.
13684 (print_insn): Likewise.
13685 (rtl_dump_bb_for_graph): Likewise.
13686 * tree-pretty-print.c (dump_function_declaration): Likewise.
13687 (dump_array_domain): Likewise.
13688 (dump_omp_clause): Likewise.
13689 (dump_location): Likewise.
13690 (dump_generic_node): Likewise.
13691 (print_struct_decl): Likewise.
13692 * diagnostic.c (diagnostic_show_locus): Use pp_space.
13694 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
13696 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
13697 candidate table when replacing a candidate statement.
13698 (replace_rhs_if_not_dup): Likewise.
13699 (replace_one_candidate): Likewise.
13701 2013-08-02 Jan Hubicka <jh@suse.cz>
13702 Martin Liska <marxin.liska@gmail.com>
13704 * cgraphunit.c (add_new_function): Fix logic when adding from
13706 (assemble_thunk): Rename to ...
13707 (expand_thunk); .. this one; export; get it working with
13708 general functions; make produced gimple valid.
13709 * cgraph.h (expand_thunk): Declare.
13711 2013-08-02 Jan Hubicka <jh@suse.cz>
13713 * ipa-cp.c (gather_context_independent_values): Use
13714 ipa_get_param_move_cost.
13715 (get_replacement_map): Remove PARAM; move parameter folding
13717 (create_specialized_node): Update.
13718 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
13719 assert that we have gimple body; update move_cost.
13720 (count_formal_params): Assert that we have gimple body.
13721 (ipa_dump_param): New function.
13722 (ipa_alloc_node_params): Break out from ...
13723 (ipa_initialize_node_params): ... here.
13724 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
13725 (ipa_write_node_info): Stream move costs.
13726 (ipa_read_node_info): Read move costs.
13727 (ipa_update_after_lto_read): Do not recompute node params.
13728 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
13729 (ipa_get_param): Check we are not in WPA.
13730 (ipa_get_param_move_cost): New.
13731 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
13732 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
13733 parm numbers to be present.
13735 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
13737 PR rtl-optimization/58048
13738 * lra-constraints.c (process_alt_operands): Don't check asm
13739 operand on register.
13741 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
13743 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
13744 the implied StoreLoad barrier for atomic operations if before.
13746 2013-08-02 Jan Hubicka <jh@suse.cz>
13747 Martin Liska <marxin.liska@gmail.com>
13749 * cgraph.c (cgraph_function_body_availability): Do not check
13751 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
13752 symtab_node_availability): Declare.
13753 * ipa.c (can_replace_by_local_alias): New.
13754 (function_and_variable_visibility): Use it.
13755 * symtab.c (symtab_for_node_and_aliases,
13756 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
13758 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
13760 PR rtl-optimization/57963
13761 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
13762 (lra_constraints): Use them.
13764 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
13766 * config/arm/types.md (define_attr "type"): Add "load_acq"
13768 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
13770 (cortex_a53_store1): Likewise.
13772 2013-08-01 Jan Hubicka <jh@suse.cz>
13774 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
13775 partitions are not needed.
13777 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
13779 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
13780 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
13781 MAYBE_NON_Q_CLASS_P where appropriate.
13783 2013-08-01 Jan Hubicka <jh@suse.cz>
13785 * cgraph.h (release_function_body): Declare.
13786 * tree.c (free_lang_data_in_decl): Free, parameters and return values
13787 of unused delcarations.
13789 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13791 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
13792 RTL form when subtracting a constant.
13794 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13796 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
13797 Generate canonical plus rtx with negated immediate instead of minus
13799 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
13801 2013-08-01 Jan Hubicka <jh@suse.cz>
13803 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
13804 (cgraph_release_function_body): Likewise.
13805 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
13806 * cgraph.h (cgrpah_node): Rename abstract_and_needed
13807 to used_as_abstract_origin.
13808 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
13809 symbols used as abstract origins.
13810 * cgraphunit.c (analyze_functions): Update.
13811 * ipa.c (symtab_remove_unreachable_nodes): Recompute
13812 used_as_abstract_origin.
13813 * tree-inline.c (tree_function_versioning): Update
13814 used_as_abstract_origin; be ready for DECL_RESULT and
13815 DECL_ARGUMENTS to be NULL.
13817 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
13818 for abstract functions.
13819 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
13822 2013-08-01 Jan Hubicka <jh@suse.cz>
13824 * profile.c (compute_value_histograms): Fix thinko.
13826 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
13828 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
13829 aarch-common-protos.h to extra_headers.
13830 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
13831 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
13832 * config/aarch64/t-aarch64 (aarch-common.o): Define.
13834 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
13836 * config/aarch64/aarch64.md (define_attr "type"): Delete.
13837 Include "../arm/types.md". Define "type" attribute for all patterns.
13838 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
13841 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
13843 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
13844 to support power8 load fusion.
13845 (fusion_gpr_mem_load): Likewise.
13847 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
13849 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
13850 declarations for power8 load fusion.
13851 (emit_fusion_gpr_load): Likewise.
13853 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
13854 tuning for power8, turn on fusion mode by default. Turn on sign
13855 extending fusion mode if normal fusion mode is on, and we are at
13857 (fusion_gpr_load_p): New function, return true if we can fuse an
13858 addis instruction with a dependent load to a GPR.
13859 (emit_fusion_gpr_load): Emit the instructions for power8 load
13862 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
13863 (VSX load fusion peepholes): New peepholes to fuse together an
13864 addi instruction with a VSX load instruction.
13866 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
13867 peepholes to fuse an addis instruction with a load to a GPR base
13868 register. If we are supporting sign extending fusions, convert
13869 sign extending loads to zero extending loads and add an explicit
13872 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
13874 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
13875 aarch-common-protos.h to extra_headers.
13876 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
13877 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
13878 (arm_early_store_addr_dep): Likewise.
13879 (arm_no_early_alu_shift_dep): Likewise.
13880 (arm_no_early_alu_shift_value_dep): Likewise.
13881 (arm_no_early_mul_dep): Likewise.
13882 (arm_no_early_store_addr_dep): Likewise.
13883 (arm_mac_accumulator_is_mul_result): Likewise.
13884 (arm_mac_accumulator_is_result): Likewise.
13885 * config/arm/aarch-common.c: ... here. New file.
13886 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
13888 (arm_early_store_addr_dep): Likewise.
13889 (arm_no_early_alu_shift_dep): Likewise.
13890 (arm_no_early_alu_shift_value_dep): Likewise.
13891 (arm_no_early_mul_dep): Likewise.
13892 (arm_no_early_store_addr_dep): Likewise.
13893 (arm_mac_accumulator_is_mul_result): Likewise.
13894 (arm_mac_accumulator_is_result): Likewise.
13895 * config/arm/aarch-common-protos.h: ... here. New file.
13896 * config/arm/t-arm (aarch-common.o): Define.
13898 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
13900 * config/arm/arm.md: Include new file "types.md".
13901 (define_attr "type"): Move from here to ...
13902 (define_attr "mul32"): Likewise.
13903 (define_attr "mul64"): Likewise.
13904 * config/arm/types.md: ... here. New file.
13906 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
13908 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
13909 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
13911 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13913 * gen-pass-instances.awk: Fix offset of substr().
13915 2013-07-31 David Malcolm <dmalcolm@redhat.com>
13917 * Makefile.in (pass-instances.def): New.
13918 (passes.o): Replace dependency on passes.def with one on
13921 * gen-pass-instances.awk: New.
13923 * passes.c (pass_manager::pass_manager): Use pass-instances.def
13924 rather than passes.def, updating local definition of NEXT_PASS
13925 macro to add an extra NUM parameter (currently unused).
13927 2013-07-30 David Malcolm <dmalcolm@redhat.com>
13929 * Makefile.in (PASS_MANAGER_H): New.
13930 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
13931 (passes.o): Likewise.
13932 (statistics.o): Likewise.
13933 (cgraphunit.o): Likewise.
13934 (context.o): Depend on PASS_MANAGER_H.
13936 * pass_manager.h: New.
13938 * cgraphunit.c (cgraph_add_new_function): Update for moves
13939 of globals to fields of pass_manager.
13940 (analyze_function): Likewise.
13941 (expand_function): Likewise.
13942 (ipa_passes): Likewise.
13943 (compile): Likewise.
13945 * context.c (context::context): New.
13946 * context.h (context::context): New.
13947 (context::get_passes): New.
13948 (context::passes_): New.
13950 * lto-cgraph.c (input_node): Update for moves of globals to
13951 fields of pass_manager.
13953 * passes.c (all_passes): Remove, in favor of a field of the
13954 same name within the new class pass_manager.
13955 (all_small_ipa_passes): Likewise.
13956 (all_lowering_passes): Likewise.
13957 (all_regular_ipa_passes): Likewise.
13958 (all_late_ipa_passes): Likewise.
13959 (all_lto_gen_passes): Likewise.
13960 (passes_by_id): Likewise.
13961 (passes_by_id_size): Likewise.
13962 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
13963 the new class pass_manager.
13964 (set_pass_for_id): Convert to...
13965 (pass_manager::set_pass_for_id): ...method.
13966 (get_pass_for_id): Convert to...
13967 (pass_manager::get_pass_for_id): ...method.
13968 (register_one_dump_file): Move body of implementation into...
13969 (pass_manager::register_one_dump_file): ...here.
13970 (register_dump_files_1): Convert to...
13971 (pass_manager::register_dump_files_1): ...method.
13972 (register_dump_files): Convert to...
13973 (pass_manager::register_dump_files): ...method.
13974 (create_pass_tab): Update for moves of globals to fields of
13976 (dump_passes): Move body of implementation into...
13977 (pass_manager::dump_passes): ...here.
13978 (register_pass): Move body of implementation into...
13979 (pass_manager::register_pass): ...here.
13980 (init_optimization_passes): Convert into...
13981 (pass_manager::pass_manager): ...constructor for new
13982 pass_manager class, and initialize the pass_lists array.
13983 (check_profile_consistency): Update for moves of globals to
13984 fields of pass_manager.
13985 (dump_profile_report): Move body of implementation into...
13986 (pass_manager::dump_profile_report): ...here.
13987 (ipa_write_summaries_1): Update for moves of pass lists from
13988 being globals to fields of pass_manager.
13989 (ipa_write_optimization_summaries): Likewise.
13990 (ipa_read_summaries): Likewise.
13991 (ipa_read_optimization_summaries): Likewise.
13992 (execute_all_ipa_stmt_fixups): Likewise.
13994 * statistics.c (statistics_fini): Update for moves of globals to
13995 fields of pass_manager.
13997 * toplev.c (general_init): Replace call to
13998 init_optimization_passes with construction of the pass_manager
14001 * tree-pass.h (all_passes): Remove, in favor of a field of the
14002 same name within the new class pass_manager.
14003 (all_small_ipa_passes): Likewise.
14004 (all_lowering_passes): Likewise.
14005 (all_regular_ipa_passes): Likewise.
14006 (all_lto_gen_passes): Likewise.
14007 (all_late_ipa_passes): Likewise.
14008 (passes_by_id): Likewise.
14009 (passes_by_id_size): Likewise.
14010 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14011 the new class pass_manager.
14012 (get_pass_for_id): Remove.
14014 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14016 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
14019 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14021 * arm.md (mulhi3): New expand pattern.
14023 2013-07-30 Jan Hubicka <jh@suse.cz>
14024 Martin Liska <marxin.liska@gmail.com>
14026 * profile.c (compute_value_histograms): Do not ICE when
14027 there is mismatch only on some counters.
14029 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14031 PR rtl-optimization/57637
14032 * function.c (move_insn_for_shrink_wrap): Also check the
14033 GEN set of the LIVE problem for the liveness analysis
14034 if it exists, otherwise give up.
14036 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14038 PR tree-optimization/57993
14039 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
14040 replaced statement in the candidate table.
14041 (phi_add_costs): Return infinite cost when the hidden basis does
14042 not dominate all phis on which the candidate is dependent.
14043 (replace_one_candidate): Record replaced statement in the
14046 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
14048 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
14049 (ashlv2si3): New expander.
14050 (*ashlv2si3_i): New define_insn_and_split.
14051 * predicates.md (float_operation): Allow patterns with three
14052 basic sub-patterns.
14054 PR rtl-optimization/58021
14055 * mode-switching.c (create_pre_exit): Always split off preceding
14056 insns if we are not at the basic block head.
14058 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
14060 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
14061 (UCLIBC_DYNAMIC_LINKER): New macro.
14062 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
14064 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
14065 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
14067 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
14068 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
14069 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
14070 for SF and DF modes. Use ieee_quad_format for TF mode.
14071 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
14072 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
14073 (mips_option_override): Handle `-mnan=legacy'.
14074 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
14075 `-mabs=2008' and `-mnan=2008'.
14076 (OPTION_DEFAULT_SPECS): Add "nan" default.
14077 (ASM_SPEC): Handle `-mnan='.
14078 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
14079 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
14080 comment accordingly.
14081 (neg<mode>2): Likewise.
14082 * config/mips/mips.opt (mabs, mnan): New options.
14083 * doc/install.texi (Configuration): Document `--with-nan=' option.
14084 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
14086 (MIPS Options): Document them.
14087 * config.gcc <mips*-*-*>: Handle `--with-nan='.
14088 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
14089 * configure: Regenerate.
14090 * config.in: Regenerate.
14092 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14094 * config/i386/i386.md (float post-reload splitters): Do not check
14095 for subregs of SSE registers.
14097 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14098 H.J. Lu <hongjiu.lu@intel.com>
14102 * config/i386/i386.md (post-reload splitter
14103 to avoid partial SSE reg dependency stalls): New pattern.
14105 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
14107 * config/s390/s390.md ("movcc"): Swap load and store instructions.
14109 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
14111 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14112 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
14114 2013-07-26 Cary Coutant <ccoutant@google.com>
14116 * dwarf2out.c (die_checksum_ordered): Don't include template
14117 instantiations in signature.
14118 (is_template_parameter): New function.
14119 (is_template_instantiation): New function.
14120 (generate_skeleton_bottom_up): Don't include template instantiations
14122 (generate_skeleton): Likewise.
14123 (break_out_comdat_types): Move recursive call to break out nested
14125 (prune_unused_types_mark_generic_parms_dies): Call
14126 is_template_parameter.
14128 2013-07-26 Ian Bolton <ian.bolton@arm.com>
14130 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
14131 uses vector registers.
14132 * config/aarch64/iterators.md: Add attributes rtn and vas.
14134 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14135 Richard Earnshaw <richard.earnshaw@arm.com>
14137 * combine.c (simplify_comparison): Re-canonicalize operands
14139 * config/arm/arm.md (movcond_addsi): New splitter.
14141 2013-07-25 Sterling Augustine <saugustine@google.com>
14143 * dwarf2out.c (size_of_pubnames): Move code to...
14144 (include_pubname_in_output): ...here. New.
14145 (want_pubnames): Rearrange.
14146 (output_pubnames): Call include_pubname_in_output. Move assertion.
14148 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14150 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
14152 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14155 * doc/extend.texi: Remove obsolete builtins. Fix
14156 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
14158 2013-07-25 Jan Hubicka <jh@suse.cz>
14160 * cgraph.c (release_function_body): Break out from ...
14161 (cgraph_release_function_body): ... this one; also release DECL_RESULT
14162 and DECL_ARGUMENTS.
14163 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
14164 old_tree in the map.
14165 (create_specialized_node): Update.
14166 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
14168 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
14169 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
14170 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
14171 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
14174 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14176 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
14177 addsi3_carryin_alt2_<optab>): Correct output template.
14179 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14181 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
14182 Adjust for arm_restrict_it.
14183 Remove trailing whitespace.
14185 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
14187 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
14188 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
14190 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
14192 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
14194 PR rtl-optimization/57960
14195 * lra-constraints.c (process_alt_operands): Use the right mode
14196 when checking strict_low.
14198 2013-07-25 Jan Hubicka <jh@suse.cz>
14200 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
14201 * cgraph.c (cgraph_remove_node): Do not release function body
14202 when in cgraph streaming.
14203 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
14204 in other partitions are not considered reachable; fix handling of
14207 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14209 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
14211 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14216 * config/arm/arm.md ("*sibcall_insn): Replace use of
14217 Ss with US. Adjust output for v5 and v4t.
14218 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
14220 * config/arm/constraints.md ("Ss"): Rename to US.
14222 2013-07-25 Terry Guo <terry.guo@arm.com>
14224 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
14225 shift_add/shift_sub0/shift_sub1 RTXs.
14227 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
14228 Anton Blanchard <anton@au1.ibm.com>
14230 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
14231 (altivec_vpks<VI_char>ss): Likewise.
14232 (altivec_vpks<VI_char>us): Likewise.
14233 (altivec_vpku<VI_char>us): Likewise.
14234 (altivec_vpku<VI_char>um): Likewise.
14236 2013-07-24 David Malcolm <dmalcolm@redhat.com>
14238 Introduce context class.
14240 * Makefile.in (CONTEXT_H): New.
14241 (OBJS): Add context.o.
14242 (toplev.o): Add CONTEXT_H to dependencies.
14245 * toplev.c (general_init): Create the singleton gcc::context instance.
14251 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
14253 PR rtl-optimization/57968
14254 * mode-switching.c (create_pre_exit): Allow instructions that
14255 don't set a return register to need a non-exit mode.
14257 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14258 Anton Blanchard <anton@au1.ibm.com>
14260 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
14261 operands to vperm for little endian.
14262 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
14263 of lvsl to create the control mask for a vperm for little endian.
14265 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14266 Anton Blanchard <anton@au1.ibm.com>
14268 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
14269 two operands for little-endian.
14271 2013-07-23 Steve Ellcey <sellcey@mips.com>
14273 * config/mips/mips.c (mips_case_values_threshold): New.
14274 (TARGET_CASE_VALUES_THRESHOLD): Define.
14276 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14277 Anton Blanchard <anton@au1.ibm.com>
14279 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
14280 selection of field for vector splat in little endian mode.
14282 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
14284 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
14285 expanders to rs6000.md.
14286 (ior<mode>3): Likewise.
14287 (and<mode>3): Likewise.
14288 (one_cmpl<mode>2): Likewise.
14289 (nor<mode>3): Likewise.
14290 (andc<mode>3): Likewise.
14291 (eqv<mode>3): Likewise.
14292 (nand<mode>3): Likewise.
14293 (orc<mode>3): Likewise.
14295 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
14298 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
14299 to split multi-word logical operations.
14300 (rs6000_split_logical_di): Likewise.
14301 (rs6000_split_logical): Likewise.
14303 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
14304 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
14305 and allow TImode operations in 32-bit.
14306 (vsx_and<mode>3_64bit): Likewise.
14307 (vsx_ior<mode>3_32bit): Likewise.
14308 (vsx_ior<mode>3_64bit): Likewise.
14309 (vsx_xor<mode>3_32bit): Likewise.
14310 (vsx_xor<mode>3_64bit): Likewise.
14311 (vsx_one_cmpl<mode>2_32bit): Likewise.
14312 (vsx_one_cmpl<mode>2_64bit): Likewise.
14313 (vsx_nor<mode>3_32bit): Likewise.
14314 (vsx_nor<mode>3_64bit): Likewise.
14315 (vsx_andc<mode>3_32bit): Likewise.
14316 (vsx_andc<mode>3_64bit): Likewise.
14317 (vsx_eqv<mode>3_32bit): Likewise.
14318 (vsx_eqv<mode>3_64bit): Likewise.
14319 (vsx_nand<mode>3_32bit): Likewise.
14320 (vsx_nand<mode>3_64bit): Likewise.
14321 (vsx_orc<mode>3_32bit): Likewise.
14322 (vsx_orc<mode>3_64bit): Likewise.
14324 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
14325 logical types in GPRs.
14327 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
14328 logical insns to rs6000.md, and allow TImode operations in
14330 (altivec_ior<mode>3): Likewise.
14331 (altivec_xor<mode>3): Likewise.
14332 (altivec_one_cmpl<mode>2): Likewise.
14333 (altivec_nor<mode>3): Likewise.
14334 (altivec_andc<mode>3): Likewise.
14336 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
14337 attributes for moving the 128-bit logical operations into
14339 (BOOL_REGS_OUTPUT): Likewise.
14340 (BOOL_REGS_OP1): Likewise.
14341 (BOOL_REGS_OP2): Likewise.
14342 (BOOL_REGS_UNARY): Likewise.
14343 (BOOL_REGS_AND_CR0): Likewise.
14344 (one_cmpl<mode>2): Add support for DI logical operations on
14345 32-bit, splitting the operations to 32-bit.
14346 (anddi3): Likewise.
14347 (iordi3): Likewise.
14348 (xordi3): Likewise.
14349 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
14350 changes to combine the 32/64-bit code, allow logical operations on
14351 TI mode in 32-bit, and to use similar match_operator patterns like
14352 scalar mode uses. Combine the Altivec and VSX code for logical
14353 operations, and move it here.
14354 (ior<mode>3, 128-bit types): Likewise.
14355 (xor<mode>3, 128-bit types): Likewise.
14356 (one_cmpl<mode>3, 128-bit types): Likewise.
14357 (nor<mode>3, 128-bit types): Likewise.
14358 (andc<mode>3, 128-bit types): Likewise.
14359 (eqv<mode>3, 128-bit types): Likewise.
14360 (nand<mode>3, 128-bit types): Likewise.
14361 (orc<mode>3, 128-bit types): Likewise.
14362 (and<mode>3_internal): Likewise.
14363 (bool<mode>3_internal): Likewise.
14364 (boolc<mode>3_internal1): Likewise.
14365 (boolc<mode>3_internal2): Likewise.
14366 (boolcc<mode>3_internal1): Likewise.
14367 (boolcc<mode>3_internal2): Likewise.
14368 (eqv<mode>3_internal1): Likewise.
14369 (eqv<mode>3_internal2): Likewise.
14370 (one_cmpl1<mode>3_internal): Likewise.
14372 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14374 * config/microblaze/microblaze.c (microblaze_expand_prologue):
14375 Rename flag_stack_usage to flag_stack_usage_info.
14377 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14379 * config/microblaze/sync.md: New file.
14380 * config/microblaze/microblaze.md: Include sync.md
14381 * config/microblaze/microblaze.c: Add print_operand 'y'.
14382 * config/microblaze/constraints.md: Add memory_contraint
14383 'Q' which is a single register.
14385 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
14387 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
14389 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14391 * reload.c (find_reloads): Exit loop once we find this operand
14392 cannot be reloaded somehow for this alternative.
14394 * reload.c (find_reloads): Exit loop once we find a hard register.
14396 * rtlanal.c (computed_jump_p): Exit loop once we find label
14399 * i386.c (ix86_pad_returns): Exit loop after setting replace.
14401 * cfgloopmanip.c (remove_path): Exit loop after setting
14404 * gensupport.c (subst_dup): Avoid loop if code is not
14405 MATCH_DUP nor MATCH_OP_DUP.
14407 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
14409 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
14411 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14413 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
14414 true for SP_REGNUM if mode == ptr_mode.
14415 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
14416 with value R0_REGNUM + 31.
14418 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14420 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
14421 pad pointer-typed argument downward.
14423 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14425 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
14426 and __ILP32__ when the ILP32 model is in use.
14428 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14430 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
14431 (aarch64_load_symref_appropriately): In the case of
14432 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
14433 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
14434 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
14435 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
14436 if 'mode' doesn't equal to 'ptr_mode'.
14437 (aarch64_output_mi_thunk): Add an assertion on the alignment of
14438 'vcall_offset'; change to call aarch64_emit_move differently depending
14439 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
14440 to calculate the upper bound of 'vcall_offset'.
14441 (aarch64_cannot_force_const_mem): Change to also return true if
14443 (aarch64_legitimize_reload_address): In the case of large
14444 displacements, add new local variable 'xmode' and an assertion
14445 based on it; change to use 'xmode' to generate the new rtx and
14447 (aarch64_asm_trampoline_template): Change to generate the template
14448 differently depending on TARGET_ILP32 or not; change to use
14449 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
14450 (aarch64_trampoline_size): Removed.
14451 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
14452 and replace immediate literals with it. Change to use 'ptr_mode'
14453 instead of 'DImode' and call convert_memory_address if the mode
14454 of 'fnaddr' doesn't equal to 'ptr_mode'.
14455 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
14457 (aarch64_elf_asm_destructor): Likewise.
14458 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
14459 on TARGET_ILP32 instead of aarch64_trampoline_size.
14460 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
14461 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
14462 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
14463 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
14464 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
14465 (storewb_pair<GPI:mode>_<P:mode>): ... this.
14466 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
14467 depending on the value of 'mode'.
14468 (add_losym_<mode>): New.
14469 (ldr_got_small_<mode>): New, based on ldr_got_small.
14470 (ldr_got_small): Remove.
14471 (ldr_got_small_sidi): New.
14472 * config/aarch64/iterators.md (P): New.
14473 (PTR): Change to 'ptr_mode' in the condition.
14475 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14477 * config.gcc (aarch64*-*-*): Support --with-abi.
14478 (aarch64*-*-elf): Support --with-multilib-list.
14479 (aarch64*-*-linux*): Likewise.
14480 (supported_defaults): Add abi to aarch64*-*-*.
14481 * configure.ac: Mention AArch64 for --with-multilib-list.
14482 * configure: Re-generated.
14483 * config/aarch64/biarchilp32.h: New file.
14484 * config/aarch64/biarchlp64.h: New file.
14485 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
14487 (MULTILIB_DEFAULTS): Ditto.
14488 (DRIVER_SELF_SPECS): Ditto.
14489 (ASM_SPEC): Update to also substitute -mabi.
14490 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
14491 file whose name depends on -mabi= and -mbig-endian.
14492 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
14494 (POINTER_SIZE): New define.
14495 (POINTERS_EXTEND_UNSIGNED): Ditto.
14496 (enum aarch64_abi_type): New enumeration tag.
14497 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
14498 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
14499 (TARGET_ILP32): New define.
14500 * config/aarch64/aarch64.opt (mabi): New.
14501 (aarch64_abi): New.
14502 (ilp32, lp64): New values for -mabi.
14503 * config/aarch64/t-aarch64 (comma): New define.
14504 (MULTILIB_OPTIONS): Ditto.
14505 (MULTILIB_DIRNAMES): Ditto.
14506 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
14507 * doc/invoke.texi: Document -mabi for AArch64.
14509 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
14511 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
14513 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14514 Anton Blanchard <anton@au1.ibm.com>
14516 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
14517 endianness when selecting field to splat.
14519 2013-07-22 Eric Christopher <echristo@gmail.com>
14521 * dwarf2out.c (die_odr_checksum): New function to use
14522 CHECKSUM_ macros and ULEB128 for DIE tag.
14523 (generate_type_signature): Use.
14525 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
14527 * config.gcc (sparc*-*-*): Accept leon3 processor.
14528 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
14529 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
14530 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
14531 * config/sparc/sparc.opt (enum processor_type): Add leon3.
14532 (mfix-ut699): Adjust comment.
14533 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
14534 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
14535 (CPP_CPU_SPEC): Likewise.
14536 (ASM_CPU_SPEC): Likewise.
14537 * config/sparc/sparc.c (leon3_cost): New constant.
14538 (sparc_option_override): Add leon3 support.
14539 (mem_ref): New function.
14540 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
14541 (sparc_do_work_around_errata): Look into the instruction in the delay
14542 slot and adjust accordingly. Add fix for the data cache nullify issues
14543 of the UT699. Change insertion position for the NOP.
14544 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
14545 (leon3_load): New reservation.
14546 (leon_store): Bump latency to 2.
14547 (grfpu): New automaton.
14548 (grfpu_alu): New unit.
14549 (grfpu_ds): Likewise.
14550 (leon_fp_alu): Adjust.
14551 (leon_fp_mult): Delete.
14552 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
14553 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
14554 * config/sparc/sparc.md (cpu): Add leon3.
14555 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
14556 (swapsi): Likewise.
14557 (atomic_test_and_set): Likewise.
14558 (ldstub): Likewise.
14560 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
14562 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
14563 default for R5900 targets.
14564 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
14565 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
14566 * config/mips/mips.c (mips_option_override): Report an error for
14567 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
14568 for -march=r5900 -mhard-float.
14570 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14572 * df-problems.c (can_move_insns_across): Exit loop once we
14573 find a non-fixed, non-global register.
14575 * ipa-pure-const.c (propagate_nothrow): Exit loop after
14578 * omega.c (omega_eliminate_red): Break after setting red_found.
14579 (omega_problem_has_red_equations): Similarly after setting found.
14580 (omega_query_variable): Similarly after setting coupled.
14582 2013-07-22 Marek Polacek <polacek@redhat.com>
14584 * gimplify.c: Don't include gimple.h twice.
14586 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14588 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
14589 instead of TARGET_THUMB1.
14590 (Pz): New constraint.
14591 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
14593 (compare_negsi_si): Likewise.
14594 (compare_addsi2_op0): Likewise.
14595 (compare_addsi2_op1): Likewise.
14596 (addsi3_carryin_<optab>): Likewise.
14597 (addsi3_carryin_alt2_<optab>): Likewise.
14598 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
14599 for arm_restrict_it.
14600 (subsi3_carryin): Likewise.
14601 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
14602 (minmax_arithsi): Disable for arm_restrict_it.
14603 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
14604 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
14605 (satsi_<SAT:code>_shift): Likewise.
14606 (arm_shiftsi3): Add alternative for 16-bit encoding.
14607 (arm32_movhf): Disable for arm_restrict_it.
14608 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
14609 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
14611 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
14613 * config/arm/arm.md (attribute "insn"): Delete.
14614 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
14615 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
14616 (not_shiftsi): Update for attribute change.
14617 (not_shiftsi_compare0): Likewise.
14618 (not_shiftsi_compare0_scratch): Likewise.
14619 (arm_one_cmplsi2): Likewise.
14620 (thumb1_one_cmplsi2): Likewise.
14621 (notsi_compare0): Likewise.
14622 (notsi_compare0_scratch): Likewise.
14623 (thumb1_movdi_insn): Likewise.
14624 (arm_movsi_insn): Likewise.
14625 (movhi_insn_arch4): Likewise.
14626 (movhi_bytes): Likewise.
14627 (arm_movqi_insn): Likewise.
14628 (thumb1_movqi_insn): Likewise.
14629 (arm32_movhf): Likewise.
14630 (thumb1_movhf): Likewise.
14631 (arm_movsf_soft_insn): Likewise.
14632 (thumb1_movsf_insn): Likewise.
14633 (thumb_movdf_insn): Likewise.
14634 (movsicc_insn): Likewise.
14635 (movsfcc_soft_insn): Likewise.
14636 (and_scc): Likewise.
14637 (cond_move): Likewise.
14638 (if_move_not): Likewise.
14639 (if_not_move): Likewise.
14640 (if_shift_move): Likewise.
14641 (if_move_shift): Likewise.
14642 (if_shift_shift): Likewise.
14643 (if_not_arith): Likewise.
14644 (if_arith_not): Likewise.
14645 (cond_move_not): Likewise.
14646 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
14647 (neon_mov<mode>): Likewise.
14648 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
14649 (thumb2_movsi_vfp): Likewise.
14650 (movsf_vfp): Likewise.
14651 (thumb2_movsf_vfp): Likewise.
14652 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
14654 (cortexa7_older_only): Likewise.
14655 (cortexa7_younger): Likewise.
14656 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
14657 (1020alu_shift_op): Likewise.
14658 (1020alu_shift_reg_op): Likewise.
14659 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
14660 (alu_shift_op): Likewise.
14661 (alu_shift_reg_op): Likewise.
14662 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
14663 (11_alu_shift_op): Likewise.
14664 (11_alu_shift_reg_op): Likewise.
14665 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
14666 (9_alu_shift_reg_op): Likewise.
14667 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
14669 (cortex_a15_alu_shift): Likewise.
14670 (cortex_a15_alu_shift_reg): Likewise.
14671 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
14672 (cortex_a5_alu_shift): Likewise.
14673 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
14675 (cortex_a53_alu_shift): Likewise.
14676 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
14678 (cortex_a7_alu_reg): Likewise.
14679 (cortex_a7_alu_shift): Likewise.
14680 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
14681 (cortex_a8_alu_shift): Likewise.
14682 (cortex_a8_alu_shift_reg): Likewise.
14683 (cortex_a8_mov): Likewise.
14684 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
14685 (cortex_a9_dp_shift): Likewise.
14686 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
14687 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
14688 (cortex_r4_mov): Likewise.
14689 (cortex_r4_alu_shift): Likewise.
14690 (cortex_r4_alu_shift_reg): Likewise.
14691 * config/arm/fa526.md (526_alu_op): Update for attribute change.
14692 (526_alu_shift_op): Likewise.
14693 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
14694 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
14695 (626te_alu_shift_op): Likewise.
14696 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
14697 (726te_alu_op): Likewise.
14698 (726te_alu_shift_op): Likewise.
14699 (726te_alu_shift_reg_op): Likewise.
14700 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
14701 (mp626_alu_shift_op): Likewise.
14702 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
14703 (pj4_alu_e1_conds): Likewise.
14704 (pj4_alu): Likewise.
14705 (pj4_alu_conds): Likewise.
14706 (pj4_shift): Likewise.
14707 (pj4_shift_conds): Likewise.
14708 (pj4_alu_shift): Likewise.
14709 (pj4_alu_shift_conds): Likewise.
14711 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14713 * config/arm/predicates.md (shiftable_operator_strict_it):
14715 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
14716 Disable cond_exec version for arm_restrict_it.
14717 (thumb2_smaxsi3): Convert to generate cond_exec.
14718 (thumb2_sminsi3): Likewise.
14719 (thumb32_umaxsi3): Likewise.
14720 (thumb2_uminsi3): Likewise.
14721 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
14722 (thumb2_neg_abssi2): Likewise.
14723 (thumb2_mov_scc): Add alternative for 16-bit encoding.
14724 (thumb2_movsicc_insn): Adjust alternatives.
14725 (thumb2_mov_negscc): Disable for arm_restrict_it.
14726 (thumb2_mov_negscc_strict_it): New pattern.
14727 (thumb2_mov_notscc_strict_it): New pattern.
14728 (thumb2_mov_notscc): Disable for arm_restrict_it.
14729 (thumb2_ior_scc): Likewise.
14730 (thumb2_ior_scc_strict_it): New pattern.
14731 (thumb2_cond_move): Adjust for arm_restrict_it.
14732 (thumb2_cond_arith): Disable for arm_restrict_it.
14733 (thumb2_cond_arith_strict_it): New pattern.
14734 (thumb2_cond_sub): Adjust for arm_restrict_it.
14735 (thumb2_movcond): Likewise.
14736 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
14737 (thumb2_zero_extendhisi2_v6): Likewise.
14738 (thumb2_zero_extendqisi2_v6): Likewise.
14739 (orsi_notsi_si): Likewise.
14740 (orsi_not_shiftsi_si): Likewise.
14742 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
14744 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
14745 instruction sequence is 1 byte shorter.
14747 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
14749 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
14750 it is not needed after split.
14752 2013-07-20 Iain Sandoe <iain@codesourcery.com>
14755 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
14756 second label for nonlocal goto receivers. Don't output pic base labels
14757 unless we're producing PIC; mark that action unreachable().
14758 (ix86_save_reg): If the function contains a nonlocal label, save the
14760 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
14761 * config/darwin.c (emitted_pic_label_num): New GTY.
14762 (update_pic_label_number_if_needed): New.
14763 (machopic_output_function_base_name): Adjust for nonlocal receiver
14765 (machopic_should_output_picbase_label): New.
14766 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
14767 (nonlocal_goto_receiver): New insn and split.
14769 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
14771 * config/aarch64/aarch64-builtins.c
14772 (aarch64_fold_builtin): Fold abs in all modes.
14773 * config/aarch64/aarch64-simd-builtins.def
14774 (abs): Enable for all modes.
14775 * config/aarch64/arm_neon.h
14776 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
14777 (vabs_f64): Add missing intrinsic.
14779 2013-07-19 Ian Bolton <ian.bolton@arm.com>
14781 * config/aarch64/arm_neon.h (vabs_s64): New function
14783 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
14786 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
14787 * config/avr/avr.md (adjust_len): Add `round'.
14788 * config/avr/avr-protos.h (avr_out_round): New prototype.
14789 (avr_out_plus): Add `out_label' argument.
14790 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
14791 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
14792 Handle the case where `insn' is just a pattern.
14793 (avr_out_bitop): Handle the case where `insn' is just a pattern.
14794 (avr_out_round): New function.
14795 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
14797 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
14799 * config/microblaze/microblaze.c (microblaze_expand_prologue):
14800 Add check for flag_stack_usage to handle -fstack-usage support
14802 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
14804 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
14805 interaction for new Power8 flags and VSX.
14807 2013-07-18 Sriraman Tallam <tmsriram@google.com>
14809 PR middle-end/57698
14810 * tree-inline.c (expand_call_inline): Emit errors during
14811 early_inlining only if optimization is not turned on.
14813 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14817 * passes.c (init_optimization_passes): Move the construction of
14818 the pass hierarchy into a new passes.def file.
14820 * Makefile.in (passes.o): Add dependency on passes.def.
14822 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14824 * passes.c (init_optimization_passes): Introduce macros for
14825 constructing the tree of passes (INSERT_PASSES_AFTER,
14826 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
14827 TERMINATE_PASS_LIST).
14829 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
14830 Wei Mi <wmi@google.com>
14832 PR rtl-optimization/57878
14833 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
14835 (reload_pseudo_compare_func): Check nregs first for reload
14838 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14840 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
14842 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
14844 * read-rtl.c (validate_const_int): Once an invalid character is
14845 seen, quit the loop.
14847 * gengtype.c (write_roots): Similarly once we find the "deletable"
14848 or "if_marked" option.
14850 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
14852 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
14853 "xtab" and "sat". Move value "clz" from here to ...
14854 (attriubte "type"): ... here.
14855 (satsi_<SAT:code>): Delete "insn" attribute.
14856 (satsi_<SAT:code>_shift): Likewise.
14857 (arm_zero_extendqisi2addsi): Likewise.
14858 (arm_extendqisi2addsi): Likewise.
14859 (clzsi2): Update for attribute changes.
14860 (rbitsi2): Likewise.
14861 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
14863 (arm_usatsihi): Likewise.
14864 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
14866 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
14868 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
14869 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
14870 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
14871 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
14872 in alphabetical order.
14873 (attribute "core_cycles"): Update for attribute changes.
14874 (arm_addsi3): Likewise.
14875 (addsi3_compare0): Likewise.
14876 (addsi3_compare0_scratch): Likewise.
14877 (addsi3_compare_op1): Likewise.
14878 (addsi3_compare_op2): Likewise.
14879 (compare_addsi2_op0): Likewise.
14880 (compare_addsi2_op1): Likewise.
14881 (addsi3_carryin_shift_<optab>): Likewise.
14882 (subsi3_carryin_shift): Likewise.
14883 (rsbsi3_carryin_shift): Likewise.
14884 (arm_subsi3_insn): Likewise.
14885 (subsi3_compare0): Likewise.
14886 (subsi3_compare): Likewise.
14887 (arm_andsi3_insn): Likewise.
14888 (thumb1_andsi3_insn): Likewise.
14889 (andsi3_compare0): Likewise.
14890 (andsi3_compare0_scratch): Likewise.
14891 (zeroextractsi_compare0_scratch
14892 (andsi_not_shiftsi_si): Likewise.
14893 (iorsi3_insn): Likewise.
14894 (iorsi3_compare0): Likewise.
14895 (iorsi3_compare0_scratch): Likewise.
14896 (arm_xorsi3): Likewise.
14897 (thumb1_xorsi3_insn): Likewise.
14898 (xorsi3_compare0): Likewise.
14899 (xorsi3_compare0_scratch): Likewise.
14900 (satsi_<SAT:code>_shift): Likewise.
14902 (arm_shiftsi3): Likewise.
14903 (shiftsi3_compare0): Likewise.
14904 (not_shiftsi): Likewise.
14905 (not_shiftsi_compare0): Likewise.
14906 (not_shiftsi_compare0_scratch): Likewise.
14907 (arm_one_cmplsi2): Likewise.
14908 (thumb_one_complsi2): Likewise.
14909 (notsi_compare0): Likewise.
14910 (notsi_compare0_scratch): Likewise.
14911 (thumb1_zero_extendhisi2): Likewise.
14912 (arm_zero_extendhisi2): Likewise.
14913 (arm_zero_extendhisi2_v6): Likewise.
14914 (arm_zero_extendhisi2addsi): Likewise.
14915 (thumb1_zero_extendqisi2): Likewise.
14916 (thumb1_zero_extendqisi2_v6): Likewise.
14917 (arm_zero_extendqisi2): Likewise.
14918 (arm_zero_extendqisi2_v6): Likewise.
14919 (arm_zero_extendqisi2addsi): Likewise.
14920 (thumb1_extendhisi2): Likewise.
14921 (arm_extendhisi2): Likewise.
14922 (arm_extendhisi2_v6): Likewise.
14923 (arm_extendqisi): Likewise.
14924 (arm_extendqisi_v6): Likewise.
14925 (arm_extendqisi2addsi): Likewise.
14926 (thumb1_extendqisi2): Likewise.
14927 (thumb1_movdi_insn): Likewise.
14928 (arm_movsi_insn): Likewise.
14929 (movsi_compare0): Likewise.
14930 (movhi_insn_arch4): Likewise.
14931 (movhi_bytes): Likewise.
14932 (arm_movqi_insn): Likewise.
14933 (thumb1_movqi_insn): Likewise.
14934 (arm32_movhf): Likewise.
14935 (thumb1_movhf): Likewise.
14936 (arm_movsf_soft_insn): Likewise.
14937 (thumb1_movsf_insn): Likewise.
14938 (movdf_soft_insn): Likewise.
14939 (thumb_movdf_insn): Likewise.
14940 (arm_cmpsi_insn): Likewise.
14941 (cmpsi_shiftsi): Likewise.
14942 (cmpsi_shiftsi_swp): Likewise.
14943 (arm_cmpsi_negshiftsi_si): Likewise.
14944 (movsicc_insn): Likewise.
14945 (movsfcc_soft_insn): Likewise.
14946 (arith_shiftsi): Likewise.
14947 (arith_shiftsi_compare0
14948 (arith_shiftsi_compare0_scratch
14949 (sub_shiftsi): Likewise.
14950 (sub_shiftsi_compare0
14951 (sub_shiftsi_compare0_scratch
14952 (and_scc): Likewise.
14953 (cond_move): Likewise.
14954 (if_plus_move): Likewise.
14955 (if_move_plus): Likewise.
14956 (if_move_not): Likewise.
14957 (if_not_move): Likewise.
14958 (if_shift_move): Likewise.
14959 (if_move_shift): Likewise.
14960 (if_shift_shift): Likewise.
14961 (if_not_arith): Likewise.
14962 (if_arith_not): Likewise.
14963 (cond_move_not): Likewise.
14964 (thumb1_ashlsi3): Set type attribute.
14965 (thumb1_ashrsi3): Likewise.
14966 (thumb1_lshrsi3): Likewise.
14967 (thumb1_rotrsi3): Likewise.
14968 (shiftsi3_compare0_scratch): Likewise.
14969 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
14970 (neon_mov<mode>): Likewise.
14971 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
14973 (thumb2_movsi_insn): Likewise.
14974 (thumb2_cmpsi_neg_shiftsi): Likewise.
14975 (thumb2_extendqisi_v6): Likewise.
14976 (thumb2_zero_extendhisi2_v6): Likewise.
14977 (thumb2_zero_extendqisi2_v6): Likewise.
14978 (thumb2_shiftsi3_short): Likewise.
14979 (thumb2_addsi3_compare0_scratch): Likewise.
14980 (orsi_not_shiftsi_si): Likewise.
14981 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
14982 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
14984 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
14985 (1020alu_shift_op): Likewise.
14986 (1020alu_shift_reg_op): Likewise.
14987 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
14988 (alu_shift_op): Likewise.
14989 (alu_shift_reg_op): Likewise.
14990 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
14991 (11_alu_shift_op): Likewise.
14992 (11_alu_shift_reg_op): Likewise.
14993 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
14994 (9_alu_shift_reg_op): Likewise.
14995 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
14997 (cortex_a15_alu_shift): Likewise.
14998 (cortex_a15_alu_shift_reg): Likewise.
14999 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
15001 (cortex_a5_alu_shift): Likewise.
15002 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
15004 (cortex_a53_alu_shift): Likewise.
15005 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15007 (cortex_a7_alu_reg): Likewise.
15008 (cortex_a7_alu_shift): Likewise.
15009 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
15011 (cortex_a8_alu_shift): Likewise.
15012 (cortex_a8_alu_shift_reg): Likewise.
15013 (cortex_a8_mov): Likewise.
15014 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
15015 (cortex_a9_dp_shift): Likewise.
15016 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
15018 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
15020 (cortex_r4_mov): Likewise.
15021 (cortex_r4_alu_shift): Likewise.
15022 (cortex_r4_alu_shift_reg): Likewise.
15023 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
15024 (526_alu_shift_op): Likewise.
15025 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
15026 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
15027 (626te_alu_shift_op): Likewise.
15028 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
15029 (726te_alu_op): Likewise.
15030 (726te_alu_shift_op): Likewise.
15031 (726te_alu_shift_reg_op): Likewise.
15032 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
15033 (mp626_alu_shift_op): Likewise.
15034 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
15035 (pj4_alu_e1_conds): Likewise.
15036 (pj4_alu): Likewise.
15037 (pj4_alu_conds): Likewise.
15038 (pj4_shift): Likewise.
15039 (pj4_shift_conds): Likewise.
15040 (pj4_alu_shift): Likewise.
15041 (pj4_alu_shift_conds): Likewise.
15042 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15044 (cortexa7_older_only): Likewise.
15045 (cortexa7_younger): Likewise.
15047 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15049 * ipa-pure-const.c (generate_summary): Rename to...
15050 (pure_const_generate_summary): ... this.
15052 2013-07-17 Iain Sandoe <iain@codesourcery.com>
15054 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
15056 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
15059 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
15062 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15064 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
15065 enabled without -march=zEC12.
15066 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
15069 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15071 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
15072 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
15073 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
15074 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
15075 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
15076 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
15077 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
15079 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
15080 and ISA_HAS_NMADD3_NMSUB3.
15081 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
15082 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
15083 (nmsub4<mode>, nmsub3<mode>): Likewise.
15084 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
15086 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15088 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
15089 TARGET_MIPS5400 checking.
15091 2013-07-16 Jakub Jelinek <jakub@redhat.com>
15092 Peter Bergner <bergner@vnet.ibm.com>
15094 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
15095 registers in the comment.
15096 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
15097 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
15098 rather than FIRST_PSEUDO_REGISTERS.
15100 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
15102 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
15103 enable extra ISA flags with TARGET_HTM.
15105 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15107 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
15110 2013-07-15 Cong Hou <congh@google.com>
15112 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
15113 in compare function for sorting.
15115 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
15117 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
15118 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
15119 * config/rs6000/rs6000.opt: Add -mhtm option.
15120 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
15121 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
15122 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
15123 __HTM__ if the HTM instructions are available.
15124 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
15125 htm_spr_reg_operand): New define_predicates.
15126 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
15127 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
15129 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
15130 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
15131 HTM builtin functions.
15132 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
15133 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
15134 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
15135 (rs6000_builtin_mask_calculate): Likewise.
15136 (rs6000_option_override_internal): Likewise.
15137 (bdesc_htm): Add new HTM builtin support.
15138 (htm_spr_num): New function.
15139 (htm_spr_regno): Likewise.
15140 (rs6000_htm_spr_icode): Likewise.
15141 (htm_expand_builtin): Likewise.
15142 (htm_init_builtins): Likewise.
15143 (rs6000_expand_builtin): Add support for HTM builtin functions.
15144 (rs6000_init_builtins): Likewise.
15145 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
15147 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
15148 (TARGET_HTM, MASK_HTM): Define macros.
15149 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
15150 (FIXED_REGISTERS): Likewise.
15151 (CALL_USED_REGISTERS): Likewise.
15152 (CALL_REALLY_USED_REGISTERS): Likewise.
15153 (REG_ALLOC_ORDER): Likewise.
15154 (enum reg_class): Likewise.
15155 (REG_CLASS_NAMES): Likewise.
15156 (REG_CLASS_CONTENTS): Likewise.
15157 (REGISTER_NAMES): Likewise.
15158 (ADDITIONAL_REGISTER_NAMES): Likewise.
15159 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
15160 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
15161 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
15162 * config/rs6000/htm.md: New file.
15163 * config/rs6000/htmintrin.h: New file.
15164 * config/rs6000/htmxlintrin.h: New file.
15166 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
15168 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15169 Define SYMBOL_TINY_GOT, update comment.
15170 * config/aarch64/aarch64.c
15171 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
15172 (aarch64_expand_mov_immediate): Likewise.
15173 (aarch64_print_operand): Likewise.
15174 (aarch64_classify_symbol): Likewise.
15175 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
15176 (ldr_got_tiny): Define.
15178 2013-07-13 Tobias Grosser <tobias@grosser.es>
15180 PR tree-optimization/54094
15181 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
15182 scheduling dimension for the parallelism check from the polyhedral
15183 information in the AST.
15184 * graphite-dependences.c (carries_deps): Do not assume the schedule is
15187 2013-07-13 Jason Merrill <jason@redhat.com>
15189 * print-tree.c (debug_vec_tree): Use debug_raw.
15190 (debug_raw (vec<tree, va_gc> &)): New.
15191 (debug_raw (vec<tree, va_gc> *)): New.
15192 * tree.h: Declare them.
15194 2013-07-13 Bin Cheng <bin.cheng@arm.com>
15196 * ifcvt.c (ifcvt_after_combine): New static variable.
15197 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
15199 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
15200 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
15201 rest_of_handle_if_after_reload): Pass new argument for if_convert.
15203 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
15205 * config/mips/mips.c (mips_expand_call): Remove empty statement.
15207 2013-07-12 Michael Matz <matz@suse.de>
15209 PR middle-end/55771
15210 * convert.c (convert_to_real): Reject non-float inner types.
15212 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
15214 * config/aarch64/aarch64-protos.h
15215 (aarch64_simd_immediate_valid_for_move): Remove.
15216 * config/aarch64/aarch64.c (simd_immediate_info): New member.
15217 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
15219 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
15221 2013-07-11 Steve Ellcey <sellcey@mips.com>
15223 * config/mips/mips.c (mips_conditional_register_usage): Do not
15224 use t[0-7] registers in MIPS16 mode when optimizing for size.
15226 2013-07-11 Sriraman Tallam <tmsriram@google.com>
15228 * config/i386/i386.c (dispatch_function_versions): Fix array
15229 indexing of function_version_info to match actual_versions.
15231 2013-07-11 Teresa Johnson <tejohnson@google.com>
15233 * vec.h (struct va_gc): Move release out-of-line.
15234 (va_gc::release): Call ggc_free on released vec.
15236 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15238 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
15239 Require GOT register as additional operand in UNSPEC.
15240 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
15241 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15242 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15243 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
15244 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
15245 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
15246 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
15248 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
15251 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
15252 name seen by assembler/linker rather if available.
15254 2013-07-11 Andreas Schwab <schwab@suse.de>
15256 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
15258 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
15260 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
15262 2013-07-10 Joseph Myers <joseph@codesourcery.com>
15264 * doc/tm.texi.in: Move hook documentation to ....
15265 * target.def: ... here.
15267 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
15268 text on @hook line.
15269 * doc/tm.texi: Regenerate.
15271 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
15274 * doc/invoke.texi: Document -Wconditionally-supported.
15276 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15279 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
15282 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15285 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
15286 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
15287 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
15288 Remove duplicate devices.
15289 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
15290 * config/avr/t-multilib: Regenerate.
15291 * config/avr/avr-tables.opt: Regenerate.
15292 * doc/avr-mmcu.texi: Regenerate.
15294 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15297 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
15299 2013-07-10 Graham Stott <graham.stott@btinternet.com>
15301 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
15302 the cost of MULT when optimizing for size.
15304 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
15306 * config/cr16/cr16-protos.h: Don't include target.h.
15308 2013-07-09 Joseph Myers <joseph@codesourcery.com>
15310 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
15311 adjust register size for TDmode and TFmode for VSX registers.
15313 2013-07-08 Kai Tietz <ktietz@redhat.com>
15316 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
15317 hook_bool_const_tree_true.
15319 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15321 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
15322 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
15323 * config/s390/s390.md: Define FPR*_REGNUM constants.
15324 Fix FPR2_REGNUM constant (18 -> 17).
15325 ("*trunc<BFP:mode><DFP_ALL:mode>2")
15326 ("*trunc<DFP_ALL:mode><BFP:mode>2")
15327 ("trunc<BFP:mode><DFP_ALL:mode>2")
15328 ("trunc<DFP_ALL:mode><BFP:mode>2")
15329 ("*extend<BFP:mode><DFP_ALL:mode>2")
15330 ("*extend<DFP_ALL:mode><BFP:mode>2")
15331 ("extend<BFP:mode><DFP_ALL:mode>2")
15332 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
15335 2013-07-08 Graham Stott <graham.stott@btinternet.com>
15337 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
15339 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15341 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
15342 and cfun_fpr_bit_p to cfun_fpr_save_p.
15343 (s390_frame_area, s390_register_info, s390_frame_info)
15344 (s390_emit_prologue, s390_emit_epilogue)
15345 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
15347 * config/s390/s390.h: Define *_REGNUM macros for floating point
15350 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
15352 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
15354 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
15356 PR rtl-optimization/57786
15357 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
15358 and break out of the loop when it is set to false.
15360 2013-07-08 Jakub Jelinek <jakub@redhat.com>
15363 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
15364 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
15365 (const_int 63)) 0)).
15366 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
15367 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
15368 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
15370 PR rtl-optimization/57829
15371 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
15372 mask bits outside of mode are just sign-extension from mode to HWI.
15374 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
15376 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
15377 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
15378 adjust_address instead of change_address to keep info about alignment.
15379 (emit_strmov): Remove.
15380 (emit_memmov): New function.
15381 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
15382 (expand_movmem_epilogue): Likewise and return updated rtx for
15384 (expand_constant_movmem_prologue): Likewise and return updated rtx for
15385 destination and source.
15386 (decide_alignment): Refactor, handle vector_loop.
15387 (ix86_expand_movmem): Likewise.
15388 (ix86_expand_setmem): Likewise.
15389 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
15391 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
15393 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
15394 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
15396 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15398 * config/i386/sse.md (sse_movlhps): Change alternative 3
15399 of operand 2 to "m".
15401 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15404 * config/i386/sse.md (iptr): New mode attribute.
15405 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
15406 (<sse>_vm<plusminus_insn><mode>3): Ditto.
15407 (<sse>_vmmul<mode>3): Ditto.
15408 (<sse>_vmdiv<mode>3): Ditto.
15409 (sse_vmrcpv4sf2): Ditto.
15410 (<sse>_vmsqrt<mode>2): Ditto.
15411 (sse_vmrsqrtv4sf2): Ditto.
15412 (<sse>_vm<code><mode>3): Ditto.
15413 (avx_vmcmp<mode>3): Ditto.
15414 (<sse>_vmmaskcmp<mode>3): Ditto.
15415 (<sse>_comi): Ditto.
15416 (<sse>_ucomi): Ditto.
15417 (*xop_vmfrcz_<mode>): Ditto.
15418 (*fmai_fmadd_<mode>): Ditto.
15419 (*fmai_fmsub_<mode>): Ditto.
15420 (*fmai_fnmadd_<mode>): Ditto.
15421 (*fmai_fnmsub_<mode>): Ditto.
15422 (*fma4i_vmfmadd_<mode>): Ditto.
15423 (*fma4i_vmfmsub_<mode>): Ditto.
15424 (*fma4i_vmfnmadd_<mode>): Ditto.
15425 (*fma4i_vmfnmsub_<mode>): Ditto.
15426 (*xop_vmfrcz_<mode>): Ditto.
15427 (sse_cvtps2pi): Ditto.
15428 (sse_cvttps2pi): Ditto.
15429 (sse_cvtss2si): Ditto.
15430 (sse_cvtss2si_2): Ditto.
15431 (sse_cvtss2siq_2): Ditto.
15432 (sse_cvttss2si): Ditto.
15433 (sse_cvttss2siq): Ditto.
15434 (sse_cvtsd2si): Ditto.
15435 (sse_cvtsd2si_2): Ditto.
15436 (sse_cvtsd2siq_2): Ditto.
15437 (sse_cvttsd2si): Ditto.
15438 (sse_cvttsd2siq): Ditto.
15439 (sse_cvtsd2ss): Ditto.
15440 (sse_cvtss2sd): Ditto.
15441 (avx2_pbroadcast<mode>): Ditto.
15442 (avx2_pbroadcast<mode>_1): Ditto.
15443 (*avx_vperm_broadcast_v4sf): Ditto.
15445 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
15446 (sse_movlhps): Ditto.
15447 (sse_storehps): Ditto.
15448 (sse_loadhps): Ditto.
15449 (sse_storelps): Ditto.
15450 (sse_loadlps): Ditto.
15451 (*vec_concatv4sf): Ditto.
15452 (*vec_interleave_highv2df): Ditto.
15453 (*vec_interleave_lowv2df): Ditto.
15454 (*vec_extractv2df_1_sse): Ditto.
15455 (*vec_extractv2df_0_sse): Ditto.
15456 (sse2_storelpd): Ditto.
15457 (sse2_loadlpd): Ditto.
15458 (sse2_movsd): Ditto.
15459 (*vec_concatv4si): Ditto.
15460 (vec_concatv2di): Ditto.
15462 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
15463 for Intel asm dialect.
15464 (mmx_punpcklwd): Ditto.
15465 (mmx_punpckldq): Ditto.
15467 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
15468 for intel assembler dialect.
15470 2013-07-06 Jakub Jelinek <jakub@redhat.com>
15473 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
15474 for BUILT_IN_C{LZ,LRSB}*.
15475 * tree.h (CASE_INT_FN): Add FN##IMAX case.
15476 * tree-vrp.c (extract_range_basic): Handle
15477 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
15478 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
15479 fall thru to code calling set_value*.
15480 * builtins.c (expand_builtin): Remove *IMAX cases.
15481 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
15482 if width is bigger than 2*HWI.
15484 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
15486 PR rtl-optimization/55342
15487 * lra-int.h (lra_subreg_reload_pseudos): New.
15488 * lra.c: Add undoing optional reloads to the block diagram.
15489 (lra_subreg_reload_pseudos): New.
15490 (lra_optional_reload_pseudos): Change comments.
15491 (lra): Init and clear lra_subreg_reload_pseudos. Clear
15492 lra_optional_reload_pseudos after undo transformations.
15493 * lra-assigns.c (pseudo_prefix_title): New.
15494 (lra_setup_reg_renumber): Use it.
15495 (spill_for): Ditto. Check subreg reload pseudos too.
15496 (assign_by_spills): Consider subreg reload pseudos too.
15497 * lra-constraints.c (simplify_operand_subreg): Use
15498 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
15499 (curr_insn_transform): Recognize and do optional reloads.
15500 (undo_optional_reloads): New.
15501 (lra_undo_inheritance): Call undo_optional_reloads.
15503 2013-07-05 Thomas Quinot <quinot@adacore.com>
15505 * tree-complex.c (expand_complex_operations_1): Fix typo.
15507 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
15509 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
15510 (tune_params): New member 'const vec_costs'.
15511 * config/aarch64/aarch64.c (generic_vector_cost): New.
15512 (generic_tunings): New member 'generic_vector_cost'.
15513 (aarch64_builtin_vectorization_cost): New.
15514 (aarch64_add_stmt_cost): New.
15515 (TARGET_VECTORIZE_ADD_STMT_COST): New.
15516 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
15518 2013-07-03 Jakub Jelinek <jakub@redhat.com>
15521 * config/i386/predicates.md (vsib_address_operand): Disallow
15522 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
15524 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
15526 PR middle-end/55030
15527 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
15528 expand_builtin_setjmp_receiver.
15529 (expand_label): Adjust, call expand_builtin_setjmp_receiver
15530 with NULL for the label parameter.
15531 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
15532 the frame-pointer. Adjust comments.
15533 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
15534 only if LABEL is non-NULL.
15536 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
15538 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
15539 (ARM_ABI_AAPCS64): Ditto.
15541 (ARM_DEFAULT_ABI): Ditto.
15543 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
15545 * config/aarch64/aarch64-builtins.c
15546 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
15547 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
15549 * config/aarch64/aarch64-simd.md
15550 (aarch64_ld1<VALL:mode>): New.
15551 (aarch64_st1<VALL:mode>): Likewise.
15552 * config/aarch64/arm_neon.h
15553 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
15555 2013-07-02 Sriraman Tallam <tmsriram@google.com>
15557 * config/i386/i386.c (gate_insert_vzeroupper): Check if
15559 (ix86_option_override_internal):Turn on all -mavx target flags by
15560 default as they are dependent on AVX anyway.
15562 2013-07-02 Cary Coutant <ccoutant@google.com>
15564 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
15565 deterministic hash.
15566 (loc_checksum_ordered): Likewise.
15567 (hash_loc_operands): Remove inline keyword.
15569 2013-07-02 Jakub Jelinek <jakub@redhat.com>
15571 PR tree-optimization/57741
15572 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
15573 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
15574 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
15575 Allow REAL_CST step_exprs if flag_associative_math.
15576 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
15578 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15580 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
15582 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15584 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
15586 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15588 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
15590 (iorsi3_insn): Likewise.
15591 (arm_xorsi3): Likewise.
15593 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
15595 * arm.md (attribute "wtype"): Delete. Move attribute values from here
15597 (attribute "type"): ... here, and prefix with "wmmx_".
15598 (attribute "core_cycles"): Update for attribute changes.
15599 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
15600 (tbcstv4hi): Likewise.
15601 (tbcstv2si): Likewise.
15602 (iwmmxt_iordi3): Likewise.
15603 (iwmmxt_xordi3): Likewise.
15604 (iwmmxt_anddi3): Likewise.
15605 (iwmmxt_nanddi3): Likewise.
15606 (iwmmxt_arm_movdi): Likewise.
15607 (iwmmxt_movsi_insn): Likewise.
15608 (mov<mode>_internal): Likewise.
15609 (and<mode>3_iwmmxt): Likewise.
15610 (ior<mode>3_iwmmxt): Likewise.
15611 (xor<mode>3_iwmmxt): Likewise.
15612 (add<mode>3_iwmmxt): Likewise.
15613 (ssaddv8qi3): Likewise.
