1 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
4 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
5 immediately after a label for a tablejump pattern.
7 * config/arm/arm.c (is_jump_table): Remove.
8 (create_fix_barrier): Use tablejump_p instead.
10 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
11 (thumb2_output_casesi): Likewise.
12 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
13 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
14 casesi_shift_media, casesi_load_media): Likewise.
15 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
16 * config/microblaze/microblaze.md: Likewise.
18 2013-11-05 Tobias Burnus <burnus@net-b.de>
20 * doc/invoke.texi (-Wdate-time): Document.
22 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
24 * double-int.c (lshift_double, rshift_double): Remove
25 SHIFT_COUNT_TRUNCATED handling.
27 2013-11-05 Jeff Law <law@redhat.com>
29 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
30 * common.opt (-fisolate-erroneous-paths): Add option and
32 * gimple-ssa-isolate-paths.c: New file.
33 * gimple.c (check_loadstore): New function.
34 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
35 Verify OP is in the argument list and the argument corresponding
36 to OP is a pointer type. Use operand_equal_p rather than
37 pointer equality when testing if OP is on the nonnull list.
38 Use check_loadstore rather than count_ptr_derefs. Handle
39 GIMPLE_RETURN statements.
40 * tree-vrp.c (infer_nonnull_range): Remove.
41 * gimple.h (infer_nonnull_range): Declare.
42 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
43 * passes.def: Add pass_isolate_erroneous_paths.
44 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
45 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
46 * tree-ssa.c (struct count_ptr_d): Remove.
47 (count_ptr_derefs, count_uses_and_derefs): Remove.
48 * tree-ssa.h (count_uses_and_derefs): Remove.
50 2013-11-05 Jakub Jelinek <jakub@redhat.com>
52 PR rtl-optimization/58997
53 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
54 get_iv_value to be in iv->mode rather than iv->extend_mode.
55 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
56 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
57 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
60 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
62 * gimple.h: Move some prototypes to gimple-expr.h and add to include
64 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
65 Move to gimple-expr.h.
66 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
67 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
68 is_gimple_id, virtual_operand_p, is_gimple_addressable,
69 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
71 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
72 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
73 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
74 is_gimple_address, is_gimple_invariant_address,
75 is_gimple_ip_invariant_address, is_gimple_min_invariant,
76 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
77 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
78 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
79 gimple_decl_printable_name, useless_type_conversion_p,
80 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
82 * gimple-expr.c: New File.
83 (useless_type_conversion_p, gimple_set_body, gimple_body,
84 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
85 gimple_can_coalesce_p, extract_ops_from_tree_1,
86 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
87 is_gimple_address, is_gimple_invariant_address,
88 is_gimple_ip_invariant_address, is_gimple_min_invariant,
89 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
90 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
91 is_gimple_mem_ref_addr): Relocate here.
92 * Makefile.in (OBJS): Add gimple-expr.o.
94 2013-11-05 David Malcolm <dmalcolm@redhat.com>
96 * gengtype-parse.c (struct_field_seq): Support empty structs.
98 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
100 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
102 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
104 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
106 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
108 (LONG_DOUBLE_TYPE_SIZE): New define.
109 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
111 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
114 * doc/extend.texi [visibility ("visibility_type")]: Add example
115 about visibility attribute on namespace declaration.
117 2013-11-05 Richard Biener <rguenther@suse.de>
120 * passes.def (all_passes): Start with pass_fixup_cfg again.
122 2013-11-05 Richard Biener <rguenther@suse.de>
124 PR tree-optimization/58955
125 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
128 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
130 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
132 (vec_pack_ufix_trunc_v2df): Likewise.
134 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
137 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
138 pattern, instead of word_mode.
140 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
142 (set_storage_via_setmem): Likewise.
144 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
146 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
147 gimple.h and the rest of the condition in eliminate_build.
148 (eliminate_build): Call new routine.
149 * gimple.h (phi_ssa_name_p): Delete.
151 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
153 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
154 case of no prefix and reserving zero slots, because when that's the
155 case we'll never get here.
156 * vec.h (va_heap::reserve): Don't try and handle
157 vec_prefix::calculate_allocation returning zero because that should
160 2013-11-05 Richard Biener <rguenther@suse.de>
163 * tree-dfa.c (get_ref_base_and_extent): Merge common code
164 in MEM_REF and TARGET_MEM_REF handling. Make sure to
165 process trailing array detection before diving into the
166 view-converted object (and possibly apply some extra offset).
168 2013-11-05 Joseph Myers <joseph@codesourcery.com>
170 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
172 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
174 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
176 PR tree-optimization/58958
177 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
178 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
180 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
182 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
183 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
185 2013-11-05 Jakub Jelinek <jakub@redhat.com>
187 PR tree-optimization/58984
188 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
189 set *SIZE_P if non-NULL on success.
190 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
192 (ipcp_transform_function): Likewise. Punt if size of access
193 is different from TYPE_SIZE on v->value's type.
195 2013-11-05 Tobias Burnus <burnus@net-b.de>
197 * doc/invoke.texi (-fopenmp-simd): Document new option.
198 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
199 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
200 * tree.c (attribute_value_equal): Ditto.
202 2013-11-04 Wei Mi <wmi@google.com>
204 * sched-rgn.c (add_branch_dependences): Keep insns in
205 a SCHED_GROUP at the end of BB to remain their location.
207 2013-11-04 Wei Mi <wmi@google.com>
209 * gcc/config/i386/i386.c (memory_address_length): Extract a part
210 of code to rip_relative_addr_p.
211 (rip_relative_addr_p): New Function.
212 (ix86_macro_fusion_p): Ditto.
213 (ix86_macro_fusion_pair_p): Ditto.
214 * gcc/config/i386/i386.h: Add new tune features about macro-fusion.
215 * gcc/config/i386/x86-tune.def (DEF_TUNE): Ditto.
216 * gcc/doc/tm.texi: Generated.
217 * gcc/doc/tm.texi.in: Ditto.
218 * gcc/haifa-sched.c (try_group_insn): New Function.
219 (group_insns_for_macro_fusion): Ditto.
220 (sched_init): Call group_insns_for_macro_fusion.
221 * gcc/target.def: Add two hooks: macro_fusion_p and
224 2013-11-04 Kostya Serebryany <kcc@google.com>
226 Update to match the changed asan API.
227 * asan.c (asan_function_start): New function.
228 (asan_emit_stack_protection): Update the string stored in the
229 stack red zone to match new API. Store the PC of the current
230 function in the red zone.
231 (asan_global_struct): Update the __asan_global definition to match
233 (asan_add_global): Ditto.
234 * asan.h (asan_function_start): New prototype.
235 * final.c (final_start_function): Call asan_function_start.
236 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
239 2013-11-04 Wei Mi <wmi@google.com>
241 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): Separate
242 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
243 * gcc/config/i386/i386.c (ix86_option_override_internal): Ditto.
244 (ix86_issue_rate): Ditto.
245 (ix86_adjust_cost): Ditto.
246 (ia32_multipass_dfa_lookahead): Ditto.
247 (ix86_sched_init_global): Ditto.
248 (get_builtin_code_for_version): Ditto.
249 * gcc/config/i386/i386.h (enum target_cpu_default): Ditto.
250 (enum processor_type): Ditto.
251 * gcc/config/i386/x86-tune.def (DEF_TUNE): Ditto.
253 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
255 PR rtl-optimization/58967
256 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
257 !lra_in_progress for mode sizes bigger word.
259 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
261 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
262 arguments to merge instruction for little endian.
263 (vec_widen_umult_lo_v16qi): Likewise.
264 (vec_widen_smult_hi_v16qi): Likewise.
265 (vec_widen_smult_lo_v16qi): Likewise.
266 (vec_widen_umult_hi_v8hi): Likewise.
267 (vec_widen_umult_lo_v8hi): Likewise.
268 (vec_widen_smult_hi_v8hi): Likewise.
269 (vec_widen_smult_lo_v8hi): Likewise.
271 2013-11-04 Ian Lance Taylor <iant@google.com>
273 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
274 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
275 builtins that take pointers.
276 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
277 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
278 (run_gcc): Pass -fnon-call-exceptions.
280 2013-11-04 Jakub Jelinek <jakub@redhat.com>
282 * optabs.c (expand_vec_perm): Revert one incorrect line from
285 PR tree-optimization/58978
286 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
287 use_stmt by single_imm_use directly. Only call single_imm_use
290 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
292 PR rtl-optimization/58968
293 * lra-spills.c (return_regno_p): New function.
294 (lra_final_code_change): Use it.
296 2013-11-04 Joseph Myers <joseph@codesourcery.com>
298 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
299 * target.def (float_exceptions_rounding_supported_p): New hook.
300 * targhooks.c (default_float_exceptions_rounding_supported_p): New
302 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
303 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
305 * doc/tm.texi: Regenerate.
306 * config.gcc (powerpc*-*-linux*): Set extra_objs.
307 * config/rs6000/rs6000-linux.c: New file.
308 * config/rs6000/rs6000-protos.h
309 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
310 * config/rs6000/linux.h
311 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
312 * config/rs6000/linux64.h
313 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
314 * config/rs6000/t-linux (rs6000-linux.o): New rule.
315 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
317 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
319 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
320 Replace the define_insn_and_split with a define_insn and two
321 define_splits, with the split after reload re-permuting the source
322 register to its original value.
323 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
324 (*vsx_le_perm_store_v8hi): Likewise.
325 (*vsx_le_perm_store_v16qi): Likewise.
327 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
329 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
332 2013-11-04 Jakub Jelinek <jakub@redhat.com>
334 PR tree-optimization/58946
335 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
336 bbs with bbinfo[idx].op != NULL before all blocks with
337 bbinfo[idx].op == NULL.
339 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
341 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
342 (avr_dump_double_int_hex): Likewise.
343 (avr_log_vadump): Remove %D and %X handling.
344 * config/avr/avr.c (avr_double_int_push_digit): Delete.
345 (avr_map_op_t): Change map from double_int to unsigned int.
346 (avr_map_op): Update accordingly.
347 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
348 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
349 unsigned ints rather than double_ints.
351 2013-11-03 Marek Polacek <polacek@redhat.com>
353 Implement -fsanitize=vla-bound.
354 * opts.c (common_handle_option): Handle vla-bound.
355 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
356 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
357 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
359 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
361 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
364 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
366 * config/i386/constraints.md (Ts, Tv): New address constrains.
367 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
368 constraint for address_no_seg_operand.
369 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
370 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
371 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
372 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
373 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
374 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
375 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
376 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
377 constraint for vsib_address_operand.
379 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
381 * gcse.c (pre_delete): Remove references to regmove from comments.
382 * recog.c: (validate_replace_rtx_1): Likewise.
383 * config/rl78/rl78.c: Likewise.
384 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
385 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
386 * common/config/mmix/mmix-common.c: Likewise.
388 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
390 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
391 * gimplify.c (gimplify_compound_lval): Likewise.
392 * graphite-clast-to-gimple.c (gloog): Likewise.
393 * graphite-dependences.c (loop_is_parallel_p): Likewise.
394 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
395 (limit_scop); Likewise.
396 (build_scops): Likewise.
397 (dot_scop): Likewise.
398 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
399 (build_scop_drs): Likewise.
400 (insert_stmts): Likewise.
401 (insert_out_of_ssa_copy): Likewise.
402 (remove_phi): Likewise.
403 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
404 * hw-doloop.c (discover_loop): Likewise.
405 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
406 * tree-dfa.c (dump_enumerated_decls): Likewise.
407 * tree-if-conv.c (if_convertable_loop_p): Likewise.
408 * tree-inline.c (tree_function_versioning): Likewise.
409 * tree-loop-distribution.c (build_rdg): Likewise.
410 (rdg_flag_vertex_and_dependent): Likewise.
411 (distribute_loop): Likewise.
412 * tree-parloops.c (loop_parallel_p): Likewise.
413 (eliminate_local_variables): Likewise.
414 (separate_decls_in_region): Likewise.
415 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
416 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
417 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
418 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
419 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
420 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
421 (vectorizable_condition): Likewise.
423 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
425 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
426 * configure: Regenerate.
427 * config/i386/i386.md (*movdi_internal): Change
428 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
429 (*movdf_internal): Ditto.
430 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
431 * config/i386/sse.md (vec_concatv2di): Output interunit movq
432 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
434 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
436 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
439 2013-10-31 David Malcolm <dmalcolm@redhat.com>
441 Automated part of renaming of symtab_node_base to symtab_node.
443 Patch autogenerated by rename_symtab.py from
444 https://github.com/davidmalcolm/gcc-refactoring-scripts
445 revision 58bb219cc090b2f4516a9297d868c245495ee622
446 with ChangeLog entry fixed up by hand.
448 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
450 (cgraph_node_for_asm): Likewise.
451 * cgraph.h (symtab_node_base): Likewise.
452 (cgraph_node): Likewise.
453 (varpool_node): Likewise.
454 (is_a_helper <cgraph_node>::test): Likewise.
455 (is_a_helper <varpool_node>::test): Likewise.
456 (symtab_nodes): Likewise.
457 (symtab_register_node): Likewise.
458 (symtab_unregister_node): Likewise.
459 (symtab_remove_node): Likewise.
460 (symtab_get_node): Likewise.
461 (symtab_node_for_asm): Likewise.
462 (symtab_node_asm_name): Likewise.
463 (symtab_node_name): Likewise.
464 (symtab_insert_node_to_hashtable): Likewise.
465 (symtab_add_to_same_comdat_group): Likewise.
466 (symtab_dissolve_same_comdat_group_list): Likewise.
467 (dump_symtab_node): Likewise.
468 (debug_symtab_node): Likewise.
469 (dump_symtab_base): Likewise.
470 (verify_symtab_node): Likewise.
471 (verify_symtab_base): Likewise.
472 (symtab_used_from_object_file_p): Likewise.
473 (symtab_alias_ultimate_target): Likewise.
474 (symtab_resolve_alias): Likewise.
475 (fixup_same_cpp_alias_visibility): Likewise.
476 (symtab_for_node_and_aliases): Likewise.
477 (symtab_nonoverwritable_alias): Likewise.
478 (availability symtab_node_availability): Likewise.
479 (symtab_semantically_equivalent_p): Likewise.
480 (fixup_same_cpp_alias_visibility): Likewise.
481 (symtab_prevail_in_asm_name_hash): Likewise.
484 (varpool_first_variable): Likewise.
485 (varpool_next_variable): Likewise.
486 (varpool_first_static_initializer): Likewise.
487 (varpool_next_static_initializer): Likewise.
488 (varpool_first_defined_variable): Likewise.
489 (varpool_next_defined_variable): Likewise.
490 (cgraph_first_defined_function): Likewise.
491 (cgraph_next_defined_function): Likewise.
492 (cgraph_first_function): Likewise.
493 (cgraph_next_function): Likewise.
494 (cgraph_first_function_with_gimple_body): Likewise.
495 (cgraph_next_function_with_gimple_body): Likewise.
496 (symtab_alias_target): Likewise.
497 (symtab_real_symbol_p): Likewise.
498 (symtab_can_be_discarded): Likewise.
499 * cgraphbuild.c (mark_address): Likewise.
500 (mark_load): Likewise.
501 (mark_store): Likewise.
502 * cgraphunit.c (decide_is_symbol_needed): Likewise.
504 (enqueue_node): Likewise.
505 (referred_to_p): Likewise.
506 (cgraph_process_same_body_aliases): Likewise.
507 (analyze_functions): Likewise.
508 (handle_alias_pairs): Likewise.
509 (output_weakrefs): Likewise.
511 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
512 * ipa-inline-analysis.c (inline_write_summary): Likewise.
513 * ipa-prop.c (remove_described_reference): Likewise.
514 (try_decrement_rdesc_refcount): Likewise.
515 (ipa_edge_duplication_hook): Likewise.
516 * ipa-ref.c (ipa_record_reference): Likewise.
517 (ipa_maybe_record_reference): Likewise.
518 (ipa_clone_ref): Likewise.
519 (ipa_clone_references): Likewise.
520 (ipa_clone_referring): Likewise.
521 (ipa_find_reference): Likewise.
522 (ipa_remove_stmt_references): Likewise.
523 (ipa_clear_stmts_in_references): Likewise.
524 * ipa-ref.h (symtab_node_base): Likewise.
526 (ipa_record_reference): Likewise.
527 (ipa_maybe_record_reference): Likewise.
528 (ipa_clone_references): Likewise.
529 (ipa_clone_referring): Likewise.
530 (ipa_clone_ref): Likewise.
531 (ipa_find_reference): Likewise.
532 (ipa_remove_stmt_references): Likewise.
533 (ipa_clear_stmts_in_references): Likewise.
534 * ipa-reference.c (ipa_reference_write_optimization_summary):
536 * ipa.c (enqueue_node): Likewise.
537 (process_references): Likewise.
538 (walk_polymorphic_call_targets): Likewise.
539 (symtab_remove_unreachable_nodes): Likewise.
540 (address_taken_from_non_vtable_p): Likewise.
541 (comdat_can_be_unshared_p_1): Likewise.
542 (comdat_can_be_unshared_p): Likewise.
543 (can_replace_by_local_alias): Likewise.
544 (function_and_variable_visibility): Likewise.
545 * is-a.h: Likewise (within example in comment).
546 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
547 (lto_symtab_encoder_encode): Likewise.
548 (lto_symtab_encoder_delete_node): Likewise.
549 (lto_symtab_encoder_in_partition_p): Likewise.
550 (lto_set_symtab_encoder_in_partition): Likewise.
551 (output_refs): Likewise.
552 (compute_ltrans_boundary): Likewise.
553 (output_symtab): Likewise.
554 (input_node): Likewise.
555 (input_ref): Likewise.
556 (input_edge): Likewise.
557 (input_cgraph_1): Likewise.
558 (input_refs): Likewise.
559 (output_cgraph_opt_summary): Likewise.
560 (input_node_opt_summary): Likewise.
561 (input_cgraph_opt_section): Likewise.
562 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
564 * lto-streamer-out.c (lto_output): Likewise.
565 (output_symbol_p): Likewise.
566 (produce_symtab): Likewise.
567 * lto-streamer.h (lto_encoder_entry): Likewise.
568 (lto_free_function_in_decl_state_for_node): Likewise.
569 (lto_symtab_encoder_encode): Likewise.
570 (lto_symtab_encoder_delete_node): Likewise.
571 (lto_symtab_encoder_in_partition_p): Likewise.
572 (lto_set_symtab_encoder_in_partition): Likewise.
573 (lto_symtab_encoder_lookup): Likewise.
574 (lsei_node): Likewise.
575 (lto_symtab_encoder_deref): Likewise.
576 * symtab.c (symtab_hash): Likewise.
577 (assembler_name_hash): Likewise.
578 (symtab_nodes): Likewise.
579 (hash_node): Likewise.
581 (hash_node_by_assembler_name): Likewise.
582 (eq_assembler_name): Likewise.
583 (insert_to_assembler_name_hash): Likewise.
584 (unlink_from_assembler_name_hash): Likewise.
585 (symtab_prevail_in_asm_name_hash): Likewise.
586 (symtab_register_node): Likewise.
587 (symtab_insert_node_to_hashtable): Likewise.
588 (symtab_unregister_node): Likewise.
589 (symtab_get_node): Likewise.
590 (symtab_remove_node): Likewise.
591 (symtab_initialize_asm_name_hash): Likewise.
592 (symtab_node_for_asm): Likewise.
593 (symtab_add_to_same_comdat_group): Likewise.
594 (symtab_dissolve_same_comdat_group_list): Likewise.
595 (symtab_node_asm_name): Likewise.
596 (symtab_node_name): Likewise.
597 (dump_symtab_base): Likewise.
598 (dump_symtab_node): Likewise.
599 (dump_symtab): Likewise.
600 (debug_symtab_node): Likewise.
601 (verify_symtab_base): Likewise.
602 (verify_symtab_node): Likewise.
603 (verify_symtab): Likewise.
604 (symtab_used_from_object_file_p): Likewise.
605 (symtab_node_availability): Likewise.
606 (symtab_alias_ultimate_target): Likewise.
607 (fixup_same_cpp_alias_visibility): Likewise.
608 (symtab_resolve_alias): Likewise.
609 (symtab_for_node_and_aliases): Likewise.
610 (symtab_for_node_and_aliases): Likewise.
611 (symtab_nonoverwritable_alias_1): Likewise.
612 (symtab_nonoverwritable_alias): Likewise.
613 (symtab_semantically_equivalent_p): Likewise.
614 * value-prof.c (init_node_map): Likewise.
615 * varasm.c (find_decl): Likewise.
616 * varpool.c (varpool_node_for_asm): Likewise.
617 (varpool_remove_unreferenced_decls): Likewise.
619 2013-10-31 David Malcolm <dmalcolm@redhat.com>
621 Manual part of renaming of symtab_node_base to symtab_node.
623 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
624 clashes with the preferred name for the base class.
625 (const_symtab_node): Remove redundant typedef.
627 2013-10-31 Jakub Jelinek <jakub@redhat.com>
629 * optabs.c (expand_vec_perm): Avoid vector mode punning
631 * expmed.c (store_bit_field_1): Likewise.
632 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
633 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
634 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
635 * config/i386/i386.c (ix86_expand_vector_move_misalign,
636 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
637 ix86_expand_sse_unpack, ix86_expand_args_builtin,
638 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
639 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
640 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
641 expand_vec_perm_vpshufb2_vpermq,
642 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
643 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
644 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
645 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
648 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
650 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
651 Use gcc_unreachable for unhandled modes. Do not check results of
652 expand_simple_binop. If not expanded to target, move the result.
654 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
655 Shiva Chen <shiva0217@gmail.com>
657 * config.gcc (nds32*-*-*): Add nds32 target.
658 * config/nds32/nds32.c: New file.
659 * config/nds32/nds32.h: New file.
660 * config/nds32/nds32.md: New file.
661 * config/nds32/constants.md: New file.
662 * config/nds32/constraints.md: New file.
663 * config/nds32/iterators.md: New file.
664 * config/nds32/nds32-doubleword.md: New file.
665 * config/nds32/nds32-intrinsic.md: New file.
666 * config/nds32/nds32_intrinsic.h: New file.
667 * config/nds32/nds32-modes.def: New file.
668 * config/nds32/nds32-multiple.md: New file.
669 * config/nds32/nds32.opt: New file.
670 * config/nds32/nds32-opts.h: New file.
671 * config/nds32/nds32-protos.h: New file.
672 * config/nds32/nds32-peephole2.md: New file.
673 * config/nds32/pipelines.md: New file.
674 * config/nds32/predicates.md: New file.
675 * config/nds32/t-mlibs: New file.
676 * common/config/nds32: New directory and files.
678 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
679 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
680 * doc/install.texi (Cross-Compiler-Specific Options): Document
681 --with-nds32-lib for nds32 target.
682 * doc/extend.texi (Function Attributes, Target Builtins): Document
683 nds32 specific attributes.
685 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
687 * lra-constraints (process_alt_operands): Use the result
688 elimination register for operand when matching constraints.
690 2013-10-31 Jakub Jelinek <jakub@redhat.com>
692 * tree-vrp.c (maybe_set_nonzero_bits): New function.
693 (remove_range_assertions): Call it.
695 * tree.c (tree_ctz): New function.
696 * tree.h (tree_ctz): New prototype.
697 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
698 first argument from tree to const_tree.
699 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
700 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
701 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
703 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
704 * tree-vect-loop.c (vect_analyze_loop_operations,
705 vect_transform_loop): Don't force scalar loop for bound just because
706 number of iterations is unknown, only do it if it is not known to be
707 a multiple of vectorization_factor.
708 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
710 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
711 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
712 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
713 (set_nonzero_bits, get_nonzero_bits): New prototypes.
714 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
715 a default def isn't partially constant.
716 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
717 is known to be partially zero.
718 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
719 to see if a default def isn't partially constant.
720 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
721 creation of a range, if VR_RANGE, try to improve nonzero_bits from
723 (set_nonzero_bits, get_nonzero_bits): New functions.
725 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
726 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
727 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
728 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
729 uses but in the condition and ASSERT_EXPR and the other successor of
730 the predecessor bb is __builtin_unreachable (), set_range_info of the
731 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
733 2013-10-31 Martin Jambor <mjambor@suse.cz>
735 PR rtl-optimization/58934
737 2013-10-30 Martin Jambor <mjambor@suse.cz>
738 PR rtl-optimization/10474
739 * ira.c (find_moveable_pseudos): Do not calculate dominance info
741 (interesting_dest_for_shprep): New function.
742 (split_live_ranges_for_shrink_wrap): Likewise.
743 (ira): Calculate dominance info and df analysis. Call
744 split_live_ranges_for_shrink_wrap.
746 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
747 Yury Gribov <y.gribov@samsung.com>
750 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
751 shadow pointer to avoid clobbering the main one.
753 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
755 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
757 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
760 * ira-color.c (update_costs_from_copies): Add new parameter. Use
761 it for calling update_costs_from_allocno.
762 (assign_hard_reg): Call restore_costs_from_copies only for
763 !retry_p. Pass new argument to update_costs_from_copies.
764 (color_pass): Pass new argument to update_costs_from_copies.
765 (ira_mark_allocation_change): Ditto.
767 2013-10-30 Sharad Singhai <singhai@google.com>
770 * opts.c (common_handle_option): Remove deprecated option
771 -ftree-vectorizer-verbose.
772 * doc/invoke.texi (Debugging Options): Ditto.
773 * opts-global.c (handle_common_deferred_options): Ditto.
774 (dump_remap_tree_vectorizer_verbose): Delete.
775 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
777 2013-10-30 DJ Delorie <dj@redhat.com>
779 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
780 that take no arguments.
782 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
785 * reload1.c (update_eliminables_and_spill): New function, broken
787 (reload): Use it. Check for frame size change after frame size
788 alignment, and call update_eliminables_and_spill first if continue-ing.
790 2013-10-30 Cong Hou <congh@google.com>
793 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
794 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
795 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
797 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
799 PR rtl-optimization/58369
800 * reload1.c (compute_reload_subreg_offset): New function.
801 (choose_reload_regs): Use it to pass endian-correct
802 offset to subreg_regno_offset.
804 2013-10-30 Tobias Burnus <burnus@net-b.de>
807 * gcc/tree-cfg.c (replace_loop_annotate): Replace warning by
810 2013-10-30 Jason Merrill <jason@redhat.com>
812 * configure.ac (loose_warn): Add -Wno-format if
813 --disable-build-format-warnings.
815 2013-10-30 David Malcolm <dmalcolm@redhat.com>
817 * cgraphunit.c (analyze_functions): Split symtab_node declarations
818 onto multiple lines to make things easier for rename_symtab.py.
820 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
821 (symtab_semantically_equivalent_p): Likewise.
823 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
826 * lra.c (check_rtl): Remove address check before LRA work.
828 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
830 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
831 POINTER_PLUS_EXPR in the defining statement.
833 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
836 * tree-pass.h (make_pass_regmove): Remove.
837 * timevar.def (TV_REGMOVE): Remove.
838 * passes.def (pass_regmove): Remove.
839 * opts.c (default_options_table): Remove entry for regmove.
840 * doc/passes.texi: Remove regmove pass description.
841 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
843 (-fdump-rtl-regmove): Ditto.
844 * common.opt (foptimize-register-move, fregmove): Ignore.
845 * Makefile.in (OBJS): Remove regmove.o.
847 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
849 (struct ira_allocno) New member allocno_prefs.
850 (ALLOCNO_PREFS): New macro.
851 (ira_prefs, ira_prefs_num): New external vars.
852 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
853 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
854 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
856 (ira_add_allocno_copy_to_list): Remove prototype.
857 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
858 (ira_pref_iterator): New type.
859 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
860 (FOR_EACH_PREF): New macro.
861 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
862 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
864 (ira_setup_alts): New function.
865 (decrease_live_ranges_number): New function.
866 (ira): Call the above function.
867 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
868 (ira_create_allocno): Initialize allocno prefs.
869 (pref_pool, pref_vec): New static vars.
870 (initiate_prefs, find_allocno_pref, ira_create_pref): New
872 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
873 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
874 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
875 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
876 (ira_add_allocno_copy_to_list): Make static. Rename to
877 add_allocno_copy_to_list.
878 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
879 swap_allocno_copy_ends_if_necessary.
880 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
881 ira_remove_allocno_prefs.
882 (ira_flattening): Ditto.
883 (ira_build): Call initiate_prefs, print_prefs.
884 (ira_destroy): Call finish_prefs.
885 * ira-color.c (struct update_cost_record): New.
886 (struct allocno_color_data): Add new member update_cost_records.
887 (update_cost_record_pool): New static var.
888 (init_update_cost_records, get_update_cost_record): New functions.
889 (free_update_cost_record_list, finish_update_cost_records): Ditto.
890 (struct update_cost_queue_elem): Add member from.
891 (initiate_cost_update): Call init_update_cost_records.
892 (finish_cost_update): Call finish_update_cost_records.
893 (queue_update_cost, get_next_update_cost): Add new param from.
894 (Update_allocno_cost, update_costs_from_allocno): New functions.
895 (update_costs_from_prefs): Ditto.
896 (update_copy_costs): Rename to update_costs_from_copies.
897 (restore_costs_from_copies): New function.
898 (update_conflict_hard_regno_costs): Don't go back.
899 (assign_hard_reg): Call restore_costs_from_copies. Add printing
901 (pop_allocnos): Add priniting more debug info.
902 (color_allocnos): Remove prefs for conflicting hard regs.
903 Call update_costs_from_prefs.
904 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
905 (get_dup_num): Rename, modify, and move to ira.c
906 (process_regs_for_copy): Add prefs.
907 (add_insn_allocno_copies): Put src as first arg of
908 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
909 * ira-costs.c (record_reg_classes): Modify and move code into
910 record_operands_costs.
911 (find_costs_and_classes): Create prefs for the hard reg of small
913 (process_bb_node_for_hard_reg_moves): Add prefs.
915 2013-10-30 Richard Biener <rguenther@suse.de>
918 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
919 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
921 (pre_and_rev_post_order_compute): ... which now wraps it.
922 * graph.c (draw_cfg_nodes_no_loops): Use
923 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
926 2013-10-30 Christian Bruel <christian.bruel@st.com>
928 * gcc/config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
929 (sh_expand_cmpstr): Handle known align and schedule improvements.
930 * gcc/config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
931 * gcc/config/sh/sh.md (cmpstrnsi): New pattern.
933 2013-10-30 Martin Jambor <mjambor@suse.cz>
935 PR rtl-optimization/10474
936 * ira.c (find_moveable_pseudos): Do not calculate dominance info
938 (interesting_dest_for_shprep): New function.
939 (split_live_ranges_for_shrink_wrap): Likewise.
940 (ira): Calculate dominance info and df analysis. Call
941 split_live_ranges_for_shrink_wrap.
943 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
946 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
948 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
950 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
951 * tree.h (POINTER_BOUNDS_P): New.
952 (BOUNDED_TYPE_P): New.
954 (pointer_bounds_type_node): New.
955 * tree.c (build_common_tree_nodes): Initialize
956 pointer_bounds_type_node.
957 * gimple.h (gimple_call_get_nobnd_arg_index): New.
958 (gimple_call_num_nobnd_args): New.
959 (gimple_call_nobnd_arg): New.
960 (gimple_return_retbnd): New.
961 (gimple_return_set_retbnd): New
962 * gimple.c (gimple_build_return): Increase number of ops
963 for return statement.
964 (gimple_call_get_nobnd_arg_index): New.
965 * gimple-pretty-print.c (dump_gimple_return): Print second op.
967 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
969 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
970 with "chkp ctor" and "bnd_legacy" attributes.
971 * gimplify.c (gimplify_init_constructor): Avoid infinite
972 loop during gimplification of bounds initializer.
974 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
976 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
977 (handle_bnd_legacy): New.
978 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
979 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
982 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
984 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
985 (BT_FN_PTR_CONST_PTR): New.
986 (BT_FN_CONST_PTR_CONST_PTR): New.
987 (BT_FN_PTR_CONST_PTR_SIZE): New.
988 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
989 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
990 (BT_FN_VOID_CONST_PTR_SIZE): New.
991 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
992 * chkp-builtins.def: New.
993 * builtins.def: include chkp-builtins.def.
994 (DEF_CHKP_BUILTIN): New.
995 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
996 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
997 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
998 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
999 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
1000 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
1001 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
1002 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
1003 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
1004 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
1005 * common.opt (fcheck-pointer-bounds): New.
1006 * toplev.c (process_options): Check Pointer Bounds Checker is
1008 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
1010 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1012 * target.def (builtin_chkp_function): New.
1013 (chkp_bound_type): New.
1014 (chkp_bound_mode): New.