15614 (ssaddv4hi3): Likewise.
15615 (ssaddv2si3): Likewise.
15616 (usaddv8qi3): Likewise.
15617 (usaddv4hi3): Likewise.
15618 (usaddv2si3): Likewise.
15619 (sub<mode>3_iwmmxt): Likewise.
15620 (sssubv8qi3): Likewise.
15621 (sssubv4hi3): Likewise.
15622 (sssubv2si3): Likewise.
15623 (ussubv8qi3): Likewise.
15624 (ussubv4hi3): Likewise.
15625 (ussubv2si3): Likewise.
15626 (mulv4hi3_iwmmxt): Likewise.
15627 (smulv4hi3_highpart): Likewise.
15628 (umulv4hi3_highpart): Likewise.
15629 (iwmmxt_wmacs): Likewise.
15630 (iwmmxt_wmacsz): Likewise.
15631 (iwmmxt_wmacu): Likewise.
15632 (iwmmxt_wmacuz): Likewise.
15633 (iwmmxt_clrdi): Likewise.
15634 (iwmmxt_clrv8qi): Likewise.
15635 (iwmmxt_clr4hi): Likewise.
15636 (iwmmxt_clr2si): Likewise.
15637 (iwmmxt_uavgrndv8qi3): Likewise.
15638 (iwmmxt_uavgrndv4hi3): Likewise.
15639 (iwmmxt_uavgv8qi3): Likewise.
15640 (iwmmxt_uavgv4hi3): Likewise.
15641 (iwmmxt_tinsrb): Likewise.
15642 (iwmmxt_tinsrh): Likewise.
15643 (iwmmxt_tinsrw): Likewise.
15644 (iwmmxt_textrmub): Likewise.
15645 (iwmmxt_textrmsb): Likewise.
15646 (iwmmxt_textrmuh): Likewise.
15647 (iwmmxt_textrmsh): Likewise.
15648 (iwmmxt_textrmw): Likewise.
15649 (iwmxxt_wshufh): Likewise.
15650 (eqv8qi3): Likewise.
15651 (eqv4hi3): Likewise.
15652 (eqv2si3): Likewise.
15653 (gtuv8qi3): Likewise.
15654 (gtuv4hi3): Likewise.
15655 (gtuv2si3): Likewise.
15656 (gtv8qi3): Likewise.
15657 (gtv4hi3): Likewise.
15658 (gtv2si3): Likewise.
15659 (smax<mode>3_iwmmxt): Likewise.
15660 (umax<mode>3_iwmmxt): Likewise.
15661 (smin<mode>3_iwmmxt): Likewise.
15662 (umin<mode>3_iwmmxt): Likewise.
15663 (iwmmxt_wpackhss): Likewise.
15664 (iwmmxt_wpackwss): Likewise.
15665 (iwmmxt_wpackdss): Likewise.
15666 (iwmmxt_wpackhus): Likewise.
15667 (iwmmxt_wpackwus): Likewise.
15668 (iwmmxt_wpackdus): Likewise.
15669 (iwmmxt_wunpckihb): Likewise.
15670 (iwmmxt_wunpckihh): Likewise.
15671 (iwmmxt_wunpckihw): Likewise.
15672 (iwmmxt_wunpckilb): Likewise.
15673 (iwmmxt_wunpckilh): Likewise.
15674 (iwmmxt_wunpckilw): Likewise.
15675 (iwmmxt_wunpckehub): Likewise.
15676 (iwmmxt_wunpckehuh): Likewise.
15677 (iwmmxt_wunpckehuw): Likewise.
15678 (iwmmxt_wunpckehsb): Likewise.
15679 (iwmmxt_wunpckehsh): Likewise.
15680 (iwmmxt_wunpckehsw): Likewise.
15681 (iwmmxt_wunpckelub): Likewise.
15682 (iwmmxt_wunpckeluh): Likewise.
15683 (iwmmxt_wunpckeluw): Likewise.
15684 (iwmmxt_wunpckelsb): Likewise.
15685 (iwmmxt_wunpckelsh): Likewise.
15686 (iwmmxt_wunpckelsw): Likewise.
15687 (ror<mode>3): Likewise.
15688 (ashr<mode>3_iwmmxt): Likewise.
15689 (lshr<mode>3_iwmmxt): Likewise.
15690 (ashl<mode>3_iwmmxt): Likewise.
15691 (ror<mode>3_di): Likewise.
15692 (ashr<mode>3_di): Likewise.
15693 (lshr<mode>3_di): Likewise.
15694 (ashl<mode>3_di): Likewise.
15695 (iwmmxt_wmadds): Likewise.
15696 (iwmmxt_wmaddu): Likewise.
15697 (iwmmxt_tmia): Likewise.
15698 (iwmmxt_tmiaph): Likewise.
15699 (iwmmxt_tmiabb): Likewise.
15700 (iwmmxt_tmiatb): Likewise.
15701 (iwmmxt_tmiabt): Likewise.
15702 (iwmmxt_tmiatt): Likewise.
15703 (iwmmxt_tmovmskb): Likewise.
15704 (iwmmxt_tmovmskh): Likewise.
15705 (iwmmxt_tmovmskw): Likewise.
15706 (iwmmxt_waccb): Likewise.
15707 (iwmmxt_wacch): Likewise.
15708 (iwmmxt_waccw): Likewise.
15709 (iwmmxt_waligni): Likewise.
15710 (iwmmxt_walignr): Likewise.
15711 (iwmmxt_walignr0): Likewise.
15712 (iwmmxt_walignr1): Likewise.
15713 (iwmmxt_walignr2): Likewise.
15714 (iwmmxt_walignr3): Likewise.
15715 (iwmmxt_wsadb): Likewise.
15716 (iwmmxt_wsadh): Likewise.
15717 (iwmmxt_wsadbz): Likewise.
15718 (iwmmxt_wsadhz): Likewise.
15719 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
15720 (iwmmxt_wabsdiffb): Likewise.
15721 (iwmmxt_wabsdiffh): Likewise.
15722 (iwmmxt_wabsdiffw): Likewise.
15723 (iwmmxt_waddsubhx): Likewise
15724 (iwmmxt_wsubaddhx): Likewise.
15725 (addc<mode>3): Likewise.
15726 (iwmmxt_avg4): Likewise.
15727 (iwmmxt_avg4r): Likewise.
15728 (iwmmxt_wmaddsx): Likewise.
15729 (iwmmxt_wmaddux): Likewise.
15730 (iwmmxt_wmaddsn): Likewise.
15731 (iwmmxt_wmaddun): Likewise.
15732 (iwmmxt_wmulwsm): Likewise.
15733 (iwmmxt_wmulwum): Likewise.
15734 (iwmmxt_wmulsmr): Likewise.
15735 (iwmmxt_wmulumr): Likewise.
15736 (iwmmxt_wmulwsmr): Likewise.
15737 (iwmmxt_wmulwumr): Likewise.
15738 (iwmmxt_wmulwl): Likewise.
15739 (iwmmxt_wqmulm): Likewise.
15740 (iwmmxt_wqmulwm): Likewise.
15741 (iwmmxt_wqmulmr): Likewise.
15742 (iwmmxt_wqmulwmr): Likewise.
15743 (iwmmxt_waddbhusm): Likewise.
15744 (iwmmxt_waddbhusl): Likewise.
15745 (iwmmxt_wqmiabb): Likewise.
15746 (iwmmxt_wqmiabt): Likewise.
15747 (iwmmxt_wqmiatb): Likewise.
15748 (iwmmxt_wqmiatt): Likewise.
15749 (iwmmxt_wqmiabbn): Likewise.
15750 (iwmmxt_wqmiabtn): Likewise.
15751 (iwmmxt_wqmiatbn): Likewise.
15752 (iwmmxt_wqmiattn): Likewise.
15753 (iwmmxt_wmiabb): Likewise.
15754 (iwmmxt_wmiabt): Likewise.
15755 (iwmmxt_wmiatb): Likewise.
15756 (iwmmxt_wmiatt): Likewise.
15757 (iwmmxt_wmiabbn): Likewise.
15758 (iwmmxt_wmiabtn): Likewise.
15759 (iwmmxt_wmiatbn): Likewise.
15760 (iwmmxt_wmiattn): Likewise.
15761 (iwmmxt_wmiawbb): Likewise.
15762 (iwmmxt_wmiawbt): Likewise.
15763 (iwmmxt_wmiawtb): Likewise.
15764 (iwmmxt_wmiawtt): Likewise.
15765 (iwmmxt_wmiawbbn): Likewise.
15766 (iwmmxt_wmiawbtn): Likewise.
15767 (iwmmxt_wmiawtbn): Likewise.
15768 (iwmmxt_wmiawttn): Likewise.
15769 (iwmmxt_wmerge): Likewise.
15770 (iwmmxt_tandc<mode>3): Likewise.
15771 (iwmmxt_torc<mode>3): Likewise.
15772 (iwmmxt_torvsc<mode>3): Likewise.
15773 (iwmmxt_textrc<mode>3): Likewise.
15774 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
15775 (wmmxt_pack): Likewise.
15776 (wmmxt_mult_c1): Likewise.
15777 (wmmxt_mult_c2): Likewise.
15778 (wmmxt_alu_c1): Likewise.
15779 (wmmxt_alu_c2): Likewise.
15780 (wmmxt_alu_c3): Likewise.
15781 (wmmxt_transfer_c1): Likewise.
15782 (wmmxt_transfer_c2): Likewise.
15783 (wmmxt_transfer_c3): Likewise.
15784 (marvell_f_iwmmxt_wstr): Likewise.
15785 (marvell_f_iwmmxt_wldr): Likewise.
15787 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
15789 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
15791 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15794 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15795 * lra-constraints.c (need_for_split_p): Check call used hard regs
15796 living through calls.
15798 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
15799 call used regs for call insn.
15801 2013-06-28 Jakub Jelinek <jakub@redhat.com>
15804 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
15805 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
15807 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15809 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
15811 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15813 * lra-constraints.c (need_for_split_p): Check call used hard regs
15814 living through calls.
15816 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
15819 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
15820 to tie with any other modes. Eliminate Altivec vector mode tests,
15821 since these are a subset of ALTIVEC or VSX vector modes. Simplify
15822 code, to return 0 if testing MODE2 for a condition, if we've
15823 already tested MODE1 for the same condition.
15825 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
15827 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
15830 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
15832 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15833 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
15835 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
15837 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
15839 (aarch64_symbolic_constant_p): Remove.
15840 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
15841 static. Fix line length and white space.
15842 (aarch64_symbolic_constant_p): Remove.
15843 * config/aarch64/predicates.md (aarch64_valid_symref):
15844 Use aarch64_classify_symbol_expression.
15846 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15848 * config/arm/constraints.md (Ts): New constraint.
15849 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
15851 (compare_scc): Use "Ts" constraint for operand 0.
15852 (ior_scc_scc): Likewise.
15853 (and_scc_scc): Likewise.
15854 (and_scc_scc_nodom): Likewise.
15855 (ior_scc_scc_cmp): Likewise for operand 7.
15856 (and_scc_scc_cmp): Likewise.
15857 * config/arm/thumb2.md (thumb2_movsi_insn):
15858 Add alternatives for 16-bit encodings.
15859 (thumb2_movhi_insn): Likewise.
15860 (thumb2_movsicc_insn): Likewise.
15861 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
15862 (thumb2_negscc): Use "Ts" constraint.
15863 Move mvn instruction outside cond_exec block.
15864 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
15865 for 16-bit encodings.
15867 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15869 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
15871 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
15872 (mulsi3subsi): Likewise.
15873 (mulsidi3adddi): Likewise.
15874 (mulsidi3_v6): Likewise.
15875 (umulsidi3_v6): Likewise.
15876 (umulsidi3adddi_v6): Likewise.
15877 (smulsi3_highpart_v6): Likewise.
15878 (umulsi3_highpart_v6): Likewise.
15879 (mulhisi3tb): Likewise.
15880 (mulhisi3bt): Likewise.
15881 (mulhisi3tt): Likewise.
15882 (maddhisi4): Likewise.
15883 (maddhisi4tb): Likewise.
15884 (maddhisi4tt): Likewise.
15885 (maddhidi4): Likewise.
15886 (maddhidi4tb): Likewise.
15887 (maddhidi4tt): Likewise.
15888 (zeroextractsi_compare0_scratch): Likewise.
15889 (insv_zero): Likewise.
15890 (insv_t2): Likewise.
15891 (anddi_notzesidi_di): Likewise.
15892 (anddi_notsesidi_di): Likewise.
15893 (andsi_notsi_si): Likewise.
15894 (iordi_zesidi_di): Likewise.
15895 (xordi_zesidi_di): Likewise.
15896 (andsi_iorsi3_notsi): Likewise.
15897 (smax_0): Likewise.
15898 (smax_m1): Likewise.
15899 (smin_0): Likewise.
15900 (not_shiftsi): Likewise.
15901 (unaligned_loadsi): Likewise.
15902 (unaligned_loadhis): Likewise.
15903 (unaligned_loadhiu): Likewise.
15904 (unaligned_storesi): Likewise.
15905 (unaligned_storehi): Likewise.
15906 (extv_reg): Likewise.
15907 (extzv_t2): Likewise.
15908 (divsi3): Likewise.
15909 (udivsi3): Likewise.
15910 (arm_zero_extendhisi2addsi): Likewise.
15911 (arm_zero_extendqisi2addsi): Likewise.
15912 (compareqi_eq0): Likewise.
15913 (arm_extendhisi2_v6): Likewise.
15914 (arm_extendqisi2addsi): Likewise.
15915 (arm_movt): Likewise.
15916 (thumb2_ldrd): Likewise.
15917 (thumb2_ldrd_base): Likewise.
15918 (thumb2_ldrd_base_neg): Likewise.
15919 (thumb2_strd): Likewise.
15920 (thumb2_strd_base): Likewise.
15921 (thumb2_strd_base_neg): Likewise.
15922 (arm_negsi2): Add alternative for 16-bit encoding.
15923 (arm_one_cmplsi2): Likewise.
15925 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15927 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
15928 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
15929 (movdfcc): Likewise.
15930 * config/arm/vfp.md (*thumb2_movsf_vfp):
15931 Disable predication for arm_restrict_it.
15932 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
15933 (*thumb2_movdfcc_vfp): Likewise.
15934 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
15935 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
15936 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
15937 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
15938 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
15939 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
15940 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
15941 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
15942 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
15943 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
15944 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
15945 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
15946 Disable predication for arm_restrict_it.
15948 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
15950 * config/i386/bmiintrin.h (_bextr_u32): New.
15951 (_bextr_u64): Ditto.
15953 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
15955 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
15956 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
15957 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
15958 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
15959 * config/mips/n32-elf.h: ...this new file.
15961 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
15964 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
15965 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
15967 2013-06-27 Catherine Moore <clm@codesourcery.com>
15969 * config/mips/mips-tables.opt: Regenerate.
15970 * config/mips/mips-cpus.def: Add m14ke and m14kec.
15971 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
15972 * doc/invoke.texi: Add -m14kc.
15974 2013-06-27 Jakub Jelinek <jakub@redhat.com>
15977 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
15978 constraints of operand 1 and 2.
15981 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
15982 to match RTL canonicalization. Swap predicates and
15983 constraints of operand 1 and 2.
15985 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
15987 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
15988 Process OP_INOUT regs for splitting too.
15990 2013-06-27 Jakub Jelinek <jakub@redhat.com>
15992 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
15993 decl before the loop, initialize to NULL.
15994 (vectorizable_load): Initialize ptr_incr to NULL.
15996 2013-06-27 Martin Jambor <mjambor@suse.cz>
15999 * ipa-ref.h (ipa_maybe_record_reference): Declare.
16000 * ipa-ref.c (ipa_maybe_record_reference): New function.
16001 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
16002 * ipa-cp.c (create_specialized_node): Record potential references from
16004 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
16006 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16008 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
16009 parameter 'mode' of type 'enum machine_mode mode'; change to pass
16010 'mode' to force_reg.
16011 (aarch64_add_offset): Update calls to aarch64_force_temporary.
16012 (aarch64_expand_mov_immediate): Likewise.
16014 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16016 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
16017 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
16019 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16021 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
16022 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16023 (struct machine_function): Add tbegin_p.
16024 (s390_canonicalize_comparison): Fold CC mode compares to
16025 conditional jump if possible.
16026 (s390_emit_jump): Return the emitted jump.
16027 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
16028 Handle CCRAWmode compares.
16029 (s390_option_override): Default to -mhtm if available.
16030 (s390_reg_clobbered_rtx): Handle floating point regs as well.
16031 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
16032 FPRs instead of df_regs_ever_live_p.
16033 (s390_optimize_nonescaping_tx): New function.
16034 (s390_init_frame_layout): Extend clobbered_regs array to cover
16036 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
16037 (s390_expand_tbegin): New function.
16038 (enum s390_builtin): New enum definition.
16039 (code_for_builtin): New array definition.
16040 (s390_init_builtins): New function.
16041 (s390_expand_builtin): New function.
16042 (TARGET_INIT_BUILTINS): Define.
16043 (TARGET_EXPAND_BUILTIN): Define.
16044 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
16045 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
16046 (s390_alc_comparison): Likewise.
16047 * config/s390/s390-modes.def: Add CCRAWmode.
16048 * config/s390/s390.h (processor_flags): Add PF_TX.
16049 (TARGET_CPU_HTM): Define macro.
16050 (TARGET_HTM): Define macro.
16051 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
16052 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
16053 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16054 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
16055 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
16056 (TBEGIN_MASK, TBEGINC_MASK): New constants.
16057 ("*cc_to_int"): Move up.
16058 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
16059 constants other than 0.
16060 ("*ccraw_to_int"): New insn and splitter definition.
16061 ("tbegin", "tbegin_nofloat", "tbegin_retry")
16062 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
16063 ("tx_assist"): New expander.
16064 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
16065 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
16066 * config/s390/s390.opt: Add -mhtm option.
16067 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
16068 * config/s390/htmxlintrin.h: New file.
16069 * config/s390/htmintrin.h: New file.
16070 * config/s390/s390intrin.h: New file.
16071 * doc/extend.texi: Document htm builtins.
16072 * config.gcc: Add the new header files to extra_headers.
16074 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
16076 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
16077 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
16079 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
16080 Pat Haugen <pthaugen@us.ibm.com>
16081 Peter Bergner <bergner@vnet.ibm.com>
16083 * config/rs6000/power8.md: New.
16084 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
16085 setting for power8 entry.
16086 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
16087 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
16088 test for Power4/Power5 only.
16089 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
16091 (force_new_group): Adjust comment.
16092 * config/rs6000/rs6000.md: Include power8.md.
16094 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
16096 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
16097 * config/arm/arm-protos.h (arm_max_conditional_execute): New
16099 (tune_params): Update comment.
16100 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
16101 (arm_max_conditional_execute): New function.
16102 (thumb2_final_prescan_insn): Use max_insn_skipped and
16103 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
16105 2013-06-25 Jakub Jelinek <jakub@redhat.com>
16107 PR tree-optimization/57705
16108 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
16109 SSA_NAME step, provided that it is not defined inside the loop.
16110 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
16111 (get_initial_def_for_induction): Handle SSA_NAME IV step.
16113 2013-06-25 Martin Jambor <mjambor@suse.cz>
16115 PR middle-end/57670
16116 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
16117 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
16119 (ipa_note_param_call): Initialize member_ptr flag.
16120 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
16121 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
16122 (ipa_write_indirect_edge_info): Stream member_ptr flag.
16123 (ipa_read_indirect_edge_info): Likewise.
16125 2013-06-25 Richard Biener <rguenther@suse.de>
16127 PR middle-end/56977
16128 * passes.c (init_optimization_passes): Move pass_fold_builtins
16129 and pass_dce earlier with -Og.
16131 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
16133 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
16134 <BIT_FIELD_REF>: Remove trailing TAB.
16135 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
16138 2013-06-24 Martin Jambor <mjambor@suse.cz>
16140 PR tree-optimization/57358
16141 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
16142 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
16143 (ipa_analyze_params_uses): Generate pessimistic info when true.
16145 2013-06-24 Martin Jambor <mjambor@suse.cz>
16147 PR tree-optimization/57539
16148 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
16149 global.inlined_to of the new node to it. All callers changed.
16150 * ipa-inline-transform.c (clone_inlined_nodes): New variable
16151 inlining_into, pass it to cgraph_clone_node.
16152 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
16153 ipa_free_edge_args_substructures.
16154 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
16155 rdesc linked list. Do not assert rdesc edges have inlined caller.
16156 Assert we have found an rdesc in the rdesc list.
16158 2013-06-24 Richard Biener <rguenther@suse.de>
16160 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
16161 (pointer_set_lookup): Declare.
16162 (class pointer_map): New template class implementing a
16163 generic pointer to T map.
16164 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
16165 pointer_map<T>::contains, pointer_map<T>::insert,
16166 pointer_map<T>::traverse): New functions.
16167 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
16168 (pointer_set_lookup): New function.
16169 (pointer_set_contains): Use pointer_set_lookup.
16170 (pointer_set_insert): Likewise.
16171 (insert_aux): Remove.
16172 (struct pointer_map_t): Embed a pointer_set_t.
16173 (pointer_map_create): Adjust.
16174 (pointer_map_destroy): Likewise.
16175 (pointer_map_contains): Likewise.
16176 (pointer_map_insert): Likewise.
16177 (pointer_map_traverse): Likewise.
16178 * tree-streamer.h (struct streamer_tree_cache_d): Use a
16179 pointer_map<unsigned> instead of a pointer_map_t.
16180 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
16181 (streamer_tree_cache_lookup): Likewise.
16182 (streamer_tree_cache_create): Likewise.
16183 (streamer_tree_cache_delete): Likewise.
16184 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
16185 pointer_map<unsigned> instead of a pointer_map_t.
16186 (lto_init_tree_ref_encoder): Adjust.
16187 (lto_destroy_tree_ref_encoder): Likewise.
16188 * lto-section-out.c (lto_output_decl_index): Likewise.
16189 (lto_record_function_out_decl_state): Likewise.
16190 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
16192 2013-06-24 Richard Biener <rguenther@suse.de>
16194 PR tree-optimization/57488
16195 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
16197 2013-06-24 Alan Modra <amodra@gmail.com>
16199 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
16200 (gen_easy_altivec_constant): Likewise.
16201 * config/rs6000/predicates.md (easy_vector_constant_add_self,
16202 easy_vector_constant_msb): Likewise.
16204 2013-06-23 Jakub Jelinek <jakub@redhat.com>
16207 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
16208 add missing return true.
16210 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
16213 * config/sh/predicates.md (general_extend_operand): Invoke
16214 general_movsrc_operand for memory operands.
16215 (general_movsrc_operand): Allow reg+reg addressing, do not use
16216 general_operand for memory operands.
16218 2013-06-23 Sriraman Tallam <tmsriram@google.com>
16220 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
16221 when current target options does not apply.
16222 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
16223 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
16224 * config/i386/bmiintrin.h: Pass appropriate target
16225 attributes to header.
16226 * config/i386/mmintrin.h: Ditto.
16227 * config/i386/nmmintrin.h: Ditto.
16228 * config/i386/avx2intrin.h: Ditto.
16229 * config/i386/fxsrintrin.h: Ditto.
16230 * config/i386/tbmintrin.h: Ditto.
16231 * config/i386/xsaveintrin.h: Ditto.
16232 * config/i386/f16cintrin.h: Ditto.
16233 * config/i386/xtestintrin.h: Ditto.
16234 * config/i386/xsaveoptintrin.h: Ditto.
16235 * config/i386/bmi2intrin.h: Ditto.
16236 * config/i386/lzcntintrin.h: Ditto.
16237 * config/i386/smmintrin.h: Ditto.
16238 * config/i386/wmmintrin.h: Ditto.
16239 * config/i386/x86intrin.h: Remove all header include guards.
16240 * config/i386/prfchwintrin.h: Ditto.
16241 * config/i386/pmmintrin.h: Ditto.
16242 * config/i386/tmmintrin.h: Ditto.
16243 * config/i386/xmmintrin.h: Ditto.
16244 * config/i386/popcntintrin.h: Ditto.
16245 * config/i386/rdseedintrin.h: Ditto.
16246 * config/i386/ammintrin.h: Ditto.
16247 * config/i386/emmintrin.h: Ditto.
16248 * config/i386/immintrin.h: Remove all header include guards.
16249 * config/i386/fma4intrin.h: Ditto.
16250 * config/i386/lwpintrin.h: Ditto.
16251 * config/i386/xopintrin.h: Ditto.
16252 * config/i386/ia32intrin.h: Ditto.
16253 * config/i386/avxintrin.h: Ditto.
16254 * config/i386/rtmintrin.h: Ditto.
16255 * config/i386/fmaintrin.h: Ditto.
16256 * config/i386/mm3dnow.h: Ditto.
16258 2013-06-22 Sriraman Tallam <tmsriram@google.com>
16260 * common/config/i386/i386-common.c: Handle LZCNT.
16262 2013-06-22 Andi Kleen <ak@linux.intel.com>
16264 * doc/extend.texi: Use __atomic_store_n instead of
16265 __atomic_store in HLE example.
16267 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
16269 * config/sh/sh.c: Remove <cstdlib> workaround.
16271 2013-06-21 Andi Kleen <ak@linux.intel.com>
16273 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
16275 2013-06-21 Andi Kleen <ak@linux.intel.com>
16277 * doc/extend.texi: Document that __atomic_clear and
16278 __atomic_test_and_set should only be used with bool.
16280 2013-06-20 Jan Hubicka <jh@suse.cz>
16282 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
16283 types_same_for_odr.
16284 * tree.c (decls_same_for_odr): New function.
16285 (same_for_edr): New function.
16286 (types_same_for_odr): New function.
16287 (get_binfo_at_offset): Use it.
16288 * tree.h (types_same_for_odr): Declare.
16290 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
16291 Jason Merrill <jason@redhat.com>
16293 * system.h: Include <cstdlib> as well as <stdlib.h>.
16295 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
16298 * config/i386/i386.c (construct_container): Report error if
16299 long double is used with disabled x87 float returns.
16301 2013-06-20 Jan Hubicka <jh@suse.cz>
16303 * lto-cgraph.c (input_symtab): Do not set cgraph state.
16305 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
16307 PR rtl-optimization/57425
16308 PR rtl-optimization/57569
16309 * alias.c (write_dependence_p): Remove parameters mem_mode and
16310 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
16311 Changed all callers.
16312 (canon_anti_dependence): Get comments and semantics in sync.
16313 Add parameter mem_canonicalized. Changed all callers.
16314 * rtl.h (canon_anti_dependence): Update prototype.
16316 2013-06-20 Richard Biener <rguenther@suse.de>
16318 * data-streamer-in.c (streamer_read_uhwi): Optimize single
16319 byte case, inline streamer_read_uchar and defer section
16322 2013-06-20 Richard Biener <rguenther@suse.de>
16324 PR tree-optimization/57584
16325 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
16326 SSA names into the expanded expression that take part in
16327 abnormal coalescing.
16329 2013-06-19 Sharad Singhai <singhai@google.com>
16331 * gcov.c (print_usage): Handle new option.
16332 (process_args): Ditto.
16333 (get_gcov_intermediate_filename): New function.
16334 (output_intermediate_file): New function.
16335 (output_gcov_file): New function
16336 (generate_results): Handle new option.
16337 (release_function): Relase demangled name.
16338 (read_graph_file): Handle demangled name.
16339 (output_lines): Ditto.
16340 * doc/gcov.texi: Document gcov intermediate format.
16342 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
16345 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
16346 (lra_emit_add): Use the functions. Add comment about Y as an
16349 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
16352 * collect2.c (collect_atexit): New.
16353 (collect_exit): Delete.
16354 (main): Register collect_atexit with atexit.
16355 (collect_wait): Change collect_exit to exit.
16357 * collect2.h (collect_exit): Delete.
16358 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
16360 2013-06-19 Wei Mi <wmi@google.com>
16362 PR rtl-optimization/57518
16363 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
16364 if regno is used in paradoxical subreg.
16365 (update_equiv_regs): Check pdx_subregs[regno] before
16366 set a reg to be equivalent with a mem.
16368 2013-06-19 Matthias Klose <doko@ubuntu.com>
16371 * file-find.h (find_a_file): Add a mode parameter.
16372 * file-find.c (find_a_file): Likewise.
16373 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
16374 with X_OK for the executables.
16375 * collect2.c (main): Call find_a_file with X_OK.
16377 2013-06-19 Steve Ellcey <sellcey@mips.com>
16380 * config/mips/mips.md (casesi_internal_mips16_<mode>):
16381 Use NEXT_INSN instead of next_real_insn.
16383 2013-06-19 Jan Hubicka <jh@suse.cz>
16385 * cgraph.h (const_value_known_p): Replace by ...
16386 (ctor_for_folding): .. this one.
16387 * cgraphunit.c (process_function_and_variable_attributes): Use it.
16388 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
16389 * expr.c (expand_expr_real_1): Likewise.
16390 (string_constant): Likewise.
16391 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
16392 * ipa.c (process_references): Likewise.
16393 (symtab_remove_unreachable_nodes): Likewise.
16394 * ipa-inline-analysis.c (param_change_prob): Likewise.
16395 * gimple-fold.c (canonicalize_constructor_val): Likewise.
16396 (get_base_constructor): Likwise.
16397 * varpool.c (varpool_remove_node): Likewise.
16398 (varpool_remove_initializer): LIkewise.
16399 (dump_varpool_node): LIkwise.
16400 (const_value_known_p): Rewrite to ...
16401 (ctor_for_folding): ... this one.
16403 2013-06-19 Jakub Jelinek <jakub@redhat.com>
16406 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
16407 PERSONALITY in $PATH derived prefixes.
16409 2013-06-19 Jeff Law <law@redhat.com>
16411 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
16414 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
16415 (simplify_bitwise_binary): Use it to simpify certain binary ops on
16418 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
16420 * config/arm/vfp.md: Move VFP instruction classification documentation
16422 * config/arm/arm.md: ... here. Update instruction classification
16425 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
16427 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
16428 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
16429 pattern. Use more efficient sequences on ARMv5 and Thumb2.
16431 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
16434 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16435 with NEXT_INSN. Use tablejump_p to check for jump table data
16438 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
16441 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
16442 that now in C++ the value is correct per the C++ standards.
16444 2013-06-19 Richard Biener <rguenther@suse.de>
16446 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
16447 for global context.
16449 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16452 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16455 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16456 with next_active_insn.
16458 2013-06-18 Sriraman Tallam <tmsriram@google.com>
16460 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
16461 functions are inlined during failures to flag an error.
16462 * tree-inline.c (expand_call_inline): Allow the error to be flagged
16463 in early inline pass.
16465 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
16467 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
16470 2013-06-18 Julian Brown <julian@codesourcery.com>
16472 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
16473 Permit virtual register pre-reload if !strict.
16474 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
16476 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
16478 * config/arm/neon.md (movmisalign<mode>): Use
16479 neon_perm_struct_or_reg_operand instead of
16480 neon_struct_or_register_operand.
16481 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
16482 neon_permissive_struct_operand instead of neon_struct_operand.
16483 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
16484 neon_vector_mem_operand.
16485 * config/arm/predicates.md (neon_struct_operand): Adjust call to
16486 neon_vector_mem_operand.
16487 (neon_permissive_struct_operand): New.
16488 (neon_struct_or_register_operand): Rename to...
16489 (neon_perm_struct_or_reg_operand): This. Adjust call to
16490 neon_vector_mem_operand.
16492 2013-06-18 Richard Biener <rguenther@suse.de>
16494 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
16495 * lto-streamer.h: Include pointer-set.h.
16496 (struct lto_decl_slot): Remove.
16497 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
16498 Remove next_index entry.
16499 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16500 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16501 (lto_init_tree_ref_encoder): Adjust.
16502 (lto_destroy_tree_ref_encoder): Likewise.
16503 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16504 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16505 (lto_output_decl_index): Adjust.
16506 (lto_new_out_decl_state): Likewise.
16507 (lto_record_function_out_decl_state): Likewise.
16508 * lto-streamer-out.c (copy_function): Likewise.
16510 2013-06-18 Richard Biener <rguenther@suse.de>
16512 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
16513 * cgraphunit.c: Include cfgloop.h.
16514 (init_lowered_empty_function): Initialize the loop tree.
16515 (assemble_thunk): Insert new BBs into loops.
16517 2013-06-18 Richard Biener <rguenther@suse.de>
16519 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
16520 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
16521 the map from cache entry to cache index optional.
16522 (streamer_tree_cache_replace_tree): Adjust accordingly.
16523 (streamer_tree_cache_append): Likewise.
16524 (streamer_tree_cache_delete): Likewise.
16525 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
16526 streamer cache map from cache entry to cache index.
16527 * lto-streamer-out.c (create_output_block): Adjust.
16529 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
16531 * config/arm/arm.md (attribute "insn"): Move multiplication and
16532 division attributes to...
16533 (attribute "type"): ... here. Remove mult.
16534 (attribute "mul32"): New attribute.
16535 (attribute "mul64"): Add umaal.
16536 (*arm_mulsi3): Update attributes.
16537 (*arm_mulsi3_v6): Likewise.
16538 (*thumb_mulsi3): Likewise.
16539 (*thumb_mulsi3_v6): Likewise.
16540 (*mulsi3_compare0): Likewise.
16541 (*mulsi3_compare0_v6): Likewise.
16542 (*mulsi_compare0_scratch): Likewise.
16543 (*mulsi_compare0_scratch_v6): Likewise.
16544 (*mulsi3addsi): Likewise.
16545 (*mulsi3addsi_v6): Likewise.
16546 (*mulsi3addsi_compare0): Likewise.
16547 (*mulsi3addsi_compare0_v6): Likewise.
16548 (*mulsi3addsi_compare0_scratch): Likewise.
16549 (*mulsi3addsi_compare0_scratch_v6): Likewise.
16550 (*mulsi3subsi): Likewise.
16551 (*mulsidi3adddi): Likewise.
16552 (*mulsi3addsi_v6): Likewise.
16553 (*mulsidi3adddi_v6): Likewise.
16554 (*mulsidi3_nov6): Likewise.
16555 (*mulsidi3_v6): Likewise.
16556 (*umulsidi3_nov6): Likewise.
16557 (*umulsidi3_v6): Likewise.
16558 (*umulsidi3adddi): Likewise.
16559 (*umulsidi3adddi_v6): Likewise.
16560 (*smulsi3_highpart_nov6): Likewise.
16561 (*smulsi3_highpart_v6): Likewise.
16562 (*umulsi3_highpart_nov6): Likewise.
16563 (*umulsi3_highpart_v6): Likewise.
16564 (mulhisi3): Likewise.
16565 (*mulhisi3tb): Likewise.
16566 (*mulhisi3bt): Likewise.
16567 (*mulhisi3tt): Likewise.
16568 (maddhisi4): Likewise.
16569 (*maddhisi4tb): Likewise.
16570 (*maddhisi4tt): Likewise.
16571 (maddhidi4): Likewise.
16572 (*maddhidi4tb): Likewise.
16573 (*maddhidi4tt): Likewise.
16574 (divsi3): Likewise.
16575 (udivsi3): Likewise.
16576 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
16577 (thumb2_mulsi_short_compare0): Likewise.
16578 (thumb2_mulsi_short_compare0_scratch): Likewise.
16579 * config/arm/arm1020e.md (1020mult1): Update attribute change.
16580 (1020mult2): Likewise.
16581 (1020mult3): Likewise.
16582 (1020mult4): Likewise.
16583 (1020mult5): Likewise.
16584 (1020mult6): Likewise.
16585 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
16587 (cortex_a15_mult64): Likewise.
16588 (cortex_a15_sdiv): Likewise.
16589 (cortex_a15_udiv): Likewise.
16590 * config/arm/arm1026ejs.md (mult1): Update attribute change.
16596 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
16597 (pj4_ir_div): Likewise.
16598 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
16599 (11_mult2): Likewise.
16600 (11_mult3): Likewise.
16601 (11_mult4): Likewise.
16602 (11_mult5): Likewise.
16603 (11_mult6): Likewise.
16604 (11_mult7): Likewise.
16605 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
16606 (cortex_a8_mla): Likewise.
16607 (cortex_a8_mull): Likewise.
16608 (cortex_a8_smulwy): Likewise.
16609 (cortex_a8_smlald): Likewise.
16610 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
16611 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
16612 (cortex_r4_mul_3): Likewise.
16613 (cortex_r4_mla_4): Likewise.
16614 (cortex_r4_mla_3): Likewise.
16615 (cortex_r4_smlald): Likewise.
16616 (cortex_r4_mull): Likewise.
16617 (cortex_r4_sdiv): Likewise.
16618 (cortex_r4_udiv): Likewise.
16619 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
16620 (cortex_a7_idiv): Likewise.
16621 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
16622 (9_mult2): Likewise.
16623 (9_mult3): Likewise.
16624 (9_mult4): Likewise.
16625 (9_mult5): Likewise.
16626 (9_mult6): Likewise.
16627 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
16628 (cortex_a53_sdiv): Likewise.
16629 (cortex_a53_udiv): Likewise.
16630 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
16631 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
16632 (mp626_mult2): Likewise.
16633 (mp626_mult3): Likewise.
16634 (mp626_mult4): Likewise.
16635 * config/arm/fa526.md (526_mult1): Update attribute change.
16636 (526_mult2): Likewise.
16637 * config/arm/arm-generic.md (mult): Update attribute change.
16638 (mult_ldsched_strongarm): Likewise.
16639 (mult_ldsched): Likewise.
16640 (multi_cycle): Likewise.
16641 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
16642 * config/arm/fa606te.md (606te_mult1): Update attribute change.
16643 (606te_mult2): Likewise.
16644 (606te_mult3): Likewise.
16645 (606te_mult4): Likewise.
16646 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
16647 (cortex_a9_mac16): Likewise.
16648 (cortex_a9_multiply): Likewise.
16649 (cortex_a9_mac): Likewise.
16650 (cortex_a9_multiply_long): Likewise.
16651 * config/arm/fa626te.md (626te_mult1): Update attribute change.
16652 (626te_mult2): Likewise.
16653 (626te_mult3): Likewise.
16654 (626te_mult4): Likewise.
16656 2013-06-18 Richard Biener <rguenther@suse.de>
16659 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
16661 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16664 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16665 with next_active_insn.
16667 2013-06-18 Alan Modra <amodra@gmail.com>
16669 * config/rs6000/rs6000.h (enum data_align): New.
16670 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
16671 (DATA_ABI_ALIGNMENT): Define.
16672 (CONSTANT_ALIGNMENT): Correct comment.
16673 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
16674 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
16676 2013-06-17 David Malcolm <dmalcolm@redhat.com>
16678 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
16679 ATTRIBUTE_UNUSED marking.
16681 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
16683 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
16684 alternative and update.
16685 (aarch64_dup_lanedi): Delete.
16686 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
16687 * config/aarch64/aarch64-simd-builtins.def: Update.
16689 2013-06-17 Richard Biener <rguenther@suse.de>
16691 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
16692 (lto_input_scc): Declare.
16693 (lto_input_tree_1): Likewise.
16694 (struct lto_stats_d): Add num_tree_bodies_output and
16695 num_pickle_refs_output.
16696 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
16697 (lto_read_tree_1): Split out from ...
16698 (lto_read_tree): ... this.
16699 (lto_input_scc): New function.
16700 (lto_input_tree_1): Split out from ...
16701 (lto_input_tree): ... this. Handle LTO_tree_scc.
16702 (lto_data_in_create): Create the streamer cache without hashes.
16703 * lto-streamer-out.c (create_output_block): Create the streamer
16704 cache with hashes when not doing WPA.
16705 (lto_write_tree_1): Split out from ...
16706 (lto_write_tree): ... this.
16707 (get_symbol_initial_value): New function.
16708 (lto_output_tree_1): Split out from ...
16709 (lto_output_tree): ... this. Write trees as series of SCCs
16710 using a DFS walk via DFS_write_tree.
16711 (struct sccs, struct scc_entry): New types.
16712 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
16713 (DFS_write_tree_body): New function.
16714 (DFS_write_tree): Likewise.
16715 (hash_tree): Likewise.
16716 (scc_entry_compare): Likewise.
16717 (hash_scc): Likewise.
16718 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
16719 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
16720 TREE_CHAIN as regular reference.
16721 (streamer_read_integer_cst): Remove.
16722 (streamer_get_pickled_tree): Adjust.
16723 * tree-streamer-out.c (streamer_write_chain): Disable streaming
16724 of DECL_EXTERNALs in BLOCK_VARS for now.
16725 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
16727 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
16728 Add hash value argument and record that if hashes are recorded
16730 (streamer_tree_cache_insert_1): Adjust.
16731 (streamer_tree_cache_insert): Likewise.
16732 (streamer_tree_cache_insert_at): Rename to ...
16733 (streamer_tree_cache_replace_tree): ... this and adjust.
16734 (streamer_tree_cache_append): Adjust.
16735 (record_common_node): Likewise.
16736 (streamer_tree_cache_create): Add argument whether to
16737 record hash values together with trees.
16738 (streamer_tree_cache_delete): Adjust.
16739 * tree-streamer.h (struct streamer_tree_cache_d): Add
16741 (streamer_read_integer_cst): Remove.
16742 (streamer_tree_cache_insert): Adjust.
16743 (streamer_tree_cache_append): Likewise.
16744 (streamer_tree_cache_insert_at): Rename to ...
16745 (streamer_tree_cache_replace_tree): ... this and adjust.
16746 (streamer_tree_cache_create): Add argument whether to record hashes.
16747 (streamer_tree_cache_get): Rename to ...
16748 (streamer_tree_cache_get_tree): ... this.
16749 (streamer_tree_cache_get_hash): New function.
16750 * tree.c (cache_integer_cst): New function.
16751 * tree.h (cache_integer_cst): Declare.
16752 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
16753 * lto-symtab.c (lto_varpool_replace_node): Only release
16754 DECL_INITIAL of non-prevailing decls.
16755 * varpool.c (varpool_remove_initializer): Do not release
16756 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
16758 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
16760 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
16761 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
16762 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
16763 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
16764 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
16765 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
16766 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
16767 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
16768 instead of TARGET_64BIT.
16769 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
16770 Require ISA_HAS_<D>DIV.
16772 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
16774 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
16775 (mips*-*-linux*): Move default with_llsc setting to where other
16777 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
16779 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
16780 Likewise. Remove default with_tune setting. Move default float
16781 setting to its own block. Handle with_llsc in the same block as above.
16783 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
16785 PR rtl-optimization/57425
16786 PR rtl-optimization/57569
16787 * alias.c (write_dependence_p): Add new parameters mem_mode,
16788 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
16789 Changed all callers.
16790 (canon_anti_dependence): New function.
16791 * cse.c (check_dependence): Use canon_anti_dependence.
16792 * cselib.c (cselib_invalidate_mem): Likewise.
16793 * rtl.h (canon_anti_dependence): Declare.
16795 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
16797 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
16798 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
16799 ".set mips3" for 64-bit targets.
16801 2013-06-15 Dehao Chen <dehao@google.com>
16803 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
16804 * gimple-low.c (gimple_check_call_matching_types): Likewise.
16805 (gimple_check_call_args): Likewise.
16806 * value-prof.c (check_ic_target): Likewise.
16807 * ipa-inline.c (early_inliner): Likewise.
16808 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
16809 * cgraph.c (cgraph_create_edge_1): Likewise.
16810 (cgraph_make_edge_direct): Likewise.
16812 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
16815 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
16816 rs6000_output_move_128bit to handle emitting quad memory
16817 operations. Set attribute length to 8 bytes.
16819 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
16821 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
16823 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
16824 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
16825 (aarch64_<su>mlsl<mode>): Likewise.
16827 2013-06-14 Mike Stump <mikestump@comcast.net>
16829 * Makefile.in (TARGET_H): Add insn-codes.h.
16831 2013-06-14 Alan Modra <amodra@gmail.com>
16833 PR middle-end/57134
16834 PR middle-end/57586
16835 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
16836 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
16837 bitfield expansion when EXPAND_MEMORY.
16838 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
16840 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
16842 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
16843 test for clearing quad memory on 32-bit later.
16845 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
16847 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
16848 (fold_negate_expr): Likewise.
16849 (fold_real_zero_addition_p): Handle vectors.
16850 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
16852 2013-06-14 Alan Modra <amodra@gmail.com>
16854 * varasm.c (force_const_mem): Revert 2013-06-07 change.
16856 2013-06-13 Jan Hubicka <jh@suse.cz>
16858 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
16859 Local comdats are not externally visible.
16860 * symtab.c (dump_symtab_base): Dump externally visible.
16861 (verify_symtab_base): Verify back links in the symtab hash.
16863 2013-06-13 Bin Cheng <bin.cheng@arm.com>
16865 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
16866 CONVERT_EXPR as equal nodes.
16868 2013-06-13 Bin Cheng <bin.cheng@arm.com>
16870 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
16872 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
16874 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
16875 Generalize to complex and vector.
16876 * tree.c (build_all_ones_cst): New function.
16877 * tree.h (build_all_ones_cst): Declare it.
16879 2013-06-13 Alan Modra <amodra@gmail.com>
16881 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
16882 * config/rs6000/rs6000.md (signbittf2): New insn.
16883 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
16884 (abstf2_internal, cmptf_internal2): Likewise.
16885 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
16887 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
16888 Pat Haugen <pthaugen@us.ibm.com>
16889 Peter Bergner <bergner@vnet.ibm.com>
16891 * config/rs6000/rs6000.c (emit_load_locked): Add support for
16892 power8 byte, half-word, and quad-word atomic instructions.
16893 (emit_store_conditional): Likewise.
16894 (rs6000_expand_atomic_compare_and_swap): Likewise.
16895 (rs6000_expand_atomic_op): Likewise.
16897 * config/rs6000/sync.md (larx): Add new modes for power8.
16899 (AINT): New mode iterator to include TImode as well as normal
16900 integer modes on power8.
16901 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
16902 that VSX registers are not considered. Use AINT mode iterator
16903 instead of INT1 to allow inclusion of quad word atomic operations
16905 (load_locked<mode>): Likewise.
16906 (store_conditional<mode>): Likewise.
16907 (atomic_compare_and_swap<mode>): Likewise.
16908 (atomic_exchange<mode>): Likewise.
16909 (atomic_nand<mode>): Likewise.
16910 (atomic_fetch_<fetchop_name><mode>): Likewise.
16911 (atomic_nand_fetch<mode>): Likewise.
16912 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
16914 (ATOMIC): On power8, add QImode, HImode modes.
16915 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
16916 modes that promote to SImode.
16917 (load_lockedti): Convert TImode arguments to PTImode, so that we
16918 get a guaranteed even/odd register pair.
16919 (load_lockedpti): Likewise.
16920 (store_conditionalti): Likewise.
16921 (store_conditionalpti): Likewise.
16923 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
16924 atomic load/store instructions.
16927 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
16929 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
16931 (insn_count): New attribute, with most cases extracted from...
16932 (length): ...here. Redefine most cases in terms of insn_count.
16933 (single_insn): Delete.
16934 (can_delay): Use insn_count to check for single instructions.
16935 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
16936 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
16937 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
16938 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
16939 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
16940 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
16941 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
16942 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
16943 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
16944 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
16945 rather than "length".
16946 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
16947 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
16948 Use "insn_count" rather than "length".
16949 * config/mips/mips-dsp.md
16950 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
16951 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
16954 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
16956 PR tree-optimization/57361
16957 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
16959 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
16961 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
16963 (aarch64_simd_combine<mode>): New instruction expansion.
16964 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
16965 function prototype.
16966 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
16967 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
16969 2013-06-12 Jan Hubicka <jh@suse.cz>
16971 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
16972 decl has when in streaming stage.
16973 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
16974 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
16976 2013-06-12 Roland Stigge <stigge@antcom.de>
16979 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
16981 2013-06-12 Jakub Jelinek <jakub@redhat.com>
16983 PR tree-optimization/57537
16984 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
16985 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
16987 2013-06-12 Richard Biener <rguenther@suse.de>
16989 * data-streamer.h (streamer_write_char_stream): CSE
16990 obs->current_pointer.
16991 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
16992 streamer_write_char_stream manually and optimize the resulting loop.
16993 (streamer_write_hwi_stream): Likewise.
16995 2013-06-12 Jan Hubicka <jh@suse.cz>
16997 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
16998 * cgraph.h (varpool_create_empty_node): Declare.
16999 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
17001 * symtab.c (symtab_unregister_node): Be lax about missin entries
17003 (symtab_get_node): Update comment.
17004 * varpool.c (varpool_create_empty_node): Break out from ...
17005 (varpool_node_for_decl): ... here.
17006 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
17008 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
17010 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
17011 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
17012 part. Use straight-line flow at the end.
17013 <COMPONENT_REF>: Remove superfluous else.
17014 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
17016 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17019 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
17020 target hook even for !TREE_PUBLIC decls. If no resolution info
17021 is available, return false for common and external decls.
17023 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17025 * config/rl78/constraints.md (U): New constraint.
17026 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
17029 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17032 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
17033 to allow returning address to AT_PLATFORM name.
17035 2013-06-11 Jan Hubicka <jh@suse.cz>
17037 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
17038 * cgraph.h (symtab_node_base): Add weakref flag.
17039 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
17040 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
17041 (output_weakrefs): Use weakref flag.
17042 * fold-const.c (simple_operand_p): Handle WEAK.
17043 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
17044 * ipa.c (varpool_externally_visible_p): Drop weakref.
17045 (function_and_variable_visibility): Update comment; fix weakref
17046 sanity checks; do not clear DECL_WEAK on them.
17047 * lto-cgraph.c (lto_output_node): update.
17048 (lto_output_varpool_node): Update.
17049 (input_overwrite_node): Update.
17050 (input_node): Update.
17051 (input_varpool_node): Update.
17052 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
17053 (lto_symtab_merge_symbols): Add sanity check.
17054 (lto_symtab_prevailing_decl): Do not special case weakrefs.
17055 * passes.c (rest_of_decl_compilation): Set static flag, too.
17056 * symtab.c (dump_symtab_base): Dump weakref.
17057 (verify_symtab_base): Sanity check weakrefs.
17058 (symtab_make_decl_local): Remove duplicated code.
17059 (symtab_alias_ultimate_target): Simplify.
17060 * varpool.c (varpool_create_variable_alias): Set weakref flag.
17062 2013-06-11 Tom de Vries <tom@codesourcery.com>
17064 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
17065 sequence_vect == NULL.
17067 2013-06-11 DJ Delorie <dj@redhat.com>
17069 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
17070 (rl78_unwind_word_mode): New.
17072 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17074 * final.c (debug_prefix_maps): Make static.
17076 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17078 * function.c (initial_trampoline): Remove stray copy.
17080 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
17082 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
17084 2013-06-11 Martin Jambor <mjambor@suse.cz>
17086 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
17087 within bounds at the beginning of the function.
17089 2013-06-11 Alan Modra <amodra@gmail.com>
17091 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
17093 (get_named_section): Don't NULL !DECL_P decl.
17095 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
17097 * doc/invoke.texi (core-avx2): Document.
17099 (atom): Updated with MOVBE.
17101 2013-06-11 Richard Biener <rguenther@suse.de>
17103 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
17105 2013-06-11 Anton Blanchard <anton@samba.org>
17107 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
17108 correct shift value in little-endian mode.
17110 2013-06-11 Jakub Jelinek <jakub@redhat.com>
17113 * varasm.c (get_variable_align): Move #endif to the right place.
17115 2013-06-10 Cary Coutant <ccoutant@google.com>
17117 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
17118 for hash so that hash table traversal order is deterministic.
17120 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
17121 Pat Haugen <pthaugen@us.ibm.com>
17122 Peter Bergner <bergner@vnet.ibm.com>
17124 * config/rs6000/vector.md (GPR move splitter): Do not split moves
17125 of vectors in GPRS if they are direct moves or quad word load or
17128 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
17130 (direct_move_p): Likewise.
17131 (quad_load_store_p): Likewise.
17133 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
17134 classes into bins based on the physical register type.
17135 (reg_class_to_reg_type): Likewise.
17136 (IS_STD_REG_TYPE): Likewise.
17137 (IS_FP_VECT_REG_TYPE): Likewise.
17138 (reload_fpr_gpr): Arrays to determine what insn to use if we can
17139 use direct move instructions.
17140 (reload_gpr_vsx): Likewise.
17141 (reload_vsx_gpr): Likewise.
17142 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
17143 information that is a simplification of register classes. Also
17144 precalculate direct move reload helpers.
17145 (direct_move_p): New function to return true if the operation can
17146 be done as a direct move instruciton.
17147 (quad_load_store_p): New function to return true if the operation
17148 is a quad memory operation.
17149 (rs6000_legitimize_address): If quad memory, only allow register
17150 indirect for TImode addresses.
17151 (rs6000_legitimate_address_p): Likewise.
17152 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
17153 (rs6000_reload_register_type): Likewise.
17154 (register_to_reg_type): Return register type.
17155 (rs6000_secondary_reload_simple_move): New helper function for
17156 secondary reload and secondary memory needed to identify anything
17157 that is a simple move, and does not need reloading.
17158 (rs6000_secondary_reload_direct_move): New helper function for
17159 secondary reload to identify cases that can be done with several
17160 instructions via the direct move instructions.
17161 (rs6000_secondary_reload_move): New helper function for secondary
17162 reload to identify moves between register types that can be done.
17163 (rs6000_secondary_reload): Add support for quad memory operations
17164 and for direct move.
17165 (rs6000_secondary_memory_needed): Likewise.
17166 (rs6000_debug_secondary_memory_needed): Change argument names.
17167 (rs6000_output_move_128bit): New function to return the move to
17168 use for 128-bit moves, including knowing about the various
17169 limitations of quad memory operations.
17171 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
17172 memory operations. call rs6000_output_move_128bit for the actual
17173 instruciton(s) to generate.
17174 (vsx_movti_64bit): Likewise.
17176 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
17177 (UNSPEC_P8V_MTVSRWZ): Likewise.
17178 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
17179 (UNSPEC_P8V_MTVSRD): Likewise.
17180 (UNSPEC_P8V_XXPERMDI): Likewise.
17181 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
17182 (UNSPEC_FUSION_GPR): Likewise.
17183 (FMOVE128_GPR): New iterator for direct move.
17184 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
17185 (f32_sv): Likewise.
17186 (f32_dm): Likewise.
17187 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
17188 loads and direct move instructions.
17189 (zero_extendsidi2_lfiwzx): Likewise.
17190 (extendsidi2_lfiwax): Likewise.
17191 (extendsidi2_nocell): Likewise.
17192 (floatsi<mode>2_lfiwax): Likewise.
17193 (lfiwax): Likewise.
17194 (floatunssi<mode>2_lfiwzx): Likewise.
17195 (lfiwzx): Likewise.
17196 (fix_trunc<mode>_stfiwx): Likewise.
17197 (fixuns_trunc<mode>_stfiwx): Likewise.
17198 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
17199 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
17200 (parity<mode>2_cmpb): Set length/type attr.
17201 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
17202 for 'mr.' to fast_compare.
17203 (bpermd_<mode>): Change type attr to popcnt.
17204 (p8_fmrgow_<mode>): New insns for power8 direct move support.
17205 (p8_mtvsrwz_1): Likewise.
17206 (p8_mtvsrwz_2): Likewise.
17207 (reload_fpr_from_gpr<mode>): Likewise.
17208 (p8_mtvsrd_1): Likewise.
17209 (p8_mtvsrd_2): Likewise.
17210 (p8_xxpermdi_<mode>): Likewise.
17211 (reload_vsx_from_gpr<mode>): Likewise.
17212 (reload_vsx_from_gprsf): Likewise.
17213 (p8_mfvsrd_3_<mode>): LIkewise.
17214 (reload_gpr_from_vsx<mode>): Likewise.
17215 (reload_gpr_from_vsxsf): Likewise.
17216 (p8_mfvsrd_4_disf): Likewise.
17217 (multi-word GPR splits): Do not split direct moves or quad memory
17220 2013-06-10 David Malcolm <dmalcolm@redhat.com>
17222 * tree-into-ssa.c (interesting_blocks): Make static.
17224 2013-06-10 Jakub Jelinek <jakub@redhat.com>
17227 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
17228 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
17229 Use DATA_ABI_ALIGNMENT for that case instead if defined.
17230 (get_variable_align): New function.
17231 (get_variable_section, emit_bss, emit_common,
17232 assemble_variable_contents, place_block_symbol): Use
17233 get_variable_align instead of DECL_ALIGN.
17234 (assemble_noswitch_variable): Add align argument, use it
17235 instead of DECL_ALIGN.
17236 (assemble_variable): Adjust caller. Use get_variable_align
17237 instead of DECL_ALIGN.
17238 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
17240 (DATA_ABI_ALIGNMENT): Define.
17241 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
17242 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
17243 opt is false, only return the psABI mandated alignment increase.
17244 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
17245 (DATA_ABI_ALIGNMENT): ... this.
17246 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
17247 (DATA_ABI_ALIGNMENT): ... this.
17248 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
17249 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
17250 (DATA_ABI_ALIGNMENT): ... this.
17251 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
17252 * doc/tm.texi: Regenerated.
17254 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
17256 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
17257 cmp_code to construct REG_EQUAL note.
17259 2013-06-09 Jakub Jelinek <jakub@redhat.com>
17262 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
17263 that operands[2] doesn't overlap with operands[0].
17265 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
17266 Jan Hubicka <jh@suse.cz>
17268 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
17269 hack to mark symbols as used.
17271 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
17273 PR rtl-optimization/57559
17274 * lra-constraints.c (process_alt_operands): Don't discourage
17275 memory with known offset for offsetable memory constraint.
17276 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
17278 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
17280 * varasm.c (struct oc_local_state): Reorder fields.
17281 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
17282 and adjust accordingly.
17283 (output_constructor): Reorder initialization code and adjust call to
17284 output_constructor_bitfield.
17286 2013-06-07 Jan Hubicka <jh@suse.cz>
17288 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
17290 2013-06-07 David Malcolm <dmalcolm@redhat.com>
17292 * tree-object-size.c (unknown): Make const.
17294 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17296 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
17297 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
17298 for last alternative in the cpu_facility attribute.
17300 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17303 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
17304 (xordi3): Change operand 2 constraint to arm_xordi_operand.
17305 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
17306 * config/arm/constraints.md (Dg): New constraint.
17307 * config/arm/neon.md (xordi3_neon): Remove.
17308 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
17309 * config/arm/predicates.md (arm_xordi_operand): New predicate.
17311 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17313 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
17314 Clean up alternatives.
17316 2013-06-07 Alan Modra <amodra@gmail.com>
17318 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
17321 2013-06-07 Alan Modra <amodra@gmail.com>
17323 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
17325 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17327 * config/arm/constraints.md (Df): New constraint.
17328 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
17329 Correct length attribute for last two alternatives.
17331 2013-06-07 Alan Modra <amodra@gmail.com>
17333 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
17334 override user -mfp-in-toc.
17335 (offsettable_ok_by_alignment): Consider just the current access
17336 rather than the whole object, unless BLKmode. Handle
17337 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
17338 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
17339 for -mcmodel=medium.
17340 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
17341 override user -mfp-in-toc or -msum-in-toc. Default to
17342 -mno-fp-in-toc for -mcmodel=medium.
17344 2013-06-06 DJ Delorie <dj@redhat.com>
17346 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
17347 TARGET_VALID_POINTER_MODE.
17349 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
17350 Pat Haugen <pthaugen@us.ibm.com>
17351 Peter Bergner <bergner@vnet.ibm.com>
17353 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
17354 Document new power8 builtins.
17356 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
17357 condition code register, to allow 128-bit logical operations to be
17358 done in the VSX or GPR registers.
17359 (nor<mode>3): Use the canonical form for nor.
17360 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
17361 vclz*, and vpopcnt* vector instructions.
17362 (nand<mode>3): Likewise.
17363 (orc<mode>3): Likewise.
17364 (clz<mode>2): LIkewise.
17365 (popcount<mode>2): Likewise.
17367 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
17368 that only the GPRs are recognized.
17370 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
17371 support for new power8 builtins.
17373 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
17375 (xscvdpspn): Likewise.
17381 (vpopcnt): Likewise.
17382 (vpopcntb): Likewise.
17383 (vpopcnth): Likewise.
17384 (vpopcntw): Likewise.
17385 (vpopcntd): Likewise.
17387 (vmrgew): Likewise.
17388 (vmrgow): Likewise.
17390 (eqv_v16qi3): Likewise.
17391 (eqv_v8hi3): Likewise.
17392 (eqv_v4si3): Likewise.
17393 (eqv_v2di3): Likewise.
17394 (eqv_v4sf3): Likewise.
17395 (eqv_v2df3): Likewise.
17397 (nand_v16qi3): Likewise.
17398 (nand_v8hi3): Likewise.
17399 (nand_v4si3): Likewise.
17400 (nand_v2di3): Likewise.
17401 (nand_v4sf3): Likewise.
17402 (nand_v2df3): Likewise.
17404 (orc_v16qi3): Likewise.
17405 (orc_v8hi3): Likewise.
17406 (orc_v4si3): Likewise.
17407 (orc_v2di3): Likewise.
17408 (orc_v4sf3): Likewise.
17409 (orc_v2df3): Likewise.
17411 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
17412 allow power8 quad mode in 64-bit.
17413 (rs6000_builtin_vectorized_function): Add support to vectorize
17414 ISA 2.07 count leading zeros, population count builtins.
17415 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
17416 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
17417 (builtin_function_type): Add vgbbd builtin function which takes an
17419 (altivec_expand_vec_perm_const): Add support for new power8 merge
17422 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
17423 that does not include TImdoe for use with 32-bit.
17424 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
17426 (UNSPEC_VSX_CVDPSPN): Likewise.
17427 (vsx_xscvdpspn): Likewise.
17428 (vsx_xscvspdpn): Likewise.
17429 (vsx_xscvdpspn_scalar): Likewise.
17430 (vsx_xscvspdpn_directmove): Likewise.
17431 (vsx_and<mode>3): Split logical operations into 32-bit and
17432 64-bit. Add support to do logical operations on TImode as well as
17433 VSX vector types. Allow logical operations to be done in either
17434 VSX registers or in general purpose registers in 64-bit mode. Add
17435 splitters if GPRs were used. For AND, add clobber of CCmode to
17436 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
17438 (vsx_and<mode>3_32bit): Likewise.
17439 (vsx_and<mode>3_64bit): Likewise.
17440 (vsx_ior<mode>3): Likewise.
17441 (vsx_ior<mode>3_32bit): Likewise.
17442 (vsx_ior<mode>3_64bit): Likewise.
17443 (vsx_xor<mode>3): Likewise.
17444 (vsx_xor<mode>3_32bit): Likewise.
17445 (vsx_xor<mode>3_64bit): Likewise.
17446 (vsx_one_cmpl<mode>2): Likewise.
17447 (vsx_one_cmpl<mode>2_32bit): Likewise.
17448 (vsx_one_cmpl<mode>2_64bit): Likewise.
17449 (vsx_nor<mode>3): Likewise.
17450 (vsx_nor<mode>3_32bit): Likewise.
17451 (vsx_nor<mode>3_64bit): Likewise.
17452 (vsx_andc<mode>3): Likewise.
17453 (vsx_andc<mode>3_32bit): Likewise.
17454 (vsx_andc<mode>3_64bit): Likewise.
17455 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
17456 and xxlorc instructions.
17457 (vsx_eqv<mode>3_64bit): Likewise.
17458 (vsx_nand<mode>3_32bit): Likewise.
17459 (vsx_nand<mode>3_64bit): Likewise.
17460 (vsx_orc<mode>3_32bit): Likewise.
17461 (vsx_orc<mode>3_64bit): Likewise.
17463 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
17465 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
17467 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
17468 (p8_vmrgow): Likewise.
17469 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
17470 GPRs to be split under VSX.
17471 (p8v_clz<mode>2): Add power8 count leading zero support.
17472 (p8v_popcount<mode>2): Add power8 population count support.
17473 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
17476 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
17479 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
17481 (vec_nand): Likewise.
17482 (vec_vclz): Likewise.
17483 (vec_vclzb): Likewise.
17484 (vec_vclzd): Likewise.
17485 (vec_vclzh): Likewise.
17486 (vec_vclzw): Likewise.
17487 (vec_vgbbd): Likewise.
17488 (vec_vmrgew): Likewise.
17489 (vec_vmrgow): Likewise.
17490 (vec_vpopcnt): Likewise.
17491 (vec_vpopcntb): Likewise.
17492 (vec_vpopcntd): Likewise.
17493 (vec_vpopcnth): Likewise.
17494 (vec_vpopcntw): Likewise.
17496 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17498 PR rtl-optimization/57468
17499 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
17502 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17504 PR rtl-optimization/57459
17505 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
17506 type when setting live regs.
17508 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17510 * config/s390/s390.opt (mlra): New option.
17511 * config/s390/s390.c (s390_decompose_address): Check displacement
17512 for all registers for LRA.
17513 (s390_secondary_reload): Don't used secondary reloads for LRA.
17514 (s390_lra_p): New function.
17515 (TARGET_LRA_P): Define.
17516 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
17517 of attribute cpu_facility to zarch for the last alternative.
17518 (*cmpmem_short): Ditto.
17520 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
17522 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
17523 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
17524 (arm_expand_prologue): Likewise.
17526 2013-06-06 Teresa Johnson <tejohnson@google.com>
17529 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
17530 as this is now done by redirect_edge_and_branch_force.
17531 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
17532 barriers, and fix interaction with splitting.
17533 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
17534 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
17535 reflect changes made in the routine.
17536 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
17537 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
17538 since this is called in cfglayout mode, and replace partition fixup
17539 with assert as that is now done by force_nonfallthru_and_redirect.
17540 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
17541 already be marked with region crossing note.
17542 (insert_section_boundary_note): Make non-static, gate on flag
17543 has_bb_partition, rewrite to also check for multiple partitions.
17544 (rest_of_handle_reorder_blocks): Remove call to
17545 insert_section_boundary_note, now done later during free_cfg.
17546 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
17547 * bb-reorder.h (insert_section_boundary_note): Declare.
17548 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
17549 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
17550 invoke insert_section_boundary_note.
17551 (try_redirect_by_replacing_jump): Remove unnecessary
17552 check for region crossing note.
17553 (fixup_partition_crossing): New function.
17554 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
17555 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
17556 in non-cfglayout mode.
17557 (force_nonfallthru_and_redirect): Fixup partition boundaries,
17558 remove old code that tried to do this. Emit barrier correctly
17559 when we are in cfglayout mode.
17560 (last_bb_in_partition): New function.
17561 (rtl_split_edge): Correctly fixup partition boundaries.
17562 (commit_one_edge_insertion): Remove old code that tried to
17563 fixup region crossing edge since this is now handled in
17564 split_block, and set up insertion point correctly since
17565 block may now end in a jump.
17566 (verify_hot_cold_block_grouping): Guard against checking when not in
17567 linearized RTL mode.
17568 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
17570 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
17571 rtl_verify_flow_info, so not called in cfglayout mode.
17572 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
17573 (fixup_reorder_chain): Remove old code that attempted to fixup region
17574 crossing note as this is now handled in force_nonfallthru_and_redirect.
17575 (duplicate_insn_chain): Don't duplicate switch section notes.
17576 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
17578 * basic-block.h (emit_barrier_after_bb): Declare.
17580 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17582 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
17583 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
17584 arm_usatsihi): Adjust alternatives for arm_restrict_it.
17586 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17588 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
17590 * config/arm/ldmstm.md: Regenerate.
17592 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17594 * config/arm/sync.md (atomic_loaddi_1):
17595 Disable predication for arm_restrict_it.
17596 (arm_load_exclusive<mode>): Likewise.
17597 (arm_load_exclusivesi): Likewise.
17598 (arm_load_exclusivedi): Likewise.
17599 (arm_load_acquire_exclusive<mode>): Likewise.
17600 (arm_load_acquire_exclusivesi): Likewise.
17601 (arm_load_acquire_exclusivedi): Likewise.
17602 (arm_store_exclusive<mode>): Likewise.
17603 (arm_store_exclusive<mode>): Likewise.
17604 (arm_store_release_exclusivedi): Likewise.
17605 (arm_store_release_exclusive<mode>): Likewise.
17607 2013-06-06 Richard Biener <rguenther@suse.de>
17609 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
17611 (lto_tag_is_tree_code_p): Adjust.
17612 (lto_tag_is_gimple_code_p): Likewise.
17613 (lto_gimple_code_to_tag): Likewise.
17614 (lto_tag_to_gimple_code): Likewise.
17615 (lto_tree_code_to_tag): Likewise.
17616 (lto_tag_to_tree_code): Likewise.
17617 * data-streamer.h (streamer_write_hwi_in_range): Use
17618 uhwi streaming to stream the normalized range.
17619 (streamer_read_hwi_in_range): Likewise.
17621 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17623 * config/arm/arm.md (enabled_for_depr_it): New attribute.
17624 (predicable_short_it): Likewise.
17625 (predicated): Likewise.
17626 (enabled): Handle above.
17627 (define_cond_exec): Set predicated attribute to yes.
17629 2013-06-05 Mike Stump <mikestump@comcast.net>
17631 * gdbinit.in (__FUNCTION__): Add.
17633 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
17635 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
17636 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
17638 2013-06-05 Jan Hubicka <jh@suse.cz>
17640 * varasm.c (mark_decl_referenced): Revert the removal until targets
17643 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
17645 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
17646 instead of mark_decl_referenced.
17648 2013-06-05 Jan Hubicka <jh@suse.cz>
17650 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
17651 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
17652 and symtab_used_from_object_file_p.
17653 (cgraph_make_node_local_1): Clear forced_by_abi.
17654 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
17655 * cgraph.h (symtab_node_base): Add forced_by_abi.
17656 (decide_is_variable_needed): Remove.
17657 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
17658 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
17659 (decide_is_symbol_needed): ... this one; handle symbols in general;
17660 always analyze virtuals; honnor forced_by_abi.
17661 (cgraph_finalize_function): Update.
17662 (varpool_finalize_decl): Update.
17663 (symbol_defined_and_needed): Remove.
17664 (analyze_functions): Update.
17665 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17666 output_refs, input_overwrite_node): Handle forced_by_abi.
17667 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
17668 (address_taken_from_non_vtable_p): ... this one.
17669 (comdat_can_be_unshared_p_1): New function.
17670 (cgraph_comdat_can_be_unshared_p): Rename to ...
17671 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
17672 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
17673 (function_and_variable_visibility): Clear forced_by_abi as needed.
17674 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
17675 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
17676 * symtab.c (dump_symtab_base): Dump forced_by_abi.
17677 * varpool.c (decide_is_variable_needed): Remove.
17679 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17681 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
17682 (arm_option_override): Override arm_restrict_it where appropriate.
17683 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
17684 * config/arm/arm.opt (mrestrict-it): New command-line option.
17685 * doc/invoke.texi: Document -mrestrict-it.
17687 2013-06-05 David Malcolm <dmalcolm@redhat.com>
17689 * tsan.c (tsan_atomic_table): Make const.
17691 2013-06-05 Richard Biener <rguenther@suse.de>
17693 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
17694 index associated with the tree we are supposed to replace.
17695 * tree-streamer-out.c (pack_ts_base_value_fields): Output
17696 TREE_ASM_WRITTEN as zero for everything but SSA names.
17698 2013-06-05 David Malcolm <dmalcolm@redhat.com>
17700 * tree-ssa-structalias.c (call_stmt_vars): Make static.
17702 2013-06-04 Jan Hubicka <jh@suse.cz>
17704 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
17705 (input_node, input_varpool_node): Handle correctly external same
17707 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
17708 nodes at ltrans stage.
17710 2013-06-04 Jan Hubicka <jh@suse.cz>
17712 * ipa-inline.c (update_caller_keys): Fix availability test.
17713 (update_callee_keys): Likewise.
17714 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
17715 to follow ELF standard.
17717 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
17719 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
17720 (mips64r5900el-*-elf*): New configurations.
17721 * config/mips/mips-cpus.def (r5900): New processor.
17722 * config/mips/mips-tables.opt: Regenerate.
17723 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
17724 (mips_issue_rate): Handle PROCESSOR_R5900.
17725 (mips_reorg_process_insns): Force reorder mode for the R5900.
17726 * config/mips/mips.h (TARGET_MIPS5900): Define.
17727 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
17729 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
17731 * config/mips/mips.md (processor): Add r5900.
17732 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
17734 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17736 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
17737 into function to generate MOVI instruction.
17738 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
17739 (aarch64_preferred_simd_mode): Turn into wrapper.
17740 (aarch64_output_scalar_simd_mov_immediate): New function.
17741 * config/aarch64/aarch64-protos.h: Add prototype for above.
17743 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17745 * config/aarch64/aarch64.c (simd_immediate_info): Remove
17746 element_char member.
17747 (sizetochar): Return signed char.
17748 (aarch64_simd_valid_immediate): Remove elchar and other
17749 unnecessary variables.
17750 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
17751 Calculate element_char as required.
17752 * config/aarch64/aarch64-protos.h: Update and move prototype
17753 for aarch64_output_simd_mov_immediate.
17754 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
17757 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17759 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
17760 information completed by aarch64_simd_valid_immediate.
17761 (aarch64_legitimate_constant_p): Update arguments.
17762 (aarch64_simd_valid_immediate): Work with struct rather than many
17764 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
17765 (aarch64_simd_make_constant): Update arguments.
17766 (aarch64_output_simd_mov_immediate): Work with struct rather than
17767 many pointers. Output immediate directly rather than as operand.
17768 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
17770 * config/aarch64/constraints.md (Dn): Update arguments.
17772 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17774 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
17776 (aarch64_simd_immediate_valid_for_move): Remove.
17777 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
17778 (aarch64_simd_make_constant): Update call.
17779 (aarch64_output_simd_mov_immediate): Update call.
17780 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
17782 * config/aarch64/constraints.md (Dn): Update call.
17784 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17786 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
17787 return type to bool for prototype.
17788 (aarch64_legitimate_constant_p): Check for true instead of not -1.
17789 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
17790 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
17792 2013-06-04 Catherine Moore <clm@codesourcery.com>
17794 * config/mips/mips.opt (meva): New.
17795 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
17796 (ASM_SPEC): Handle -meva.
17797 * doc/invoke.texi (meva): Document.
17799 2013-06-04 Alan Modra <amodra@gmail.com>
17801 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
17804 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17806 * rtl.def: Add extra fourth optional field to define_cond_exec.
17807 * gensupport.c (process_one_cond_exec): Process attributes from
17809 * doc/md.texi: Document fourth field in define_cond_exec.
17811 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
17813 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
17814 out the processing order as in store_bit_field_1.
17816 2013-06-04 Jan Hubicka <jh@suse.cz>
17818 PR middle-end/57500
17819 * cgraphunit.c (cgraph_process_same_body_aliases): Create
17820 non-VAR_DECL node if it does not exist yet.
17822 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
17824 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
17825 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
17826 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
17827 target_cpu_default setting.
17829 2013-06-03 Teresa Johnson <tejohnson@google.com>
17831 * dumpfile.c (opt_info_switch_p): Change -fopt-info
17832 default to -fopt-info=optimized instead of all.
17833 * doc/invoke.texi: Ditto.
17834 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
17835 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
17836 (execute_vect_slp): Emit BB vectorization success under
17837 MSG_OPTIMIZED_LOCATIONS.
17838 * tree-vect-slp.c (vect_slp_transform_bb): Change
17839 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
17840 * tree-vect-loop.c (vect_transform_loop): Ditto.
17842 2013-06-03 Jason Merrill <jason@redhat.com>
17845 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
17846 Use TARGET_EXPR for C++.
17848 2013-06-03 Jakub Jelinek <jakub@redhat.com>
17850 PR rtl-optimization/57268
17851 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
17852 if DEBUG_INSN_P (insn).
17855 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
17857 PR rtl-optimization/57268
17858 * sched-deps.c (sched_analyze_2): Flush dependence lists if
17859 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
17861 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
17863 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
17864 (ix86_avoid_lea_for_addr): Likewise.
17865 (exact_dependency_1): Likewise.
17866 (ix86_adjust_cost): Likewise.
17867 (swap_top_of_ready_list): Fix formatting and !reload_completed check
17869 (do_reorder_for_imul): Fix typo, formatting and
17870 !reload_completed check removed.
17871 (ix86_sched_reorder): Fix typo and formatting.
17872 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
17875 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
17877 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
17879 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
17881 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
17882 <STRING_CST>: Likewise.
17883 <VECTOR_CST>: Likewise.
17885 2013-06-01 Janus Weil <janus@gcc.gnu.org>
17886 Mikael Morin <mikael@gcc.gnu.org>
17888 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
17889 * config.in: Regenerated.
17890 * configure: Regenerated.
17892 2013-06-01 Jan Hubicka <jh@suse.cz>
17894 PR middle-end/57366
17895 * cgraphunit.c (compile): When weakref is not supported,
17896 set up transparent aliases before final output pass.
17897 * varasm.c (assemble_alias): Do not try to do it here.
17899 2013-06-01 Jan Hubicka <jh@suse.cz>
17901 PR middle-end/57467
17902 * passes.c (for_per_function): Skip unanalyzed functions.
17904 2013-06-01 Jan Hubicka <jh@suse.cz>
17906 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
17907 (lto_symtab_merge_symbols_1): ... this one.
17908 (lto_symtab_merge_cgraph_nodes): Rename to ...
17909 (lto_symtab_merge_symbols): ... this one; simplify.
17910 * cgraph.c (same_body_aliases_done): Rename to ...
17911 (cpp_implicit_aliases_done): ... this one.
17912 (cgraph_create_function_alias): Update.
17913 (cgraph_same_body_alias): Update.
17914 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
17915 (verify_edge_corresponds_to_fndecl): Simplify.
17916 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
17917 (cgraph_node): Remove same_body_alias.
17918 (varpool_node): Remove alias_of and extra_name_alias.
17919 (same_body_aliases_done): Rename to ..
17920 (cpp_implicit_aliases_done): ... this one.
17921 (symtab_alias_ultimate_target): Add default parameter.
17922 (symtab_resolve_alias): New function.
17923 (fixup_same_cpp_alias_visibility): Declare.
17924 (cgraph_function_node): Add default parameter.
17925 (cgraph_node_asm_name): Likewise.
17926 (cgraph_function_or_thunk_node): Add default parameter; do
17927 not ICE when it is NULL.
17928 (varpool_variable_node): Likewise.
17929 * tree-emutls.c (create_emultls_var): Update.
17930 (ipa_lower_emutls): Update.
17931 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
17932 (cgraph_reset_node): Reset alias info.
17933 (cgraph_finalize_function): Update.
17934 (fixup_same_cpp_alias_visibility): Move to symtab.c.
17935 (analyze_function): Simplify.
17936 (cgraph_process_same_body_aliases): Simplify.
17937 (analyze_functions): Fixup same body aliases.
17938 (handle_alias_pairs): Simplify.
17939 (assemble_thunk): Update.
17940 (assemble_thunks_and_aliases): Update.
17941 (output_weakrefs): Rewrite.
17942 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
17943 (lto_output_varpool_node): Likewise.
17944 (compute_ltrans_boundary): Remve assert.
17945 (get_alias_symbol): New functoin.
17946 (input_node): Rewrite alias handling.
17947 (input_varpool_node): Likewise.
17948 * ipa-pure-const.c (propagate_pure_const): Fix formating.
17949 * ipa.c (process_references): Handle weakrefs correctly.
17950 (symtab_remove_unreachable_nodes): Likewise.
17951 * trans-mem.c (get_cg_data): Update.
17952 (ipa_tm_create_version_alias): Update.
17953 (ipa_tm_execute): Update.
17954 * symtab.c (dump_symtab_base): Dump aliases.
17955 (verify_symtab_base): Verify aliases.
17956 (symtab_node_availability): New function.
17957 (symtab_alias_ultimate_target): Simplify.
17958 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
17959 handle all the fixup cases.
17960 (symtab_resolve_alias): New function.
17961 * passes.c (ipa_write_summaries): Handle weakrefs.
17962 * varpool.c (varpool_analyze_node): Simplify.
17963 (assemble_aliases): Update.
17964 (varpool_create_variable_alias): Simplify.
17965 (varpool_extra_name_alias): Simplify.
17966 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
17967 (lto_symtab_merge_symbols): ... this one.
17969 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
17972 PR rtl-optimization/57268
17973 * sched-deps.c (sched_analyze_2): Flush dependence lists if
17974 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
17976 2013-06-01 Tobias Burnus <burnus@net-b.de>
17978 Partially reverted:
17979 2013-05-31 Tobias Burnus <burnus@net-b.de>
17981 PR middle-end/57073
17982 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
17985 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
17987 PR rtl-optimization/57268
17988 * sched-deps.c (sched_analyze_2): Flush dependence lists if
17989 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
17991 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
17993 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
17994 unordered comparison operators when -fno-trapping-math is in effect
17996 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
17997 and implement unordered comparison operators properly on the e500.
17999 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18001 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
18002 for constant scalar integers.
18003 (simplify_relational_operation_1): Likewise.
18005 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
18007 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
18008 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
18011 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18012 Igor Zamyatin <igor.zamyatin@intel.com>
18014 Silvermont (SLM) architecture performance tuning.
18015 * config/i386/i386.h (enum ix86_tune_indices): Add
18016 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
18017 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
18019 * config/i386/i386.c (initial_ix86_tune_features)
18020 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
18021 (ix86_lea_outperforms): Handle Silvermont tuning.
18022 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
18024 (ix86_use_lea_for_mov): Likewise.
18025 (ix86_avoid_lea_for_addr): Likewise.
18026 (ix86_lea_for_add_ok): Likewise.
18027 (exact_dependency_1): New function.
18028 (exact_store_load_dependency): Likewise.
18029 (ix86_adjust_cost): Handle Silvermont tuning.
18030 (do_reoder_for_imul): Likewise.
18031 (swap_top_of_ready_list): New function.
18032 (ix86_sched_reorder): Changed to handle Silvermont tuning.
18034 * config/i386/i386.md (peepholes that split memory operand in fp
18037 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18039 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
18040 Remove un-necessary braces.
18042 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18044 * config/aarch64/aarch64.c (aarch64_classify_symbol):
18045 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
18047 2013-05-31 Tobias Burnus <burnus@net-b.de>
18049 PR middle-end/57073
18050 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
18052 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18055 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
18056 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
18057 * config/arm/neon.md (iordi3_neon): Remove.
18058 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
18059 * config/arm/predicates.md (imm_for_neon_logic_operand):
18060 Move to earlier in the file.
18061 (neon_logic_op2): Likewise.
18062 (arm_iordi_operand_neon): New predicate.
18064 2013-05-31 Richard Biener <rguenther@suse.de>
18066 PR tree-optimization/57478
18067 PR tree-optimization/57453
18068 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
18071 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18073 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
18074 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
18076 2013-05-30 Tobias Burnus <burnus@net-b.de>
18077 Thomas Koenig <tkoenig@gcc.gnu.org>
18079 PR middle-end/57073
18080 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
18081 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
18083 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
18085 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
18087 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
18089 * target.def (register_usage_leveling_p): New hook.
18090 * targhooks.c (default_register_usage_leveling_p): New.
18091 * targhooks.h (default_register_usage_leveling_p): New prototype.
18092 * lra-assigns.c (register_usage_leveling_p): Use the hook.
18093 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
18094 * doc/tm.texi: Update.
18095 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
18097 2013-05-30 Ian Bolton <ian.bolton@arm.com>
18099 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
18100 (*insv_reg<mode>): New define_insn.
18102 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
18104 PR rtl-optimization/57439
18105 * postreload.c (move2add_valid_value_p): Check that we have
18106 a zero subreg_regno_offset when accessing the register in
18107 the requested mode.
18109 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18110 Igor Zamyatin <igor.zamyatin@intel.com>
18112 Silvermont (SLM) architecture pipeline model, tuning and
18114 * config.gcc: Add slm config options and target.
18116 * config/i386/slm.md: New.
18118 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
18120 * config/i386/i386-c.c (ix86_target_macros_internal): New case
18122 (ix86_target_macros_internal): Likewise.
18124 * config/i386/i386.c (slm_cost): New cost.
18125 (m_SLM): New macro flag.
18126 (initial_ix86_tune_features): Set m_SLM.
18127 (x86_accumulate_outgoing_args): Likewise.
18128 (x86_arch_always_fancy_math_387): Likewise.
18129 (processor_target_table): Add slm cost.
18130 (cpu_names): Add slm cpu name.
18131 (x86_option_override_internal): Set SLM ISA.
18132 (ix86_issue_rate): New case PROCESSOR_SLM.
18133 (ia32_multipass_dfa_lookahead): Likewise.
18134 (fold_builtin_cpu): Add slm.
18136 * config/i386/i386.h (TARGET_SLM): New target macro.
18137 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
18138 (processor_type): Add PROCESSOR_SLM.
18140 * config/i386/i386.md (cpu): Add new value "slm".
18141 (slm.md): Include slm.md.
18143 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
18144 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18146 * config/arm/arm-protos.h: Add and update function protos.
18147 * config/arm/arm.c (use_simple_return_p): New added.
18148 (thumb2_expand_return): Check simple_return flag.
18149 * config/arm/arm.md: Add simple_return and conditional simple_return.
18150 * config/arm/iterators.md: Add iterator for return and simple_return.
18152 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18154 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
18155 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
18156 (arm_emit_vfp_multi_reg_pop): Likewise.
18157 (thumb2_emit_ldrd_pop): Likewise.
18158 (arm_expand_epilogue): Add misc REG_CFA notes.
18159 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
18161 2013-05-29 Lawrence Crowl <crowl@google.com>
18163 * config/arm/t-arm: Update for below.
18165 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
18166 Change type to hash_table. Update dependent calls and types.
18168 * config/i386/t-cygming: Update for below.
18170 * config/i386/t-interix: Update for below.
18172 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
18173 Change type to hash_table. Update dependent calls and types.
18174 (i386_find_on_wrapper_list::wrappers): Likewise.
18176 * config/ia64/t-ia64: Update for below.
18178 * config/ia64/ia64.c (bundle_state_table):
18179 Change type to hash_table. Update dependent calls and types.
18181 * config/mips/mips.c (mips_reorg_process_insns::htab):
18182 Change type to hash_table. Update dependent calls and types.
18184 * config/sol2.c (solaris_comdat_htab):
18185 Change type to hash_table. Update dependent calls and types.
18187 * config/t-sol2: Update for above.
18189 2013-05-29 Teresa Johnson <tejohnson@google.com>
18191 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
18192 functions are not yet marked as defined.
18194 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
18195 Pat Haugen <pthaugen@us.ibm.com>
18196 Peter Bergner <bergner@vnet.ibm.com>
18198 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
18202 (vrotl<mode>3): Likewise.
18203 (vashl<mode>3): Likewise.
18204 (vlshr<mode>3): Likewise.
18205 (vashr<mode>3): Likewise.
18207 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18208 support for power8 V2DI builtins.
18210 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
18211 power8 V2DI builtins.
18212 (vupkhsw): Likewise.
18213 (vupklsw): Likewise.
18214 (vaddudm): Likewise.
18215 (vminsd): Likewise.
18216 (vmaxsd): Likewise.
18217 (vminud): Likewise.
18218 (vmaxud): Likewise.
18219 (vpkudum): Likewise.
18220 (vpksdss): Likewise.
18221 (vpkudus): Likewise.
18222 (vpksdus): Likewise.
18227 (vsubudm): Likewise.
18228 (vcmpequd): Likewise.
18229 (vcmpgtsd): Likewise.
18230 (vcmpgtud): Likewise.
18231 (vcmpequd_p): Likewise.
18232 (vcmpgtsd_p): Likewise.
18233 (vcmpgtud_p): Likewise.
18234 (vupkhsw): Likewise.
18235 (vupklsw): Likewise.
18236 (vaddudm): Likewise.
18237 (vmaxsd): Likewise.
18238 (vmaxud): Likewise.
18239 (vminsd): Likewise.
18240 (vminud): Likewise.
18241 (vpksdss): Likewise.
18242 (vpksdus): Likewise.
18243 (vpkudum): Likewise.
18244 (vpkudus): Likewise.
18249 (vsubudm): Likewise.
18251 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
18252 support for power8 V2DI instructions.
18254 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
18255 power8 V2DI instructions. Combine pack and unpack insns to use an
18256 iterator for each mode. Check whether a particular mode supports
18257 Altivec instructions instead of just checking TARGET_ALTIVEC.
18258 (UNSPEC_VPKUWUM): Likewise.
18259 (UNSPEC_VPKSHSS): Likewise.
18260 (UNSPEC_VPKSWSS): Likewise.
18261 (UNSPEC_VPKUHUS): Likewise.
18262 (UNSPEC_VPKSHUS): Likewise.
18263 (UNSPEC_VPKUWUS): Likewise.
18264 (UNSPEC_VPKSWUS): Likewise.
18265 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
18266 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
18267 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
18268 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
18269 (UNSPEC_VUPKHSB): Likewise.
18270 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
18271 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
18272 (UNSPEC_VUPKHSH): Likewise.
18273 (UNSPEC_VUPKLSB): Likewise.
18274 (UNSPEC_VUPKLSH): Likewise.
18276 (VI_char): Likewise.
18277 (VI_scalar): Likewise.
18278 (VI_unit): Likewise.
18280 (VP_small): Likewise.
18281 (VP_small_lc): Likewise.
18282 (VU_char): Likewise.
18283 (add<mode>3): Likewise.
18284 (altivec_vaddcuw): Likewise.
18285 (altivec_vaddu<VI_char>s): Likewise.
18286 (altivec_vadds<VI_char>s): Likewise.
18287 (sub<mode>3): Likewise.
18288 (altivec_vsubcuw): Likewise.
18289 (altivec_vsubu<VI_char>s): Likewise.
18290 (altivec_vsubs<VI_char>s): Likewise.
18291 (altivec_vavgs<VI_char>): Likewise.
18292 (altivec_vcmpbfp): Likewise.
18293 (altivec_eq<mode>): Likewise.
18294 (altivec_gt<mode>): Likewise.
18295 (altivec_gtu<mode>): Likewise.
18296 (umax<mode>3): Likewise.
18297 (smax<mode>3): Likewise.
18298 (umin<mode>3): Likewise.
18299 (smin<mode>3): Likewise.
18300 (altivec_vpkuhum): Likewise.
18301 (altivec_vpkuwum): Likewise.
18302 (altivec_vpkshss): Likewise.
18303 (altivec_vpkswss): Likewise.
18304 (altivec_vpkuhus): Likewise.
18305 (altivec_vpkshus): Likewise.
18306 (altivec_vpkuwus): Likewise.
18307 (altivec_vpkswus): Likewise.
18308 (altivec_vpks<VI_char>ss): Likewise.
18309 (altivec_vpks<VI_char>us): Likewise.
18310 (altivec_vpku<VI_char>us): Likewise.
18311 (altivec_vpku<VI_char>um): Likewise.
18312 (altivec_vrl<VI_char>): Likewise.
18313 (altivec_vsl<VI_char>): Likewise.
18314 (altivec_vsr<VI_char>): Likewise.
18315 (altivec_vsra<VI_char>): Likewise.
18316 (altivec_vsldoi_<mode>): Likewise.
18317 (altivec_vupkhsb): Likewise.
18318 (altivec_vupkhs<VU_char>): Likewise.
18319 (altivec_vupkls<VU_char>): Likewise.
18320 (altivec_vupkhsh): Likewise.
18321 (altivec_vupklsb): Likewise.
18322 (altivec_vupklsh): Likewise.
18323 (altivec_vcmpequ<VI_char>_p): Likewise.
18324 (altivec_vcmpgts<VI_char>_p): Likewise.
18325 (altivec_vcmpgtu<VI_char>_p): Likewise.
18326 (abs<mode>2): Likewise.
18327 (vec_unpacks_hi_v16qi): Likewise.
18328 (vec_unpacks_hi_v8hi): Likewise.
18329 (vec_unpacks_lo_v16qi): Likewise.
18330 (vec_unpacks_hi_<VP_small_lc>): Likewise.
18331 (vec_unpacks_lo_v8hi): Likewise.
18332 (vec_unpacks_lo_<VP_small_lc>): Likewise.
18333 (vec_pack_trunc_v8h): Likewise.
18334 (vec_pack_trunc_v4si): Likewise.
18335 (vec_pack_trunc_<mode>): Likewise.
18337 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
18339 (vec_vmaxsd): Likewise.
18340 (vec_vmaxud): Likewise.
18341 (vec_vminsd): Likewise.
18342 (vec_vminud): Likewise.
18343 (vec_vpksdss): Likewise.
18344 (vec_vpksdus): Likewise.
18345 (vec_vpkudum): Likewise.
18346 (vec_vpkudus): Likewise.
18347 (vec_vrld): Likewise.
18348 (vec_vsld): Likewise.
18349 (vec_vsrad): Likewise.
18350 (vec_vsrd): Likewise.
18351 (vec_vsubudm): Likewise.
18352 (vec_vupkhsw): Likewise.
18353 (vec_vupklsw): Likewise.
18355 2013-05-29 Jan Hubicka <jh@suse.cz>
18357 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
18358 flags; reorder rest of fields in more consistent way.
18359 (varpool_node): Remove analyzed, finalized and alias.
18360 (cgraph_ndoe): Likewise.
18361 (symtab_alias_ultimate_target): New function.
18362 (cgraph_function_node): Move offline.
18363 (cgraph_reset_node): Declare.
18364 (cgraph_comdat_can_be_unshared_p): Remove.
18365 (varpool_remove_initializer): Declare.
18366 (varpool_first_defined_variable, varpool_next_defined_variable
18367 cgraph_first_defined_function, cgraph_next_defined_function): Update.
18368 (cgraph_function_with_gimple_body_p): Update.
18369 (varpool_all_refs_explicit_p): Update.
18370 (symtab_alias_target): New function.
18371 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
18372 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
18373 (cgraph_function_or_thunk_node): Simplify using
18374 symtab_alias_ultimate_target.
18375 (varpool_variable_node): Likewise.
18376 * cgraph.c (cgraph_create_function_alias): Update.
18377 (cgraph_add_thunk): Update.
18378 (cgraph_remove_node): Update.
18379 (dump_cgraph_node): Do not dump removed flags.
18380 (cgraph_function_body_availability): Update.
18381 (cgraph_propagate_frequency): Update.
18382 (verify_cgraph_node): Check sanity of local flag.
18383 (cgraph_function_node): Move here from cgraph.h; revamp for
18384 cgraph_function_or_thunk_node.
18385 * lto-symtab.c (lto_varpool_replace_node): Update.
18386 (lto_symtab_resolve_can_prevail_p): Update.
18387 (lto_symtab_merge_cgraph_nodes): Update.
18388 * ipa-cp.c (determine_versionability, initialize_node_lattices,
18389 propagate_constants_accross_call, devirtualization_time_bonus,
18390 ipcp_propagate_stage): Update.
18391 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
18392 * ipa-inline-transform.c (clone_inlined_nodes,
18393 preserve_function_body_p): Update.
18394 * ipa-reference.c (propagate): Update.
18395 (write_node_summary_p): Update.
18396 * toplev.c (wrapup_global_declaration_2): Update.
18397 * cgraphunit.c (cgraph_analyze_function): Rename to ...
18398 (analyze_function) ... this one.
18399 (cgraph_process_new_functions): Update.
18400 (cgraph_reset_node): Export.
18401 (cgraph_finalize_function): Update.
18402 (cgraph_add_new_function): Update.
18403 (process_function_and_variable_attributes): Update.
18404 (varpool_finalize_decl): Update.
18405 (symbol_finalized): Remove.
18406 (symbol_finalized_and_needed): Rename to ...
18407 (symbol_defined_and_needed): ... update.
18408 (cgraph_analyze_functions): Update.
18409 (handle_alias_pairs): Update.
18410 (mark_functions_to_output): Update.
18411 (assemble_thunk): Update.
18412 (output_in_order): Update.
18413 (output_weakrefs): Update.
18414 (finalize_compilation_unit): Update.
18415 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
18416 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
18417 input_node, input_varpool_node): Update.
18418 * dbxout.c (dbxout_expand_expr): Update.
18419 * cgraphclones.c (cgraph_clone_node): Update.
18420 (cgraph_copy_node_for_versioning): Update.
18421 (cgraph_materialize_clone): Update.
18422 (cgraph_materialize_all_clones): Update.
18423 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
18424 propagate_pure_const, propagate_nothrow): Update.
18425 * lto-streamer-out.c (lto_output, write_symbol): Update.
18426 * ipa-utils.c (ipa_reverse_postorder): Update.
18427 * ipa-inline.c (can_inline_edge_p): Update.
18428 (update_caller_keys, ipa_inline): Update.
18429 * dwarf2out.c (reference_to_unused,
18430 premark_types_used_by_global_vars_helper): Update.
18431 * tree-eh.c (tree_could_trap_p): Update.
18432 * ipa-split.c (consider_split, execute_split_functions): Update.
18433 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
18434 has_addr_references_p): Update; move ahead in file for better
18436 (process_references): Simplify.
18437 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
18438 bodies are removed.
18439 (cgraph_comdat_can_be_unshared_p): Make static.
18440 (cgraph_externally_visible_p): Update.
18441 (varpool_externally_visible_p): Update.
18442 (function_and_variable_visibility): Update.
18443 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
18444 ipa_tm_mark_force_output_node): Update.
18445 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
18446 estimate_edge_devirt_benefit, inline_generate_summary,
18447 inline_write_summary): Update.
18448 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
18449 * ipa-prop.c (ipa_compute_jump_functions): Update.
18450 (ipa_print_node_params, ipa_prop_read_section,
18451 ipa_update_after_lto_read, read_replacements_section): Update.
18452 * varasm.c (mark_decl_referenced): Update.
18453 (assemble_alias, dump_tm_clone_pairs): Update.
18454 * tree-inline.c (copy_bb): Update.
18455 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
18457 * symtab.c (dump_symtab_base): Print new flags.
18458 (verify_symtab_base): Verify new flags.
18459 (symtab_alias_ultimate_target): New function.
18460 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
18461 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
18463 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
18465 * i386.c (ix86_get_function_versions_dispatcher,
18466 ix86_generate_version_dispatcher_body): Update.
18467 (fold_builtin_cpu): Use varpool_add_new_variable.
18468 * varpool.c (varpool_remove_initializer): Break out from ...
18469 (varpool_remove_node): ... this one.
18470 (dump_varpool_node, varpool_node_for_asm,
18471 cgraph_variable_initializer_availability, varpool_analyze_node,
18472 varpool_assemble_decl, varpool_remove_unreferenced_decls,
18473 varpool_finalize_named_section_flags, varpool_create_variable_alias):
18476 2013-05-29 Jan Hubicka <jh@suse.cz>
18478 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
18480 2013-05-29 Easwaran Raman <eraman@google.com>
18482 PR tree-optimization/57442
18483 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
18484 when control exits the main loop.
18486 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
18488 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
18490 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
18491 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
18492 * rx/t-rx: Add rx100 under multi library matches option for nofpu
18495 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18497 PR tree-optimization/57441
18498 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
18499 Don't limit size of incr_vec to number of candidates.
18501 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
18503 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
18504 and mips16 directories.
18505 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
18506 (MULTILIB_DIRNAMES): Ditto.
18507 (MULTILIB_EXCEPTIONS): Add new exceptions.
18508 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
18509 (MULTILIB_DIRNAMES): Ditto.
18510 (MULTILIB_EXCEPTIONS): Add new exceptions.
18512 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18513 Marcus Shawcroft <marcus.shawcroft@arm.com>
18515 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
18516 SYMBOL_TINY_ABSOLUTE.
18517 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
18518 SYMBOL_TINY_ABSOLUTE.
18519 (aarch64_expand_mov_immediate): Likewise.
18520 (aarch64_classify_symbol): Likewise.
18521 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
18522 Permit SYMBOL_TINY_ABSOLUTE.
18523 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
18525 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18526 Marcus Shawcroft <marcus.shawcroft@arm.com>
18528 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
18529 Refactor if/switch. Replace gcc_assert with if.
18531 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
18533 * config/i386/i386.c (initial_ix86_tune_features): Enable
18534 FP Reassociation for AMD bdver1 and bdver2.
18536 2013-05-29 Martin Jambor <mjambor@suse.cz>
18538 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
18539 and IMAGPART_EXPR do not occur within other handled_components.
18541 2013-05-29 Richard Biener <rguenther@suse.de>
18543 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
18544 access on whether the use is in the BB we currently try to
18546 (vect_bb_vectorization_profitable_p): Pass the BB we currently
18547 vectorize to vect_bb_slp_scalar_cost.
18549 2013-05-29 Richard Biener <rguenther@suse.de>
18551 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
18552 computing scalar cost offsetted by stmts that are kept live
18554 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
18555 for computation of scalar cost.
18557 2013-05-28 Steve Ellcey <sellcey@mips.com>
18559 * config/mips/mips-cpus.def (mips32r2): Change processor type.
18561 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
18563 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
18564 array notation built-in reduction functions.
18565 * doc/passes.texi (Passes): Added documentation about changes done
18567 * doc/invoke.texi (C Dialect Options): Added documentation about
18568 the -fcilkplus flag.
18569 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
18570 (BUILTINS_DEF): Depend on cilkplus.def.
18571 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
18572 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
18573 * cilkplus.def: New file.
18575 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
18577 PR rtl-optimization/57439
18578 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
18580 2013-05-28 Easwaran Raman <eraman@google.com>
18582 PR tree-optimization/57337
18583 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
18584 (find_insert_point): Correctly identify the insertion point
18585 when two statements with the same UID is compared.
18587 2013-05-28 Richard Biener <rguenther@suse.de>
18589 PR tree-optimization/56787
18590 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
18591 from the list of data references.
18592 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
18594 (vect_analyze_loop_operations): Likewise.
18595 (vect_transform_loop): Remove clobbers.
18597 2013-05-28 Martin Jambor <mjambor@suse.cz>
18599 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
18600 and REALPART_EXPRs have scalar type.
18602 2013-05-28 Richard Biener <rguenther@suse.de>
18604 PR tree-optimization/57411
18605 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
18607 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
18608 virtual operand propagation.
18610 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18612 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
18613 destination register for bmasksi_vis.
18614 (vector_init_bshuffle): Likewise.
18615 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
18617 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18619 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
18620 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
18621 mode if the instruction isn't available in the original mode.
18622 * config/sparc/sparc.opt (mfix-ut699): New option.
18623 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
18624 (divdf3): Turn into expander.
18625 (divdf3_nofix): New insn.
18626 (divdf3_fix): Likewise.
18627 (divsf3): Disable if -mfix-ut699.
18628 (sqrtdf2): Turn into expander.
18629 (sqrtdf2_nofix): New insn.
18630 (sqrtdf2_fix): Likewise.
18631 (sqrtsf2): Disable if -mfix-ut699.
18633 2013-05-27 Richard Biener <rguenther@suse.de>
18635 PR middle-end/57412
18636 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
18637 block for the new loop.
18639 2013-05-27 Richard Biener <rguenther@suse.de>
18641 PR tree-optimization/57343
18642 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
18643 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
18644 (number_of_iterations_cond): Do not build the folded tree.
18646 2013-05-27 Richard Biener <rguenther@suse.de>
18649 PR middle-end/57381
18650 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
18651 OEP_CONSTANT_ADDRESS_OF retained.
18653 PR tree-optimization/57417
18654 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
18655 for unchanged base.
18656 (set_ssa_val_to): Compare addresses using
18657 get_addr_base_and_unit_offset.
18659 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
18661 PR rtl-optimization/56833
18662 * postreload.c (move2add_record_mode): New function.
18663 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
18664 (move2add_use_add2_insn): Use move2add_record_sym_value.
18665 (move2add_use_add3_insn): Likewise.
18666 (reload_cse_move2add): Use move2add_valid_value_p and
18667 move2add_record_mode. Invalidate call-clobbered and REG_INC
18668 affected regs by setting reg_mode to VOIDmode.
18669 (move2add_note_store): Don't pretend the inside of a SUBREG is
18670 the actual destination. Invalidate single/leading registers by
18671 setting reg_mode to VOIDmode.
18672 Use move2add_record_sym_value, move2add_valid_value_p and
18673 move2add_record_mode.
18675 2013-05-27 Richard Biener <rguenther@suse.de>
18677 PR tree-optimization/57396
18678 * tree-affine.c (double_int_constant_multiple_p): Properly
18679 return false for val == 0 and div != 0.
18681 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18683 * config/mips/mips.h: Use #elif in preprocessor conditions.
18685 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18688 * config/mips/constraints.md (kl): New constraint.
18689 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
18690 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
18691 constraint for operand 0. Split after CSE for MIPS16. Emit a move
18692 from LO for MIPS16.
18693 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
18695 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18698 * config/mips/mips.c (mips_can_inline_p): New function.
18699 (TARGET_CAN_INLINE_P): Define.
18701 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
18703 * sched-int.h (ds_t, dw_t): Make unsigned int.
18704 Fix documentation that describes how all the ds_t bits are used.
18705 Reserve the last bit for delayed-branch scheduling.
18706 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
18707 (BITS_PER_DEP_WEAK): Fix definition and documentation.
18708 (gen_dep_weak_1): Remove prototype.
18709 * sched-deps.c (get_dep_weak_1): Make static.
18710 * target.def (speculate_insn, needs_block_p, gen_spec_check,
18711 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
18712 * doc/tm.texi: Regenerate.
18713 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
18715 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
18718 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
18720 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
18721 Sandra Loosemore <sandra@codesourcery.com>
18723 * config.gcc (powerpc-*): Allow native for with-cpu.
18725 2013-05-24 Jeff Law <law@redhat.com>
18727 PR tree-optimization/57124
18728 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
18729 conversion feeding a condition if the range has an overflow
18730 if -fstrict-overflow. Add warnings for when we do make the
18733 2013-05-24 Dehao Chen <dehao@google.com>
18735 * tree-cfg.c (locus_discrim_map): Fix the typo.
18736 (locus_discrim_hasher): Likewise.
18737 (locus_discrim_hasher::hash): Likewise.
18738 (locus_discrim_hasher::equal): Likewise.
18740 2013-05-24 Martin Jambor <mjambor@suse.cz>
18742 PR tree-optimization/57294
18743 * cgraph.h (ipa_record_stmt_references): Declare.
18744 * cgraphbuild.c (ipa_record_stmt_references): New function.
18745 (build_cgraph_edges): Use ipa_record_stmt_references.
18746 (rebuild_cgraph_edges): Likewise.
18747 (cgraph_rebuild_references): Likewise.
18748 * ipa-prop.c (ipa_modify_call_arguments): Discard references
18749 associated with the old statement and build references from the
18750 newly built statements.
18751 * ipa-ref.c (ipa_remove_stmt_references): New function.
18752 * ipa-ref.h (ipa_remove_stmt_references): Declare.
18754 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
18756 * lra-constraints.c (emit_spill_move): Use smaller mode for
18758 (check_and_process_move): Consider mem-reg moves for secondary
18760 (curr_insn_transform): Don't lose insns emitted before for
18761 secondary memory moves.
18762 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
18763 reg set up in the current insn.
18765 2013-05-24 Dehao Chen <dehao@google.com>
18767 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
18769 (locus_descrim_hasher::equal): Likewise.
18770 (build_gimple_cfg): New discriminator assignment algorithm.
18771 (make_edges): Likewise.
18772 (next_discriminator_for_locus): Likewise.
18773 (same_line_p): Likewise.
18774 (assign_discriminators): Likewise.
18775 (make_cond_expr_edges): Likewise.
18776 (make_gimple_switch_edges): Likewise.
18777 (make_goto_expr_edges): Likewise.
18778 (make_gimple_asm_edges): Likewise.
18780 2013-05-24 Ian Bolton <ian.bolton@arm.com>
18782 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
18783 X format specifier to only display bottom 16 bits.
18784 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
18785 immediate to match for operand 2, since it will be masked.
18787 2013-05-24 Richard Biener <rguenther@suse.de>
18789 PR tree-optimization/57287
18790 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
18791 all SSA names that occur in abnormal PHIs.
18793 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
18796 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
18797 that index is not negative.
18799 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
18801 PR rtl-optimization/55177
18802 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
18803 (simplify_byte_swapping_operation): New.
18804 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
18805 (simplify_relational_operation_1): Deal with BSWAP.
18807 2013-05-23 Richard Henderson <rth@redhat.com>
18810 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
18811 (ix86_reorg): Call it.
18813 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
18816 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
18817 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
18818 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
18820 2013-05-23 Christian Bruel <christian.bruel@st.com>
18823 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
18825 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18826 Marcus Shawcroft <marcus.shawcroft@arm.com>
18828 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
18829 * config/aarch64/constraints.md (Usa): Remove.
18830 * doc/md.texi (AArch64 Usa): Remove.
18832 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18833 Marcus Shawcroft <marcus.shawcroft@arm.com>
18835 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
18836 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
18837 * config/aarch64/predicates.md (aarch64_const_address): Remove.
18838 (aarch64_mov_operand): Use aarch64_mov_operand_p.
18840 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
18842 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
18843 instruction (AdvSIMD).
18844 * config/aarch64/aarch64-builtins.c
18845 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
18846 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
18848 2013-05-23 Martin Jambor <mjambor@suse.cz>
18850 PR middle-end/57347
18851 * tree.h (contains_bitfld_component_ref_p): Declare.
18852 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
18853 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
18855 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
18856 not access a bit-field. Assert all final offsets are byte-aligned.
18858 2013-05-23 Richard Biener <rguenther@suse.de>
18860 PR tree-optimization/57380
18861 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
18862 least one invariant or re-used load.
18863 * passes.c (init_optimization_passes): Move pass_phiprop before
18866 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
18868 * config/aarch64/aarch64-simd.md
18869 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
18871 2013-05-23 Richard Biener <rguenther@suse.de>
18873 PR middle-end/57381
18874 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
18875 OEP_CONSTANT_ADDRESS_OF retained.
18877 2013-05-23 Jakub Jelinek <jakub@redhat.com>
18879 PR middle-end/57344
18880 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
18881 don't lower unit. Handle unit not being always BITS_PER_WORD.
18883 2013-05-23 Richard Biener <rguenther@suse.de>
18885 PR rtl-optimization/57341
18886 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
18887 instead of true_dependence.
18889 2013-05-22 David Malcolm <dmalcolm@redhat.com>
18891 * bb-reorder.c (branch_threshold): Make const.
18892 (exec_threshold): Ditto.
18894 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
18895 Pat Haugen <pthaugen@us.ibm.com>
18896 Peter Bergner <bergner@vnet.ibm.com>
18898 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
18899 documentation for the power8 crypto builtins.
18901 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
18903 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
18904 macros for defining power8 builtin functions.
18905 (BU_P8V_AV_2): Likewise.
18906 (BU_P8V_AV_P): Likewise.
18907 (BU_P8V_VSX_1): Likewise.
18908 (BU_P8V_OVERLOAD_1): Likewise.
18909 (BU_P8V_OVERLOAD_2): Likewise.
18910 (BU_CRYPTO_1): Likewise.
18911 (BU_CRYPTO_2): Likewise.
18912 (BU_CRYPTO_3): Likewise.
18913 (BU_CRYPTO_OVERLOAD_1): Likewise.
18914 (BU_CRYPTO_OVERLOAD_2): Likewise.
18915 (XSCVSPDP): Fix typo, point to the correct instruction.
18916 (VCIPHER): Add power8 crypto builtins.
18917 (VCIPHERLAST): Likewise.
18918 (VNCIPHER): Likewise.
18919 (VNCIPHERLAST): Likewise.
18920 (VPMSUMB): Likewise.
18921 (VPMSUMH): Likewise.
18922 (VPMSUMW): Likewise.
18923 (VPERMXOR_V2DI): Likewise.
18924 (VPERMXOR_V4SI: Likewise.
18925 (VPERMXOR_V8HI: Likewise.
18926 (VPERMXOR_V16QI: Likewise.
18927 (VSHASIGMAW): Likewise.
18928 (VSHASIGMAD): Likewise.
18929 (VPMSUM): Likewise.
18930 (VPERMXOR): Likewise.
18931 (VSHASIGMA): Likewise.
18933 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
18934 __CRYPTO__ if the crypto instructions are available.
18935 (altivec_overloaded_builtins): Add support for overloaded power8
18938 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
18939 support for power8 crypto builtins.
18940 (builtin_function_type): Likewise.
18941 (altivec_init_builtins): Add support for builtins that take vector
18942 long long (V2DI) arguments.
18944 * config/rs6000/crypto.md: New file, define power8 crypto
18947 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
18948 Pat Haugen <pthaugen@us.ibm.com>
18949 Peter Bergner <bergner@vnet.ibm.com>
18951 * doc/invoke.texi (Option Summary): Add power8 options.
18952 (RS/6000 and PowerPC Options): Likewise.
18954 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
18955 constraints.md instead of rs6000.h. Reorder w* constraints. Add
18956 wm, wn, wr documentation.
18958 * config/rs6000/constraints.md (wm): New constraint for VSX
18959 registers if direct move instructions are enabled.
18960 (wn): New constraint for no registers.
18961 (wq): New constraint for quad word even GPR registers.
18962 (wr): New constraint if 64-bit instructions are enabled.
18963 (wv): New constraint if power8 vector instructions are enabled.
18964 (wQ): New constraint for quad word memory locations.
18966 * config/rs6000/predicates.md (const_0_to_15_operand): New
18967 constraint for 0..15 for crypto instructions.
18968 (gpc_reg_operand): If VSX allow registers in VSX registers as well
18969 as GPR and floating point registers.
18970 (int_reg_operand): New predicate to match only GPR registers.
18971 (base_reg_operand): New predicate to match base registers.
18972 (quad_int_reg_operand): New predicate to match even GPR registers
18973 for quad memory operations.
18974 (vsx_reg_or_cint_operand): New predicate to allow vector logical
18975 operations in both GPR and VSX registers.
18976 (quad_memory_operand): New predicate for quad memory operations.
18977 (reg_or_indexed_operand): New predicate for direct move support.
18979 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
18980 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
18981 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
18982 (POWERPC_MASKS): Add power8 options.
18983 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
18986 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
18987 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
18989 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
18990 (-mpower8-fusion): New power8 options.
18991 (-mpower8-fusion-sign): Likewise.
18992 (-mpower8-vector): Likewise.
18993 (-mcrypto): Likewise.
18994 (-mdirect-move): Likewise.
18995 (-mquad-memory): Likewise.
18997 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
18998 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
19000 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
19001 (rs6000_debug_vector_unit): Add p8_vector.
19002 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
19003 definitions. Also print fusion state.
19004 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
19005 (rs6000_builtin_mask_calculate): Add power8 builtin support.
19006 (rs6000_option_override_internal): Add support for power8.
19007 (rs6000_common_init_builtins): Add debugging for skipped builtins
19008 if -mdebug=builtin.
19009 (rs6000_adjust_cost): Add power8 support.
19010 (rs6000_issue_rate): Likewise.
19011 (insn_must_be_first_in_group): Likewise.
19012 (insn_must_be_last_in_group): Likewise.
19013 (force_new_group): Likewise.
19014 (rs6000_register_move_cost): Likewise.
19015 (rs6000_opt_masks): Likewise.
19017 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
19018 power8 capable assembler, default to power7 options.
19019 (TARGET_DIRECT_MOVE): Likewise.
19020 (TARGET_CRYPTO): Likewise.
19021 (TARGET_P8_VECTOR): Likewise.
19022 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
19023 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
19024 (VECTOR_MEM_P8_VECTOR_P): Likewise.
19025 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
19026 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
19027 (TARGET_XSCVDPSPN): Likewise.
19028 (TARGET_XSCVSPDPN): Likewsie.
19029 (TARGET_SYNC_HI_QI): Likewise.
19030 (TARGET_SYNC_TI): Likewise.
19031 (MASK_CRYPTO): Likewise.
19032 (MASK_DIRECT_MOVE): Likewise.
19033 (MASK_P8_FUSION): Likewise.
19034 (MASK_P8_VECTOR): Likewise.
19035 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
19036 temporary used by some of the direct move instructions to get two FP
19037 temporary registers does not force creation of a stack frame.
19038 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
19039 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
19040 that any VSX registers are tieable, even if they are also an
19041 Altivec vector mode.
19042 (r6000_reg_class_enum): Add wm, wr, wv constraints.
19043 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
19044 (RS6000_BTM_CRYPTO): Likewise.
19045 (RS6000_BTM_COMMON): Likewise.
19047 * config/rs6000/rs6000.md (cpu attribute): Add power8.
19048 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
19049 (enum rs6000_vector): Add power8 vector support.
19051 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19055 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
19056 (any_sibcall_could_use_r3): this and handle indirect calls.
19057 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
19059 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19061 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
19063 2013-05-22 Richard Biener <rguenther@suse.de>
19065 PR middle-end/57349
19066 * profile.c (branch_prob): Do not split blocks that are
19067 abnormally receiving from ECF_RETURNS_TWICE functions.
19069 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19071 * recog.c (offsettable_address_addr_space_p): Fix calculation of
19072 address mode. Move pointer mode initialization to the same place.
19074 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
19076 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
19077 while it has any effect.
19079 2013-05-21 Easwaran Raman <eraman@google.com>
19081 PR tree-optimization/57322
19082 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
19083 UID of the statement added to the BB to be 1.
19085 2013-05-21 Jakub Jelinek <jakub@redhat.com>
19087 PR tree-optimization/57331
19088 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
19089 of conversion from pointer type to integral type with integer.
19091 2013-05-21 Martin Jambor <mjambor@suse.cz>
19094 * ipa-prop.c (ipa_read_node_info): Process param_used and
19095 controlled_uses in the same order as when writing.
19097 2013-05-21 Magnus Granberg <baldrick@free.fr>
19100 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
19102 2013-05-21 Richard Biener <rguenther@suse.de>
19104 PR tree-optimization/57318
19105 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
19106 estimate stmts with side-effects as likely eliminated.
19108 2013-05-21 Richard Biener <rguenther@suse.de>
19110 PR tree-optimization/57330
19111 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
19112 preserve the call stmts fntype.
19114 2013-05-21 Richard Biener <rguenther@suse.de>
19116 PR tree-optimization/57303
19117 * tree-ssa-sink.c (statement_sink_location): Improve killing
19118 stmt detection and properly handle self-assignments.
19120 2013-05-21 Christian Bruel <christian.bruel@st.com>
19122 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
19123 spanning registers. LEAF_REG_REMAP is supported only for contiguous
19124 registers. Set register size out of the PARALLEL loop.
19126 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
19129 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
19130 (*fmasf4, *fmasf4_media): New insns.
19132 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19134 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
19135 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
19136 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
19137 (mips_idiv_insns): Update the comments to say that the returned
19138 instruction counts are in units of BASE_INSN_LENGTH.
19139 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
19140 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
19141 using 2 rather than 4 as the length of indirect MIPS16 and
19142 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
19143 length of a NOP. Don't divide MIPS16 lengths by 2.
19144 (mips16_split_long_branches): Assume a branch is long if the
19145 length is greater than 4 rather than 8.
19146 * config/mips/mips.md (length): Give MIPS16 lengths directly,
19147 rather than multiplying them by 2. Multiply instruction counts
19148 by BASE_INSN_LENGTH rather than 4.
19149 (*jump_mips16, tls_get_tp_mips16_<mode>)
19150 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
19152 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19154 * config/mips/mips.md (extended_mips16): Remove branch case.
19155 (length): Remove duplicated extended_mips16 test.
19157 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19159 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
19161 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
19163 * recog.h (Recog_data): Rename to...
19164 (recog_data_d): ...this.
19165 (recog_data): Update accordingly.
19166 * recog.c (recog_data): Likewise.
19167 * reload.c (save_recog_data): Likewise.
19168 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
19169 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
19171 2013-05-17 Julian Brown <julian@codesourcery.com>
19173 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
19174 found in a REG_EQUAL note, invalidate it.
19176 2013-05-17 Easwaran Raman <eraman@google.com>
19178 * tree-ssa-reassoc.c (find_insert_point): New function.
19179 (insert_stmt_after): Likewise.
19180 (get_def_stmt): Likewise.
19181 (ensure_ops_are_available): Likewise.
19182 (not_dominated_by): Likewise.
19183 (rewrite_expr_tree): Do not move statements beyond what is
19184 necessary. Remove call to swap_ops_for_binary_stmt...
19185 (reassociate_bb): ... and move it here.
19186 (build_and_add_sum): Assign UIDs for new statements.
19187 (linearize_expr): Likewise.
19188 (do_reassoc): Renumber gimple statement UIDs.
19190 2013-05-17 Jan Hubicka <jh@suse.cz>
19192 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
19194 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
19195 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
19197 (output_weakrefs): Update.
19199 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
19200 Martin Jambor <mjambor@suse.cz>
19202 PR middle-end/57276
19203 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
19204 value that corresponds to the given aggval is found in values vector.
19206 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
19208 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
19209 sse, sse2, sse3, ssse3 and sse4a flags to options.
19211 2013-05-17 David Malcolm <dmalcolm@redhat.com>
19213 * gengtype-state.c: (s_expr_writer): New class, to handle
19214 prettifying of output layout of s-expressions.
19215 (state_writer): New class, to write out gtype.state.
19216 (state_written_type_count): Move this variable into member data of
19218 (s_expr_writer::s_expr_writer): New code: constructor for new class
19219 (state_writer::state_writer(): ditto
19220 (s_expr_writer::write_new_line): New function
19221 (s_expr_writer::write_any_indent): ditto
19222 (s_expr_writer::begin_s_expr): ditto
19223 (s_expr_writer::end_s_expr): ditto
19224 (write_state_fileloc): convert to method of state_writer...
19225 (state_writer:: write_state_fileloc): ...and use methods of
19226 s_expr_writer to write indentation into the gtype.state output file
19227 to visually represent the hierarchical structure of the list
19229 (write_state_fields): ditto, renaming to...
19230 (state_writer::write_state_fields)
19231 (write_state_a_string): ditto, renaming to...
19232 (state_writer::write_state_a_string)
19233 (write_state_string_option): ditto, renaming to...
19234 (state_writer::write_state_string_option)
19235 (write_state_type_option): ditto, renaming to...
19236 (state_writer::write_state_type_option)
19237 (write_state_nested_option): ditto, renaming to...
19238 (state_writer::write_state_nested_option)
19239 (write_state_option): ditto, renaming to...
19240 (state_writer::write_state_option)
19241 (write_state_options): ditto, renaming to...
19242 (state_writer::write_state_options)
19243 (write_state_lang_bitmap): ditto, renaming to...
19244 (state_writer::write_state_lang_bitmap)
19245 (write_state_version): ditto, renaming to...
19246 (state_writer::write_state_version)
19247 (write_state_scalar_type): ditto, renaming to...
19248 (state_writer::write_state_scalar_type)
19249 (write_state_string_type): ditto, renaming to...
19250 (state_writer::write_state_string_type)
19251 (write_state_undefined_type): ditto, renaming to...
19252 (state_writer::write_state_undefined_type)
19253 (write_state_struct_union_type): ditto, renaming to...
19254 (state_writer::write_state_struct_union_type)
19255 (write_state_struct_type): ditto, renaming to...
19256 (state_writer::write_state_struct_type)
19257 (write_state_user_struct_type): ditto, renaming to...
19258 (state_writer::write_state_user_struct_type)
19259 (write_state_lang_struct_type): ditto, renaming to...
19260 (state_writer::write_state_lang_struct_type)
19261 (write_state_param_struct_type): ditto, renaming to...
19262 (state_writer::write_state_param_struct_type)
19263 (write_state_pointer_type): ditto, renaming to...
19264 (state_writer::write_state_pointer_type)
19265 (write_state_array_type): ditto, renaming to...
19266 (state_writer::write_state_array_type)
19267 (write_state_gc_used): ditto, renaming to...
19268 (state_writer::write_state_gc_used)
19269 (write_state_common_type_content): ditto, renaming to...
19270 (state_writer::write_state_common_type_content)
19271 (write_state_type): ditto, renaming to...
19272 (state_writer::write_state_type)
19273 (write_state_pair_list): ditto, renaming to...
19274 (state_writer::write_state_pair_list)
19275 (write_state_pair): ditto, renaming to...
19276 (state_writer::write_state_pair)
19277 (write_state_typedefs): ditto, renaming to...
19278 (state_writer::write_state_typedefs)
19279 (write_state_structures): ditto, renaming to...
19280 (state_writer::write_state_structures)
19281 (write_state_param_structs): ditto, renaming to...
19282 (state_writer::write_state_param_structs)
19283 (write_state_variables): ditto, renaming to...
19284 (state_writer::write_state_variables)
19285 (write_state_srcdir): ditto, renaming to...
19286 (state_writer::write_state_srcdir)
19287 (write_state_files_list): ditto, renaming to...