1015 (fn_abi_va_list_bounds_size): New.
1016 (load_bounds_for_arg): New.
1017 (store_bounds_for_arg): New.
1018 * targhooks.h (default_load_bounds_for_arg): New.
1019 (default_store_bounds_for_arg): New.
1020 (default_fn_abi_va_list_bounds_size): New.
1021 (default_chkp_bound_type): New.
1022 (default_chkp_bound_mode): New.
1023 (default_builtin_chkp_function): New.
1024 * targhooks.c (default_load_bounds_for_arg): New.
1025 (default_store_bounds_for_arg): New.
1026 (default_fn_abi_va_list_bounds_size): New.
1027 (default_chkp_bound_type): New.
1028 (default_chkp_bound_mode); New.
1029 (default_builtin_chkp_function): New.
1030 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
1031 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
1032 (TARGET_STORE_BOUNDS_FOR_ARG): New.
1033 (TARGET_BUILTIN_CHKP_FUNCTION): New.
1034 (TARGET_CHKP_BOUND_TYPE): New.
1035 (TARGET_CHKP_BOUND_MODE): New.
1036 * doc/tm.texi: Regenerated.
1037 * langhooks.h (lang_hooks): Add chkp_supported field.
1038 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
1039 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
1041 2013-10-29 Andrew Pinski <apinski@cavium.com>
1043 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
1044 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
1045 fails, combining the branches anyways.
1046 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
1047 increases the number of combinings.
1048 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
1050 2013-10-29 Mike Stump <mikestump@comcast.net>
1052 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
1053 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
1054 (make_vector_mode): Increase namebuf to 16.
1055 (emit_insn_modes_h): When processing BImode, don't
1056 also match partial int modes.
1057 (emit_class_narrowest_mode): Likewise.
1059 * config/bfin/bfin-modes.def: Add precision to PDI.
1060 * config/m32c/m32c-modes.def: Add precision to PSI.
1061 * config/msp430/msp430-modes.def: Add precision to PSI.
1062 * config/rs6000/rs6000-modes.def: Add precision to PTI.
1063 * config/sh/sh-modes.def: Add precision to PSI and PDI.
1065 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
1068 * config/sh/sh.md (*addc): Rename existing variations to ...
1069 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
1070 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
1071 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
1072 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
1074 2013-10-29 Teresa Johnson <tejohnson@google.com>
1077 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
1078 insanities when updating probabilities.
1080 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1082 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
1083 prettyprinter for cgraph_node to reflect the conversion of the
1084 symtable types to a C++ class hierarchy: it now *is* a
1085 symtab_node_base, rather than having one (named "symbol").
1087 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
1089 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
1090 __cilkrts_pop_frame. If matched, then return true for built-in
1092 (expand_builtin): Added BUILT_IN_CILK_DETACH and
1093 BUILT_IN_CILK_POP_FRAME case.
1094 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
1095 (lhs_cilk_detect_spawn): Likewise.
1096 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
1097 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
1098 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
1099 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
1100 (LANG_HOOKS_CILKPLUS): Likewise.
1101 * tree.h (CILK_SPAWN_FN): Likewise.
1102 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
1103 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
1104 (OBJS): Added cilk-common.o.
1105 (BUILTINS_DEF): Added cilk-builtins.def.
1106 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
1107 (lhd_cilk_detect_spawn): Likewise.
1108 * langhooks.h (lang_hooks_for_cilkplus): New struct.
1109 (struct lang_hooks): Added new field called "cilkplus."
1110 * cilk-common.c: New file.
1112 * cilk-builtins.def: Likewise.
1113 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
1114 "__cilk" macro and set it to 200.
1115 * function.h (struct function::cilk_frame_decl): New field.
1116 (struct function::is_cilk_function): Likewise.
1117 (struct function::calls_cilk_spawn): Likewise.
1118 * gimplify.c (gimplify_call_expr): Added a check if the function call
1119 being gimplified is a spawn detach point. If so, then add pop_frame
1120 and detach function calls.
1121 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
1122 for gimplifying _Cilk_spawn and _Cilk_sync statements.
1123 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
1124 function. If so, added a _Cilk_sync and gimplified it.
1125 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
1126 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
1127 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
1129 (can_inline_edge_p): Prevent inling of spawnee function.
1130 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
1131 for functions that use Cilk keywords.
1132 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
1133 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
1134 CILK_SYNC_STMT cases.
1135 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
1137 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
1138 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
1139 * passes.texi (Cilk Keywords): New section that describes the compiler
1140 code changes for handling Cilk Keywords.
1142 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1144 Patch autogenerated by refactor_symtab.py from
1145 https://github.com/davidmalcolm/gcc-refactoring-scripts
1146 revision 58bb219cc090b2f4516a9297d868c245495ee622
1148 * asan.c (asan_finish_file): Update for conversion of symtab types to
1149 a true class hierarchy.
1150 * cfgexpand.c (estimated_stack_frame_size): Likewise.
1151 * cgraph.c (cgraph_get_body): Likewise.
1152 (cgraph_get_create_real_symbol_node): Likewise.
1153 (verify_cgraph_node): Likewise.
1154 (verify_edge_corresponds_to_fndecl): Likewise.
1155 (verify_edge_count_and_frequency): Likewise.
1156 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
1157 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1158 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
1159 (cgraph_node_cannot_return): Likewise.
1160 (cgraph_set_pure_flag_1): Likewise.
1161 (cgraph_set_const_flag_1): Likewise.
1162 (cgraph_set_nothrow_flag_1): Likewise.
1163 (cgraph_make_node_local_1): Likewise.
1164 (cgraph_for_node_and_aliases): Likewise.
1165 (cgraph_for_node_thunks_and_aliases): Likewise.
1166 (cgraph_node_can_be_local_p): Likewise.
1167 (cgraph_node_cannot_be_local_p_1): Likewise.
1168 (cgraph_function_body_availability): Likewise.
1169 (dump_cgraph_node): Likewise.
1170 (cgraph_rtl_info): Likewise.
1171 (cgraph_mark_address_taken_node): Likewise.
1172 (cgraph_remove_node): Likewise.
1173 (cgraph_release_function_body): Likewise.
1174 (cgraph_update_edges_for_call_stmt_node): Likewise.
1175 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
1176 (cgraph_make_edge_direct): Likewise.
1177 (cgraph_resolve_speculation): Likewise.
1178 (cgraph_speculative_call_info): Likewise.
1179 (cgraph_turn_edge_to_speculative): Likewise.
1180 (cgraph_create_edge_1): Likewise.
1181 (cgraph_set_call_stmt): Likewise.
1182 (cgraph_node_for_asm): Likewise.
1183 (cgraph_add_thunk): Likewise.
1184 (cgraph_same_body_alias): Likewise.
1185 (cgraph_create_function_alias): Likewise.
1186 (cgraph_create_node): Likewise.
1187 (cgraph_create_empty_node): Likewise.
1188 (record_function_versions): Likewise.
1189 (used_from_object_file_p): Likewise.
1190 * cgraph.h (symtab_can_be_discarded): Likewise.
1191 (symtab_real_symbol_p): Likewise.
1192 (cgraph_mark_force_output_node): Likewise.
1193 (cgraph_edge_recursive_p): Likewise.
1194 (symtab_alias_target): Likewise.
1195 (varpool_all_refs_explicit_p): Likewise.
1196 (varpool_can_remove_if_no_refs): Likewise.
1197 (cgraph_only_called_directly_or_aliased_p): Likewise.
1198 (cgraph_next_function_with_gimple_body): Likewise.
1199 (cgraph_first_function_with_gimple_body): Likewise.
1200 (cgraph_function_with_gimple_body_p): Likewise.
1201 (cgraph_next_function): Likewise.
1202 (cgraph_first_function): Likewise.
1203 (cgraph_next_defined_function): Likewise.
1204 (cgraph_first_defined_function): Likewise.
1205 (varpool_next_defined_variable): Likewise.
1206 (varpool_first_defined_variable): Likewise.
1207 (varpool_next_static_initializer): Likewise.
1208 (varpool_first_static_initializer): Likewise.
1209 (varpool_next_variable): Likewise.
1210 (varpool_first_variable): Likewise.
1211 (varpool_node_name): Likewise.
1212 (varpool): Likewise.
1214 (is_a_helper <varpool_node>::test): Likewise.
1215 (is_a_helper <cgraph_node>::test): Likewise.
1216 (varpool_variable_node): Likewise.
1217 (cgraph_function_or_thunk_node): Likewise.
1218 (varpool_alias_target): Likewise.
1219 (cgraph_alias_target): Likewise.
1220 (cgraph_node_name): Likewise.
1221 (varpool_node_asm_name): Likewise.
1222 (cgraph_node_asm_name): Likewise.
1223 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
1224 (cgraph_rebuild_references): Likewise.
1225 (rebuild_cgraph_edges): Likewise.
1226 (record_eh_tables): Likewise.
1227 (build_cgraph_edges): Likewise.
1228 (mark_store): Likewise.
1229 (mark_load): Likewise.
1230 (mark_address): Likewise.
1231 (record_type_list): Likewise.
1232 (record_reference): Likewise.
1233 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
1234 (cgraph_materialize_clone): Likewise.
1235 (cgraph_function_versioning): Likewise.
1236 (cgraph_copy_node_for_versioning): Likewise.
1237 (update_call_expr): Likewise.
1238 (cgraph_find_replacement_node): Likewise.
1239 (cgraph_create_virtual_clone): Likewise.
1240 (cgraph_clone_node): Likewise.
1241 * cgraphunit.c (compile): Likewise.
1242 (output_weakrefs): Likewise.
1243 (output_in_order): Likewise.
1244 (expand_function): Likewise.
1245 (assemble_thunks_and_aliases): Likewise.
1246 (expand_thunk): Likewise.
1247 (mark_functions_to_output): Likewise.
1248 (handle_alias_pairs): Likewise.
1249 (analyze_functions): Likewise.
1250 (walk_polymorphic_call_targets): Likewise.
1251 (varpool_finalize_decl): Likewise.
1252 (process_function_and_variable_attributes): Likewise.
1253 (cgraph_process_same_body_aliases): Likewise.
1254 (analyze_function): Likewise.
1255 (cgraph_add_new_function): Likewise.
1256 (cgraph_finalize_function): Likewise.
1257 (referred_to_p): Likewise.
1258 (cgraph_reset_node): Likewise.
1259 (cgraph_process_new_functions): Likewise.
1260 (enqueue_node): Likewise.
1261 (decide_is_symbol_needed): Likewise.
1262 * coverage.c (coverage_compute_profile_id): Likewise.
1263 * dbxout.c (dbxout_expand_expr): Likewise.
1264 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
1265 (reference_to_unused): Likewise.
1266 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1267 * gimplify.c (unvisit_body): Likewise.
1268 (unshare_body): Likewise.
1269 * ipa-cp.c (ipcp_generate_summary): Likewise.
1270 (ipcp_decision_stage): Likewise.
1271 (identify_dead_nodes): Likewise.
1272 (decide_whether_version_node): Likewise.
1273 (decide_about_value): Likewise.
1274 (perhaps_add_new_callers): Likewise.
1275 (create_specialized_node): Likewise.
1276 (update_profiling_info): Likewise.
1277 (ipcp_propagate_stage): Likewise.
1278 (estimate_local_effects): Likewise.
1279 (good_cloning_opportunity_p): Likewise.
1280 (devirtualization_time_bonus): Likewise.
1281 (propagate_constants_accross_call): Likewise.
1282 (initialize_node_lattices): Likewise.
1283 (ipcp_cloning_candidate_p): Likewise.
1284 (determine_versionability): Likewise.
1285 (print_all_lattices): Likewise.
1286 (print_lattice): Likewise.
1287 (ipcp_discover_new_direct_edges): Likewise.
1288 * ipa-devirt.c (ipa_devirt): Likewise.
1289 (likely_target_p): Likewise.
1290 (update_type_inheritance_graph): Likewise.
1291 (possible_polymorphic_call_target_p): Likewise.
1292 (dump_possible_polymorphic_call_targets): Likewise.
1293 (devirt_variable_node_removal_hook): Likewise.
1294 (record_binfo): Likewise.
1295 (maybe_record_node): Likewise.
1296 (build_type_inheritance_graph): Likewise.
1297 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1298 (inline_generate_summary): Likewise.
1299 (inline_analyze_function): Likewise.
1300 (do_estimate_growth): Likewise.
1301 (simple_edge_hints): Likewise.
1302 (estimate_node_size_and_time): Likewise.
1303 (estimate_edge_devirt_benefit): Likewise.
1304 (compute_inline_parameters): Likewise.
1305 (estimate_function_body_sizes): Likewise.
1306 (compute_bb_predicates): Likewise.
1307 (initialize_inline_failed): Likewise.
1308 (dump_inline_summary): Likewise.
1309 (dump_inline_edge_summary): Likewise.
1310 * ipa-inline-transform.c (inline_transform): Likewise.
1311 (preserve_function_body_p): Likewise.
1312 (save_inline_function_body): Likewise.
1313 (inline_call): Likewise.
1314 (clone_inlined_nodes): Likewise.
1315 (can_remove_node_now_p): Likewise.
1316 (can_remove_node_now_p_1): Likewise.
1317 * ipa-inline.c (early_inliner): Likewise.
1318 (early_inline_small_functions): Likewise.
1319 (inline_always_inline_functions): Likewise.
1320 (ipa_inline): Likewise.
1321 (flatten_function): Likewise.
1322 (inline_small_functions): Likewise.
1323 (speculation_useful_p): Likewise.
1324 (recursive_inlining): Likewise.
1325 (update_caller_keys): Likewise.
1326 (reset_edge_caches): Likewise.
1327 (update_edge_key): Likewise.
1328 (edge_badness): Likewise.
1329 (relative_time_benefit): Likewise.
1330 (want_inline_self_recursive_call_p): Likewise.
1331 (want_inline_small_function_p): Likewise.
1332 (want_early_inline_function_p): Likewise.
1333 (num_calls): Likewise.
1334 (can_early_inline_edge_p): Likewise.
1335 (can_inline_edge_p): Likewise.
1336 (report_inline_failed_reason): Likewise.
1337 * ipa-profile.c (ipa_profile): Likewise.
1338 (ipa_propagate_frequency): Likewise.
1339 (ipa_propagate_frequency_1): Likewise.
1340 (ipa_profile_generate_summary): Likewise.
1341 * ipa-prop.c (ipcp_transform_function): Likewise.
1342 (read_replacements_section): Likewise.
1343 (ipa_prop_read_section): Likewise.
1344 (ipa_modify_call_arguments): Likewise.
1345 (ipa_print_node_params): Likewise.
1346 (propagate_controlled_uses): Likewise.
1347 (update_indirect_edges_after_inlining): Likewise.
1348 (remove_described_reference): Likewise.
1349 (ipa_make_edge_direct_to_target): Likewise.
1350 (ipa_analyze_node): Likewise.
1351 (ipa_analyze_params_uses): Likewise.
1352 (ipa_compute_jump_functions): Likewise.
1353 (ipa_get_callee_param_type): Likewise.
1354 (ipa_print_node_jump_functions): Likewise.
1355 (ipa_initialize_node_params): Likewise.
1356 (ipa_populate_param_decls): Likewise.
1357 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
1358 (write_agg_replacement_chain): Likewise.
1359 (ipa_write_node_info): Likewise.
1360 (ipa_edge_duplication_hook): Likewise.
1361 (try_decrement_rdesc_refcount): Likewise.
1362 * ipa-pure-const.c (propagate_nothrow): Likewise.
1363 (propagate_pure_const): Likewise.
1364 (pure_const_read_summary): Likewise.
1365 (pure_const_write_summary): Likewise.
1366 (analyze_function): Likewise.
1367 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
1368 (ipa_ref_referring_ref_list): Likewise.
1369 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
1370 (ipa_remove_stmt_references): Likewise.
1371 (ipa_find_reference): Likewise.
1372 (ipa_dump_referring): Likewise.
1373 (ipa_dump_references): Likewise.
1374 (ipa_record_reference): Likewise.
1375 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
1376 (ipa_reference_write_optimization_summary): Likewise.
1377 (write_node_summary_p): Likewise.
1378 (propagate): Likewise.
1379 (read_write_all_from_decl): Likewise.
1380 (generate_summary): Likewise.
1381 (analyze_function): Likewise.
1382 (propagate_bits): Likewise.
1383 (ipa_reference_get_not_written_global): Likewise.
1384 (ipa_reference_get_not_read_global): Likewise.
1385 * ipa-split.c (execute_split_functions): Likewise.
1386 (split_function): Likewise.
1387 * ipa-utils.c (ipa_merge_profiles): Likewise.
1388 (dump_cgraph_node_set): Likewise.
1389 (ipa_reverse_postorder): Likewise.
1390 (ipa_edge_within_scc): Likewise.
1391 (ipa_get_nodes_in_cycle): Likewise.
1392 (ipa_free_postorder_info): Likewise.
1393 (ipa_reduced_postorder): Likewise.
1394 (searchc): Likewise.
1395 (recursive_call_p): Likewise.
1396 * ipa.c (ipa_cdtor_merge): Likewise.
1397 (record_cdtor_fn): Likewise.
1398 (function_and_variable_visibility): Likewise.
1399 (varpool_externally_visible_p): Likewise.
1400 (cgraph_externally_visible_p): Likewise.
1401 (comdat_can_be_unshared_p): Likewise.
1402 (comdat_can_be_unshared_p_1): Likewise.
1403 (address_taken_from_non_vtable_p): Likewise.
1404 (ipa_discover_readonly_nonaddressable_vars): Likewise.
1405 (symtab_remove_unreachable_nodes): Likewise.
1406 (walk_polymorphic_call_targets): Likewise.
1407 (process_references): Likewise.
1408 (enqueue_node): Likewise.
1409 (has_addr_references_p): Likewise.
1410 (cgraph_non_local_node_p_1): Likewise.
1411 * is-a.h (varpool_analyze_node): Likewise.
1412 * lto-cgraph.c (input_symtab): Likewise.
1413 (merge_profile_summaries): Likewise.
1414 (input_cgraph_1): Likewise.
1415 (input_edge): Likewise.
1416 (input_varpool_node): Likewise.
1417 (input_node): Likewise.
1418 (input_overwrite_node): Likewise.
1419 (compute_ltrans_boundary): Likewise.
1420 (output_refs): Likewise.
1421 (lto_output_varpool_node): Likewise.
1422 (lto_output_node): Likewise.
1423 (reachable_from_other_partition_p): Likewise.
1424 (referenced_from_other_partition_p): Likewise.
1425 (lto_output_edge): Likewise.
1426 (output_node_opt_summary): Likewise.
1427 (add_node_to): Likewise.
1428 (reachable_from_this_partition_p): Likewise.
1429 (lto_set_symtab_encoder_in_partition): Likewise.
1430 (lto_symtab_encoder_in_partition_p): Likewise.
1431 (lto_set_symtab_encoder_encode_initializer): Likewise.
1432 (lto_symtab_encoder_encode_initializer_p): Likewise.
1433 (lto_set_symtab_encoder_encode_body): Likewise.
1434 (lto_symtab_encoder_encode_body_p): Likewise.
1435 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1437 * lto-streamer-in.c (lto_read_body): Likewise.
1438 (fixup_call_stmt_edges): Likewise.
1439 (fixup_call_stmt_edges_1): Likewise.
1440 * lto-streamer-out.c (produce_symtab): Likewise.
1441 (output_symbol_p): Likewise.
1442 (write_symbol): Likewise.
1443 (lto_output): Likewise.
1444 (copy_function): Likewise.
1445 (output_function): Likewise.
1446 * passes.c (function_called_by_processed_nodes_p): Likewise.
1447 (ipa_write_optimization_summaries): Likewise.
1448 (ipa_write_summaries): Likewise.
1449 (do_per_function_toporder): Likewise.
1450 (do_per_function): Likewise.
1451 (dump_passes): Likewise.
1452 * symtab.c (symtab_semantically_equivalent_p): Likewise.
1453 (symtab_nonoverwritable_alias): Likewise.
1454 (symtab_nonoverwritable_alias_1): Likewise.
1455 (symtab_for_node_and_aliases): Likewise.
1456 (symtab_resolve_alias): Likewise.
1457 (fixup_same_cpp_alias_visibility): Likewise.
1458 (symtab_alias_ultimate_target): Likewise.
1459 (symtab_used_from_object_file_p): Likewise.
1460 (verify_symtab_base): Likewise.
1461 (dump_symtab_base): Likewise.
1462 (symtab_node_name): Likewise.
1463 (symtab_node_asm_name): Likewise.
1464 (symtab_dissolve_same_comdat_group_list): Likewise.
1465 (symtab_add_to_same_comdat_group): Likewise.
1466 (symtab_unregister_node): Likewise.
1467 (symtab_insert_node_to_hashtable): Likewise.
1468 (symtab_register_node): Likewise.
1469 (unlink_from_assembler_name_hash): Likewise.
1470 (insert_to_assembler_name_hash): Likewise.
1471 (eq_assembler_name): Likewise.
1472 (hash_node_by_assembler_name): Likewise.
1473 (eq_node): Likewise.
1474 (hash_node): Likewise.
1475 * toplev.c (wrapup_global_declaration_2): Likewise.
1476 * trans-mem.c (ipa_tm_execute): Likewise.
1477 (ipa_tm_transform_clone): Likewise.
1478 (ipa_tm_transform_transaction): Likewise.
1479 (ipa_tm_transform_calls_redirect): Likewise.
1480 (ipa_tm_insert_gettmclone_call): Likewise.
1481 (ipa_tm_insert_irr_call): Likewise.
1482 (ipa_tm_create_version): Likewise.
1483 (ipa_tm_create_version_alias): Likewise.
1484 (ipa_tm_mark_forced_by_abi_node): Likewise.
1485 (ipa_tm_mark_force_output_node): Likewise.
1486 (ipa_tm_diagnose_tm_safe): Likewise.
1487 (ipa_tm_mayenterirr_function): Likewise.
1488 (ipa_tm_scan_irr_function): Likewise.
1489 (ipa_tm_note_irrevocable): Likewise.
1490 (ipa_tm_scan_calls_clone): Likewise.
1491 (get_cg_data): Likewise.
1492 * tree-eh.c (tree_could_trap_p): Likewise.
1493 * tree-emutls.c (ipa_lower_emutls): Likewise.
1494 (create_emultls_var): Likewise.
1495 (lower_emutls_function_body): Likewise.
1496 (gen_emutls_addr): Likewise.
1497 (emutls_decl): Likewise.
1498 (new_emutls_decl): Likewise.
1499 * tree-inline.c (tree_function_versioning): Likewise.
1500 (optimize_inline_calls): Likewise.
1501 (expand_call_inline): Likewise.
1502 (estimate_num_insns): Likewise.
1503 (copy_bb): Likewise.
1504 (delete_unreachable_blocks_update_callgraph): Likewise.
1505 * tree-nested.c (gimplify_all_functions): Likewise.
1506 (create_nesting_tree): Likewise.
1507 (check_for_nested_with_variably_modified): Likewise.
1508 * tree-pretty-print.c (dump_function_header): Likewise.
1509 * tree-profile.c (tree_profiling): Likewise.
1510 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
1511 (modify_function): Likewise.
1512 (convert_callers): Likewise.
1513 (convert_callers_for_node): Likewise.
1514 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
1515 (associate_varinfo_to_alias): Likewise.
1516 (create_variable_info_for): Likewise.
1517 (get_constraint_for_ssa_var): Likewise.
1518 * tree-vectorizer.c (increase_alignment): Likewise.
1519 * tree.c (find_decls_types_in_var): Likewise.
1520 (find_decls_types_in_node): Likewise.
1521 (free_lang_data_in_decl): Likewise.
1522 * value-prof.c (gimple_ic_transform): Likewise.
1523 (gimple_ic): Likewise.
1524 (check_ic_target): Likewise.
1525 (init_node_map): Likewise.
1526 * varasm.c (decl_binds_to_current_def_p): Likewise.
1527 (default_binds_local_p_1): Likewise.
1528 (dump_tm_clone_pairs): Likewise.
1529 (assemble_alias): Likewise.
1530 (find_decl): Likewise.
1531 (mark_decl_referenced): Likewise.
1532 * varpool.c (varpool_for_node_and_aliases): Likewise.
1533 (varpool_extra_name_alias): Likewise.
1534 (varpool_create_variable_alias): Likewise.
1535 (add_new_static_var): Likewise.
1536 (varpool_finalize_named_section_flags): Likewise.
1537 (varpool_remove_unreferenced_decls): Likewise.
1538 (enqueue_node): Likewise.
1539 (varpool_assemble_decl): Likewise.
1540 (assemble_aliases): Likewise.
1541 (varpool_analyze_node): Likewise.
1542 (cgraph_variable_initializer_availability): Likewise.
1543 (varpool_add_new_variable): Likewise.
1544 (ctor_for_folding): Likewise.
1545 (dump_varpool_node): Likewise.
1546 (varpool_remove_initializer): Likewise.
1547 (varpool_remove_node): Likewise.
1548 (varpool_node_for_decl): Likewise.
1549 (varpool_create_empty_node): Likewise.
1550 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
1551 (ix86_get_function_versions_dispatcher): Likewise.
1553 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1555 * cgraph.h (symtab_node_base): Convert to a class;
1556 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
1557 chain_next/prev from symtab_node_def.
1558 (cgraph_node): Inherit from symtab_node; add GTY option
1559 tag ("SYMTAB_FUNCTION").
1560 (varpool_node): Inherit from symtab_node; add GTY option
1561 tag ("SYMTAB_VARIABLE").
1562 (symtab_node_def): Remove.
1563 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
1564 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
1565 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
1566 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
1568 * ipa-ref.h (symtab_node_def): Drop.
1569 (symtab_node): Change underlying type from symtab_node_def to
1571 (const_symtab_node): Likwise.
1573 * is-a.h: Update examples in comment.
1575 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
1576 (assembler_name_hash): Likewise.
1578 2013-10-29 Martin Liska <marxin.liska@gmail.com>
1580 * doc/tree-ssa.texi (gimple_phi_result): Document.
1581 (gimple_phi_num_args, gimple_phi_arg): Likewise.
1582 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
1583 (PHI_RESULT, PHI_NUM_ARGS): Remove.
1584 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
1586 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1588 * expr.h: Revert change and include tree-core.h.
1589 * rtl.h: Revert change and don't include tree-core.h.
1591 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1593 * config/darwin.c: Include gimple.h.
1594 * config/i386/winnt.c: Likewise.
1596 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
1598 PR tree-optimization/19831
1599 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
1601 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1603 * tree-outof-ssa.h: Remove include files.
1604 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
1606 * tree-ssa-coalesce.c: Likewise.
1607 * cfgexpand.c: Likewise.
1608 * tree-ssa-ter.c: Likewise.
1609 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
1610 * lto-streamer.h: Likewise.
1611 * cgraphbuild.c: Add gimple.h to include list.
1612 * data-streamer-in.c: Likewise.
1613 * ipa-cp.c: Likewise.
1614 * tree-streamer.c: Likewise.
1615 * lto-compress.c: Likewise.
1616 * ipa-reference.c: Likewise.
1617 * data-streamer-out.c: Likewise.
1618 * lto-cgraph.c: Likewise.
1619 * cgraphclones.c: Likewise.
1620 * ipa-utils.c: Likewise.
1621 * data-streamer.c: Likewise.
1622 * ipa-split.c: Likewise.
1623 * lto-section-in.c: Likewise.
1624 * tree-streamer-out.c: Likewise.
1625 * ipa-prop.c: Likewise.
1626 * tree-streamer-in.c: Likewise.
1627 * symtab.c: Likewise.
1628 * opts-global.c: Likewise.
1629 * lto-opts.c: Likewise.
1630 * lto-section-out.c: Likewise.
1631 * lto-streamer.c: Likewise.
1632 * rtl.h: Add tree-core.h to include list.
1633 * expr.h: Remove tree-core.h from include list.
1634 * gimple.h: Likewise.
1635 * ipa-utils.h: Likewise.
1636 * streamer-hooks.h: Likewise.
1637 * streamer-hooks.c: Include input.h.
1639 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1641 * config/arm/arm.c (cortexa7_extra_costs): New table.
1642 (arm_cortex_a7_tune): New.
1643 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
1645 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
1647 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
1649 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1651 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
1652 to use autogenerated markers for a class-hierarchy, every class
1653 must have a GTY marker.
1654 * gengtype.h (struct type): Add linked list of subclasses to
1655 the "s" member of the union.
1656 (add_subclass): New decl.
1657 * gengtype-state.c (read_state_struct_type): Set up subclass
1659 * gengtype.c (get_ultimate_base_class): New.
1660 (add_subclass): New.
1661 (new_structure): Set up subclass linked list.
1662 (set_gc_used_type): Propagate usage information to subclasses.
1663 (output_mangled_typename): Use get_ultimate_base_class.
1664 (walk_subclasses): Use the subclass linked list, avoiding an
1665 O(N^2) when writing out all types.
1666 (walk_type): Issue an error if the base class is missing a tag,
1667 rather than generating bogus C code. Add a gcc_unreachable
1668 default case, in case people omit tags from concrete subclasses,
1669 or get the values wrong.
1670 (write_func_for_structure): Issue an error for subclasses for
1671 which the base doesn't have a "desc", since otherwise the
1672 autogenerated routines for the base would silently fail to visit
1673 any subclass fields.
1674 (write_root): Use get_ultimate_base_class, tweaking constness of
1675 tp to match that function's signature.
1677 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1679 * doc/gty.texi (GTY Options): Add note about inheritance to
1680 description of desc and tag.
1681 (Inheritance and GTY): New.
1683 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1685 * gengtype-parse.c (opts_have): Drop "static" so that
1686 we can use this from gengtype.c.
1687 * gengtype.c (set_gc_used_type): Mark any base class as used;
1688 update field traversal to visit inherited fields.
1689 (output_mangled_typename): Convert references to classes within
1690 an inheritance hierarchy to reference the ultimate base class,
1691 since only it will have gt_ functions.
1692 (get_string_option): New.
1693 (walk_subclasses): New.
1694 (walk_type): Treat GTY structs that have a "desc" as being the
1695 root of an inheritance hierarchy. Generate a switch on it
1696 within the marking function which walks all subclasses, adding
1697 cases for them via walk_subclasses. For subclasses, visit all
1698 fields of the type (including inherited ones).
1699 (write_func_for_structure): Don't write fns for subclasses, only
1700 for the ultimate base class within an inheritance hierarchy.
1701 Subclasses-marking will be handled by the base class marking functions.
1702 (write_types): Likewise.
1703 (write_local_func_for_structure): Likewise.
1704 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
1705 a "tag" option (and are thus concrete subclasses).
1706 (write_root): Use the marker function for the ultimate base class.
1707 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
1708 (opts_have): Add declaration.
1710 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
1712 * lra-spills.c (lra_final_code_change): Remove useless move insns
1713 originated from moves of pseudos.
1715 2013-10-28 Jeff Law <law@redhat.com>
1717 * lower-subreg.c (resolve_simple_move): Fix comment typo.
1719 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
1721 * df-scan.c (df_collection_rec): Adjust.
1722 (copy_defs): New constant.
1723 (copy_uses): Likewise.
1724 (copy_eq_uses): Likewise.
1725 (copy_mw): Likewise.
1726 (copy_all): Likewise.
1727 (df_insn_rescan): Adjust.
1728 (df_notes_rescan): Likewise.
1729 (df_swap_refs): Likewise.
1730 (df_sort_and_compress_refs): Likewise.
1731 (df_sort_and_compress_mws): Likewise.
1732 (df_install_refs): Likewise.
1733 (df_install_mws): Likewise.
1734 (df_refs_add_to_chains): Add flags parameter controlling which vectors
1736 (df_bb_refs_record): Adjust.
1737 (df_record_entry_block_defs): Likewise.
1738 (df_record_exit_block_defs): Likewise.
1739 (df_refs_verify): Likewise.
1740 (df_mws_verify): Likewise.
1741 (df_insn_refs_verify): Likewise.
1742 (df_bb_verify): Likewise.
1743 * ipa-pure-const.c (finish_state): Remove.
1744 (propagate): Adjust.
1745 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
1746 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
1747 var-tracking.c: Adjust.
1748 * vec.c (stack_vecs): Remove.
1749 (register_stack_vec): Likewise.
1750 (stack_vec_register_index): Likewise.
1751 (unregister_stack_vec): Likewise.
1752 * vec.h (struct va_stack): Remove.
1753 (struct vec<T, A, vl_ptr>): Specialize as
1754 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
1755 allocation strategy compatable with the vl_ptr layout.
1756 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
1757 specialization anyway.
1758 (class stack_vec): New class.
1759 (vec_stack_alloc): Remove.
1760 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
1762 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1763 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1764 Sergey Lega <sergey.s.lega@intel.com>
1765 Anna Tikhonova <anna.tikhonova@intel.com>
1766 Ilya Tocar <ilya.tocar@intel.com>
1767 Andrey Turetskiy <andrey.turetskiy@intel.com>
1768 Ilya Verbin <ilya.verbin@intel.com>
1769 Kirill Yukhin <kirill.yukhin@intel.com>
1770 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1772 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
1773 (*prefetch_avx512pf_<mode>): New.
1774 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
1775 (avx512f_maskcmp<mode>3): Ditto.
1776 (vashrv16si3): Ditto.
1778 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1779 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1780 Sergey Lega <sergey.s.lega@intel.com>
1781 Anna Tikhonova <anna.tikhonova@intel.com>
1782 Ilya Tocar <ilya.tocar@intel.com>
1783 Andrey Turetskiy <andrey.turetskiy@intel.com>
1784 Ilya Verbin <ilya.verbin@intel.com>
1785 Kirill Yukhin <kirill.yukhin@intel.com>
1786 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1788 * config/i386/i386.md (any_truncate): New.
1789 (trunsuffix): Ditto.
1790 * config/i386/predicates.md (const_8_to_9_operand): New.
1791 (const_10_to_11_operand): Ditto.
1792 (const_12_to_13_operand): Ditto.
1793 (const_14_to_15_operand): Ditto.
1794 (const_16_to_19_operand): Ditto.
1795 (const_20_to_23_operand): Ditto.
1796 (const_24_to_27_operand): Ditto.
1797 (const_28_to_31_operand): Ditto.
1798 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
1799 (cvtusi2<ssescalarmodesuffix>32): New.
1800 (cvtusi2<ssescalarmodesuffix>64): Ditto.
1801 (ufloatv16siv16sf2): Ditto.
1802 (avx512f_fix_notruncv16sfv16si): Ditto.
1803 (avx512f_ufix_notruncv16sfv16si): Ditto.
1804 (avx512f_vcvtss2usi): Ditto.
1805 (avx512f_vcvtss2usiq): Ditto.
1806 (avx512f_vcvttss2usi): Ditto.
1807 (avx512f_vcvttss2usiq): Ditto.
1808 (avx512f_vcvtsd2usi): Ditto.
1809 (avx512f_vcvtsd2usiq): Ditto.
1810 (avx512f_vcvttsd2usi): Ditto.
1811 (avx512f_vcvttsd2usiq): Ditto.
1812 (ufloatv8siv8df): Ditto.
1813 (avx512f_cvtdq2pd512_2): Ditto.
1814 (avx512f_cvtpd2dq512): Ditto.
1815 (avx512f_ufix_notruncv8dfv8si): Ditto.
1816 (avx512f_cvtpd2ps512): Ditto.
1817 (vec_unpacks_lo_v16sf): Ditto.
1818 (vec_unpacks_hi_v16sf): Ditto.
1819 (vec_unpacks_float_hi_v16si): Ditto.
1820 (vec_unpacks_float_lo_v16si): Ditto.
1821 (avx512f_unpckhps512): Ditto.
1822 (avx512f_unpcklps512): Ditto.
1823 (avx512f_movshdup512): Ditto.
1824 (avx512f_movsldup512): Ditto.
1825 (vec_extract_lo_v32hi): Ditto.
1826 (vec_extract_hi_v32hi): Ditto.
1827 (vec_extract_lo_v64qi): Ditto.
1828 (vec_extract_hi_v64qi): Ditto.
1829 (avx512f_unpckhpd512): Ditto.
1830 (avx512f_movddup512): Ditto.
1831 (avx512f_unpcklpd512): Ditto.
1832 (*avx512f_unpcklpd512): Ditto.
1833 (avx512f_shufps512_1): Ditto.
1834 (avx512f_shufpd512_1): Ditto.
1835 (avx512f_interleave_highv8di): Ditto.
1836 (avx512f_interleave_lowv8di): Ditto.
1837 (PMOV_DST_MODE): Ditto.
1838 (pmov_src_mode): Ditto.
1839 (pmov_src_lower): Ditto.
1841 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
1842 (*avx512f_<code>v8div16qi2): Ditto.
1843 (*avx512f_<code>v8div16qi2_store): Ditto.
1844 (vec_widen_umult_even_v16si): Ditto.
1845 (*vec_widen_umult_even_v16si): Ditto.
1846 (vec_widen_smult_even_v16si): Ditto.
1847 (*vec_widen_smult_even_v16si): Ditto.
1848 (avx512f_interleave_highv16si): Ditto.
1849 (avx512f_interleave_lowv16si): Ditto.
1850 (avx512f_<code>v16qiv16si2): Ditto.
1851 (avx512f_<code>v16hiv16si2): Ditto.
1852 (avx512f_<code>v8qiv8di2): Ditto.
1853 (avx512f_<code>v8hiv8di2): Ditto.
1854 (avx512f_<code>v8siv8di2): Ditto.
1855 (avx512cd_maskb_vec_dupv8di): Ditto.
1856 (avx512cd_maskw_vec_dupv16si): Ditto.
1857 (avx512f_vcvtph2ps512): Ditto.
1858 (avx512f_vcvtps2ph512): Ditto.
1859 (VEC_EXTRACT_MODE): Extened with wider modes.
1860 (VEC_PERM_AVX2): Ditto.
1861 (VEC_PERM_CONST): Ditto.
1863 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
1865 * config/arc/arc.c (arc_ccfsm_post_advance):
1866 Add comment about TYPE_RETURN.
1868 2013-10-28 Bin Cheng <bin.cheng@arm.com>
1870 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
1871 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
1872 (strip_offset): Convert offset to unsigned number.
1874 2013-10-27 Tom de Vries <tom@codesourcery.com>
1876 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
1877 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
1879 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
1881 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
1882 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
1883 (TARGET_BIG_ENDIAN): New macro.
1884 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
1885 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
1886 * config/sh/sh.c: Likewise.
1887 * config/sh/sh.h: Likewise.
1889 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
1891 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
1892 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
1894 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1896 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
1898 (cmpstr_t, cmpstrsi): Fix formatting.
1900 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1903 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
1904 addressing, do not use general_operand for memory operands.
1906 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
1909 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1910 * lra-spills.c (lra_final_code_change): Remove useless move insns.
1912 2013-10-26 Jeff Law <law@redhat.com>
1914 * predict.c (PRED_MUDFLAP): Remove.
1915 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
1917 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
1918 (OBJS): Remove tree-nomudflap.o
1919 (GTFILES): Remove tree-mudflap.c
1920 * builtins.c (expand_builtin_alloc): Remove mudflap support.
1921 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
1922 (mfwrap_spec, mflib_spec): Likewise.
1923 (cpp_unique_options, cc1_options, static_specs): Likewise.
1924 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
1925 * passes.def: Likewise.
1926 * toplev.c (compile_file, process_options): Likewise.
1927 * tree-inline.c (copy_tree_r): Likewise.
1928 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
1929 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
1930 (build_constant_desc, output_constant_def_contents): Likewise.
1931 (categorize_decl_for_section): Likewise.
1932 * tree-mudflap.c: Removed.
1933 * tree-mudflap.h: Removed.
1934 * tree-nomudflap.c: Removed.
1935 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
1936 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
1937 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
1938 * config/sol2.h (MFLIB_SPEC): Likewise.
1939 * doc/install.texi: Remove mudflap references.
1940 * doc/passes.texi: Similarly.
1941 * doc/sourcebuild.texi: Similarly.
1942 * doc/invoke.texi: Remove mudlfap related options.
1944 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1946 PR rtl-optimization/58759
1947 * lra-constraints.c (lra_constraints): Remove wrong condition to
1948 remove insn setting up an equivalent pseudo.
1950 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1952 * config/rs6000/rs6000-protos.h
1953 (rs6000_secondary_memory_needed_mode): New prototype.
1954 * config/rs6000/rs6000.c: Include ira.h.
1955 (TARGET_LRA_P): Redefine.
1956 (rs6000_legitimate_offset_address_p): Call
1957 legitimate_constant_pool_address_p in strict mode for LRA.
1958 (rs6000_legitimate_address_p): Ditto.
1959 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
1960 (rs6000_emit_move): Add LRA version of code to generate load/store
1962 (rs6000_secondary_memory_needed_mode): New.
1963 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
1964 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
1965 constants, memory, and FP registers.
1966 (rs6000_lra_p): New.
1967 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
1968 * config/rs6000/rs6000.opt (mlra): New option.
1969 * lra-spills.c (lra_final_code_change): Remove useless move insns.
1971 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
1973 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
1974 has_single_use () and not do the conversion if has_single_use ()
1975 returns false for the multiplication result.
1977 2013-10-25 David Malcolm <dmalcolm@redhat.com>
1979 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
1980 to "NODE", since this works on a "tree", not an
1982 (CONSTANT_CLASS_P): Likewise.
1985 (INDIRECT_REF_P): Likewise.
1986 (REFERENCE_CLASS_P): Likewise.
1987 (COMPARISON_CLASS_P): Likewise.
1988 (UNARY_CLASS_P): Likewise.
1989 (BINARY_CLASS_P): Likewise.
1990 (STATEMENT_CLASS_P): Likewise.
1991 (VL_EXP_CLASS_P): Likewise.
1992 (EXPRESSION_CLASS_P): Likewise.
1993 (IS_TYPE_OR_DECL_P): Likewise.
1995 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
1997 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
1998 ADDR_EXPR in the defining statement.
2000 2013-10-25 Richard Biener <rguenther@suse.de>
2002 PR tree-optimization/58626
2003 * tree-loop-distribution.c (enum rdg_dep_type): Remove
2004 anti_dd, output_dd and input_dd.
2005 (struct rdg_edge): Remove level and relation members.
2006 (RDGE_LEVEL, RDGE_RELATION): Remove.
2007 (dot_rdg_1): Adjust.
2008 (create_rdg_edge_for_ddr): Remove.
2009 (create_rdg_edges_for_scalar): Adjust.
2010 (create_edge_for_control_dependence): Likewise.
2011 (create_rdg_edges): Split into ...
2012 (create_rdg_flow_edges): ... this
2013 (create_rdg_cd_edges): ... and this.
2015 (build_rdg): Likewise, do not compute data dependences or
2017 (pg_add_dependence_edges): New function.
2019 (distribute_loop): First apply all non-dependence based
2020 partition mergings. Then compute dependences between partitions
2021 and merge and order partitions according to them.
2023 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2025 PR rtl-optimization/58831
2026 * alias.c (init_alias_analysis): At the beginning of each iteration,
2027 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
2029 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2031 * recog.c (search_ofs): New static variable moved from...
2032 (peep2_find_free_register): ...here.
2033 (peephole2_optimize): Initialize it.
2035 2013-10-25 Tobias Burnus <burnus@net-b.de>
2037 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
2039 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
2041 * config/i386/i386.h (TARGET_MPX): New define.
2042 (TARGET_MPX_P): Ditto.
2044 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2046 * config/i386/constraints.md (B): New.
2049 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
2050 * config/i386/i386-modes.def (BND32): New.
2052 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
2053 * config/i386/i386.c (isa_opts): Add mmpx.
2054 (regclass_map): Add bound registers.
2055 (dbx_register_map): Likewise.
2056 (dbx64_register_map): Likewise.
2057 (svr4_dbx_register_map): Likewise.
2059 (ix86_option_override_internal): Support MPX ISA.
2060 (ix86_conditional_register_usage): Support bound registers.
2061 (print_reg): Likewise.
2062 (ix86_code_end): Add MPX bnd prefix.
2063 (output_set_got): Likewise.
2064 (ix86_output_call_insn): Likewise.
2065 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
2066 (ix86_print_operand_punct_valid_p): Likewise.
2067 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
2069 (ix86_class_likely_spilled_p): Add bound regs support.
2070 (ix86_hard_regno_mode_ok): Likewise.
2071 (x86_order_regs_for_local_alloc): Likewise.
2072 (ix86_bnd_prefixed_insn_p): New.
2073 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
2074 (FIXED_REGISTERS): Add bound registers.
2075 (CALL_USED_REGISTERS): Likewise.
2076 (REG_ALLOC_ORDER): Likewise.
2077 (HARD_REGNO_NREGS): Likewise.
2079 (VALID_BND_REG_MODE): New.
2080 (FIRST_BND_REG): New.
2081 (LAST_BND_REG): New.
2082 (reg_class): Add BND_REGS.
2083 (REG_CLASS_NAMES): Likewise.
2084 (REG_CLASS_CONTENTS): Likewise.
2086 (ANY_BND_REG_P): New.
2088 (HI_REGISTER_NAMES): Add bound registers.
2089 * config/i386/i386.md (UNSPEC_BNDMK): New.
2090 (UNSPEC_BNDMK_ADDR): New.
2091 (UNSPEC_BNDSTX): New.
2092 (UNSPEC_BNDLDX): New.
2093 (UNSPEC_BNDLDX_ADDR): New.
2094 (UNSPEC_BNDCL): New.
2095 (UNSPEC_BNDCU): New.
2096 (UNSPEC_BNDCN): New.
2097 (UNSPEC_MPX_FENCE): New.
2100 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
2101 (length_immediate): Likewise.
2102 (prefix_0f): Likewise.
2104 (prefix_rep): Check for bnd prefix.
2105 (length_nobnd): New.
2106 (length): Use length_nobnd if specified.
2111 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
2114 (simple_return_internal): Likewise.
2115 (simple_return_pop_internal): Likewise.
2116 (*indirect_jump): Add MPX bnd prefix.
2117 (*tablejump_1): Likewise.
2118 (simple_return_internal_long): Likewise.
2119 (simple_return_indirect_internal): Likewise.
2123 (*mov<mode>_internal_mpx): New.
2124 (<mode>_<bndcheck>): New.
2125 (*<mode>_<bndcheck>): New.
2130 * config/i386/predicates.md (lea_address_operand): Rename to...
2131 (address_no_seg_operand): ... this.
2132 (address_mpx_no_base_operand): New.
2133 (address_mpx_no_index_operand): New.
2134 (bnd_mem_operator): New.
2135 * config/i386/i386.opt (mmpx): New.
2136 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
2137 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
2139 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2141 * mode-classes.def (MODE_POINTER_BOUNDS): New.
2142 * tree.def (POINTER_BOUNDS_TYPE): New.
2143 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
2144 (POINTER_BOUNDS_MODE): New.
2145 (make_pointer_bounds_mode): New.
2146 * machmode.h (POINTER_BOUNDS_MODE_P): New.
2147 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
2148 (layout_type): Support POINTER_BOUNDS_TYPE.
2149 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
2150 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
2151 (type_contains_placeholder_1): Likewise.
2152 * tree.h (POINTER_BOUNDS_TYPE_P): New.
2153 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
2154 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
2156 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2158 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
2159 mode of address computation when calling memory_address_addr_space.
2161 2013-08-24 Richard Henderson <rth@twiddle.net>
2164 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
2165 instead of create_convert_operand_to.
2166 (maybe_emit_sync_lock_test_and_set): Likewise.
2167 (expand_atomic_compare_and_swap): Likewise.
2168 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
2170 2013-08-24 Sriraman Tallam <tmsriram@google.com>
2172 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
2173 Change param_is to use the struct and not the pointer to the struct.
2175 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
2177 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
2178 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2179 Move to targhooks.c.
2180 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
2181 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2182 Relocate from gimplify.c.
2183 * targhooks.h: Add 2 prototypes.
2184 * tree.h: Delete 2 prototypes.
2186 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2188 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
2189 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
2191 2013-10-24 Cong Hou <congh@google.com>
2193 * convert.c (convert_to_real): Guard those unsafe math function
2194 convertions with flag_unsafe_math_optimizations. Handle sqrt()
2197 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
2200 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
2202 (cgraph_create_edge): Use the new argument.
2203 (cgraph_create_indirect_edge): Likewise.
2205 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
2207 * gcc/config/arc/arc.c (arc_ccfsm_post_advance): Also handle
2209 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
2210 changes statep->state.
2212 2013-10-24 Tobias Burnus <burnus@net-b.de>
2215 * tree-cfg.c (replace_loop_annotate): New function.
2216 (execute_build_cfg): Call it.
2217 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
2218 * internal-fn.c (expand_ANNOTATE): New function.
2219 * internal-fn.def (ANNOTATE): Define as new internal function.
2220 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
2221 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
2222 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
2223 * doc/extend.texi (Pragmas): Document #pragma ivdep.
2224 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
2226 2013-10-17 Ian Bolton <ian.bolton@arm.com>
2227 Marcus Shawcroft <marcus.shawcroft@arm.com>
2229 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
2230 Special case reload SP+C into none GENERAL_REGS.
2232 2013-10-24 Michael Matz <matz@suse.de>
2234 * gengtype.c (is_file_equal): Check that files will be same length.
2236 2013-10-25 Christian Bruel <christian.bruel@st.com>
2238 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
2239 * gcc/config/sh/t-sh (sh-mem.o): New rule.
2240 * gcc/config/sh/sh-mem.cc (expand_block_move): Moved here.
2241 (sh_expand_cmpstr): New function.
2242 * gcc/config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
2243 * gcc/config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
2244 * gcc/config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
2245 (rotlhi3_8): Rename.
2247 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2249 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
2250 * Makefile.in (CXXDEPMODE): Assign and change users.
2251 (CCDEPMODE): Delete.
2252 * configure: Regenerate.
2254 2013-10-23 David Malcolm <dmalcolm@redhat.com>
2256 * gengtype-parse.c (require_without_advance): New.
2257 (type): For GTY-marked types that are not GTY((user)), parse any
2258 base classes, requiring them to be single-inheritance, and not
2259 be templates. For non-GTY-marked types and GTY((user)),
2260 continue to skip over any C++ inheritance specification.
2261 * gengtype-state.c (state_writer::write_state_struct_type):
2262 Write base class of type (if any).
2263 (read_state_struct_type): Read base class of type (if any).
2264 * gengtype.c (new_structure): Add a "base_class" parameter.
2265 (create_optional_field_): Update for new parameter to new_structure.
2266 (adjust_field_rtx_def): Likewise.
2267 (adjust_field_tree_exp): Likewise.
2268 * gengtype.h (struct type): Add "base_class" field to the s
2270 (new_structure): Add "base" parameter.
2272 2013-10-23 Sriraman Tallam <tmsriram@google.com>
2275 * config/i386/i386.c (ix86_option_override_internal):
2276 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
2277 (ix86_valid_target_attribute_tree):
2278 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
2279 Change TARGET_SSE to TARGET_SSE_P (opts->...)
2281 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2283 * tree-ssa-loop.h: Remove include files.
2284 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2285 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
2286 * graphite-clast-to-gimple.c: Likewise.
2287 * graphite-scop-detection.c: Likewise.
2288 * graphite-sese-to-poly.c: Likewise.
2289 * ipa-inline-analysis.c: Likewise.
2290 * ipa-pure-const.c: Likewise.
2291 * loop-init.c: Likewise.
2292 * passes.c: Likewise.
2293 * predict.c: Likewise.
2294 * tree-cfg.c: Likewise.
2295 * tree-cfgcleanup.c: Likewise.
2296 * tree-chrec.c: Likewise.
2297 * tree-data-ref.c: Likewise.
2298 * tree-loop-distribution.c: Likewise.
2299 * tree-parloops.c: Likewise.
2300 * tree-predcom.c: Likewise.
2301 * tree-scalar-evolution.c: Likewise.
2302 * tree-ssa-address.c: Likewise.
2303 * tree-ssa.c: Likewise.
2304 * tree-ssa-dce.c: Likewise.
2305 * tree-ssa-loop.c: Likewise.
2306 * tree-ssa-loop-im.c: Likewise.
2307 * tree-ssa-loop-ivcanon.c: Likewise.
2308 * tree-ssa-loop-ivopts.c: Likewise.
2309 * tree-ssa-loop-manip.c: Likewise.
2310 * tree-ssa-loop-niter.c: Likewise.
2311 * tree-ssa-loop-prefetch.c: Likewise.
2312 * tree-ssa-loop-unswitch.c: Likewise.
2313 * tree-ssa-reassoc.c: Likewise.
2314 * tree-vect-data-refs.c: Likewise.
2315 * tree-vect-loop.c: Likewise.
2316 * tree-vect-loop-manip.c: Likewise.
2317 * tree-vectorizer.c: Likewise.
2318 * tree-vect-stmts.c: Likewise.
2319 * tree-vrp.c: Likewise.
2321 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2323 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
2325 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2327 PR tree-optimization/58775
2328 PR tree-optimization/58791
2329 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
2330 (insert_stmt_after): Rewritten, don't move the stmt, but really
2332 (get_stmt_uid_with_default): Remove.
2333 (build_and_add_sum): Use insert_stmt_after and
2334 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
2335 (update_range_test): Set uid on stmts added by
2336 force_gimple_operand_gsi. Don't immediately modify statements
2337 in inter-bb optimization, just update oe->op values.
2338 (optimize_range_tests): Return bool whether any changed have been made.
2339 (update_ops): New function.
2340 (struct inter_bb_range_test_entry): New type.
2341 (maybe_optimize_range_tests): Perform statement changes here.
2342 (not_dominated_by, appears_later_in_bb, get_def_stmt,
2343 ensure_ops_are_available): Remove.
2344 (find_insert_point): Rewritten.
2345 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
2346 return LHS of the (new resp. old) stmt. Don't call
2347 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
2348 instead of last, move new stmt at the right place.
2349 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
2350 (negate_value): Likewise. Set uids.
2351 (break_up_subtract_bb): Initialize uids.
2352 (reassociate_bb): Adjust rewrite_expr_tree caller.
2353 (do_reassoc): Don't call renumber_gimple_stmt_uids.
2355 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
2358 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
2359 TARGET_32BIT final condition.
2360 (mulsi3_internal2 and splitter): Same.
2362 2013-10-23 Jeff Law <law@redhat.com>
2364 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
2365 through joiner blocks with abnormal outgoing edges.
2367 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
2368 Add parameter JOINERS, to allow/disallow threading through joiner
2370 (thread_block): New. Call thread_block_1.
2371 (mark_threaded_blocks): Remove code to filter out certain cases
2372 of threading through joiner blocks.
2373 (thread_through_all_blocks): Document how we can have a dangling
2374 edge AUX field and clear it.
2376 2013-10-23 Ian Lance Taylor <iant@google.com>
2378 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
2379 (C++ Dialect Options): Likewise.
2380 (Optimize Options): Likewise.
2382 2013-10-23 Tom de Vries <tom@codesourcery.com>
2384 PR tree-optimization/58805
2385 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
2387 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2389 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
2390 sequence based on get_range_info returned range.
2392 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2394 * tree-ssa.h: Remove all #include's
2395 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2396 * alias.c: Move required includes from tree-ssa.h.
2398 * builtins.c: Likewise.
2399 * calls.c: Likewise.
2400 * cfgexpand.c: Likewise.
2401 * cfghooks.c: Likewise.
2402 * cfgloop.c: Likewise.
2403 * cfgloopmanip.c: Likewise.
2404 * cgraph.c: Likewise.
2405 * cgraphbuild.c: Likewise.
2406 * cgraphclones.c: Likewise.
2407 * cgraphunit.c: Likewise.
2409 * except.c: Likewise.
2411 * final.c: Likewise.
2412 * fold-const.c: Likewise.
2413 * ggc-page.c: Likewise.
2414 * gimple-builder.c: Likewise.
2415 * gimple-fold.c: Likewise.
2416 * gimple-iterator.c: Likewise.
2417 * gimple-low.c: Likewise.
2418 * gimple-pretty-print.c: Likewise.
2419 * gimple-ssa-strength-reduction.c: Likewise.
2420 * gimple-streamer-in.c: Likewise.
2421 * gimple-streamer-out.c: Likewise.
2422 * gimplify.c: Likewise.
2423 * graphite-blocking.c: Likewise.
2424 * graphite-clast-to-gimple.c: Likewise.
2425 * graphite-dependences.c: Likewise.
2426 * graphite-interchange.c: Likewise.
2427 * graphite-optimize-isl.c: Likewise.
2428 * graphite-poly.c: Likewise.
2429 * graphite-scop-detection.c: Likewise.
2430 * graphite-sese-to-poly.c: Likewise.
2431 * graphite.c: Likewise.
2432 * ipa-cp.c: Likewise.
2433 * ipa-inline-analysis.c: Likewise.
2434 * ipa-inline-transform.c: Likewise.
2435 * ipa-inline.c: Likewise.
2436 * ipa-prop.c: Likewise.
2437 * ipa-pure-const.c: Likewise.
2438 * ipa-reference.c: Likewise.
2439 * ipa-split.c: Likewise.
2440 * ipa-utils.c: Likewise.
2441 * loop-init.c: Likewise.
2442 * lto-cgraph.c: Likewise.
2443 * lto-section-in.c: Likewise.
2444 * lto-section-out.c: Likewise.
2445 * lto-streamer-in.c: Likewise.
2446 * lto-streamer-out.c: Likewise.
2447 * lto-streamer.c: Likewise.
2448 * omp-low.c: Likewise.
2449 * passes.c: Likewise.
2450 * predict.c: Likewise.
2451 * print-tree.c: Likewise.
2452 * profile.c: Likewise.
2454 * targhooks.c: Likewise.
2455 * tracer.c: Likewise.
2456 * trans-mem.c: Likewise.
2457 * tree-call-cdce.c: Likewise.
2458 * tree-cfg.c: Likewise.
2459 * tree-cfgcleanup.c: Likewise.
2460 * tree-chrec.c: Likewise.
2461 * tree-complex.c: Likewise.
2462 * tree-data-ref.c: Likewise.
2463 * tree-dfa.c: Likewise.
2464 * tree-eh.c: Likewise.
2465 * tree-emutls.c: Likewise.
2466 * tree-if-conv.c: Likewise.
2467 * tree-inline.c: Likewise.
2468 * tree-into-ssa.c: Likewise.
2469 * tree-loop-distribution.c: Likewise.
2470 * tree-mudflap.c: Likewise.
2471 * tree-nested.c: Likewise.
2472 * tree-nrv.c: Likewise.
2473 * tree-object-size.c: Likewise.
2474 * tree-outof-ssa.c: Likewise.
2475 * tree-parloops.c: Likewise.
2476 * tree-phinodes.c: Likewise.
2477 * tree-predcom.c: Likewise.
2478 * tree-pretty-print.c: Likewise.
2479 * tree-profile.c: Likewise.
2480 * tree-scalar-evolution.c: Likewise.
2481 * tree-sra.c: Likewise.
2482 * tree-ssa-address.c: Likewise.
2483 * tree-ssa-alias.c: Likewise.
2484 * tree-ssa-ccp.c: Likewise.
2485 * tree-ssa-coalesce.c: Likewise.
2486 * tree-ssa-copy.c: Likewise.
2487 * tree-ssa-copyrename.c: Likewise.
2488 * tree-ssa-dce.c: Likewise.
2489 * tree-ssa-dom.c: Likewise.
2490 * tree-ssa-dse.c: Likewise.
2491 * tree-ssa-forwprop.c: Likewise.
2492 * tree-ssa-ifcombine.c: Likewise.
2493 * tree-ssa-live.c: Likewise.
2494 * tree-ssa-loop-ch.c: Likewise.
2495 * tree-ssa-loop-im.c: Likewise.
2496 * tree-ssa-loop-ivcanon.c: Likewise.
2497 * tree-ssa-loop-ivopts.c: Likewise.
2498 * tree-ssa-loop-manip.c: Likewise.
2499 * tree-ssa-loop-niter.c: Likewise.
2500 * tree-ssa-loop-prefetch.c: Likewise.
2501 * tree-ssa-loop-unswitch.c: Likewise.
2502 * tree-ssa-loop.c: Likewise.
2503 * tree-ssa-math-opts.c: Likewise.
2504 * tree-ssa-operands.c: Likewise.
2505 * tree-ssa-phiopt.c: Likewise.
2506 * tree-ssa-phiprop.c: Likewise.
2507 * tree-ssa-pre.c: Likewise.
2508 * tree-ssa-propagate.c: Likewise.
2509 * tree-ssa-reassoc.c: Likewise.
2510 * tree-ssa-sccvn.c: Likewise.
2511 * tree-ssa-sink.c: Likewise.
2512 * tree-ssa-strlen.c: Likewise.
2513 * tree-ssa-structalias.c: Likewise.
2514 * tree-ssa-tail-merge.c: Likewise.
2515 * tree-ssa-ter.c: Likewise.
2516 * tree-ssa-threadedge.c: Likewise.
2517 * tree-ssa-threadupdate.c: Likewise.
2518 * tree-ssa-uncprop.c: Likewise.
2519 * tree-ssa-uninit.c: Likewise.
2520 * tree-ssa.c: Likewise.
2521 * tree-ssanames.c: Likewise.
2522 * tree-stdarg.c: Likewise.
2523 * tree-streamer-in.c: Likewise.
2524 * tree-switch-conversion.c: Likewise.
2525 * tree-tailcall.c: Likewise.
2526 * tree-vect-data-refs.c: Likewise.
2527 * tree-vect-generic.c: Likewise.
2528 * tree-vect-loop-manip.c: Likewise.
2529 * tree-vect-loop.c: Likewise.
2530 * tree-vect-patterns.c: Likewise.
2531 * tree-vect-slp.c: Likewise.
2532 * tree-vect-stmts.c: Likewise.
2533 * tree-vectorizer.c: Likewise.
2534 * tree-vrp.c: Likewise.
2537 * value-prof.c: Likewise.
2538 * var-tracking.c: Likewise.
2539 * varpool.c: Likewise.
2540 * vtable-verify.c: Likewise.
2542 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2544 * config/tilegx/tilegx.c: Include "tree.h".
2546 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2548 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
2549 the info, not after it.
2550 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
2551 dump_ssaname_info call.
2552 (pp_gimple_stmt_1): Adjust caller.
2553 (dump_phi_nodes): Likewise. Don't print "# " here.
2555 2013-10-22 Jan Hubicka <jh@suse.cz>
2557 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
2559 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
2560 * i386.c (expand_small_movmem_or_setmem): New function.
2561 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
2563 (alg_usable_p): Add support for value ranges; cleanup.
2564 (ix86_expand_set_or_movmem): Add support for misaligned moves.
2566 2013-10-22 Sterling Augustine <saugustine@google.com>
2568 * doc/invoke.texi: Document -ggnu-pubnames.
2569 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
2571 * dwarf2out.c: Include gdb/gdb-index.h.
2572 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
2573 debug_generate_pub_sections.
2574 (is_java, output_pubtables, output_pubname): New functions.
2575 (include_pubname_in_output): Handle debug_generate_pub_sections at
2577 (size_of_pubnames): Use new local space_for_flags based on
2578 debug_generate_pub_sections.
2579 (output_pubnames): Unify pubnames and pubtypes output logic.
2580 Genericize comments. Call output_pubname.
2581 (dwarf2out_finish): Move logic to output_pubnames and call it.
2583 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
2586 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
2587 Remove CCCmode handling.
2588 <case LTU>: Return 'c' suffix for CCCmode.
2589 <case GEU>: Return 'nc' suffix for CCCmode.
2590 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
2591 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
2592 (*sub<mode>3_cc_overflow): Ditto.
2593 (*subsi3_zext_cc_overflow): Ditto.
2595 2013-10-22 Steve Ellcey <sellcey@mips.com>
2597 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
2598 (AND (NOT OP1) (NOT OP2)).
2600 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
2602 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2603 meaning of merge-high and merge-low masks for little endian; avoid
2604 use of vector-pack masks for little endian for mismatched modes.
2606 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2608 * config/tilepro/tilepro.c: Include "tree.h".
2610 2013-10-22 Andreas Schwab <schwab@suse.de>
2612 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
2615 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
2617 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
2618 [Ed Smith-Rowland]): New entries.
2619 ([Stephen M. Webb]): Update.
2621 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
2623 * tree-ssa-ter.h: Remove duplicate copy of file contents.
2625 2013-10-21 Marek Polacek <polacek@redhat.com>
2628 * fold-const.c (fold_range_test): Return 0 if the type is not
2631 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
2633 * system.h: Move hwint.h include further down.
2634 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
2635 gcc_checking_asserts.
2636 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
2638 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2640 Fix volatile issues in optimize_bit_field_compare.
2641 * fold-const.c (optimize_bit_field_compare): Bail out if
2642 lvolatilep or rvolatilep.
2644 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2646 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
2647 and get_inner_reference returning different pmode for non-volatile
2648 bit-field members dependent on flag_strict_volatile_bitfields.