19288 (state_writer::write_state_files_list)
19289 (write_state_languages): ditto, renaming to...
19290 (state_writer::write_state_languages)
19291 (write_state): create a state_writer instance and use it when
19292 writing out the state file
19294 2013-05-17 Mike Stump <mikestump@comcast.net>
19296 PR rtl-optimization/57304
19297 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
19298 accessing DF_REF_REAL_LOC.
19300 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19302 PR rtl-optimization/57281
19303 PR rtl-optimization/57300
19304 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
19305 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
19306 what the other splitter did if the registers are dead.
19308 2013-05-17 Richard Biener <rguenther@suse.de>
19310 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
19313 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19315 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
19318 2013-05-17 Marek Polacek <polacek@redhat.com>
19320 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
19321 length when doing non-zero store of storing '\0' to '\0'.
19323 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19325 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
19326 vect_external_def oprnd1 with loop_vinfo, try to emit
19327 optional cast, negation and and stmts on the loop preheader
19328 edge instead of into the pattern def seq.
19330 PR tree-optimization/57051
19331 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
19332 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
19334 2013-05-16 Nick Clifton <nickc@redhat.com>
19336 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
19337 (rl78_is_naked_func): New function.
19338 (rl78_expand_prologue): Skip prologue generation for naked functions.
19339 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
19340 * doc/extend.texi (naked): Add RL78 to the list of processors
19341 that supports this attribute.
19343 2013-05-16 Jeff Law <law@redhat.com>
19345 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
19347 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19349 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
19350 cache parameters using detect_caches_amd also for CYRIX,
19351 NSC and TM2 signatures.
19353 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19354 Dzianis Kahanovich <mahatma@eu.by>
19358 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
19359 VIA/Centaur processors and determine their cache parameters
19360 using detect_caches_amd.
19362 2013-05-16 Teresa Johnson <tejohnson@google.com>
19364 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
19365 (rtl_verify_edges): New function.
19366 (rtl_verify_bb_insns): Ditto.
19367 (rtl_verify_bb_pointers): Ditto.
19368 (rtl_verify_bb_insn_chain): Ditto.
19369 (rtl_verify_fallthru): Ditto.
19370 (rtl_verify_bb_layout): Ditto.
19371 (rtl_verify_flow_info_1): Outline checks into new functions.
19372 (rtl_verify_flow_info): Ditto.
19374 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
19376 * cfghooks.c (copy_bbs): Add update_dominance argument.
19377 * cfghooks.h (copy_bbs): Update prototype.
19378 * tree-cfg.c (gimple_duplicate_sese_region):
19379 Add update_dominance argument.
19380 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
19381 * tree-ssa-loop-ch.c (copy_loop_headers): Update
19382 gimple_duplicate_sese_region call.
19383 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
19384 Update copy_bbs call.
19385 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
19386 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
19388 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19390 * tree-vectorizer.h (NUM_PATTERNS): Increment.
19391 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
19392 vect_recog_rotate_pattern.
19393 (vect_recog_rotate_pattern): New function.
19395 2013-05-16 Jason Merrill <jason@redhat.com>
19397 * Makefile.in (LLINKER): New variable.
19398 (mostlyclean): Remove link mutex.
19399 * configure.ac: Handle --enable-link-mutex.
19400 * lock-and-run.sh: New script.
19402 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19405 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
19408 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19410 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
19412 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
19414 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
19415 * config/arm/arm.c (next_consecutive_mem): New function.
19416 (gen_movmem_ldrd_strd): Likewise.
19417 * config/arm/arm.md (movmemqi): Update condition and code.
19418 (unaligned_loaddi, unaligned_storedi): New patterns.
19420 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19422 * config.gcc: Obsolete *-*-solaris2.9*.
19423 * doc/install.texi (Specific, *-*-solaris2*): Document it.
19425 2013-05-16 Richard Biener <rguenther@suse.de>
19427 * passes.c (init_optimization_passes): Move pass_parallelize_loops
19428 earlier, after GRAPHITE transforms and IV canonicalization.
19430 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19432 * omp-low.c (extract_omp_for_data): For collapsed loops,
19433 if at least one of the loops is known at compile time to
19434 iterate zero times, set count to 0.
19435 (expand_omp_regimplify_p): New function.
19436 (expand_omp_for_generic): For collapsed loops, if at least
19437 one of the loops isn't known to iterate at least once,
19438 add runtime check with setting count to 0.
19439 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
19440 For unsigned types if it isn't known at compile time that
19441 the loop will iterate at least once, add runtime check to bypass
19442 the whole loop if initial condition isn't true.
19444 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
19446 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
19448 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
19450 PR middle-end/57286
19451 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
19452 transformations to avoid an infinite loop.
19454 2013-05-16 Marek Polacek <polacek@redhat.com>
19456 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
19458 2013-05-15 Leif Ekblad <leif@rdos.net>
19460 * config/i386/i386.c (ix86_decompose_address): Use
19461 DEFAULT_TLS_SEG_REG to access TLS segment register.
19462 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
19463 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
19464 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
19466 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
19469 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
19470 sibling calls to functions that would normally be lazily bound,
19471 unless $gp is call-clobbered.
19473 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
19475 * config/i386/i386.c (ix86_option_override_internal): Update
19476 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
19477 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
19478 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
19479 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
19481 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
19483 2013-05-15 Andreas Schwab <schwab@suse.de>
19485 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
19486 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
19489 2013-05-15 Teresa Johnson <tejohnson@google.com>
19491 * loop-unroll.c (report_unroll_peel): Check decision before
19492 emitting unroll/peel message.
19494 2013-05-15 Teresa Johnson <tejohnson@google.com>
19496 * function.h (has_bb_partition): New rtl_data flag.
19497 (bb_reorder_complete): Ditto.
19498 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
19499 instead of flag_reorder_blocks_and_partition.
19500 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
19501 with some enhancements.
19502 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
19503 * bb-reorder.c (connect_traces): Check for has_bb_partition
19504 instead of flag_reorder_blocks_and_partition.
19505 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
19506 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
19507 verify_hot_cold_block_grouping.
19508 (partition_hot_cold_basic_blocks): Set has_bb_partition.
19510 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19513 * config/arm/predicates.md (call_insn_operand): New predicate.
19514 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
19515 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
19516 if insn is not a tail call.
19517 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
19519 * config/arm/arm.h (enum reg_class): New caller save register class.
19520 (REG_CLASS_NAMES): Likewise.
19521 (REG_CLASS_CONTENTS): Likewise.
19522 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
19525 2013-05-15 Richard Biener <rguenther@suse.de>
19527 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
19528 of MSG_OPTIMIZED_LOCATIONS.
19529 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
19530 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
19532 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
19533 of MSG_OPTIMIZED_LOCATIONS.
19534 (execute_vect_slp): Likewise.
19535 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
19536 (vect_create_cond_for_alias_checks): Likewise.
19537 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
19538 (vect_recog_widen_mult_pattern): Likewise.
19539 (vect_recog_widen_sum_pattern): Likewise.
19540 (vect_recog_over_widening_pattern): Likewise.
19541 (vect_recog_widen_shift_pattern): Likewise.
19542 (vect_recog_vector_vector_shift_pattern): Likewise.
19543 (vect_recog_divmod_pattern): Likewise.
19544 (vect_recog_mixed_size_cond_pattern): Likewise.
19545 (vect_recog_bool_pattern): Likewise.
19546 (vect_pattern_recog_1): Likewise.
19548 2013-05-15 Martin Jambor <mjambor@suse.cz>
19550 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
19551 non-functions to builtin_unreachable.
19552 * ipa-inline-transform.c (inline_call): Do not assert estimates were
19553 correct when new direct edges were discovered.
19555 2013-05-15 Martin Jambor <mjambor@suse.cz>
19557 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
19558 header, print symbol order instead of node uid, print more information
19559 about indirect edge targets.
19560 (ipa_make_edge_direct_to_target): Print symbol order instead of node
19562 (ipa_make_edge_direct_to_target): Likewise.
19563 (remove_described_reference): Likewise.
19564 (propagate_controlled_uses): Likewise.
19565 (ipa_print_node_params): Also print symbol order.
19566 (ipcp_transform_function): Print symbol order instead of node uids.
19567 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
19568 (cgraph_get_create_real_symbol_node): Likewise.
19569 * ipa-cp.c (print_lattice): Likewise.
19570 (print_all_lattices): Likewise.
19571 (determine_versionability): Likewise.
19572 (initialize_node_lattices): Likewise.
19573 (estimate_local_effects): Likewise.
19574 (update_profiling_info): Likewise.
19575 (create_specialized_node): Likewise.
19576 (perhaps_add_new_callers): Likewise.
19577 (decide_about_value): Likewise.
19578 (decide_whether_version_node): Likewise.
19579 (identify_dead_nodes): Likewise.
19580 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
19581 (dump_inline_summary): Likewise.
19582 (estimate_node_size_and_time): Likewise.
19583 (inline_analyze_function): Likewise.
19584 * ipa-inline.c (report_inline_failed_reason): Likewise.
19585 (want_early_inline_function_p): Likewise.
19586 (edge_badness): Likewise.
19587 (update_edge_key): Likewise.
19588 (inline_small_functions): Likewise. Add dumping of order to two other
19590 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
19591 instead of node uids.
19592 (propagate_pure_const): Likewise.
19593 (propagate_pure_const): Likewise.
19594 * ipa-utils.c (dump_cgraph_node_set): Likewise.
19595 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
19596 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
19598 * tree-pretty-print.c (dump_function_header): Likewise.
19599 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
19600 Print symbol order instead of node uids.
19602 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19604 * config/s390/s390.c (s390_register_move_cost): Don't impose the
19605 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
19607 2013-05-15 Richard Biener <rguenther@suse.de>
19609 PR tree-optimization/57275
19610 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
19611 return value for fail to do runtime alias checks for gather loads.
19613 2013-05-15 Jan Hubicka <jh@suse.cz>
19617 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
19618 weakrefs are not external.
19619 (lto_symtab_merge_decls): Fix thinko when dealing with
19620 non-lto_symtab decls.
19621 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
19622 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
19623 * varpool.c (dump_varpool_node): Dump more flags.
19625 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19627 * config/i386/i386.c (processor_alias_table): Add instruction
19628 FSGSBASE for AMD bdver3 architecture.
19630 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19632 * tree.c (warn_deprecated_use): Print file:line using locus color.
19633 * diagnostic.c (diagnostic_report_current_module): Print file:line
19634 and file:line:column using locus color.
19636 2013-05-14 Mike Stump <mikestump@comcast.net>
19638 * gdbinit.in: Add __null.
19640 2013-05-14 Mike Stump <mikestump@comcast.net>
19642 * recog.h: Rename struct recog_data to Recog_data.
19643 * recog.c: Likewise.
19644 * reload.c (can_reload_into): Likewise.
19645 * config/picochip/picochip.c: Likewise.
19647 2013-05-14 Mike Stump <mikestump@comcast.net>
19649 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
19651 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
19653 * resource.h (struct resources): Remove unch_memory member.
19654 (CLEAR_RESOURCE): Don't clear unch_memory.
19655 * resource.c (mark_referenced_resources): Don't set it.
19656 (mark_set_resources): Likewise.
19657 (mark_target_live_regs): Don't clear it.
19658 (init_resource_info): Likewise.
19659 * reorg.c (resource_conflicts_p): Don't compare it.
19660 (redundant_insn): Don't set it.
19662 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
19664 * emit-rtl.c (next_label): Remove unused function.
19665 (skip_consecutive_labels, link_cc0_insns): Move to ...
19666 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
19667 only place where these functions are used, and make them static.
19669 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
19671 * fold-const.c (fold_negate_expr): Handle vectors.
19672 (fold_truth_not_expr): Make it static.
19673 (fold_invert_truthvalue): New static function.
19674 (invert_truthvalue_loc): Handle vectors. Do not call
19675 fold_truth_not_expr directly.
19676 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
19677 <TRUTH_NOT_EXPR>: Do not cast to boolean.
19678 (fold_comparison): Handle vector constants.
19679 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
19680 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
19681 * tree.h (fold_truth_not_expr): Remove declaration.
19683 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
19685 * config/aarch64/aarch64-simd.md
19686 (aarch64_vcond_internal<mode>): Rename to...
19687 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
19688 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
19689 float modes. Clarify all iterator modes.
19690 (vcond<mode><mode>): Use new name for vcond expanders.
19691 (vcond<v_cmp_result><mode>): Likewise.
19692 (vcondu<mode><mode>: Likewise.
19693 * config/aarch64/iterators.md (VDQF_COND): New.
19695 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
19698 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
19699 variable for the shift amount. Check that we shift by non-negative
19702 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
19705 * config/arm/arm.h (EPILOGUE_USES): Only return true
19706 for LR_REGNUM after epilogue_completed.
19708 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
19710 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
19711 is error_mark_node.
19713 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19716 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
19717 and Solaris 11+/x86 with gld.
19718 * configure: Regenerate.
19720 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19722 * expmed.c (expand_shift_1): Canonicalize rotates by
19723 constant bitsize / 2 to bitsize - 1.
19724 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
19725 case ROTATERT>: Likewise.
19728 2013-05-10 Jakub Jelinek <jakub@redhat.com>
19730 * config/i386/i386.md (rotateinv): New code attr.
19731 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
19732 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
19733 roll $31, %eax, etc.
19735 2013-05-14 Richard Biener <rguenther@suse.de>
19737 PR middle-end/57235
19738 * tree-eh.c (sink_clobbers): Give up for successors with
19739 multiple predecessors and no virtual uses.
19741 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
19743 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
19744 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
19746 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19748 PR middle-end/57251
19749 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
19750 the case when both op0 and op1 have VOIDmode.
19752 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
19754 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
19755 in multiply-accumulate mode.
19757 2013-05-13 Guozhi Wei <carrot@google.com>
19759 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
19761 2013-05-13 Kai Tietz <ktietz@redhat.com>
19764 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
19765 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
19766 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
19767 * config/i386/i386.c (ix86_option_override_internal): Likewise.
19768 (ix86_expand_prologue): Likewise.
19769 (ix86_expand_split_stack_prologue): Likewise.
19770 (legitimate_pic_address_disp_p): Likewise.
19771 (legitimize_pic_address): Likewise.
19772 (legitimize_tls_address): Likewise.
19773 (legitimize_pe_coff_symbol): Likewise.
19774 (output_pic_addr_const): Likewise.
19775 (construct_plt_address): Likewise.
19776 (ix86_expand_call): Likewise.
19777 (x86_output_mi_thunk): Likewise.
19778 (x86_function_profiler): Likewise.
19780 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
19782 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
19783 similar switch cases.
19784 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
19785 (aarch64_simd_mov_to_<mode>low): Delete.
19786 (aarch64_simd_mov_to_<mode>high): Delete.
19787 (move_lo_quad_<mode>): Add w<-r alternative.
19788 (aarch64_simd_move_hi_quad_<mode>): Likewise.
19789 (aarch64_simd_mov_from_*): Update type attribute.
19790 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
19793 2013-05-13 Jan Hubicka <jh@suse.cz>
19795 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
19796 * config/i386/i386.c (ix86_compute_frame_layout,
19797 ix86_expand_epilogue, emit_i387_cw_initialization,
19798 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
19799 ix86_local_alignment): Fix use of size/speed predicates.
19801 2013-05-13 Jakub Jelinek <jakub@redhat.com>
19803 PR tree-optimization/45216
19804 PR tree-optimization/57157
19805 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
19806 the (-Y) & (B - 1) variant if OP is |.
19807 * expmed.c (expand_shift_1): For rotations by const0_rtx just
19808 return shifted. Use (-op1) & (prec - 1) as other_amount
19809 instead of prec - op1.
19811 2013-05-13 Martin Jambor <mjambor@suse.cz>
19813 PR middle-end/42371
19814 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
19815 (ipa_constant_data): New type.
19816 (ipa_jump_func): Use ipa_constant_data to hold information about
19817 constant jump functions.
19818 (ipa_get_jf_constant): Adjust to jump function type changes.
19819 (ipa_get_jf_constant_rdesc): New function.
19820 (ipa_param_descriptor): New field controlled_uses.
19821 (ipa_get_controlled_uses): New function.
19822 (ipa_set_controlled_uses): Likewise.
19823 * ipa-ref.h (ipa_find_reference): Declare.
19824 * ipa-prop.c (ipa_cst_ref_desc): New type.
19825 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
19827 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
19828 New parameter cs. Adjust all callers.
19829 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
19830 (remove_described_reference): New function.
19831 (jfunc_rdesc_usable): Likewise.
19832 (try_make_edge_direct_simple_call): Decrement controlled use count,
19833 attempt to remove reference if it hits zero.
19834 (combine_controlled_uses_counters): New function.
19835 (propagate_controlled_uses): Likewise.
19836 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
19837 (ipa_edge_duplication_hook): Duplicate reference descriptions.
19838 (ipa_print_node_params): Print described use counter.
19839 (ipa_write_jump_function): Adjust to jump function type changes.
19840 (ipa_read_jump_function): New parameter CS, pass it to
19841 ipa_set_jf_constant. Adjust caller.
19842 (ipa_write_node_info): Stream controlled use count
19843 (ipa_read_node_info): Likewise.
19844 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
19846 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
19847 count. Remove cloning-added reference if it reaches zero.
19848 * ipa-ref.c (ipa_find_reference): New function.
19850 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19852 * config/i386/i386.c (processor_target_table): Modified default
19853 alignment values for AMD BD and BT architectures.
19855 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
19857 * tree-vect-generic.c (uniform_vector_p): Move ...
19858 * tree.c (uniform_vector_p): ... here.
19859 * tree.h (uniform_vector_p): Declare it.
19860 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
19863 2013-05-13 Jakub Jelinek <jakub@redhat.com>
19865 PR tree-optimization/57230
19866 * tree-ssa-strlen.c (handle_char_store): Record length for
19867 array store from STRING_CST.
19869 PR tree-optimization/57230
19870 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
19873 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
19875 * config/epiphany/epiphany.c (epiphany_init): Check size of
19876 NUM_MODES_FOR_MODE_SWITCHING.
19877 (epiphany_expand_prologue):
19878 Remove CONFIG_REGNUM initial value handling code.
19879 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
19880 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
19881 (emit_set_fp_mode, epiphany_mode_after): Likewise.
19882 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
19883 Don't return 1 for FP_MODE_NONE.
19884 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
19885 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
19886 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
19887 * config/epiphany/epiphany.md (save_config): New pattern.
19889 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
19891 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
19893 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
19895 * config/i386/i386.md (memory): Handle sseishft1.
19896 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
19897 (*vec_extractv2di_1): Ditto.
19899 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
19901 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
19904 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
19906 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
19907 Add mthumb/march=armv7-a multilib.
19908 Add mthumb/march=armv7-r multilib.
19909 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
19911 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
19913 * config/v850/t-rtems: Add more multilibs.
19915 2013-05-10 Richard Biener <rguenther@suse.de>
19917 PR tree-optimization/57214
19918 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
19919 not propagate from SSA names that occur in abnormal PHI nodes.
19921 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
19923 * stor-layout.c (element_precision): New function.
19924 * machmode.h (element_precision): Declare it.
19925 * tree.c (build_minus_one_cst): New function.
19926 (element_precision): Likewise.
19927 * tree.h (build_minus_one_cst): Declare new function.
19928 (element_precision): Likewise.
19929 * fold-const.c (operand_equal_p): Use element_precision.
19930 (fold_binary_loc): Handle vector types.
19931 * convert.c (convert_to_integer): Use element_precision.
19932 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
19935 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
19937 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
19938 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
19939 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
19940 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
19941 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
19942 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
19943 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
19944 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
19945 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
19946 (Uuw8): New constraints.
19947 (Usb4): Move into alphabetical order.
19948 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
19949 (sd8_operand, ub8_operand, uw8_operand): New predicates.
19950 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
19951 previously unnamed patterns.
19952 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
19953 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
19954 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
19955 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
19956 of set_attr_alternative/if_then_else. Use extended_mips16 instead
19957 of specific lengths.
19959 2013-05-10 Jakub Jelinek <jakub@redhat.com>
19961 * config/i386/i386.md (rotateinv): New code attr.
19962 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
19963 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
19964 roll $31, %eax, etc.
19966 PR tree-optimization/45216
19967 PR tree-optimization/57157
19968 * tree-ssa-forwprop.c (simplify_rotate): New function.
19969 (ssa_forward_propagate_and_combine): Call it.
19971 2013-05-10 Richard Biener <rguenther@suse.de>
19973 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
19974 disable peeling when we version for aliasing.
19975 (vector_alignment_reachable_p): Honor explicit user alignment.
19976 (vect_supportable_dr_alignment): Likewise.
19977 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
19978 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
19979 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
19980 then peeling to arrange for the cost-model check to come first.
19982 2013-05-10 Alan Modra <amodra@gmail.com>
19984 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
19985 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
19986 * configure: Regenerate.
19988 2013-05-10 Alan Modra <amodra@gmail.com>
19991 * varasm.c (default_elf_select_section): Move !DECL_P check..
19992 (get_named_section): ..to here before calling get_section_name.
19994 (default_section_type_flags): Add DECL_P check.
19995 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
19996 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
19998 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
20000 * config/epiphany/epiphany.c (epiphany_expand_prologue):
20001 When using gen_stack_adjust_str with a register offset, add a
20002 REG_FRAME_RELATED_EXPR note.
20004 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
20006 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
20007 (*vec_extractv4si_zext_mem): Ditto.
20008 (*vec_extractv2di): Add 0->x and x->x alternatives.
20009 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
20010 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
20012 2013-05-09 Jason Merrill <jason@redhat.com>
20014 N3639 C++1y VLA support
20015 * gimplify.c (gimplify_vla_decl): Don't touch an existing
20018 * tree.c (build_constructor_va): New.
20019 * tree.h: Declare it.
20021 2013-05-09 Martin Jambor <mjambor@suse.cz>
20024 * gimple-fold.c (canonicalize_constructor_val): Call
20025 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
20027 2013-05-09 Jan Hubicka <jh@suse.cz>
20028 Richard Biener <rguenther@suse.de>
20031 * symtab.c (symtab_make_decl_local): Do not add private names.
20033 2013-05-09 Jan Hubicka <jh@suse.cz>
20036 * symtab.c (insert_to_assembler_name_hash): Handle clones.
20037 (unlink_from_assembler_name_hash): Likewise.
20038 (symtab_prevail_in_asm_name_hash, symtab_register_node,
20039 symtab_unregister_node, symtab_initialize_asm_name_hash,
20040 change_decl_assembler_name): Update.
20042 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
20044 * config/aarch64/aarch64.md: New movtf split.
20045 (*movtf_aarch64): Update.
20046 (aarch64_movdi_tilow): Handle TF modes and rename to
20047 aarch64_movdi_<mode>low.
20048 (aarch64_movdi_tihigh): Handle TF modes and rename to
20049 aarch64_movdi_<mode>high
20050 (aarch64_movtihigh_di): Handle TF modes and rename to
20051 aarch64_mov<mode>high_di
20052 (aarch64_movtilow_di): Handle TF modes and rename to
20053 aarch64_mov<mode>low_di
20054 (aarch64_movtilow_tilow): Remove spurious whitespace.
20055 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
20057 (aarch64_print_operand): Update.
20059 2013-05-09 Alan Modra <amodra@gmail.com>
20061 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
20063 * configure: Regenerate.
20065 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20067 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
20068 splitter preparation statements.
20069 * config/i386/sse.md (*vec_extract* splitters): Ditto.
20070 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
20073 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20075 * gimple-ssa-strength-reduction.c (count_candidates): Change
20076 return value to int.
20077 (analyze_candidates_and_replace): Change type of length to int.
20079 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20081 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
20082 (*vec_extract<mode>): Use VI12_128 mode iterator.
20083 (*vec_extract<mode>_mem): Ditto.
20084 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
20087 2013-05-08 Diego Novillo <dnovillo@google.com>
20092 2012-08-17 Diego Novillo <dnovillo@google.com>
20095 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
20096 * config.in: Regenerate.
20097 * configure: Regenerate.
20098 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
20100 2013-05-08 Jan Hubicka <jh@suse.cz>
20103 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
20104 * cgraph.h (symtab_node_base): Add unique_name.
20105 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20106 input_overwrite_node, input_varpool_node): Stream unique_name.
20107 * cgraphclones.c (cgraph_create_virtual_clone,
20108 cgraph_function_versioning): Set unique_name.
20109 * ipa.c (function_and_variable_visibility): Set unique_name.
20111 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20113 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
20114 (alloc_cand_and_find_basis): Restrict conditional candidate
20115 processing to CAND_MULTs.
20117 2013-05-08 Jan Hubicka <jh@suse.cz>
20120 lto-symtab.c (lto_symtab_symbol_p): New function.
20121 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
20122 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
20123 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
20124 Skip static symbols.
20126 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
20128 PR tree-optimization/57200
20129 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
20130 Only call inform if the preceding warning_at returns true.
20132 2013-05-07 Han Shen <shenhan@google.com>
20134 * cfgexpand.c (record_or_union_type_has_array_p): New function.
20135 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
20136 * common.opt (fstack-protector-strong): New option.
20137 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
20138 * doc/invoke.texi (Optimization Options): Document
20139 "-fstack-protector-strong".
20140 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
20142 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20144 * config/mips/mips.c (mips_machine_reorg2): Return 0.
20146 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
20148 * ira.c (update_equiv_regs): Add insn having equiv memory even if
20149 it is not lhs of the insn.
20150 (setup_reg_equiv): Remove insn having equiv memory which it is not
20152 * lra-constraints.c (process_address): Try to improve generation
20153 code for address base + disp.
20154 (lra_constraints): Make correct the code for checking insn setting
20155 up backward equivalence. Remove insn only if it is in the init
20157 * lra-eliminations.c (update_reg_eliminate): Change return value.
20158 (lra_eliminate): Use the result.
20160 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
20162 * config/i386/sse.md (ssescalarnummask): New mode attribute.
20163 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
20164 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
20165 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
20166 register target operands.
20167 (*vec_extractv8hi_sse2): New pattern.
20168 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
20169 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
20170 (*vec_extract<mode>_mem): New insn and split pattern.
20172 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
20174 * config/arm/arm.c (arm_asan_shadow_offset): New function.
20175 (TARGET_ASAN_SHADOW_OFFSET): Define.
20176 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
20177 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
20179 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20181 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
20182 (incr_vec_index): Return -1 if increment not found.
20183 (create_add_on_incoming_edge): Assert if increment not found.
20184 (record_increment): Limit number of increments recorded.
20185 (all_phi_incrs_profitable): Return false if an increment not found.
20186 (replace_profitable_candidates): Don't process increments that were
20188 (analyze_candidates_and_replace): Limit size of incr_vec.
20190 2013-05-07 Richard Biener <rguenther@suse.de>
20192 * calls.c (special_function_p): setjmp-like functions are leaf.
20193 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
20194 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
20196 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
20198 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
20199 (aarch64_simd_mov<mode>): New expander.
20200 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
20201 (aarch64_simd_mov_to_<mode>high): Likewise.
20202 (aarch64_simd_mov_from_<mode>low): Likewise.
20203 (aarch64_simd_mov_from_<mode>high): Likewise.
20204 (aarch64_dup_lane<mode>): Update.
20205 (aarch64_dup_lanedi): New instruction pattern.
20206 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
20207 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
20209 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20211 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
20212 (replace_mult_candidate): Remove unnecessary argument; remove
20213 unnecessary parameter from call to introduce_cast_before_cand.
20214 (replace_unconditional_candidate): Remove unnecessary parameter
20215 from call to replace_mult_candidate.
20216 (replace_conditional_candidate): Likewise.
20217 (insert_initializers): Use make_temp_ssa_name.
20218 (introduce_cast_before_cand): Remove unnecessary argument; use
20219 make_temp_ssa_name.
20220 (replace_one_candidate): Remove unnecessary argument; remove
20221 unnecessary parameter from calls to introduce_cast_before_cand.
20222 (replace_profitable_candidates): Remove unnecessary parameters
20223 from calls to replace_one_candidate.
20225 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20227 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
20228 phi def as possibly hiding a basis for a CAND_ADD whose operands
20229 have been commuted in the analysis.
20230 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
20232 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20234 * config/aarch64/aarch64.md
20235 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
20236 shift value between 0-4.
20238 2013-05-07 Richard Biener <rguenther@suse.de>
20240 * double-int.h (rshift): New overload.
20241 * double-int.c (rshift): New function.
20242 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
20243 (create_reference_ops_from_ref): Remove.
20244 (vn_reference_insert): Use shared ops for constructing the
20245 reference and copy it.
20247 2013-05-07 Richard Biener <rguenther@suse.de>
20249 PR middle-end/57190
20250 * tree-eh.c (sink_clobbers): Properly propagate
20251 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
20253 2013-05-07 Jakub Jelinek <jakub@redhat.com>
20255 PR tree-optimization/57149
20256 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
20257 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
20258 collect_phi_def_edges, execute_late_warn_uninitialized): Use
20259 uninit_undefined_value_p instead of ssa_undefined_value_p.
20262 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
20263 for modifier == EXPAND_INITIALIZER.
20265 2013-05-07 Anton Blanchard <anton@samba.org>
20267 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
20268 for powerpc64 little endian.
20269 * configure: Regenerate.
20271 2013-05-06 Graham Stott <grahams@btinternet.com>
20273 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
20274 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
20275 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
20276 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
20278 2013-05-06 Graham Stott <grahams@btinternet.com>
20280 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
20281 codes which allow non-lvalues.
20283 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
20285 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
20286 components are all 1s.
20287 (integer_minus_onep): New function.
20288 * tree.h (integer_minus_onep): Declare it.
20289 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
20290 integer_minus_onep instead of integer_all_onesp.
20292 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20295 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
20296 variations of these patterns.
20298 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20300 * config/i386/i386.md (isa): Add x64_sse4 member.
20301 (enabled): Handle x64_sse4.
20302 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
20303 instruction for 64bit SSE4_1 targets. Update insn attributes.
20304 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
20305 instruction for SSE4_1 targets. Update insn attributes.
20306 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20307 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
20308 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
20310 (*vec_extractv4si): Rename from *sse4_1_pextrd.
20311 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
20312 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
20314 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20317 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
20319 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
20321 * final.c (do_assembler_dialects): Don't handle curly braces and
20322 vertical bar escaped by % as dialect delimiters.
20323 (output_asm_insn): Print curly braces and vertical bar if escaped
20324 by % and ASSEMBLER_DIALECT defined.
20325 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
20326 * doc/tm.texi: Regenerated.
20328 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20330 * config/mips/mips.c: Include tree-pass.h.
20331 (mips_reorg): Split in pre- and post-dbr_schedule parts.
20332 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
20333 (pass_mips_machine_reorg2): New machine specific pass.
20334 (insert_pass_mips_machine_reorg2): New pass plugin definition.
20335 (mips_option_override): Register the new pass.
20336 * rtl.h (cleanup_barriers): Remove prototype.
20337 (dbr_schedule): Likewise.
20338 * jump.c (cleanup_barriers): Make static.
20339 * reorg.c (dbr_schedule): Likewise.
20341 2013-05-06 Richard Biener <rguenther@suse.de>
20343 PR tree-optimization/57185
20344 * tree-parloops.c (add_field_for_reduction): Handle anonymous
20345 SSA names properly.
20347 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20350 * config/i386/i386.c (add_parameter_dependencies): Add dependence
20351 between "first_arg" and "insn", not "last" and "insn".
20353 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20355 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
20356 (find_candidates_in_block): Re-enable slsr_process_phi.
20357 (create_phi_basis): Fix double counting of candidate adjustment.
20359 2013-05-06 Richard Biener <rguenther@suse.de>
20361 PR middle-end/57147
20362 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
20363 the edge is also fallthru, preserve it and just clear the
20365 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
20366 also complex, preserve that and just clear the fallthru flag.
20367 * tree-inline.c (update_ssa_across_abnormal_edges): Also
20368 update virtual operands.
20370 2013-05-06 Alan Modra <amodra@gmail.com>
20372 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
20373 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
20374 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
20375 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
20376 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20377 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
20379 2013-05-06 Alan Modra <amodra@gmail.com>
20381 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
20382 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
20383 (DEFAULT_ASM_ENDIAN): Define.
20384 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20385 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
20386 Update -K PIC clause from sysv4.h.
20387 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
20388 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
20390 2013-05-06 Alan Modra <amodra@gmail.com>
20392 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
20393 twice for little-endian.
20394 (ashrdi3_no_power, ashrdi3): Support little-endian.
20396 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20399 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
20400 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
20402 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
20403 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
20404 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
20407 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
20408 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20410 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
20411 * config/pa/pa.opt: Make mbig-switch a no-op.
20412 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
20413 (CASE_VECTOR_MODE): Always return SImode.
20414 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
20415 for the !TARGET_BIG_SWITCH case.
20416 * config/pa/pa-linux.h: Likewise.
20417 * config/pa/pa-openbsd.h: Likewise.
20418 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
20419 * config/pa/pa.md (short_jump): Remove define_insn.
20420 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
20421 (casesi0): Remove define_insn.
20422 (type): Remove btable_branch.
20423 (pa_combine_type): Likewise.
20424 (in_nullified_branch_delay): Likewise.
20425 (in_call_delay): Likewise.
20426 (define_delay): Likewise.
20427 (define_insn_reservation "Z3"): Likewise.
20428 (define_insn_reservation "Z4"): Likewise.
20429 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
20430 (pa_adjust_insn_length): Remove adjustment for btable branches.
20431 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
20434 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
20436 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20437 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
20438 Add m->r,x alternatives.
20439 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
20440 splitters using SWI48x mode iterator.
20441 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
20442 TARGET_64BIT. Add m->x alternative.
20443 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
20444 Add o->x alternative. Enable for TARGET_SSE.
20445 (sse_storeq): Remove expander.
20446 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
20447 with memory input operand.
20448 (*vec_extractv2di_1 splitter): New.
20449 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
20450 * config/i386/i386.md (ssevecmodelower): New mode attribute.
20452 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20454 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
20455 (INT_LOWPART): Delete.
20456 (extract_MB): Adjust.
20457 (extract_ME): Adjust.
20458 (print_operand): Adjust.
20460 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20462 * config/rs6000/predicates.md (reg_or_add_cint_operand,
20463 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20464 (reg_or_logical_cint_operand, easy_fp_constant,
20465 logical_const_operand): Delete "CONST_DOUBLE" case.
20466 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
20467 "HOST_BITS_PER_WIDE_INT == 64" test.
20468 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
20469 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
20470 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
20471 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
20473 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
20474 CONST_DOUBLE DImode/VOIDmode case.
20475 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
20476 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20477 CONST_DOUBLE VOIDmode case.
20478 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20479 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
20480 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
20481 Delete CONST_DOUBLE case.
20482 (splitters for mov FMOVE64 const_double): Delete
20483 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20484 "HOST_BITS_PER_WIDE_INT >= 64" test.
20485 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
20487 (mov DI const_double): Delete.
20489 2013-05-04 Jakub Jelinek <jakub@redhat.com>
20491 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
20492 on op shows all bits zero in mode of a lowpart subreg, return zero.
20494 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
20497 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
20498 to save TFmode registers and DImode to save TImode registers for
20499 caller save operations.
20500 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
20501 mark being partially clobbered since they only use the first
20504 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
20505 and TDmode only use the upper 64-bits of each VSX register.
20507 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20509 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
20510 (find_candidates_in_block): Disable slsr_process_phi.
20512 2013-05-03 Guozhi Wei <carrot@google.com>
20514 * coverage.c (coverage_obj_init): Move the construction of gcov
20516 (build_init_ctor): ... here.
20518 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20520 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
20521 (slsr_cand_d): Redefine def_phi.
20522 (stride_status, phi_adjust_status, count_phis_status): New enums.
20523 (find_phi_def): New.
20524 (find_basis_for_base_expr): New.
20525 (find_basis_for_candidate): Handle hidden bases.
20526 (alloc_cand_and_find_basis): Handle phi candidates.
20527 (slsr_process_phi): New.
20528 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
20529 (create_mul_imm_cand): Likewise.
20530 (create_add_ssa_cand): Exclude phi base candidates.
20531 (create_add_imm_cand): Likewise.
20532 (slsr_process_cast): Likewise.
20533 (slsr_process_copy): Likewise.
20534 (find_candidates_in_block): Handle phi candidates.
20535 (dump_candidate): Likewise.
20536 (unconditional_cands): Delete.
20537 (unconditional_cands_with_known_stride_p): Delete.
20538 (phi_dependent_cand_p): New.
20539 (cand_increment): Handle phi-dependent candidates.
20540 (replace_dependent): Delete.
20541 (replace_mult_candidate): New.
20542 (replace_unconditional_candidate): New.
20543 (incr_vec_index): Move to avoid forward reference.
20544 (create_add_on_incoming_edge): New.
20545 (create_phi_basis): New.
20546 (replace_dependents): Delete.
20547 (replace_conditional_candidate): New.
20548 (phi_add_costs): New.
20549 (replace_uncond_cands_and_profitable_phis): New.
20550 (record_increment): Handle phi adjustments.
20551 (record_phi_increments): New.
20552 (record_increments): Handle phi adjustments.
20553 (phi_incr_cost): New.
20554 (lowest_cost_path): Handle phis.
20555 (total_savings): Likewise.
20556 (analyze_increments): Likewise.
20557 (ncd_with_phi): New.
20558 (ncd_of_cand_and_phis): New.
20559 (nearest_common_dominator_for_cands): Handle phi increments.
20560 (all_phi_incrs_profitable): New.
20561 (replace_profitable_candidates): Handle phi-dependent candidates.
20562 (analyze_candidates_and_replace): Likewise.
20564 2013-05-03 Teresa Johnson <tejohnson@google.com>
20567 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
20568 do not exceed REG_BR_PROB_BASE.
20570 2013-05-03 Jeff Law <law@redhat.com>
20572 PR tree-optimization/57144
20573 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
20574 operand of the condition will bit into the new type when eliminating
20575 a cast feeding a condition.
20577 2013-05-03 Jakub Jelinek <jakub@redhat.com>
20579 PR rtl-optimization/57130
20580 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
20581 of COMPARE as in_code to the recursive call if needed.
20583 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
20585 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
20586 (enabled): Handle new members.
20587 * config/i386/sse.md (*vec_concatv2si): Merge from
20588 *vec_concatv2si_sse2 and vec_concatv2si_sse.
20589 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
20591 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
20593 PR tree-optimization/57027
20594 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
20595 for fnms opportunity, check we got the prerequisite kind
20596 of tree / gimple before using accessor functions.
20598 2013-05-03 Richard Biener <rguenther@suse.de>
20600 * double-int.h (lshift): New overload without precision
20601 and arith argument.
20602 (operator *=, operator +=, operator -=): Move ...
20603 * double-int.c (operator *=, operator +=, operator -=): ... here
20604 and implement more efficiently.
20605 (mul_double_with_sign): Remove.
20606 (lshift_double): Adjust to take unsinged shift argument, push
20607 dispatching code to callers.
20608 (mul_double_wide_with_sign): Add early out for callers that
20609 are not interested in high parts or overflow.
20610 (lshift): New function.
20611 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
20612 dispatch code here.
20613 (lrotate, rrotate): Use logical shifts.
20614 * expr.c (get_inner_reference): Use lshift.
20615 * fixed-value.c (do_fixed_divide): Likewise.
20616 * tree-dfa.c (get_ref_base_and_extent): Likewise.
20617 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
20618 (indirect_refs_may_alias_p): Likewise.
20619 (stmt_kills_ref_p_1): Likewise.
20621 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20623 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
20625 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20627 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
20628 scalar form of FABD instruction.
20630 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
20632 * lra-constraints.c (process_alt_operands): Add checking alt
20633 number to choose the best alternative.
20635 2013-05-02 Richard Biener <rguenther@suse.de>
20637 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
20638 bitmap and its handling.
20639 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
20641 2013-05-02 Richard Biener <rguenther@suse.de>
20643 PR middle-end/57140
20644 * tree-inline.c (copy_loops): Properly handle removed loops.
20645 (copy_cfg_body): Mark destination loops for fixup if source
20646 loops needed fixup.
20648 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
20651 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
20652 generating simple_return for naked functions.
20654 2013-05-02 Martin Jambor <mjambor@suse.cz>
20656 PR middle-end/56988
20657 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
20658 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
20660 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
20661 ipa_agg_replacement_value structures.
20662 (known_aggs_to_agg_replacement_list): Likewise.
20663 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
20664 (read_agg_replacement_chain): Likewise.
20665 (ipcp_transform_function): Also check that by_ref flags match.
20667 2013-05-02 Richard Biener <rguenther@suse.de>
20669 * graphds.h (struct graph): Add obstack member.
20670 * graphds.c (new_graph): Initialize obstack and allocate
20672 (add_edge): Allocate edge from the obstack.
20673 (free_graph): Free the obstack instead of all edges and vertices.
20675 2013-05-02 Teresa Johnson <tejohnson@google.com>
20677 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
20679 * cfg.c (update_bb_profile_for_threading): Ditto.
20680 * tree-inline.c (copy_bb): Ditto.
20681 (copy_edges_for_bb): Ditto.
20682 (initialize_cfun): Ditto.
20683 (copy_cfg_body): Ditto.
20684 (expand_call_inline): Ditto.
20685 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
20686 (estimate_node_size_and_time): Ditto.
20687 (inline_merge_summary): Ditto.
20688 * cgraphclones.c (cgraph_clone_edge): Ditto.
20689 (cgraph_clone_node): Ditto.
20690 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20691 (compute_trg_info): Ditto.
20693 2013-05-02 Ian Bolton <ian.bolton@arm.com>
20695 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
20696 S reg when fp attribute set.
20697 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
20699 2013-05-02 Ian Bolton <ian.bolton@arm.com>
20701 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
20703 (*and_one_cmplsi3_compare0_uxtw): Likewise.
20704 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
20705 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
20707 2013-05-02 Richard Biener <rguenther@suse.de>
20709 * tree-scalar-evolution.c (scev_info_hasher): Remove.
20710 (struct instantiate_cache_entry): New type.
20711 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
20712 (struct instantiate_cache_type): New type.
20713 (set_instantiated_value, get_instantiated_value): Remove.
20714 (get_instantiated_value_entry): New function.
20715 (instantiate_scev_name): Use the new cache and adjust.
20716 (instantiate_scev_poly): Adjust.
20717 (instantiate_scev_binary): Likewise.
20718 (instantiate_array_ref): Likewise.
20719 (instantiate_scev_convert): Likewise.
20720 (instantiate_scev_not): Likewise.
20721 (instantiate_scev_3): Likewise.
20722 (instantiate_scev_2): Likewise.
20723 (instantiate_scev_r): Likewise.
20724 (instantiate_scev): Likewise.
20725 (resolve_mixers): Likewise.
20727 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
20730 * lra-constraints.c (best_small_class_operands_num): Remove.
20731 (process_alt_operands): Remove small_class_operands_num. Take
20732 small classes operands into losers and only if the operand is not
20733 matched. Modify debugging output.
20734 (curr_insn_transform): Remove best_small_class_operands_num.
20737 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20739 * config/aarch64/aarch64-builtins.c
20740 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
20741 * config/aarch64/aarch64-simd-builtins.def
20742 (reduc_splus_): Add new modes.
20743 (reduc_uplus_): New.
20744 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
20745 (reduc_uplus_v4sf): Likewise.
20746 (reduc_splus_v4sf): Likewise.
20747 (aarch64_addv<mode>): Likewise.
20748 (reduc_uplus_<mode>): Likewise.
20749 (reduc_splus_<mode>): Likewise.
20750 (aarch64_addvv2di): Likewise.
20751 (reduc_uplus_v2di): Likewise.
20752 (reduc_splus_v2di): Likewise.
20753 (aarch64_addvv2si): Likewise.
20754 (reduc_uplus_v2si): Likewise.
20755 (reduc_splus_v2si): Likewise.
20756 (reduc_<sur>plus_<mode>): New.
20757 (reduc_<sur>plus_v2di): Likewise.
20758 (reduc_<sur>plus_v2si): Likewise.
20759 (reduc_<sur>plus_v4sf): Likewise.
20760 (aarch64_addpv4sf): Likewise.
20761 * config/aarch64/arm_neon.h
20762 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
20763 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
20764 add UNSPEC_SADDV, UNSPEC_UADDV.
20766 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
20768 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20770 * config/aarch64/arm_neon.h
20771 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
20773 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20775 * config/aarch64/aarch64-builtins
20776 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
20778 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20780 * config/aarch64/aarch64-simd-builtins.def
20781 (reduc_smax_): New.
20782 (reduc_smin_): Likewise.
20783 (reduc_umax_): Likewise.
20784 (reduc_umin_): Likewise.
20785 (reduc_smax_nan_): Likewise.
20786 (reduc_smin_nan_): Likewise.
20789 (smax): Update for V2SF, V4SF and V2DF modes.
20792 (smin_nan): Likewise.
20793 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
20794 (<su><maxmin><mode>3): ...This, refactor.
20795 (s<maxmin><mode>3): New.
20796 (<maxmin_uns><mode>3): Likewise.
20797 (reduc_<maxmin_uns>_<mode>): Refactor.
20798 (reduc_<maxmin_uns>_v4sf): Likewise.
20799 (reduc_<maxmin_uns>_v2si): Likewise.
20800 (aarch64_<fmaxmin><mode>: Remove.
20801 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
20803 (vmin<q>_f<32,64>): Likewise.
20804 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
20806 (su): Add mappings for smax, smin, umax, umin.
20808 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
20809 (FMAXMIN): Rename as...
20810 (FMAXMIN_UNS): ...This.
20812 (fmaxminv): Likewise.
20813 (fmaxmin): Likewise.
20815 (maxmin_uns_op): Likewise.
20817 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20819 * config/aarch64/arm_neon.h
20820 (vac<ge, gt><sd>_f<32, 64>): Rename to...
20821 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
20822 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
20824 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20826 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
20827 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
20829 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20831 * config/aarch64/aarch64-simd.md
20832 (vcond<mode>_internal): Handle special cases for constant masks.
20833 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
20834 (vcondu<mode><mode>): Likewise.
20835 (vcond<v_cmp_result><mode>): New.
20837 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20839 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
20840 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
20841 * config/aarch64/aarch64-simd-builtins.def
20842 (cmeq): Update to BUILTIN_VALLDI.
20847 * config/aarch64/arm_neon.h
20848 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
20849 to builtins or C as appropriate.
20851 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20853 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
20855 (cmhi): Rename to...
20857 * config/aarch64/aarch64-simd.md
20858 (simd_mode): Add SF.
20859 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
20860 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
20861 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
20862 (cstore<mode>_neg): ...This.
20863 * config/aarch64/iterators.md
20865 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
20866 (COMPARISONS): New.
20867 (UCOMPARISONS): Likewise.
20868 (optab): Add missing comparisons.
20874 (VCMP_S): Likewise.
20875 (VCMP_U): Likewise.
20876 (V_cmp_result): Add DF, SF modes.
20877 (v_cmp_result): Likewise.
20879 (vmtype): Likewise.
20880 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
20882 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
20884 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
20885 define_insn to define_insn_and_split.
20886 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
20887 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
20888 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
20889 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
20890 (thumb2_negscc): Likewise.
20892 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
20894 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
20896 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
20898 * config/arm/thumb2.md: Remove trailing whitespaces.
20900 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
20902 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
20903 Use gen_int_mode rather than GEN_INT.
20905 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
20907 * value-prof.c (stream_in_histogram_value): Remove the strayed
20910 2013-04-30 Richard Biener <rguenther@suse.de>
20912 PR middle-end/57122
20913 * cfghooks.c (split_edge): Properly check for the loop latch edge.
20915 2013-04-30 Richard Biener <rguenther@suse.de>
20917 PR middle-end/57107
20918 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
20920 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
20922 PR rtl-optimization/56957
20923 PR rtl-optimization/57105
20924 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
20925 variable. Use just INSN_UID for determining whether an insn
20926 should be only disconnected from the insn stream.
20927 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
20929 2013-04-30 Jakub Jelinek <jakub@redhat.com>
20931 PR tree-optimization/57104
20932 * tsan.c (instrument_expr): Don't instrument accesses to
20933 DECL_HARD_REGISTER VAR_DECLs.
20935 2013-04-30 Richard Biener <rguenther@suse.de>
20937 * function.h (loops_for_fn): New inline function.
20938 (set_loops_for_fn): Likewise.
20939 * cfgloop.h (place_new_loop): Add struct function parameter.
20940 (get_loop): Likewise.
20941 (get_loops): Likewise.
20942 (number_of_loops): Likewise.
20943 (fel_next): Adjust.
20944 (fel_init): Likewise.
20945 * cfg.c (get_loop_copy): Adjust.
20946 * cfgloop.c (flow_loops_dump): Likewise.
20947 (record_loop_exits): Likewise.
20948 (verify_loop_structure): Likewise.
20949 * cfgloopanal.c (mark_irreducible_loops): Likewise.
20950 (estimate_reg_pressure_cost): Likewise.
20951 (mark_loop_exit_edges): Likewise.
20952 * cfgloopmanip.c (place_new_loop): Likewise.
20953 (add_loop): Likewise.
20954 (duplicate_loop): Likewise.
20955 * graph.c (draw_cfg_nodes): Likewise.
20956 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
20957 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
20958 (extract_affine_chrec): Likewise.
20959 (build_scop_iteration_domain): Likewise.
20960 * graphite.c (graphite_initialize): Likewise.
20961 * ira-build.c (create_loop_tree_nodes): Likewise.
20962 (more_one_region_p): Likewise.
20963 (rebuild_regno_allocno_maps): Likewise.
20964 (mark_loops_for_removal): Likewise.
20965 (mark_all_loops_for_removal): Likewise.
20966 (remove_unnecessary_regions): Likewise.
20967 (ira_build): Likewise.
20968 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
20969 * loop-init.c (fix_loop_structure): Likewise.
20970 (gate_rtl_move_loop_invariants): Likewise.
20971 (gate_rtl_unswitch): Likewise.
20972 (gate_rtl_unroll_and_peel_loops): Likewise.
20973 (rtl_doloop): Likewise.
20974 * lto-streamer-in.c (input_cfg): Likewise.
20975 * lto-streamer-out.c (output_cfg): Likewise.
20976 * modulo-sched.c (sms_schedule): Likewise.
20977 * predict.c (tree_estimate_probability): Likewise.
20978 (tree_estimate_probability_driver): Likewise.
20979 (estimate_loops): Likewise.
20980 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
20981 (move_sese_region_to_fn): Likewise.
20982 (debug_loop_num): Likewise.
20983 * tree-chrec.c (chrec_evaluate): Likewise.
20984 (hide_evolution_in_other_loops_than_loop): Likewise.
20985 (chrec_component_in_loop_num): Likewise.
20986 (reset_evolution_in_loop): Likewise.
20987 (evolution_function_is_invariant_rec_p): Likewise.
20988 * tree-if-conv.c (main_tree_if_conversion): Likewise.
20989 * tree-inline.c (copy_loops): Likewise.
20990 (copy_cfg_body): Likewise.
20991 (tree_function_versioning): Likewise.
20992 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
20993 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
20995 (add_to_evolution_1): Likewise.
20996 (scev_const_prop): Likewise.
20997 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
20998 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
20999 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
21000 (tree_ssa_lim_initialize): Likewise.
21001 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
21002 (verify_loop_closed_ssa): Likewise.
21003 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
21004 (tree_ssa_loop_im): Likewise.
21005 (tree_ssa_loop_unswitch): Likewise.
21006 (tree_vectorize): Likewise.
21007 (check_data_deps): Likewise.
21008 (tree_ssa_loop_ivcanon): Likewise.
21009 (tree_ssa_loop_bounds): Likewise.
21010 (tree_complete_unroll): Likewise.
21011 (tree_complete_unroll_inner): Likewise.
21012 (tree_parallelize_loops): Likewise.
21013 (tree_ssa_loop_prefetch): Likewise.
21014 (tree_ssa_loop_ivopts): Likewise.
21015 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
21016 * tree-vectorizer.c (vectorize_loops): Likewise.
21018 2013-04-29 Mike Frysinger <vapier@gentoo.org>
21020 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
21021 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
21022 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
21023 with EABI_LINK_SPEC.
21025 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21028 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
21031 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
21034 * lra-constraints.c (process_alt_operands): Discourage a bit more
21035 using memory for pseudos. Print cost dump for alternatives.
21036 Modify cost values for conflicts with early clobbers.
21037 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
21039 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21042 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
21044 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21046 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
21047 from/to S register.
21048 (movdi_aarch64): Support LDR/STR from/to D register.
21050 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21052 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
21053 or higher by default.
21055 2013-04-29 Richard Biener <rguenther@suse.de>
21057 PR middle-end/57075
21058 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
21059 even if not adding abnormal edges for calls that can make
21062 2013-04-29 Richard Biener <rguenther@suse.de>
21064 PR middle-end/57103
21065 * tree-cfg.c (move_stmt_op): Fix condition under which to update
21067 (move_stmt_r): Remove redundant checking.
21069 2013-04-29 Teresa Johnson <tejohnson@google.com>
21072 * basic-block.h (apply_scale): New function.
21073 (apply_probability): Use apply_scale.
21074 * gimple-streamer-in.c (input_bb): Ditto.
21075 * lto-streamer-in.c (input_cfg): Ditto.
21076 * lto-cgraph.c (merge_profile_summaries): Ditto.
21077 * tree-optimize.c (execute_fixup_cfg): Ditto.
21078 * tree-inline.c (copy_bb): Update comment to use apply_scale.
21079 (copy_edges_for_bb): Ditto.
21080 (copy_cfg_body): Ditto.
21082 2013-04-29 Tom de Vries <tom@codesourcery.com>
21084 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
21085 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
21086 (tail_merge_optimize): Handle current_loops == NULL.
21088 2013-04-26 Jeff Law <law@redhat.com>
21090 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
21091 (simplify_cond_using_ranges): Generalize code to simplify
21092 COND_EXPRs where one argument is a constant and the other
21093 is an SSA_NAME created by an integral type conversion.
21095 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21097 * config/arm/arm.md (store_minmaxsi): Use only when
21098 optimize_insn_for_size_p.
21100 2013-04-29 Christian Bruel <christian.bruel@st.com>
21103 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
21105 2013-04-29 Richard Biener <rguenther@suse.de>
21107 PR middle-end/57089
21108 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
21109 loop tree make sure to schedule a fixup for the child as well.
21110 (expand_omp_for_generic): Properly add loops.
21111 (expand_omp_for_static_nochunk): Likewise.
21112 (expand_omp_for_static_chunk): Likewise.
21113 (expand_omp_for): For the degenerate case fixup loops.
21114 (expand_omp_sections): Fix default bb placement in loops.
21115 (expand_omp_atomic_pipeline): Properly add loops.
21117 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21119 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
21121 2013-04-29 Tom de Vries <tom@codesourcery.com>
21123 * tree-ssa-tail-merge.c: Update header comment.
21125 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21127 * config/aarch64/arm_neon.h
21128 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
21129 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
21130 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
21131 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
21132 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
21133 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
21134 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
21135 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
21137 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21139 * config/aarch64/aarch64-simd.md
21140 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
21141 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
21142 fix_trunc, fixuns_trunc.
21143 (ftrunc<VDQF:mode>2): New.
21144 * config/aarch64/iterators.md (optab): Add fix, fixuns.
21145 (fix_trunc_optab): New.
21147 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21149 * config/aarch64/aarch64-builtins.c
21150 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
21151 iceilf, lround, iroundf.
21153 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21156 * config/i386/i386.h (enum ix86_tune_indices)
21157 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
21158 New, split from X86_TUNE_INTER_UNIT_MOVES.
21159 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
21160 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
21161 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
21162 (TARGET_INTER_UNIT_MOVES): Remove.
21163 * config/i386/i386.c (initial_ix86_tune_features): Update.
21164 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
21165 (ix86_expand_convert_uns_didf_sse): Use
21166 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21167 (ix86_expand_vector_init_one_nonzero): Ditto.
21168 (ix86_expand_vector_init_interleave): Ditto.
21169 (inline_secondary_memory_needed): Return true for moves from SSE class
21170 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
21171 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
21172 * config/i386/constraints.md (Yi, Ym): Depend on
21173 TARGET_INTER_UNIT_MOVES_TO_VEC.
21174 (Yj, Yn): New constraints.
21175 * config/i386/i386.md (*movdi_internal): Change constraints of
21176 operand 1 from Yi to Yj and from Ym to Yn.
21177 (*movsi_internal): Ditto.
21178 (*movdf_internal): Ditto.
21179 (*movsf_internal): Ditto.
21180 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
21181 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21182 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
21183 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
21184 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
21185 * config/i386/sse.md (movdi_to_sse): Ditto.
21186 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
21187 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
21188 TARGET_INTER_UNIT_MOVES.
21189 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
21190 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
21191 instead of TARGET_INTER_UNIT_MOVES.
21192 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
21193 operand 1 from Yi to Yj and from Ym to Yn.
21195 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21197 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
21198 (float_truncate_hi_): Likewise.
21199 (float_extend_lo_): Likewise.
21200 (float_truncate_lo_): Likewise.
21201 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
21202 (aarch64_float_extend_lo_v2df): Likewise.
21203 (vec_unpacks_hi_v4sf): Likewise.
21204 (aarch64_float_truncate_lo_v2sf): Likewise.
21205 (aarch64_float_truncate_hi_v4sf): Likewise.
21206 (vec_pack_trunc_v2df): Likewise.
21207 (vec_pack_trunc_df): Likewise.
21209 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21211 * config/aarch64/aarch64-builtins.c
21212 (aarch64_fold_builtin): Fold float conversions.
21213 * config/aarch64/aarch64-simd-builtins.def
21214 (floatv2si, floatv4si, floatv2di): New.
21215 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
21216 * config/aarch64/aarch64-simd.md
21217 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
21218 * config/aarch64/iterators.md (FLOATUORS): New.
21219 (optab): Add float, floatuns.
21220 (su_optab): Likewise.
21222 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21224 * config/aarch64/aarch64-builtins.c
21225 (aarch64_builtin_vectorized_function): Use new names for
21227 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
21228 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
21229 (fcvtzu): Split as...
21230 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
21231 (fcvtas): Split as...
21232 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
21233 (fcvtau): Split as...
21234 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
21235 (fcvtps): Split as...
21236 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
21237 (fcvtpu): Split as...
21238 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
21239 (fcvtms): Split as...
21240 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
21241 (fcvtmu): Split as...
21242 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
21243 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
21244 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
21245 (lfrintnusf, lfrintnudf): Likewise.
21246 * config/aarch64/aarch64-simd.md
21247 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
21249 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
21250 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
21251 (fcvt_pattern): Likewise.
21253 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21255 * config/aarch64/aarch64-simd.md
21256 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
21257 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
21259 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21261 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
21262 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
21263 (vrnd<a,m,n,p>_f32): Implement using builtins.
21264 (vrnd<i,x><q>_f<32, 64>): New.
21266 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21268 * config/aarch64/aarch64-builtins.c
21269 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
21270 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
21271 (frintz): Rename to...
21273 (frintp): Rename to...
21275 (frintm): Rename to...
21277 (frinti): Rename to...
21278 (nearbyint): ...this.
21279 (frintx): Rename to...
21281 (frinta): Rename to...
21283 * config/aarch64/aarch64-simd.md
21284 (aarch64_frint<frint_suffix><mode>): Delete.
21285 (<frint_pattern><mode>2): Convert to insn.
21286 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
21287 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
21288 (frint_pattern): Likewise.
21289 (frint_suffix): Likewise.
21291 2013-04-29 Richard Biener <rguenther@suse.de>
21293 PR tree-optimization/57081
21294 * loop-init.c: Include tree-flow.h.
21295 (loop_optimizer_finalize): Free number of iteration estimates.
21296 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
21298 2013-04-29 Jakub Jelinek <jakub@redhat.com>
21300 PR tree-optimization/57083
21301 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
21302 non-singleton shift count range, zero extend low_bound for uns case.
21304 * config/i386/predicates.md (general_vector_operand): New predicate.
21305 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
21306 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
21307 if they aren't nonimmediate operands. If their original values
21308 satisfy const_vector_equal_evenodd_p, don't shift them.
21309 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
21310 predicates. For the SSE4.1 case force operands[{1,2}] into registers
21311 if not nonimmediate_operand.
21312 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
21313 instead of register_operand.
21314 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
21316 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
21318 * stor-layout.c (finalize_size_functions): Allocate a structure and
21319 reset cfun before dumping the functions.
21321 2013-04-27 Jakub Jelinek <jakub@redhat.com>
21323 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
21326 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
21327 use xop_pmacsdqh if uns_p.
21328 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
21329 the immediate rotate count.
21331 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
21333 * rtl.h (struct rtx_def): Add comment for field jump.
21334 (LRA_SUBREG_P): New macro.
21335 * recog.c (register_operand): Check LRA_SUBREG_P.
21336 * lra.c (lra): Add note at the end of RTL code. Align non-empty
21338 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
21339 (lra_final_code_change): Skip subreg change for operators.
21340 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
21341 if there are no operand changes.
21342 * lra-constraints.c (curr_insn_set): New.
21343 (match_reload): Set LRA_SUBREG_P.
21344 (emit_spill_move): Ditto.
21345 (check_and_process_move): Use curr_insn_set. Process only single
21346 set insns. Don't initialize sec_mem_p and change_p.
21347 (simplify_operand_subreg): Use LRA_SUBREG_P.
21348 (reg_in_class_p): New function.
21349 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
21350 of #ifdef. Add code to remove cycling.
21351 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
21352 non-null disp. Reload inner instead of disp when base and index
21353 are null. Try to put lo_sum into register.
21354 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
21355 (check_and_process_move): Move code for move cost check to
21356 simple_move_p. Remove equiv_substitution.
21357 (simple_move_p): New function.
21358 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
21359 curr_insn_set. Call check_and_process_move only for single set
21360 insns. Use the new function. Move call of check_and_process_move
21361 after operand equiv substitution and address process.
21363 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21366 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
21367 with nonlocal goto receivers or returns twice calls, ignore
21368 unininitialized values from abnormal edges to nl goto receiver
21369 or returns twice call.
21371 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21373 PR tree-optimization/57051
21374 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
21375 and VEC_RSHIFT_EXPR if shift count is a multiple of element
21378 2013-04-26 Richard Biener <rguenther@suse.de>
21380 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
21381 (expand_omp_taskreg): Likewise. Mark loops for fixup.