2649 * stor-layout.c (layout_decl): Remove special handling of
2650 flag_strict_volatile_bitfields.
2651 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
2652 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
2654 2013-10-21 Paulo Matos <pmatos@broadcom.com>
2656 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
2659 2013-10-21 Jeff Law <law@redhat.com>
2661 * tree-ssa-threadedge.c (thread_through_normal_block): New
2662 argument VISITED. Remove VISISTED as a local variable. When we
2663 have a threadable jump, verify the destination of the jump has not
2665 (thread_across_edge): Allocate VISITED bitmap once at function
2666 scope and use it throughout. Make sure to set appropriate bits in
2667 VISITED for E (start of jump thread path).
2668 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
2669 through a joiner if any edge on the path has a recorded jump thread.
2671 2013-10-21 Ian Lance Taylor <iant@google.com>
2673 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
2674 don't imply that attributes can solve all problems.
2675 (Directory Options): Fix typo.
2677 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2679 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
2680 extend and extend_add.
2682 2013-10-21 Richard Biener <rguenther@suse.de>
2684 PR tree-optimization/58794
2685 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
2686 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
2688 2013-10-21 Richard Biener <rguenther@suse.de>
2691 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
2692 to (T) X for sign-changing conversions (or no conversion).
2694 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
2696 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
2698 2013-10-20 Jan Hubicka <jh@suse.cz>
2700 * config/i386/i386-tune.def: Add comment; organize into categories
2702 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2704 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
2705 argument. Update function comment.
2706 (expand_set_or_movmem_via_rep): New function combining
2707 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
2708 (expand_movmem_via_rep_mov): Remove.
2709 expand_setmem_via_rep_stos): Remove.
2710 (expand_movmem_epilogue): Update calls correspondingly.
2711 (expand_setmem_epilogue_via_loop): Likewise.
2713 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
2714 (expand_set_or_movmem_prologue): New function combining
2715 expand_movmem_prologue and expand_setmem_prologue.
2716 (expand_movmem_prologue): Remove.
2717 (expand_setmem_prologue): Remove.
2718 (expand_set_or_movmem_constant_prologue): New function combining
2719 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
2720 (expand_constant_movmem_prologue): Remove.
2721 (expand_constant_setmem_prologue): Remove.
2722 (promote_duplicated_reg): Allow vector-const0 value.
2723 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
2724 and ix86_expand_setmem.
2725 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
2726 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
2728 2013-10-21 Diego Novillo <dnovillo@google.com>
2730 * asan.c: Include tree.h
2731 * bb-reorder.c: Likewise.
2732 * cfgcleanup.c: Likewise.
2733 * cfgloopmanip.c: Likewise.
2734 * data-streamer-in.c: Likewise.
2735 * data-streamer-out.c: Likewise.
2736 * data-streamer.c: Likewise.
2737 * dwarf2cfi.c: Likewise.
2738 * graphite-blocking.c: Likewise.
2739 * graphite-clast-to-gimple.c: Likewise.
2740 * graphite-dependences.c: Likewise.
2741 * graphite-interchange.c: Likewise.
2742 * graphite-optimize-isl.c: Likewise.
2743 * graphite-poly.c: Likewise.
2744 * graphite-scop-detection.c: Likewise.
2745 * graphite-sese-to-poly.c: Likewise.
2746 * graphite.c: Likewise.
2747 * ipa-devirt.c: Likewise.
2748 * ipa-profile.c: Likewise.
2751 * loop-init.c: Likewise.
2752 * loop-unroll.c: Likewise.
2753 * lower-subreg.c: Likewise.
2754 * lto/lto-object.c: Likewise.
2755 * recog.c: Likewise.
2756 * reginfo.c: Likewise.
2757 * tree-loop-distribution.c: Likewise.
2758 * tree-parloops.c: Likewise.
2759 * tree-ssa-strlen.c: Likewise.
2760 * tree-streamer.c: Likewise.
2761 * value-prof.c: Likewise.
2762 * target-globals.c: Likewise.
2763 * expr.h: Include tree-core.h instead of tree.h.
2764 * gimple.h: Likewise.
2765 * ipa-prop.h: Likewise.
2766 * ipa-utils.h: Likewise.
2767 * lto-streamer.h: Likewise.
2768 * streamer-hooks.h: Likewise.
2769 * ipa-reference.h: Include cgraph.h instead of tree.h.
2770 * cgraph.h: Include basic-block.h instead of tree.h.
2771 * tree-streamer.h: Do not include tree.h.
2772 * genattrtab.c (write_header): Generate inclusion of
2774 * genautomata.c (main): Likewise.
2775 * genemit.c: Likewise.
2776 * genopinit.c: Likewise.
2777 * genoutput.c (output_prologue): Likewise.
2778 * genpeep.c: Likewise.
2780 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2782 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
2784 (vec_unpacku_hi_v8hi): Likewise.
2785 (vec_unpacku_lo_v16qi): Likewise.
2786 (vec_unpacku_lo_v8hi): Likewise.
2788 2013-10-20 Jan Hubicka <jh@suse.cz>
2790 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
2791 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
2792 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
2794 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
2796 * config/mips/mips.h (ISA_HAS_WSBH): Define.
2797 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
2799 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
2801 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
2804 * system.h: Undef m_slot.
2806 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2808 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
2809 all elements for both endian flavors.
2811 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2814 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
2815 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
2816 and SI_REG for 64bit SYSV ABI targets.
2818 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2820 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
2821 to multi_reg_return. Clarify that we are skipping USEs of multiple
2822 return registers. Use bool type where appropriate.
2824 2013-10-18 Jan Hubicka <jh@suse.cz>
2826 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
2828 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
2829 (X86_TUNE_PUSH_MEMORY): Likewise.
2830 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
2831 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
2832 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
2834 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
2835 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
2837 (ix86_option_override_internal): Update to use tune features instead
2840 2013-10-18 Cong Hou <congh@google.com>
2842 PR tree-optimization/58508
2843 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
2844 statement that contains data refs with zero-step.
2846 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2848 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
2849 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
2850 * gimple-low.c (gimple_check_call_arg,
2851 gimple_check_call_matching_types): Move to cgraph.c.
2852 * gimple-low.h: Remove prototype.
2853 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
2854 Relocate from gimple-low.c.
2855 * cgraph.h: Add prototype. Don't include basic-block.h.
2856 * gimplify.c: Add gimple-low to include list.
2857 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
2858 * tree-eh.c: Add gimple-low to include list.
2859 * tree-nested.c: Likewise.
2860 * cfgexpand.c: Add tree-ssa-address.h to include list.
2862 * gimple-fold.c: Likewise.
2863 * gimple-ssa-strength-reduction.c: Likewise.
2864 * trans-mem.c: Likewise.
2865 * tree-mudflap.c: Likewise.
2866 * tree-ssa-loop-ivopts.c: Likewise.
2867 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
2868 (degenerate_phi_result): Move to tree-phinodes.c.
2869 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
2870 * tree-ssa-threadedge.c: Likewise.
2871 * tree-vrp.c: Likewise.
2872 * tree-phinodes.c (degenerate_phi_result): Relocate here.
2873 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
2874 * tree-phinodes.h (degenerate_phi_result): Add prototype.
2875 * tree-ssa-copy.c: Include tree-ssa-dom.h.
2876 * tree-ssa-forwprop.c: Likewise.
2877 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
2878 pass_data_cleanup_cfg_post_optimizing,
2879 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
2880 * tree-optimize.c: Delete File.
2881 * graphite.c: Include tree-cfgcleanup.h.
2882 * passes.c: Likewise.
2883 * tree-cfg.c: Likewise.
2884 * tree-profile.c: Likewise.
2885 * tree-ssa-dse.c: Likewise.
2886 * tree-ssa-loop-ivcanon.c: Likewise.
2887 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
2888 * tree-outof-ssa.c: Include sbitmap.h.
2889 * tree-ssa-live.c: Likewise.
2890 * tree-ssa-propagate.c: Likewise.
2891 * tree-ssa-structalias.c: Likewise.
2892 * tree-stdarg.c: Likewise.
2893 * Makefile.in (OBJS): Delete tree-optimize.o.
2894 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
2895 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
2896 * varasm.c: Include basic-block.h.
2897 * cfgloop.h: Include function.h instead of basic-block.h
2898 (bb_loop_depth): Move to cfgloop.c.
2899 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
2901 2013-10-18 Teresa Johnson <tejohnson@google.com>
2903 * predict.c (probably_never_executed): Compare frequency-based
2904 count to number of training runs.
2905 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
2907 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2909 * config/arm/arm.c (cortexa9_extra_costs): New table.
2910 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
2912 2013-10-18 Jeff Law <law@redhat.com>
2914 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
2916 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
2917 (register_jump_thread): Add "registered_jump_thread" debug
2919 * dbgcnt.def (registered_jump_thread): New debug counter.
2921 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2923 * config/rs6000/rs6000.c: Include cgraph.h.
2925 2013-10-18 Teresa Johnson <tejohnson@google.com>
2927 * tree-ssa-tail-merge.c (replace_block_by): Update edge
2928 weights during merging.
2930 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2932 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
2933 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
2934 * tree-ssa-operands.h: Remove prototype.
2935 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
2936 * gimple.c (virtual_operand_p): Relocate from gimple.c.
2937 * gimple.h: Add prototype.
2938 * gimple-ssa.h: Include tree-ssa-operands.h.
2939 * tree-dump.c: Add tree-cfg.h to include list.
2940 * tree-ssa-alias.c: Add ipa-reference.h to include list.
2941 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
2942 * config/i386/i386.c: Don't include tree-flow.h.
2943 * config/rs6000/rs6000.c: Likewise.
2945 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2947 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
2949 2013-10-18 Richard Biener <rguenther@suse.de>
2951 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
2952 or TYPE_UNSIGNED of integral types.
2953 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
2954 NULL_TREE for zero-precision integral types.
2956 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
2958 * config/aarch64/arm_neon.h
2959 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
2961 2013-10-17 Sriraman Tallam <tmsriram@google.com>
2964 * opth-gen.awk: Define target_flags_explicit.
2966 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2968 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
2969 fields to the reg_addr array that describes the valid addressing
2970 mode for any register, general purpose registers, floating point
2971 registers, and Altivec registers.
2972 (FIRST_RELOAD_REG_CLASS): Likewise.
2973 (LAST_RELOAD_REG_CLASS): Likewise.
2974 (struct reload_reg_map_type): Likewise.
2975 (reload_reg_map_type): Likewise.
2976 (RELOAD_REG_VALID): Likewise.
2977 (RELOAD_REG_MULTIPLE): Likewise.
2978 (RELOAD_REG_INDEXED): Likewise.
2979 (RELOAD_REG_OFFSET): Likewise.
2980 (RELOAD_REG_PRE_INCDEC): Likewise.
2981 (RELOAD_REG_PRE_MODIFY): Likewise.
2982 (reg_addr): Likewise.
2983 (mode_supports_pre_incdec_p): New helper functions to say whether
2984 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
2985 (mode_supports_pre_modify_p): Likewise.
2986 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
2987 print the valid address mode bits for each mode.
2988 (rs6000_debug_print_mode): Likewise.
2989 (rs6000_debug_reg_global): Likewise.
2990 (rs6000_setup_reg_addr_masks): New function to set up the address
2991 mask bits for each type.
2992 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
2993 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
2994 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
2995 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
2996 PRE_MODIFY are supported.
2997 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
2998 registers, instead of {src,dest}_av_p.
2999 (rs6000_print_options_internal): Tweak the debug output slightly.
3001 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3003 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
3006 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
3008 * tree-flow.h (struct omp_region): Move to omp-low.c.
3009 Remove omp_ prototypes and variables.
3010 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
3011 (copy_var_decl): Relocate prototype from tree-flow.h.
3012 * gimple.c (copy_var_decl): Relocate from omp-low.c.
3013 * tree.h: Move prototype to omp-low.h.
3014 * omp-low.h: New File. Relocate prototypes here.
3015 * omp-low.c (struct omp_region): Make local here.
3016 (root_omp_region): Make static.
3017 (copy_var_decl) Move to gimple.c.
3018 (new_omp_region): Make static.
3019 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
3020 * tree-cfg.c: Include omp-low.h.
3021 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
3022 * gimplify.c: Include omp-low.h.
3023 * tree-parloops.c: Likewise.
3025 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3027 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
3028 operands are in memory, prefer to force non-matched operand 1 to
3031 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3034 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
3035 restrict TImode addresses to single indirect registers if both
3036 -mquad-memory and -mvsx-timode are used.
3037 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
3038 we should emit load/store quad. Remove using %y for quad memory
3041 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
3042 constraints to allow load/store quad on machines where TImode is
3043 not allowed in VSX registers. Use 'n' instead of 'F' constraint
3044 for TImode to load integer constants.
3046 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3048 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
3050 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
3052 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
3053 handling of STACK_REG.
3055 2013-10-17 Richard Biener <rguenther@suse.de>
3057 PR tree-optimization/58143
3058 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
3060 (rewrite_to_defined_overflow): Likewise.
3061 (move_computations_dom_walker::before_dom): Rewrite stmts
3062 with undefined signed overflow that are not always executed
3063 into unsigned arithmetic.
3065 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3068 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
3069 explicit isa flag to be an options variable, instead of using
3070 global_options_set. Remove define from rs6000.h.
3071 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
3073 * config/rs6000/rs6000.c (rs6000_option_override_internal):
3074 Initialize rs6000_isa_flags_explicit.
3075 (rs6000_function_specific_save): Add gcc_options* parameter, so
3076 that the powerpc builds after the 2013-10-15 changes.
3077 (rs6000_function_specific_restore): Likewise.
3079 2013-10-16 DJ Delorie <dj@redhat.com>
3081 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
3082 op is a REG before checking REGNO.
3083 (rl78_alloc_physical_registers): Verify pattern is a SET before
3086 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3088 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
3090 (vec_unpacks_lo_v4sf): Likewise.
3091 (vec_unpacks_float_hi_v4si): Likewise.
3092 (vec_unpacks_float_lo_v4si): Likewise.
3093 (vec_unpacku_float_hi_v4si): Likewise.
3094 (vec_unpacku_float_lo_v4si): Likewise.
3096 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3098 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
3099 (vsx_concat_v2sf): Likewise.
3101 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
3103 * config/aarch64/aarch64.md
3104 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
3106 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
3108 PR tree-optimization/58697
3109 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
3110 estimated_loop_iterations_int.
3111 (max_stmt_executions_int): Call get_max_loop_iterations_int.
3112 (get_max_loop_iterations_int): New. HWINT version of
3113 get_max_loop_iterations.
3114 * cfgloop.h: Add prototypes.
3115 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
3116 * loop-unroll.c (decide_peel_once_rolling): Call
3117 get_estimated_loop_iterations_int.
3118 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
3119 * tree-ssa-loop-niter.h: Tweak prototypes.
3121 2013-10-16 David Malcolm <dmalcolm@redhat.com>
3123 * gengtype-parse.c (struct_field_seq): Ignore access-control
3124 keywords ("public:" etc).
3126 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
3128 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
3129 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
3131 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
3133 * config/arm/arm.opt (mlra): New option.
3134 * config/arm/arm.c (arm_lra_p): New function.
3135 (TARGET_LRA_P): Define.
3137 2013-10-16 Paulo Matos <pmatos@broadcom.com>
3139 * tree-core.h (tree_code_name): Remove.
3140 * tree.h (get_tree_code_name): New prototype.
3141 * tree.c (tree_code_name): Make static.
3142 (get_tree_code_name): New function.
3143 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
3144 tree_class_check_failed, tree_range_check_failed,
3145 tree_not_class_check_failed, omp_clause_check_failed,
3146 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
3147 wrapper get_tree_code_name instead of calling tree_code_name directly.
3148 * tree-vrp.c (dump_asserts_for): Likewise.
3149 * tree-dump.c (dequeue_and_dump): Likewise.
3150 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
3151 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
3152 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
3153 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
3154 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
3155 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
3156 dump_gimple_omp_for): Likewise.
3157 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
3158 * tree-ssa-pre.c (print_pre_expr): Likewise.
3159 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
3160 * print-tree.c (print_node_brief, print_node): Likewise.
3161 * gimple.c (gimple_check_failed): Likewise.
3162 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
3163 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
3164 * config/mep/mep.c (mep_validate_vliw): Likewise.
3165 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
3166 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
3168 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
3170 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
3173 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
3175 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
3176 (MULTILIB_MATCHES): Add multilib for -march=v8.
3178 2013-10-15 Sriraman Tallam <tmsriram@google.com>
3181 * optc-save-gen.awk: Add extra parameter to the save and restore
3183 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
3184 * tree.c (build_optimization_node): New parameter. Add extra parameter
3185 to call to cl_optimization_save.
3186 (build_target_option_node): New parameter. Add extra parameter
3187 to call to cl_target_option_save.
3188 * tree.h (build_optimization_node): New parameter.
3189 (build_target_option_node): New parameter.
3190 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
3191 build_optimization_node and build_target_option_node.
3192 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
3193 (handle_pragma_push_options): Ditto.
3194 * toplev.c (process_options): Ditto.
3195 * opts.c (init_options_struct): Check for opts_set non-null.
3196 * target.def (target_option.save): New parameter.
3197 (target_option.restore): New parameter.
3198 * tm.texi: Generate.
3199 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
3200 (ix86_pragma_target_parse): Ditto.
3201 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
3203 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
3204 to build_optimization_node and build_target_option_node.
3205 (rs6000_valid_attribute_p): Ditto.
3206 (rs6000_pragma_target_parse): Ditto.
3207 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
3209 * config/i386/i386.h:
3210 TARGET_64BIT_P: New Macro
3211 TARGET_MMX_P: New Macro.
3212 TARGET_3DNOW_P: New Macro.
3213 TARGET_3DNOW_A_P: New Macro.
3214 TARGET_SSE_P: New Macro.
3215 TARGET_SSE2_P: New Macro.
3216 TARGET_SSE3_P: New Macro.
3217 TARGET_SSSE3_P: New Macro.
3218 TARGET_SSE4_1_P: New Macro.
3219 TARGET_SSE4_2_P: New Macro.
3220 TARGET_AVX_P: New Macro.
3221 TARGET_AVX2_P: New Macro.
3222 TARGET_AVX512F_P: New Macro.
3223 TARGET_AVX512PF_P: New Macro.
3224 TARGET_AVX512ER_P: New Macro.
3225 TARGET_AVX512CD_P: New Macro.
3226 TARGET_FMA_P: New Macro.
3227 TARGET_SSE4A_P: New Macro.
3228 TARGET_FMA4_P: New Macro.
3229 TARGET_XOP_P: New Macro.
3230 TARGET_LWP_P: New Macro.
3231 TARGET_ABM_P: New Macro.
3232 TARGET_BMI_P: New Macro.
3233 TARGET_BMI2_P: New Macro.
3234 TARGET_LZCNT_P: New Macro.
3235 TARGET_TBM_P: New Macro.
3236 TARGET_POPCNT_P: New Macro.
3237 TARGET_SAHF_P: New Macro.
3238 TARGET_MOVBE_P: New Macro.
3239 TARGET_CRC32_P: New Macro.
3240 TARGET_AES_P: New Macro.
3241 TARGET_PCLMUL_P: New Macro.
3242 TARGET_CMPXCHG16B_P: New Macro.
3243 TARGET_FSGSBASE_P: New Macro.
3244 TARGET_RDRND_P: New Macro.
3245 TARGET_F16C_P: New Macro.
3246 TARGET_RTM_P: New Macro.
3247 TARGET_HLE_P: New Macro.
3248 TARGET_RDSEED_P: New Macro.
3249 TARGET_PRFCHW_P: New Macro.
3250 TARGET_ADX_P: New Macro.
3251 TARGET_FXSR_P: New Macro.
3252 TARGET_XSAVE_P: New Macro.
3253 TARGET_XSAVEOPT_P: New Macro.
3254 TARGET_LP64_P: New Macro.
3255 TARGET_X32_P: New Macro.
3256 TARGET_FPMATH_DEFAULT_P: New Macro.
3257 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
3258 * config/i386/i386.c (ix86_option_override_internal): New parameters.
3260 Change ix86_tune_string to access opts->x_ix86_tune_string.
3261 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
3262 Change ix86_arch_string to access opts->x_ix86_arch_string.
3263 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
3264 Change ix86_pmode to access opts->x_ix86_pmode.
3265 Change ix86_abi to access opts->x_ix86_abi.
3266 Change ix86_cmodel to access opts->x_ix86_cmodel.
3267 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
3268 Change ix86_isa_flags_explicit to access
3269 opts->x_ix86_isa_flags_explicit.
3270 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
3271 Change ix86_regparm to access opts->x_ix86_regparm.
3272 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
3273 Change ix86_preferred_stack_boundary_arg to access
3274 opts->x_ix86_preferred_stack_boundary_arg.
3275 Change ix86_force_align_arg_pointer to access
3276 opts->x_ix86_force_align_arg_pointer.
3277 Change ix86_incoming_stack_boundar_arg to access
3278 opts->x_ix86_incoming_stack_boundar_arg.
3279 Change ix86_fpmath to access opts->x_ix86_fpmath.
3280 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
3281 Change ix86_recip_name to access opts->x_ix86_recip_name.
3282 Change ix86_stack_protector_guard to access
3283 opts->x_ix86_stack_protector_guard.
3284 Change ix86_tune_memcpy_strategy to access
3285 opts->x_ix86_tune_memcpy_strategy.
3286 Change ix86_tune_memset_strategy to access
3287 opts->x_ix86_tune_memset_strategy.
3288 Change global_options to access opts.
3289 Change global_options_set to access opts_set.
3290 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
3291 Change TARGET_MMX to TARGET_MMX_P (opts->...).
3292 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
3293 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
3294 Change TARGET_SSE to TARGET_SSE_P (opts->...).
3295 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
3296 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
3297 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
3298 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
3299 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
3300 Change TARGET_AVX to TARGET_AVX_P (opts->...).
3301 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
3302 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
3303 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
3304 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
3305 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
3306 Change TARGET_FMA to TARGET_FMA_P (opts->...).
3307 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
3308 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
3309 Change TARGET_XOP to TARGET_XOP_P (opts->...).
3310 Change TARGET_LWP to TARGET_LWP_P (opts->...).
3311 Change TARGET_ABM to TARGET_ABM_P (opts->...).
3312 Change TARGET_BMI to TARGET_BMI_P (opts->...).
3313 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
3314 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
3315 Change TARGET_TBM to TARGET_TBM_P (opts->...).
3316 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
3317 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
3318 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
3319 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
3320 Change TARGET_AES to TARGET_AES_P (opts->...).
3321 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
3322 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
3323 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
3324 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
3325 Change TARGET_F16C to TARGET_F16C_P (opts->...).
3326 Change TARGET_RTM to TARGET_RTM_P (opts->...).
3327 Change TARGET_HLE to TARGET_HLE_P (opts->...).
3328 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
3329 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
3330 Change TARGET_ADX to TARGET_ADX_P (opts->...).
3331 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
3332 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
3333 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
3334 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
3335 Change TARGET_X32 to TARGET_X32_P (opts->...).
3336 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
3337 Change TARGET_FLOAT_RETURNS_IN_80387 to
3338 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
3339 (ix86_function_specific_save): New parameter. Use opts-> fields
3340 to replace global fields.
3341 (ix86_function_specific_restore): Ditto.
3342 (ix86_valid_target_attribute_inner_p): New parameters.
3344 Fix call to ix86_handle_option and set_option.
3345 (ix86_valid_target_attribute_tree): New parameters.
3346 Change global_options to access opts.
3347 Change global_options_set to access opts_set.
3348 Fix call to ix86_valid_target_attribute_inner_p.
3349 Change ix86_tune_string to access opts->x_ix86_tune_string.
3350 Change ix86_arch_string to access opts->x_ix86_arch_string.
3351 Change ix86_fpmath to access opts->x_ix86_fpmath
3352 Fix call to ix86_option_override_internal.
3353 Fix call to ix86_add_new_builtins.
3354 Fix calls to build_optimization_node and build_target_option_node.
3355 (ix86_valid_target_attribute_p): Remove access to global_options.
3356 Use new gcc_options structure func_options.
3357 Fix call to ix86_valid_target_attribute_tree.
3358 Fix call to build_optimization_node.
3359 (get_builtin_code_for_version): Fix call to
3360 ix86_valid_target_attribute_tree.
3362 2013-10-15 David Malcolm <dmalcolm@redhat.com>
3364 * Makefile.in (PICFLAG): New.
3365 (enable_host_shared): New.
3366 (INTERNAL_CFLAGS): Use PICFLAG.
3367 (LIBIBERTY): Use pic build of libiberty.a if configured with
3368 --enable-host-shared.
3369 * configure.ac: Add --enable-host-shared, setting up new
3371 * configure: Regenerate.
3372 * doc/install.texi (--enable-shared): Add note contrasting it with ...
3373 (--enable-host-shared): New option.
3375 2013-10-15 Richard Biener <rguenther@suse.de>
3377 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
3378 for built-in functions.
3380 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
3382 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
3383 (optimize_range_tests_1): New function,
3384 extracted from optimize_range_tests.
3385 (optimize_range_tests_xor): Similarly.
3386 (optimize_range_tests_diff): New function.
3387 (optimize_range_tests): Use optimize_range_tests_1.
3389 2013-10-15 Cong Hou <congh@google.com>
3391 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
3392 requirement of the reduction pattern so that one operand of the
3393 reduction operation can come from outside of the loop.
3395 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3397 * config/arm/neon-schedgen.ml: Remove.
3398 * config/arm/cortex-a9-neon.md: Remove comment regarding
3401 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3403 * config/arm/types: Remove old neon types.
3405 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3407 * config/arm/cortex-a7.md
3408 (cortex_a7_neon_type): New.
3409 (cortex_a7_neon_mul): Update for new types.
3410 (cortex_a7_neon_mla): Likewise.
3411 (cortex_a7_neon): Likewise.
3413 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3415 * config/arm/cortex-a15-neon.md
3416 (cortex_a15_neon_type): New,
3418 (cortex_a15_neon_int_1): Remove.
3419 (cortex_a15_neon_int_2): Likewise.
3420 (cortex_a15_neon_int_3): Likewise.
3421 (cortex_a15_neon_int_4): Likewise.
3422 (cortex_a15_neon_int_5): Likewise.
3423 (cortex_a15_neon_vqneg_vqabs): Likewise.
3424 (cortex_a15_neon_vmov): Likewise.
3425 (cortex_a15_neon_vaba): Likewise.
3426 (cortex_a15_neon_vaba_qqq): Likewise.
3427 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3428 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3429 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3431 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3432 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3433 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
3434 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3435 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3436 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3437 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3438 (cortex_a15_neon_shift_1): Likewise.
3439 (cortex_a15_neon_shift_2): Likewise.
3440 (cortex_a15_neon_shift_3): Likewise.
3441 (cortex_a15_neon_vshl_ddd): Likewise.
3442 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3443 (cortex_a15_neon_vsra_vrsra): Likewise.
3444 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3445 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3446 (cortex_a15_neon_bp_3cycle): Likewise.
3447 (cortex_a15_neon_ldm_2): Likewise.
3448 (cortex_a15_neon_stm_2): Likewise.
3449 (cortex_a15_neon_mcr): Likewise.
3450 (cortex_a15_neon_mrc): Likewise.
3451 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3452 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3453 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3454 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3455 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3456 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3457 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3458 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3459 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3460 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3461 (cortex_a15_neon_bp_simple): Likewise.
3462 (cortex_a15_neon_bp_2cycle): Likewise.
3463 (cortex_a15_neon_bp_3cycle): Likewise.
3464 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3465 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3466 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3467 (cortex_a15_neon_vld2_4_regs): Likewise.
3468 (cortex_a15_neon_vld3_vld4): Likewise.
3469 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3470 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3471 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
3472 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
3473 (cortex_a15_neon_vst3_vst4): Rename to...
3474 (cortex_a15_neon_vst4): This, update for new attributes.
3475 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
3476 (cortex_a15_neon_vld3_vld4_lane): Likewise.
3477 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3478 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3479 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3480 (cortex_a15_neon_ldm_2): Likewise.
3481 (cortex_a15_neon_stm_2): Likewise.
3482 (cortex_a15_neon_mcr): Likewise.
3483 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3484 (cortex_a15_neon_mrc): Likewise.
3485 (cortex_a15_neon_mrrc): Likewise.
3487 (cortex_a15_neon_abd): New.
3488 (cortex_a15_neon_abd_q): Likewise.
3489 (cortex_a15_neon_aba): Likewise.
3490 (cortex_a15_neon_aba_q): Likewise.
3491 (cortex_a15_neon_acc): Likewise.
3492 (cortex_a15_neon_acc_q): Likewise.
3493 (cortex_a15_neon_arith_basic): Likewise.
3494 (cortex_a15_neon_arith_complex): Likewise.
3495 (cortex_a15_neon_multiply): Likewise.
3496 (cortex_a15_neon_multiply_q): Likewise.
3497 (cortex_a15_neon_mla): Likewise.
3498 (cortex_a15_neon_mla_q): Likewise.
3499 (cortex_a15_neon_sat_mla_long): Likewise.
3500 (cortex_a15_neon_shift_acc): Likewise.
3501 (cortex_a15_neon_shift_imm_basic): Likewise.
3502 (cortex_a15_neon_shift_imm_complex): Likewise.
3503 (cortex_a15_neon_shift_reg_basic): Likewise.
3504 (cortex_a15_neon_shift_reg_basic_q): Likewise.
3505 (cortex_a15_neon_shift_reg_complex): Likewise.
3506 (cortex_a15_neon_shift_reg_complex_q): Likewise.
3507 (cortex_a15_neon_fp_negabs): Likewise
3508 (cortex_a15_neon_fp_arith): Likewise
3509 (cortex_a15_neon_fp_arith_q): Likewise
3510 (cortex_a15_neon_fp_cvt_int): Likewise
3511 (cortex_a15_neon_fp_cvt_int_q): Likewise
3512 (cortex_a15_neon_fp_cvt_16): Likewise
3513 (cortex_a15_neon_fp_mul): Likewise
3514 (cortex_a15_neon_fp_mul_q): Likewise
3515 (cortex_a15_neon_fp_mla): Likewise
3516 (cortex_a15_neon_fp_mla_q): Likewise
3517 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
3518 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
3519 (cortex_a15_neon_bitops): Likewise.
3520 (cortex_a15_neon_bitops_q): Likewise.
3521 (cortex_a15_neon_from_gp): Likewise.
3522 (cortex_a15_neon_from_gp_q): Likewise.
3523 (cortex_a15_neon_tbl3_tbl4): Likewise.
3524 (cortex_a15_neon_zip_q): Likewise.
3525 (cortex_a15_neon_to_gp): Likewise.
3526 (cortex_a15_neon_load_a): Likewise.
3527 (cortex_a15_neon_load_b): Likewise.
3528 (cortex_a15_neon_load_c): Likewise.
3529 (cortex_a15_neon_load_d): Likewise.
3530 (cortex_a15_neon_load_e): Likewise.
3531 (cortex_a15_neon_load_f): Likewise.
3532 (cortex_a15_neon_store_a): Likewise.
3533 (cortex_a15_neon_store_b): Likewise.
3534 (cortex_a15_neon_store_c): Likewise.
3535 (cortex_a15_neon_store_d): Likewise.
3536 (cortex_a15_neon_store_e): Likewise.
3537 (cortex_a15_neon_store_f): Likewise.
3538 (cortex_a15_neon_store_g): Likewise.
3539 (cortex_a15_neon_store_h): Likewise.
3540 (cortex_a15_vfp_to_from_gp): Likewise.
3542 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3544 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
3546 (cortex_a9_neon_vshl_ddd): Remove.
3547 (cortex_a9_neon_vst3_vst4): Likewise.
3548 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3550 (cortex_a9_neon_bit_ops_q): New.
3552 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
3553 (cortex_a9_neon_int_2): Likewise.
3554 (cortex_a9_neon_int_3): Likewise.
3555 (cortex_a9_neon_int_4): Likewise.
3556 (cortex_a9_neon_int_5): Likewise.
3557 (cortex_a9_neon_vqneg_vqabs): Likewise.
3558 (cortex_a9_neon_vmov): Likewise.
3559 (cortex_a9_neon_vaba): Likewise.
3560 (cortex_a9_neon_vaba_qqq): Likewise.
3561 (cortex_a9_neon_shift_1): Likewise.
3562 (cortex_a9_neon_shift_2): Likewise.
3563 (cortex_a9_neon_shift_3): Likewise.
3564 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3565 (cortex_a9_neon_vsra_vrsra): Likewise.
3566 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3567 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3568 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3570 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3571 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3572 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3574 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3575 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3576 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3577 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3578 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3579 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3580 (cortex_a9_neon_fp_vsum): Likewise.
3581 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3582 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3583 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3584 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3585 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3586 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3587 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3588 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3589 (cortex_a9_neon_bp_simple): Likewise.
3590 (cortex_a9_neon_bp_2cycle): Likewise.
3591 (cortex_a9_neon_bp_3cycle): Likewise.
3592 (cortex_a9_neon_ldr): Likewise.
3593 (cortex_a9_neon_str): Likewise.
3594 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3595 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3596 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3597 (cortex_a9_neon_vld2_4_regs): Likewise.
3598 (cortex_a9_neon_vld3_vld4): Likewise.
3599 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3600 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3601 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3602 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3603 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3604 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3605 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3606 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3607 (cortex_a9_neon_mcr): Likewise.
3608 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3609 (cortex_a9_neon_mrc): Likewise.
3610 (cortex_a9_neon_mrrc): Likewise.
3612 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3614 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
3616 (cortex_a8_neon_vshl_ddd): Remove.
3617 (cortex_a8_neon_vst3_vst4): Likewise.
3618 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3620 (cortex_a8_neon_bit_ops_q): New.
3622 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
3623 (cortex_a8_neon_int_2): Likewise..
3624 (cortex_a8_neon_int_3): Likewise.
3625 (cortex_a8_neon_int_5): Likewise.
3626 (cortex_a8_neon_vqneg_vqabs): Likewise.
3627 (cortex_a8_neon_int_4): Likewise.
3628 (cortex_a8_neon_vaba): Likewise.
3629 (cortex_a8_neon_vaba_qqq): Likewise.
3630 (cortex_a8_neon_shift_1): Likewise.
3631 (cortex_a8_neon_shift_2): Likewise.
3632 (cortex_a8_neon_shift_3): Likewise.
3633 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3634 (cortex_a8_neon_vsra_vrsra): Likewise.
3635 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3636 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3637 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3639 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3640 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3641 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3643 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3644 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3645 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3646 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3647 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3648 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3649 (cortex_a8_neon_fp_vsum): Likewise.
3650 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3651 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3652 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3653 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3654 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3655 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3656 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3657 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3658 (cortex_a8_neon_bp_simple): Likewise.
3659 (cortex_a8_neon_bp_2cycle): Likewise.
3660 (cortex_a8_neon_bp_3cycle): Likewise.
3661 (cortex_a8_neon_ldr): Likewise.
3662 (cortex_a8_neon_str): Likewise.
3663 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3664 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3665 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3666 (cortex_a8_neon_vld2_4_regs): Likewise.
3667 (cortex_a8_neon_vld3_vld4): Likewise.
3668 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3669 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3670 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3671 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3672 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3673 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3674 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3675 (cortex_a8_neon_mcr): Likewise.
3676 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3677 (cortex_a8_neon_mrc): Likewise.
3678 (cortex_a8_neon_mrrc): Likewise.
3680 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3682 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
3684 * config/aarch64/aarch64-simd.md (neon_type): Remove.
3685 (aarch64_simd_dup<mode>): Add "type" attribute.
3686 (aarch64_dup_lane<mode>): Likewise.
3687 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
3688 (*aarch64_simd_mov<mode>): Likewise.
3689 (aarch64_simd_mov_from_<mode>low): Likewise.
3690 (aarch64_simd_mov_from_<mode>high): Likewise.
3691 (orn<mode>3): Likewise.
3692 (bic<mode>3): Likewise.
3693 (add<mode>3): Likewise.
3694 (sub<mode>3): Likewise.
3695 (mul<mode>3): Likewise.
3696 (*aarch64_mul3_elt<mode>): Likewise.
3697 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
3698 (*aarch64_mul3_elt_to_128df): Likewise.
3699 (*aarch64_mul3_elt_to_64v2df): Likewise.
3700 (neg<mode>2): Likewise.
3701 (abs<mode>2): Likewise.
3702 (abd<mode>_3): Likewise.
3703 (aba<mode>_3): Likewise.
3704 (fabd<mode>_3): Likewise.
3705 (*fabd_scalar<mode>3): Likewise.
3706 (and<mode>3): Likewise.
3707 (ior<mode>3): Likewise.
3708 (xor<mode>3): Likewise.
3709 (one_cmpl<mode>2): Likewise.
3710 (aarch64_simd_vec_set<mode>): Likewise.
3711 (aarch64_simd_lshr<mode>): Likewise.
3712 (aarch64_simd_ashr<mode>): Likewise.
3713 (aarch64_simd_imm_shl<mode>): Likewise.
3714 (aarch64_simd_reg_sshl<mode): Likewise.
3715 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
3716 (aarch64_simd_reg_shl<mode>_signed): Likewise.
3717 (aarch64_simd_vec_setv2di): Likewise.
3718 (aarch64_simd_vec_set<mode>): Likewise.
3719 (aarch64_mla<mode>): Likewise.
3720 (*aarch64_mla_elt<mode>): Likewise.
3721 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
3722 (aarch64_mls<mode>): Likewise.
3723 (*aarch64_mls_elt<mode>): Likewise.
3724 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
3725 (<su><maxmin><mode>3): Likewise.
3726 (move_lo_quad_<mode>): Likewise.
3727 (aarch64_simd_move_hi_quad_<mode>): Likewise.
3728 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
3729 (vec_pack_trunc_<mode>): Likewise.
3730 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
3731 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
3732 (*aarch64_<su>mlal_lo<mode>): Likewise.
3733 (*aarch64_<su>mlal_hi<mode>): Likewise.
3734 (*aarch64_<su>mlsl_lo<mode>): Likewise.
3735 (*aarch64_<su>mlsl_hi<mode>): Likewise.
3736 (*aarch64_<su>mlal<mode>): Likewise.
3737 (*aarch64_<su>mlsl<mode>): Likewise.
3738 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
3739 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
3740 (add<mode>3): Likewise.
3741 (sub<mode>3): Likewise.
3742 (mul<mode>3): Likewise.
3743 (div<mode>3): Likewise.
3744 (neg<mode>2): Likewise.
3745 (abs<mode>2): Likewise.
3746 (fma<mode>4): Likewise.
3747 (*aarch64_fma4_elt<mode>): Likewise.
3748 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
3749 (*aarch64_fma4_elt_to_128df): Likewise.
3750 (*aarch64_fma4_elt_to_64v2df): Likewise.
3751 (fnma<mode>4): Likewise.
3752 (*aarch64_fnma4_elt<mode>): Likewise.
3753 (*aarch64_fnma4_elt_<vswap_width_name><mode>
3754 (*aarch64_fnma4_elt_to_128df): Likewise.
3755 (*aarch64_fnma4_elt_to_64v2df): Likewise.
3756 (<frint_pattern><mode>2): Likewise.
3757 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
3758 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
3759 (vec_unpacks_lo_v4sf): Likewise.
3760 (aarch64_float_extend_lo_v2df): Likewise.
3761 (vec_unpacks_hi_v4sf): Likewise.
3762 (aarch64_float_truncate_lo_v2sf): Likewise.
3763 (aarch64_float_truncate_hi_v4sf): Likewise.
3764 (aarch64_vmls<mode>): Likewise.
3765 (<su><maxmin><mode>3): Likewise.
3766 (<maxmin_uns><mode>3): Likewise.
3767 (reduc_<sur>plus_<mode>): Likewise.
3768 (reduc_<sur>plus_v2di): Likewise.
3769 (reduc_<sur>plus_v2si): Likewise.
3770 (reduc_<sur>plus_<mode>): Likewise.
3771 (aarch64_addpv4sf): Likewise.
3772 (clz<mode>2): Likewise.
3773 (reduc_<maxmin_uns>_<mode>): Likewise.
3774 (reduc_<maxmin_uns>_v2di): Likewise.
3775 (reduc_<maxmin_uns>_v2si): Likewise.
3776 (reduc_<maxmin_uns>_<mode>): Likewise.
3777 (reduc_<maxmin_uns>_v4sf): Likewise.
3778 (aarch64_simd_bsl<mode>_internal): Likewise.
3779 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
3780 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
3781 (aarch64_get_lane<mode>): Likewise.
3782 (*aarch64_combinez<mode>): Likewise.
3783 (aarch64_combine<mode>): Likewise.
3784 (aarch64_simd_combine<mode>): Likewise.
3785 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
3786 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
3787 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
3788 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
3789 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
3790 (aarch64_<sur>h<addsub><mode>): Likewise.
3791 (aarch64_<sur><addsub>hn<mode>): Likewise.
3792 (aarch64_<sur><addsub>hn2<mode>): Likewise.
3793 (aarch64_pmul<mode>): Likewise.
3794 (aarch64_<su_optab><optab><mode>): Likewise.
3795 (aarch64_<sur>qadd<mode>): Likewise.
3796 (aarch64_sqmovun<mode>): Likewise.
3797 (aarch64_<sur>qmovn<mode>): Likewise.
3798 (aarch64_s<optab><mode>): Likewise.
3799 (aarch64_sq<r>dmulh<mode>): Likewise.
3800 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3801 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
3802 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3803 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
3804 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3805 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3806 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
3807 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
3808 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
3809 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
3810 (aarch64_sqdmull<mode>): Likewise.
3811 (aarch64_sqdmull_lane<mode>_internal): Likewise.
3812 (aarch64_sqdmull_n<mode>): Likewise.
3813 (aarch64_sqdmull2<mode>_internal): Likewise.
3814 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
3815 (aarch64_sqdmull2_n<mode>_internal): Likewise.
3816 (aarch64_<sur>shl<mode>): Likewise.
3817 (aarch64_<sur>q<r>shl<mode>
3818 (aarch64_<sur>shll_n<mode>): Likewise.
3819 (aarch64_<sur>shll2_n<mode>): Likewise.
3820 (aarch64_<sur>shr_n<mode>): Likewise.
3821 (aarch64_<sur>sra_n<mode>): Likewise.
3822 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
3823 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
3824 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
3825 (aarch64_cm<optab><mode>): Likewise.
3826 (aarch64_cm<optab>di): Likewise.
3827 (aarch64_cm<optab><mode>): Likewise.
3828 (aarch64_cm<optab>di): Likewise.
3829 (aarch64_cmtst<mode>): Likewise.
3830 (aarch64_cmtstdi): Likewise.
3831 (aarch64_cm<optab><mode>): Likewise.
3832 (*aarch64_fac<optab><mode>): Likewise.
3833 (aarch64_addp<mode>): Likewise.
3834 (aarch64_addpdi): Likewise.
3835 (sqrt<mode>2): Likewise.
3836 (vec_load_lanesoi<mode>): Likewise.
3837 (vec_store_lanesoi<mode>): Likewise.
3838 (vec_load_lanesci<mode>): Likewise.
3839 (vec_store_lanesci<mode>): Likewise.
3840 (vec_load_lanesxi<mode>): Likewise.
3841 (vec_store_lanesxi<mode>): Likewise.
3842 (*aarch64_mov<mode>): Likewise.
3843 (aarch64_ld2<mode>_dreg): Likewise.
3844 (aarch64_ld2<mode>_dreg): Likewise.
3845 (aarch64_ld3<mode>_dreg): Likewise.
3846 (aarch64_ld3<mode>_dreg): Likewise.
3847 (aarch64_ld4<mode>_dreg): Likewise.
3848 (aarch64_ld4<mode>_dreg): Likewise.
3849 (aarch64_tbl1<mode>): Likewise.
3850 (aarch64_tbl2v16qi): Likewise.
3851 (aarch64_combinev16qi): Likewise.
3852 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
3853 (aarch64_st2<mode>_dreg): Likewise.
3854 (aarch64_st2<mode>_dreg): Likewise.
3855 (aarch64_st3<mode>_dreg): Likewise.
3856 (aarch64_st3<mode>_dreg): Likewise.
3857 (aarch64_st4<mode>_dreg): Likewise.
3858 (aarch64_st4<mode>_dreg): Likewise.
3859 (*aarch64_simd_ld1r<mode>): Likewise.
3860 (aarch64_frecpe<mode>): Likewise.
3861 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
3862 (aarch64_frecps<mode>): Likewise.
3864 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3866 * config/arm/iterators.md (V_elem_ch): New.
3868 (VQH_type): Likewise.
3869 * config/arm/arm.md (is_neon_type): New.
3870 (conds): Use is_neon_type.
3871 (anddi3_insn): Update type attribute.
3872 (xordi3_insn): Likewise.
3873 (one_cmpldi2): Likewise.
3874 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
3875 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
3876 (*movmisalign<mode>_neon_store): Likewise.
3877 (*movmisalign<mode>_neon_load): Likewise.
3878 (vec_set<mode>_internal): Likewise.
3879 (vec_set<mode>_internal): Likewise.
3880 (vec_setv2di_internal): Likewise.
3881 (vec_extract<mode>): Likewise.
3882 (vec_extract<mode>): Likewise.
3883 (vec_extractv2di): Likewise.
3884 (*add<mode>3_neon): Likewise.
3885 (adddi3_neon): Likewise.
3886 (*sub<mode>3_neon): Likewise.
3887 (subdi3_neon): Likewise.
3888 (fma<VCVTF:mode>4): Likewise.
3889 (fma<VCVTF:mode>4_intrinsic): Likewise.
3890 (*fmsub<VCVTF:mode>4): Likewise.
3891 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3892 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3893 (ior<mode>3): Likewise.
3894 (and<mode>3): Likewise.
3895 (orn<mode>3_neon): Likewise.
3896 (orndi3_neon): Likewise.
3897 (bic<mode>3_neon): Likewise.
3898 (bicdi3_neon): Likewise.
3899 (xor<mode>3): Likewise.
3900 (one_cmpl<mode>2): Likewise.
3901 (abs<mode>2): Likewise.
3902 (neg<mode>2): Likewise.
3903 (negdi2_neon): Likewise.
3904 (*umin<mode>3_neon): Likewise.
3905 (*umax<mode>3_neon): Likewise.
3906 (*smin<mode>3_neon): Likewise.
3907 (*smax<mode>3_neon): Likewise.
3908 (vashl<mode>3): Likewise.
3909 (vashr<mode>3_imm): Likewise.
3910 (vlshr<mode>3_imm): Likewise.
3911 (ashl<mode>3_signed): Likewise.
3912 (ashl<mode>3_unsigned): Likewise.
3913 (neon_load_count): Likewise.
3914 (ashldi3_neon_noclobber): Likewise.
3915 (ashldi3_neon): Likewise.
3916 (signed_shift_di3_neon): Likewise.
3917 (unsigned_shift_di3_neon): Likewise.
3918 (ashrdi3_neon_imm_noclobber): Likewise.
3919 (lshrdi3_neon_imm_noclobber): Likewise.
3920 (<shift>di3_neon): Likewise.
3921 (widen_ssum<mode>3): Likewise.
3922 (widen_usum<mode>3): Likewise.
3923 (quad_halves_<code>v4si): Likewise.
3924 (quad_halves_<code>v4sf): Likewise.
3925 (quad_halves_<code>v8hi): Likewise.
3926 (quad_halves_<code>v16qi): Likewise.
3927 (reduc_splus_v2di): Likewise.
3928 (neon_vpadd_internal<mode>): Likewise.
3929 (neon_vpsmin<mode>): Likewise.
3930 (neon_vpsmax<mode>): Likewise.
3931 (neon_vpumin<mode>): Likewise.
3932 (neon_vpumax<mode>): Likewise.
3933 (*ss_add<mode>_neon): Likewise.
3934 (*us_add<mode>_neon): Likewise.
3935 (*ss_sub<mode>_neon): Likewise.
3936 (*us_sub<mode>_neon): Likewise.
3937 (neon_vadd<mode>_unspec): Likewise.
3938 (neon_vaddl<mode>): Likewise.
3939 (neon_vaddw<mode>): Likewise.
3940 (neon_vhadd<mode>): Likewise.
3941 (neon_vqadd<mode>): Likewise.
3942 (neon_vaddhn<mode>): Likewise.
3943 (neon_vmul<mode>): Likewise.
3944 (neon_vfms<VCVTF:mode>): Likewise.
3945 (neon_vmlal<mode>): Likewise.
3946 (neon_vmls<mode>): Likewise.
3947 (neon_vmlsl<mode>): Likewise.
3948 (neon_vqdmulh<mode>): Likewise.
3949 (neon_vqdmlal<mode>): Likewise.
3950 (neon_vqdmlsl<mode>): Likewise.
3951 (neon_vmull<mode>): Likewise.
3952 (neon_vqdmull<mode>): Likewise.
3953 (neon_vsub<mode>_unspec): Likewise.
3954 (neon_vsubl<mode>): Likewise.
3955 (neon_vsubw<mode>): Likewise.
3956 (neon_vqsub<mode>): Likewise.
3957 (neon_vhsub<mode>): Likewise.
3958 (neon_vsubhn<mode>): Likewise.
3959 (neon_vceq<mode>): Likewise.
3960 (neon_vcge<mode>): Likewise.
3961 (neon_vcgeu<mode>): Likewise.
3962 (neon_vcgt<mode>): Likewise.
3963 (neon_vcgtu<mode>): Likewise.
3964 (neon_vcle<mode>): Likewise.
3965 (neon_vclt<mode>): Likewise.
3966 (neon_vcage<mode>): Likewise.
3967 (neon_vcagt<mode>): Likewise.
3968 (neon_vtst<mode>): Likewise.
3969 (neon_vabd<mode>): Likewise.
3970 (neon_vabdl<mode>): Likewise.
3971 (neon_vaba<mode>): Likewise.
3972 (neon_vabal<mode>): Likewise.
3973 (neon_vmax<mode>): Likewise.
3974 (neon_vmin<mode>): Likewise.
3975 (neon_vpaddl<mode>): Likewise.
3976 (neon_vpadal<mode>): Likewise.
3977 (neon_vpmax<mode>): Likewise.
3978 (neon_vpmin<mode>): Likewise.
3979 (neon_vrecps<mode>): Likewise.
3980 (neon_vrsqrts<mode>): Likewise.
3981 (neon_vqabs<mode>): Likewise.
3982 (neon_vqneg<mode>): Likewise.
3983 (neon_vcls<mode>): Likewise.
3984 (clz<mode>2): Likewise.
3985 (popcount<mode>2): Likewise.
3986 (neon_vrecpe<mode>): Likewise.
3987 (neon_vrsqrte<mode>): Likewise.
3988 (neon_vget_lane<mode>_sext_internal): Likewise.
3989 (neon_vget_lane<mode>_zext_internal): Likewise.
3990 (neon_vdup_n<mode>): Likewise.
3991 (neon_vdup_n<mode>): Likewise.
3992 (neon_vdup_nv2di): Likewise.
3993 (neon_vdup_lane<mode>_interal): Likewise.
3994 (*neon_vswp<mode>): Likewise.
3995 (neon_vcombine<mode>): Likewise.
3996 (float<mode><V_cvtto>2): Likewise.
3997 (floatuns<mode><V_cvtto>2): Likewise.
3998 (fix_trunc<mode><V_cvtto>2): Likewise.
3999 (fixuns_trunc<mode><V_cvtto>2
4000 (neon_vcvt<mode>): Likewise.
4001 (neon_vcvt<mode>): Likewise.
4002 (neon_vcvtv4sfv4hf): Likewise.
4003 (neon_vcvtv4hfv4sf): Likewise.
4004 (neon_vcvt_n<mode>): Likewise.
4005 (neon_vcvt_n<mode>): Likewise.
4006 (neon_vmovn<mode>): Likewise.
4007 (neon_vqmovn<mode>): Likewise.
4008 (neon_vqmovun<mode>): Likewise.
4009 (neon_vmovl<mode>): Likewise.
4010 (neon_vmul_lane<mode>): Likewise.
4011 (neon_vmul_lane<mode>): Likewise.
4012 (neon_vmull_lane<mode>): Likewise.
4013 (neon_vqdmull_lane<mode>): Likewise.
4014 (neon_vqdmulh_lane<mode>): Likewise.
4015 (neon_vqdmulh_lane<mode>): Likewise.
4016 (neon_vmla_lane<mode>): Likewise.
4017 (neon_vmla_lane<mode>): Likewise.
4018 (neon_vmlal_lane<mode>): Likewise.
4019 (neon_vqdmlal_lane<mode>): Likewise.
4020 (neon_vmls_lane<mode>): Likewise.
4021 (neon_vmls_lane<mode>): Likewise.
4022 (neon_vmlsl_lane<mode>): Likewise.
4023 (neon_vqdmlsl_lane<mode>): Likewise.
4024 (neon_vext<mode>): Likewise.
4025 (neon_vrev64<mode>): Likewise.
4026 (neon_vrev32<mode>): Likewise.
4027 (neon_vrev16<mode>): Likewise.
4028 (neon_vbsl<mode>_internal): Likewise.
4029 (neon_vshl<mode>): Likewise.
4030 (neon_vqshl<mode>): Likewise.
4031 (neon_vshr_n<mode>): Likewise.
4032 (neon_vshrn_n<mode>): Likewise.
4033 (neon_vqshrn_n<mode>): Likewise.
4034 (neon_vqshrun_n<mode>): Likewise.
4035 (neon_vshl_n<mode>): Likewise.
4036 (neon_vqshl_n<mode>): Likewise.
4037 (neon_vqshlu_n<mode>): Likewise.
4038 (neon_vshll_n<mode>): Likewise.
4039 (neon_vsra_n<mode>): Likewise.
4040 (neon_vsri_n<mode>): Likewise.
4041 (neon_vsli_n<mode>): Likewise.
4042 (neon_vtbl1v8qi): Likewise.
4043 (neon_vtbl2v8qi): Likewise.
4044 (neon_vtbl3v8qi): Likewise.
4045 (neon_vtbl4v8qi): Likewise.
4046 (neon_vtbl1v16qi): Likewise.
4047 (neon_vtbl2v16qi): Likewise.
4048 (neon_vcombinev16qi): Likewise.
4049 (neon_vtbx1v8qi): Likewise.
4050 (neon_vtbx2v8qi): Likewise.
4051 (neon_vtbx3v8qi): Likewise.
4052 (neon_vtbx4v8qi): Likewise.
4053 (*neon_vtrn<mode>_insn): Likewise.
4054 (*neon_vzip<mode>_insn): Likewise.
4055 (*neon_vuzp<mode>_insn): Likewise.
4056 (neon_vld1<mode>): Likewise.
4057 (neon_vld1_lane<mode>): Likewise.
4058 (neon_vld1_lane<mode>): Likewise.
4059 (neon_vld1_dup<mode>): Likewise.
4060 (neon_vld1_dup<mode>): Likewise.
4061 (neon_vld1_dupv2di): Likewise.
4062 (neon_vst1<mode>): Likewise.
4063 (neon_vst1_lane<mode>): Likewise.
4064 (neon_vst1_lane<mode>): Likewise.
4065 (neon_vld2<mode>): Likewise.
4066 (neon_vld2<mode>): Likewise.
4067 (neon_vld2_lane<mode>): Likewise.
4068 (neon_vld2_lane<mode>): Likewise.
4069 (neon_vld2_dup<mode>): Likewise.
4070 (neon_vst2<mode>): Likewise.
4071 (neon_vst2<mode>): Likewise.
4072 (neon_vst2_lane<mode>): Likewise.
4073 (neon_vst2_lane<mode>): Likewise.
4074 (neon_vld3<mode>): Likewise.
4075 (neon_vld3qa<mode>): Likewise.
4076 (neon_vld3qb<mode>): Likewise.
4077 (neon_vld3_lane<mode>): Likewise.
4078 (neon_vld3_lane<mode>): Likewise.
4079 (neon_vld3_dup<mode>): Likewise.
4080 (neon_vst3<mode>): Likewise.
4081 (neon_vst3qa<mode>): Likewise.
4082 (neon_vst3qb<mode>): Likewise.
4083 (neon_vst3_lane<mode>): Likewise.
4084 (neon_vst3_lane<mode>): Likewise.
4085 (neon_vld4<mode>): Likewise.
4086 (neon_vld4qa<mode>): Likewise.
4087 (neon_vld4qb<mode>): Likewise.
4088 (neon_vld4_lane<mode>): Likewise.
4089 (neon_vld4_lane<mode>): Likewise.
4090 (neon_vld4_dup<mode>): Likewise.
4091 (neon_vst4<mode>): Likewise.
4092 (neon_vst4qa<mode>): Likewise.
4093 (neon_vst4qb<mode>): Likewise.
4094 (neon_vst4_lane<mode>): Likewise.
4095 (neon_vst4_lane<mode>): Likewise.
4096 (neon_vec_unpack<US>_lo_<mode>): Likewise.
4097 (neon_vec_unpack<US>_hi_<mode>): Likewise.
4098 (neon_vec_<US>mult_lo_<mode>): Likewise.
4099 (neon_vec_<US>mult_hi_<mode>): Likewise.
4100 (neon_vec_<US>shiftl_<mode>): Likewise.
4101 (neon_unpack<US>_<mode>): Likewise.
4102 (neon_vec_<US>mult_<mode>): Likewise.
4103 (vec_pack_trunc_<mode>): Likewise.
4104 (neon_vec_pack_trunc_<mode>): Likewise.
4105 (neon_vabd<mode>_2): Likewise.
4106 (neon_vabd<mode>_3): Likewise.
4108 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4110 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
4111 (load_pair): Update type attribute.
4112 (store_pair): Update type attribute.
4113 * config/aarch64/iterators.md (q): New.
4115 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4117 * config/arm/types.md: Add new types for Neon insns.
4119 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4120 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4121 Sergey Lega <sergey.s.lega@intel.com>
4122 Anna Tikhonova <anna.tikhonova@intel.com>
4123 Ilya Tocar <ilya.tocar@intel.com>
4124 Andrey Turetskiy <andrey.turetskiy@intel.com>
4125 Ilya Verbin <ilya.verbin@intel.com>
4126 Kirill Yukhin <kirill.yukhin@intel.com>
4127 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4129 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
4130 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
4131 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
4133 (srcp14<mode>): Ditto.
4134 (rsqrt14<mode>): Ditto.
4135 (rsqrt14<mode>): Ditto.
4136 (avx512f_vmscalef<mode>): Ditto.
4137 (avx512f_scalef<mode>): Ditto.
4138 (avx512f_getexp<mode>): Ditto.
4139 (avx512f_sgetexp<mode>): Ditto.
4140 (avx512f_fixupimm<mode>): Ditto.
4141 (avx512f_sfixupimm<mode>): Ditto.
4142 (avx512f_rndscale<mode>): Ditto.
4143 (*avx512er_exp2<mode>): Ditto.
4144 (*avx512er_rcp28<mode>): Ditto.
4145 (avx512er_rsqrt28<mode>): Ditto.
4146 (avx512f_getmant<mode>): Ditto.
4147 (avx512f_getmant<mode>): Ditto.
4148 (avx512f_rndscale<mode>): Fix formatting.
4150 2013-10-15 Martin Jambor <mjambor@suse.cz>
4152 * ipa-utils.h (ipa_edge_within_scc): Declare.
4153 * ipa-cp.c (edge_within_scc): Moved...
4154 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
4156 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4157 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4158 Sergey Lega <sergey.s.lega@intel.com>
4159 Anna Tikhonova <anna.tikhonova@intel.com>
4160 Ilya Tocar <ilya.tocar@intel.com>
4161 Andrey Turetskiy <andrey.turetskiy@intel.com>
4162 Ilya Verbin <ilya.verbin@intel.com>
4163 Kirill Yukhin <kirill.yukhin@intel.com>
4164 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4166 * config/i386/predicates.md (const_8_to_15_operand): New.
4167 (const_16_to_31_operand): Ditto.
4168 * config/i386/sse.md (V8FI): New.
4170 (reduc_splus_v8df): Ditto.
4171 (reduc_splus_v16sf): Ditto.
4172 (avx512f_vextract<shuffletype>32x4_1): Ditto.
4173 (vec_extract_hi_<mode>): Ditto.
4174 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
4175 (vec_set_lo_<mode>): Ditto.
4176 (vec_set_hi_<mode>): Ditto.
4177 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
4178 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
4179 (avx512f_pshufd_1): Ditto.
4180 (avx512f_broadcast<mode>): Ditto.
4181 (avx512f_broadcast<mode>): Ditto.
4182 (define_split): Split vec_extract_lo into move.
4183 (ssequartermode): Ditto.
4184 (ssedoublemode): Extened with wider modes.
4185 (vec_extract_lo_<mode>): Ditto.
4187 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4188 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4189 Sergey Lega <sergey.s.lega@intel.com>
4190 Anna Tikhonova <anna.tikhonova@intel.com>
4191 Ilya Tocar <ilya.tocar@intel.com>
4192 Andrey Turetskiy <andrey.turetskiy@intel.com>
4193 Ilya Verbin <ilya.verbin@intel.com>
4194 Kirill Yukhin <kirill.yukhin@intel.com>
4195 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4197 * config/i386/predicates.md (register_or_constm1_operand): New.
4198 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
4199 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
4200 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
4201 UNSPEC_SCATTER_PREFETCH
4203 (avx512f_ucmp<mode>3): Ditto.
4204 (avx512f_vternlog<mode>): Ditto.
4205 (avx512f_align<mode>): Ditto.
4206 (<shift_insn><mode>3): Ditto.
4207 (avx512f_<rotate>v<mode>): Ditto.
4208 (avx512f_<rotate><mode>): Ditto.
4209 (avx512f_eq<mode>3): Ditto.
4210 (avx512f_eq<mode>3_1): Ditto.
4211 (avx512f_gt<mode>3): Ditto.
4212 (avx512f_testm<mode>3): Ditto.
4213 (avx512f_testnm<mode>3): Ditto.
4214 (avx512pf_gatherpf<mode>): Ditto.
4215 (*avx512pf_gatherpf<mode>_mask): Ditto.
4216 (*avx512pf_gatherpf<mode>): Ditto.
4217 (avx512pf_scatterpf<mode>): Ditto.
4218 (*avx512pf_scatterpf<mode>_mask): Ditto.
4219 (*avx512pf_scatterpf<mode>): Ditto.
4220 (avx512f_vec_dup_gpr<mode>): Ditto.
4221 (clz<mode>2): Ditto.
4222 (conflict<mode>): Ditto.
4223 (REDUC_SMINMAX_MODE): Extened with wider modes.
4224 (reduc_<code>_<mode>): Ditto.
4225 (vlshr<mode>3): Ditto.
4226 (vashl<mode>3): Ditto.
4228 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4229 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4230 Sergey Lega <sergey.s.lega@intel.com>
4231 Anna Tikhonova <anna.tikhonova@intel.com>
4232 Ilya Tocar <ilya.tocar@intel.com>
4233 Andrey Turetskiy <andrey.turetskiy@intel.com>
4234 Ilya Verbin <ilya.verbin@intel.com>
4235 Kirill Yukhin <kirill.yukhin@intel.com>
4236 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4238 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
4241 (avx512fmaskmode): Ditto.
4242 (bcstscalarsuff): Ditto.
4243 (avx512f_blendm<mode>): Ditto.
4244 (cmp_imm_predicate): Ditto.
4245 (avx512f_cmp<mode>3): Ditto.
4246 (avx512f_vec_dup<mode>): Ditto.
4247 (avx512f_vec_dup_mem<mode>): Ditto.
4248 (avx512f_vpermi2var<mode>3): Ditto.
4249 (avx512f_vpermt2var<mode>3): Ditto.
4250 (vec_init<mode>): Ditto.
4251 (avx512f_gathersi<mode>): Ditto.
4252 (*avx512f_gathersi<mode>): Ditto.
4253 (*avx512f_gathersi<mode>_2): Ditto.
4254 (avx512f_gatherdi<mode>): Ditto.
4255 (*avx512f_gatherdi<mode>): Ditto.
4256 (*avx512f_gatherdi<mode>_2): Ditto.
4257 (avx512f_scattersi<mode>): Ditto.
4258 (*avx512f_scattersi<mode>): Ditto.
4259 (avx512f_scatterdi<mode>): Ditto.
4260 (*avx512f_scatterdi<mode>): Ditto.
4261 (sseintprefix): Extened with wider modes.
4262 (VEC_GATHER_IDXSI): Ditto.
4263 (VEC_GATHER_IDXDI): Ditto.
4264 (VEC_GATHER_SRCDI): Ditto.
4266 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
4267 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4269 * config/arm/t-aprofile: New file.
4270 * config.gcc: Handle --with-multilib-list option.
4272 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
4274 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
4277 2013-10-15 Richard Biener <rguenther@suse.de>
4279 * tree-loop-distribution.c (build_empty_rdg): Inline into
4281 (rdg_flag_vertex): Inline into single user.
4282 (rdg_flag_vertex_and_dependent): Likewise.