21382 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
21383 (fixup_loop_arrays_after_move): New function.
21384 (move_sese_region_to_fn): Properly outline the loop tree parts
21385 of the SESE region.
21387 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
21389 * config/i386/i386.md (type, unit): Fix long lines.
21391 2013-04-26 Richard Biener <rguenther@suse.de>
21393 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
21394 (lto-streamer-out.o): Likewise.
21395 * cfgloop.c (init_loops_structure): Export, add struct function
21396 argument and adjust.
21397 (flow_loops_find): Adjust.
21398 * cfgloop.h (enum loop_estimation): Add EST_LAST.
21399 (init_loops_structure): Declare.
21400 * lto-streamer-in.c: Include cfgloop.h.
21401 (input_cfg): Input the loop tree.
21402 * lto-streamer-out.c: Include cfgloop.h.
21403 (output_cfg): Output the loop tree.
21404 (output_struct_function_base): Do not drop PROP_loops.
21406 2013-03-26 Richard Biener <rguenther@suse.de>
21408 * tree-cfg.c (execute_build_cfg): Build the loop tree.
21409 (pass_build_cfg): Provide PROP_loops.
21410 (move_sese_region_to_fn): Remove loops that are outlined into fn
21412 * tree-inline.c: Include cfgloop.h.
21413 (initialize_cfun): Do not drop PROP_loops.
21414 (copy_loops): New function.
21415 (copy_cfg_body): Copy loop structure.
21416 (tree_function_versioning): Initialize destination loop tree.
21417 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
21418 (pass_parallelize_loops): Do IL verification.
21419 * loop-init.c (loop_optimizer_init): Fixup loops if required.
21420 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
21421 the CFG make sure we fixup loops as well.
21422 * tree-ssa-tail-merge.c: Include cfgloop.h.
21423 (replace_block_by): When merging loop latches mark loops for fixup.
21424 * lto-streamer-out.c (output_struct_function_base): Drop
21425 PROP_loops for now.
21426 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
21427 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
21428 * ipa-split.c: Include cfgloop.h.
21429 (split_function): Add the new return block to the loop tree root.
21430 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
21431 whether we have removed the forwarder block.
21432 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
21433 * cfgloop.h (place_new_loop): Declare.
21434 * cfgloopmanip.c (place_new_loop): Export.
21435 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
21436 (tree-switch-conversion.o): Likewise.
21437 (tree-complex.o): Likewise.
21438 (tree-inline.o): Likewise.
21439 (tree-ssa-tailmerge.o): Likewise.
21440 (ipa-split.o): Likewise.
21441 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
21442 (tree-ssa-copy.o): Likewise.
21443 * tree-switch-conversion.c: Include cfgloop.h
21444 (process_switch): If we emit a bit-test cascade, schedule loops
21446 * tree-complex.c: Include cfgloop.h.
21447 (expand_complex_div_wide): Properly add new basic-blocks to loops.
21448 * asan.c: Include cfgloop.h.
21449 (create_cond_insert_point): Properly add new basic-blocks to
21450 loops, schedule loop fixup.
21451 * cfgloop.c (verify_loop_structure): Check that looks are not
21453 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
21455 (expand_omp_for_generic): Likewise.
21456 (expand_omp_sections): Likewise.
21457 (expand_omp_atomic_pipeline): Schedule loops for fixup.
21458 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
21459 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
21460 is initialized, not when loops are present.
21461 * tree-parloops.c (parallelize_loops): Remove checking here.
21462 * passes.c (init_optimization_passes): Schedule a copy-propagation
21463 pass before complete unrolling of inner loops.
21465 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21467 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
21468 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
21469 (colorize_init): Add argument to _WIN32 version.
21470 * toplev.c: Include diagnostic-color.h.
21471 (process_options): Default to -fdiagnostics-color=auto if
21472 GCC_COLORS env var is in the environment.
21473 * common.opt (fdiagnostics-color=): Add Var and Init.
21474 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
21475 env var is in the environment, the default is auto rather than never.
21477 * diagnostic.h (file_name_as_prefix): Add context argument.
21478 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
21479 the string as locus.
21480 * langhooks.c (lhd_print_error_function): Adjust caller.
21482 2013-04-25 Lawrence Crowl <crowl@google.com>
21484 * var-tracking.c (shared_hash_def::htab):
21485 Change type to hash_table. Update dependent calls and types.
21487 2013-04-25 Lawrence Crowl <crowl@google.com>
21489 * Makefile.in: Update as needed below.
21491 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
21492 Move declaration to after the type's method definitons.
21494 * attribs.c (htab_t scoped_attributes::attribute_hash):
21495 Change type to hash_table. Update dependent calls and types.
21497 * bitmap.c (htab_t bitmap_desc_hash):
21498 Change type to hash_table. Update dependent calls and types.
21500 * cselib.c (htab_t cselib_hash_table):
21501 Change type to hash_table. Update dependent calls and types.
21503 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
21504 (hash_string_slot_node): Move implementation into lto-streamer.h
21505 struct string_slot_hasher.
21506 (eq_string_slot_node): Likewise.
21508 * data-streamer-out.c: Update output_block::string_hash_table
21509 dependent calls and types.
21511 * dwarf2cfi.c (htab_t trace_index):
21512 Change type to hash_table. Update dependent calls and types.
21514 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
21515 Change type to hash_table. Update dependent calls and types.
21516 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
21517 (htab_t optimize_external_refs::map): Likewise.
21518 (htab_t output_comp_unit::extern_map): Likewise.
21519 (htab_t output_comdat_type_unit::extern_map): Likewise.
21520 (htab_t output_macinfo::macinfo_htab): Likewise.
21521 (htab_t optimize_location_lists::htab): Likewise.
21522 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
21524 * except.c (htab_t ehspec_hash_type):
21525 Change type to hash_table. Update dependent calls and types.
21526 (assign_filter_values::ttypes): Likewise.
21527 (assign_filter_values::ehspec): Likewise.
21528 (sjlj_assign_call_site_values::ar_hash): Likewise.
21529 (convert_to_eh_region_ranges::ar_hash): Likewise.
21531 * gcse.c (htab_t pre_ldst_table):
21532 Change type to hash_table. Update dependent calls and types.
21534 * ggc-common.c (htab_t saving_htab):
21535 Change type to hash_table. Update dependent calls and types.
21536 (htab_t loc_hash): Likewise.
21537 (htab_t ptr_hash): Likewise.
21538 (call_count): Rename ggc_call_count.
21539 (call_alloc): Rename ggc_call_alloc.
21540 (loc_descriptor): Rename make_loc_descriptor.
21541 (add_statistics): Rename ggc_add_statistics.
21543 * ggc-common.c (saving_htab):
21544 Change type to hash_table. Update dependent calls and types.
21546 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
21547 (push_gimplify_context): Likewise.
21548 (pop_gimplify_context): Likewise.
21549 (struct gimple_temp_hash_elt): Added.
21550 (struct gimplify_hasher): Likewise.
21551 (struct gimplify_ctx.temp_htab):
21552 Change type to hash_table. Update dependent calls and types.
21554 * gimple-fold.c: Include gimplify-ctx.h.
21556 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
21557 Change type to hash_table. Update dependent calls and types.
21558 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
21559 avoid potential global name collision.
21561 * gimplify.c: Include gimplify-ctx.h.
21562 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
21563 (htab_t gimplify_ctx::temp_htab):
21564 Update dependent calls and types for new type hash_table.
21565 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
21566 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
21568 * gimplify-ctx.h: New.
21569 (struct gimple_temp_hash_elt): Move from gimplify.c.
21570 (class gimplify_hasher): New.
21571 (struct gimplify_ctx): Move from gimple.h.
21572 (htab_t gimplify_ctx::temp_htab):
21573 Change type to hash_table. Update dependent calls and types.
21575 * graphite-clast-to-gimple.c: Include graphite-htab.h.
21576 (htab_t ivs_params::newivs_index):
21577 Change type to hash_table. Update dependent calls and types.
21578 (htab_t ivs_params::params_index): Likewise.
21579 (htab_t print_generated_program::params_index): Likewise.
21580 (htab_t gloog::newivs_index): Likewise.
21581 (htab_t gloog::params_index): Likewise.
21583 * graphite.c: Include graphite-htab.h.
21584 4htab_t graphite_transform_loops::bb_pbb_mapping):
21585 Change type to hash_table. Update dependent calls and types.
21587 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
21588 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
21589 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
21591 * graphite-dependences.c: Include graphite-htab.h.
21592 (loop_is_parallel_p): Change hash table type of parameter.
21594 * graphite-htab.h: New.
21595 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
21596 (extern find_pbb_via_hash): Move from graphite-poly.h.
21597 (extern loop_is_parallel_p): Move from graphite-poly.h.
21598 (extern get_loop_body_pbbs): Move from graphite-poly.h.
21600 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
21601 (extern loop_is_parallel_p): Move to graphite-htab.h.
21602 (extern get_loop_body_pbbs): Move to graphite-htab.h.
21604 * haifa-sched.c (htab_t delay_htab):
21605 Change type to hash_table. Update dependent calls and types.
21606 (htab_t delay_htab_i2): Likewise.
21608 * ira-color.c (htab_t allocno_hard_regs_htab):
21609 Change type to hash_table. Update dependent calls and types.
21611 * ira-costs.c (htab_t cost_classes_htab):
21612 Change type to hash_table. Update dependent calls and types.
21614 * loop-invariant.c (htab_t merge_identical_invariants::eq):
21615 Change type to hash_table. Update dependent calls and types.
21617 * loop-iv.c (htab_t bivs):
21618 Change type to hash_table. Update dependent calls and types.
21620 * loop-unroll.c (htab_t opt_info::insns_to_split):
21621 Change type to hash_table. Update dependent calls and types.
21622 (htab_t opt_info::insns_with_var_to_expand): Likewise.
21624 * lto-streamer.h (struct string_slot): Move from data-streamer.h
21625 (struct string_slot_hasher): New.
21626 (htab_t output_block::string_hash_table):
21627 Change type to hash_table. Update dependent calls and types.
21629 * lto-streamer-in.c (freeing_string_slot_hasher): New.
21630 (htab_t file_name_hash_table):
21631 Change type to hash_table. Update dependent calls and types.
21633 * lto-streamer-out.c: Update output_block::string_hash_table dependent
21636 * lto-streamer.c (htab_t tree_htab):
21637 Change type to hash_table. Update dependent calls and types.
21639 * omp-low.c: Include gimplify-ctx.h.
21641 * passes.c (htab_t name_to_pass_map):
21642 Change type to hash_table. Update dependent calls and types.
21643 (pass_traverse): Rename to passes_pass_traverse.
21645 * plugin.c (htab_t event_tab):
21646 Change type to hash_table. Update dependent calls and types.
21648 * postreload-gcse.c (htab_t expr_table):
21649 Change type to hash_table. Update dependent calls and types.
21650 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
21652 * sese.c (debug_rename_map_1): Make extern.
21653 (htab_t copy_bb_and_scalar_dependences::rename_map):
21654 Change type to hash_table. Update dependent calls and types.
21656 * sese.h (extern debug_rename_map): Move to .c file.
21658 * store-motion.c (htab_t store_motion_mems_table):
21659 Change type to hash_table. Update dependent calls and types.
21661 * trans-mem.c (htab_t tm_new_mem_hash):
21662 Change type to hash_table. Update dependent calls and types.
21664 * tree-browser.c (htab_t TB_up_ht):
21665 Change type to hash_table. Update dependent calls and types.
21667 * tree-cfg.c (htab_t discriminator_per_locus):
21668 Change type to hash_table. Update dependent calls and types.
21670 * tree-complex.c: Include tree-hasher.h
21671 (htab_t complex_variable_components):
21672 Change type to hash_table. Update dependent calls and types.
21674 * tree-eh.c (htab_t finally_tree):
21675 Change type to hash_table. Update dependent calls and types.
21677 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
21678 struct int_tree_hasher.
21679 (extern int_tree_map_eq): Likewise.
21680 (uid_decl_map_hash): Removed.
21681 (extern decl_tree_map_eq): Likewise.
21683 * tree-hasher.h: New.
21684 (struct int_tree_hasher): New.
21685 (typedef int_tree_htab_type): New.
21687 * tree-inline.c: Include gimplify-ctx.h.
21689 * tree-mudflap.c: Include gimplify-ctx.h.
21691 * tree-parloops.c: Include tree-hasher.h.
21692 (htab_t eliminate_local_variables_stmt::decl_address):
21693 Change type to hash_table. Update dependent calls and types.
21694 (htab_t separate_decls_in_region::decl_copies): Likewise.
21696 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
21697 Change type to hash_table. Update dependent calls and types.
21699 * tree-sra.c (candidates):
21700 Change type to hash_table. Update dependent calls and types.
21702 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
21704 (int_tree_map_hash): Likewise.
21706 * tree-ssa-dom.c (htab_t avail_exprs):
21707 Change type to hash_table. Update dependent calls and types.
21709 * tree-ssa-live.c (var_map_base_init::tree_to_index):
21710 Change type to hash_table. Update dependent calls and types.
21712 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
21713 Change type to hash_table. Update dependent calls and types.
21715 * tree-ssa-phiopt.c (seen_ssa_names):
21716 Change type to hash_table. Update dependent calls and types.
21718 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
21719 Change type to hash_table. Update dependent calls and types.
21721 * tree-ssa-uncprop.c (equiv):
21722 Change type to hash_table. Update dependent calls and types.
21724 2013-04-25 Jakub Jelinek <jakub@redhat.com>
21726 PR rtl-optimization/57003
21727 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
21728 call note_stores with kill_clobbered_value callback again after
21729 killing regs_invalidated_by_call.
21731 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21733 * config/aarch64/aarch64-simd.md
21734 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
21735 (aarch64_simd_bsl<mode>): Likewise.
21736 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
21738 2013-04-25 Marek Polacek <polacek@redhat.com>
21740 PR tree-optimization/57066
21741 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
21743 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21745 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
21747 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21749 * config/aarch64/aarch64-builtins.c
21750 (aarch64_fold_builtin): New.
21751 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
21752 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
21753 * config/aarch64/aarch64-simd-builtins.def (abs): New.
21754 * config/aarch64/arm_neon.h
21755 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
21757 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21758 Tejas Belagod <tejas.belagod@arm.com>
21760 * config/aarch64/aarch64-builtins.c
21761 (aarch64_gimple_fold_builtin): New.
21762 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
21763 * config/aarch64/aarch64-simd-builtins.def (addv): New.
21764 * config/aarch64/aarch64-simd.md (addpv4sf): New.
21765 (addvv4sf): Update.
21766 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
21768 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21770 * config/aarch64/aarch64.md
21771 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
21773 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21775 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
21776 (*ngcsi_uxtw): New pattern.
21778 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21779 Julian Brown <julian@codesourcery.com>
21781 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
21782 (TB_DREG): Add T_V4HF.
21783 (v4hf_UP): New macro.
21784 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21785 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21786 Handle initialisation of V4HF. Adjust initialisation of reinterpret
21788 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21789 (arm_vector_mode_supported_p): Handle V4HF.
21790 (arm_mangle_map): Handle V4HFmode.
21791 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
21792 * config/arm/arm_neon_builtins.def: Add entries for
21793 vcvtv4hfv4sf, vcvtv4sfv4hf.
21794 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
21795 (neon_vcvtv4hfv4sf): Likewise.
21796 * config/arm/neon-gen.ml: Handle half-precision floating point
21798 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
21799 * config/arm/arm_neon.h: Regenerate.
21800 * config/arm/neon.ml (type elts): Add F16.
21801 (type vectype): Add T_float16x4, T_floatHF.
21802 (type vecmode): Add V4HF.
21803 (type features): Add Requires_FP_bit feature.
21804 (elt_width): Handle F16.
21805 (elt_class): Likewise.
21806 (elt_of_class_width): Likewise.
21807 (mode_of_elt): Refactor.
21808 (type_for_elt): Handle F16, fix error messages.
21809 (vectype_size): Handle T_float16x4.
21810 (vcvt_sh): New function.
21811 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
21812 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
21813 (string_of_mode): Handle V4HF.
21814 * doc/arm-neon-intrinsics.texi: Regenerate.
21816 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21818 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
21819 format specifier in 'X' case.
21821 2013-04-25 Alan Modra <amodra@gmail.com>
21824 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
21825 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
21826 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
21827 Repeat for many other rotate/shift and mask patterns using subregs.
21828 Name lshiftrt insns.
21829 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
21830 on WORDS_BIG_ENDIAN.
21832 2013-04-25 Alan Modra <amodra@gmail.com>
21834 * config.gcc: Support little-endian powerpc-linux targets.
21835 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
21836 (LINK_OS_LINUX_SPEC): Define.
21837 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
21838 Preserve MASK_LITTLE_ENDIAN.
21839 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
21840 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
21841 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
21842 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
21843 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
21844 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
21845 Correct fp word order for little-endian. Don't shift toc entries
21846 smaller than a word for little-endian.
21847 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
21848 (bswapdi2 splits): Correct low-part subreg for little-endian.
21849 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
21850 low/high where such is correct only for be.
21851 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
21852 little-endian for -mcall-aixdesc.
21854 2013-04-25 Alan Modra <amodra@gmail.com>
21856 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
21857 replace_equiv_address_nv.
21859 2013-04-25 Alan Modra <amodra@gmail.com>
21861 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
21863 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
21866 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
21867 * rtl.h (struct rtx_def): ...
21869 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
21871 PR rtl-optimizations/57046
21872 * lra-constraints (split_reg): Set up lra_risky_transformations_p
21873 for multi-reg splits.
21875 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
21877 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
21879 2013-04-24 Sterling Augustine <saugustine@google.com>
21881 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
21882 (comp_dir_string, debug_str_dwo_section): New.
21883 (DEBUG_STR_DWO_SECTION): Rename to ...
21884 (DEBUG_DWO_STR_SECTION): ... this.
21885 (DEBUG_NORM_STR_SECTION): Delete.
21886 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
21887 (DEBUG_STR_DWO_SECTION_FLAGS): New.
21888 (find_AT_string): Move most logic to ...
21889 (find_AT_string_in_table): ... here. New.
21890 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
21891 add_skeleton_AT_string. Delete logic.
21892 (output_skeleton_debug_sections): Remove call to
21893 add_top_level_skeleton_die_attrs.
21894 (add_comp_dir_attribute): Move logic to comp_dir_string.
21895 (dwarf2out_init): Initialize debug_str_dwo_section.
21896 (output_indirect_string): Call find_string_form.
21897 (output_indirect_strings): Rewrite.
21898 (prune_unused_types): Empty skeleton_debug_str_hash.
21899 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
21900 (dwarf2out_finish): Call output_indirect_strings.
21902 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
21904 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
21906 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
21908 * rtl.h (struct rtx_def): Add comment for field jump.
21909 (LRA_SUBREG_P): New macro.
21910 * recog.c (register_operand): Check LRA_SUBREG_P.
21911 * lra.c (lra): Add note at the end of RTL code. Align non-empty
21913 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
21914 (lra_final_code_change): Skip subreg change for operators.
21915 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
21916 if there are no operand changes.
21917 * lra-constraints.c (curr_insn_set): New.
21918 (match_reload): Set LRA_SUBREG_P.
21919 (emit_spill_move): Ditto.
21920 (check_and_process_move): Use curr_insn_set. Process only single
21921 set insns. Don't initialize sec_mem_p and change_p.
21922 (simplify_operand_subreg): Use LRA_SUBREG_P.
21923 (reg_in_class_p): New function.
21924 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
21925 of #ifdef. Add code to remove cycling.
21926 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
21927 non-null disp. Reload inner instead of disp when base and index
21928 are null. Try to put lo_sum into register.
21929 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
21930 (check_and_process_move): Move code for move cost check to
21931 simple_move_p. Remove equiv_substitution.
21932 (simple_move_p): New function.
21933 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
21934 curr_insn_set. Call check_and_process_move only for single set
21935 insns. Use the new function. Move call of check_and_process_move
21936 after operand equiv substitution and address process.
21938 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
21940 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
21941 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
21942 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
21944 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
21946 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
21948 2013-04-24 Marek Polacek <polacek@redhat.com>
21950 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
21951 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
21952 (select_loops_exit_conditions): Likewise.
21953 (number_of_iterations_for_all_loops): Likewise.
21954 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
21955 (scev_analysis): Likewise.
21957 2013-04-02 Catherine Moore <clm@codesourcery.com>
21958 Chao-ying Fu <fu@mips.com>
21960 * config/mips/micromips.md (jraddiusp): New pattern.
21961 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
21962 instruction if possible.
21964 2013-04-24 Alan Modra <amodra@gmail.com>
21966 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
21968 2013-04-24 Julian Brown <julian@codesourcery.com>
21969 Chung-Lin Tang <cltang@codesourcery.com>
21971 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
21972 dependency behavior in enumeration type DIE generation. Add TODO note
21973 to comments about future DW_FORM_sdata/udata re-work of related code.
21975 2013-04-23 Lawrence Crowl <crowl@google.com>
21977 * Makefile.in: Update as needed below.
21979 * hash-table.h (class hash_table):
21980 Correct many methods with parameter types compare_type to the correct
21981 value_type. (Correct code was unlikely to notice the change.)
21982 (hash_table::elements_with_deleted) New.
21983 (class hashtable::iterator): New.
21984 (hashtable::begin()): New.
21985 (hashtable::end()): New.
21986 (FOR_EACH_HASH_TABLE_ELEMENT): New.
21988 * statistics.c (statistics_hashes):
21989 Change type to hash_table. Update dependent calls and types.
21991 * tree-into-ssa.c (var_infos):
21992 Change type to hash_table. Update dependent calls and types.
21994 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
21995 Change type to hash_table. Update dependent calls and types.
21997 * tree-ssa-loop-im.c (struct mem_ref.refs):
21998 Change type to hash_table. Update dependent calls and types.
22000 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
22001 Change type to hash_table. Update dependent calls and types.
22003 * tree-ssa-sccvn.c (vn_tables_s::nary):
22004 Change type to hash_table. Update dependent calls and types.
22005 (vn_tables_s::phis): Likewise.
22006 (vn_tables_s::references): Likewise.
22008 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
22009 (vn_reference_eq): Update parameter and return types.
22011 * tree-ssa-structalias.c (pointer_equiv_class_table):
22012 Change type to hash_table. Update dependent calls and types.
22013 (location_equiv_class_table): Likewise.
22015 * tree-vect-data-refs.c: Consequential changes for making
22016 peeling a hash_table.
22018 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
22019 (destroy_loop_vec_info): Dependent hash_table update.
22021 * tree-vectorizer.h (peeling_htab):
22022 Change type to hash_table. Update dependent calls and types.
22024 2013-04-23 Shiva Chen <shiva0217@gmail.com>
22026 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
22027 to check the register content is equal or not.
22028 * lra-constraints.c (match_reload): Use lra_assign_reg_val
22029 to assign register content record.
22030 * lra-eliminations.c (update_reg_eliminate): Use
22031 lra_update_reg_val_offset to update register content offset.
22032 * lra-int.h (struct lra_reg): Add offset member.
22033 (lra_reg_val_equal_p): New static inline function.
22034 (lra_update_reg_val_offset): New static inline function.
22035 (lra_assign_reg_val): New static inline function.
22036 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
22037 to assign register content record.
22038 (initialize_lra_reg_info_element): Initial offset to zero.
22040 2013-04-23 Catherine Moore <clm@codesourcery.com>
22042 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
22043 operands. Record compression.
22045 2013-04-23 Xinliang David Li <davidxl@google.com>
22047 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
22049 2013-04-23 Richard Biener <rguenther@suse.de>
22051 PR middle-end/57036
22052 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
22053 parameter, only add abnormal goto edges from the copied body
22054 if the call could perform abnormal gotos.
22055 (copy_cfg_body): Adjust.
22057 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
22059 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
22061 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
22063 * coretypes.h (gimple_stmt_iterator): Add struct to make
22066 2013-04-23 Richard Biener <rguenther@suse.de>
22068 PR tree-optimization/57026
22069 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
22070 from SSA names occuring in abnormal PHI nodes.
22072 2013-04-22 Andi Kleen <ak@linux.intel.com>
22074 * lto/lto.c (print_lto_report_1): Fix LTO report names.
22076 2013-04-22 Andi Kleen <ak@linux.intel.com>
22078 * lto/lto.c (print_lto_report_1): Declare early.
22079 (read_cgraph_and_symbols): Call print_lto_report_1 early.
22081 2013-04-22 Andi Kleen <ak@linux.intel.com>
22083 * common.opt (-flto-report-wpa): Add.
22084 * doc/invoke.texi (-flto-report-wpa): Add.
22085 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
22088 2013-04-22 Xinliang David Li <davidxl@google.com>
22090 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
22091 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
22092 * Makefile.in: New dependency
22094 David Daney <ddaney.cavm@gmail.com>
22096 * configure.ac (gcc_cv_as_micromips_support): Use the
22097 --fatal-warnings option.
22098 * configure: Regenerate.
22100 2013-04-22 Marek Polacek <polacek@redhat.com>
22103 * tsan.c (instrument_expr): Don't instrument expression
22104 in case its size is zero.
22106 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
22110 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22112 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22114 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22116 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
22117 (gimple_stmt_iterator): New typedef.
22118 * gimple.h (gimple_stmt_iterator): Rename to...
22119 (gimple_stmt_iterator_d): ... This.
22120 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
22121 trees be valid for GIMPLE and GENERIC.
22122 (TARGET_GIMPLE_FOLD_BUILTIN): New.
22123 * gimple-fold.c (gimple_fold_call): Call target hook
22124 gimple_fold_builtin.
22125 * hooks.c (hook_bool_gsiptr_false): New.
22126 * hooks.h (hook_bool_gsiptr_false): New.
22127 * target.def (fold_stmt): New.
22128 * doc/tm.texi: Regenerate.
22130 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
22133 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
22134 a set sp if no stack realignment.
22136 2013-04-22 Nick Clifton <nickc@redhat.com>
22138 * config.gcc (tilegx-linux): Extend extra_objs rather than
22140 (tilepro-linux): Likewise.
22142 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22144 * config/aarch64/aarch64-builtins.c
22146 (CF0, CF1, CF2, CF3, CF4, CF10): New.
22147 (VAR<1-12>): Add MAP parameter.
22148 (BUILTIN_*): Likewise.
22149 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
22150 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
22151 (aarch64_ushl_n<mode>): Likewise.
22152 (aarch64_sshr_n<mode>): Likewise.
22153 (aarch64_ushr_n<mode>): Likewise.
22154 (aarch64_<maxmin><mode>): Likewise.
22155 (aarch64_sqrt<mode>): Likewise.
22156 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
22157 (vshr<q>_n_*): Likewise.
22159 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22161 * config/aarch64/aarch64-builtins.c
22162 (aarch64_simd_builtin_type_mode): Handle SF types.
22164 (BUILTIN_GPF): Define.
22165 (aarch64_init_simd_builtins): Handle SF types.
22166 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
22167 (frecps): Likewise.
22168 (frecpx): Likewise.
22169 * config/aarch64/aarch64-simd.md
22170 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
22171 (aarch64_frecpe<mode>): New.
22172 (aarch64_frecps<mode>): Likewise.
22173 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
22174 (v8type): Add frecp<esx>.
22175 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
22176 (aarch64_frecps<mode>): Likewise.
22177 * config/aarch64/iterators.md (FRECP): New.
22178 (frecp_suffix): Likewise.
22179 * config/aarch64/arm_neon.h
22180 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
22182 2013-04-22 Christian Bruel <christian.bruel@st.com>
22185 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
22186 (REG_CLASS_NAMES): Idem.
22187 (REG_CLASS_CONTENTS): Idem.
22188 (REGCLASS_HAS_FP_REG): Idem.
22189 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
22190 (sh_conditional_register_usage): Idem.
22192 2013-04-21 Jeff Law <law@redhat.com>
22194 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
22195 (ssa_forward_propagate_and_combine): Use it.
22197 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22199 * lra.c: Update the flow chart diagram.
22201 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22203 PR rtl-optimization/56847
22204 * lra-constraints.c (process_alt_operands): Discourage alternative
22205 with non-matche doffsettable memory constraint fro memory with
22208 2013-04-19 Richard Biener <rguenther@suse.de>
22210 PR tree-optimization/56982
22211 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
22213 * gimplify.c (gimplify_call_expr): Notice special calls.
22214 (gimplify_modify_expr): Likewise.
22215 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
22216 abnormal control flow receivers.
22217 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
22218 in the same way as cfun->has_nonlocal_labels.
22219 (gimple_purge_dead_abnormal_call_edges): Likewise.
22220 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
22221 receivers start a basic-block.
22223 2013-04-19 Richard Biener <rguenther@suse.de>
22225 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
22227 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
22228 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
22229 (SLP_TREE_LOAD_PERMUTATION): Add.
22230 (vect_transform_slp_perm_load): Adjust prototype.
22231 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
22232 (vect_free_slp_instance): Likewise.
22233 (vect_create_new_slp_node): Likewise.
22234 (vect_supported_slp_permutation_p): Remove.
22235 (vect_slp_rearrange_stmts): Adjust.
22236 (vect_supported_load_permutation_p): Likewise. Inline
22237 vect_supported_slp_permutation_p here.
22238 (vect_analyze_slp_instance): Compute load permutations per
22239 slp node instead of per instance.
22240 (vect_get_slp_defs): Adjust.
22241 (vect_transform_slp_perm_load): Likewise.
22242 (vect_schedule_slp_instance): Remove redundant code.
22243 (vect_schedule_slp): Remove hack for PR56270, add it ...
22244 * tree-vect-stmts.c (vectorizable_load): ... here, do not
22245 CSE loads for SLP. Adjust.
22247 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22249 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
22250 spelling in two comments.
22252 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22255 * config/arm/arm.c (load_multiple_sequence): Require SP
22256 as base register for loads if SP is in the register list.
22258 2013-04-19 Martin Jambor <mjambor@suse.cz>
22260 PR tree-optimization/56718
22261 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
22262 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
22263 and made public. Adjusted all callers.
22264 (ipa_intraprocedural_devirtualization): New function.
22265 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
22266 (ipa_intraprocedural_devirtualization): Likewise.
22267 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
22269 2013-04-19 Richard Biener <rguenther@suse.de>
22271 PR tree-optimization/57000
22272 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
22274 2013-04-19 Terry Guo <terry.guo@arm.com>
22276 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
22278 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
22279 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
22280 (cortex_m4_fmacs): Use new reservations.
22281 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
22283 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
22285 PR rtl-optimization/56999
22286 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
22288 (lra_coalesce): Remove split_origin_bitmap and related code.
22289 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
22290 ranges if necessary.
22292 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
22294 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
22296 (ix86_expand_call): Remove clobbered_registers array and use
22297 x86_64_ms_sysv_extra_clobbered_registers instead.
22298 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
22300 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
22302 * config/i386/i386.md (*call_rex64_ms_sysv): Use
22303 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
22304 (*call_value_rex64_ms_sysv): Ditto.
22306 2013-04-18 Cary Coutant <ccoutant@google.com>
22308 * dwarf2out.c (output_pubnames): Check die_perennial_p of
22309 parent instead of die_mark.
22311 2013-04-18 Diego Novillo <dnovillo@google.com>
22313 * gimple.c (create_gimple_tmp): New.
22314 (get_expr_type): New.
22315 (build_assign): New.
22316 (build_type_cast): New.
22317 * gimple.h (enum ssa_mode): Define.
22318 (gimple_seq_set_location): New.
22319 * asan.c (build_check_stmt): Change some gimple_build_* calls
22320 to use build_assign and build_type_cast.
22322 2013-04-18 Richard Biener <rguenther@suse.de>
22324 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
22325 handle negative step. Remove redundant checks.
22326 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
22327 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
22328 for negative step and grouped loads fail to vectorize.
22330 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22332 * emit-rtl.c (reset_insn_used_flags): New function.
22333 (reset_all_used_flags): Use it.
22334 (verify_insn_sharing): New function.
22335 (verify_rtl_sharing): Fix verification for SEQUENCEs.
22337 2013-04-18 Jakub Jelinek <jakub@redhat.com>
22339 PR tree-optimization/56984
22340 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
22341 and (x >> M) >= N don't register any assertion if N << M is the
22344 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22346 * lower-subreg.c (resolve_simple_move): If called self-recursive,
22347 do not delete_insn insns that have not yet been emitted, only
22348 unlink them with remove_insn.
22349 * df-scan.c (df_insn_delete): Revert r197492.
22351 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
22353 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
22354 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
22356 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22358 * config/arm/arm.md (movsicc_insn): Convert define_insn into
22359 define_insn_and_split.
22360 (and_scc,ior_scc,negscc): Likewise.
22361 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
22363 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22365 * config/arm/arm.c (use_return_insn): Return 0 for targets that
22366 can benefit from using a sequence of LDRD instructions in epilogue
22367 instead of a single LDM instruction.
22369 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
22372 * doc/extend.texi: Fix typo.
22374 2013-04-17 Richard Biener <rguenther@suse.de>
22376 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
22377 (vect_build_slp_tree): ... here.
22378 (vect_build_slp_tree_1): Compute which stmts of the SLP group
22379 match. Remove special-casing of mismatched complex loads.
22380 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
22381 re-try the match with swapped commutative operands.
22382 (vect_supported_load_permutation_p): Remove special-casing of
22383 mismatched complex loads.
22384 (vect_analyze_slp_instance): Adjust.
22386 2013-04-17 Richard Biener <rguenther@suse.de>
22388 PR rtl-optimization/56921
22389 * cfgloop.h (struct loop): Add simple_loop_desc member.
22390 (struct niter_desc): Mark with GTY(()).
22391 (simple_loop_desc): Do not use aux field but simple_loop_desc.
22392 * loop-iv.c (get_simple_loop_desc): Likewise.
22393 (free_simple_loop_desc): Likewise.
22396 2013-04-16 Richard Biener <rguenther@suse.de>
22398 PR rtl-optimization/56921
22399 * loop-init.c (pass_rtl_move_loop_invariants): Add
22400 TODO_do_not_ggc_collect to todo_flags_finish.
22401 (pass_rtl_unswitch): Same.
22402 (pass_rtl_unroll_and_peel_loops): Same.
22403 (pass_rtl_doloop): Same.
22405 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
22407 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
22408 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
22409 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
22411 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
22412 * tree-streamer.c (record_common_node): Adjust reference in comment.
22414 2013-04-17 Terry Guo <terry.guo@arm.com>
22416 * config/arm/cortex-m4.md: Add a new bypass.
22418 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22420 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
22422 (*subs_<optab><mode>_multp2): New pattern.
22423 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22424 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22426 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22428 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
22429 (*subs_mul_imm_<mode>): New pattern.
22431 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
22434 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
22435 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
22436 (vsx_movti_32bit): Change j->wa to O->wa.
22438 2013-04-16 Richard Biener <rguenther@suse.de>
22440 PR rtl-optimization/56921
22441 * loop-init.c (pass_rtl_move_loop_invariants): Add
22442 TODO_do_not_ggc_collect to todo_flags_finish.
22443 (pass_rtl_unswitch): Same.
22444 (pass_rtl_unroll_and_peel_loops): Same.
22445 (pass_rtl_doloop): Same.
22447 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
22449 * config/arm/arm.c (emit_multi_reg_push): New declaration
22450 for an existing function.
22451 (arm_emit_strd_push): New function.
22452 (arm_expand_prologue): Used here.
22453 (arm_emit_ldrd_pop): New function.
22454 (arm_expand_epilogue): Used here.
22455 (arm_get_frame_offsets): Update condition.
22456 (arm_emit_multi_reg_pop): Add a special case for load of a single
22457 register with writeback.
22459 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
22461 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
22464 2013-04-16 Richard Biener <rguenther@suse.de>
22466 PR tree-optimization/56756
22467 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
22468 (first_mem_ref_loc): New.
22469 (execute_sm): Place the load temporarily before a previous
22470 access instead of in the latch edge to ensure its SSA dependencies
22471 are defined at points dominating the load.
22473 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
22475 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
22476 correct fix by moving header and footer insn to the footer of
22477 the merged basic block. Clear BB_END of the merged-away block.
22479 PR middle-end/43631
22480 * emit-rtl.c (make_note_raw): New function.
22481 (link_insn_into_chain): New static inline function.
22482 (add_insn): Use it.
22483 (add_insn_before, add_insn_after): Factor insn chain linking code...
22484 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
22485 using link_insn_into_chain.
22486 (note_outside_basic_block_p): New helper function for emit_note_after
22487 and emit_note_before.
22488 (emit_note_after): Use nobb variant of add_insn_after if the note
22489 should not be contained in a basic block.
22490 (emit_note_before): Use nobb variant of add_insn_before if the note
22491 should not be contained in a basic block.
22492 (emit_note_copy): Use make_note_raw.
22493 (emit_note): Likewise.
22494 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
22495 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
22496 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
22497 the moved barrier the tail of the basic block it follows.
22498 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
22500 2013-04-15 Jakub Jelinek <jakub@redhat.com>
22502 PR tree-optimization/56962
22503 * gimple-ssa-strength-reduction.c (record_increment): Only set
22504 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
22505 either rhs1 or rhs2 is equal to c->base_expr.
22507 2013-04-15 Richard Biener <rguenther@suse.de>
22509 PR tree-optimization/56933
22510 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
22512 (GROUP_READ_WRITE_DEPENDENCE): Remove.
22513 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
22514 * tree-vect-data-refs.c (vect_analyze_group_access): Move
22515 dependence check ...
22516 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
22518 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
22519 GROUP_READ_WRITE_DEPENDENCE.
22521 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22523 * emit-rtl.c (reset_all_used_flags): New function.
22524 (verify_rtl_sharing): Call reset_all_used_flags before and after
22525 performing the checks.
22527 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22529 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
22530 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
22531 * config/arm/constraints.md (De): New constraint.
22532 * config/arm/neon.md (anddi3_neon): Delete.
22533 (neon_vand<mode>): Expand to standard anddi3 pattern.
22534 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
22535 Move earlier in the file.
22536 (neon_inv_logic_op2): Likewise.
22537 (arm_anddi_operand_neon): New predicate.
22539 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22541 * configure.ac (gcc_cv_ld_as_needed): Set
22542 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
22543 Use -z ignore, -z record on *-*-solaris2*.
22544 (HAVE_LD_AS_NEEDED): Update comment.
22545 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
22546 * configure: Regenerate.
22547 * config.in: Regenerate.
22548 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
22549 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
22550 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
22551 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
22552 equivalents. Fix markup.
22553 * doc/tm.texi: Regenerate.
22555 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
22557 * config/i386/i386.opt: New option mstack-protector-guard=.
22558 * config/i386/i386-opts.h: Add enum stack_protector_guard.
22559 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
22560 TARGET_SSP_TLS_GUARD.
22561 * config/i386/i386.c (ix86_option_override_internal): Set
22562 ix86_stack_protector_guard.
22563 * config/i386/i386.md (stack_protect_set): Enable for
22564 TARGET_SSP_TLS_GUARD only.
22565 (stack_protect_set_<mode>): Ditto.
22566 (stack_protect_test): Ditto.
22567 (stack_protect_test_<mode>): Ditto.
22568 * doc/invoke.texi (i386 Option): Document.
22570 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
22573 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
22574 (S_MODES): Set H_MODE bit.
22575 (SF_MODES): Set only S_MODE and SF_MODE bits.
22576 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
22577 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
22578 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
22579 <MODE_FLOAT>: Likewise.
22581 2013-04-15 Joey Ye <joey.ye@arm.com>
22583 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
22585 2013-04-15 Joey Ye <joey.ye@arm.com>
22587 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
22589 (thumb_far_jump_used_p): Count instruction size and set
22592 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
22594 * reorg.c (fill_simple_delay_slots): Reindent block of code.
22595 * resource.c (mark_target_live_regs): Reformat conditional block.
22597 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22599 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
22600 notes, they are emitted only just before final.
22601 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
22603 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22605 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
22606 * cfgrtl.c (delete_insn): Call it here instead.
22607 * lra-spills.c (lra_final_code_change): Use delete_insn.
22608 * haifa-sched.c (sched_remove_insn): Likewise.
22609 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
22610 returning to the nop pool.
22611 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
22612 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
22614 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22616 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
22617 * doc/tm.texi: Regenerated.
22619 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
22621 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
22624 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22626 * df-core.c (df_find_def): Compare register numbers.
22627 (df_find_use): Likewise.
22629 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
22632 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
22633 lra_in_progress for return.
22635 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22637 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
22638 define_insn into define_insn_and_split and emit movsicc patterns.
22640 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22642 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
22644 2013-04-12 Richard Biener <rguenther@suse.de>
22646 * tree-pass.h (TODO_do_not_ggc_collect): New.
22647 * passes.c (execute_one_ipa_transform_pass): Honor
22648 TODO_do_not_ggc_collect.
22649 (execute_one_pass): Likewise.
22652 2013-04-10 Richard Biener <rguenther@suse.de>
22654 * passes.c (init_optimization_passes): Remove reload pass.
22655 * ira.c (do_reload): Merge into ...
22657 (rest_of_handle_reload): Remove.
22658 (pass_reload): Likewise.
22659 * config/i386/i386.c (ix86_option_override): Refer to ira instead
22660 of reload for vzeroupper pass placement.
22662 2013-04-12 Jakub Jelinek <jakub@redhat.com>
22664 PR tree-optimization/56918
22665 PR tree-optimization/56920
22666 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
22667 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
22668 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
22669 use wide_mul_with_sign method.
22671 2013-04-12 Richard Biener <rguenther@suse.de>
22673 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
22674 not be considered a gimple constant.
22676 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
22678 * fold-const.c (const_binop): Handle vector shifts by a scalar.
22679 (fold_binary_loc): Call const_binop also for mixed vector-scalar
22682 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
22683 Jakub Jelinek <jakub@redhat.com>
22685 * opts.c: Include diagnostic-color.h.
22686 (common_handle_option): Handle OPT_fdiagnostics_color_.
22687 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
22688 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
22689 (diagnostic-color.o): New.
22690 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
22691 (diagnostic_color_rule): New enum.
22692 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
22693 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
22694 the location string.
22695 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
22696 either NULL, or color kind.
22697 * diagnostic-color.c: New file.
22698 * diagnostic-color.h: New file.
22699 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
22701 * doc/invoke.texi (-fdiagnostics-color): Document.
22702 * pretty-print.h (pp_show_color): Define.
22703 (struct pretty_print_info): Add show_color field.
22704 * diagnostic.c: Include diagnostic-color.h.
22705 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
22706 macros. Colorize error:, warning: etc. strings and also the location
22708 (diagnostic_show_locus): Colorize the caret line.
22709 * pretty-print.c: Include diagnostic-color.h.
22710 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
22711 inside of %< %> quotes or quoted through q format modifier.
22713 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22715 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
22717 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22719 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
22720 code in CC_NZ mode.
22721 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
22724 2013-04-11 Marek Polacek <polacek@redhat.com>
22726 PR tree-optimization/48184
22727 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
22729 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
22731 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
22732 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
22733 (skip_simple_arithmetic): Tidy up.
22734 * tree.h (skip_simple_constant_arithmetic): Declare.
22736 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22738 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
22740 2013-04-11 Richard Biener <rguenther@suse.de>
22742 * tree-vect-loop.c (get_initial_def_for_induction): Properly
22743 generate vector constants.
22745 2013-04-11 Richard Biener <rguenther@suse.de>
22747 PR tree-optimization/56878
22748 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
22749 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
22751 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
22752 Prefer to align the DR with the most invariant base address.
22754 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
22756 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
22759 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
22761 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
22762 floating-point vector comparisons against 0.
22764 2013-04-11 Jakub Jelinek <jakub@redhat.com>
22766 PR tree-optimization/56899
22767 * fold-const.c (extract_muldiv_1): Apply distributive law
22768 only if TYPE_OVERFLOW_WRAPS (ctype).
22770 2013-04-11 Bin Cheng <bin.cheng@arm.com>
22773 * ira-costs.c (scan_one_insn): Check whether the source rtx of
22774 loading has side effect.
22776 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
22778 * config/sparc/sparc.c: Include tree-pass.h.
22779 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
22780 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
22781 head of file. Change return type. Split off gate function.
22782 (sparc_gate_work_around_errata): New function.
22783 (pass_work_around_errata): New pass definition.
22784 (insert_pass_work_around_errata) New pass insert definition to
22785 insert pass_work_around_errata just after delayed-branch scheduling.
22786 (sparc_option_override): Insert the pass.
22787 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
22789 2013-04-10 David S. Miller <davem@davemloft.net>
22791 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
22792 or -mcpu=hypersparc.
22794 * target.def (cstore_mode): New hook.
22795 * target.h: Include insn-codes.h
22796 * targhooks.c: Likewise.
22797 (default_cstore_mode): New function.
22798 * targhooks.h: Declare it.
22799 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
22800 * doc/tm.texi: Rebuild.
22801 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
22802 target hook, rather than inspecting the insn_data.
22803 * config/sparc/sparc.c (sparc_cstore_mode): New function.
22804 (TARGET_CSTORE_MODE): Redefine.
22805 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
22807 * config/sparc/predicates.md (cstore_result_operand): New special
22809 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
22810 Use it for operand 0.
22811 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
22812 (*snesi_special): Likewise.
22813 (*snesi_zero): Likewise.
22814 (*seqsi_zero): Likewise.
22815 (*sltu_insn): Likewise.
22816 (*sgeu_insn): Likewise.
22817 (*seqdi_special): Make operand 0 and comparison operation be of
22819 (*snedi_special): Likewise.
22820 (*snedi_special_vis3): Likewise.
22821 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
22822 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
22823 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
22824 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
22825 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
22826 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
22827 (*sltu_extend_sp64): Likewise.
22828 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
22829 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
22830 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
22831 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
22832 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
22834 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
22836 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
22837 (aarch64_start_file): Use the new function.
22839 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
22840 Jason Merrill <jason@redhat.com>
22842 * common.opt: Add -gdwarf.
22843 * opts.c (common_handle_option): Handle it.
22844 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
22846 2013-04-10 Richard Biener <rguenther@suse.de>
22848 * passes.c (execute_todo): Do not call ggc_collect conditional here.
22849 (execute_one_ipa_transform_pass): But unconditionally here.
22850 (execute_one_pass): And here.
22851 (init_optimization_passes): Remove reload pass.
22852 * tree-pass.h (TODO_ggc_collect): Remove.
22853 (pass_reload): Likewise.
22854 * ira.c (do_reload): Merge into ...
22856 (rest_of_handle_reload): Remove.
22857 (pass_reload): Likewise.
22858 * config/i386/i386.c (ix86_option_override): Refer to ira instead
22859 of reload for vzeroupper pass placement.
22860 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
22861 and todo_flags_finish of all passes.
22863 2013-04-10 Richard Biener <rguenther@suse.de>
22865 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
22866 first_const_oprnd field, rename first_def_type to first_op_type.
22867 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
22868 (vect_get_and_check_slp_defs): Always use the type of the
22869 operand. Allow mixed vect_external_def, vect_constant_def types.
22870 (vect_get_constant_vectors): Handle mixed vect_external_def,
22871 vect_constant_def types.
22873 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
22875 PR tree-optimization/55524
22876 * tree-ssa-math-opts.c
22877 (convert_mult_to_fma): Don't use an fms construct
22878 when we don't have an fms operation, but fnma, and it looks
22879 likely that we'll be able to use the latter.
22881 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
22883 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
22885 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
22886 inline fail caused by overwritable functions.
22888 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
22890 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
22891 unnecessary bits in the constant power of two case.
22893 2013-04-10 Richard Biener <rguenther@suse.de>
22895 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
22896 broken code swapping operands.
22897 (vect_build_slp_tree): Do not compute load permutations here.
22898 (vect_analyze_slp_instance): Compute load permutations here,
22899 after building the SLP tree.
22901 2013-04-09 Christian Bruel <christian.bruel@st.com>
22903 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
22904 of next/prev_real_insn.
22906 2013-04-09 Jan Hubicka <jh@suse.cz>
22908 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
22909 Drop aliased parameter.
22910 (function_and_variable_visibility): Do not handle alias pairs.
22911 * cgraph.c (varpool_externally_visible_p): Update prototype.
22912 * varpool.c (varpool_add_new_variable): Update.
22914 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22916 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
22918 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
22920 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
22922 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
22924 2013-04-09 Marek Polacek <polacek@redhat.com>
22926 PR tree-optimization/48762
22927 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
22929 2013-04-09 Richard Biener <rguenther@suse.de>
22931 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
22933 (vect_build_slp_tree): Likewise.
22934 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
22935 calculating the cost of a SLP instance.
22936 (vect_analyze_slp_instance): Use it from here, after building
22939 2013-04-09 Jakub Jelinek <jakub@redhat.com>
22941 PR middle-end/56883
22942 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
22943 expand_omp_for_static_chunk): Use simple_p = true in
22944 force_gimple_operand_gsi calls when assigning to addressable decls.
22946 2013-04-09 Jeff Law <law@redhat.com>
22948 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
22949 when the boolean was created by converting a wider object which
22950 had a boolean range.
22952 2013-04-09 Richard Biener <rguenther@suse.de>
22954 * tree-vectorizer.h (slp_void_p): Remove.
22955 (slp_tree): Typedef before _slp_tree declaration.
22956 (struct _slp_tree): Use a vector of slp_tree as children.
22957 (vect_get_place_in_interleaving_chain): Remove.
22958 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
22960 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
22962 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
22963 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
22964 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
22965 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
22966 Use slp_node instead of slp_void_p and adjust.
22968 2013-04-09 Richard Biener <rguenther@suse.de>
22970 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
22971 work that is not necessary.
22973 2013-04-09 Jakub Jelinek <jakub@redhat.com>
22975 PR tree-optimization/56854
22976 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
22977 forward into clobber stmts if it would change MEM_REF lhs into
22980 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
22982 * tree.c (type_hash_lookup, type_hash_add): Make static.
22983 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
22985 2013-04-09 Richard Biener <rguenther@suse.de>
22987 * tree.h (unsave_expr_now): Remove.
22988 * tree-inline.c (mark_local_for_remap_r): Remove.
22989 (unsave_expr_1): Likewise.
22990 (unsave_r): Likewise.
22991 (unsave_expr_now): Likewise.
22992 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
22993 (propagate_tree_value): Likewise.
22995 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
22997 * doc/rtl.texi (sequence): Rewrite documentation to match the
22998 current use of SEQUENCE rtl objects.
22999 * rtl.def (SEQUENCE): Likewise.
23001 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
23002 Update documentation.
23003 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
23004 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
23006 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
23008 2013-04-08 Teresa Johnson <tejohnson@google.com>
23010 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
23011 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
23013 (estimate_edge_size_and_time): Add comment to suggest using rounding
23015 (estimate_node_size_and_time): Ditto.
23016 (remap_edge_change_prob): Use helper rounding divide methods.
23017 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
23018 (gimple_mod_pow2_value_transform): Ditto.
23019 (gimple_mod_subtract_transform): Ditto.
23020 (gimple_ic_transform): Ditto.
23021 (gimple_stringops_transform): Ditto.
23022 * stmt.c (conditional_probability): Ditto.
23023 (emit_case_dispatch_table): Ditto.
23024 * lto-cgraph.c (merge_profile_summaries): Ditto.
23025 * tree-optimize.c (execute_fixup_cfg): Ditto.
23026 * cfgcleanup.c (try_forward_edges): Ditto.
23027 * cfgloopmanip.c (scale_loop_profile): Ditto.
23029 (duplicate_loop_to_header_edge): Ditto.
23030 (lv_adjust_loop_entry_edge): Ditto.
23031 * tree-vect-loop.c (vect_transform_loop): Ditto.
23032 * profile.c (compute_branch_probabilities): Ditto.
23033 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
23034 * lto-streamer-in.c (input_cfg): Ditto.
23035 * gimple-streamer-in.c (input_bb): Ditto.
23036 * ipa-cp.c (update_profiling_info): Ditto.
23037 (update_specialized_profile): Ditto.
23038 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
23039 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
23041 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23042 (compute_trg_info): Ditto.
23043 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
23044 (purge_dead_edges): Ditto.
23045 * loop-unswitch.c (unswitch_loop): Ditto.
23046 * cgraphclones.c (cgraph_clone_edge): Ditto.
23047 (cgraph_clone_node): Ditto.
23048 * tree-inline.c (copy_bb): Ditto.
23049 (copy_edges_for_bb): Ditto.
23050 (initialize_cfun): Ditto.
23051 (copy_cfg_body): Ditto.
23052 (expand_call_inline): Ditto.
23054 2013-04-08 Kai Tietz <ktietz@redhat.com>
23056 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
23057 TARGET_CYGWIN64 by TARGET_64BIT.
23059 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23061 * config/epiphany/epiphany.md (GPR_1): New constant.
23062 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
23063 * config/epiphany/epiphany.c (gen_compare_reg):
23064 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
23065 is already in place.
23066 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
23067 Don't require being called during rtl expansion; If y operlaps r0,
23069 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
23070 (epiphany_expand_epilogue): Likewise.
23072 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
23073 Don't use CC_FPmode for ORDERED / UNORDERED.
23074 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
23076 * config/epiphany/constraints.md (CnL): New constraint.
23077 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
23078 * config/epiphany/predicates.md (add_operand): Allow 1024.
23080 * config/epiphany/epiphany.md (logical_op): New code iterator.
23081 (op_mnc): New code attribute.
23082 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
23083 (mov_f+1, mov_f+2): New peephole2 patterns.
23085 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
23086 (cstoresi4): Also allow re-use of zero result when doing a NE
23087 comparison to a non-zero operand.
23088 Use (clobber (scratch)) for first insn if the gpr output is not needed.
23090 * config/epiphany/epiphany.md (<insn_opname>v2si3):
23091 Use gen_addsi3_i / gen_subsi3_i.
23093 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23097 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
23098 contain anything but clobbers, at most one __builtin_stack_restore,
23099 optionally debug stmts and final resx, and if it has at least one
23100 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
23101 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
23102 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
23103 which isn't defaut definition, remove them.
23104 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
23105 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
23106 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
23107 with MEM_REF LHS with SSA_NAME address.
23109 2013-04-08 Jeff Law <law@redhat.com>
23111 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
23113 2013-04-08 Richard Biener <rguenther@suse.de>
23115 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
23117 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
23118 determined vector type.
23119 (vect_analyze_data_refs): Likewise.
23120 (vect_get_new_vect_var): Adjust.
23121 (vect_create_destination_var): Preserve SSA name versions.
23122 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
23123 not dump anything here.
23125 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23127 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
23128 Add member lr_slot_known.
23129 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
23131 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
23132 Remove code that sets lr_slot_offset according to what a previous
23133 version of epiphany_emit_save_restore used to do.
23134 (epiphany_emit_save_restore): When doing an lr save or restore,
23135 set/verify lr_slot_known and lr_slot_offset.
23137 2013-04-08 Xinyu Qi <xyqi@marvell.com>
23140 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
23143 2013-04-08 Richard Biener <rguenther@suse.de>
23145 * alias.c (find_base_term): Fix thinko in previous change.
23147 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23149 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
23150 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
23151 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
23152 if possible to compute val.
23153 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
23154 For QImode integers don't require anything about precision. Use
23155 const_with_all_bytes_same to find out if the constant doesn't have
23156 repeated bytes in it.
23158 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23160 * config/s390/s390.c (s390_expand_insv): Only accept insertions
23163 2013-04-08 Marek Polacek <polacek@redhat.com>
23165 PR rtl-optimization/48182
23166 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
23169 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23172 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
23173 nuses, make sure we have a label.
23175 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23178 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
23179 (rs6000_emit_swdiv_low_precision): Remove.
23180 (rs6000_emit_swdiv): Rewrite to handle between one and four
23181 iterations of Newton-Raphson generally; modify required number of
23182 iterations for some cases.
23183 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
23185 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
23187 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
23188 set-but-unused variable.
23190 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
23191 basic blocks of released function bodies garbage-collectable.
23193 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
23194 (struct rtl_opt_pass): Add TODO_df_finish.
23196 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
23198 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23200 * config/arm/constraints.md (q): New constraint.
23201 * config/arm/ldrdstrd.md: New file.
23202 * config/arm/arm.md (ldrdstrd.md) New include.
23203 (arm_movdi): Use "q" instead of "r" constraint
23204 for double-word memory access.
23205 (movdf_soft_insn): Likewise.
23206 * config/arm/vfp.md (movdi_vfp): Likewise.
23207 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
23208 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
23209 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
23210 (mem_ok_for_ldrd_strd): Likewise.
23211 (output_move_double): Update assertion.
23213 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23215 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
23217 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23219 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
23220 define_insn_and_split.
23221 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
23223 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23225 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
23226 define_insn_and_split.
23227 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
23228 (shiftsi3_compare): New pattern.
23229 (rrx): New pattern.
23230 * config/arm/unspecs.md (UNSPEC_RRX): New.
23232 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23234 * config/arm/arm.md (negdi_extendsidi): New pattern.
23235 (negdi_zero_extendsidi): Likewise.
23237 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23239 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
23240 define_insn_and_split.
23241 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
23242 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
23244 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23246 * config/arm/arm.md (arm_subdi3): Convert define_insn into
23247 define_insn_and_split.
23248 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
23249 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
23251 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23253 * config/arm/arm.md (subsi3_carryin): New pattern.
23254 (subsi3_carryin_const): Likewise.
23255 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
23256 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
23258 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23260 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
23262 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23264 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
23265 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
23267 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23269 * config/arm/arm.c (arm_expand_builtin): Change fcode
23270 type to unsigned int.
23272 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23274 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
23276 2013-04-04 Ian Lance Taylor <iant@google.com>
23278 * doc/standards.texi (Standards): The Go frontend supports the Go 1
23281 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
23283 PR middle-end/56729
23284 * df-scan.c (df_insn_delete): Disable failing assert.
23286 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23288 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
23289 New function prototype.
23290 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
23291 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
23292 (arm_builtin_vectorized_function): New function.
23294 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23296 * config/arm/arm_neon_builtins.def: New file.
23297 * config/arm/arm.c (neon_builtin_data): Move contents to
23298 arm_neon_builtins.def.
23299 (enum arm_builtins): Include neon builtin definitions.
23300 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
23301 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
23303 2013-04-04 Marek Polacek <polacek@redhat.com>
23305 PR tree-optimization/48186
23306 * predict.c (maybe_hot_frequency_p): Return false if
23307 HOT_BB_FREQUENCY_FRACTION is 0.
23308 (cgraph_maybe_hot_edge_p): Likewise.
23310 2013-04-04 Richard Biener <rguenther@suse.de>
23312 PR tree-optimization/56826
23313 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
23316 2013-04-04 Richard Biener <rguenther@suse.de>
23318 PR tree-optimization/56213
23319 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
23320 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
23322 2013-04-04 Richard Biener <rguenther@suse.de>
23324 PR tree-optimization/56837
23325 * tree-loop-distribution.c (classify_partition): For non-zero
23326 values require that the value has the same precision as its
23327 mode to be useful as memset value.
23329 2013-04-03 Nick Clifton <nickc@redhat.com>
23331 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
23332 (fmssf4): Use fmsf.s on E3V5 architectures.
23333 (fnmasf4): Use fnmaf.s on E3V5 architectures.
23334 (fnmssf4): Use fnmsf.s on E3V5 architectures.
23336 2013-04-03 Jeff Law <law@redhat.com>
23338 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
23339 (lra-eliminations.o): Likewise.
23341 2013-04-03 Teresa Johnson <tejohnson@google.com>
23343 * gcov-io.c (compute_working_sets): Moved most of body of old
23344 compute_working_sets here from profile.c.
23345 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
23346 (gcov_working_set_t): Moved typedef here from basic-block.h
23347 (compute_working_set): Declare.
23348 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
23349 (get_working_sets): Renamed from compute_working_set,
23350 replace most of body with call to new compute_working_sets.
23351 (get_exec_counts): Replace call to compute_working_sets
23352 to get_working_sets.
23353 * profile.h (get_working_sets): Renamed from compute_working_set.
23354 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
23355 to get_working_sets.
23356 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
23357 * gcov-dump.c (dump_working_sets): New function.
23359 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
23361 * hwint.c (sext_hwi, zext_hwi): New functions.
23362 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
23363 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
23364 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
23365 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
23366 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
23367 (sext_hwi, zext_hwi): New functions.
23369 2013-04-03 Jeff Law <law@redhat.com>
23371 PR tree-optimization/56799
23372 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
23373 back test for widening conversion erroneously dropped in prior change.
23375 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23378 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
23379 instead of next_real_insn.
23381 2013-04-03 Marek Polacek <polacek@redhat.com>
23384 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
23386 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23389 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
23391 (thumb1_output_casesi): Likewise.
23392 (thumb2_output_casesi): Likewise.
23394 2013-04-03 Richard Biener <rguenther@suse.de>
23396 PR tree-optimization/56817
23397 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
23399 (tree_unroll_loops_completely_1): ... new function to manually
23400 walk the loop tree, properly defering outer loops of unrolled
23401 loops to later iterations.
23403 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23405 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
23406 (vectorizable_load): Likewise.
23407 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
23408 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
23410 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23412 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
23415 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
23417 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
23419 2013-04-03 Bin Cheng <bin.cheng@arm.com>
23421 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
23423 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23425 PR tree-optimization/56790
23426 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
23429 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23431 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
23433 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
23434 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
23437 2013-04-03 Jakub Jelinek <jakub@redhat.com>
23440 * tree.h (force_folding_builtin_constant_p): New decl.
23441 * builtins.c (force_folding_builtin_constant_p): New variable.
23442 (fold_builtin_constant_p): Fold immediately also if
23443 force_folding_builtin_constant_p.
23445 2013-04-03 Richard Biener <rguenther@suse.de>
23447 PR tree-optimization/56812
23448 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
23449 DRs of the same interleaving chain are independent.
23451 2013-04-02 Jason Merrill <jason@redhat.com>
23453 * gdbinit.in (pbb): Use debug fn.
23455 2013-04-02 Lawrence Crowl <crowl@google.com>
23457 * sese.h (struct ivtype_map_elt_s): Remove unused.
23458 (extern debug_ivtype_map): Remove unused.
23459 (extern eq_ivtype_map_elts): Remove unused.
23460 * sese.c (debug_ivtype_map): Removed unused.
23461 (debug_ivtype_map_1): Removed unused.
23462 (debug_ivtype_elt): Remove unused.
23463 (eq_ivtype_map_elts): Remove unused.
23466 2013-04-02 Kai Tietz <ktietz@redhat.com>
23469 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
23470 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
23471 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
23473 (legitimize_pe_coff_symbol): Likewise.