4283 (build_rdg_partition_for_vertex): Remove processed bitmap.
4284 (rdg_build_partitions): Simplify.
4286 2013-10-15 Richard Biener <rguenther@suse.de>
4288 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
4289 Restructure forwarding through conversions and copies to
4290 avoid performing copy-propagation the wrong way. Adjust
4291 recursion invocations.
4292 (forward_propagate_addr_expr): Add argument stating if we
4293 are recursing from a single-use.
4294 (ssa_forward_propagate_and_combine): Adjust.
4296 2013-10-14 David Malcolm <dmalcolm@redhat.com>
4298 * dumpfile.h (gcc::dump_manager): New class, to hold state
4299 relating to dumpfile management.
4300 (get_dump_file_name): Remove in favor of method of dump_manager.
4301 (dump_initialized_p): Likewise.
4302 (dump_start): Likewise.
4303 (dump_finish): Likewise.
4304 (dump_switch_p): Likewise.
4305 (dump_register): Likewise.
4306 (get_dump_file_info): Likewise.
4307 * context.c (gcc::context::context): Construct the dump_manager
4309 * context.h (gcc::context::get_dumps): New.
4310 (gcc::context::m_dumps): New.
4311 * coverage.c (coverage_init): Port to dump_manager API.
4312 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
4313 (extra_dump_files_in_use): Likewise.
4314 (extra_dump_files_alloced): Likewise.
4315 (gcc::dump_manager::dump_manager): New.
4316 (dump_register): Convert to...
4317 (gcc::dump_manager::dump_register): ...method, replacing
4318 function-static next_dump with m_next_dump field.
4319 (get_dump_file_info): Convert to...
4320 (gcc::dump_manager::get_dump_file_info): ...method.
4321 (get_dump_file_name): Convert to...
4322 (gcc::dump_manager::get_dump_file_name): ...method.
4323 (dump_start): Convert to...
4324 (gcc::dump_manager::dump_start): ...method.
4325 (dump_finish): Convert to...
4326 (gcc::dump_manager::dump_finish): ...method.
4327 (dump_begin): Replace body with...
4328 (gcc::dump_manager::dump_begin): ...new method.
4329 (dump_phase_enabled_p): Convert to...
4330 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4331 (dump_phase_enabled_p): Convert to...
4332 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4333 (dump_initialized_p): Convert to...
4334 (gcc::dump_manager::dump_initialized_p): ...method.
4335 (dump_flag_name): Replace body with...
4336 (gcc::dump_manager::dump_flag_name): ...new method.
4337 (dump_enable_all): Convert to...
4338 (gcc::dump_manager::dump_enable_all): ...new method.
4339 (opt_info_enable_passes): Convert to...
4340 (gcc::dump_manager::opt_info_enable_passes): ...new method.
4341 (dump_switch_p_1): Convert to...
4342 (gcc::dump_manager::dump_switch_p_1): ...new method.
4343 (dump_switch_p): Convert to...
4344 (gcc::dump_manager::dump_switch_p): ...new method.
4345 (opt_info_switch_p): Port to dump_manager API.
4346 (enable_rtl_dump_file): Likewise.
4347 * opts-global.c (handle_common_deferred_options): Port to new
4349 * passes.c (pass_manager::finish_optimization_passes): Likewise.
4350 (pass_manager::register_one_dump_file): Likewise.
4351 (pass_manager::register_pass): Likewise.
4352 (pass_init_dump_file): Likewise.
4353 (pass_fini_dump_file): Likewise.
4354 * statistics.c (statistics_early_init): Likewise.
4356 2013-10-14 Richard Biener <rguenther@suse.de>
4358 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
4359 iterative_hash_canonical_type, gimple_canonical_type_hash,
4360 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
4361 gimple_register_canonical_type, print_gimple_types_stats,
4362 free_gimple_type_tables): Move to lto/lto.c
4363 (gt-gimple.h): Do not include.
4364 * gimple.h (gimple_register_canonical_type,
4365 print_gimple_types_stats, free_gimple_type_tables): Remove.
4366 * Makefile.in (GTFILES): Remove gimple.c.
4368 2013-10-14 Travis Snoozy <quandary@remstate.com>
4371 * config/msp430/msp430.c (msp430_option_override): Correct thinko
4372 scanning for msp430x targets.
4374 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
4377 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
4378 (registers_ok_for_ldd_peep): Move around.
4379 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
4380 * config/sparc/sparc.md (widening peepholes): Use it.
4382 2013-10-14 Richard Biener <rguenther@suse.de>
4386 * gimple.c (iterative_hash_canonical_type): Make sure to
4387 record the hash into the correct hashtable slot.
4389 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
4391 PR rtl-optimization/58662
4392 * combine.c (try_combine): Take into account death nodes on I2 when
4393 splitting a PARALLEL of two independent SETs. Fix dump message.
4395 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
4398 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
4399 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
4400 * config/sh/t-sh (sh_treg_combine.o): New entry.
4401 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
4402 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
4403 (register_sh_passes): New function. Register sh_treg_combine pass.
4404 (sh_option_override): Invoke it.
4405 (sh_canonicalize_comparison): Handle op0_preserve_value.
4406 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
4407 opportunities. Canonicalize branch condition.
4408 (nott): Allow only if pseudos can be created for non-SH2A.
4410 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
4413 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
4414 (ix86_expand_movmem): Replace copy_addr_to_reg with
4415 ix86_copy_addr_to_reg.
4416 (ix86_expand_setmem): Likewise.
4418 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
4420 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
4421 provided by ix86_fp_compare_mode instead of CCFPUmode.
4423 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
4425 * config/aarch64/arm_neon.h
4426 (vtbx<1,3>_<psu>8): Fix register constriants.
4428 2013-10-11 Jeff Law <law@redhat.com>
4430 PR tree-optimization/58640
4431 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
4432 threading paths that cross over two loop entry points.
4434 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4436 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
4437 handle vector float as well.
4438 (*vsx_le_perm_load_v4si): Likewise.
4439 (*vsx_le_perm_store_v2di): Likewise.
4440 (*vsx_le_perm_store_v4si): Likewise.
4442 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4444 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
4445 directly to circumvent subtract from splat{31} workaround.
4446 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
4448 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
4449 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
4450 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
4451 (altivec_vperm_<mode>): Convert to define_insn_and_split to
4452 separate big and little endian logic.
4453 (*altivec_vperm_<mode>_internal): New define_insn.
4454 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
4455 separate big and little endian logic.
4456 (*altivec_vperm_<mode>_uns_internal): New define_insn.
4457 (vec_permv16qi): Add little endian logic.
4459 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
4461 * doc/extend.texi (returns_nonnull): Remove arguments.
4463 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4464 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4465 Sergey Lega <sergey.s.lega@intel.com>
4466 Anna Tikhonova <anna.tikhonova@intel.com>
4467 Ilya Tocar <ilya.tocar@intel.com>
4468 Andrey Turetskiy <andrey.turetskiy@intel.com>
4469 Ilya Verbin <ilya.verbin@intel.com>
4470 Kirill Yukhin <kirill.yukhin@intel.com>
4471 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4473 * config/i386/sse.md (VI48F_256_512): New.
4474 (avx2_permvar<mode>): Change to ...
4475 (<avx2_avx512f>_permvar<mode>): This.
4477 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4478 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4479 Sergey Lega <sergey.s.lega@intel.com>
4480 Anna Tikhonova <anna.tikhonova@intel.com>
4481 Ilya Tocar <ilya.tocar@intel.com>
4482 Andrey Turetskiy <andrey.turetskiy@intel.com>
4483 Ilya Verbin <ilya.verbin@intel.com>
4484 Kirill Yukhin <kirill.yukhin@intel.com>
4485 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4487 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
4488 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
4489 * config/i386/sse.md (VI4_AVX): New.
4490 (sf2simodelower): Ditto.
4491 (sse2_cvtps2dq): Change to ...
4492 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
4494 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4495 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4496 Sergey Lega <sergey.s.lega@intel.com>
4497 Anna Tikhonova <anna.tikhonova@intel.com>
4498 Ilya Tocar <ilya.tocar@intel.com>
4499 Andrey Turetskiy <andrey.turetskiy@intel.com>
4500 Ilya Verbin <ilya.verbin@intel.com>
4501 Kirill Yukhin <kirill.yukhin@intel.com>
4502 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4504 * config/i386/sse.md (V_512): New.
4506 (vcond<V_512:mode><VF_512:mode>): Ditto.
4507 (vcond<V_512:mode><VI_512:mode>): Ditto.
4508 (vcondu<V_512:mode><VI_512:mode>): Ditto.
4510 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4511 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4512 Sergey Lega <sergey.s.lega@intel.com>
4513 Anna Tikhonova <anna.tikhonova@intel.com>
4514 Ilya Tocar <ilya.tocar@intel.com>
4515 Andrey Turetskiy <andrey.turetskiy@intel.com>
4516 Ilya Verbin <ilya.verbin@intel.com>
4517 Kirill Yukhin <kirill.yukhin@intel.com>
4518 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4520 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
4521 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
4523 (fma<mode>4): Removed condition.
4524 (fms<mode>4): Ditto.
4525 (fnma<mode>4): Ditto.
4526 (fnms<mode>4): Ditto.
4527 (fma4i_fmadd_<mode>): Ditto.
4528 (*fma_fmadd_<mode>): Ditto.
4529 (*fma_fmsub_<mode>): Ditto.
4530 (*fma_fnmadd_<mode>): Ditto.
4531 (*fma_fnmsub_<mode>): Ditto.
4532 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
4533 (*fma_fmaddsub_<mode>): Ditto.
4534 (*fma_fmsubadd_<mode>): Ditto.
4535 (*fmai_fmadd_<mode>): Ditto.
4536 (*fmai_fmsub_<mode>): Ditto.
4537 (*fmai_fnmadd_<mode>): Ditto.
4538 (*fmai_fnmsub_<mode>): Ditto.
4540 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4541 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4542 Sergey Lega <sergey.s.lega@intel.com>
4543 Anna Tikhonova <anna.tikhonova@intel.com>
4544 Ilya Tocar <ilya.tocar@intel.com>
4545 Andrey Turetskiy <andrey.turetskiy@intel.com>
4546 Ilya Verbin <ilya.verbin@intel.com>
4547 Kirill Yukhin <kirill.yukhin@intel.com>
4548 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4550 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
4551 (VI124_256): Changed to ...
4552 (VI124_256_48_512): This.
4553 (ssepackmode): Extended with wider modes.
4554 (<code><mode>3): Changed iterator.
4555 (*avx2_<code><mode>3): Ditto.
4556 (vec_pack_trunc_<mode>): Ditto.
4558 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4559 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4560 Sergey Lega <sergey.s.lega@intel.com>
4561 Anna Tikhonova <anna.tikhonova@intel.com>
4562 Ilya Tocar <ilya.tocar@intel.com>
4563 Andrey Turetskiy <andrey.turetskiy@intel.com>
4564 Ilya Verbin <ilya.verbin@intel.com>
4565 Kirill Yukhin <kirill.yukhin@intel.com>
4566 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4568 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
4569 (VI8F_256_512): Ditto.
4570 (abs<mode>2): Changed iterator.
4571 (avx2_perm<mode>): Changed to ...
4572 (<avx2_avx512f>_perm<mode>): This.
4573 (avx2_perm<mode>_1): Changed to ...
4574 (<avx2_avx512f>_perm<mode>_1): This.
4576 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4577 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4578 Sergey Lega <sergey.s.lega@intel.com>
4579 Anna Tikhonova <anna.tikhonova@intel.com>
4580 Ilya Tocar <ilya.tocar@intel.com>
4581 Andrey Turetskiy <andrey.turetskiy@intel.com>
4582 Ilya Verbin <ilya.verbin@intel.com>
4583 Kirill Yukhin <kirill.yukhin@intel.com>
4584 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4586 * config/i386/sse.md (VI48_AVX512F): New.
4587 (VI48_AVX2): Changed to ...
4588 (VI48_AVX2_48_AVX512F): This.
4589 (avx2_ashrv<mode>): Changed to ...
4590 (<avx2_avx512f>_ashrv<mode>): This.
4591 (avx2_<shift_insn>v<mode>): Changed to ...
4592 (<avx2_avx512f>_<shift_insn>v<mode>): This.
4594 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4595 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4596 Sergey Lega <sergey.s.lega@intel.com>
4597 Anna Tikhonova <anna.tikhonova@intel.com>
4598 Ilya Tocar <ilya.tocar@intel.com>
4599 Andrey Turetskiy <andrey.turetskiy@intel.com>
4600 Ilya Verbin <ilya.verbin@intel.com>
4601 Kirill Yukhin <kirill.yukhin@intel.com>
4602 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4604 * config/i386/sse.md (VI4_AVX512F): New.
4605 (VI8_AVX2_AVX512F): Ditto.
4606 (mul<mode>3): Extended with wider modes.
4607 (*<sse4_1_avx2>_mul<mode>3): Ditto.
4608 (mul<mode>3): Ditto.
4609 (vec_widen_<s>mult_odd_<mode>): Ditto.
4611 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4612 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4613 Sergey Lega <sergey.s.lega@intel.com>
4614 Anna Tikhonova <anna.tikhonova@intel.com>
4615 Ilya Tocar <ilya.tocar@intel.com>
4616 Andrey Turetskiy <andrey.turetskiy@intel.com>
4617 Ilya Verbin <ilya.verbin@intel.com>
4618 Kirill Yukhin <kirill.yukhin@intel.com>
4619 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4621 * config/i386/sse.md (VI2_AVX512F): New.
4622 (VI124_AVX512F): Ditto.
4623 (sseunpackmode): Extended with wider modes.
4624 (sseunpackfltmode): Ditto.
4625 (vec_unpacks_float_hi_<mode>): Ditto.
4626 (vec_unpacks_float_lo_<mode>): Ditto.
4627 (vec_unpacku_float_hi_<mode>): Ditto.
4628 (vec_unpacku_float_lo_<mode>): Ditto.
4629 (vec_unpacks_lo_<mode>): Ditto.
4630 (vec_unpacks_hi_<mode>): Ditto.
4631 (vec_unpacku_lo_<mode>): Ditto.
4632 (vec_unpacku_hi_<mode>): Ditto.
4634 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4635 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4636 Sergey Lega <sergey.s.lega@intel.com>
4637 Anna Tikhonova <anna.tikhonova@intel.com>
4638 Ilya Tocar <ilya.tocar@intel.com>
4639 Andrey Turetskiy <andrey.turetskiy@intel.com>
4640 Ilya Verbin <ilya.verbin@intel.com>
4641 Kirill Yukhin <kirill.yukhin@intel.com>
4642 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4644 * config/i386/i386.md (multdiv): New.
4645 (multdiv_mnemonic): Ditto.
4646 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
4647 (<sse>_vm<multdiv_mnemonic><mode>3): This.
4648 (<sse>_vmdiv<mode>3): Removed.
4650 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4651 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4652 Sergey Lega <sergey.s.lega@intel.com>
4653 Anna Tikhonova <anna.tikhonova@intel.com>
4654 Ilya Tocar <ilya.tocar@intel.com>
4655 Andrey Turetskiy <andrey.turetskiy@intel.com>
4656 Ilya Verbin <ilya.verbin@intel.com>
4657 Kirill Yukhin <kirill.yukhin@intel.com>
4658 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4660 * config/i386/sse.md (V): Extended with wider modes.
4662 (ssehalfvecmode): Ditto.
4664 (ssepackfltmode): Ditto.
4665 (avx_vec_concat<mode>): Ditto.
4666 (V_256_512): New iterator.
4667 (VF2_512_256): Ditto.
4668 (si2dfmode): New attribute.
4669 (si2dfmodelower): Ditto.
4671 (concat_tg_mode): Ditto.
4672 (floatv4siv4df2): Changed to ...
4673 (float<si2dfmodelower><mode>2): This.
4674 (avx_cvtps2pd256): Changed to ...
4675 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
4676 (vec_pack_trunc_v4df): Changed to ...
4677 (vec_pack_trunc_<mode>): This.
4678 (avx_vpermil<mode>): Changed to ...
4679 (<sse2_avx_avx512f>_vpermil<mode>): This.
4680 (<fixsuffix>fix_truncv8dfv8si2): New.
4681 (vec_pack_sfix_trunc_v8df): Ditto.
4682 (avx512f_rndscale<mode>): Ditto.
4683 (avx512f_roundpd512): Ditto.
4684 (vec_pack_ufix_trunc_<mode>): Updated iterator.
4686 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4687 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4688 Sergey Lega <sergey.s.lega@intel.com>
4689 Anna Tikhonova <anna.tikhonova@intel.com>
4690 Ilya Tocar <ilya.tocar@intel.com>
4691 Andrey Turetskiy <andrey.turetskiy@intel.com>
4692 Ilya Verbin <ilya.verbin@intel.com>
4693 Kirill Yukhin <kirill.yukhin@intel.com>
4694 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4696 * config/i386/i386.md (any_fix): New iterator.
4697 (fixsuffix): New attribute.
4698 * config/i386/sse.md (VF1): Extened with wider modes.
4702 (sseintvecmodelower): Ditto.
4703 (ssescalarmode): Ditto.
4704 (ssescalarnum): Ditto.
4706 (ssexmmmode): Ditto.
4707 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
4708 (<sse>_rcp<mode>2): Change iterator.
4709 (rsqrt<mode>2): Ditto.
4710 (<sse>_rsqrt<mode>2): Ditto.
4711 (avx2_vec_dup<mode>): Ditto.
4712 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
4713 (round<mode>2_sfix): Ditto.
4714 (avx2_pbroadcast<mode>): Ditto.
4715 (*andnot<mode>3): Handle XI mode.
4716 (*<code><mode>3): Ditto.
4717 (AVXTOSSEMODE): Removed.
4718 (avx_vpermil<mode>): Changed to ...
4719 (<sse2_avx_avx512f>_vpermil<mode>): This.
4721 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4722 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4723 Sergey Lega <sergey.s.lega@intel.com>
4724 Anna Tikhonova <anna.tikhonova@intel.com>
4725 Ilya Tocar <ilya.tocar@intel.com>
4726 Andrey Turetskiy <andrey.turetskiy@intel.com>
4727 Ilya Verbin <ilya.verbin@intel.com>
4728 Kirill Yukhin <kirill.yukhin@intel.com>
4729 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4731 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
4732 (<sse>_comi): Ditto.
4733 (<sse>_ucomi): Ditto.
4734 (sse_cvtss2siq_2): Ditto.
4735 (sse2_cvtsd2si): Ditto.
4736 (sse2_cvtsd2siq): Ditto.
4737 (sse2_cvttsd2si): Ditto.
4738 (sse2_cvttsd2siq): Ditto.
4739 (<shift_insn><mode>3): Ditto.
4740 (sse2_cvtsi2sdq): Update constraint and prefix.
4741 (sse_cvtsi2ss): Update prefix.
4742 (sse_cvtsi2ssq): Ditto.
4744 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4746 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
4747 ignore internal calls.
4749 2013-10-11 Richard Biener <rguenther@suse.de>
4751 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
4752 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
4753 TREE_OVERFLOW is set.
4755 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
4757 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
4759 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
4760 * gimple.h: Likewise.
4762 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
4764 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
4767 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
4769 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
4770 * configure: Regenerate.
4772 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4774 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
4775 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
4776 formatting fixes, use pp_colon instead of pp_character (..., ':'),
4777 similarly pp_right_paren.
4778 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
4779 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
4780 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
4781 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
4783 (omp_declare_simd_clauses_equal,
4784 omp_remove_redundant_declare_simd_attrs): New functions.
4785 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
4786 (walk_tree_1): Handle new OpenMP 4.0 clauses.
4787 * tree.h (OMP_LOOP_CHECK): Define.
4788 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
4789 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
4790 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
4791 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
4792 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
4793 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
4794 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
4795 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
4796 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
4797 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
4798 OMP_CLAUSE_SIMDLEN_EXPR): Define.
4799 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
4800 (omp_remove_redundant_declare_simd_attrs): New prototype.
4801 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
4802 GIMPLE_OMP_TEAMS): New codes.
4803 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
4804 * omp-low.c (struct omp_context): Add cancel_label and cancellable
4806 (target_nesting_level): New variable.
4807 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
4808 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
4809 for collapse > 1 static schedule unless ordered.
4810 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
4811 (determine_parallel_type): Adjust get_ws_args_for caller.
4812 (install_var_field): Handle mask & 4 for double indirection.
4813 (scan_sharing_clauses): Ignore shared clause on teams construct.
4814 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
4815 (create_omp_child_function): If inside target or declare target
4816 constructs, set "omp declare target" attribute on the child function.
4817 (find_combined_for): New function.
4818 (scan_omp_parallel): Handle combined loops.
4819 (scan_omp_target, scan_omp_teams): New functions.
4820 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
4821 restrictions and set ctx->cancellable for cancellable constructs.
4822 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
4823 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
4824 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
4825 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
4826 (omp_clause_aligned_alignment): New function.
4827 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
4828 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
4829 on teams constructs. Handle user defined reductions and new
4831 (lower_reduction_clauses): Don't set placeholder to address of ref
4832 if it has already the right type.
4833 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
4834 (expand_parallel_call): Use the new non-_start suffixed builtins,
4835 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
4836 and GOMP_parallel_end after the call.
4837 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
4838 (expand_omp_for_init_counts): Handle combined loops.
4839 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
4841 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
4842 end of cancellable loops.
4843 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
4844 Likewise. Handle collapse > 1 loops.
4845 (expand_omp_simd): Handle combined loops.
4846 (expand_omp_for): Add inner_stmt argument, adjust callers of
4847 expand_omp_for* functions, use expand_omp_for_static*chunk even
4848 for collapse > 1 unless ordered.
4849 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
4850 of cancellable sections.
4851 (expand_omp_single): Remove need_barrier variable, just rely on
4852 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
4853 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
4854 (expand_omp_atomic_load, expand_omp_atomic_store,
4855 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
4856 (expand_omp_target): New function.
4857 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
4858 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
4859 (build_omp_regions_1): Immediately close region for
4860 GF_OMP_TARGET_KIND_UPDATE.
4861 (maybe_add_implicit_barrier_cancel): New function.
4862 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
4864 (lower_omp_single): Likewise. Add clobber after the barrier.
4865 (lower_omp_taskgroup): New function.
4866 (lower_omp_for): Handle combined loops. Adjust
4867 lower_rec_input_clauses caller. Handle cancellation.
4868 (lower_depend_clauses): New function.
4869 (lower_omp_taskreg): Lower depend clauses. Adjust
4870 lower_rec_input_clauses caller. Add clobber after the call. Handle
4872 (lower_omp_target, lower_omp_teams): New functions.
4873 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
4874 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
4875 and GOMP_cancellation_point calls.
4876 (lower_omp): Fold stmts inside of target region.
4877 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
4878 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4879 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
4880 (BT_FN_VOID_OMPFN_PTR_UINT,
4881 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
4882 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
4883 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
4884 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
4885 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
4886 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
4887 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
4888 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
4889 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
4890 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
4891 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
4892 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
4893 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
4894 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
4895 BUILT_IN_GOMP_PARALLEL_END.
4896 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
4897 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4898 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
4899 GF_OMP_FOR_KIND_DISTRIBUTE.
4900 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
4901 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
4902 (dump_gimple_omp_return): Print lhs if it has any.
4903 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
4904 gimple_omp_atomic_seq_cst_p.
4905 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
4906 and GIMPLE_OMP_TEAMS.
4907 * langhooks.c (lhd_omp_mappable_type): New function.
4908 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
4909 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
4911 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
4912 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
4913 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
4914 (struct gimplify_omp_ctx): Add combined_loop field.
4915 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
4916 on stmts inside of target region.
4917 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
4918 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
4919 ORT_TARGET and ORT_TARGET_DATA.
4920 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
4922 (omp_notice_threadprivate_variable): Complain about threadprivate
4923 variables in target region.
4924 (omp_notice_variable): Complain about vars with non-mappable type
4925 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
4926 (omp_check_private): Ignore ORT_TARGET* regions.
4927 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
4928 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
4929 (find_combined_omp_for): New function.
4930 (gimplify_omp_for): Handle gimplification of combined loops.
4931 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
4933 (gimplify_omp_target_update): New function.
4934 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
4935 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
4936 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
4937 (gimplify_body): If fndecl has "omp declare target" attribute, add
4938 implicit ORT_TARGET context around it.
4939 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
4940 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
4941 * tree-nested.c (convert_nonlocal_reference_stmt,
4942 convert_local_reference_stmt, convert_gimple_call): Handle
4943 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4944 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
4945 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
4946 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
4947 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
4948 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
4949 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
4950 BUILT_IN_GOMP_LOOP_END_CANCEL,
4951 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
4952 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
4953 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
4954 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
4955 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
4956 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
4957 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
4958 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
4959 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
4960 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
4961 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
4962 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
4963 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
4964 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
4965 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
4966 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4967 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
4968 gimple_build_omp_teams): New functions.
4969 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
4970 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
4971 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
4972 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4973 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
4974 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
4975 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
4976 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
4977 GF_OMP_ATOMIC_SEQ_CST): New.
4978 (gimple_build_omp_taskgroup, gimple_build_omp_target,
4979 gimple_build_omp_teams): New prototypes.
4980 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
4981 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4982 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
4983 GIMPLE_OMP_SINGLE as end of range.
4984 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
4985 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
4986 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
4987 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
4988 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
4989 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
4990 gimple_omp_target_kind, gimple_omp_target_set_kind,
4991 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
4992 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
4993 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
4994 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
4995 gimple_omp_teams_set_clauses): New inlines.
4996 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
4997 and GIMPLE_OMP_TASKGROUP.
4998 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
4999 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
5000 enum omp_clause_proc_bind_kind): New.
5001 (union omp_clause_subcode): Add depend_kind, map_kind and
5002 proc_bind_kind fields.
5003 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
5004 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5005 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
5006 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
5007 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
5009 2013-10-10 Teresa Johnson <tejohnson@google.com>
5011 * predict.c (tree_estimate_probability): Add new parameter
5012 for estimate_bb_frequencies.
5013 (estimate_bb_frequencies): Add new parameter to force estimation.
5014 (rebuild_frequencies): When max frequency in function is small,
5015 recompute counts from frequencies.
5016 * predict.h (estimate_bb_frequencies): New parameter.
5018 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5020 * ipa-inline.c (ipa_inline): Fix leak of "order" when
5021 optimizations are disabled.
5023 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5025 * coverage.c (coverage_finish): Fix leak of da_file_name.
5027 2013-10-10 Jan Hubicka <jh@suse.cz>
5029 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
5030 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
5033 2013-10-10 Jakub Jelinek <jakub@redhat.com>
5036 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
5037 if any labels are in FALLTHRU_BB, use a special label emitted
5038 immediately after the asm goto insn rather than label_rtx
5040 (expand_asm_stmt): Adjust caller.
5041 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
5042 edge if the last insn in predecessor is a jump with single successor,
5043 but it isn't simplejump_p.
5045 2013-10-10 Richard Biener <rguenther@suse.de>
5047 PR tree-optimization/58656
5048 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
5050 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5052 * gimplify.c: Include expr.h and tm_p.h for targets with special
5053 va-arg padding requirements.
5055 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5057 * tree-flow.h: Move some prototypes to gimple.h.
5058 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
5059 * gimple.h: Relocate some prototypes from tree-flow.h
5060 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
5062 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
5063 (build_va_arg_indirect_ref): Relocate and make static.
5064 (std_gimplify_va_arg_expr): Relocate here.
5065 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
5066 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
5068 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5070 * doc/md.texi: Document the mnemonic attribute.
5072 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5075 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
5076 (set_attr_alternative x ...) when searching for user defined
5079 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5081 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
5083 2013-10-09 Easwaran Raman <eraman@google.com>
5085 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
5086 * cfgexpand.c (defer_stack_allocation): ...use here
5087 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
5089 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
5091 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
5092 (operand_equal_for_value_replacement): New function, extracted from
5093 value_replacement and enhanced to catch more cases.
5094 (value_replacement): Use operand_equal_for_value_replacement.
5096 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5098 * loop-doloop.c (doloop_modify, doloop_optimize): Use
5099 get_max_loop_iterations.
5101 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5103 * config/arm/aarch-common.c (arm_early_load_addr_dep):
5104 Place comment above function.
5106 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5108 * tree-flow.h: Remove all remaining prototypes, enums and structs that
5109 are not related to tree-cfg.c.
5110 * tree-ssa-address.h: New file. Relocate prototypes.
5111 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
5112 (addr_for_mem_ref): New. Combine call to get_address_description and
5113 return addr_for_mem_ref.
5114 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
5115 * tree-ssa-live.h: Adjust prototypes.
5116 * passes.c: Include tree-ssa-live.h.
5117 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
5118 * graphite.c (graphite_transform_loops): Make static.
5119 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5120 make_pass_graphite, pass_data_graphite_transforms,
5121 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
5122 * ipa-pure-const.c (warn_function_noreturn): Make static.
5123 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5124 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5125 Relocate from tree-cfg.c
5126 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
5128 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5129 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5130 Move to ipa-pure-const.c.
5131 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
5132 Relocate from tree-optimize.c.
5133 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
5134 make_pass_fixup_cfg): Move to tree-cfg.c.
5135 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
5136 Relocate some prototypes.
5137 * tree-data-ref.h (tree_check_data_deps) Add prototype.
5138 * tree-dump.c (dump_function_to_file): Remove prototype.
5139 Add tree-flow.h to the include file.
5140 * tree-dump.h: Remove prototype.
5141 * tree-parloops.h: New File. Add prototypes.
5142 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
5143 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
5144 from tree-ssa-loop.c.
5145 * tree-predcom.c (run_tree_predictive_commoning,
5146 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
5147 Relocate here from tree-ssa-loop.c.
5148 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
5149 ssa_name_values.release ().
5150 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
5151 (ssa_name_values): Relocate from tree-flow.h.
5152 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
5153 * tree-ssa-loop.c (run_tree_predictive_commoning,
5154 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
5155 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5156 make_pass_graphite, pass_data_graphite_transforms,
5157 make_pass_graphite_transforms, gate_tree_parallelize_loops,
5158 tree_parallelize_loops, pass_data_parallelize_loops,
5159 make_pass_parallelize_loops): Move to other files.
5160 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
5162 * tree.h: Remove prototypes from tree-address.c.
5164 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5166 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
5167 (struct int_tree_map): Move to tree-hasher.h
5168 (SCALE, LABEL, PERCENT): Move to gimple.h
5169 * tree-flow-inline.h: Delete. Move functions to other files.
5170 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
5171 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
5172 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
5173 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
5174 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
5175 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
5176 * tree-hasher.h: Don't include tree-flow.h.
5177 (struct int_tree_map): Relocate from tree-flow.h.
5178 * tree-sra.c (contains_view_convert_expr_p): Relocate from
5179 tree-flow-inline.h and make static.
5180 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
5182 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
5183 tree-flow-inline.h and make static.
5184 * tree.h (is_global_var, may_be_aliased): Relocate from
5186 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
5187 * value-prof.c: No longer include tree-flow-inline.h.
5188 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
5190 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5192 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
5193 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
5194 (enum move_pos): Move to tree-ssa-loop-im.h
5195 * cfgloop.h: Move some prototypes.
5196 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
5197 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
5198 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
5199 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
5200 (loop_containing_stmt): Relocate from tree-flow-inline.h.
5201 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
5202 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
5203 (enum move_pos): Relocate here.
5204 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
5206 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
5207 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
5208 make_pass_lim): Relocate here from tree-ssa-loop.c.
5209 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
5211 (loop_edge_to_cancel, unloop_loops): Make static.
5212 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5213 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
5214 (tree_complete_unroll, gate_tree_complete_unroll,
5215 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
5216 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5217 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
5218 * tree-ssa-loop-ivopts.c: Remove local prototypes.
5219 (stmt_invariant_in_loop_p): Remove unused function.
5220 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
5221 * tree-ssa-loop-manip.h: New file. Add prototypes.
5222 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
5223 (gcov_type_to_double_int): Move to cfgloop.h.
5224 (double_int_cmp, bound_index,
5225 estimate_numbers_of_iterations_loop): Make static.
5226 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
5227 (max_loop_iterations): Factor out get_max_loop_iterations.
5228 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
5230 * tree-ssa-loop-niter.h: New file. Add prototypes.
5231 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
5232 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
5233 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
5234 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
5235 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
5236 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
5237 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
5238 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
5239 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
5240 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
5241 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5242 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
5243 pass_data_complete_unroll, make_pass_complete_unroll,
5244 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5245 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
5246 tree-ssa-loop-ivcanon.c.
5247 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
5248 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
5249 tree-ssa-loop-prefetch.c.