23474 (is_imported_p): New helper-function.
23475 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
23476 for Windows x64 targets.
23477 (ix86_expand_prologue): Optimize for pe-coff targets.
23478 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
23479 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
23480 medium/large code-model.
23481 (legitimize_pic_address): Likewise.
23482 (legitimize_tls_address): Likewise.
23483 (ix86_expand_call): Likewise.
23484 (x86_output_mi_thunk): Likewise.
23485 (get_dllimport_decl): Add new beimport argument.
23486 (construct_plt_address): Don't assert for x64 pe-coff targets.
23487 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
23489 (SYMBOL_FLAG_STUBVAR): New macro.
23490 (SYMBOL_REF_STUBVAR_P): Likewise.
23491 * config/i386/winnt.c (stub_list): New structure.
23492 (stub_head): New local variable.
23493 (i386_pe_record_stub): New function.
23494 (i386_pe_file_end): Emit refptr-stubs.
23496 2013-04-02 Jakub Jelinek <jakub@redhat.com>
23498 PR rtl-optimization/56745
23499 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
23500 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
23503 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
23504 and both of them are MEM_REFs, just compare first argument for
23505 equality and attempt to deal even with differing offsets.
23508 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
23509 of gimple_clobber_p to be MEM_REF.
23510 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
23511 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
23512 after gimplification.
23513 * asan.c (get_mem_ref_of_assignment): Don't instrument
23514 gimple_clobber_p stmts.
23515 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
23516 gimple_clobber_p stmt if they have MEM_REF lhs and
23517 are dead because of another gimple_clobber_p stmt.
23518 * tree-ssa-live.c (clear_unused_block_pointer): Treat
23519 gimple_clobber_p stmts like debug stmts.
23520 (remove_unused_locals): Remove clobbers with MEM_REF lhs
23521 that refer to unused VAR_DECLs or uninitialized values.
23522 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
23523 gimple_clobber_p stmts if they refer to removed parameters.
23524 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
23527 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
23529 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
23530 using SWI48 mode attribute.
23532 2013-04-02 Wei Mi <wmi@google.com>
23534 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
23535 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
23536 *<rotate_insn><mode>3_mask in i386.md.
23538 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
23540 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
23542 2013-04-02 Richard Biener <rguenther@suse.de>
23544 PR tree-optimization/56778
23545 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
23546 Runtime alias tests are not supported for gather loads.
23547 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
23548 stmts referenced from SSA operands before updating SSA form.
23550 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
23551 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23553 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
23554 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
23555 * config/arm/cortex-a53.md: New file.
23556 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
23557 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
23558 * config/arm/arm.c (arm_issue_rate): Likewise.
23559 * config/arm/arm-tune.md: Regenerate
23560 * config/arm/arm-tables.opt: Regenerate.
23561 * config/arm/arm-cores.def: Add cortex-a53.
23563 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
23565 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
23568 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23570 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
23571 scalar load/store operations using B/H registers.
23572 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
23574 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23576 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
23578 * config/aarch64/aarch64.c
23579 (aarch64_simd_scalar_immediate_valid_for_move): New.
23580 * config/aarch64/aarch64-protos.h
23581 (aarch64_simd_scalar_immediate_valid_for_move): New.
23582 * config/aarch64/constraints.md (Dh, Dq): New.
23583 * config/aarch64/iterators.md (hq): New.
23585 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
23587 * reorg.c (get_branch_condition): Deal with conditional returns.
23588 (fill_simple_delay_slots): Remove dead code dealing with jumps.
23590 2013-04-01 Wei Mi <wmi@google.com>
23592 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
23593 Truncate operand 2 using %b asm operand modifier.
23594 (*<shift_insn><mode>3_mask): Ditto.
23595 (*<rotate_insn><mode>3_mask): Ditto.
23597 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
23599 PR middle-end/56798
23600 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
23602 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
23604 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
23606 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
23608 2013-03-30 Lawrence Crowl <crowl@google.com>
23610 * dse.c (clear_alias_sets): Remove never set.
23611 (disqualified_clear_alias_sets): Remove never set.
23612 (clear_alias_mode_pool): Remove never set.
23613 (dse_step0): Remove condition that is never true.
23614 (canon_address): Remove condition that is never true.
23615 (dse_step7): Remove condition that is never true.
23616 (rest_of_handle_dse): Remove condition that is never true.
23617 (rest_of_handle_dse::did_global): Remove never read from above.
23618 (dse_step2_spill): Remove never called from above.
23619 (dse_step5_spill): Remove never called from above.
23621 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
23623 * doc/md.texi (Standard Names) <casesi>: Update documentation for
23624 JUMP_TABLE_DATA changes.
23625 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
23626 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
23627 (Insns) <jump_table_data>: New entry.
23628 * doc/tm.texi: Regenerate.
23630 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
23632 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
23633 for table jump at the end of a basic block using tablejump_p.
23634 * targhooks.c (default_invalid_within_doloop): Likewise.
23635 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
23636 target hook implementation that is identical to the default hook.
23637 (rs6000_invalid_within_doloop): Remove.
23639 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
23640 unused variable from tablejump_p call.
23642 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
23643 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
23644 (INSN_DELETED_P): Likewise.
23645 (emit_jump_table_data): New prototype.
23646 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
23647 after 4th as unused.
23648 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
23649 * sched-vis.c (print_insn): Likewise.
23650 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
23651 insn for compatibility with back ends that use next_active_insn to
23652 identify jump table data.
23653 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
23654 (remove_insn): Likewise.
23655 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
23657 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
23658 (emit_jump_table_data): New function.
23660 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
23661 basic block, a JUMP_TABLE_DATA never is.
23662 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
23663 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
23664 off from code handling real insns.
23665 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
23666 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
23667 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
23668 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
23669 is not a NONDEBUG_INSN_P.
23670 * ira-costs.c (scan_one_insn): Likewise.
23671 * jump.c (mark_all_labels): Likewise.
23672 (mark_jump_label_1): Likewise.
23673 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
23674 * lra.c (get_insn_freq): Expect all insns reaching here to be in
23676 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
23677 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
23678 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
23679 JUMP_TABLE_DATA_P insns.
23680 (calculate_elim_costs_all_insns): Likewise.
23681 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
23682 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
23683 (delete_output_reload): Code style fixups.
23684 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
23685 insn flags on this non-insn.
23686 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
23687 as scheduling barriers, for pre-change compatibility.
23688 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
23689 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
23691 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
23692 redundant JUMP_TABLE_DATA_P test.
23693 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
23694 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
23695 (frv_for_each_packet): Likewise.
23696 * config/i386/i386.c (min_insn_size): Likewise.
23697 (ix86_avoid_jump_mispredicts): Likewise.
23698 * config/m32r/m32r.c (m32r_is_insn): Likewise.
23699 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
23700 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
23701 (mips16_insn_length): Robustify.
23702 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
23703 (mips16_split_long_branches): Likewise.
23704 * config/pa/pa.c (pa_combine_instructions): Likewise.
23705 * config/rs6000/rs6000.c (get_next_active_insn): Treat
23706 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
23707 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
23708 as contributing to pool range lengths.
23709 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
23710 Remove redundant JUMP_TABLE_DATA_P test.
23711 (sh_loop_align): Likewise.
23712 (split_branches): Likewise.
23713 (sh_insn_length_adjustment): Likewise.
23714 * config/spu/spu.c (get_branch_target): Likewise.
23716 2013-03-29 Jan Hubicka <jh@suse.cz>
23718 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
23719 gcov streaming; stream hot bb threshold to ltrans.
23720 * predict.c (get_hot_bb_threshold): Break out from ....
23721 (maybe_hot_count_p): ... here.
23722 (set_hot_bb_threshold): New function.
23723 * lto-section-in.c (lto_section_name): Add profile.
23724 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
23725 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
23726 and data-streamer.h
23727 (histogram_entry): New structure.
23728 (histogram, histogram_pool): New global vars.
23729 (histogram_hash): New structure.
23730 (histogram_hash::hash): New method.
23731 (histogram_hash::equal): Likewise.
23732 (account_time_size): New function.
23733 (cmp_counts): New function.
23734 (dump_histogram): New function.
23735 (ipa_profile_generate_summary): New function.
23736 (ipa_profile_write_summary): New function.
23737 (ipa_profile_read_summary): New function.
23738 (ipa_profile): Decide on threshold.
23739 (pass_ipa_profile): Add ipa_profile_write_summary and
23740 ipa_profile_read_summary.
23741 * Makefile.in (ipa.o): Update dependencies.
23742 * lto-streamer.h (LTO_section_ipa_profile): New section.
23744 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
23746 * tree.h (VAR_P): New.
23748 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
23751 * doc/invoke.texi ([-fwhole-program]): Fix typo.
23753 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
23755 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
23756 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
23757 (control_flow_insn_p): Likewise.
23758 * cfgrtl.c (duplicate_insn_chain): Likewise.
23759 * final.c (get_attr_length_1): Likewise.
23760 (shorten_branches): Likewise.
23761 (final_scan_insn): Likewise.
23762 * function.c (instantiate_virtual_regs): Likewise.
23763 * gcse.c (insert_insn_end_basic_block): Likewise.
23764 * ira-costs.c (scan_one_insn): Likewise.
23765 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
23766 * lra.c (check_rtl): Likewise.
23767 * reload1.c (elimination_costs_in_insn): Likewise.
23768 * reorg.c (follow_jumps): Likewise.
23770 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
23771 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
23772 (thumb_far_jump_used_p): Likewise.
23773 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
23774 (workaround_speculation): Likewise.
23775 (add_sched_insns_for_speculation): Likewise.
23776 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
23777 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
23778 (frv_for_each_packet): Likewise.
23779 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
23780 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
23781 (final_emit_insn_group_barriers): Likewise.
23782 * config/m32r/m32r.c (m32r_is_insn): Likewise.
23783 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
23784 (mips16_insn_length): Likewise.
23785 * config/pa/pa.c (pa_reorg): Likewise.
23786 (pa_combine_instructions): Likewise.
23787 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
23788 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
23789 (sh_reorg): Likewise.
23790 (split_branches): Likewise.
23791 * config/spu/spu.c (get_branch_target): Likewise.
23793 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
23796 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
23798 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
23799 Fix declaration name.
23801 2013-03-28 Lawrence Crowl <crowl@google.com>
23803 * graphds.h (struct graph.indicies): Remove unused.
23804 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
23805 (SCOP_ORIGINAL_PDDRS): Remove unused.
23806 * sese.h (extern insert_loop_close_phis): Removed unused.
23807 (extern insert_guard_phis): Removed unused.
23808 (extern ivtype_map_elt_info): Removed unused.
23809 (new_ivtype_map_elt): Removed unused.
23810 * sese.c (ivtype_map_elt_info): Removed unused.
23812 2013-03-28 Lawrence Crowl <crowl@google.com>
23814 * Makefile.in: Add several missing include dependences.
23816 (test-dump.o): New. This object is not added to any executable,
23817 but is present for ad-hoc testing.
23819 (debug (const bitmap_head_def &)): New.
23820 (debug (const bitmap_head_def *)): New.
23822 (extern debug (const bitmap_head_def &)): New.
23823 (extern debug (const bitmap_head_def *)): New.
23825 (debug (edge_def &)): New.
23826 (debug (edge_def *)): New.
23828 (debug (basic_block_def &)): New.
23829 (debug (basic_block_def *)): New.
23831 (dump_node (const_tree, int, FILE *)): Correct source file.
23833 (debug (die_struct &)): New.
23834 (debug (die_struct *)): New.
23836 (extern debug (die_struct &)): New.
23837 (extern debug (die_struct *)): New.
23838 * gimple-pretty-print.c
23839 (debug (gimple_statement_d &)): New.
23840 (debug (gimple_statement_d *)): New.
23841 * gimple-pretty-print.h
23842 (extern debug (gimple_statement_d &)): New.
23843 (extern debug (gimple_statement_d *)): New.
23845 (debug (ira_allocno_copy &)): New.
23846 (debug (ira_allocno_copy *)): New.
23847 (debug (ira_allocno &)): New.
23848 (debug (ira_allocno *)): New.
23850 (extern debug (ira_allocno_copy &)): New.
23851 (extern debug (ira_allocno_copy *)): New.
23852 (extern debug (ira_allocno &)): New.
23853 (extern debug (ira_allocno *)): New.
23855 (debug (live_range &)): New.
23856 (debug (live_range *)): New.
23858 (debug (lra_live_range &)): New.
23859 (debug (lra_live_range *)): New.
23861 (debug (lra_live_range &)): New.
23862 (debug (lra_live_range *)): New.
23864 (debug (omega_pb_d &)): New.
23865 (debug (omega_pb_d *)): New.
23867 (extern debug (omega_pb_d &)): New.
23868 (extern debug (omega_pb_d *)): New.
23870 (debug (const rtx_def &)): New.
23871 (debug (const rtx_def *)): New.
23873 (debug_tree (tree): Move within file.
23874 (debug_raw (const tree_node &)): New.
23875 (debug_raw (const tree_node *)): New.
23876 (dump_tree_via_hooks (const tree_node *, int)): New.
23877 (debug (const tree_node &)): New.
23878 (debug (const tree_node *)): New.
23879 (debug_verbose (const tree_node &)): New.
23880 (debug_verbose (const tree_node *)): New.
23881 (debug_head (const tree_node &)): New.
23882 (debug_head (const tree_node *)): New.
23883 (debug_body (const tree_node &)): New.
23884 (debug_body (const tree_node *)): New.
23885 (debug_vec_tree (tree): Move and reimplement in terms of dump.
23886 (debug (vec<tree, va_gc> &)): New.
23887 (debug (vec<tree, va_gc> *)): New.
23889 (extern debug (const rtx_def &)): New.
23890 (extern debug (const rtx_def *)): New.
23892 (debug_raw (simple_bitmap_def &)): New.
23893 (debug_raw (simple_bitmap_def *)): New.
23894 (debug (simple_bitmap_def &)): New.
23895 (debug (simple_bitmap_def *)): New.
23897 (extern debug (simple_bitmap_def &)): New.
23898 (extern debug (simple_bitmap_def *)): New.
23899 (extern debug_raw (simple_bitmap_def &)): New.
23900 (extern debug_raw (simple_bitmap_def *)): New.
23902 (debug (vinsn_def &)): New.
23903 (debug (vinsn_def *)): New.
23904 (debug_verbose (vinsn_def &)): New.
23905 (debug_verbose (vinsn_def *)): New.
23906 (debug (expr_def &)): New.
23907 (debug (expr_def *)): New.
23908 (debug_verbose (expr_def &)): New.
23909 (debug_verbose (expr_def *)): New.
23910 (debug (vec<rtx> &)): New.
23911 (debug (vec<rtx> *)): New.
23913 (extern debug (vinsn_def &)): New.
23914 (extern debug (vinsn_def *)): New.
23915 (extern debug_verbose (vinsn_def &)): New.
23916 (extern debug_verbose (vinsn_def *)): New.
23917 (extern debug (expr_def &)): New.
23918 (extern debug (expr_def *)): New.
23919 (extern debug_verbose (expr_def &)): New.
23920 (extern debug_verbose (expr_def *)): New.
23921 (extern debug (vec<rtx> &)): New.
23922 (extern debug (vec<rtx> *)): New.
23924 (_list_iter_cond_expr): Make inline instead of static.
23926 (debug (sreal &)): New.
23927 (debug (sreal *)): New.
23929 (extern debug (sreal &)): New.
23930 (extern debug (sreal *)): New.
23932 (extern debug_raw (const tree_node &)): New.
23933 (extern debug_raw (const tree_node *)): New.
23934 (extern debug (const tree_node &)): New.
23935 (extern debug (const tree_node *)): New.
23936 (extern debug_verbose (const tree_node &)): New.
23937 (extern debug_verbose (const tree_node *)): New.
23938 (extern debug_head (const tree_node &)): New.
23939 (extern debug_head (const tree_node *)): New.
23940 (extern debug_body (const tree_node &)): New.
23941 (extern debug_body (const tree_node *)): New.
23942 (extern debug (vec<tree, va_gc> &)): New.
23943 (extern debug (vec<tree, va_gc> *)): New.
23945 (debug (struct loop &)): New.
23946 (debug (struct loop *)): New.
23947 (debug_verbose (struct loop &)): New.
23948 (debug_verbose (struct loop *)): New.
23949 * tree-dump.c: Add header dependence.
23951 (extern debug (struct loop &)): New.
23952 (extern debug (struct loop *)): New.
23953 (extern debug_verbose (struct loop &)): New.
23954 (extern debug_verbose (struct loop *)): New.
23956 (debug (data_reference &)): New.
23957 (debug (data_reference *)): New.
23958 (debug (vec<data_reference_p> &)): New.
23959 (debug (vec<data_reference_p> *)): New.
23960 (debug (vec<ddr_p> &)): New.
23961 (debug (vec<ddr_p> *)): New.
23963 (extern debug (data_reference &)): New.
23964 (extern debug (data_reference *)): New.
23965 (extern debug (vec<data_reference_p> &)): New.
23966 (extern debug (vec<data_reference_p> *)): New.
23967 (extern debug (vec<ddr_p> &)): New.
23968 (extern debug (vec<ddr_p> *)): New.
23970 (debug (pt_solution &)): New.
23971 (debug (pt_solution *)): New.
23973 (extern debug (pt_solution &)): New.
23974 (extern debug (pt_solution *)): New.
23976 (debug (_var_map &)): New.
23977 (debug (_var_map *)): New.
23978 (debug (tree_live_info_d &)): New.
23979 (debug (tree_live_info_d *)): New.
23981 (extern debug (_var_map &)): New.
23982 (extern debug (_var_map *)): New.
23983 (extern debug (tree_live_info_d &)): New.
23984 (extern debug (tree_live_info_d *)): New.
23986 2013-03-28 Jan Hubicka <jh@suse.cz>
23988 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
23990 2013-03-28 Ian Bolton <ian.bolton@arm.com>
23992 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
23993 record only when desired or required.
23995 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
23997 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
23998 *vec_extractv2di_1_rex64. Use x64 isa attribute.
24000 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24002 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
24003 (*andsi3_compare0_uxtw): New pattern.
24004 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
24005 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
24007 2013-03-28 Jan Hubicka <jh@suse.cz>
24009 * data-streamer-in.c (streamer_read_gcov_count): New function.
24010 * gimple-streamer-out.c: Include value-prof.h.
24011 (output_gimple_stmt): Output histogram.
24012 (output_bb): Use streamer_write_gcov_count.
24013 * value-prof.c: Include data-streamer.h
24014 (dump_histogram_value): Add HIST_TYPE_MAX.
24015 (stream_out_histogram_value): New function.
24016 (stream_in_histogram_value): New function.
24017 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
24018 (stream_out_histogram_value, stream_in_histogram_value): Declare.
24019 * data-streamer-out.c (streamer_write_gcov_count): New function.
24020 (streamer_write_gcov_count_stream): New function.
24021 * lto-cgraph.c (lto_output_edge): Update counter streaming.
24022 (lto_output_node): Likewise.
24023 (input_node, input_edge): Likewise.
24024 * lto-streamer-out.c (output_cfg): Update streaming.
24025 * lto-streamer-in.c (input_cfg): Likewise.
24026 * data-streamer.h (streamer_write_gcov_count,
24027 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
24028 * gimple-streamer-in.c: Include value-prof.h
24029 (input_gimple_stmt): Input histograms.
24030 (input_bb): Update profile streaming.
24032 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24034 * genmodes.c (emit_max_int): New function.
24035 (emit_insn_modes_h): Added call to emit_max_function.
24036 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
24038 * machmode.def: Fixed comment.
24040 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24042 * combine.c (try_combine): Removed useless assert.
24043 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
24045 2013-03-28 Marek Polacek <polacek@redhat.com>
24046 Richard Biener <rguenther@suse.de>
24048 PR tree-optimization/56695
24049 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
24050 build signed result of a vector comparison.
24051 * tree-cfg.c (verify_gimple_comparison): Check that a result
24052 of a vector comparison has signed type.
24054 2013-03-28 Richard Biener <rguenther@suse.de>
24056 PR tree-optimization/37021
24057 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
24058 do not restrict gaps between groups.
24059 * tree-vect-stmts.c (vectorizable_load): Properly account for
24060 a gap between groups.
24062 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
24064 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
24065 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
24068 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
24070 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
24071 * doc/extend.texi (Named Address Spaces): Ditto.
24072 (Variable Attributes): Ditto.
24074 2013-03-27 Kai Tietz <ktietz@redhat.com>
24076 * config.build: Add support for cygwin x64 target.
24077 * config.gcc: Likewise.
24078 * config.host: Likewise.
24079 * configure.ac: Likewise
24080 * configure: Regenerated.
24082 2013-03-27 Kai Tietz <ktietz@redhat.com>
24084 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
24085 * config/i386/t-cygwin-w64: New file.
24086 * config/i386/cygwin-w64.h: New file.
24087 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
24088 and add support for x64-cygwin target.
24089 (CPP_SPEC): Likewise.
24090 (CXX_WRAP_SPEC_LIST): Undefine before define.
24091 (LIBGCJ_SONAME): Use 15 as version.
24093 2013-03-27 Richard Biener <rguenther@suse.de>
24095 PR tree-optimization/56716
24096 * tree-ssa-structalias.c (perform_var_substitution): Adjust
24097 dumping for ref nodes.
24099 2013-03-27 Martin Jambor <mjambor@suse.cz>
24101 PR tree-optimization/55334
24102 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
24103 restricted pointers to arrays.
24105 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
24107 * Makefile.in (.SUFFIXES): Add .cc.
24108 (.c.o): Apply same recipe for implicit rule .cc.o.
24110 2013-03-27 Richard Biener <rguenther@suse.de>
24112 PR tree-optimization/37021
24113 * tree-vect-data-refs.c (vect_check_strided_load): Allow
24114 REALPART/IMAGPART_EXPRs around the supported refs.
24115 * tree-ssa-structalias.c (find_func_aliases): Assume that
24116 floating-point values are not used to transfer pointers.
24118 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
24120 * target.def (TARGET_HAS_IFUNC_P): New target hook.
24121 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
24122 * doc/tm.texi: Regenerate.
24123 * targhooks.h (default_has_ifunc_p): New.
24124 * targhooks.c (default_has_ifunc_p): Ditto.
24125 * config/linux-protos.h: New file.
24126 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
24127 hook for linux which disables support of indirect functions in android.
24128 * config/linux-android.c: New file.
24129 * config/t-linux-android.c: Ditto.
24130 * config.gcc: Added new object file linux-android.o.
24131 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
24132 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
24133 * varasm.c (do_assemble_alias): Likewise.
24134 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
24135 doesn't support indirect functions.
24136 * configure: Regenerate.
24138 2013-03-27 Bin Cheng <bin.cheng@arm.com>
24141 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
24142 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
24145 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24147 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
24149 2013-03-27 Terry Guo <terry.guo@arm.com>
24151 * config/arm/arm-cores.def: Added core cortex-r7.
24152 * config/arm/arm-tune.md: Regenerated.
24153 * config/arm/arm-tables.opt: Regenerated.
24154 * doc/invoke.texi: Added entry for core cortex-r7.
24156 2013-03-27 Walter Lee <walt@tilera.com>
24158 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
24159 double-decrement of next_scratch_regno.
24161 2013-03-27 Walter Lee <walt@tilera.com>
24163 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
24165 (insn_v1mulus): Ditto.
24166 (insn_v2muls): Ditto.
24168 2013-03-27 Walter Lee <walt@tilera.com>
24170 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
24171 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
24173 2013-03-27 Walter Lee <walt@tilera.com>
24175 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
24176 (*sibcall_value): Ditto.
24178 2013-03-27 Walter Lee <walt@tilera.com>
24180 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
24181 (insn_mnz_v8qi): ... this ...
24182 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
24184 (insn_v<n>mnz): Replaced by ...
24185 (insn_v1mnz): ... this ...
24186 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
24188 (insn_mz_<mode>): Replaced by ...
24189 (insn_mz_v8qi): ... this ...
24190 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
24192 (insn_v<n>mz): Replaced by ...
24193 (insn_v1mz): ... this ...
24194 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
24197 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
24199 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
24201 2013-03-26 Roland McGrath <mcgrathr@google.com>
24203 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
24204 than fprintf with a non-constant, non-format string.
24206 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24208 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
24209 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
24210 operand 0 predicate.
24211 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
24212 attribute. Use general_x64nomem_operand as operand 1 predicate.
24213 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
24214 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24215 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
24216 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24217 (mov<mode>_insv_1): Remove expander. Merge insn with
24218 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
24219 Use general_x64nomem_operand as operand 1 predicate.
24220 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
24221 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
24222 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24223 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
24224 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
24225 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
24226 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24227 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
24228 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
24229 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
24230 (general_x64nomem_operand): Ditto.
24232 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
24234 * config/rtems.opt: Add -pthread option.
24236 2013-03-26 Richard Biener <rguenther@suse.de>
24238 * alias.c (find_base_term): Avoid redundant and not used recursion.
24239 (base_alias_check): Get the initial base term from the caller.
24240 (true_dependence_1): Compute and pass base terms to base_alias_check.
24241 (write_dependence_p): Likewise.
24242 (may_alias_p): Likewise.
24244 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
24246 * config/aarch64/aarch64.c (aarch64_classify_address): Support
24247 PC-relative load in SI modes and above only.
24249 2013-03-26 Xinyu Qi <xyqi@marvell.com>
24251 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
24252 * config/arm/iwmmxt.md (WCGR0): Update.
24253 (WCGR1, WCGR2, WCGR3): Likewise.
24255 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24257 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
24258 Use x64 and nox64 isa attributes.
24260 2013-03-26 Richard Biener <rguenther@suse.de>
24262 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
24263 alignment computations and rely on get_object_alignment_1
24264 for the !TYPE_P case.
24265 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
24267 2013-03-26 Walter Lee <walt@tilera.com>
24269 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
24270 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
24272 2013-03-25 Jeff Law <law@redhat.com>
24274 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
24275 check for INTEGRAL_TYPE_P that was missing due to checking in
24276 wrong version of prior patch.
24278 2013-03-25 Walter Lee <walt@tilera.com>
24280 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
24281 TILEGX_INSN_SHUFFLEBYTES1.
24282 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
24284 (tilegx_builtins): Ditto.
24285 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
24287 2013-03-25 Walter Lee <walt@tilera.com>
24289 * config/tilegx/tilegx.md (floatsisf2): New pattern.
24290 (floatunssisf2): New pattern.
24291 (floatsidf2): New pattern.
24292 (floatunssidf2): New pattern.
24294 2013-03-25 Walter Lee <walt@tilera.com>
24296 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
24297 tests for constraint J, K, N, P.
24299 2013-03-25 Walter Lee <walt@tilera.com>
24301 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
24302 Use indirect/pcrel encoding.
24303 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
24306 2013-03-25 Steve Ellcey <sellcey@mips.com>
24308 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
24309 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
24310 * config/mips/mips.c (mips_option_override): Set IMADD default.
24311 * config/mips/mips.h (PTF_AVOID_IMADD): New.
24312 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
24313 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
24314 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
24315 * doc/invoke.texi (-mimadd/-mno-imadd): New.
24317 2013-03-25 Jeff Law <law@redhat.com>
24319 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
24320 slightly to avoid creating and folding useless trees. Simplify
24321 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
24323 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
24325 * config/i386/i386.md (*zero_extendsidi2): Merge with
24326 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
24327 * config/i386/predicates.md (x86_64_zext_operand): Rename from
24328 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
24329 targets. Clarify comment.
24331 2013-03-25 Martin Jambor <mjambor@suse.cz>
24333 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
24334 pass-through jump functions differently.
24335 (ipa_read_jump_function): Likewise. Also use setter functions to set
24338 2013-03-25 Martin Jambor <mjambor@suse.cz>
24340 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
24341 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
24343 (ipa_get_indirect_edge_target): New function.
24344 (devirtualization_time_bonus): New parameter known_aggs, pass it to
24345 ipa_get_indirect_edge_target. Update all callers.
24346 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
24347 ipa_get_indirect_edge_target_1 instead of calling
24348 ipa_get_indirect_edge_target.
24349 (create_specialized_node): Pass aggvlas to
24350 ipcp_discover_new_direct_edges.
24352 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24354 * config/arm/arm.md (f_sels, f_seld): New types.
24355 (*cmov<mode>): New pattern.
24356 * config/arm/predicates.md (arm_vsel_comparison_operator): New
24359 2013-03-25 Kai Tietz <ktietz@redhat.com>
24361 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
24362 POSIX-printf for mingw-hosted builds.
24364 2013-03-25 Richard Biener <rguenther@suse.de>
24366 PR middle-end/56694
24367 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
24368 must-not-throw stmt location.
24370 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24372 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
24373 Emit load-acquire versions when acq is true.
24374 (arm_emit_store_exclusive): Add rel parameter.
24375 Emit store-release versions when rel is true.
24376 (arm_split_compare_and_swap): Use acquire-release instructions
24378 of barriers when appropriate.
24379 (arm_split_atomic_op): Likewise.
24380 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
24381 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
24382 (VUNSPEC_SLX): Likewise.
24383 (VUNSPEC_LDA): Likewise.
24384 (VUNSPEC_STL): Likewise.
24385 * config/arm/sync.md (atomic_load<mode>): New pattern.
24386 (atomic_store<mode>): Likewise.
24387 (arm_load_acquire_exclusive<mode>): Likewise.
24388 (arm_load_acquire_exclusivesi): Likewise.
24389 (arm_load_acquire_exclusivedi): Likewise.
24390 (arm_store_release_exclusive<mode>): Likewise.
24392 2013-03-25 Catherine Moore <clm@codesourcery.com>
24394 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
24395 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
24396 * config/mip/predicates.md (lwsp_swsp_operand,
24397 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
24398 sb16_operand, db4_operand, db7_operand, ib3_operand,
24399 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
24400 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
24401 andi16_operand): New predicates.
24402 * config/mips/mips.md (compression): New attribute.
24403 (enabled): New attribute.
24404 (length): Consider compression in computing length.
24405 (shift_compression): New code attribute.
24406 (*add<mode>3): New operands. Record compression.
24407 (sub<mode>3): Likewise.
24408 (one_cmpl<mode>2): Likewise.
24409 (*and<mode>3): Likewise.
24410 (*ior<mode>3): Likewise.
24411 (unnamed pattern for xor): Likewise.
24412 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
24413 (*<optab><mode>3): Likewise.
24414 (*mov<mode>_internal: Likewise.
24415 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
24416 (mips_unsigned_immediate_p): New.
24417 (umips_lwsp_swsp_address_p): New.
24418 (m16_based_address_p): New.
24419 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
24420 (mips_unsigned_immediate_p): New prototype.
24421 (lwsp_swsp_address_p): New prototype.
24422 (m16_based_address_p): New prototype.
24423 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
24424 (mips_signed_immediate_p): New function.
24425 (m16_based_address_p): New function.
24426 (lwsp_swsp_address_p): New function.
24427 (mips_print_operand_punctuation): Recognize short delay slot insns
24428 for microMIPS.add<mode>3"
24430 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24433 * config/arm/iterators.md (v_cmp_result): New mode attribute.
24434 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
24436 2013-03-25 Richard Biener <rguenther@suse.de>
24438 PR tree-optimization/56689
24439 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
24442 2013-03-25 Richard Biener <rguenther@suse.de>
24444 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
24446 (memory_references): Likewise.
24447 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
24448 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
24449 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
24450 (gather_mem_refs_in_loops): Fold into ...
24451 (analyze_memory_references): ... this. Move initialization
24452 to tree_ssa_lim_initialize.
24453 (fill_always_executed_in): Rename to ...
24454 (fill_always_executed_in_1): ... this.
24455 (fill_always_executed_in): Move contains_call computation to
24456 this new function from ...
24457 (tree_ssa_lim_initialize): ... here.
24458 (tree_ssa_lim): Call fill_always_executed_in.
24460 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24462 * postreload.c (reload_combine): Fix code detecting returns.
24464 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24466 * function.c (emit_use_return_register_into_block): On cc0 targets,
24467 do not emit the sequence between cc0 setter and user.
24469 2013-03-25 Kai Tietz <ktietz@redhat.com>
24471 * config/i386/predicates.md (local_symbolic_operand): Interpret
24472 dll-imported symbols as none-local.
24474 2013-03-25 Richard Biener <rguenther@suse.de>
24476 * tree-ssa-loop-im.c (struct depend): Remove.
24477 (struct lim_aux_data): Make depends a vec of gimples.
24478 (free_lim_aux_data): Adjust.
24479 (add_dependency): Likewise.
24480 (set_level): Likewise.
24482 2013-03-25 Richard Biener <rguenther@suse.de>
24484 PR middle-end/56434
24485 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
24486 the pointer returned by calls with ECF_MALLOC set.
24488 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24490 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
24492 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24494 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
24495 using MMXMODE mode iterator.
24496 (*move<mode>_internal): Merge with *movv2sf_internal and
24497 *movv2sf_internal_rex64 using MMXMODE mode iterator.
24499 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24501 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
24502 (record_last_mem_set_info): Likewise.
24504 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
24505 of XNEWVEC followed by memset.
24506 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
24508 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24510 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
24511 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
24512 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
24513 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
24514 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
24515 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
24516 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
24517 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
24518 BARRIER_P instead of GET_CODE.
24520 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24522 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
24523 inaccuracy in the probing code.
24525 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
24526 (ctrapdi4): Likewise.
24528 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24530 * calls.c (expand_call): Add missing guard to code handling return
24531 of non-BLKmode structures in MSB.
24532 * function.c (expand_function_end): Likewise.
24534 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24536 * combine.c (try_combine): Adjust comment. Do not add the set of
24537 insn #0 if the destination indirectly is set or dies in insn #2.
24538 Tidy up code to distribute a new note.
24540 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24542 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
24543 also for alternatives 16 and 17.
24545 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24547 * config/i386/sse.md (*mov<mode>_internal): Merge with
24548 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
24549 Emit insn template depending on type attribute. Use
24550 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24551 movd instead of movq mnemonic for interunit moves. Rewrite mode
24552 attribute calculation. Remove unit attribute calculation.
24553 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
24554 Set prefix_data16 attribute for DImode ssemov types.
24555 Use Ym instead of y for SSE-MMX conversion alternatives.
24556 Reorder operand constraints.
24558 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
24560 * df.h (df_insn_delete): Adjust prototype.
24561 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
24562 and let it decide whether mark the basic block dirty.
24563 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
24564 * df-scan.c (df_insn_info_delete): New helper function, split
24565 off from df_insn_delete.
24566 (df_scan_free_bb_info): Use it.
24567 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
24569 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
24570 that the insn is actually an insn and it has a non-NULL basic block.
24571 Do not mark basic block dirty if only deleting a DEBUG_INSN.
24573 2013-03-22 Richard Biener <rguenther@suse.de>
24575 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
24577 (mem_ref_alloc): Do not allocate them.
24578 (refs_independent_p): Do not query or maintain a cache.
24580 2013-03-22 Richard Biener <rguenther@suse.de>
24582 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
24583 (gather_mem_refs_in_loops): Do not compute it.
24584 (analyze_memory_references): Do not allocate it.
24585 (tree_ssa_lim_finalize): Do not free it.
24586 (for_all_locs_in_loop): Do not query all_refs_in_loop.
24588 2013-03-22 Richard Biener <rguenther@suse.de>
24590 * is-a.h (as_a): Use gcc_checking_assert.
24592 2013-03-22 Ian Bolton <ian.bolton@arm.com>
24594 * config/aarch64/aarch64.c (aarch64_print_operand): New
24595 format specifier for printing a constant in hex.
24596 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
24597 format specifier for printing second operand.
24599 2013-03-22 Richard Biener <rguenther@suse.de>
24601 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
24603 (gather_mem_refs_in_loops): Perform store accumulation here.
24604 (create_vop_ref_mapping_loop): Remove.
24605 (create_vop_ref_mapping): Likewise.
24606 (analyze_memory_references): Initialize refs_stored_in_loop.
24607 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
24608 (record_indep_loop): Remove.
24609 (record_dep_loop): New function.
24610 (ref_indep_loop_p_1): Adjust to only walk over references
24611 in the loop, not its subloops.
24612 (ref_indep_loop_p): Rename to ...
24613 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
24614 maintaining a more fine-grained cache.
24615 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
24616 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
24618 2013-03-22 Richard Biener <rguenther@suse.de>
24620 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
24621 (struct mem_ref): Make accesses_in_loop a vec of a vec of
24622 aggregate mem_ref_loc.
24623 (free_mem_ref_locs): Inline into ...
24624 (memref_free): ... this and adjust.
24625 (mem_ref_alloc): Adjust.
24626 (mem_ref_locs_alloc): Remove.
24627 (record_mem_ref_loc): Adjust.
24628 (get_all_locs_in_loop): Rewrite into ...
24629 (for_all_locs_in_loop): ... this iterator.
24630 (rewrite_mem_ref_loc): New functor.
24631 (rewrite_mem_refs): Use for_all_locs_in_loop.
24632 (sm_set_flag_if_changed): New functor.
24633 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
24634 (ref_always_accessed): New functor.
24635 (ref_always_accessed_p): Use for_all_locs_in_loop.
24637 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24639 * tree-pass.h (PROP_gimple_lvec): New.
24640 * passes.c (dump_properties): Handle PROP_gimple_lvec.
24641 (init_optimization_passes): Move pass_lower_vector.
24642 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
24644 (pass_lower_vector): Provide PROP_gimple_lvec.
24645 (pass_lower_vector_ssa): Likewise.
24646 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
24648 2013-03-21 Mark Wielaard <mjw@redhat.com>
24650 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
24652 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
24654 * config/i386/i386.md (*movdi_internal): Disparage slightly
24655 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
24656 conversion alternatives.
24658 2013-03-21 Jakub Jelinek <jakub@redhat.com>
24660 PR middle-end/48087
24661 * diagnostic.def (DK_WERROR): New kind.
24662 * diagnostic.h (werrorcount): Define.
24663 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
24664 promoted to DK_ERROR, increment DK_WERROR counter instead of
24666 * toplev.c (toplev_main): Call print_ignored_options even if
24667 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
24668 even if just werrorcount is non-zero.
24671 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
24673 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
24674 (string_cst_pool_decl): New function.
24675 (optimize_one_addr_into_implicit_ptr): New function.
24676 (resolve_addr_in_expr): Optimize DWARF location expression
24677 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
24678 which doesn't live in memory, but has DW_AT_location or
24679 DW_AT_const_value, or refers to a string literal, into
24680 DW_OP_GNU_implicit_pointer.
24681 (optimize_location_into_implicit_ptr): New function.
24682 (resolve_addr): If removing DW_AT_location of a variable because
24683 it was DW_OP_addr of address of the variable, but the variable doesn't
24684 live in memory, try to emit const value attribute for the initializer.
24686 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24688 * tree.h (VECTOR_TYPE_P): New macro.
24689 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
24690 TYPE_MODE): Use it.
24691 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
24692 VEC_COND_EXPR cannot be lvalues.
24693 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
24695 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24697 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
24698 Restrict the transformation to equal modes.
24700 2013-03-21 Richard Biener <rguenther@suse.de>
24702 PR tree-optimization/39326
24703 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
24704 (MEM_ANALYZABLE): Adjust.
24705 (record_mem_ref_loc): Move bitmap ops ...
24706 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
24707 unanalyzable refs, do not record locations for it.
24708 (analyze_memory_references): Allocate ref zero as shared
24710 (refs_independent_p): Do not test for unanalyzed mems here.
24711 (ref_indep_loop_p_1): Special-case disambiguation against
24712 the unanalyzed ref.
24713 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
24715 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
24717 * config/arm/arm-protos.h (tune_params): Add
24718 prefer_neon_for_64bits field.
24719 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
24720 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
24721 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
24722 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
24723 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
24724 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
24725 (arm_option_override): Handle -mneon-for-64bits new option.
24726 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
24727 (prefer_neon_for_64bits): Declare new variable.
24728 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
24729 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
24730 (arch_enabled): Handle new arch types. Remove support for onlya8
24732 (one_cmpldi2): Use new arch names.
24733 (zero_extend<mode>di2, extend<mode>di2): Ditto.
24734 * config/arm/arm.opt (mneon-for-64bits): Add option.
24735 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
24736 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
24737 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
24739 * doc/invoke.texi (-mneon-for-64bits): Document.
24741 2013-03-21 Richard Biener <rguenther@suse.de>
24743 PR tree-optimization/39326
24744 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
24745 (sort_bbs_in_loop_postorder_cmp): New function.
24746 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
24748 2013-03-21 Richard Biener <rguenther@suse.de>
24750 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
24751 (vect_insert_into_interleaving_chain): Likewise.
24752 (vect_drs_dependent_in_basic_block): Inline ...
24753 (vect_slp_analyze_data_ref_dependence): ... here. New function,
24755 (vect_analyze_data_ref_dependence): ... here. Simplify.
24756 (vect_check_interleaving): Simplify.
24757 (vect_analyze_data_ref_dependences): Likewise. Split out ...
24758 (vect_slp_analyze_data_ref_dependences): ... this new function.
24759 (dr_group_sort_cmp): New function.
24760 (vect_analyze_data_ref_accesses): Compute data-reference groups
24761 here instead of in vect_analyze_data_ref_dependence. Use
24762 a more efficient algorithm.
24763 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
24764 vect_slp_analyze_data_ref_dependences. Call
24765 vect_analyze_data_ref_accesses earlier.
24766 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
24767 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
24768 (vect_slp_analyze_data_ref_dependences): New prototype.
24770 2013-03-21 Richard Biener <rguenther@suse.de>
24772 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
24773 ref is stored in the loop.
24774 (find_refs_for_sm): Walk only over all stores.
24775 (store_motion_loop): Allocate from lim_bitmap_obstack.
24776 (store_motion): Likewise.
24778 2013-03-21 Richard Biener <rguenther@suse.de>
24780 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
24781 Update virtual SSA form.
24783 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
24785 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
24786 * configure: Regenerate.
24787 * config.in: Regenerate.
24788 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
24789 if !HAVE_LD_EH_FRAME_CIEV3.
24791 2013-03-21 Richard Biener <rguenther@suse.de>
24793 * tree-cfg.c (verify_expr_no_block): New function.
24794 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
24795 nor DECL_VALUE_EXPR have locations with associated blocks.
24796 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
24797 (clear_unused_block_pointer): Remove code dealing with
24798 blocks in DECL_DEBUG_EXPR locations.
24800 2013-03-21 Richard Biener <rguenther@suse.de>
24802 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
24803 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
24804 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
24805 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
24806 instead of DECL_DEBUG_EXPR_IS_FROM.
24807 * gimplify.c (gimplify_modify_expr): Likewise.
24808 * tree-cfg.c (verify_expr_location_1): Likewise.
24809 * tree-complex.c (create_one_component_var): Likewise.
24810 * tree-sra.c (create_access_replacement): Likewise.
24811 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
24812 (clear_unused_block_pointer): Likewise.
24813 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
24814 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
24815 * var-tracking.c (var_debug_decl): Likewise.
24816 (track_expr_p): Likewise.
24817 * tree-inline.c (add_local_variables): Likewise. Set
24818 DECL_HAS_DEBUG_EXPR_P after copying it.
24819 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
24820 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
24822 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
24825 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
24826 * configure: Regenerate.
24827 * config.in: Regenerate.
24828 * config/i386/i386.md (*movdf_internal): Use
24829 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24830 movd instead of movq mnemonic for interunit moves.
24831 (*movdi_internal): Ditto.
24833 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24835 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
24836 (abd<mode>_3): New pattern.
24837 (aba<mode>_3): New pattern.
24838 (fabd<mode>_3): New pattern.
24840 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24842 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
24843 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
24844 occurrence of REGISTER_PREFIX as its empty string.
24846 2013-03-20 Jeff Law <law@redhat.com>
24848 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
24849 addititional equivalences for equality comparisons between an SSA_NAME
24850 and a constant where the SSA_NAME was set from a widening conversion.
24852 2013-03-20 Walter Lee <walt@tilera.com>
24854 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
24856 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
24858 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
24859 depending on type attribute.
24860 (*movti_internal): Ditto.
24861 (*movtf_internal): Ditto.
24862 (*movxf_internal): Ditto.
24863 (*movdf_internal): Ditto.
24864 (*movsf_internal): Ditto.
24866 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
24868 * config/i386/i386.md (*movti_internal): Set prefix attribute to
24869 maybe_vex for sselog1 and ssemov types.
24870 (*movdi_internal): Reorder operand constraints.
24871 (*movsi_internal): Ditto. Set prefix attribute to
24872 maybe_vex for sselog1 and ssemov types.
24873 (*movtf_internal): Set prefix attribute to maybe_vex
24874 for sselog1 and ssemov types.
24875 (*movdf_internal): Ditto. Set prefix_data16 attribute for
24876 DImode ssemov types. Reorder operand constraints.
24877 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
24878 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
24879 attribute for SImode ssemov types. Reorder operand constraints.
24881 2013-03-20 Martin Jambor <mjambor@suse.cz>
24883 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
24884 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
24886 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
24888 * config/rs6000/predicates.md (indexed_address, update_address_mem
24889 update_indexed_address_mem): New predicates.
24890 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
24891 attribute for load/store instructions.
24892 * config/rs6000/dfp.md (movsd_store): Likewise.
24893 (movsd_load): Likewise.
24894 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
24895 (unnamed HI->DI extend define_insn): Likewise.
24896 (unnamed SI->DI extend define_insn): Likewise.
24897 (unnamed QI->SI extend define_insn): Likewise.
24898 (unnamed QI->HI extend define_insn): Likewise.
24899 (unnamed HI->SI extend define_insn): Likewise.
24900 (unnamed HI->SI extend define_insn): Likewise.
24901 (extendsfdf2_fpr): Likewise.
24902 (movsi_internal1): Likewise.
24903 (movsi_internal1_single): Likewise.
24904 (movhi_internal): Likewise.
24905 (movqi_internal): Likewise.
24906 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
24907 attribute for load/store instructions.
24908 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
24910 (mov<mode>_softfloat): Likewise.
24911 (mov<mode>_hardfloat32): Likewise.
24912 (mov<mode>_hardfloat64): Likewise.
24913 (mov<mode>_softfloat64): Likewise.
24914 (movdi_internal32): Likewise.
24915 (movdi_internal64): Likewise.
24916 (probe_stack_<mode>): Likewise.
24918 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
24920 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
24921 floating point, and decimal floating point to reload iterator.
24923 * config/rs6000/constraints.md (wl constraint): New constraints to
24924 return FLOAT_REGS if certain options are used to reduce the number
24925 of separate patterns that exist in the file.
24926 (wx constraint): Likewise.
24927 (wz constraint): Likewise.
24929 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
24930 -mdebug=reg, print wg, wl, wx, and wz constraints.
24931 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
24932 Initialize the reload functions for 64-bit binary/decimal floating
24934 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
24935 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
24936 create the buffer on the stack to overcome not having a 32-bit
24938 (rs6000_emit_move): Likewise.
24939 (rs6000_secondary_memory_needed_rtx): Likewise.
24940 (rs6000_alloc_sdmode_stack_slot): Likewise.
24941 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
24942 via xxlxor, just like DFmode 0.0.
24944 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
24945 define as 1 if we are running on a power7 or newer.
24946 (enum r6000_reg_class_enum): Add new constraints.
24948 * config/rs6000/dfp.md (movsd): Delete, combine with binary
24949 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
24950 with other moves by using conditional constraits (wg). Use LFIWZX
24951 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
24952 (movsd splitter): Likewise.
24953 (movsd_hardfloat): Likewise.
24954 (movsd_softfloat): Likewise.
24956 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
24957 binary and decimal floating point moves.
24958 (fmove_ok): New attributes to combine binary and decimal floating
24959 point moves, and to combine power6x (mfpgpr) moves along normal
24961 (real_value_to_target): Likewise.
24962 (f32_lr): Likewise.
24963 (f32_lm): Likewise.
24964 (f32_li): Likewise.
24965 (f32_sr): Likewise.
24966 (f32_sm): Likewise.
24967 (f32_si): Likewise.
24968 (movsf): Combine binary and decimal floating point moves. Combine
24969 power6x (mfpgpr) moves with other moves by using conditional
24970 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
24971 (mov<mode> for SFmode/SDmode); Likewise.
24972 (SFmode/SDmode splitters): Likewise.
24973 (movsf_hardfloat): Likewise.
24974 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
24975 (movsf_softfloat): Likewise.
24976 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
24978 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
24979 wx and wz constraints.
24981 * config/rs6000/constraints.md (wg constraint): New constraint to
24982 return FLOAT_REGS if -mmfpgpr (power6x) was used.
24984 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
24987 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
24988 -mdebug=reg, print wg, wl, wx, and wz constraints.
24989 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
24990 Initialize the reload functions for 64-bit binary/decimal floating
24992 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
24993 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
24994 create the buffer on the stack to overcome not having a 32-bit
24996 (rs6000_emit_move): Likewise.
24997 (rs6000_secondary_memory_needed_rtx): Likewise.
24998 (rs6000_alloc_sdmode_stack_slot): Likewise.
24999 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25000 via xxlxor, just like DFmode 0.0.
25002 * config/rs6000/dfp.md (movdd): Delete, combine with binary
25003 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25004 with other moves by using conditional constraits (wg). Use LFIWZX
25005 and STFIWX for loading SDmode on power7.
25006 (movdd splitters): Likewise.
25007 (movdd_hardfloat32): Likewise.
25008 (movdd_softfloat32): Likewise.
25009 (movdd_hardfloat64_mfpgpr): Likewise.
25010 (movdd_hardfloat64): Likewise.
25011 (movdd_softfloat64): Likewise.
25013 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
25014 64-bit binary and decimal floating point moves.
25015 (FMOVE64X): Likewise.
25016 (movdf): Combine 64-bit binary and decimal floating point moves.
25017 Combine power6x (mfpgpr) moves with other moves by using
25018 conditional constraits (wg).
25019 (mov<mode> for DFmode/DDmode): Likewise.
25020 (DFmode/DDmode splitters): Likewise.
25021 (movdf_hardfloat32): Likewise.
25022 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
25023 (movdf_softfloat32): Likewise.
25024 (movdf_hardfloat64_mfpgpr): Likewise.
25025 (movdf_hardfloat64): Likewise.
25026 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
25027 (movdf_softfloat64): Likewise.
25028 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
25029 (reload_<mode>_load): Move to later in the file so they aren't in
25030 the middle of the floating point move insns.
25031 (reload_<mode>_store): Likewise.
25033 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
25036 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
25037 constraint if -mdebug=reg.
25038 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
25039 Enable using dd reload support if needed.
25041 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
25042 binary and decimal floating point moves in rs6000.md.
25043 (movtd_internal): Likewise.
25045 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
25046 decimal floating point moves.
25048 (movtf_internal): Likewise.
25049 (mov<mode>_internal, TDmode/TFmode): Likewise.
25050 (movtf_softfloat): Likewise.
25051 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
25053 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
25054 movdi_internal64, using wg constraint for move direct operations.
25055 (movdi_internal64): Likewise.
25057 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
25058 MODES_TIEABLE_P for selected modes. Print the numerical value of
25059 the various virtual registers. Use GPR/FPR first/last values,
25060 instead of hard coding the register numbers. Print which modes
25061 have reload functions registered.
25062 (rs6000_option_override_internal): If -mdebug=reg, trace the options
25063 settings before/after setting cpu, target and subtarget settings.
25064 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
25065 and for secondary reload failures in rs6000_secondary_reload_inner.
25066 (rs6000_secondary_reload_fail): Likewise.
25067 (rs6000_secondary_reload_inner): Likewise.
25069 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
25070 macros for first/last GPR and FPR registers.
25071 (LAST_GPR_REGNO): Likewise.
25072 (FIRST_FPR_REGNO): Likewise.
25073 (LAST_FPR_REGNO): Likewise.
25075 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
25076 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
25077 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
25078 (vcond<mode><mode>): Likewise.
25079 (vcondu<mode><mode>): Likewise.
25080 (vector_gtu<mode>): Likewise.
25081 (vector_gte<mode>): Likewise.
25082 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
25083 to prevent the compiler from converting DImode operations to TImode.
25084 (ior<mode>3): Likewise.
25085 (and<mode>3): Likewise.
25086 (one_cmpl<mode>2): Likewise.
25087 (nor<mode>3): Likewise.
25088 (andc<mode>3): Likewise.
25090 * config/rs6000/constraints.md (wt constraint): New constraint
25091 that returns VSX_REGS if TImode is allowed in VSX registers.
25093 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
25094 constant under VSX.
25096 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
25097 similar to TImode, but it is restricted to being in the GPRs.
25099 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
25100 TImode to occupy a single VSX register.
25102 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
25103 -mvsx-timode for power7/power8.
25104 (power7 cpu): Likewise.
25105 (power8 cpu): Likewise.
25107 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
25108 sure that TFmode/TDmode take up two registers if they are ever
25109 allowed in the upper VSX registers.
25110 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
25112 (rs6000_init_hard_regno_mode_ok): Likewise.
25113 (rs6000_debug_reg_global): Add debugging for PTImode and wt
25114 constraint. Print if LRA is turned on.
25115 (rs6000_option_override_internal): Give an error if -mvsx-timode
25116 and VSX is not enabled.
25117 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
25118 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
25119 to reg+offset addressing. Use PTImode when checking offset
25120 addresses for validity.
25121 (reg_offset_addressing_ok_p): Likewise.
25122 (rs6000_legitimate_offset_address_p): Likewise.
25123 (rs6000_legitimize_address): Likewise.
25124 (rs6000_legitimize_reload_address): Likewise.
25125 (rs6000_legitimate_address_p): Likewise.
25126 (rs6000_eliminate_indexed_memrefs): Likewise.
25127 (rs6000_emit_move): Likewise.
25128 (rs6000_secondary_reload): Likewise.
25129 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
25130 reloads to fpr registers to continue to use reg+offset addressing,
25131 but 64-bit reloads to altivec registers need reg+reg addressing.
25132 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
25133 it. Treat LO_SUM like a PLUS operation.
25134 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
25135 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
25136 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
25137 registers to share a register with a smaller sized type, since VSX
25138 puts scalars in the upper 64-bits.
25139 (print_operand): Add support for PTImode.
25140 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
25141 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
25142 registers, but don't have arithmetic support.
25143 (rs6000_memory_move_cost): Add test for VSX.
25144 (rs6000_opt_masks): Add -mvsx-timode.
25146 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
25149 (VSr): Use wt constraint for TImode.
25150 (VSv): Drop TImode support.
25151 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
25152 (vsx_movti_64bit): Likewise.
25153 (vsx_movti_32bit): Likewise.
25154 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
25155 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
25156 one '?' on the appropriate output constraint. Do not allow TImode
25157 logical operations on 32-bit systems.
25158 (vsx_ior<mode>3): Likewise.
25159 (vsx_xor<mode>3): Likewise.
25160 (vsx_one_cmpl<mode>2): Likewise.
25161 (vsx_nor<mode>3): Likewise.
25162 (vsx_andc<mode>3): Likewise.
25163 (vsx_concat_<mode>): Likewise.
25164 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
25166 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
25167 OPTION_MASK_VSX_TIMODE.
25168 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
25169 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
25171 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
25172 (TI2 iterator): New iterator for TImode, PTImode.
25173 (wd mode attribute): Add values for vector types.
25174 (movti_string): Replace TI move operations with operations for TImode
25175 and PTImode. Add support for TImode being allowed in VSX registers.
25176 (mov<mode>_string, TImode/PTImode): Likewise.
25177 (movti_ppc64): Likewise.
25178 (mov<mode>_ppc64, TImode/PTImode): Likewise.
25179 (TI mode splitters): Likewise.
25181 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
25184 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
25186 PR tree-optimization/56355
25187 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
25188 Also handle integers with undefined overflow.
25190 2013-03-20 Catherine Moore <clm@codesourcery.com>
25191 Maciej W. Rozycki <macro@codesourcery.com>
25192 Tom de Vries <tom@codesourcery.com>
25193 Nathan Sidwell <nathan@codesourcery.com>
25194 Iain Sandoe <iain@codesourcery.com>
25195 Nathan Froyd <froydnj@codesourcery.com>
25196 Chao-ying Fu <fu@mips.com>
25198 * doc/extend.texi: (micromips, nomicromips, nocompression):
25199 Document new function attributes.
25200 * doc/invoke.texi (minterlink-compressed, mmicromips,
25201 m14k, m14ke, m14kec): Document new options.
25202 (minterlink-mips16): Update documentation.
25203 * doc/md.texi (ZC, ZD): Document new constraints.
25204 * configure.ac (gcc_cv_as_micromips): Check if linker
25205 supports the .set micromips directive.
25206 * configure: Regenerate.
25207 * config.in: Regenerate.
25208 * config/mips/mips-tables.opt: Regenerate.
25209 * config/mips/micromips.md: New file.
25210 * constraints.md (ZC, ZD): New constraints.
25211 * config/mips/predicates.md (movep_src_register): New predicate.
25212 (movep_src_operand): New predicate.
25213 (non_volatile_mem_operand): New predicate.
25214 * config/mips/mips.md (multimem): New type.
25215 (length): Differentiate between 17-bit and 18-bit branch offsets.
25216 (MOVEP1, MOVEP2): New mode iterator.
25217 (mov_<load>l): Use ZC constraint.
25218 (mov_<load>r): Likewise.
25219 (mov_<store>l): Likewise.
25220 (mov_<store>r): Likewise.
25221 (*branch_equality<mode>_inverted): Add microMIPS support.
25222 (*branch_equality<mode>): Likewise.
25223 (*jump_absolute): Likewise.
25224 (indirect_jump_<mode>): Likewise.
25225 (tablejump_<mode>): Likewise.
25226 (<optab>_internal): Likewise.
25227 (sibcall_internal): Likewise.
25228 (sibcall_value_internal): Likewise.
25229 (prefetch): Use constraint ZD.
25230 * config/mips/mips.opt (minterlink-compressed): New option.
25231 (minterlink-mips16): Now an alias for minterlink-compressed.
25232 (mmicromips): New option.
25233 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
25234 (compare_and_swap_12): Likewise.
25235 (sync_add<mode>): Likewise.
25236 (sync_<optab>_12): Likewise.
25237 (sync_old_<optab>_12): Likewise.
25238 (sync_new_<optab>_12): Likewise.
25239 (sync_nand_12): Likewise.
25240 (sync_old_nand_12): Likewise.
25241 (sync_new_nand_12): Likewise.
25242 (sync_sub<mode>): Likewise.
25243 (sync_old_add<mode>): Likewise.
25244 (sync_old_sub<mode>): Likewise.
25245 (sync_new_add<mode>): Likewise.
25246 (sync_new_sub<mode>): Likewise.
25247 (sync_<optab><mode>): Likewise.
25248 (sync_old_<optab><mode>): Likewise.
25249 (sync_new_<optab><mode>): Likewise.
25250 (sync_nand<mode>): Likewise.
25251 (sync_old_nand<mode>): Likewise.
25252 (sync_new_nand<mode>): Likewise.
25253 (sync_lock_test_and_set<mode>): Likewise.
25254 (test_and_set_12): Likewise.
25255 (atomic_compare_and_swap<mode>): Likewise.
25256 (atomic_exchange<mode>_llsc): Likewise.
25257 (atomic_fetch_add<mode>_llsc): Likewise.
25258 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
25259 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
25260 (umips_save_restore_pattern_p): Likewise.
25261 (umips_load_store_pair_p): Likewise.
25262 (umips_output_load_store_pair): Likewise.
25263 (umips_movep_target_p): Likewise.
25264 (umips_12bit_offset_address_p): Likewise.
25265 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
25266 (mips_base_mips16): Rename this...
25267 (mips_base_compression_flags): ...to this. Update all uses.
25268 (mips_attribute_table): Add micromips, nomicromips and nocompression.
25269 (mips_mips16_decl_p): Delete.
25270 (mips_nomips16_decl_p): Delete.
25271 (mips_get_compress_on_flags): New function.
25272 (mips_get_compress_off_flags): New function.
25273 (mips_get_compress_mode): New function.
25274 (mips_get_compress_on_name): New function.
25275 (mips_get_compress_off_name): New function.
25276 (mips_insert_attributes): Support multiple compression types.
25277 (mips_merge_decl_attributes): Likewise.
25278 (umips_12bit_offset_address_p): New function.
25279 (mips_start_function_definition): Emit .set micromips directive.
25280 (mips_call_may_need_jalx_p): New function.
25281 (mips_function_ok_for_sibcall): Add microMIPS support.
25282 (mips_print_operand_punctuation): Support short delay slots and
25284 (umips_swm_mask, umips_swm_encoding): New.
25285 (umips_build_save_restore): New function.
25286 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
25287 (was_mips16_p): Remove.
25288 (old_compression_mode): New.
25289 (mips_set_compression_mode): New function.
25290 (mips_set_current_function): Add microMIPS support.
25291 (mips_option_override): Likewise.
25292 (umips_save_restore_pattern_p): New function.
25293 (umips_output_save_restore): New function.
25294 (umips_load_store_pair_p_1): New function.
25295 (umips_load_store_pair_p): New function.
25296 (umips_output_load_store_pair_1): New function.
25297 (umips_output_load_store_pair): New function.
25298 (umips_movep_target_p) New function.
25299 (mips_prepare_pch_save): Add microMIPS support.
25300 * config/mips/mips.h (TARGET_COMPRESSION): New.
25301 (TARGET_CPU_CPP_BUILTINS): Update macro
25302 to use new compression flags and to support microMIPS.
25303 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
25304 (MIPS_ARCH_FLOAT_SPEC): Likewise.
25305 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
25306 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
25307 (ASM_SPEC): Support mmicromips and mno-micromips.
25308 (M16STORE_REG_P): New macro.
25309 (MIPS_CALL): Support TARGET_MICROMIPS.
25310 (MICROMIPS_J): New macro.
25311 (mips_base_mips16): Rename this...
25312 (mips_base_compression_flags): ...to this.
25313 (UMIPS_12BIT_OFFSET_P): New macro.
25314 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
25315 (MULTILIB_DIRNAMES): Likewise.
25316 2013-03-20 Richard Biener <rguenther@suse.de>
25318 PR tree-optimization/56661
25319 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
25320 the result does not have to be distinct.
25322 2013-03-20 Richard Biener <rguenther@suse.de>
25324 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
25327 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
25328 Steven Bosscher <steven@gcc.gnu.org>
25330 PR rtl-optimization/56605
25331 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
25333 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25336 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
25337 that require movd instead of movq.