5250 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
5252 (get_lsm_tmp_name): Relocate and add suffix parameter.
5253 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
5254 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
5255 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
5256 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
5257 (get_estimated_loop_iterations): Factor out accessor from
5258 estimated_loop_iterations in tree-ssa-loop-niter.c.
5259 (get_max_loop_iterations): Factor out accessor from
5260 _max_loop_iterations in tree-ssa-niter.c.
5261 * loop-unroll.c (decide_unroll_constant_iterations,
5262 decide_unroll_runtime_iterations, decide_peel_simple,
5263 decide_unroll_stupid): Use new get_* accessors.
5265 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
5267 PR tree-optimization/20318
5268 * doc/extend.texi (returns_nonnull): New function attribute.
5269 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
5271 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
5272 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
5274 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
5277 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
5278 false if both components are bitfields.
5280 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5282 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
5283 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
5284 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
5285 * config/aarch64/aarch64.h
5286 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
5287 * config/aarch64/aarch64-simd-builtins.def
5288 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
5290 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5292 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
5294 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5296 * config/aarch64/arm_neon.h (vdiv_f64): Added.
5298 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5300 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
5301 (vneg_f64): New intrinsic.
5302 (vneg_s8): Asm replaced with C.
5303 (vneg_s16): Likewise.
5304 (vneg_s32): Likewise.
5305 (vneg_s64): New intrinsic.
5306 (vnegq_f32): Asm replaced with C.
5307 (vnegq_f64): Likewise.
5308 (vnegq_s8): Likewise.
5309 (vnegq_s16): Likewise.
5310 (vnegq_s32): Likewise.
5311 (vnegq_s64): Likewise.
5313 2013-10-09 Renlin Li <Renlin.Li@arm.com>
5315 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
5317 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5319 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
5320 packed stack special handling.
5321 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
5322 back to fixed stack slots for FPRs saved due to stdarg.
5324 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5326 * config/s390/s390.c (s390_frame_info): Restructure function.
5328 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5330 * config/s390/s390.c (struct s390_frame_layout): New field
5332 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
5333 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
5334 regs_ever_clobbered to char*.
5335 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
5336 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
5337 (s390_register_info): Enable FPR save slots. Move/Copy some
5338 functionality into ...
5339 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
5340 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
5342 (s390_frame_info): Do gpr slot allocation here now. stdarg does
5343 not imply a stack frame.
5344 (s390_init_frame_layout): Remove variable clobbered_regs.
5345 (s390_update_register_info): Remove function.
5346 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
5347 cannot be used for register renaming.
5348 (s390_hard_regno_scratch_ok): New function.
5349 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
5350 (s390_initial_elimination_offset): Change offset calculation of
5351 the return address pointer.
5352 (save_gprs): Deal with only r6 being saved from the call-saved regs.
5353 (restore_gprs): Set frame related flag.
5354 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
5355 (s390_emit_prologue): Call s390_register_info instead of
5356 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
5357 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
5358 (s390_optimize_prologue): Call s390_optimize_register_info.
5359 Try to remove also FPR slot save/restore INSNs. Remove frame
5360 related flags from restore INSNs.
5362 2013-10-08 DJ Delorie <dj@redhat.com>
5364 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
5367 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
5368 avoid conflict with the MI use of %c.
5369 * config/rl78/rl78-real.md: change %c to %C throughout.
5370 * config/rl78/rl78-virt.md: Likewise.
5372 2013-10-08 Jan Hubicka <jh@suse.cz>
5374 * config/i386/i386.c (ix86_option_override_internal): Switch
5375 to SSE math for -ffast-math when target ISA supports SSE2.
5377 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
5379 * tree-flow.h: Remove some prototypes.
5380 * tree.h: Remove some protypes, add a couple.
5381 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
5382 using_eh_for_cleanups_p): Add interface routines for front ends.
5383 * tree-eh.h: New file. Add protoptyes.
5384 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
5385 (add_stmt_to_eh_lp_fn): Make static.
5386 (lower_try_finally): Use new using_eh_for_cleanups_p.
5387 * emit-rtl.c: Include tree-eh.h.
5388 * gimple.h: Include tree-eh.h.
5390 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
5392 PR tree-optimization/58480
5393 * tree-vrp.c (infer_nonnull_range): New function.
5394 (infer_value_range): Call infer_nonnull_range.
5396 2013-10-08 Dehao Chen <dehao@google.com>
5398 PR tree-optimization/58619
5399 * tree-inline.c (copy_phis_for_bb): Combine location data
5402 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5405 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
5406 RTX_FRAME_RELATED_P on INSN.
5408 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5410 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
5411 (altivec_expand_vec_perm_const): Call it.
5413 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5415 * config/rs6000/vector.md (mov<mode>): Emit permuted move
5416 sequences for LE VSX loads and stores at expand time.
5417 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
5419 * config/rs6000/rs6000.c (rs6000_const_vec): New.
5420 (rs6000_gen_le_vsx_permute): New.
5421 (rs6000_gen_le_vsx_load): New.
5422 (rs6000_gen_le_vsx_store): New.
5423 (rs6000_gen_le_vsx_move): New.
5424 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
5425 (*vsx_le_perm_load_v4si): New.
5426 (*vsx_le_perm_load_v8hi): New.
5427 (*vsx_le_perm_load_v16qi): New.
5428 (*vsx_le_perm_store_v2di): New.
5429 (*vsx_le_perm_store_v4si): New.
5430 (*vsx_le_perm_store_v8hi): New.
5431 (*vsx_le_perm_store_v16qi): New.
5432 (*vsx_xxpermdi2_le_<mode>): New.
5433 (*vsx_xxpermdi4_le_<mode>): New.
5434 (*vsx_xxpermdi8_le_V8HI): New.
5435 (*vsx_xxpermdi16_le_V16QI): New.
5436 (*vsx_lxvd2x2_le_<mode>): New.
5437 (*vsx_lxvd2x4_le_<mode>): New.
5438 (*vsx_lxvd2x8_le_V8HI): New.
5439 (*vsx_lxvd2x16_le_V16QI): New.
5440 (*vsx_stxvd2x2_le_<mode>): New.
5441 (*vsx_stxvd2x4_le_<mode>): New.
5442 (*vsx_stxvd2x8_le_V8HI): New.
5443 (*vsx_stxvd2x16_le_V16QI): New.
5445 2013-10-07 Renlin Li <Renlin.Li@arm.com>
5447 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
5449 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5451 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
5452 loop to work also for 31bit ABI.
5453 Save the stack pointer for frame_size > 0.
5455 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5457 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
5458 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
5459 constraint letters from expanders.
5460 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
5461 retry count to general_operand.
5462 ("tabort"): Give operand 0 a mode.
5463 ("tabort_1"): Add mode and constraint letter for operand 0.
5464 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
5466 2013-10-04 Jeff Law <law@redhat.com>
5468 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
5470 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
5472 (thread_across_edge): Here. Call it.
5474 2013-10-04 Cary Coutant <ccoutant@google.com>
5476 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
5477 discarding a location list expression (or a piece of one).
5479 2013-10-03 Jan Hubicka <jh@suse.cz>
5481 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
5482 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
5483 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
5485 2013-10-03 Jan Hubicka <jh@suse.cz>
5487 * config/i386/i386.c (ix86_option_override_internal): Do not enable
5488 accumulate-outgoing-args when producing unwind info.
5490 2013-10-03 Wei Mi <wmi@google.com>
5492 * lra-constraints.c (insert_move_for_subreg): New function
5493 extracted from simplify_operand_subreg.
5494 (simplify_operand_subreg): Add reload for paradoxical subreg.
5496 2013-10-03 Rong Xu <xur@google.com>
5498 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
5499 the candidate of builtin_expect such that we should fix the
5500 size/time estimation.
5501 (estimate_function_body_sizes): Do the acutally size/time fix-up
5504 2013-10-03 Rong Xu <xur@google.com>
5506 * predict.c (tree_predict_by_opcode): Get the probability
5507 for builtin_expect from param builtin_expect_probability.
5508 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
5509 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
5510 * doc/invoke.texi: Add documentation for builtin-expect-probability.
5512 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
5515 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
5516 * calls.c (alloca_call_p): Use get_callee_fndecl.
5517 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
5518 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
5520 (vrp_visit_stmt): Remove duplicated code.
5522 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
5524 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
5525 ceildf2, btruncdf2, instead of vsx_* name.
5527 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
5528 iterators to only do V2DF and V4SF here. Move the DF code to
5529 rs6000.md where it is combined with SF mode. Replace <VSv> with
5530 just 'v' since only vector operations are handled with these insns
5531 after moving the DF support to rs6000.md.
5532 (vsx_sub<mode>3): Likewise.
5533 (vsx_mul<mode>3): Likewise.
5534 (vsx_div<mode>3): Likewise.
5535 (vsx_fre<mode>2): Likewise.
5536 (vsx_neg<mode>2): Likewise.
5537 (vsx_abs<mode>2): Likewise.
5538 (vsx_nabs<mode>2): Likewise.
5539 (vsx_smax<mode>3): Likewise.
5540 (vsx_smin<mode>3): Likewise.
5541 (vsx_sqrt<mode>2): Likewise.
5542 (vsx_rsqrte<mode>2): Likewise.
5543 (vsx_fms<mode>4): Likewise.
5544 (vsx_nfma<mode>4): Likewise.
5545 (vsx_copysign<mode>3): Likewise.
5546 (vsx_btrunc<mode>2): Likewise.
5547 (vsx_floor<mode>2): Likewise.
5548 (vsx_ceil<mode>2): Likewise.
5549 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
5550 (vsx_sminsf3): Likewise.
5551 (vsx_fmadf4): Likewise.
5552 (vsx_fmsdf4): Likewise.
5553 (vsx_nfmadf4): Likewise.
5554 (vsx_nfmsdf4): Likewise.
5555 (vsx_cmpdf_internal1): Likewise.
5557 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
5558 simpler to select whether a target has SPE or traditional floating
5559 point support in iterators.
5560 (TARGET_DF_SPE): Likewise.
5561 (TARGET_SF_FPR): Likewise.
5562 (TARGET_DF_FPR): Likewise.
5563 (TARGET_SF_INSN): Macros to say whether floating point support
5564 exists for a given operation for expanders.
5565 (TARGET_DF_INSN): Likewise.
5567 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
5568 combining of SF/DF mode operations, using both traditional and VSX
5576 (abs<mode>2): Combine SF/DF modes using traditional floating point
5577 instructions. Add support for using the upper DF registers with
5578 VSX support, and SF registers with power8-vector support. Update
5579 expanders for operations supported by both the SPE and traditional
5580 floating point units.
5581 (abs<mode>2_fpr): Likewise.
5582 (nabs<mode>2): Likewise.
5583 (nabs<mode>2_fpr): Likewise.
5584 (neg<mode>2): Likewise.
5585 (neg<mode>2_fpr): Likewise.
5586 (add<mode>3): Likewise.
5587 (add<mode>3_fpr): Likewise.
5588 (sub<mode>3): Likewise.
5589 (sub<mode>3_fpr): Likewise.
5590 (mul<mode>3): Likewise.
5591 (mul<mode>3_fpr): Likewise.
5592 (div<mode>3): Likewise.
5593 (div<mode>3_fpr): Likewise.
5594 (sqrt<mode>3): Likewise.
5595 (sqrt<mode>3_fpr): Likewise.
5596 (fre<Fs>): Likewise.
5597 (rsqrt<mode>2): Likewise.
5598 (cmp<mode>_fpr): Likewise.
5599 (smax<mode>3): Likewise.
5600 (smin<mode>3): Likewise.
5601 (smax<mode>3_vsx): Likewise.
5602 (smin<mode>3_vsx): Likewise.
5603 (negsf2): Delete SF operations that are merged with DF.
5610 (fmasf4_fpr): Likewise.
5611 (fmssf4_fpr): Likewise.
5612 (nfmasf4_fpr): Likewise.
5613 (nfmssf4_fpr): Likewise.
5614 (sqrtsf2): Likewise.
5615 (rsqrtsf_internal1): Likewise.
5616 (smaxsf3): Likewise.
5617 (sminsf3): Likewise.
5618 (cmpsf_internal1): Likewise.
5619 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
5620 (negdf2): Delete DF operations that are merged with SF.
5622 (nabsdf2): Likewise.
5628 (rsqrtdf_internal1): Likewise.
5629 (fmadf4_fpr): Likewise.
5630 (fmsdf4_fpr): Likewise.
5631 (nfmadf4_fpr): Likewise.
5632 (nfmsdf4_fpr): Likewise.
5633 (sqrtdf2): Likewise.
5634 (smaxdf3): Likewise.
5635 (smindf3): Likewise.
5636 (cmpdf_internal1): Likewise.
5637 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
5638 (btrunc<mode>2): Delete separate expander, and combine with the
5639 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
5640 (btrunc<mode>2_fpr): Likewise.
5641 (ceil<mode>2): Likewise.
5642 (ceil<mode>2_fpr): Likewise.
5643 (floor<mode>2): Likewise.
5644 (floor<mode>2_fpr): Likewise.
5645 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
5646 Add support for using the upper registers with VSX and
5647 power8-vector. Move insns to be closer to the define_expands. On
5648 VSX systems, prefer the traditional form of FMA over the VSX
5649 version, since the traditional form allows the target not to
5650 overlap with the inputs.
5651 (fms<mode>4_fpr): Likewise.
5652 (nfma<mode>4_fpr): Likewise.
5653 (nfms<mode>4_fpr): Likewise.
5655 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5656 Richard Earnshaw <richard.earnshaw@arm.com>
5658 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
5659 (struct mult_cost_table): Likewise.
5660 (struct mem_cost_table): Likewise.
5661 (struct fp_cost_table): Likewise.
5662 (struct vector_cost_table): Likewise.
5663 (cpu_cost_table): Likewise.
5664 * config/arm/arm.opt (mold-rts-costs): New option.
5665 (mnew-generic-costs): Likewise.
5666 * config/arm/arm.c (generic_extra_costs): New table.
5667 (cortexa15_extra_costs): Likewise.
5668 (arm_slowmul_tune): Use NULL as new costs.
5669 (arm_fastmul_tune): Likewise.
5670 (arm_strongarm_tune): Likewise.
5671 (arm_xscale_tune): Likewise.
5672 (arm_9e_tune): Likewise.
5673 (arm_v6t2_tune): Likewise.
5674 (arm_cortex_a5_tune): Likewise.
5675 (arm_cortex_a9_tune): Likewise.
5676 (arm_v6m_tune): Likewise.
5677 (arm_fa726te_tune): Likewise.
5678 (arm_cortex_a15_tune): Use cortex15_extra_costs.
5679 (arm_cortex_tune): Use generict_extra_costs.
5680 (shifter_op_p): New function.
5681 (arm_unspec_cost): Likewise.
5682 (LIBCALL_COST): Define.
5683 (arm_new_rtx_costs): New function.
5684 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
5685 table is available. Use old costs otherwise unless mnew-generic-costs
5687 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
5688 (cpu_cost_table): Declare.
5690 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
5693 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
5694 (*subs_mul_imm_<mode>)
5695 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
5696 (*sub_<shift>_<mode>)
5697 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
5698 Remove k constraint.
5700 2013-10-03 Ian Bolton <ian.bolton@arm.com>
5702 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
5704 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
5706 2013-10-02 Teresa Johnson <tejohnson@google.com>
5708 * predict.c (probably_never_executed): New function.
5709 (probably_never_executed_bb_p): Invoke probably_never_executed.
5710 (probably_never_executed_edge_p): Ditto.
5711 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
5712 Treat profile insanities conservatively.
5714 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
5716 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
5718 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
5720 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
5721 use it. Use smaller increase for scratch. Don't increase reject
5722 for early clobber scratch.
5723 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
5724 setting eliminated regs except setting fp from hfp.
5725 (lra_eliminate): Check lra_insn_recog_data on NULL.
5727 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5730 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
5731 setting -mvsx-timode by default until the underlying problem is fixed.
5732 (RS6000_CPU, power7 defaults): Likewise.
5734 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5736 * config/x-linux (host-linux.o): Remove header dependencies.
5737 Use $(COMPILE) and $(POSTCOMPILE).
5738 * config/t-linux-android (linux-android.o): Ditto.
5740 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5742 * Makefile.in (expmed.o-warn): Remove.
5744 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5746 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
5747 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
5749 2013-10-02 Teresa Johnson <tejohnson@google.com>
5751 * dojump.c (do_jump_1): Divide probability between
5752 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
5754 2013-10-02 Tom Tromey <tromey@redhat.com>
5756 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
5758 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5760 * tree-flow.h: Remove some prototypes.
5761 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
5762 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
5763 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
5764 mark_virtual_phi_result_for_renaming): Relocate here.
5765 * tree-into-ssa.h: Add prototypes.
5766 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
5767 single_pred_before_succ_order.
5768 (blocks_in_phiopt_order): Rename and move to cfganal.c.
5769 (nonfreeing_call_p) Move to gimple.c.
5770 * cfganal.c (single_pred_before_succ_order): Move and renamed from
5772 * basic-block.h (single_pred_before_succ_order): Add prototype.
5773 * gimple.c (nonfreeing_call_p): Relocate here.
5774 * gimple.h: Add prototype.
5775 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
5776 * tree-ssa-dom.h: New file. Relocate prototypes here.
5777 * tree-ssa.h: Include tree-ssa-dom.h.
5779 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5781 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
5782 Use $(COMPILE) and $(POSTCOMPILE).
5784 * config/alpha/x-alpha (driver-alpha.o): Ditto.
5786 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5788 * tree-flow.h: Remove some prototypes.
5789 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
5790 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5791 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
5792 propagate_tree_value*): Move from here to...
5793 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
5794 propagate_tree_value*): Relocate here.
5795 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5796 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
5797 * gimple-fold.c: Remove gimple-fold.h from include list.
5798 * tree-vrp.c: Remove gimple-fold.h from include list.
5799 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
5800 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
5801 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
5802 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
5803 * sese.c: Add tree-ssa-propagate.h to include list.
5805 2013-10-02 Richard Biener <rguenther@suse.de>
5807 * tree-loop-distribution.c: Include tree-vectorizer.h for
5809 (enum partition_kind): Remove PKIND_REDUCTION.
5810 (struct partition_s): Remove has_writes member, add reduction_p member.
5811 (partition_alloc): Adjust.
5812 (partition_builtin_p): Likewise.
5813 (partition_has_writes): Remove.
5814 (partition_reduction_p): New function.
5815 (partition_merge_into): Likewise.
5816 (generate_code_for_partition): Commonize builtin partition
5818 (rdg_cannot_recompute_vertex_p): Remove.
5819 (already_processed_vertex_p): Likewise.
5820 (rdg_flag_vertex): Do not set has_writes.
5821 (classify_partition): Adjust.
5822 (rdg_build_partitions): Do not set has_writes, treat all
5823 partitions as useful.
5824 (distribute_loop): Record number of library calls generated. Adjust.
5825 (tree_loop_distribution): Report number of loops and library
5826 calls generated as opt-info.
5828 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5830 * tree-flow.h: Include new .h files. Move prototypes.
5831 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
5832 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
5833 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
5834 * tree-pretty-print.h: Add prototypes from tree-flow.h.
5835 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
5836 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
5837 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
5838 * tree.h (get_ref_base_and_extent): Move prototype out.
5839 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
5841 * gimple-low.h: New File. Add prototypes from tree-flow.h.
5842 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
5843 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
5844 * tree-scalar-evolution.c: Include tree.h.
5845 * sese.c: Include tree.h.
5846 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
5847 * dwarf2out.c: Include tree-dfa.h.
5848 * tree-chrec.c: Include tree.h.
5849 * tree-data-ref.c: Include tree.h.
5851 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
5853 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
5856 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5858 * config/t-sol2 (sol2-c.o): Remove header dependencies.
5859 Use $(COMPILE) and $(POSTCOMPILE).
5860 (sol2-cxx.o): Likewise.
5861 (sol2-stubs.o): Likewise.
5863 * config/x-solaris (host-solaris.o): Likewise.
5865 * config/sparc/t-sparc (sparc.o): Remove.
5866 (sparc-c.o): Remove header dependencies.
5867 Use $(COMPILE) and $(POSTCOMPILE).
5868 * config/sparc/x-sparc: Likewise.
5870 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
5872 * config/arc/arc-opts.h: Add 2013 to Copyright years.
5873 * config/arc/arc700.md: Likewise.
5874 * config/arc/arc-modes.def: Likewise.
5875 * config/arc/arc-simd.h: Likewise.
5876 * config/arc/t-arc-uClibc: Likewise.
5877 * config/arc/t-arc-newlib: Likewise.
5879 2013-10-02 Renlin Li <renlin.li@arm.com>
5881 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
5883 (aarch64_expand_epilogue): Likewise.
5885 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5886 Yufeng Zhang <yufeng.zhang@arm.com>
5888 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
5890 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
5891 'base_in' represent a conversion and legal_cast_p_1 holds; set
5892 'base_in' with the returned value from get_unwidened.
5894 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5896 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
5897 plus_constant is not used.
5899 2013-10-01 Wei Mi <wmi@google.com>
5901 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
5902 * config/i386/i386.md: Add define_peephole2 to
5903 break partial reg stall for cvtss2sd/cvtsd2ss.
5905 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5907 * config/arc/arc.c (pass_arc_ifcvt::clone):
5908 Update for ctxt_ -> m_ctxt change.
5910 2013-10-01 Jeff Law <law@redhat.com>
5912 * tree-ssa-threadupdate.c (struct redirection_data): Delete
5913 outgoing_edge and intermediate_edge fields. Instead store the path.
5914 (redirection_data::hash): Hash on the last edge's destination index.
5915 (redirection_data::equal): Check the entire thread path.
5916 (lookup_redirectio_data): Corresponding changes.
5917 (create_edge_and_update_destination_phis): Likewise.
5918 (thread_single_edge): Likewise.
5920 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5921 Diego Novillo <dnovillo@google.com>
5923 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
5924 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
5925 (vld32wh_insn, vld32wl_insn): Delete commented-out old
5926 versions of these patterns.
5928 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
5929 (__builtin_arc_aligned): Likewise.
5931 * config/arc/arc.md: Expand adc_0 comment stating the intended
5932 purpose and why it isn't ready.
5933 Replace commented out call_value_via_label_mixed with a
5934 plain comment about bl_s.
5936 * config/arc/arc.c (stdio.h): Don't include directly.
5937 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
5938 Remove if (1) condition.
5939 (arc_encode_section_info): Fix comment.
5941 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5943 * config/arc/arc.c (arc_conditional_register_usage):
5944 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
5945 Also set reg_alloc_order for DMA config regs.
5947 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5948 Jeremy Bennett <jeremy.bennett@embecosm.com>
5950 * doc/install.texi (--with-cpu): Mention ARC.
5951 (arc-*-elf32): New paragraph.
5952 (arc-linux-uclibc): Likewise.
5953 * doc/md.texi (Machine Constraints): Add ARC part.
5954 * doc/invoke.texi: (menu): Add ARC Options.
5955 (Machine Dependent Options) <ARC Options>: Add synopsis.
5956 (node ARC Options): Add.
5957 * doc/extend.texi (long_call / short_call attribute): Add ARC.
5958 (ARC Built-in Functions): New section defining
5959 generic ARC built-in functions.
5960 (ARC SIMD Built-in Functions): New section defining SIMD specific
5962 (Declaring Attributes of Functions): Extended
5963 description of short_call and long_call attributes for ARC and
5964 added index entries.
5966 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
5967 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
5968 Joern Rennecke <joern.rennecke@embecosm.com>
5969 Muhammad Khurram Riaz <khurram.riaz@arc.com>
5970 Brendan Kehoe <brendan@zen.org>
5971 Michael Eager <eager@eagercon.com>
5972 Simon Cook <simon.cook@embecosm.com>
5973 Jeremy Bennett <jeremy.bennett@embecosm.com>
5975 * config/arc, common/config/arc: New directories.
5977 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5978 Brendan Kehoe <brendan@zen.org>
5979 Simon Cook <simon.cook@embecosm.com>
5981 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
5983 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
5985 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
5986 * tree-ssa-coalesce.h: New. Move prototype to here.
5987 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
5988 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
5989 (gimple_can_coalesce_p): Move to...
5990 * gimple.c (gimple_can_coalesce_p): Here.
5992 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
5994 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
5995 (dump_decl_set): Move to gimple.c.
5996 * gimple.h: Don't include tree-ssa-operands.h.
5997 (dump_decl_set): Add prototype.
5998 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
5999 Move to gimple-ssa.h.
6000 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
6001 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
6002 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
6003 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
6004 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
6006 (dump_decl_set): Relocate here.
6007 * gimple-ssa.h: New file.
6008 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6009 Relocate from gimple.h.
6010 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
6011 * tree-ssa-operands.c (swap_ssa_operands): Rename from
6012 swap_tree_operands and remove non-ssa path.
6013 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
6014 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
6016 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
6017 vect_is_simple_reduction_1): Use swap_ssa_operands.
6018 * tree-flow.h: Move various prototypes to tree-phinodes.h.
6019 (enum need_phi_state): Move to tree-into-ssa.c.
6020 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6021 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
6022 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
6023 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
6024 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6025 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6026 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6027 num_imm_uses): Move to ssa-iterators.h.
6028 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
6029 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
6031 (op_iter_done, op_iter_next_def, op_iter_next_tree,
6032 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
6033 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
6034 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
6035 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
6036 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
6037 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
6038 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
6039 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
6040 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
6041 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
6042 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
6043 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
6044 (set_phi_nodes): Move to tree-phinodes.h.
6045 * tree-ssa-operands.h (enum ssa_op_iter_type,
6046 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
6047 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
6048 (dump_decl_set): Remove prototype.
6049 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
6050 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
6051 (set_phi_nodes): Relocate from tree-flow-inline.h.
6052 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
6054 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
6055 include list. Temporarily add gimple.h to include list.
6056 * ssa-iterators.h: New file.
6057 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6058 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
6059 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
6060 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
6061 Relocate from tree-ssa-operands.h.
6062 (delink_imm_use, link_imm_use_to_list, link_imm_use,
6063 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6064 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6065 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6066 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
6067 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
6068 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
6069 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
6070 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
6071 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
6072 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
6073 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
6074 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
6075 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
6076 Relocate from tree-flow-inline.h.
6077 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
6079 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
6082 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
6083 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
6084 Insert '\t' to output template.
6085 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
6086 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
6087 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
6088 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
6090 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
6092 * doc/install.texi (Host/target specific installation notes for GCC):
6093 Put @anchor before @heading.
6094 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
6095 Use @email for email addresses.
6097 2013-10-01 Jeff Law <law@redhat.com>
6099 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
6100 a vec. Only delete the path if we create one without successfully
6101 registering a jump thread.
6102 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
6104 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
6105 (paths): New vector of jump threading paths.
6106 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
6107 (THREAD_PATH): New accessor macro for the entire thread path.
6108 (lookup_redirection_data): Get intermediate and final outgoing edge
6109 from the thread path.
6110 (create_edge_and_update_destination_phis): Copy the threading path.
6111 (ssa_fix_duplicate_block_edges): Get edges and block types from the
6112 jump threading path.
6113 (ssa_redirect_edges): Get edges and block types from the jump threading
6114 path. Free the path vector.
6115 (thread_block): Get edges from the jump threading path. Look at the
6116 entire path to see if we thread to a loop exit. If we cancel a jump
6117 thread request, then free the path vector.
6118 (thread_single_edge): Get edges and block types from the jump threading
6119 path. Free the path vector.
6120 (thread_through_loop_header): Get edges and block types from the jump
6121 threading path. Free the path vector.
6122 (mark_threaded_blocks): Iterate over the vector of paths and store
6123 the path on the appropriate edge. Get edges and block types from the
6124 jump threading path.
6125 (mark_threaded_blocks): Get edges and block types from the jump
6126 threading path. Free the path vector.
6127 (thread_through_all_blocks): Use the vector of paths rather than
6128 a vector of 3-edge sets.
6129 (register_jump_thread): Accept pointer to a path vector rather
6130 than the path vector itself. Store the path vector for later use.
6133 2013-10-01 Jakub Jelinek <jakub@redhat.com>
6134 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6137 * config/s390/s390.c (s390_split_branches): Modify check for table
6139 (s390_chunkify_start): Rearrange table jump insn check in order to
6140 deal with compare and branch insns correctly.
6142 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
6146 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
6147 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
6148 define_insn_and_split.
6149 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
6150 (shiftsi3_compare): New pattern.
6152 * config/arm/unspecs.md (UNSPEC_RRX): New.
6154 2013-10-01 Alan Modra <amodra@gmail.com>
6156 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
6157 casing inout operands.
6159 2013-10-01 Richard Biener <rguenther@suse.de>
6161 PR tree-optimization/58553
6162 * tree-loop-distribution.c (struct partition_s): Add niter member.
6163 (classify_partition): Populate niter member for the partition
6164 and properly identify whether the relevant store happens before
6165 or after the loop exit.
6166 (generate_memset_builtin): Use niter member from the partition.
6167 (generate_memcpy_builtin): Likewise.
6169 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6171 * vec.h (vec_prefix, vec): Prefix member names with "m_".
6172 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
6174 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6176 * basic-block.h (edge_list): Prefix member names with "m_".
6177 * context.h (context): Likewise.
6178 * domwalk.h (dom_walker): Likewise.
6179 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
6180 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
6181 * hash-table.h (hash_table): Likewise.
6182 * machmode.h (bit_field_mode_iterator): Likewise.
6183 * pass_manager.h (pass_list): Likewise.
6184 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
6185 * tree-pass.h (pass_data): Likewise.
6186 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
6187 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
6188 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
6189 * asan.c (pass_data_asan): Update accordingly.
6190 * cfganal.c (control_dependences::find_control_dependence): Likewise.
6191 (control_dependences::control_dependences): Likewise.
6192 (control_dependences::~control_dependences): Likewise.
6193 (control_dependences::~control_dependences): Likewise.
6194 (control_dependences::get_edges_dependent_on): Likewise.
6195 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
6196 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
6197 * context.c (gcc::context::context): Likewise.
6198 * cprop.c (pass_rtl_cprop::clone): Likewise.
6199 * domwalk.c (dom_walker::walk): Likewise.
6200 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
6201 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
6202 * mode-switching.c (pass_mode_switching::clone): Likewise.
6203 * passes.c (opt_pass::opt_pass): Likewise.
6204 (pass_manager::pass_manager): Likewise.
6205 * predict.c (pass_strip_predict_hints::clone): Likewise.
6206 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
6207 (pass_split_all_insns::clone): Likewise.
6208 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
6210 (bit_field_mode_iterator::next_mode): Likewise.
6211 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
6212 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
6213 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
6214 * tree-complex.c (pass_lower_complex::clone): Likewise.
6215 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
6216 * tree-object-size.c (pass_object_sizes::clone): Likewise.
6217 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
6218 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
6219 (pass_fold_builtins::clone): Likewise.
6220 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
6221 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
6222 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
6223 (pass_cd_dce::clone): Likewise.
6224 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
6225 (pass_phi_only_cprop::clone): Likewise.
6226 * tree-ssa-dse.c (pass_dse::clone): Likewise.
6227 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
6228 * tree-ssa-loop.c (pass_lim::clone): Likewise.
6229 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
6230 * tree-ssa-pre.c (pass_fre::clone): Likewise.
6231 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
6232 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
6233 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
6234 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
6235 * tree-vrp.c (pass_vrp::clone): Likewise.
6236 * tsan.c (pass_tsan::clone): Likewise.
6238 2013-09-30 Jakub Jelinek <jakub@redhat.com>
6241 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
6242 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
6245 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
6246 optimization, punt if sign_bit_p looked through any zero extension.
6248 2013-09-30 Teresa Johnson <tejohnson@google.com>
6250 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
6251 Update redirected out edge count in joiner case.
6252 (ssa_redirect_edges): Common the joiner and non-joiner cases
6253 so that joiner case gets profile updates.
6255 2013-09-30 Richard Biener <rguenther@suse.de>
6257 PR tree-optimization/58554
6258 * tree-loop-distribution.c (classify_partition): Require
6259 unconditionally executed stores for memcpy and memset recognition.
6260 (tree_loop_distribution): Calculate dominance info.
6262 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
6264 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
6265 (NO_PROFILE_COUNTERS): Likewise.
6266 (PROFILE_HOOK): Likewise.
6267 (FUNCTION_PROFILER): Likewise.
6268 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
6270 2013-09-30 Iain Sandoe <iain@codesourcery.com>
6272 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
6273 calls and defines in TARGET_MACHO conditional.
6274 (load_macho_picbase_di): Likewise.