25339 2013-03-20 Richard Biener <rguenther@suse.de>
25341 * tree-ssa-structalias.c (struct variable_info): Add pointer
25342 to the first field of an aggregate with sub-vars. Make
25343 this and the pointer to the next subfield its ID.
25344 (vi_next): New function.
25345 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
25346 storedanything_id, integer_id): Increment by one.
25347 (new_var_info, get_call_vi, lookup_call_clobber_vi,
25348 get_call_clobber_vi): Adjust.
25349 (solution_set_expand): Simplify and speedup.
25350 (solution_set_add): Inline into ...
25351 (set_union_with_increment): ... this. Adjust accordingly.
25352 (do_sd_constraint): Likewise.
25353 (do_ds_constraint): Likewise.
25354 (do_complex_constraint): Simplify.
25355 (build_pred_graph): Adjust.
25356 (solve_graph): Likewise. Simplify and speedup.
25357 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
25358 get_constraint_for_component_ref, get_constraint_for_1,
25359 first_vi_for_offset, first_or_preceding_vi_for_offset,
25360 create_function_info_for, create_variable_info_for_1,
25361 create_variable_info_for, intra_create_variable_infos): Adjust.
25362 (init_base_vars): Push NULL for ID zero.
25363 (compute_points_to_sets): Adjust.
25365 2013-03-20 Richard Biener <rguenther@suse.de>
25367 * cfgloop.c (verify_loop_structure): Streamline and avoid
25368 ICEing on corrupt loop tree.
25369 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
25372 2013-03-20 Richard Biener <rguenther@suse.de>
25374 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
25375 check whether an SSA update is needed.
25377 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
25379 * config/mips/constraints.md (T): Rename to...
25383 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
25384 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
25386 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25388 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
25389 (*subsi3_carryin_uxtw): Likewise.
25391 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25393 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
25394 (*rorsi3_insn_uxtw): Likewise.
25396 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25398 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
25399 (*extrsi5_insn_uxtw): Likewise.
25401 2013-03-19 Richard Biener <rguenther@suse.de>
25403 PR tree-optimization/56273
25404 * passes.c (init_optimization_passes): Move second VRP after DOM.
25406 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
25408 * config/i386/i386.md (*movti_internal): Merge from
25409 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
25410 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
25411 nox64 isa attributes.
25413 2013-03-18 Richard Biener <rguenther@suse.de>
25415 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
25417 (merge_node_constraints): Likewise.
25418 (build_succ_graph): Likewise.
25419 (valid_graph_edge): Inline into single caller.
25420 (unify_nodes): Likewise. Use bitmap_set_bit return value
25422 (scc_visit): Fix formatting and variable use.
25423 (do_sd_constraint): Use gcc_checking_assert.
25424 (do_ds_constraint): Likewise.
25425 (do_complex_constraint): Likewise.
25426 (condense_visit): Likewise. Cleanup.
25427 (dump_pred_graph): New function.
25428 (perform_var_substitution): Dump the pred-graph before
25429 variable substitution.
25430 (find_equivalent_node): Use gcc_checking_assert.
25431 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
25433 2013-03-18 Richard Biener <rguenther@suse.de>
25435 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
25436 Remove cond_expr_stmt_list argument and do not gimplify the
25438 (vect_loop_versioning): Adjust.
25439 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
25440 Cleanup to use less temporaries.
25441 (vect_create_data_ref_ptr): Cleanup.
25443 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25445 PR tree-optimization/56635
25446 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
25447 require types_compatible_p types.
25449 2013-03-18 Nick Clifton <nickc@redhat.com>
25451 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
25452 spurious backslash.
25454 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
25455 Add missing line to comment describing function.
25457 2013-03-18 Richard Biener <rguenther@suse.de>
25459 PR tree-optimization/56210
25460 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
25461 Handle string / character search functions.
25462 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
25464 2013-03-18 Richard Biener <rguenther@suse.de>
25466 PR middle-end/56483
25467 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
25468 and implement properly.
25469 * gimple.h (gimple_cond_single_var_p): Remove.
25471 2013-03-18 Richard Biener <rguenther@suse.de>
25473 * tree-data-ref.h (find_data_references_in_loop): Declare.
25474 * tree-data-ref.c (get_references_in_stmt): Use a stack
25475 vector pre-allocated in the callers.
25476 (find_data_references_in_stmt): Adjust.
25477 (graphite_find_data_references_in_stmt): Likewise.
25478 (create_rdg_vertices): Likewise.
25479 (find_data_references_in_loop): Export.
25480 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
25481 Compute dependences here...
25482 (vect_analyze_data_refs): ...not here. When we encounter
25483 a non-vectorizable data reference in basic-block vectorization
25484 truncate the data reference vector. Do not bother to
25485 fixup data-dependence information for gather loads.
25486 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
25487 of data references, as reported.
25489 2013-03-18 Richard Biener <rguenther@suse.de>
25491 PR tree-optimization/3713
25492 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
25493 has_constants and expr.
25494 (stmt_has_constants): Properly valueize SSA names when deciding
25495 whether the stmt has constants.
25497 2013-03-18 Richard Biener <rguenther@suse.de>
25499 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
25500 whole function when there is nothing to do.
25501 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
25502 * tree-vectorizer.c (vectorize_loops): Update virtual and
25503 loop-closed SSA once.
25504 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
25506 2013-03-18 Richard Biener <rguenther@suse.de>
25508 PR middle-end/56113
25509 * domwalk.c (bb_postorder): New global static.
25510 (cmp_bb_postorder): New function.
25511 (walk_dominator_tree): Replace scheme imposing an order for
25512 visiting dominator sons by one sorting them at the time they
25513 are pushed on the stack.
25515 2013-03-18 Richard Biener <rguenther@suse.de>
25517 PR tree-optimization/39326
25518 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
25519 (struct mem_ref): Replace mem member with ao_ref typed member.
25520 (MEM_ANALYZABLE): Adjust.
25521 (memref_eq): Likewise.
25522 (mem_ref_alloc): Likewise.
25523 (gather_mem_refs_stmt): Likewise.
25524 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
25525 (execute_sm_if_changed_flag_set): Adjust.
25526 (execute_sm): Likewise.
25527 (ref_always_accessed_p): Likewise.
25528 (refs_independent_p): Likewise.
25529 (can_sm_ref_p): Likewise.
25531 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25534 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
25535 return 1 even for !unsignedp.
25537 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25539 * config/i386/i386.md (isa): Add x64 and nox64.
25540 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
25541 (*pushtf): Enable *roF alternative for x64 isa only.
25542 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
25543 mode attribute of integer alternatives to DImode for TARGET_64BIT.
25544 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
25545 (*movtf_internal): Merge from *movtf_internal_rex64 and
25546 *movtf_internal_sse. Use x64 and nox64 isa attributes.
25547 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
25548 nox64 isa attributes.
25549 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
25550 nox64 isa attributes.
25551 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
25553 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25555 * config/alpha/alpha.c (TARGET_LRA_P): New define.
25557 2013-03-17 Jakub Jelinek <jakub@redhat.com>
25560 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
25561 class names. Remove trailing comma after "ALL_REGS".
25563 2013-03-16 Jan Hubicka <jh@suse.cz>
25565 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
25566 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
25567 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
25568 of cgraph_get_create_node.
25569 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
25571 2013-03-16 Jason Merrill <jason@redhat.com>
25574 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
25575 with DW_AT_default_value.
25577 2013-03-16 Jakub Jelinek <jakub@redhat.com>
25579 * BASE-VER: Set to 4.9.0.
25581 2013-03-14 Andi Kleen <ak@linux.intel.com>
25584 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
25585 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
25586 Document _x* TSX intrinsics.
25588 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
25589 David Holsgrove <david.holsgrove@xilinx.com>
25591 * configure.ac: Add MicroBlaze TLS support detection.
25592 * configure: Regenerate.
25593 * config/microblaze/microblaze-protos.h
25594 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
25595 symbol_mentioned_p, label_mentioned_p): Add prototypes.
25596 * config/microblaze/microblaze.c (microblaze_address_type): Add
25597 ADDRESS_TLS and tls_reloc address types.
25598 (microblaze_address_info): Add tls_reloc.
25599 (TARGET_HAVE_TLS): Define.
25600 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
25601 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
25602 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
25603 load_tls_operand, microblaze_call_tls_get_addr,
25604 microblaze_legitimize_tls_address): New functions.
25605 (microblaze_classify_unspec): Handle UNSPEC_TLS.
25606 (get_base_reg): Use microblaze_tls_symbol_p.
25607 (microblaze_classify_address): Handle TLS.
25608 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
25609 label_mentioned_p and microblaze_tls_referenced_p.
25610 (microblaze_legitimize_address): Handle TLS.
25611 (microblaze_address_insns): Handle ADDRESS_TLS.
25612 (pic_address_needs_scratch): Handle TLS.
25613 (print_operand_address): Handle TLS.
25614 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
25615 (microblaze_expand_move): Handle TLS.
25616 (microblaze_legitimate_constant_p): Check
25617 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
25618 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
25619 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
25620 (PIC_OFFSET_TABLE_REGNUM): Set.
25621 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
25622 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
25623 (addsi3, movsi_internal2, movdf_internal): Update constraints
25624 * config/microblaze/predicates.md (arith_plus_operand): Define
25625 (move_operand): Redefine as move_src_operand,
25626 check microblaze_tls_referenced_p.
25628 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25630 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
25631 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
25633 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25635 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
25638 2013-03-14 Jakub Jelinek <jakub@redhat.com>
25640 PR tree-optimization/53265
25641 * common.opt (Waggressive-loop-optimizations): New option.
25642 * tree-ssa-loop-niter.c: Include tree-pass.h.
25643 (do_warn_aggressive_loop_optimizations): New function.
25644 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
25645 if number_of_latch_executions returned constant.
25646 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
25647 early. If number_of_latch_executions returned constant, set
25648 nb_iterations_upper_bound back to it.
25649 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
25651 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
25652 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
25654 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
25655 (MULTILIB_OSDIRNAMES): Set.
25656 * genmultilib: If defaultosdirname doesn't start with :: , set
25657 defaultosdirname2 instead, clear it and emit two . multilib_raw
25658 entries instead of just one.
25660 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
25662 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
25663 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
25664 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
25665 (SUBTARGET_OVERRIDE_OPTIONS): New.
25667 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
25670 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
25671 (musermode): Convert to Var(TARGET_USERMODE).
25672 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
25673 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
25674 * config/sh/sh.c (sh_option_override): Use
25675 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
25676 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
25678 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
25680 (udivsi3_i4_single, divsi3_i4_single): Use
25681 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
25683 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
25685 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
25688 2013-03-13 Richard Biener <rguenther@suse.de>
25690 PR tree-optimization/56608
25691 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
25692 calls when vectorizing basic-blocks.
25694 2013-03-13 Jakub Jelinek <jakub@redhat.com>
25697 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
25700 2013-03-12 Jakub Jelinek <jakub@redhat.com>
25702 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
25704 2013-03-11 Jan Hubicka <jh@suse.cz>
25707 * lto-streamer-out.c (output_symbol_p): Skip references from
25708 constructors of external variables.
25710 2013-03-11 Jan Hubicka <jh@suse.cz>
25712 PR middle-end/56571
25713 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
25715 * emit-rtl.c (verify_rtx_sharing): Likewise.
25716 (copy_insn_1): Likewise.
25717 * rtl.c (copy_rtx): Likewise.
25719 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
25722 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
25723 output_operand_lossage message.
25725 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
25728 * arm.c (shift_op): Validate RTL pattern on the fly.
25729 (arm_print_operand, case 'S'): Don't use shift_operator to validate
25732 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25735 * config/pa/pa.md (call_value): Check for calls to powf and direct to
25736 new call patterns that clobber %fr12.
25737 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
25738 split and postreload patterns.
25739 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
25740 registers %fr12 and %fr12R as call used.
25742 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
25744 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
25745 (canon_address, record_store, replace_read, check_mem_read_rtx,
25746 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
25747 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
25748 rest_of_handle_dse): Likewise.
25750 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
25752 PR middle-end/56524
25753 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
25755 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
25756 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
25757 (save_optabs_if_changed): Replace with...
25758 (init_tree_optimization_optabs): ...this.
25759 * optabs.c (save_optabs_if_changed): Rename to...
25760 (init_tree_optimization_optabs): ...this. Take the optimization node
25761 as argument. Do nothing if the base optabs are already correct.
25762 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
25763 to recompute optabs.
25764 * function.h (function): Remove optabs field.
25765 * function.c (invoke_set_current_function_hook): Call
25766 init_tree_optimization_optabs. Use the result to initialize
25769 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
25771 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
25772 if GTMA_HAS_NO_INSTRUMENTATION.
25773 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
25774 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
25775 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
25776 * gimple-pretty-print.c (dump_gimple_transaction): Handle
25777 GTMA_HAS_NO_INSTRUMENTATION.
25779 2013-03-08 Jakub Jelinek <jakub@redhat.com>
25781 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
25784 2013-03-08 Marek Polacek <polacek@redhat.com>
25785 Jakub Jelinek <jakub@redhat.com>
25787 PR tree-optimization/56478
25788 * predict.c (is_comparison_with_loop_invariant_p): Change the
25789 type of loop_step to tree.
25790 (predict_loops): Adjust.
25791 (predict_iv_comparison): Perform the computations on double_ints.
25793 2013-03-08 Richard Biener <rguenther@suse.de>
25795 PR tree-optimization/56570
25796 * tree-cfg.c (verify_expr_location_1): Verify locations for
25798 * tree-sra.c (create_access_replacement): Strip locations
25799 from DECL_DEBUG_EXPRs.
25801 2013-03-08 Richard Biener <rguenther@suse.de>
25803 * tree-inline.c (expand_call_inline): Do not associate
25804 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
25805 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
25807 2013-03-08 Richard Biener <rguenther@suse.de>
25809 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
25810 or block changes with -Og. Fix for location / block encoding
25811 changes and PHI arguments with locations.
25813 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
25815 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
25817 (struct output_info): Likewise.
25818 (register_overhead): Remove bad gcc_assert.
25819 (bitmap_find_bit): If there is only a single bitmap element, do not
25820 count a miss as a search.
25821 (print_statistics): Update for counter type changes.
25822 (dump_bitmap_statistics): Likewise. Print headers such that they
25823 are properly lined up with the printed counters.
25825 2013-03-07 Jakub Jelinek <jakub@redhat.com>
25827 PR tree-optimization/56559
25828 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
25829 check that it has only a single use.
25831 2013-03-07 Richard Biener <rguenther@suse.de>
25833 * doc/invoke.texi (fwhole-program): Discourage use in combination
25836 2013-03-06 Jakub Jelinek <jakub@redhat.com>
25838 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
25840 PR tree-optimization/56539
25841 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
25842 instead of GSI_CONTINUE_LINKING as last argument to
25843 force_gimple_operand_gsi. Adjust function comment.
25845 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
25848 PR middle-end/56548
25849 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
25850 promoted mode, convert the result back to the original mode.
25852 2013-03-06 Richard Biener <rguenther@suse.de>
25854 PR middle-end/56294
25855 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
25856 (insert_updated_phi_nodes_compare_uids): New function.
25857 (update_ssa): Sort symbols_to_rename after UID before
25858 traversing it to insert PHI nodes.
25860 2013-03-06 Richard Biener <rguenther@suse.de>
25862 PR middle-end/50494
25863 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
25864 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
25867 2013-02-13 Richard Biener <rguenther@suse.de>
25870 * varasm.c (output_constant_def_1): Get the decl representing
25871 the constant as argument.
25872 (output_constant_def): Wrap output_constant_def_1.
25873 (make_decl_rtl): Use output_constant_def_1 with the decl
25874 representing the constant.
25875 (build_constant_desc): Optionally re-use a decl already
25876 representing the constant.
25877 (tree_output_constant_def): Adjust.
25879 2013-03-06 Joey Ye <joey.ye@arm.com>
25882 * gcc.c (convert_white_space): New function.
25883 (main): Handles white space in function name.
25885 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
25888 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
25889 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
25890 to SH_DIV_CALL_TABLE for TARGET_SH2.
25891 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
25893 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
25894 call-table options.
25896 2013-03-05 Sterling Augustine <saugustine@google.com>
25897 Cary Coutant <ccoutant@google.com>
25900 * dwarf2out.c (resolve_addr): Don't call
25901 remove_loc_list_addr_table_entries a second time for the same
25904 2013-03-05 Jakub Jelinek <jakub@redhat.com>
25907 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
25908 (avoid_complex_debug_insns): New function.
25909 (expand_debug_locations): Call it.
25911 PR rtl-optimization/56484
25912 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
25913 lifetimes of hard registers on small register class machines.
25915 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
25917 * config/microblaze/microblaze-protos.h: Rename
25918 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
25919 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
25921 (microblaze_fast_interrupt_function_p): New function.
25922 (microblaze_is_interrupt_handler): Rename to
25923 microblaze_is_interrupt_variant and add fast_interrupt check.
25924 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
25925 (save_restore_insns): Likewise.
25926 (compute_frame_size): Likewise.
25927 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
25928 (microblaze_globalize_label): Likewise.
25929 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
25930 * config/microblaze/microblaze.md: Use wrapper
25931 microblaze_is_interrupt_variant.
25933 2013-03-05 Kai Tietz <ktietz@redhat.com>
25935 * sdbout.c (sdbout_one_type): Switch to current function's section
25936 supporting cold/hot.
25938 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
25940 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
25943 2013-03-05 Jakub Jelinek <jakub@redhat.com>
25945 PR middle-end/56461
25946 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
25947 if VALGRIND_GET_VBITS is defined, temporarily make object
25948 memory all defined, and restore previous valgrind addressability
25949 and definability afterwards. Free this_object at the end.
25951 PR middle-end/56461
25952 * lra.c (lra): Call lra_clear_live_ranges if live_p,
25953 right before calling lra_create_live_ranges, also call it
25954 when clearing live_p. Only call lra_clear_live_ranges
25955 at the end if live_p.
25957 PR middle-end/56461
25958 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
25960 2013-03-05 Richard Biener <rguenther@suse.de>
25962 PR tree-optimization/56521
25963 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
25966 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
25969 * except.h (remove_unreachable_eh_regions): New prototype.
25970 * except.c (remove_eh_handler_splicer): New function, split out
25971 of remove_eh_handler.
25972 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
25973 warning about running it on many EH regions one at a time.
25974 (remove_unreachable_eh_regions_worker): New function, walk the
25975 EH tree in depth-first order and remove non-marked regions.
25976 (remove_unreachable_eh_regions): New function.
25977 * tree-eh.c (mark_reachable_handlers): New function, split out
25978 from remove_unreachable_handlers.
25979 (remove_unreachable_handlers): Use mark_reachable_handlers and
25980 remove_unreachable_eh_regions.
25981 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
25982 and remove_unreachable_eh_regions.
25984 2013-03-05 Richard Biener <rguenther@suse.de>
25986 PR middle-end/56525
25987 * loop-init.c (fix_loop_structure): Remove loops in two stages,
25988 not freeing them until the end.
25990 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25992 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
25994 2013-03-05 Richard Biener <rguenther@suse.de>
25996 PR tree-optimization/56270
25997 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
25998 of loads after scheduling an SLP instance.
26000 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26002 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
26004 (check_gcc_parallelize): Run guality.exp as a separate job from
26005 vect.exp with unsorted.exp and $(dg_target_exps) separately from
26006 struct-layout-1.exp with stackalign.exp.
26008 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
26010 PR middle-end/56461
26011 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
26012 load_index sbitmap even if some bit in it isn't set.
26014 PR middle-end/56461
26015 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
26016 (discover_iteration_bound_by_body_walk): Change queues to
26017 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
26018 spelling in comment. Call safe_push on queues[bound_index] directly.
26019 Release queues[queue_index] in every iteration unconditionally.
26020 Release bounds vector.
26022 PR middle-end/56461
26023 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
26024 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
26025 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
26028 2013-03-05 Richard Biener <rguenther@suse.de>
26031 * tree-inline.c (remap_blocks_to_null): New function.
26032 (expand_call_inline): When expanding a call stmt without
26033 an associated block inline remap all callee blocks to NULL.
26035 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26037 PR rtl-optimization/56494
26038 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
26039 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
26040 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
26042 PR middle-end/56461
26043 * sel-sched-ir.c (free_sched_pools): Release
26044 succs_info_pool.stack[succs_info_pool.max_top] vectors too
26045 if succs_info_pool.max_top isn't -1.
26048 * opts.c (opts_obstack, opts_concat): Moved to...
26049 * opts-common.c (opts_obstack, opts_concat): ... here.
26051 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26053 PR middle-end/56461
26054 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
26056 2013-03-04 Martin Jambor <mjambor@suse.cz>
26058 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
26059 all appropriate places.
26061 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
26063 PR tree-optimization/56424
26064 * ipa-split.c (split_function): Do not set the RSO flag if result is
26065 not by reference and its type is a register type.
26067 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
26069 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
26070 (microblaze_legitimate_pic_operand): Likewise
26071 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
26072 new function microblaze_legitimate_pic_operand
26073 * config/microblaze/microblaze-protos.h
26074 (microblaze_legitimate_pic_operand): Declare.
26076 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26078 * config/microblaze/predicates.md (call_insn_simple_operand):
26079 New predicate for supported rtx code types.
26080 * config/microblaze/microblaze.md (call_internal1): Use
26081 call_insn_simple_operand predicate.
26083 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26085 PR middle-end/56461
26086 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
26087 partitions.ordered_remove.
26089 PR middle-end/56461
26090 * tree-vect-stmts.c (vectorizable_conversion): Don't call
26091 vec_oprnds0.create (1) for modifier == NONE.
26093 PR middle-end/56461
26094 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
26095 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
26096 vec_oprnds1 right before pushing anything to it for
26099 PR middle-end/56461
26100 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
26101 set nbbs to 0 instead of having separate code path.
26102 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
26103 instead of false as last argument if returning NULL.
26105 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
26107 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
26108 the attribute is now called "target" instead of "option".
26109 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
26110 * doc/tm.texi.in (Target Attributes): Likewise document the correct
26111 attribute/pragma name for TARGET_OPTION_VALID_P and
26112 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
26113 * doc/tm.texi: Regenerated.
26115 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
26117 * config/microblaze/microblaze.c:
26118 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
26119 * config/microblaze/microblaze.h: Add -mxl-reorder to
26121 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
26122 instructions emitted if TARGET_REORDER.
26123 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
26124 or 0 for -m/-mno case, but initialises as 2 to detect default use case
26127 2013-03-01 Xinliang David Li <davidxl@google.com>
26129 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
26132 2013-03-01 Jakub Jelinek <jakub@redhat.com>
26134 PR middle-end/56461
26135 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
26136 vector even when returning true. Fix up function comment formatting.
26138 PR middle-end/56461
26139 * ira-build.c (ira_loop_nodes_count): New variable.
26140 (create_loop_tree_nodes): Initialize it.
26141 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
26143 PR middle-end/56461
26144 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
26145 method on dr_chain and result_chain.
26146 * tree-vect-stmts.c (vectorizable_store): Only call
26147 result_chain.create if j == 0.
26149 PR middle-end/56461
26150 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
26151 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
26152 before overwriting it.
26154 2013-03-01 Tobias Burnus <burnus@net-b.de>
26156 * doc/extended.texi (C Extensions): Change order in @menu
26158 (Other MIPS Built-in Functions): Move last MIPS entry before
26159 "picoChip Built-in Functions".
26160 (SH Built-in Functions): Move after RX Built-in Functions.
26161 * doc/gcc.texi (Introduction): Change order in @menu
26163 * doc/md.texi (Constraints): Ditto.
26164 * gty.texi (Type Information): Ditto.
26165 (User-provided marking routines for template types): Make
26167 * doc/invoke.texi (AArch64 Options): Move before
26168 "Adapteva Epiphany Options".
26170 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
26171 Jakub Jelinek <jakub@redhat.com>
26174 * asan.c (gate_asan): Lookup no_sanitize_address instead of
26175 no_address_safety_analysis attribute.
26176 * doc/extend.texi (no_address_safety_attribute): Rename to
26177 no_sanitize_address attribute, mention no_address_safety_analysis
26178 attribute as deprecated alias.
26180 2013-02-28 Jakub Jelinek <jakub@redhat.com>
26182 PR middle-end/56461
26183 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
26184 type to vec<vec<tree> > *.
26185 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
26186 to be vec<tree> instead of vec<tree> *, set vec_defs
26187 to vNULL and call vec_defs.create (number_of_vects), adjust other
26189 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
26190 vectorizable_condition): Adjust vect_get_slp_defs callers.
26192 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26194 * config/aarch64/aarch64.c
26195 (aarch64_float_const_representable): Remove unused variable.
26197 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26199 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
26201 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26203 * config/aarch64/aarch64-builtins.c
26204 (aarch64_init_simd_builtins): Make static.
26206 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26208 * config/aarch64/aarch64.c
26209 (aarch64_simd_make_constant): Make static.
26211 2013-02-28 Martin Jambor <mjambor@suse.cz>
26213 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
26214 with no initialization to the RHS of debug statements.
26216 2013-02-28 Martin Jambor <mjambor@suse.cz>
26218 PR tree-optimization/56294
26219 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
26221 (get_access_replacement): Do not call create_access_replacement.
26222 Assert a replacement exists.
26223 (get_repl_default_def_ssa_name): Create the replacement declaration
26226 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26228 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
26229 final_end_function.
26231 2013-02-28 Marek Polacek <polacek@redhat.com>
26233 PR rtl-optimization/56466
26234 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
26235 if we're changing a loop.
26236 (peel_loops_completely): Likewise.
26238 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
26241 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
26243 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26246 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
26247 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
26248 INTX_FTYPE_FX, FX_FTYPE_INTX.
26249 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
26251 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26253 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
26254 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
26255 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
26256 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
26257 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
26258 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
26259 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
26260 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
26261 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
26262 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
26263 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
26264 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
26265 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
26266 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
26267 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
26268 (avrxmega6): Increase max flash segments from 5 to 6.
26269 * config/avr/t-multilib: Regenerate.
26270 * config/avr/avr-tables.opt: Regenerate.
26271 * doc/avr-mmcu.texi: Regenerate.
26273 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26275 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
26276 (avr_device_to_arch): Rename to avr_device_to_ld.
26277 (avr_device_to_as): New prototype.
26278 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
26279 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
26280 * config/avr/driver-avr.c (avr_device_to_as): New.
26281 (avr_device_to_arch): Rename to avr_device_to_ld.
26283 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26285 PR middle-end/56461
26286 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
26287 method on dr_chain and result_chain.
26289 PR middle-end/56461
26290 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
26291 pointer_set_destroy on not_executed_last_iteration.
26293 PR middle-end/56461
26294 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
26296 PR middle-end/56461
26297 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
26298 FOR_EACH_DEFINED_FUNCTION when freeing state.
26300 PR middle-end/56461
26301 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
26303 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
26306 PR middle-end/56461
26307 * ipa-cp.c (decide_whether_version_node): Call vec_free on
26308 known_aggs[i].items and release known_aggs vector.
26310 PR middle-end/56461
26311 * ipa-reference.c (propagate): Free node_info even for alias nodes.
26313 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26315 * config/microblaze/microblaze.c (microblaze_emit_compare):
26316 Use xor for EQ/NE comparisions.
26317 * config/microblaze/microblaze.md (cstoresf4): Add constraints
26318 (cbranchsf4): Adjust operator to comparison_operator.
26320 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26322 PR middle-end/56461
26323 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
26325 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
26326 vec_safe_push, always update *slot.
26327 (redirect_edge_var_map_clear): Use vec_free.
26328 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
26329 (free_var_map_entry): Use vec_free.
26330 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
26331 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
26333 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
26335 PR middle-end/45472
26336 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
26337 when the may_trap_p bit of the exprs being merged differs.
26338 Reorder tests for speculativeness in the logical and operator.
26340 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26342 * incpath.c (add_standard_paths): Use reconcat instead of concat
26343 where appropriate and avoid leaking memory.
26345 * opts.h: Include obstack.h.
26346 (opts_concat): New prototype.
26347 (opts_obstack): New declaration.
26348 * opts.c (opts_concat): New function.
26349 (opts_obstack): New variable.
26350 (init_options_struct): Call gcc_init_obstack on opts_obstack.
26351 (finish_options): Use opts_concat instead of concat
26352 and XOBNEWVEC instead of XNEWVEC.
26353 * opts-common.c (generate_canonical_option, decode_cmdline_option,
26354 generate_option): Likewise.
26355 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
26356 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
26359 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
26360 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
26362 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26364 PR middle-end/56461
26365 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
26367 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26369 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
26370 (arm_block_move_unaligned_straight): Likewise.
26371 (arm_adjust_block_mem): Likewise.
26373 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26376 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
26377 temp, cond and label.
26378 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
26381 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
26382 * config/c6x/c6x.h (dbx_register_map): Update declaration.
26385 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
26386 of prologue/epilogue functions.
26389 * config/tilegx/tilegx.c (tilegx_expand_prologue):
26390 Remove unused variable cfa_offset.
26391 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
26394 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
26395 type promotion to unsigned.
26398 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
26399 for HOST_WIDE_INT of 32 bit / same size as int.
26400 (arm_block_move_unaligned_straight): Likewise.
26401 (arm_adjust_block_mem): Likewise.
26404 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
26407 2013-02-26 Marek Polacek <polacek@redhat.com>
26409 PR tree-optimization/56426
26410 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
26412 2013-02-26 Richard Biener <rguenther@suse.de>
26415 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
26416 unused variable loops.
26418 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26420 PR tree-optimization/56448
26421 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
26422 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
26423 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
26424 later operands of the references, or even first operand for
26425 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
26427 PR tree-optimization/56443
26428 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
26429 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
26430 to type_for_mode langhook.
26432 2013-02-25 Matt Turner <mattst88@gmail.com>
26434 * doc/invoke.texi: Document r4700.
26436 2013-02-25 Richard Biener <rguenther@suse.de>
26438 PR tree-optimization/56175
26439 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
26441 (simplify_bitwise_binary): ... here. Also guard the conversion
26442 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
26444 2013-02-25 Catherine Moore <clm@codesourcery.com>
26447 2013-02-24 Catherine Moore <clm@codesourcery.com>
26448 Maciej W. Rozycki <macro@codesourcery.com>
26449 Tom de Vries <tom@codesourcery.com>
26450 Nathan Sidwell <nathan@codesourcery.com>
26451 Iain Sandoe <iain@codesourcery.com>
26452 Nathan Froyd <froydnj@codesourcery.com>
26453 Chao-ying Fu <fu@mips.com>
26455 * doc/extend.texi: (micromips, nomicromips, nocompression):
26456 Document new function attributes.
26457 * doc/invoke.texi (minterlink-compressed, mmicromips,
26458 m14k, m14ke, m14kec): Document new options.
26459 (minterlink-mips16): Update documentation.
26460 * doc/md.texi (ZC, ZD): Document new constraints.
26461 * configure.ac (gcc_cv_as_micromips): Check if linker
26462 supports the .set micromips directive.
26463 * configure: Regenerate.
26464 * config.in: Regenerate.
26465 * config/mips/mips-tables.opt: Regenerate.
26466 * config/mips/micromips.md: New file.
26467 * constraints.md (ZC, AD): New constraints.
26468 * config/mips/predicates.md (movep_src_register): New predicate.
26469 (movep_src_operand): New predicate.
26470 (non_volatile_mem_operand): New predicate.
26471 * config/mips/mips.md (multimem): New type.
26472 (length): Differentiate between 17-bit and 18-bit branch offsets.
26473 (MOVEP1, MOVEP2): New mode iterator.
26474 (mov_<load>l): Use ZC constraint.
26475 (mov_<load>r): Likewise.
26476 (mov_<store>l): Likewise.
26477 (mov_<store>r): Likewise.
26478 (*branch_equality<mode>_inverted): Add microMIPS support.
26479 (*branch_equality<mode>): Likewise.
26480 (*jump_absolute): Likewise.
26481 (indirect_jump_<mode>): Likewise.
26482 (tablejump_<mode>): Likewise.
26483 (<optab>_internal): Likewise.
26484 (sibcall_internal): Likewise.
26485 (sibcall_value_internal): Likewise.
26486 (prefetch): Use constraint ZD.
26487 * config/mips/mips.opt (minterlink-compressed): New option.
26488 (minterlink-mips16): Now an alias for minterlink-compressed.
26489 (mmicromips): New option.
26490 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26491 (compare_and_swap_12): Likewise.
26492 (sync_add<mode>): Likewise.
26493 (sync_<optab>_12): Likewise.
26494 (sync_old_<optab>_12): Likewise.
26495 (sync_new_<optab>_12): Likewise.
26496 (sync_nand_12): Likewise.
26497 (sync_old_nand_12): Likewise.
26498 (sync_new_nand_12): Likewise.
26499 (sync_sub<mode>): Likewise.
26500 (sync_old_add<mode>): Likewise.
26501 (sync_old_sub<mode>): Likewise.
26502 (sync_new_add<mode>): Likewise.
26503 (sync_new_sub<mode>): Likewise.
26504 (sync_<optab><mode>): Likewise.
26505 (sync_old_<optab><mode>): Likewise.
26506 (sync_new_<optab><mode>): Likewise.
26507 (sync_nand<mode>): Likewise.
26508 (sync_old_nand<mode>): Likewise.
26509 (sync_new_nand<mode>): Likewise.
26510 (sync_lock_test_and_set<mode>): Likewise.
26511 (test_and_set_12): Likewise.
26512 (atomic_compare_and_swap<mode>): Likewise.
26513 (atomic_exchange<mode>_llsc): Likewise.
26514 (atomic_fetch_add<mode>_llsc): Likewise.
26515 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26516 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26517 (umips_save_restore_pattern_p): Likewise.
26518 (umips_load_store_pair_p): Likewise.
26519 (umips_output_load_store_pair): Likewise.
26520 (umips_movep_target_p): Likewise.
26521 (umips_12bit_offset_address_p): Likewise.
26522 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26523 (mips_base_mips16): Rename this...
26524 (mips_base_compression_flags): ...to this. Update all uses.
26525 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26526 (mips_mips16_decl_p): Delete.
26527 (mips_nomips16_decl_p): Delete.
26528 (mips_get_compress_on_flags): New function.
26529 (mips_get_compress_off_flags): New function.
26530 (mips_get_compress_mode): New function.
26531 (mips_get_compress_on_name): New function.
26532 (mips_get_compress_off_name): New function.
26533 (mips_insert_attributes): Support multiple compression types.
26534 (mips_merge_decl_attributes): Likewise.
26535 (umips_12bit_offset_address_p): New function.
26536 (mips_start_function_definition): Emit .set micromips directive.
26537 (mips_call_may_need_jalx_p): New function.
26538 (mips_function_ok_for_sibcall): Add microMIPS support.
26539 (mips_print_operand_punctuation): Support short delay slots and
26541 (umips_swm_mask, umips_swm_encoding): New.
26542 (umips_build_save_restore): New function.
26543 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26544 (was_mips16_p): Remove.
26545 (old_compression_mode): New.
26546 (mips_set_compression_mode): New function.
26547 (mips_set_current_function): Add microMIPS support.
26548 (mips_option_override): Likewise.
26549 (umips_save_restore_pattern_p): New function.
26550 (umips_output_save_restore): New function.
26551 (umips_load_store_pair_p_1): New function.
26552 (umips_load_store_pair_p): New function.
26553 (umips_output_load_store_pair_1): New function.
26554 (umips_output_load_store_pair): New function.
26555 (umips_movep_target_p) New function.
26556 (mips_prepare_pch_save): Add microMIPS support.
26557 * config/mips/mips.h (TARGET_COMPRESSION): New.
26558 (TARGET_CPU_CPP_BUILTINS): Update macro
26559 to use new compression flags and to support microMIPS.
26560 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26561 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26562 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26563 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26564 (ASM_SPEC): Support mmicromips and mno-micromips.
26565 (M16STORE_REG_P): New macro.
26566 (MIPS_CALL): Support TARGET_MICROMIPS.
26567 (MICROMIPS_J): New macro.
26568 (mips_base_mips16): Rename this...
26569 (mips_base_compression_flags): ...to this.
26570 (UMIPS_12BIT_OFFSET_P): New macro.
26571 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26572 (MULTILIB_DIRNAMES): Likewise.
26574 2013-02-25 Tom de Vries <tom@codesourcery.com>
26576 PR rtl-optimization/56131
26577 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
26578 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
26579 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
26581 2013-02-25 Tobias Burnus <burnus@net-b.de>
26583 * doc/invoke.texi (-fsanitize=): Move from optimization
26584 to debugging options.
26586 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26588 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
26590 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26591 Alexander Monakov <amonakov@ispras.ru>
26593 PR middle-end/56077
26594 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
26595 flush pending lists also on non-jumps. Adjust comment.
26597 2013-02-24 Catherine Moore <clm@codesourcery.com>
26598 Maciej W. Rozycki <macro@codesourcery.com>
26599 Tom de Vries <tom@codesourcery.com>
26600 Nathan Sidwell <nathan@codesourcery.com>
26601 Iain Sandoe <iain@codesourcery.com>
26602 Nathan Froyd <froydnj@codesourcery.com>
26603 Chao-ying Fu <fu@mips.com>
26605 * doc/extend.texi: (micromips, nomicromips, nocompression):
26606 Document new function attributes.
26607 * doc/invoke.texi (minterlink-compressed, mmicromips,
26608 m14k, m14ke, m14kec): Document new options.
26609 (minterlink-mips16): Update documentation.
26610 * doc/md.texi (ZC, ZD): Document new constraints.
26611 * configure.ac (gcc_cv_as_micromips): Check if linker
26612 supports the .set micromips directive.
26613 * configure: Regenerate.
26614 * config.in: Regenerate.
26615 * config/mips/mips-tables.opt: Regenerate.
26616 * config/mips/micromips.md: New file.
26617 * constraints.md (ZC, AD): New constraints.
26618 * config/mips/predicates.md (movep_src_register): New predicate.
26619 (movep_src_operand): New predicate.
26620 (non_volatile_mem_operand): New predicate.
26621 * config/mips/mips.md (multimem): New type.
26622 (length): Differentiate between 17-bit and 18-bit branch offsets.
26623 (MOVEP1, MOVEP2): New mode iterator.
26624 (mov_<load>l): Use ZC constraint.
26625 (mov_<load>r): Likewise.
26626 (mov_<store>l): Likewise.
26627 (mov_<store>r): Likewise.
26628 (*branch_equality<mode>_inverted): Add microMIPS support.
26629 (*branch_equality<mode>): Likewise.
26630 (*jump_absolute): Likewise.
26631 (indirect_jump_<mode>): Likewise.
26632 (tablejump_<mode>): Likewise.
26633 (<optab>_internal): Likewise.
26634 (sibcall_internal): Likewise.
26635 (sibcall_value_internal): Likewise.
26636 (prefetch): Use constraint ZD.
26637 * config/mips/mips.opt (minterlink-compressed): New option.
26638 (minterlink-mips16): Now an alias for minterlink-compressed.
26639 (mmicromips): New option.
26640 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26641 (compare_and_swap_12): Likewise.
26642 (sync_add<mode>): Likewise.
26643 (sync_<optab>_12): Likewise.
26644 (sync_old_<optab>_12): Likewise.
26645 (sync_new_<optab>_12): Likewise.
26646 (sync_nand_12): Likewise.
26647 (sync_old_nand_12): Likewise.
26648 (sync_new_nand_12): Likewise.
26649 (sync_sub<mode>): Likewise.
26650 (sync_old_add<mode>): Likewise.
26651 (sync_old_sub<mode>): Likewise.
26652 (sync_new_add<mode>): Likewise.
26653 (sync_new_sub<mode>): Likewise.
26654 (sync_<optab><mode>): Likewise.
26655 (sync_old_<optab><mode>): Likewise.
26656 (sync_new_<optab><mode>): Likewise.
26657 (sync_nand<mode>): Likewise.
26658 (sync_old_nand<mode>): Likewise.
26659 (sync_new_nand<mode>): Likewise.
26660 (sync_lock_test_and_set<mode>): Likewise.
26661 (test_and_set_12): Likewise.
26662 (atomic_compare_and_swap<mode>): Likewise.
26663 (atomic_exchange<mode>_llsc): Likewise.
26664 (atomic_fetch_add<mode>_llsc): Likewise.
26665 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26666 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26667 (umips_save_restore_pattern_p): Likewise.
26668 (umips_load_store_pair_p): Likewise.
26669 (umips_output_load_store_pair): Likewise.
26670 (umips_movep_target_p): Likewise.
26671 (umips_12bit_offset_address_p): Likewise.
26672 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26673 (mips_base_mips16): Rename this...
26674 (mips_base_compression_flags): ...to this. Update all uses.
26675 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26676 (mips_mips16_decl_p): Delete.
26677 (mips_nomips16_decl_p): Delete.
26678 (mips_get_compress_on_flags): New function.
26679 (mips_get_compress_off_flags): New function.
26680 (mips_get_compress_mode): New function.
26681 (mips_get_compress_on_name): New function.
26682 (mips_get_compress_off_name): New function.
26683 (mips_insert_attributes): Support multiple compression types.
26684 (mips_merge_decl_attributes): Likewise.
26685 (umips_12bit_offset_address_p): New function.
26686 (mips_start_function_definition): Emit .set micromips directive.
26687 (mips_call_may_need_jalx_p): New function.
26688 (mips_function_ok_for_sibcall): Add microMIPS support.
26689 (mips_print_operand_punctuation): Support short delay slots and
26691 (umips_swm_mask, umips_swm_encoding): New.
26692 (umips_build_save_restore): New function.
26693 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26694 (was_mips16_p): Remove.
26695 (old_compression_mode): New.
26696 (mips_set_compression_mode): New function.
26697 (mips_set_current_function): Add microMIPS support.
26698 (mips_option_override): Likewise.
26699 (umips_save_restore_pattern_p): New function.
26700 (umips_output_save_restore): New function.
26701 (umips_load_store_pair_p_1): New function.
26702 (umips_load_store_pair_p): New function.
26703 (umips_output_load_store_pair_1): New function.
26704 (umips_output_load_store_pair): New function.
26705 (umips_movep_target_p) New function.
26706 (mips_prepare_pch_save): Add microMIPS support.
26707 * config/mips/mips.h (TARGET_COMPRESSION): New.
26708 (TARGET_CPU_CPP_BUILTINS): Update macro
26709 to use new compression flags and to support microMIPS.
26710 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26711 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26712 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26713 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26714 (ASM_SPEC): Support mmicromips and mno-micromips.
26715 (M16STORE_REG_P): New macro.
26716 (MIPS_CALL): Support TARGET_MICROMIPS.
26717 (MICROMIPS_J): New macro.
26718 (mips_base_mips16): Rename this...
26719 (mips_base_compression_flags): ...to this.
26720 (UMIPS_12BIT_OFFSET_P): New macro.
26721 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26722 (MULTILIB_DIRNAMES): Likewise.
26724 2013-02-24 Jakub Jelinek <jakub@redhat.com>
26727 * target-globals.c (save_target_globals): For init_reg_sets and
26728 target_reinit remporarily set this_fn_optabs to this_target_optabs.
26730 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
26732 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
26733 * config/aarch64/t-aarch64
26734 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
26736 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
26738 PR inline-asm/56148
26739 * lra-constraints.c (process_alt_operands): Reload operand
26740 conflicting with earlier clobber only if no more other conflicting
26743 2013-02-22 Jakub Jelinek <jakub@redhat.com>
26746 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
26747 if not linking a shared library.
26749 2013-02-22 Seth LaForge <sethml@google.com>
26751 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
26753 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
26755 * config/arm/arm.md (split for extendsidi): Update condition.
26756 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
26757 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
26758 (qhs_zextenddi_cstr): Likewise.
26760 2013-02-21 Jakub Jelinek <jakub@redhat.com>
26762 PR middle-end/56420
26763 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
26764 avoid signed wrapping.
26765 (expand_mult): Handle properly multiplication by
26766 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
26767 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
26768 in the compiler if coeff is HOST_WIDE_INT_MIN.
26769 (expand_divmod): Don't make ext_op1 static, change it's type to
26770 uhwi. Avoid undefined behavior in -INTVAL (op1).
26772 PR rtl-optimization/50339
26773 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
26775 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
26776 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
26777 into splitting_ashiftrt field.
26778 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
26780 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
26783 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
26785 PR middle-end/56108
26786 * trans-mem.c (execute_tm_mark): Do not expand transactions that
26787 are sure to go irrevocable.
26789 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
26791 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
26792 scalars are valid operands.
26794 2013-02-21 Martin Jambor <mjambor@suse.cz>
26796 PR tree-optimization/56310
26797 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
26798 only matching indices and non-negative final offsets.
26799 (intersect_aggregates_with_edge): Pass src_idx to
26800 agg_replacements_to_vector. Pass src_idx insstead of index to
26801 intersect_with_agg_replacements.
26803 2013-02-21 Martin Jambor <mjambor@suse.cz>
26805 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
26806 instead of hard-wired defaults.
26808 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
26810 * doc/invoke.texi (MIPS Options): Update documentation of the
26811 floating-point multiply-accumulate instruction restrictions.
26813 2013-02-21 Kostya Serebryany <kcc@google.com>
26815 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
26816 asan_shadow_offset on x86_64 linux.
26818 2013-02-21 Richard Biener <rguenther@suse.de>
26820 PR tree-optimization/56415
26822 2013-02-11 Richard Biener <rguenther@suse.de>
26824 PR tree-optimization/56273
26825 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
26828 2013-02-21 Jakub Jelinek <jakub@redhat.com>
26831 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
26834 PR inline-asm/56405
26835 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
26836 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
26838 2013-02-20 Jan Hubicka <jh@suse.cz>
26840 PR tree-optimization/56265
26841 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
26842 when target is referenced for first time.
26844 2013-02-20 Richard Biener <rguenther@suse.de>
26846 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
26847 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
26848 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
26849 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
26850 not return anything.
26851 (rename_ssa_copies): Do not remove unused locals.
26852 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
26853 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
26854 * passes.c (execute_function_todo): Do not schedule unused locals
26855 removal if cleanup_tree_cfg did something.
26856 * tree-ssa-live.c (remove_unused_locals): Dump statistics
26857 about the number of removed locals.
26859 2013-02-20 Richard Biener <rguenther@suse.de>
26861 PR tree-optimization/56398
26862 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
26864 2013-02-20 Martin Jambor <mjambor@suse.cz>
26866 PR tree-optimization/55334
26867 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
26868 restricted pointers to arrays.
26870 2013-02-20 Richard Biener <rguenther@suse.de>
26871 Jakub Jelinek <jakub@redhat.com>
26873 PR tree-optimization/56396
26874 * tree-ssa-ccp.c (n_const_val): New static variable.
26875 (get_value): Return NULL for SSA names we don't have a lattice
26877 (ccp_initialize): Initialize n_const_val.
26878 * tree-ssa-copy.c (n_copy_of): New static variable.
26879 (init_copy_prop): Initialize n_copy_of.
26880 (get_value): Return NULL_TREE for SSA names we don't have a
26883 2013-02-20 Martin Jambor <mjambor@suse.cz>
26885 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
26887 2013-02-20 Richard Biener <rguenther@suse.de>
26889 * genpreds.c (write_lookup_constraint): Do not compare first
26890 letter of the constraint again.
26892 2013-02-20 Richard Biener <rguenther@suse.de>
26894 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
26896 (get_use_iv_cost): Terminate hashtable walk when coming across
26899 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
26901 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
26902 reassociation for avx2 targets.
26904 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26906 * config/microblaze/microblaze.c: microblaze_has_clz = 0
26907 Add version check for v8.10.a to enable microblaze_has_clz
26908 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
26909 version and TARGET_PATTERN_COMPARE check
26910 * config/microblaze/microblaze.md: New clzsi2 instruction
26912 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26914 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
26915 function before branching.
26917 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
26919 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
26921 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
26923 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
26925 PR middle-end/55889
26926 * sel-sched.c: Include ira.h.
26927 (implicit_clobber_conflict_p): New function.
26928 (moveup_expr): Use it.
26929 * Makefile.in (sel-sched.o): Depend on ira.h.
26931 2013-02-19 Richard Biener <rguenther@suse.de>
26933 PR tree-optimization/56384
26934 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
26935 (vn_hash_type): Split out from ...
26936 (vn_hash_constant_with_type): ... here.
26937 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
26938 (vn_phi_eq): Compare types from vn_phi_s structure.
26939 (vn_phi_lookup): Populate vn_phi_s type.
26940 (vn_phi_insert): Likewise.
26942 2013-02-19 Jakub Jelinek <jakub@redhat.com>
26944 PR tree-optimization/56350
26945 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
26946 if haven't found reduction or nested cycle operand, rather than
26947 asserting we must find it.
26949 PR tree-optimization/56381
26950 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
26953 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
26954 Jakub Jelinek <jakub@redhat.com>
26957 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
26958 (swap_optab_enable): Same.
26959 (init_all_optabs): Use argument instead of global.
26960 * tree.h (struct tree_optimization_option): New field target_optabs.
26961 * expr.h (init_all_optabs): Add argument to prototype.
26962 (TREE_OPTIMIZATION_OPTABS): New.
26963 (save_optabs_if_changed): Protoize.
26964 * optabs.h: Declare this_fn_optabs.
26965 * optabs.c (save_optabs_if_changed): New.
26966 Declare this_fn_optabs.
26967 (init_optabs): Add argument to init_all_optabs() call.
26968 * function.c (invoke_set_current_function_hook): Handle per
26970 * function.h (struct function): New field optabs.
26971 * config/mips/mips.c (mips_set_mips16_mode): Handle when
26972 optimization_current_node has changed.
26973 * target-globals.h (save_target_globals_default_opts): Protoize.
26974 * target-globals.c (save_target_globals_default_opts): New.
26976 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26979 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
26980 registers %fr12 and %fr12R as call used.
26983 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
26984 and HImode, require all displacements to be an integer multiple of
26986 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
26987 only allow QImode and HImode when reload is in progress and strict is
26988 true. Likewise for symbolic addresses. Use base14_operand to check
26989 displacements in REG+BASE addresses.
26991 2013-02-18 Richard Biener <rguenther@suse.de>
26993 PR tree-optimization/56366
26994 * tree-vect-loop.c (get_initial_def_for_induction): Properly
26995 handle sign-conversion of outer-loop initial induction value.
26997 2013-02-18 Richard Biener <rguenther@suse.de>
26999 PR middle-end/56349
27000 * cfghooks.c (merge_blocks): If we merge a latch into another
27001 block adjust references to it.
27002 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
27003 (verify_loop_structure): Verify that a recorded latch is in fact
27006 2013-02-18 Richard Biener <rguenther@suse.de>
27008 PR tree-optimization/56321
27009 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
27010 order SSA name release and virtual operand unlinking.
27012 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27014 * config/microblaze/microblaze.md (save_stack_block): Define.
27015 (restore_stack_block): Likewise.
27017 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27019 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
27020 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
27021 * config/microblaze/microblaze.c (microblaze_option_override):
27022 Bail out early for PIC modes when target does not support PIC.
27024 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27026 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
27027 Replace with a microblaze version.
27028 (microblaze_trampoline_init): Adapt for microblaze.
27029 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
27032 2013-02-16 Jakub Jelinek <jakub@redhat.com>
27033 Dodji Seketeli <dodji@redhat.com>
27036 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
27037 (instrument_mem_region_access): Do not forget to always put
27038 instrumentation of the of 'base' and 'base + len' in a "if (len !=
27039 0) statement, even for cases where either 'base' or 'base + len'
27040 are not instrumented -- because they have been previously
27041 instrumented. Simplify the logic by putting all the statements
27042 instrument 'base + len' inside a sequence, and then insert that
27043 sequence right before the current insertion point. Then, to
27044 instrument 'base + len', just get an iterator on that statement.
27045 And do not forget to update the pointer to iterator the function
27046 received as argument.
27048 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
27050 PR rtl-optimization/56348
27051 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
27053 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
27055 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
27056 (clean_graph_dump_file): Pass base to start_graph_dump.
27058 2013-02-14 Richard Henderson <rth@redhat.com>
27061 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
27063 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
27065 * collect2-aix.h: Define F_LOADONLY.
27067 2013-02-14 Richard Biener <rguenther@suse.de>
27070 * varasm.c (output_constant_def_1): Get the decl representing
27071 the constant as argument.
27072 (output_constant_def): Wrap output_constant_def_1.
27073 (make_decl_rtl): Use output_constant_def_1 with the decl
27074 representing the constant.
27075 (build_constant_desc): Optionally re-use a decl already
27076 representing the constant.
27077 (tree_output_constant_def): Adjust.
27079 2013-02-14 Dodji Seketeli <dodji@redhat.com>
27082 * asan.c (instrument_builtin_call): Really put the length of the
27083 second source argument into src1_len.
27085 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27087 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
27088 argument. If it is false, don't create edge from then_bb to
27090 (insert_if_then_before_iter): Pass true to it.
27091 (build_check_stmt): Pass false to it.
27092 (transform_statements): Flush hash table only on extended basic
27093 block boundaries, rather than at the beginning of every bb.
27094 Don't flush hash table on nonfreeing_call_p calls.
27095 * tree-flow.h (nonfreeing_call_p): New prototype.
27096 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
27098 2013-02-13 David S. Miller <davem@davemloft.net>
27100 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
27102 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
27105 * ira.c (max_regno_before_ira): Move from ...
27107 (fix_reg_equiv_init): Use max_regno_before_ira instead of
27110 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27112 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
27114 2013-02-13 Richard Biener <rguenther@suse.de>
27117 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
27118 globals in MEM_REFs.
27120 2013-02-13 Richard Biener <rguenther@suse.de>
27122 * loop-init.c (loop_optimizer_init): Clear loop state when
27123 re-initializing preserved loops.
27124 * loop-unswitch.c (unswitch_single_loop): Return whether
27125 we unswitched the loop. Do not verify loop state here.
27126 (unswitch_loops): When we unswitched a loop discover new loops.
27128 2013-02-13 Kostya Serebryany <kcc@google.com>
27130 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
27132 * sanitizer.def: Rename __asan_init to __asan_init_v1.
27134 2013-02-12 Dodji Seketeli <dodji@redhat.com>
27136 Avoid instrumenting duplicated memory access in the same basic block
27137 * Makefile.in (asan.o): Add new dependency on hash-table.h
27138 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
27139 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
27140 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
27141 (free_mem_ref_resources, has_mem_ref_been_instrumented)
27142 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
27143 (get_mem_ref_of_assignment): New functions.
27144 (get_mem_refs_of_builtin_call): Extract from
27145 instrument_builtin_call and tweak a little bit to make it fit with
27147 (instrument_builtin_call): Use the new
27148 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
27149 of is_gimple_builtin_call.
27150 (instrument_derefs, instrument_mem_region_access): Insert the
27151 instrumented memory reference into the hash table.
27152 (maybe_instrument_assignment): Renamed instrument_assignment into
27153 this, and change it to advance the iterator when instrumentation
27154 actually happened and return true in that case. This makes it
27155 homogeneous with maybe_instrument_assignment, and thus give a
27156 chance to callers to be more 'regular'.
27157 (transform_statements): Clear the memory reference hash table
27158 whenever we enter a new BB, when we cross a function call, or when
27159 we are done transforming statements. Use
27160 maybe_instrument_assignment instead of instrumentation. No more
27161 need to special case maybe_instrument_assignment and advance the
27162 iterator after calling it; it's now handled just like
27163 maybe_instrument_call. Update comment.
27165 2013-02-13 Richard Biener <rguenther@suse.de>
27167 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
27168 Fix loop discovery code.
27170 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
27172 PR inline-asm/56148
27173 * lra-constraints.c (process_alt_operands): Match early clobber
27174 operand with itself. Check conflicts with earlyclobber only if
27175 the operand is not reloaded. Prefer to reload conflicting operand
27176 if earlyclobber and matching operands are the same.
27178 2013-02-12 Richard Biener <rguenther@suse.de>
27181 * lto-streamer-out.c (write_symbol): Do not output symbols
27182 for hard register variables.
27184 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
27187 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
27188 (umulsidi3_insn, mulsidi3_insn): New insns.
27190 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
27192 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
27193 (struct tune_params): Add vec_costs field.
27194 * config/arm/arm.c (arm_builtin_vectorization_cost)
27195 (arm_add_stmt_cost): New functions.
27196 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
27197 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
27198 (arm_default_vec_cost): New struct of type cpu_vec_costs.
27199 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
27200 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
27201 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
27202 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
27204 2013-02-12 Richard Biener <rguenther@suse.de>
27207 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
27208 decls again if possible.
27210 2013-02-12 Richard Biener <rguenther@suse.de>
27212 PR middle-end/56288
27213 * tree-ssa.c (verify_ssa_name): Fix check, move
27214 SSA_NAME_IN_FREE_LIST check up.
27216 2013-02-12 Jakub Jelinek <jakub@redhat.com>
27217 Steven Bosscher <steven@gcc.gnu.org>
27219 PR rtl-optimization/56151
27220 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
27221 equal to op0 or op1, and last_insn pattern is CODE operation
27222 with MEM dest and one of the operands matches that MEM.
27224 2013-02-11 Sriraman Tallam <tmsriram@google.com>
27226 * doc/extend.texi: Document Function Multiversioning and "default"
27227 parameter string to target attribute.
27228 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
27229 target attribute parameter is "default".
27230 (ix86_compare_version_priority): Remove checks for target attribute.
27231 (ix86_mangle_function_version_assembler_name): Change error to sorry.
27232 Remove check for target attribute equal to NULL. Add assert.
27233 (ix86_generate_version_dispatcher_body): Change error to sorry.
27235 2013-02-11 Iain Sandoe <iain@codesourcery.com>
27236 Jack Howarth <howarth@bromo.med.uc.edu>
27237 Patrick Marlier <patrick.marlier@gmail.com>
27240 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
27241 define ENDFILE_SPEC as TM_DESTRUCTOR.
27242 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
27244 2013-02-11 Alexander Potapenko <glider@google.com>
27245 Jack Howarth <howarth@bromo.med.uc.edu>
27246 Jakub Jelinek <jakub@redhat.com>
27249 * config/darwin.c (cdtor_record): Rename ctor_record.
27250 (sort_cdtor_records): Rename sort_ctor_records.
27251 (finalize_dtors): New routine to sort destructors by
27252 priority before use in assemble_integer.
27253 (machopic_asm_out_destructor): Use finalize_dtors if needed.
27255 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
27257 PR rtl-optimization/56275
27258 * simplify-rtx.c (avoid_constant_pool_reference): Check that
27259 offset is non-negative and less than cmode size before
27260 calling simplify_subreg.
27262 2013-02-11 Richard Biener <rguenther@suse.de>
27264 PR tree-optimization/56264
27265 * cfgloop.h (fix_loop_structure): Adjust prototype.
27266 * loop-init.c (fix_loop_structure): Return the number of
27267 newly discovered loops.
27268 * tree-cfgcleanup.c (repair_loop_structures): When new loops
27269 are discovered, do a full loop-closed SSA rewrite.
27271 2013-02-11 Richard Biener <rguenther@suse.de>
27273 PR tree-optimization/56273
27274 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27276 (check_array_ref): Fix missing newline in dumps.
27277 (search_for_addr_array): Likewise.
27279 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
27281 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
27283 2013-02-09 Jakub Jelinek <jakub@redhat.com>
27286 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
27288 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
27290 PR rtl-optimization/56246
27291 * lra-constraints.c (simplify_operand_subreg): Try to reuse
27293 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
27294 constraints are satisfied.
27296 2013-02-08 Jeff Law <law@redhat.com>
27299 * emit-rtl.c (reg_is_parm_p): New function.
27300 * regs.h (reg_is_parm_p): New prototype.
27301 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
27302 callee-clobbered registers.
27304 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
27307 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
27308 If there is no implicit builtin declaration, just return NULL.
27310 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27312 * config/i386/sse.md (FMAMODEM): New mode iterator.
27313 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
27314 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
27316 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27318 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
27319 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
27320 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
27322 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27324 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
27325 (microblaze*-*-elf): Likewise.
27326 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
27328 * config/microblaze/microblaze-c.c: Add builtin defines for
27329 _LITTLE_ENDIAN and _BIG_ENDIAN.
27330 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
27331 add to TARGET_DEFAULT flags.
27332 Expand ASM_SPEC and LINK_SPEC.
27333 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
27334 * config/microblaze/microblaze.md: Update extendsidi2 and
27335 movdi_internal instructions to use low-order / high-order reg
27337 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
27338 options and inversemask / mask of LITTLE_ENDIAN.
27339 * config/microblaze/t-microblaze: Expand multilib options to
27340 include mlittle-endian (le) and update exceptions patterns.
27342 2013-02-08 Jakub Jelinek <jakub@redhat.com>
27344 PR rtl-optimization/56195
27345 * lra-constraints.c (get_reload_reg): Don't reuse regs
27346 if they have smaller mode than requested, if they have
27347 wider mode than requested, try to return a SUBREG.
27349 PR tree-optimization/56250
27350 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
27351 if type is unsigned and code isn't MULT_EXPR.
27353 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27355 PR tree-optimization/56064
27356 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
27357 bits according to mode.
27358 * fixed-value.h (fixed_from_double_int)
27359 (const_fixed_from_double_int): Adjust comments.
27361 2013-02-08 Richard Biener <rguenther@suse.de>
27364 * lto-streamer.h (struct data_in): Remove current_file, current_line
27365 and current_col members.
27366 * lto-streamer-out.c (lto_output_location): Stream changed bits
27367 en-block for efficiency.
27368 * lto-streamer-in.c (clear_line_info): Remove.
27369 (lto_input_location): Cache current file, line and column
27370 globally via local statics. Read changed bits en-block.
27371 (input_function): Do not call clear_line_info.
27372 (lto_read_body): Likewise.
27373 (lto_input_toplevel_asms): Likewise.
27375 2013-02-08 Michael Matz <matz@suse.de>
27377 PR tree-optimization/52448
27378 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
27379 (nt_call_phase): New static.
27380 (add_or_mark_expr): Only mark accesses with newer phase than any
27382 (nonfreeing_call_p): New.
27383 (nt_init_block): Update nt_call_phase, mark blocks as visited.
27384 (nt_fini_block): Keep blocks marked as visited.
27385 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
27387 2013-02-08 Richard Biener <rguenther@suse.de>
27389 * ira.c (ira): Free broken dominator information.
27391 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27393 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
27395 2013-02-08 Marek Polacek <polacek@redhat.com>
27397 * cfgloop.c (verify_loop_structure): Add more checking of headers.
27399 2013-02-08 Richard Biener <rguenther@suse.de>
27401 PR middle-end/56181
27402 * cfgloop.h (flow_loops_find): Adjust.
27403 (bb_loop_header_p): Declare.
27404 * cfgloop.c (bb_loop_header_p): New function split out from ...
27405 (flow_loops_find): ... here. Adjust function signature,
27406 support incremental loop structure update.
27407 (verify_loop_structure): Cleanup. Verify a loop is a loop.
27408 * cfgloopmanip.c (fix_loop_structure): Move ...
27409 * loop-init.c (fix_loop_structure): ... here.
27410 (apply_loop_flags): Split out from ...
27411 (loop_optimizer_init): ... here.
27412 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
27413 in incremental mode, only remove dead loops here.
27415 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27418 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
27419 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
27420 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
27421 (*round<mode>3.libgcc): New insns for fixed-modes.
27422 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
27423 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
27424 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
27425 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
27426 implementations. Define to __builtin_avr_absFX,
27427 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
27428 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
27429 __builtin_avr_countlsFX, respectively.
27430 * config/avr/avr-c.c (target.h): Include it.
27431 (enum avr_builtin_id): New enum.
27432 (avr_resolve_overloaded_builtin): New static function.
27433 (avr_register_target_pragmas): Use it to set
27434 targetm.resolve_overloaded_builtin.
27435 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
27436 tree nodes used by DEF_BUILTIN.
27437 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
27438 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
27439 <AVR_BUILTIN_xxBITS>: Same.
27441 2013-02-08 Richard Biener <rguenther@suse.de>
27443 * cfgloop.c (verify_loop_structure): Properly handle
27444 a loop exiting to another loop header.
27445 * ira-int.h (ira_loops): Remove.
27446 * ira.c (ira_loops): Remove.
27447 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
27448 (do_reload): Use loop_optimizer_finalize.
27449 * ira-build.c (create_loop_tree_nodes): Use get_loops and
27450 number_of_loops to access the loop tree.
27451 (more_one_region_p): Likewise.
27452 (finish_loop_tree_nodes): Likewise.
27453 (rebuild_regno_allocno_maps): Likewise.
27454 (mark_loops_for_removal): Likewise.
27455 (mark_all_loops_for_removal): Likewise.
27456 (remove_unnecessary_regions): Likewise.
27457 (ira_build): Likewise.
27458 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
27460 2013-02-08 Richard Biener <rguenther@suse.de>
27462 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
27463 * ipa-pure-const.c (analyze_function): Avoid calling
27464 mark_irreducible_loops twice.
27465 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
27467 2013-02-07 David S. Miller <davem@davemloft.net>
27469 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
27471 * var-tracking.c (vt_add_function_parameter): Test the presence of
27472 HAVE_window_save properly and do not remap argument registers when
27473 we have a leaf function.
27475 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
27478 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
27480 * config/i386/i386.c (ix86_print_operand): Ditto.
27482 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27484 * lra-constraints.c (process_alt_operands): Fix recently added comment.
27486 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27488 PR rtl-optimization/56225
27489 * lra-constraints.c (process_alt_operands): Check that reload hard
27490 reg can hold value for strict_low_part.
27492 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27495 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
27496 dwarf2out_end_function.
27497 (in_first_function_p, maybe_at_text_label_p,
27498 first_loclabel_num_not_at_text_label): New variables.
27499 (dwarf2out_var_location): In the first function find out
27500 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
27501 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
27504 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
27506 PR rtl-optimization/56178
27507 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
27508 SUBREG of a register. Tidy up related block of code.
27509 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
27510 note if the source is a register or a SUBREG of a register.
27512 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27515 * config/rs6000/rs6000.md (ptrm): New mode attr.
27516 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
27517 call_value_indirect_aix<pttrsize>,
27518 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
27521 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
27523 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
27524 if -bnortl. Convert to strcmp and strncmp.
27526 2013-02-07 Alan Modra <amodra@gmail.com>
27529 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
27530 addresses won't wrap when offsetting.
27531 (rs6000_secondary_reload): Provide secondary reloads needed for
27532 wrapping LO_SUM addresses.
27534 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
27536 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
27537 MACH, just __MACH__.
27539 2013-02-06 Richard Biener <rguenther@suse.de>
27541 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
27542 instead of calling fix_loop_structure.
27544 2013-02-06 Jakub Jelinek <jakub@redhat.com>
27546 PR middle-end/56217
27547 * omp-low.c (use_pointer_for_field): Return false if
27548 lower_send_shared_vars doesn't generate any copy-out code.
27550 2013-02-06 Tom de Vries <tom@codesourcery.com>
27552 PR rtl-optimization/56131
27553 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
27554 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
27555 of the label is NULL. Add comment.
27557 2013-02-05 Jakub Jelinek <jakub@redhat.com>
27559 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
27562 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
27563 (STATIC_LIBTSAN_LIBS): Likewise.
27564 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
27565 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
27566 is defined, don't add anything else beyond that.
27567 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
27568 (LINK_COMMAND_SPEC): Use them.
27570 PR tree-optimization/56205
27571 * tree-stdarg.c (check_all_va_list_escapes): Return true if
27572 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
27573 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
27575 2013-02-05 Richard Biener <rguenther@suse.de>
27577 PR tree-optimization/53342
27578 PR tree-optimization/53185
27579 * tree-vectorizer.h (vect_check_strided_load): Remove.
27580 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
27581 not disallow peeling for vectorized strided loads.
27582 (vect_check_strided_load): Make static and simplify.
27583 (vect_analyze_data_refs): Adjust.
27584 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
27585 correctly when vectorizing strided loads.
27587 2013-02-05 Richard Biener <rguenther@suse.de>
27589 * doc/install.texi: Refer to ISL, not PPL.
27591 2013-02-05 Jan Hubicka <jh@suse.cz>
27593 PR tree-optimization/55789
27594 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
27596 2013-02-05 Jan Hubicka <jh@suse.cz>
27598 PR tree-optimization/55789
27599 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
27600 the dead call anyway.
27602 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
27605 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
27607 2013-02-04 Alexander Potapenko <glider@google.com>
27608 Jack Howarth <howarth@bromo.med.uc.edu>
27609 Jakub Jelinek <jakub@redhat.com>
27612 * config/darwin.c (sort_ctor_records): Stabilized qsort
27613 on constructor priority by using original position.
27614 (finalize_ctors): New routine to sort constructors by
27615 priority before use in assemble_integer.
27616 (machopic_asm_out_constructor): Use finalize_ctors if needed.
27618 2013-02-04 Jakub Jelinek <jakub@redhat.com>
27621 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
27622 about visibility on artificial decls.
27623 * config/sol2.c (solaris_assemble_visibility): Likewise.
27625 2013-02-04 Kai Tietz <ktietz@redhat.com>
27628 * config/i386/i386.c (function_value_ms_64): Add additional valtype
27629 argument and improve checking of return-argument types for 16-byte
27631 (ix86_function_value_1): Add additional valtype argument on call
27632 of function_value_64.
27633 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
27634 handling infunction_value_64 function.
27636 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
27638 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
27640 2013-02-04 Richard Biener <rguenther@suse.de>
27642 PR tree-optimization/56188
27643 * tree-ssa-structalias.c (label_visit): Consider case with
27644 initially non-empty points-to set.
27645 (perform_var_substitution): Dump node mapping and clean up.
27647 2013-02-04 Richard Guenther <rguenther@suse.de>
27650 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
27651 node prevail as last resort.
27652 (lto_symtab_merge_decls): Remove guard on LTRANS here.
27653 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
27655 2013-02-04 Richard Biener <rguenther@suse.de>
27657 PR tree-optimization/56113
27658 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
27660 (equiv_class_lookup_or_add): ... this.
27661 (label_visit): Adjust and fix error in previous patch.
27662 (perform_var_substitution): Adjust.
27664 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
27666 * config/sh/divtab.c: Fix formatting and comments throughout the file.
27667 * config/sh/sh4-300.md: Likewise.
27668 * config/sh/sh4a.md: Likewise.
27669 * config/sh/constraints.md: Likewise.
27670 * config/sh/sh.md: Likewise.
27671 * config/sh/netbsd-elf.h: Likewise.
27672 * config/sh/predicates.md: Likewise.
27673 * config/sh/sh-protos.h: Likewise.
27674 * config/sh/ushmedia.h: Likewise.
27675 * config/sh/linux.h: Likewise.
27676 * config/sh/sh.c: Likewise.
27677 * config/sh/superh.h: Likewise.
27678 * config/sh/elf.h: Likewise.
27679 * config/sh/sh4.md: Likewise.
27680 * config/sh/sh.h: Likewise.
27682 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27684 * config/pa/constraints.md: Adjust unused letters. Change "T"
27685 constraint to match_test floating_point_store_memory_operand().
27686 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
27687 (base14_operand): New.
27688 (floating_point_store_memory_operand): New.
27689 (integer_store_memory_operand): Revise to use base14_operand and
27690 reg_plus_base_memory_operand.
27691 (move_dest_operand): Allow symbolic_memory_operands.
27692 (symbolic_memory_operand): Check for LO_SOM.
27693 (symbolic_operand): Change default case to break.
27694 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
27695 CONST_DOUBLE values to be reloaded by putting them into memory when
27696 the destination is a floating point register.
27697 (movdf): Remove code to handle CONST_DOUBLE.
27699 (reload_indf_r1): New.
27700 (reload_insf_r1): New.
27701 Consistently use "Q" and "T" constraints with integer and floating
27702 point move instructions, respectively.
27703 (movdi): Remove FAIL.
27704 Change predicate for source operand unamed DImode move from
27705 general_operand to move_src_operand.
27706 (umulsidi3): Change predicate for destination operand to
27708 Likewise for similar unamed patterns.
27709 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
27710 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
27711 (hppa_legitimize_address): Simplify mask calculation.
27712 (pa_emit_move_sequence): Revised handling of secondary reloads from
27713 REG+D addresses for floating point loads and stores. Directly handle
27714 loading CONST0_RTX (mode) to a floating point register.
27715 (pa_secondary_reload): Handle reloading DF and SFmode constant values
27716 to floating point registers. Don't restrict secondary reloads to
27717 floating point registers to integer modes. Revise some comments and
27719 (TARGET_LEGITIMATE_ADDRESS_P): Define.
27720 (pa_legitimate_address_p): New.
27721 (pa_legitimize_reload_address): New.
27722 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
27723 (STRICT_REG_OK_FOR_BASE_P): New.
27724 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
27725 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
27727 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
27728 Andrew Dixie <andrewd@gentrack.com>
27730 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
27733 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
27735 * expmed.c (extract_bit_field_1): Pass the full width of the
27736 structure to get_best_reg_extraction_insn.
27738 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
27741 * configure.ac (use_cxa_atexit): Add AIX.
27742 * configure: Regenerate.
27744 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
27746 2013-02-01 Jakub Jelinek <jakub@redhat.com>
27749 * final.c (need_profile_function): New variable.
27750 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
27751 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
27752 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
27753 notes, targetm.asm_out.function_prologue doesn't emit anything,
27754 HAVE_prologue and profiler should be emitted before prologue,
27755 set need_profile_function instead of emitting it.
27756 (final_scan_insn): If need_profile_function, emit
27757 profile_function on the first NOTE_INSN_BASIC_BLOCK or
27758 NOTE_INSN_FUNCTION_BEG note.
27760 2013-02-01 Richard Henderson <rth@redhat.com>
27762 * config/rs6000/rs6000.md (smulditi3): New.
27765 * config/alpha/alpha.md (umulditi3): New.
27767 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
27769 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
27770 (ASM_OUTPUT_ALIGNED_LOCAL): New.
27772 2013-02-01 Richard Biener <rguenther@suse.de>
27774 PR tree-optimization/56113
27775 * tree-ssa-structalias.c (label_visit): Reduce work for
27776 single-predecessor nodes.
27778 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
27780 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
27781 range isn't testing for zero.
27783 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
27785 PR middle-end/56113
27786 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
27788 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
27789 Nick Clifton <nickc@redhat.com>
27791 * config/v850/constraints.md (Q): Define as a memory constraint.
27792 * config/v850/predicates.md (label_ref_operand): New predicate.
27793 (e3v5_shift_operand): New predicate.
27794 (ior_operator): New predicate.
27795 * config/v850/t-v850: Add e3v5 multilib.
27796 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
27797 (v850_gen_movdi): Prototype.
27798 * config/v850/v850.c: Add support for e3v5 architecture.
27799 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
27801 (construct_save_jarl): Add e3v5 long JARL support.
27802 (v850_adjust_insn_length): New function. Adjust length of call
27803 insns when using e3v5 instructions.
27804 (v850_gen_movdi): New function: Generate instructions to move a
27806 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
27807 (CPP_SPEC): Define __v850e3v5__ as appropriate.
27808 (TARGET_USE_FPU): Enable for e3v5.
27809 (CONST_OK_FOR_W): New macro.
27810 (ADJUST_INSN_LENGTH): Define.
27811 * config/v850/v850.md (UNSPEC_LOOP): Define.
27812 (attr cpu): Add v850e3v5.
27813 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
27814 (movdi): New pattern.
27815 (movdi_internal): New pattern.
27816 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
27817 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
27818 (cstoresf4): Likewise.
27819 (cstoredf4): Likewise.
27820 (insv): New pattern.
27821 (rotlso3_a): New pattern.
27822 (rotlsi3_b): New pattern
27823 (rotlsi3_v850e3v5): New pattern.
27824 (doloop_begin): New pattern.
27825 (fix_loop_counter): New pattern.
27826 (doloop_end): New pattern.
27827 (branch_normal): Add e3v5 long branch support.
27828 (branch_invert): Likewise.
27829 (branch_z_normal): Likewise.
27830 (branch_z_invert): Likewise.
27831 (branch_nz_normal): Likewise.
27832 (branch_nz_invert): Likewise.
27833 (call_internal_short): Add e3v5 register-indirect JARL support.
27834 (call_internal_long): Likewise.
27835 (call_value_internal_short): Likewise.
27836 (call_value_internal_long): Likewise.
27837 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
27838 (mloop): New option.
27839 * config.gcc: Add support for configuring v840e3v5 target.
27840 * doc/invoke.texi: Document new v850 specific command line options.
27842 2013-01-31 Paul Koning <ni1d@arrl.net>
27846 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
27847 children if parent is a class.
27848 (prune_unused_types_prune): Don't add DW_AT_declaration.
27850 2013-01-31 Richard Biener <rguenther@suse.de>
27852 PR tree-optimization/56157
27853 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
27854 match up operand with SLP child.
27856 2013-01-31 Jason Merrill <jason@redhat.com>
27859 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
27860 parameters the first time.
27861 (gen_scheduled_generic_parms_dies): Check completeness here.
27863 2013-01-31 Richard Biener <rguenther@suse.de>
27865 PR middle-end/53073
27866 * common.opt (faggressive-loop-optimizations): New flag,
27867 enabled by default.
27868 * doc/invoke.texi (faggressive-loop-optimizations): Document.
27869 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
27870 infer_loop_bounds_from_undefined by it.
27872 2013-01-31 Richard Biener <rguenther@suse.de>
27874 PR tree-optimization/56150
27875 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
27876 visit virtual operands.
27877 (find_uses_to_rename_bb): Likewise.
27879 2013-01-31 Richard Biener <rguenther@suse.de>
27881 PR tree-optimization/56150
27882 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
27883 mixed store non-store stmts.
27885 2013-01-30 Jakub Jelinek <jakub@redhat.com>
27888 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
27889 LIBASAN_EARLY_SPEC is defined.
27890 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
27891 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
27893 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
27896 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
27897 invalid args instead of ICEing on it.
27898 (ix86_valid_target_attribute_tree): Return error_mark_node if
27899 ix86_valid_target_attribute_inner_p failed.
27900 (ix86_valid_target_attribute_p): Return false only if
27901 ix86_valid_target_attribute_tree returned error_mark_node. Allow
27902 target("default") attribute.
27903 (sorted_attr_string): Change argument from const char * to tree,
27904 merge in all target attribute arguments rather than just one.
27905 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
27906 instead of free. Avoid using strcat.
27907 (ix86_mangle_function_version_assembler_name): Mangle
27908 target("default") as if no target attribute is present. Adjust
27909 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
27910 instead of xmalloc and XDELETEVEC instead of free.
27911 (ix86_function_versions): Don't return true if one of the decls
27912 doesn't have target attribute. If they don't and one of the decls
27913 is DECL_FUNCTION_VERSIONED, report an error. Adjust
27914 sorted_attr_string caller. Use XDELETEVEC instead of free.
27915 (ix86_supports_function_versions): Remove.
27916 (make_name): Fix up formatting.
27917 (make_dispatcher_decl): Remove resolver_name and its initialization.
27918 Avoid leaking memory.
27919 (is_function_default_version): Return true if there is
27920 target("default") attribute rather than no target attribute at all.
27921 (make_resolver_func): Avoid leaking memory.
27922 (ix86_generate_version_dispatcher_body): Likewise.
27923 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
27924 * target.def (supports_function_versions): Remove.
27925 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
27926 * doc/tm.texi: Regenerated.
27928 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
27930 PR rtl-optimization/56144
27931 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
27932 for values with side effects.
27934 2013-01-30 Richard Biener <rguenther@suse.de>
27936 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
27937 (sparseset_pop): Likewise.
27938 * cfganal.c (compute_idf): Likewise. Increase work-stack size
27939 to be able to use quick_push in the worker loop.
27941 2013-01-30 Marek Polacek <polacek@redhat.com>
27943 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
27945 2013-01-30 Richard Biener <rguenther@suse.de>
27948 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
27950 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
27952 PR tree-optimization/56064
27953 * fixed-value.c (fixed_from_double_int): New function.
27954 * fixed-value.h (fixed_from_double_int): New prototype.
27955 (const_fixed_from_double_int): New static inline function.
27956 * fold-const.c (native_interpret_fixed): New static function.
27957 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
27958 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
27959 (native_encode_fixed): New static function.
27960 (native_encode_expr) <FIXED_CST>: Use it.
27961 (native_interpret_int): Move double_int worker code to...
27962 * double-int.c (double_int::from_buffer): ...this new static method.
27963 * double-int.h (double_int::from_buffer): Prototype it.
27965 2013-01-30 Richard Biener <rguenther@suse.de>
27967 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
27968 New pointer-map and obstack.
27969 (init_alias_vars): Allocate pointer-map and obstack.
27970 (delete_points_to_sets): Free them.
27971 (find_what_var_points_to): Cache result.
27972 (find_what_p_points_to): Adjust for changed interface of
27973 find_what_var_points_to.
27974 (compute_points_to_sets): Likewise.
27975 (ipa_pta_execute): Likewise.
27977 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
27979 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
27980 * configure: Regenerate.
27981 * config.in: Regenerate.
27982 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
27983 #nobits/#progbits if supported.
27985 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
27988 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
27989 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
27990 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
27992 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
27994 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
27995 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
27997 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
27999 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
28000 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
28002 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28004 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
28006 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
28007 * config/arm/cortex-a7.md: New bypasses using
28008 arm_mac_accumulator_is_result.
28010 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28012 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
28013 (cortex_a7_neon_mla): Likewise.
28014 (cortex_a7_fpfmad): New reservation.
28015 (cortex_a7_fpmacs): Use ffmas and update required units.
28016 (cortex_a7_fpmuld): Update required units and latency.
28017 (cortex_a7_fpmacd): Likewise.
28018 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
28019 (cortex_a7_neon). Likewise.
28020 (bypass) Update participating units.
28022 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28024 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
28025 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
28027 * config/arm/vfp11.md (vfp_farith): Use ffmas.
28028 (vfp_fmul): Use ffmad.
28029 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
28030 (cortex_r4_fmacd): Use ffmad.
28031 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
28032 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
28033 (cortex_a9_fmacd): Use ffmad.
28034 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
28035 (cortex_a8_vfp_macd): Use ffmad.
28036 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
28037 (cortex_a5_fpmacd): Use ffmad.
28038 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
28039 (cortex_a15_vfp_macd): Use ffmad.
28040 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
28042 2013-01-29 Jason Merrill <jason@redhat.com>
28045 * varasm.c (default_assemble_visibility): Don't warn about
28046 visibility on artificial decls.
28048 2013-01-29 Richard Biener <rguenther@suse.de>
28050 PR tree-optimization/56113
28051 * tree-ssa-structalias.c (equiv_class_lookup): Also return
28053 (label_visit): Free duplicate bitmaps and record the leader instead.
28054 (perform_var_substitution): Adjust.
28056 2013-01-29 Richard Biener <rguenther@suse.de>
28058 PR tree-optimization/55270
28059 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
28060 the CFG, schedule loops for fixup.
28062 2013-01-29 Nick Clifton <nickc@redhat.com>
28064 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
28067 2013-01-28 Leif Ekblad <leif@rdos.net>
28069 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
28070 * config/i386/i386.h (TARGET_RDOS): New macro.
28071 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
28072 * config/i386/i386.c (ix86_option_override_internal): For 64bit
28073 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
28074 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
28075 DEFAULT_LARGE_SECTION_THRESHOLD.
28076 * config/i386/i386.md (R14_REG, R15_REG): New constants.
28077 * config/i386/rdos.h: New file.
28078 * config/i386/rdos64.h: New file.
28080 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
28083 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
28086 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28088 PR rtl-optimization/56117
28089 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
28090 call cselib_lookup_from_insn on the MEM before calling
28091 add_insn_mem_dependence.
28093 2013-01-28 Richard Biener <rguenther@suse.de>
28095 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
28096 to a stmt that didn't have one.
28097 (copy_phis_for_bb): Likewise for PHI arguments.
28098 (copy_debug_stmt): Likewise for debug stmts.
28100 2013-01-28 Richard Biener <rguenther@suse.de>
28102 PR tree-optimization/56034
28103 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
28104 (partition_builtin_p): Adjust.
28105 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
28106 it is the last partition.
28107 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
28108 up the vertex for the definition.
28109 (classify_partition): Classify whether a partition is a
28110 PKIND_REDUCTION, thus has uses outside of the loop.
28111 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
28112 Merge all PKIND_REDUCTION partitions into the last partition.
28113 (tree_loop_distribution): Seed partitions from reductions as well.
28115 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28117 PR tree-optimization/56125
28118 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
28119 pow(x,c) into sqrt(x) * powi(x, n/2) or
28120 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
28121 optimizing for size.
28122 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
28123 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
28126 PR tree-optimization/56094
28127 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
28128 to UNKNOWN_LOCATION while gimplifying expr.
28130 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
28133 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
28134 operand 0 in movabs insn template for -masm=intel asm alternative.
28135 (*movabs<mode>_2): Ditto for operand 1.
28137 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
28140 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
28143 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28145 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
28148 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
28150 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
28151 Undef to avoid warning.
28153 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28155 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
28156 * configure: Regenerate.
28158 2013-01-25 Jakub Jelinek <jakub@redhat.com>
28160 PR tree-optimization/56098
28161 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
28162 for stmts with volatile ops.
28163 (cond_store_replacement): Don't optimize if assign has volatile ops.
28164 (cond_if_else_store_replacement_1): Don't optimize if either
28165 then_assign or else_assign have volatile ops.
28166 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
28169 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28171 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
28173 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28175 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
28176 missing ':' in asm example.
28178 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
28180 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
28181 entries into lane and laneq entries.
28182 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
28183 Remove AdvSIMD scalar modes.
28184 (aarch64_sq<r>dmulh_laneq<mode>): New.
28185 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
28187 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
28188 builtin implementations to relfect changes in RTL in aarch64-simd.md.
28189 * config/aarch64/iterators.md (VCOND): New.
28192 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28195 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
28196 Add NULL LIBNAME argument to existing definitions.
28197 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
28198 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
28199 * config/avr/avr.c (DEF_BUILTIN): Same.
28200 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
28201 (avr_expand_builtin): Expand to a vanilla call if a libgcc
28202 implementation is available (DECL_ASSEMBLER_NAME is set).
28203 (avr_fold_absfx): New static function.
28204 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
28205 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
28206 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
28207 AVR_BUILTIN_ABSLLK.
28208 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
28209 (abshk, absk, abslk, absllk): Provide as static inline functions.
28211 2013-01-25 Marek Polacek <polacek@redhat.com>
28213 PR tree-optimization/56035
28214 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
28216 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
28218 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
28219 (*movtf_internal_rex64): Add (!o,C) alternative
28220 (*movxf_internal_rex64): Ditto.
28221 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
28223 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
28225 * doc/invoke.texi: fix typo.
28226 * doc/objc.texi: fix typo.
28228 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
28230 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
28231 for the first two alternatives.
28233 2013-01-24 Diego Novillo <dnovillo@google.com>
28235 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
28236 (ggc-zone.o): Remove.
28237 * configure.ac: Remove option --with-gc.
28238 * configure: Re-generate.
28239 * doc/install.texi: Remove documentation for --with-gc.
28240 * gengtype.c (write_enum_defn): Remove. Update all users.
28241 (write_Types_process_field): Remove generation of gt_e_* argument.
28242 (output_type_enum): Remove. Update all users.
28243 (write_enum_defn): Remove. Update all users.
28244 (enum alloc_zone): Remove. Update all users.
28245 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
28246 * ggc-common.c (ggc_splay_alloc): Remove first argument.
28247 Update all callers.
28248 (struct ptr_data): Remove field TYPE. Update all users.
28249 (gt_pch_note_object): Remove argument TYPE. Update all users.
28250 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
28252 * ggc-none.c (ggc_alloc_typed_stat): Remove.
28253 (struct alloc_zone): Remove.
28254 (ggc_internal_alloc_zone_stat): Remove.
28255 (ggc_internal_cleared_alloc_zone_stat): Remove.
28256 * ggc-page.c (ggc_alloc_typed_stat): Remove.
28257 (ggc_pch_count_object): Remove last argument. Update all users.
28258 (ggc_pch_alloc_object): Remove last argument. Update all users.
28259 (struct alloc_zone): Remove.
28260 * ggc-zone.c: Remove.
28261 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
28262 (struct alloc_zone): Remove.
28263 (ggc_alloc_typed_stat): Remove.
28264 (ggc_alloc_typed): Remove.
28265 (ggc_splay_alloc): Remove first argument.
28266 (rtl_zone): Remove. Update all users.
28267 (tree_zone): Remove. Update all users.
28268 (tree_id_zone): Remove. Update all users.
28269 (ggc_internal_zone_alloc_stat): Remove. Update all users.
28270 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
28271 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
28272 * tree-ssanames.c: Remove references to zone allocator in comments.
28274 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
28276 * config/avr/avr.c (avr_out_fract): Make register numbers that
28277 might be outside of source operand signed.
28279 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
28281 * config/i386/constraints.md (Yf): New constraint.
28282 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
28283 of f constraint to conditionaly disable x87 register preferences.
28284 (*movdf_internal): Ditto.
28285 (*movsf_internal): Ditto.
28287 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
28289 PR inline-asm/55934
28290 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
28291 that have operands with impossible constraints.
28292 Add a FIXME for a speed-up opportunity.
28293 * lra-constraints.c (process_alt_operands): Verify that a class
28294 selected from constraints on asms is valid for the operand mode.
28295 (curr_insn_transform): Remove incorrect comment.
28297 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
28299 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
28300 TOC operand is a valid symbol ref in the constant pool.
28302 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28304 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
28306 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
28309 * config/avr/stdfix.h: New file.
28310 * t-avr (stdfix-gcc.h): New rule to build it.
28311 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
28313 2013-01-23 Kostya Serebryany <kcc@google.com>
28315 * config/darwin.h: remove dependency on
28316 CoreFoundation (asan on Mac OS).
28318 2013-01-23 Jakub Jelinek <jakub@redhat.com>
28321 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
28322 instead of cmpdi_operand for first comparison operand.
28323 Don't assert that comparison operands aren't both constants.
28325 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28327 * doc/install.texi (Downloading the Source): Update references to
28328 downloading separate components.
28330 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28332 * doc/extend.texi (__int128): Improve grammar.
28334 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
28337 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
28338 alternative to (o,r).
28339 (*movdi_internal_rex64): Remove (!o,n) alternative.
28340 (DImode immediate->memory splitter): Remove.
28341 (DImode immediate->memory peephole2): Remove.
28342 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
28343 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
28344 alternative to (!o,*r).
28345 (*movtf_internal_sse): New pattern.
28346 (*movxf_internal_rex64): New pattern.
28347 (*movxf_internal): Disable for TARGET_64BIT.
28348 (*movdf_internal_rex64): Remove (!o,F) alternative.
28350 2013-01-22 Jakub Jelinek <jakub@redhat.com>
28352 PR middle-end/56074
28353 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
28354 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
28355 * tree-vect-loop-manip.c (find_loop_location): Also ignore
28356 stmt locations where LOCATION_LOCUS of the stmt location is
28357 UNKNOWN_LOCATION or BUILTINS_LOCATION.
28360 * config/i386/i386.md (UNSPEC_STOS): New.
28361 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
28362 *strsetqi_1): Add UNSPEC_STOS.
28364 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
28367 * doc/invoke.texi: Remove left over -Wsynth example.
28369 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28371 PR tree-optimization/56051
28372 * fold-const.c (fold_binary_loc): Don't fold
28373 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
28374 a narrowing conversion, or widening conversion from signed
28377 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28379 PR rtl-optimization/56023
28380 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
28381 dependent on debug instruction.
28383 2013-01-21 Martin Jambor <mjambor@suse.cz>
28385 PR middle-end/56022
28386 * function.c (allocate_struct_function): Call
28387 invoke_set_current_function_hook earlier.
28389 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28391 * reload1.c (init_reload): Only initialize reload_obstack
28392 during the first call.
28394 2013-01-21 Marek Polacek <polacek@redhat.com>
28396 * cfgloop.c (verify_loop_structure): Fix up grammar.
28398 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
28400 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
28401 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
28403 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28406 * config/arm/marvell-pj4.md: Update copyright year.
28407 Fix up use of alu to alu_reg and simple_alu_imm.
28409 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28411 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
28413 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
28416 * lra-constraints.c (curr_insn_transform): Don't reuse original
28417 insn for secondary memory move when memory mode should be different.
28419 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28421 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
28422 atomic_storedi_1): New patterns.
28424 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
28426 btver2 pipeline descriptions.
28427 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
28429 * config/i386/i386.md (btver2_decode): New type attributes.
28430 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
28432 * config/i386/btver2.md: New file describing btver2 pipelines.
28434 2013-01-19 Andrew Pinski <apinski@cavium.com>
28436 PR tree-optimization/52631
28437 * tree-ssa-sccvn (visit_use): Before looking up the original
28438 statement, try looking up the simplified expression.
28440 2013-01-19 Anthony Green <green@moxielogic.com>
28442 * config/moxie/moxie.c (moxie_expand_prologue): Set
28443 current_function_static_stack_size.
28445 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28447 PR tree-optimization/56029
28448 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
28449 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
28451 2013-01-18 Sharad Singhai <singhai@google.com>
28453 PR tree-optimization/55995
28454 * dumpfile.c (dump_loc): Print location only if available.
28455 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
28457 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
28460 * lra-constraints.c (curr_insn_transform): Reuse original insn for
28461 secondary memory move.
28462 (inherit_reload_reg): Use rclass instead of cl for
28463 check_secondary_memory_needed_p.
28465 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28467 PR middle-end/56015
28468 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
28469 the case where writing real complex part of target modifies op1.
28471 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
28473 * config/aarch64/aarch64-simd.md
28474 (aarch64_vcond_internal<mode>): Handle unordered cases.
28475 * config/aarch64/iterators.md (v_cmp_result): New.
28477 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
28478 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28480 * config/arm/marvell-pj4.md: New file.
28481 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
28482 * config/arm/arm.md (generic_sched): Add marvell_pj4.
28483 (generic_vfp): Likewise.
28484 * config/arm/arm-cores.def: Add marvell-pj4.
28485 * config/arm/arm-tune.md: Regenerate.
28486 * config/arm/arm-tables.opt: Regenerate.
28487 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
28488 * doc/invoke.texi: Document marvell-pj4.
28490 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
28492 * config/aarch64/arm_neon.h: Map scalar types to standard types.
28494 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28499 * var-tracking.c (negative_power_of_two_p): New.
28500 (global_get_addr_cache, local_get_addr_cache): New.
28501 (get_addr_from_global_cache, get_addr_from_local_cache): New.
28502 (vt_canonicalize_addr): Rewrite using the above. Adjust the
28504 (vt_stack_offset_p): Remove.
28505 (vt_canon_true_dep): Always canonicalize loc's address.
28506 (clobber_overlapping_mems): Make sure we have a MEM.
28507 (local_get_addr_clear_given_value): New.
28508 (val_reset): Clear local cached entries.
28509 (compute_bb_dataflow): Create and release the local cache.
28510 Disable duplicate MEMs clobbering.
28511 (emit_notes_in_bb): Clobber MEMs likewise.
28512 (vt_emit_notes): Create and release the local cache.
28513 (vt_initialize, vt_finalize): Create and release the global
28514 cache, respectively.
28515 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
28517 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28519 PR libmudflap/53359
28520 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
28521 not found in the symtab.
28523 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28526 PR rtl-optimization/55547
28527 PR rtl-optimization/53827
28530 * alias.c (offset_overlap_p): New, factored out of...
28531 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
28532 the conservative special case for symbolic constants. Don't
28533 adjust zero sizes on alignment.
28535 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
28537 PR rtl-optimization/52573
28538 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
28539 REG_UNUSED for the same register.
28541 2013-01-17 Richard Biener <rguenther@suse.de>
28542 Marek Polacek <polacek@redhat.com>
28544 PR rtl-optimization/55833
28545 * loop-unswitch.c (unswitch_loops): Move loop verification...
28546 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
28547 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
28548 Set it to true when we're removing a loop from hierarchy tree in
28549 an irreducible region.
28550 (fix_bb_placements): Adjust caller.
28551 (fix_loop_placements): Likewise.
28553 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
28555 * config/avr/builtins.def (DEF_BUILTIN): Factor out
28556 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
28557 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
28558 Remove ID. Adjust comments.
28559 * config/avr/avr-c.c (avr_builtin_name): Remove.
28560 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
28561 * config/avr/avr.c (avr_tolower): New static function.
28562 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
28563 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
28564 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
28567 2013-01-17 Jan Hubicka <jh@suse.cz>
28569 PR tree-optimization/55273
28570 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
28572 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
28575 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
28576 store through atomic_store<mode>_1.
28577 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
28579 2013-01-17 Martin Jambor <mjambor@suse.cz>
28581 PR tree-optimizations/55264
28582 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
28583 for virtual methods.
28584 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
28585 virtual methods before inlining is over.
28586 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
28588 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
28591 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
28593 PR rtl-optimization/56005
28594 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
28595 pending reads for prefetch.
28597 2013-01-16 Ian Bolton <ian.bolton@arm.com>
28599 * config/aarch64/aarch64.md
28600 (*cstoresi_neg_uxtw): New pattern.
28601 (*cmovsi_insn_uxtw): New pattern.
28602 (*<optab>si3_uxtw): New pattern.
28603 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
28604 (*<optab>si3_insn_uxtw): New pattern.
28605 (*bswapsi2_uxtw): New pattern.
28607 2013-01-16 Richard Biener <rguenther@suse.de>
28609 * tree-inline.c (tree_function_versioning): Remove set but
28610 never used variable.
28612 2013-01-16 Richard Biener <rguenther@suse.de>
28614 PR tree-optimization/55964
28615 * tree-flow.h (rename_variables_in_loop): Remove.
28616 (rename_variables_in_bb): Likewise.
28617 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
28618 (copy_loop_before): Adjust and delete update-ssa status.
28619 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
28620 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
28621 (rename_variables_in_loop): Remove.
28622 (slpeel_update_phis_for_duplicate_loop): Likewise.
28623 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
28624 use available cfg machinery instead of duplicating it.
28625 Update PHI nodes and perform poor-mans SSA update here.
28626 (slpeel_tree_peel_loop_to_edge): Adjust.
28628 2013-01-16 Richard Biener <rguenther@suse.de>
28630 PR tree-optimization/54767
28631 PR tree-optimization/53465
28632 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
28633 (vrp_visit_phi_node): For PHI arguments coming via backedges
28634 drop all symbolical range information.
28635 (execute_vrp): Compute backedges.
28637 2013-01-16 Richard Biener <rguenther@suse.de>
28639 * doc/install.texi: Update CLooG and ISL requirements to
28642 2013-01-16 Christian Bruel <christian.bruel@st.com>
28645 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
28646 (broken_move): Handle UNSPECV_SP_SWITCH_B.
28647 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
28649 2013-01-16 DJ Delorie <dj@redhat.com>
28651 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
28652 (UNSPECV_SP_SWITCH_E): New.
28653 (sp_switch_1): Change to an unspec.
28654 (sp_switch_2): Change to an unspec. Don't use post-inc when we
28657 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
28659 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
28660 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
28661 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
28662 (expand_mem_thread_fence): Ditto.
28663 (expand_mem_signal_fence): Ditto.
28664 (expand_atomic_load): Ditto.
28665 (expand_atomic_store): Ditto.
28667 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
28669 PR rtl-optimization/55547
28670 PR rtl-optimization/53827
28673 * alias.c (memrefs_conflict_p): Set sizes to negative after
28676 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28679 * function.c (thread_prologue_and_epilogue_insns): Always
28680 add crtl->drap_reg to set_up_by_prologue.set, even if
28681 stack_realign_drap is false.
28683 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
28685 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
28686 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
28687 *call): Fix indention.
28689 2013-01-15 Tom de Vries <tom@codesourcery.com>
28692 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
28695 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
28697 PR rtl-optimization/55153
28698 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
28700 2013-01-15 Martin Jambor <mjambor@suse.cz>
28702 PR tree-optimization/55920
28703 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
28704 accesses as grp_to_be_debug_replaced.
28706 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28708 PR tree-optimization/55920
28709 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
28710 there is non-useless type conversion needed from debug rhs to lhs,
28711 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
28713 2013-01-15 Joseph Myers <joseph@codesourcery.com>
28714 Mikael Pettersson <mikpe@it.uu.se>
28717 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
28719 (ASM_OUTPUT_CASE_LABEL): Remove.
28720 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
28721 * final.c (shorten_branches): Update alignment of labels before
28722 jump tables if CASE_VECTOR_SHORTEN_MODE.
28724 2013-01-15 Richard Biener <rguenther@suse.de>
28727 * system.h: Do not include gmp.h for building host tools.
28729 2013-01-15 Richard Biener <rguenther@suse.de>
28731 PR middle-end/55882
28732 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
28733 account for bitpos when computing alignment.
28735 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
28737 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
28738 (ix86_target_macros_internal): Likewise.
28740 * config/i386/i386.c (m_CORE2I7): Removed.
28741 (m_CORE_HASWELL): New macro.
28742 (m_CORE_ALL): Likewise.
28743 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
28744 (initial_ix86_arch_features): Likewise.
28745 (processor_target_table): Initializations for Core avx2.
28746 (cpu_names): New names "core-avx2".
28747 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
28748 PROCESSOR_CORE_HASWELL.
28749 (ix86_issue_rate): New case.
28750 (ia32_multipass_dfa_lookahead): Likewise.
28751 (ix86_sched_init_global): Likewise.
28753 * config/i386/i386.h (TARGET_HASWELL): New macro.
28754 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
28755 (processor_type): New PROCESSOR_HASWELL.
28757 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28759 PR tree-optimization/55955
28760 * tree-vect-loop.c (vectorizable_reduction): Give up early on
28761 *SHIFT_EXPR and *ROTATE_EXPR codes.
28763 PR tree-optimization/48766
28764 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
28765 -ftrapv disable -fwrapv.
28767 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28770 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
28771 etc. to 1 and not to __flash.
28772 Use LL suffix for __INT24_MAX__ with -mint8.
28773 Use ULL suffix for __UINT24_MAX__ with -mint8.
28775 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28777 * config/avr/avr-arch.h
28778 (struct base_arch_s): Use typedef avr_arch_t instead.
28779 (struct arch_info_s): Use typedef avr_arch_info_t instead.
28780 (struct mcu_type_s): Use typedef avr_mcu_t instead.
28781 * config/avr/avr.c: Same.
28782 * config/avr/avr-devices.c: Same.
28783 * config/avr/driver-avr.c: Same.
28784 * config/avr/gen-avr-mmcu-texi.c: Same.
28785 * config/avr/avr-mcus.def: Adjust comment.
28787 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
28789 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
28790 * config/aarch64/iterators.md (VALLDI): New.
28792 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
28793 Andi Kleen <ak@linux.intel.com>
28796 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
28797 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
28800 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28802 * config/avr/avr-stdint.h: Remove trailing blanks.
28803 * config/avr/avr-log.h: Same.
28804 * config/avr/avr-arch.h: Same.
28805 * config/avr/avr-devices.c: Same.
28806 * config/avr/avr-dimode.md: Same.
28807 * config/avr/predicates.md: Same.
28808 * config/avr/avr-c.c: Same. And fix typo.
28810 * config/avr/avr-protos.h: Same. And:
28811 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
28812 (init_cumulative_args): Rename to avr_init_cumulative_args.
28813 (expand_prologue): Rename to avr_expand_prologue.
28814 (expand_epilogue): Rename to avr_expand_epilogue.
28815 (adjust_insn_length): Rename to avr_adjust_insn_length.
28816 (notice_update_cc): Rename to avr_notice_update_cc.
28817 (final_prescan_insn): Rename to avr_final_prescan_insn.
28818 * config/avr/avr.c: Same.
28819 * config/avr/avr.h: Same.
28820 * config/avr/avr.md: Remove trailing blanks.
28821 (prologue): Use avr_expand_prologue.
28822 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
28824 2013-01-14 Richard Biener <rguenther@suse.de>
28826 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
28827 verify_location, collect_subblocks): New functions.
28828 (verify_gimple_in_cfg): Verify that locations only reference
28829 BLOCKs in the functions BLOCK tree.
28831 2013-01-14 Richard Biener <rguenther@suse.de>
28833 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
28835 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
28837 (insert_out_of_ssa_copy_on_edge): Likewise.
28838 (rewrite_close_phi_out_of_ssa): Likewise.
28839 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
28841 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
28842 propagated constants.
28843 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
28846 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28848 * config/avr/avr-modes.def: Add GPL copyright notice.
28850 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
28852 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
28853 MEMMODEL_MASK to determine memory model.
28854 (atomic_store<mode>): Ditto from operands[2].
28855 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
28857 2013-01-13 Jakub Jelinek <jakub@redhat.com>
28860 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
28861 (fold_gimple_assign): Don't call unshare_expr here.
28862 (fold_ctor_reference): Call unshare_expr.
28864 2013-01-13 Terry Guo <terry.guo@arm.com>
28866 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
28867 * doc/fragments.texi: Document MULTILIB_REUSE.
28868 * gcc.c (multilib_reuse): New internal spec.
28869 (set_multilib_dir): Also search multilib from multilib_reuse.
28870 * genmultilib (tmpmultilib3): Refactor code.
28871 (tmpmultilib4): Ditto.
28872 (multilib_reuse): New multilib argument.
28874 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
28876 * Makefile.in: Update copyright.
28878 2013-01-12 Tom de Vries <tom@codesourcery.com>
28880 PR middle-end/55890
28881 * calls.c (expand_call): Check if arg_nr is valid.
28883 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
28885 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
28886 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
28887 documentation. Add missing '__' in front of
28888 __builtin_ia32_packssdw256.
28890 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
28893 * config/s390/s390.c (s390_preferred_reload_class): Do not return
28894 NO_REGS for larl operands.
28895 (s390_reload_larl_operand): Use s390_load_address instead of
28898 2013-01-11 Richard Biener <rguenther@suse.de>
28900 * tree-cfg.c (verify_node_sharing_1): Split out from ...
28901 (verify_node_sharing): ... here.
28902 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
28904 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
28906 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
28907 Substitute TREECHECKING.
28908 * configure: Regenerate.
28909 * Makefile.in (TREECHECKING): New.
28911 2013-01-11 Richard Guenther <rguenther@suse.de>
28913 PR tree-optimization/44061
28914 * tree-vrp.c (extract_range_basic): Compute zero as
28915 value-range for __builtin_constant_p of function parameters.
28917 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
28919 Update copyright years.
28921 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
28923 PR rtl-optimization/55672
28924 * lra-eliminations.c (mark_not_eliminable): Permit addition with
28925 const to be eliminable.
28927 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
28929 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
28930 * configure: Regenerate.
28932 2013-01-10 Richard Biener <rguenther@suse.de>
28934 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
28936 2013-01-10 Richard Biener <rguenther@suse.de>
28939 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
28940 locations for virtual PHI arguments.
28941 (rewrite_update_phi_arguments): Likewise.
28943 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
28945 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
28948 2013-01-10 Jakub Jelinek <jakub@redhat.com>
28950 PR tree-optimization/55921
28951 * tree-complex.c (expand_complex_asm): New function.
28952 (expand_complex_operations_1): Call it for GIMPLE_ASM.
28954 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
28957 * config/s390/s390.c (s390_symref_operand_p)
28958 (s390_loadrelative_operand_p): Merge the two functions.
28959 (s390_check_qrst_address, print_operand_address): Add parameters
28960 to s390_loadrelative_operand_p invokation.
28961 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
28962 (s390_reload_larl_operand, s390_secondary_reload): Use
28963 s390_loadrelative_operand_p instead of s390_symref_operand_p.
28964 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
28966 2013-01-09 Mike Stump <mikestump@comcast.net>
28968 * dse.c (record_store): Remove unnecessary assert.
28970 2013-01-09 Jan Hubicka <jh@suse.cz>
28972 PR tree-optimization/55569
28973 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
28974 * cfgloop.h (scale_loop_profile): Likewise.
28976 2013-01-09 Jan Hubicka <jh@suse.cz>
28979 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
28981 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
28983 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
28985 PR middle-end/55114
28986 * expr.h (maybe_emit_group_store): Declare.
28987 * expr.c (maybe_emit_group_store): New function.
28988 * builtins.c (expand_builtin_int_roundingfn): Call it.
28989 (expand_builtin_int_roundingfn_2): Likewise.
28991 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
28993 PR rtl-optimization/55829
28994 * lra-constraints.c (match_reload): Add code for absent output.
28995 (curr_insn_transform): Add code for reloads of matched inputs
28998 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
29000 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
29001 attribute of movddup insn to DF.
29002 (*vec_interleave_lowv2df): Ditto.
29003 (vec_dupv2df): Ditto.
29005 2013-01-09 Jan Hubicka <jh@suse.cz>
29007 PR tree-optimiation/55875
29008 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
29009 EVERY_ITERATION parameter.
29010 (number_of_iterations_exit): Check if exit is executed every iteration.
29011 (idx_infer_loop_bounds): Similarly here.
29012 (n_of_executions_at_most): Simplify
29013 to only test for cases where statement is dominated by the
29014 particular bound; handle correctly the "postdominance" test.
29015 (scev_probably_wraps_p): Use max loop iterations info
29016 as a global bound first.
29018 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
29019 Nick Clifton <nickc@redhat.com>
29021 * config/v850/v850.md (cbranchsf4): New pattern.
29022 (cstoresf4): New pattern.
29023 (cbranchdf4): New pattern.
29024 (cstoredf4): New pattern.
29025 (movsicc): Disallow floating point comparisons.
29026 (cmpsf_le_insn): Fix order of operators.
29027 (cmpsf_lt_insn): Likewise.
29028 (cmpsf_eq_insn): Likewise.
29029 (cmpdf_le_insn): Likewise.
29030 (cmpdf_lt_insn): Likewise.
29031 (cmpdf_eq_insn): Likewise.
29032 (cmpsf_ge_insn): Use LE comparison.
29033 (cmpdf_ge_insn): Likewise.
29034 (cmpsf_gt_insn): Use LT comparison.
29035 (cmpdf_gt_insn): Likewise.
29036 (cmpsf_ne_insn): Delete pattern.
29037 (cmpdf_ne_insn): Delete pattern.
29038 * config/v850/v850.c (v850_gen_float_compare): Use
29039 gen_cmpdf_eq_insn for NE comparison.
29040 (v850_float_z_comparison_operator)
29041 (v850_float_nz_comparison_operator): Move from here ...
29042 * config/v850/predicates.md: ... to here. Move GT and GE
29043 comparisons into v850_float_z_comparison_operator.
29044 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
29046 (v850_float_nz_comparison_operator): Likewise.
29048 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29050 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
29051 with calls to gen_insvsi/gen_insvdi.
29053 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29055 * config/i386/i386.c (initial_ix86_tune_features): Set up
29056 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
29058 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
29059 Jakub Jelinek <jakub@redhat.com>
29061 PR tree-optimization/48189
29062 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
29063 If nitercst is 0, don't predict the exit edge.
29065 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29067 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
29068 in asm_fprintf with reg_names.
29069 (aarch64_print_operand_address): Likewise.
29070 (aarch64_return_addr): Likewise.
29071 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
29073 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29075 * config/pa/pa.h (VAL_U6_BITS_P): Define.
29076 (INT_U6_BITS): Likewise.
29077 * config/pa/predicates.md (uint6_operand): New predicate.
29078 (shift5_operand, shift6_operand): Likewise.
29079 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
29081 (lshrdi3): Use shift6_operand.
29082 (shrpsi4, shrpdi4): New insn patterns.
29083 (extzv): Delete expander.
29084 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
29085 predicates in unamed zero extract patterns. Tighten common constraint.
29086 (extv): Delete expander.
29087 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
29088 predicates in unamed sign extract patterns. Tighten common constraint.
29089 (insv): Delete expander.
29090 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
29091 predicates in unamed insert patterns. Tighten common constraint.
29092 Change uint32_operand predicate to uint6_operand predicate in unamed
29093 DImode pattern to insert constant values of type 1...1xxxx.
29095 2013-01-04 Jan Hubicka <jh@suse.cz>
29097 PR tree-optimization/55823
29098 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
29101 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29102 Uros Bizjak <ubizjak@gmail.com>
29104 PR rtl-optimization/55845
29105 * df-problems.c (can_move_insns_across): Stop scanning at
29106 volatile_insn_p source instruction or give up if
29107 across_from .. across_to range contains any volatile_insn_p
29110 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29112 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
29113 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
29115 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
29116 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
29118 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29121 * asan.c (asan_clear_shadow): New function.
29122 (asan_emit_stack_protection): Use it.
29124 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29126 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
29127 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
29128 with tab instead of space.
29130 2013-01-08 Nick Clifton <nickc@redhat.com>
29132 * config/rl78/rl78.c (rl78_expand_prologue): Always select
29133 register bank 0 at the start of an interrupt handler.
29134 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
29137 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29139 * config/aarch64/aarch64-simd.md
29140 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
29141 (aarch64_simd_bsl): Likewise.
29142 (aarch64_vcond_internal<mode>): Likewise.
29143 (vcond<mode><mode>): Likewise.
29144 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
29145 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
29147 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29149 * config/aarch64/aarch64-builtins.c
29150 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
29152 2013-01-08 Martin Jambor <mjambor@suse.cz>
29155 * tree-sra.c (analyze_access_subtree): Return true also after
29156 potentially creating a debug-only replacement.
29158 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29160 PR middle-end/55890
29161 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
29163 PR tree-optimization/54120
29164 * tree-vrp.c (range_fits_type_p): Don't allow
29165 src_precision < precision from signed vr to unsigned_p
29166 if vr->min or vr->max is negative.
29167 (simplify_float_conversion_using_ranges): Test can_float_p
29168 against CODE_FOR_nothing.
29170 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29171 Richard Biener <rguenther@suse.de>
29173 PR middle-end/55851
29174 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
29175 types instead of just INTEGER_TYPE types.
29177 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
29179 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
29182 2013-01-07 Steve Ellcey <sellcey@mips.com>
29185 * config/mips/mips.opt: Change mad to mmad to match documentation.
29187 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29190 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
29191 .progmemx.data now.
29193 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29196 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
29197 (avr_addrspace_t): Add .section_name field.
29198 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
29200 (avr_addrspace): Same. Initialize .section_name. Remove last
29201 NULL entry. Put __memx into .progmemx.data.
29202 (progmem_section_prefix): Remove.
29203 (avr_asm_init_sections): No need to initialize progmem_section.
29204 (avr_asm_named_section): Use avr_addrspace[].section_name to get
29205 section name prefix.
29206 (avr_asm_select_section): Ditto. And use get_unnamed_section to
29207 retrieve the progmem section.
29208 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
29209 boundary to run over avr_addrspace[].
29210 (avr_register_target_pragmas): Ditto.
29212 2013-01-06 Jakub Jelinek <jakub@redhat.com>
29214 * varasm.c (output_constant_def_contents): For asan_protect_global
29215 protected strings, adjust DECL_ALIGN if needed, before testing for
29217 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
29218 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
29220 (output_object_block): For asan protected decls, emit asan padding
29221 after their contents.
29222 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
29223 (asan_finish_file): Test it here instead.
29225 2013-01-07 Nick Clifton <nickc@redhat.com>
29226 Matthias Klose <doko@debian.org>
29227 Doug Kwan <dougkwan@google.com>
29228 H.J. Lu <hongjiu.lu@intel.com>
29231 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
29233 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
29235 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
29237 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
29239 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
29241 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29244 * doc/install.texi (Cross-Compiler-Specific Options): Document
29247 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
29249 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
29250 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
29251 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
29252 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
29253 vqmovun_high_s64): Fix source operand number and update copyright.
29255 2013-01-07 Richard Biener <rguenther@suse.de>
29257 PR middle-end/55890
29258 * gimple.h (gimple_call_builtin_p): New overload.
29259 * gimple.c (validate_call): New function.
29260 (gimple_call_builtin_p): Likewise.
29261 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
29262 Use gimple_call_builtin_p.
29263 (find_func_clobbers): Likewise.
29264 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
29265 (strlen_optimize_stmt): Likewise.
29267 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29269 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
29270 (vld1q_dup_*): Likewise.
29271 (vld1_*): Likewise.
29272 (vld1q_*): Likewise.
29273 (vld1_lane_*): Likewise.
29274 (vld1q_lane_*): Likewise.
29276 2013-01-07 Richard Biener <rguenther@suse.de>
29278 * lto-streamer.h (LTO_minor_version): Bump to 2.
29280 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29282 * config/aarch64/aarch64-protos.h
29283 (aarch64_const_double_zero_rtx_p): Rename to...
29284 (aarch64_float_const_zero_rtx_p): ...this.
29285 (aarch64_float_const_representable_p): New.
29286 (aarch64_output_simd_mov_immediate): Likewise.
29287 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
29288 move immediate case.
29289 * config/aarch64/aarch64.c
29290 (aarch64_const_double_zero_rtx_p): Rename to...
29291 (aarch64_float_const_zero_rtx_p): ...this.
29292 (aarch64_print_operand): Allow printing of new constants.
29293 (aarch64_valid_floating_const): New.
29294 (aarch64_legitimate_constant_p): Check for valid floating-point
29296 (aarch64_simd_valid_immediate): Likewise.
29297 (aarch64_vect_float_const_representable_p): New.
29298 (aarch64_float_const_representable_p): Likewise.
29299 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
29300 (aarch64_output_simd_mov_immediate): New.
29301 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
29302 (*movdf_aarch64): Likewise.
29303 * config/aarch64/constraints.md (Ufc): New.
29304 (Y): call aarch64_float_const_zero_rtx.
29305 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
29307 2013-01-07 Richard Biener <rguenther@suse.de>
29309 PR tree-optimization/55888
29310 PR tree-optimization/55862
29311 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
29312 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
29313 not if it is contained therein.
29315 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29317 * config/avr/t-avr: Typo.
29319 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29322 * config/avr/t-avr: Don't automatically rebuild
29323 $(srcdir)/config/avr/t-multilib
29324 $(srcdir)/config/avr/avr-tables.opt
29325 $(srcdir)/doc/avr-mmcu.texi
29326 (avr-mcus): New phony target to build them on request.
29327 (s-avr-mlib, s-avr-mmcu-texi): Remove.
29328 * avr/avr-mcus.def: Adjust comments.
29330 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
29332 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
29334 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29336 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
29338 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29340 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
29342 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
29344 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
29345 to generate profiling.
29346 * config/rs6000/aix64.h (LIB_SPEC): Same.
29348 2013-01-04 Andrew Pinski <apinski@cavium.com>
29350 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
29352 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
29354 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
29356 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
29358 (ix86_expand_move): Ditto.
29359 (ix86_zero_extend_to_Pmode): Ditto.
29360 (ix86_expand_call): Ditto.
29361 (ix86_expand_special_args_builtin): Ditto.
29362 (ix86_expand_builtin): Ditto.
29364 2013-01-04 Richard Biener <rguenther@suse.de>
29366 PR tree-optimization/55862
29367 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
29368 translating them through PHI nodes.
29370 2013-01-04 Martin Jambor <mjambor@suse.cz>
29372 PR tree-optimization/55755
29373 * tree-sra.c (sra_modify_assign): Do not check that an access has no
29374 children when trying to avoid producing a VIEW_CONVERT_EXPR.
29376 2013-01-04 Marek Polacek <polacek@redhat.com>
29378 PR middle-end/55859
29379 * opts.c (default_options_optimization): Clarify error message.
29381 2013-01-04 Richard Biener <rguenther@suse.de>
29383 PR middle-end/55863
29384 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
29387 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29390 * config/pa/pa.md (movsi): Revert previous change.
29391 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
29394 2013-01-03 Richard Henderson <rth@redhat.com>
29396 * config/i386/i386.c (ix86_expand_move): Always assign to op1
29397 after eliminating TLS symbols.
29399 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
29402 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
29403 * graphite-poly.c (debug_gmp_value): Likewise.
29405 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
29408 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
29409 selected code model, define __code_mode_small__, __code_model_medium__,
29410 __code_model_large__, __code_model_32__ or __code_model_kernel__.
29411 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
29412 xchg temporary register with %k. Declare temporary register as
29414 [__x86_64__]: For medium and large code models, preserve %rbx register.
29416 2013-01-03 Richard Biener <rguenther@suse.de>
29418 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
29419 (dump_subscript): Adjust.
29420 (finalize_ddr_dependent): Do not dump redundant info.
29421 (analyze_siv_subscript): Adjust.
29422 (subscript_dependence_tester): Likewise.
29423 (compute_affine_dependence): Likewise.
29425 2013-01-03 Richard Biener <rguenther@suse.de>
29428 2013-01-03 Richard Biener <rguenther@suse.de>
29430 PR tree-optimization/55857
29431 * tree-vect-stmts.c (vectorizable_load): Do not setup
29432 re-alignment for invariant loads.
29434 2013-01-02 Richard Biener <rguenther@suse.de>
29436 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29437 invariant load do not generate a vector load from the scalar location.
29439 2013-01-03 Richard Biener <rguenther@suse.de>
29441 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
29442 for not vectorizing.
29443 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
29444 not build INDIRECT_REFs, call get_name once only.
29445 (vect_create_data_ref_ptr): Likewise. Dump base object kind
29446 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
29448 2013-01-03 Richard Biener <rguenther@suse.de>
29450 PR tree-optimization/55857
29451 * tree-vect-stmts.c (vectorizable_load): Do not setup
29452 re-alignment for invariant loads.
29454 2013-01-03 Richard Biener <rguenther@suse.de>
29457 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
29458 prefer a built-in decl.
29460 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29462 * gcc.c (process_command): Update copyright notice dates.
29463 * gcov.c (print_version): Likewise.
29464 * gcov-dump.c (print_version): Likewise.
29466 PR rtl-optimization/55838
29467 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
29468 iv0.step, iv1.step and step.
29470 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29471 Marc Glisse <marc.glisse@inria.fr>
29473 PR tree-optimization/55832
29474 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
29475 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
29476 integer_{one,zero}_node.
29478 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29481 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
29482 * var-tracking.c (reverse_op): Don't add reverse ops to
29483 VALUEs that have already
29484 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
29486 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
29488 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
29490 2013-01-02 Teresa Johnson <tejohnson@google.com>
29492 * dumpfile.c (dump_loc): Print filename with location.
29493 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
29494 new location_t parameter to emit complete unroll message with
29495 new dump framework.
29496 (canonicalize_loop_induction_variables): Compute loops location
29497 and pass to try_unroll_loop_completely.
29498 * loop-unroll.c (report_unroll_peel): New function.
29499 (peel_loops_completely): Use new dump format with location
29500 for main dumpfile message, and invoke report_unroll_peel on success.
29501 (decide_unrolling_and_peeling): Ditto.
29502 (decide_peel_once_rolling): Remove old dumpfile message subsumed
29503 by report_unroll_peel.
29504 (decide_peel_completely): Ditto.
29505 (decide_unroll_constant_iterations): Ditto.
29506 (decide_unroll_runtime_iterations): Ditto.
29507 (decide_peel_simple): Ditto.
29508 (decide_unroll_stupid): Ditto.
29509 * cfgloop.c (get_loop_location): New function.
29510 * cfgloop.h (get_loop_location): Declare.
29512 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29514 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
29517 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29519 PR middle-end/55198
29520 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
29521 BLKmode objects when EXPAND_MEMORY is specified.
29523 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29525 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
29527 (fold_builtin_cpu): Do not share cpu model decls across statements.
29529 2013-01-02 Jason Merrill <jason@redhat.com>
29532 * tree.c (build_array_type_1): Revert earlier change.
29534 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
29536 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
29538 * config/aarch64/aarch64-tune.md: Re-generate.
29540 2013-01-02 Richard Biener <rguenther@suse.de>
29542 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29543 invariant load do not generate a vector load from the scalar location.
29545 2013-01-02 Richard Biener <rguenther@suse.de>
29548 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
29549 * configure: Regenerate.
29551 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29553 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
29554 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
29555 (expand_builtin_int_roundingfn_2): Keep the original target around
29556 for the fallback case.
29558 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29560 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
29561 to be clear for sign changes.
29563 2013-01-01 Jan Hubicka <jh@suse.cz>
29565 * ipa-inline-analysis.c: Fix formatting.
29567 2013-01-01 Jakub Jelinek <jakub@redhat.com>
29569 PR tree-optimization/55831
29570 * tree-vect-loop.c (get_initial_def_for_induction): Use
29571 gsi_after_labels instead of gsi_start_bb.
29573 Copyright (C) 2013 Free Software Foundation, Inc.
29575 Copying and distribution of this file, with or without modification,
29576 are permitted in any medium without royalty provided the copyright
29577 notice and this notice are preserved.