6275 (reload_macho_picbase): Likewise.
6276 (reload_macho_picbase_si): Likewise.
6277 (reload_macho_picbase_di): Likewise.
6278 (nonlocal_goto_receiver): Likewise.
6280 2013-09-30 Nick Clifton <nickc@redhat.com>
6282 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
6283 that use the MSP430X ISA.
6284 (msp430_option_override): Scan -mmcu command line option for any
6285 MCU name that supports the MSP430X ISA.
6286 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
6287 -mmcu options which enable the MSP430X ISA.
6289 2013-09-30 Richard Biener <rguenther@suse.de>
6292 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
6293 before looking for setjmp-like calls.
6295 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6298 * config/darwin-protos.h (machopic_get_function_picbase): New.
6299 * config/darwin.c (machopic_get_function_picbase): New.
6300 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
6301 label for a new func. (load_macho_picbase_di): Likewise.
6302 (reload_macho_picbase): New expand.
6303 (reload_macho_picbase_si): New insn.
6304 (reload_macho_picbase_di): New insn.
6305 (nonlocal_goto_receiver): New define and split.
6306 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
6307 (unspecv enum): Add UNSPECV_NLGR.
6309 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6311 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
6312 that altivec registers are correctly sized on Darwin.
6314 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6316 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
6317 darwin-driver.o): Use COMPILE and POSTCOMPILE.
6318 * config/x-darwin (host-darwin.o): Likewise.
6319 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
6320 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
6321 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
6323 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
6325 * doc/invoke.texi: Fix usage of @tie{} command.
6327 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
6329 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
6331 2013-09-28 Jan Hubicka <jh@suse.cz>
6333 * config/i386/x86-tune.def: Add documentation for each of the options;
6336 2013-09-28 Jan Hubicka <jh@suse.cz>
6338 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
6340 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
6341 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
6342 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
6343 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
6344 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
6346 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
6348 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
6349 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
6350 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
6351 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
6352 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
6353 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
6354 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
6355 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
6356 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
6357 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
6358 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
6359 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
6360 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
6361 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
6362 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
6363 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
6364 gimple.h, godump.c, graphite-clast-to-gimple.c,
6365 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
6366 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
6367 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
6368 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6369 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
6370 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
6371 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
6372 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
6373 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
6374 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
6375 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
6376 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
6377 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
6378 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
6379 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
6380 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
6381 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
6382 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
6383 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
6384 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
6385 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
6386 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
6387 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
6388 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
6389 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6390 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6391 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
6392 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
6393 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
6394 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6395 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6396 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
6397 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
6398 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6399 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6400 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
6401 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
6402 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
6403 whitespace before "(".
6405 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
6407 * expr.h (extract_bit_field): Remove packedp parameter.
6408 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
6409 from forward declaration.
6410 (store_split_bit_field): Remove packedp arg from calls to
6411 extract_fixed_bit_field.
6412 (extract_bit_field_1): Remove packedp parameter and packedp
6413 argument from recursive calls and calls to extract_fixed_bit_field.
6414 (extract_bit_field): Remove packedp parameter and corresponding
6415 arg to extract_bit_field_1.
6416 (extract_fixed_bit_field): Remove packedp parameter. Remove code
6418 (extract_split_bit_field): Remove packedp arg from call to
6419 extract_fixed_bit_field.
6420 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
6421 (copy_blkmode_from_reg): Likewise.
6422 (copy_blkmode_to_reg): Likewise.
6423 (read_complex_part): Likewise.
6424 (store_field): Likewise.
6425 (expand_expr_real_1): Likewise.
6426 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
6427 to extract_bit_field.
6428 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
6429 call to extract_bit_field.
6430 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
6431 call to extract_bit_field.
6432 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
6433 and special packedp behavior from -fstrict-volatile-bitfields
6436 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6438 * lra-eliminations.c (init_elim_table): Guard value_p.
6440 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6442 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
6443 DFmode, DImode, and SFmode in the upper VSX registers based on the
6444 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
6445 if -mpower8-vector. Combine -mvsx-timode handling with the rest
6446 of the VSX register handling.
6448 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
6450 (zero_extendsidi2_lfiwzx): Add support for loading into the
6451 Altivec registers with -mpower8-vector. Use wu/wv constraints to
6452 only do VSX memory options on Altivec registers.
6453 (extendsidi2_lfiwax): Likewise.
6454 (extendsfdf2_fpr): Likewise.
6455 (mov<mode>_hardfloat, SF/SD modes): Likewise.
6456 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
6457 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
6458 (movdi_internal64): Likewise.
6460 2013-09-27 Xinliang David Li <davidxl@google.com>
6462 * opts.c (finish_options): Adjust parameters
6463 according to vect cost model.
6464 (common_handle_option): Set dynamic vect cost
6466 targhooks.c (default_add_stmt_cost): Compute stmt cost
6468 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
6469 Use helper function.
6470 * tree-vectorizer.h (unlimited_cost_model): New function.
6471 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
6472 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
6474 (vect_enhance_data_refs_alignment): Ditto.
6475 * flag-types.h: New enum.
6476 * common/config/i386/i386-common.c (ix86_option_init_struct):
6477 No need to initialize vect_cost_model flag.
6478 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
6481 2013-09-27 Diego Novillo <dnovillo@google.com>
6483 * gimple.h (enum ssa_mode): Remove.
6485 2013-09-27 Paulo Matos <pmatos@broadcom.com>
6487 * cfgloop.h (number_of_loops): Fix typo in check for null.
6489 2013-09-27 Jakub Jelinek <jakub@redhat.com>
6492 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
6493 are children of outermost saved_cfun's loop, and set it up to
6494 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
6495 if loop != loop0 and SESE region contains bbs that belong to loop0.
6497 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6499 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
6500 (binary_scale_code_p, get_base_term, get_index_term): New functions.
6501 (set_address_segment, set_address_base, set_address_index)
6502 (set_address_disp): Accept the argument unconditionally.
6503 (baseness): Remove must_be_base_p and must_be_index_p checks.
6504 (decompose_normal_address): Classify as much as possible in the
6507 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6509 * cse.c (count_reg_usage): Handle INT_LIST.
6510 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
6511 * reginfo.c (reg_scan_mark_refs): Likewise.
6512 * reload1.c (eliminate_regs_1): Likewise.
6514 2013-09-27 Iain Sandoe <iain@codesourcery.com>
6517 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
6520 2013-09-27 Richard Biener <rguenther@suse.de>
6522 PR tree-optimization/58459
6523 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
6524 restriction not propagating into loops.
6526 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
6528 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
6529 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
6530 * doc/tree-ssa.texi (Walking use-def chains): Delete.
6532 2013-09-26 Richard Biener <rguenther@suse.de>
6534 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
6536 2013-09-26 Richard Biener <rguenther@suse.de>
6538 * alias.h (component_uses_parent_alias_set): Rename to ...
6539 (component_uses_parent_alias_set_from): ... this.
6540 * alias.c (component_uses_parent_alias_set): Rename to ...
6541 (component_uses_parent_alias_set_from): ... this and return
6543 (reference_alias_ptr_type_1): Use the result from
6544 component_uses_parent_alias_set_from instead of stripping
6545 components one at a time.
6546 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
6548 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6550 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
6551 Move prototypes to...
6552 * tree-ssa-ter.h: New File. Move prototypes here.
6553 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
6554 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
6555 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
6556 from is_replaceable_p.
6557 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
6558 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
6559 (process_replaceable): Use ter_is_replaceable_p.
6560 (find_replaceable_in_bb): Use ter_is_replaceable_p.
6561 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
6562 newly refactored ssa_is_replaceable_p.
6563 * cfgexpand.c: Include tree-outof-ssa.h.
6564 * ssaexpand.h: Delete.
6566 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6568 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
6569 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
6571 (create_gimple_tmp): Delete.
6572 (get_expr_type, build_assign, build_type_cast): Move to...
6573 * gimple-builder.c: New File.
6574 (get_expr_type): Relocate from gimple.c.
6575 (build_assign, build_type_cast): Change to only create ssanames.
6576 * gimple.h: Move prototypes to...
6577 * gimple-builder.h: New File. Here.
6578 * tree-ssa.h: And here.
6579 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
6580 count_uses_and_derefs): Relocate from gimple.c.
6581 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
6582 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
6583 * tree-ssa-math-opts (execute_cse_reciprocals): Use
6584 gimple_replace_ssa_lhs.
6585 * asan.c: Include gimple-builder.h.
6586 * Makefile.in: Add gimple-builder.o.
6588 2013-09-26 Richard Biener <rguenther@suse.de>
6590 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
6591 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
6592 (loe_visit_block): Use gcc_checking_assert.
6593 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
6594 gimple_assign_ssa_name_copy_p.
6595 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
6598 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
6600 * config/rs6000/t-rs6000 (rs6000.o): Remove.
6601 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
6603 2013-09-26 Richard Biener <rguenther@suse.de>
6605 PR tree-optimization/58539
6606 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
6607 the fact that debug statements are not taking part in loop-closed
6610 2013-09-26 Nick Clifton <nickc@redhat.com>
6612 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
6613 time warning message.
6614 (msp430_print_operand_raw): Delete unused letter parameter.
6615 (TARGET_PRINT_OPERAND_ADDRESS): Define.
6616 (msp430_print_operand_address): New function.
6617 (msp430_print_operand): Move address printing code from here to
6619 * config/msp430/msp430.md (movsipsi2): Add comment in generated
6621 (zero_extendpsisi2): Likewise.
6622 (extendpsisi2): New pattern.
6623 (andneghi3): New pattern.
6625 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
6627 * config/aarch64/aarch64.opt (mlra): New option.
6628 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
6629 (TARGET_LRA_P): Define.
6631 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
6633 * expr.c (expand_assignment): Remove obsolete comment.
6635 2013-09-25 Jeff Law <law@redhat.com>
6637 * tree-flow.h (thread_through_all_blocks): Prototype moved into
6638 tree-ssa-threadupdate.h.
6639 (register_jump_thread): Similarly.
6640 * tree-ssa-threadupdate.h: New header file.
6641 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
6642 * tree-vrp.c: Likewise.
6643 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
6644 (thread_around_empty_blocks): Change type of path vector argument to
6645 an edge,type pair from just an edge. Initialize both elements when
6646 appending to a jump threading path. Tweak references to elements
6648 (thread_across_edge): Similarly. Release memory for the elements
6650 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
6651 (dump_jump_thread_path): New function broken out from
6652 register_jump_thread.
6653 (register_jump_thread): Use dump_jump_thread_path. Change type of
6654 path vector entries. Search the path for NULL edges and dump
6655 the path if one is found. Tweak the conversion of path to 3-edge
6656 form to use the block copy type information embedded in the path.
6658 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6660 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
6662 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6663 Vladimir Makarov <vmakarov@redhat.com>
6665 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
6666 from the least significant bit.
6667 (strip_address_mutations): Add bitfield operations handling.
6668 (must_be_index_p): Add shifting and rotate operations handling.
6669 (set_address_base): Use must_be_base_p predicate.
6670 (set_address_index): Use must_be_index_p predicate.
6672 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
6673 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6674 Sergey Lega <sergey.s.lega@intel.com>
6675 Anna Tikhonova <anna.tikhonova@intel.com>
6676 Ilya Tocar <ilya.tocar@intel.com>
6677 Andrey Turetskiy <andrey.turetskiy@intel.com>
6678 Ilya Verbin <ilya.verbin@intel.com>
6679 Kirill Yukhin <kirill.yukhin@intel.com>
6680 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6682 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
6684 (ix86_expand_vector_move_misalign): Support new unaligned load and
6685 stores and use new names.
6686 (CODE_FOR_sse2_storedqu): Rename to ...
6687 (CODE_FOR_sse2_storedquv16qi): ... this.
6688 (CODE_FOR_sse2_loaddqu): Rename to ...
6689 (CODE_FOR_sse2_loaddquv16qi): ... this.
6690 (CODE_FOR_avx_loaddqu256): Rename to ...
6691 (CODE_FOR_avx_loaddquv32qi): ... this.
6692 (CODE_FOR_avx_storedqu256): Rename to ...
6693 (CODE_FOR_avx_storedquv32qi): ... this.
6694 * config/i386/i386.md (fpint_logic): New.
6695 * config/i386/sse.md (VMOVE): Extend for AVX512.
6699 (VI_UNALIGNED_LOADSTORE): Ditto.
6700 (sse2_avx_avx512f): Ditto.
6701 (sse2_avx2): Extend for AVX512.
6702 (sse4_1_avx2): Ditto.
6703 (avx2_avx512f): New.
6704 (sse): Extend for AVX512.
6707 (avxsizesuffix): Ditto.
6708 (sseintvecmode): Ditto.
6710 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
6711 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
6712 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6713 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
6714 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6715 (<sse2_avx_avx512f>_storedqu<mode): ... this.
6716 (<sse>_movnt<mode>): Replace constraint "x" with "v".
6717 (STORENT_MODE): Extend for AVX512.
6718 (*absneg<mode>2): Replace constraint "x" with "v".
6719 (*mul<mode>3): Ditto.
6720 (*ieee_smin<mode>3): Ditto.
6721 (*ieee_smax<mode>3): Ditto.
6722 (avx_cmp<mode>3): Replace VF with VF_128_256.
6723 (*<sse>_maskcmp<mode>3_comm): Ditto.
6724 (<sse>_maskcmp<mode>3): Ditto.
6725 (<sse>_andnot<mode>3): Extend for AVX512.
6726 (<code><mode>3, anylogic): Replace VF with VF_128_256.
6727 (<code><mode>3, fpint_logic): New.
6728 (*<code><mode>3): Extend for AVX512.
6729 (avx512flogicsuff): New.
6730 (avx512f_<logic><mode>): Ditto.
6731 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
6733 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
6734 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
6735 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
6736 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
6737 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
6738 (xop_vpermil2<mode>3): Ditto.
6739 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
6740 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
6741 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
6742 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
6744 2013-09-25 Tom Tromey <tromey@redhat.com>
6746 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
6747 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
6748 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
6749 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
6750 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
6751 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
6752 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
6753 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
6754 (GRAPHITE_HTAB_H): Remove.
6756 2013-09-25 Tom Tromey <tromey@redhat.com>
6758 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
6760 2013-09-25 Tom Tromey <tromey@redhat.com>
6762 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
6764 2013-09-25 Tom Tromey <tromey@redhat.com>
6766 * config/i386/t-i386 (i386.o): Remove.
6767 (i386-c.o): Use COMPILE and POSTCOMPILE.
6769 2013-09-25 Tom Tromey <tromey@redhat.com>
6771 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
6773 2013-09-25 Tom Tromey <tromey@redhat.com>
6775 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
6776 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
6777 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
6778 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
6779 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
6780 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
6781 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
6782 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
6783 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
6784 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
6785 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
6786 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
6787 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
6788 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
6789 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
6790 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
6791 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
6792 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
6793 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
6794 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
6795 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
6796 (tree-ssa-pre.o, tree-ssa-sccvn.o)
6797 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
6798 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
6799 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
6800 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
6801 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
6802 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
6803 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
6804 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
6805 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
6806 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
6807 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
6808 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
6809 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
6810 (graphite-clast-to-gimple.o, graphite-dependences.o)
6811 (graphite-interchange.o, graphite-poly.o)
6812 (graphite-scop-detection.o, graphite-sese-to-poly.o)
6813 (graphite-optimize-isl.o, tree-vect-loop.o)
6814 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
6815 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
6816 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
6817 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
6818 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
6819 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
6820 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
6821 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
6822 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
6823 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
6824 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
6825 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
6826 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
6827 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
6828 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
6829 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
6830 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
6831 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
6832 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
6833 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
6834 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
6835 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
6836 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
6837 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
6838 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
6839 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
6840 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
6841 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
6842 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
6843 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
6844 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
6845 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
6846 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
6847 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
6848 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
6849 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
6850 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
6851 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
6852 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
6853 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
6854 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
6855 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
6856 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
6857 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
6858 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
6859 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
6860 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
6861 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
6862 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
6863 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
6864 (gcov-dump.o): Remove.
6865 (default-c.o): Use COMPILE and POSTCOMPILE.
6866 (CFLAGS-gcc.o): New variable.
6867 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
6869 2013-09-25 Tom Tromey <tromey@redhat.com>
6871 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
6872 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
6873 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
6874 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
6875 (c-family/c-pragma.o, c-family/c-pretty-print.o)
6876 (c-family/c-semantics.o, c-family/c-ada-spec.o)
6877 (c-family/array-notation-common.o, c-family/stub-objc.o)
6878 (c-family/c-ubsan.o): Remove.
6880 2013-09-25 Tom Tromey <tromey@redhat.com>
6882 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
6884 2013-09-25 Tom Tromey <tromey@redhat.com>
6886 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
6887 to add -DENABLE_SHARED_LIBGCC.
6888 (gcc.o): Don't use subshell.
6890 2013-09-25 Tom Tromey <tromey@redhat.com>
6892 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
6893 * configure.ac: Don't invoke AM_PROG_CC_C_O.
6894 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
6895 * configure, config.in: Rebuild.
6897 2013-09-25 Tom Tromey <tromey@redhat.com>
6899 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
6900 (COMPILE, POSTCOMPILE): New variables.
6901 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
6902 (DEPFILES): New variable.
6903 Include ".Po" files.
6904 * configure.ac: Add checks for dependency checking.
6905 * configure, aclocal.m4: Regenerate.
6907 2013-09-25 Tom Tromey <tromey@redhat.com>
6909 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
6910 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
6912 2013-09-25 Tom Tromey <tromey@redhat.com>
6914 * Makefile.in (generated_files): Add options.h,
6915 target-hooks-def.h, insn-opinit.h,
6916 common/common-target-hooks-def.h, pass-instances.def,
6917 c-family/c-target-hooks-def.h.
6919 2013-09-25 Jeff Law <law@redhat.com>
6921 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
6922 than foo[foo.length () - 1] to access last member in a vec.
6923 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
6925 2013-09-25 Richard Biener <rguenther@suse.de>
6928 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
6930 2013-09-25 Jan Hubicka <jh@suse.cz>
6932 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
6935 2013-09-25 Marek Polacek <polacek@redhat.com>
6938 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
6939 when determining the type name.
6941 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
6943 * config/sh/sh.md: Fix formatting.
6945 2013-09-24 Xinliang David Li <davidxl@google.com>
6947 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
6948 max peel iterations parameter.
6949 * param.def: New parameter.
6950 * doc/invoke.texi: Document New parameter.
6952 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
6954 * gimple-pretty-print.c: Various whitespace tweaks.
6955 * tree-core.h: Likewise.
6956 * tree-pretty-print.c: Likewise.
6957 * tree-ssa-alias.c: Likewise.
6958 * tree-ssa-copy.c: Likewise.
6959 * tree-ssanames.c: Likewise.
6960 * tree-ssanames.h: Likewise.
6961 * tree-vrp.c: Likewise.
6963 2013-09-24 Alan Modra <amodra@gmail.com>
6967 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
6968 for output operands that disallow regs. Don't use EXPAND_WRITE on
6971 2013-09-24 Richard Biener <rguenther@suse.de>
6974 * tree.c (reference_alias_ptr_type): Move ...
6975 * alias.c (reference_alias_ptr_type): ... here and implement
6976 in terms of the new reference_alias_ptr_type_1.
6977 (ref_all_alias_ptr_type_p): New helper.
6978 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
6979 use ref_all_alias_ptr_type_p.
6980 (get_deref_alias_set): Add flag_strict_aliasing check here.
6981 (reference_alias_ptr_type_1): New function, split out from ...
6982 (get_alias_set): ... here.
6983 (alias_ptr_types_compatible_p): New function.
6984 * alias.h (reference_alias_ptr_type): Declare.
6985 (alias_ptr_types_compatible_p): Likewise.
6986 * tree.h (reference_alias_ptr_type): Remove.
6987 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
6988 to compare MEM_REF alias types.
6990 2013-09-24 Richard Biener <rguenther@suse.de>
6992 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
6994 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6996 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
6997 reload helper function arrays into a single array reg_addr.
6998 (reload_fpr_gpr): Likewise.
6999 (reload_gpr_vsx): Likewise.
7000 (reload_vsx_gpr): Likewise.
7001 (struct rs6000_reg_addr): Likewise.
7002 (reg_addr): Likewise.
7003 (rs6000_debug_reg_global): Change rs6000_vector_reload,
7004 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
7005 (rs6000_init_hard_regno_mode_ok): Likewise.
7006 (rs6000_secondary_reload_direct_move): Likewise.
7007 (rs6000_secondary_reload): Likewise.
7009 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
7010 constraints: wu, ww, and wy. Repurpose wv constraint added during
7011 power8 changes. Put wg constraint in alphabetical order.
7013 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
7014 for future work to add ISA 2.07 VSX single precision support.
7015 (-mvsx-scalar-double): Change default from -1 to 1, update
7016 documentation comment.
7017 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
7018 (-mupper-regs-df): New debug switch to control whether DF values
7019 can go in the traditional Altivec registers.
7020 (-mupper-regs-sf): New debug switch to control whether SF values
7021 can go in the traditional Altivec registers.
7023 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
7025 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
7026 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
7027 Add new constraints, wu/ww/wy. Repurpose wv constraint.
7028 (rs6000_debug_legitimate_address_p): Print if we are running
7029 before, during, or after reload.
7030 (rs6000_secondary_reload): Add a comment.
7031 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
7033 * config/rs6000/constraints.md (wa constraint): Sort w<x>
7034 constraints. Update documentation string.
7035 (wd constraint): Likewise.
7036 (wf constraint): Likewise.
7037 (wg constraint): Likewise.
7038 (wn constraint): Likewise.
7039 (ws constraint): Likewise.
7040 (wt constraint): Likewise.
7041 (wx constraint): Likewise.
7042 (wz constraint): Likewise.
7043 (wu constraint): New constraint for ISA 2.07 SFmode scalar
7045 (ww constraint): Likewise.
7046 (wy constraint): Likewise.
7047 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
7048 in the previous submissions.
7049 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
7051 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
7053 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
7054 (REG_BR_PROB): Say that the probability is stored in an int_list.
7055 * reg-notes.def: Update commentary to mention INT_LIST.
7056 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
7057 (INT_LIST): New rtx.
7058 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
7059 * rtlanal.c (int_reg_note_p): New function.
7060 (alloc_reg_note): Assert that the note does not have an int argument.
7061 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
7062 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
7063 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
7064 rather than an INSN_LIST. Handle INT_LIST.
7065 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
7066 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
7067 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
7068 Manipulate them as ints rather than rtxes.
7069 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
7070 * regmove.c (copy_src_to_dest): Likewise.
7071 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
7073 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
7074 into the cases that need it.
7075 * config/arm/arm.c (arm_unwind_emit): Likewise.
7077 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
7078 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
7079 * loop-doloop.c (add_test, doloop_modify): Likewise.
7080 * loop-unswitch.c (compare_and_jump_seq): Likewise.
7081 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
7082 * predict.c (combine_predictions_for_insn): Likewise.
7083 * print-rtl.c (print_rtx): Handle INT_LIST.
7084 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
7085 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
7086 * config/arm/arm.c (emit_unlikely_jump): Likewise.
7087 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
7088 (ix86_split_fp_branch, predict_jump): Likewise.
7089 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
7090 * config/sh/sh.c (expand_cbranchsi4): Likewise.
7091 * config/spu/spu.c (ea_load_store_inline): Likewise.
7093 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
7094 value of a REG_BR_PROB note.
7095 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
7096 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
7097 * emit-rtl.c (try_split): Likewise.
7098 * predict.c (br_prob_note_reliable_p): Likewise.
7099 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
7100 * reorg.c (mostly_true_jump): Likewise.
7101 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
7102 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
7103 * config/i386/i386.c (ix86_print_operand): Likewise.
7104 * config/ia64/ia64.c (ia64_print_operand): Likewise.
7105 * config/mmix/mmix.c (mmix_print_operand): Likewise.
7106 * config/rs6000/rs6000.c (output_cbranch): Likewise.
7107 * config/s390/s390.c (s390_expand_tbegin): Likewise.
7108 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
7109 * config/sparc/sparc.c (output_cbranch): Likewise.
7110 * config/spu/spu.c (get_branch_target): Likewise.
7111 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
7112 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
7114 2013-09-23 Jan Hubicka <jh@suse.cz>
7116 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
7118 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
7119 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
7121 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
7122 (gimple_fold_call): Dump inconsistent devirtualizations; add
7123 sanity check for type based devirtualizations.
7124 * ipa-prop.c: Include ipa-utils.h
7125 (ipa_intraprocedural_devirtualization): Add sanity check.
7126 (try_make_edge_direct_virtual_call): Likewise.
7128 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
7130 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
7131 assignment statements.
7133 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
7135 * gimple-pretty-print.c (dump_ssaname_info): New function.
7136 (dump_gimple_phi): Call it.
7137 (pp_gimple_stmt_1): Likewise.
7138 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
7139 (range_info_def): Declare.
7140 * tree-pretty-print.c (pp_double_int): New function.
7141 (dump_generic_node): Call it.
7142 * tree-pretty-print.h (pp_double_int): Declare.
7143 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
7144 * tree-ssanames.h (range_info_def): New structure.
7145 (value_range_type): Move definition here.
7146 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
7148 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
7150 (set_range_info): New function.
7151 (get_range_info): Likewise.
7152 (duplicate_ssa_name_range_info): Likewise.
7153 (duplicate_ssa_name_fn): Check pointer type and call
7154 duplicate_ssa_name_range_info.
7155 * tree-ssa-copy.c (fini_copy_prop): Likewise.
7156 * tree-vrp.c (value_range_type): Remove definition, now in
7158 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
7159 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
7160 (SSA_NAME_RANGE_INFO): New macro.
7162 2013-09-23 Richard Biener <rguenther@suse.de>
7164 PR tree-optimization/58464
7165 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7166 (phi_trans_add): Change to add conditionally on being not present.
7167 (phi_translate_1): Remove recursion detection here.
7168 (phi_translate): Pre-seed the cache with NULL to catch
7169 recursion here in a more generic way.
7170 (bitmap_find_leader): Adjust comment.
7171 (get_representative_for): Dump value-numbers.
7172 (create_expression_by_pieces): Likewise.
7173 (insert_into_preds_of_block): Likewise.
7175 2013-09-23 Christian Bruel <christian.bruel@st.com>
7178 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
7179 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
7181 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
7184 2013-09-20 Renlin Li <renlin.li@arm.com>
7186 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
7188 (aarch64_expand_epilogue): Likewise.
7189 (aarch64_legitimize_reload_address): Likewise.
7191 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
7193 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
7194 NULL_TREE before pushing them onto the vector. Likewise for labels.
7196 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
7198 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
7199 comparison operators when -fno-trapping-math is in effect.
7200 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
7201 comparison operators in TFmode and assert that unsupported operators
7203 (ia64_print_operand): Likewise.
7205 2013-09-21 Jan Hubicka <jh@suse.cz>
7207 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
7208 (sse_typeless_stores): Enable for core
7209 (sse_load0_by_pxor): Likewise.
7210 (four_jump_limit): Disable for core.
7211 (pad_returns): Likewise.
7212 (avoid_vector_decode): Likewise.
7213 (fuse_cmp_and_branch): Enable for cores.
7214 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
7216 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7219 * config/pa/pa.c (pa_option_override): Disable auto increment and
7220 decrement instructions until reload is completed.
7222 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
7223 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
7224 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
7226 2013-09-20 DJ Delorie <dj@redhat.com>
7227 Nick Clifton <nickc@redhat.com>
7229 * config/rl78/rl78.c: Various whitespace and comment tweaks.
7230 (need_to_save): Save bank 0 on interrupts.
7231 (characterize_address): Strip far address wrappers.
7232 (rl78_as_legitimate_address): Likewise.
7233 (transcode_memory_rtx): Likewise.
7234 (rl78_peep_movhi_p): Disable this peephole after devirt.
7235 (rl78_propogate_register_origins): Forget all origins when a
7237 * config/rl78/rl78-virt.md: Various whitespace tweaks.
7238 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
7240 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
7241 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
7242 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
7244 * config/rl78/constraints.md: Various whitespace and paren tweaks.
7246 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7248 * config/pa/pa.md: In "scc" insn patterns, change output template to
7249 handle const0_rtx in reg_or_0_operand operands.
7251 2013-09-20 Martin Husemann <martin@NetBSD.org>
7254 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
7255 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
7257 2013-09-20 Richard Biener <rguenther@suse.de>
7260 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
7261 remembering SSA name version and block index.
7262 (new_scev_info_str): Adjust.
7263 (hash_scev_info): Likewise. Also hash the block index.
7264 (eq_scev_info): Adjust.
7265 (find_var_scev_info): Likewise.
7266 (struct instantiate_cache_entry): Remove.
7267 (struct instantiate_cache_type): Use a htab to map name, block
7269 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7270 (get_instantiated_value_entry): Likewise.
7271 (hash_idx_scev_info, eq_idx_scev_info): New functions.
7272 (instantiate_scev_name): Adjust.
7274 2013-09-20 Jeff Law <law@redhat.com>
7276 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
7278 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
7280 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
7281 Call aarch64_simd_expand_args to update op[argc].
7283 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7285 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
7288 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7290 * gengtype.c (file_rules): Added rule for *.cc files.
7291 (get_output_file_with_visibility): Give fatal message when no
7294 2013-09-20 Renlin Li <renlin.li@arm.com>
7296 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
7297 (aarch64_expand_epilogue): Likewise.
7298 (aarch64_legitimize_reload_address): Likewise.
7300 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
7303 * expr.c (expand_assignment): Remove misalignp code path.
7305 2013-09-20 Marek Polacek <polacek@redhat.com>
7308 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
7309 TYPE_PRECISION. Add asserts.
7311 2013-09-20 Richard Biener <rguenther@suse.de>
7313 PR tree-optimization/58453
7314 * tree-loop-distribution.c (distribute_loop): Apply the cost
7315 model for -ftree-loop-distribute-patterns, too.
7317 2013-09-20 Richard Biener <rguenther@suse.de>
7320 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
7321 make type comparison less strict.
7323 2013-09-20 Alan Modra <amodra@gmail.com>
7325 * configure: Regenerate.
7326 * aclocal.m4: Regenerate.
7328 2013-09-20 Marek Polacek <polacek@redhat.com>
7331 * doc/extend.texi: Document that attribute used is meant to be used
7332 on variables with static storage duration.
7334 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7336 PR tree-optimization/58472
7337 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
7338 simd_lane_access set inv_p = false.
7339 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
7340 the simduid magic VAR_DECL.
7342 2013-09-19 Jan Hubicka <jh@suse.cz>
7344 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
7346 2013-09-17 Jeff Law <law@redhat.com>
7348 * tree-ssa-dom.c (record_temporary_equivalences): New function
7349 split out of dom_opt_dom_walker::after_dom_children.
7350 (dom_opt_dom_walker::thread_across_edge): Move common code
7351 in here from dom_opt_dom_walker::after_dom_children.
7352 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
7354 2013-09-19 Jan Hubicka <jh@suse.cz>
7356 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
7357 (TARGET_GENERIC): Use PROCESOR_GENERIC
7358 (enum processor_type): Unify generic32 and 64.
7359 * i386.md (cpu): Likewise.
7360 * x86-tune.def (use_leave): Enable for generic32.
7361 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
7362 * athlon.md: Change generic64 to generic in all occurences.
7363 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
7364 (ix86_target_macros_internal): Likewise.
7365 * driver-i386.c (host_detect_local_cpu): Likewise.
7366 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
7368 (generic_memcpy, generic_memset, generic_cost): This one.
7369 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
7370 (m_GENERIC32, m_GENERIC64): Remove.
7371 (m_GENERIC): Turn into one flag.
7372 (processor_target): Unify generic tunnings.
7373 (ix86_option_override_internal): Replace generic32/64 by generic.
7374 (ix86_issue_rate): Likewise.
7375 (ix86_adjust_cost): Likewise.
7377 2013-09-19 Jan Hubicka <jh@suse.cz>
7379 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
7380 of speculative flag.
7382 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7384 * omp-low.c (expand_omp_sections): Always pass len - 1 to
7385 GOMP_sections_start, even if !exit_reachable.
7387 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7389 * lra-constraints.c (need_for_all_save_p): Use macro
7390 HARD_REGNO_CALL_PART_CLOBBERED.
7391 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
7392 set up pseudo conflict hard regs.
7394 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7397 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
7400 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7402 PR rtl-optimization/58438
7403 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
7404 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
7405 from previous subpasses.
7407 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
7409 * arm.c (arm_get_frame_offsets): Validate architecture supports
7410 LDRD/STRD before accepting the tuning preference.
7411 (arm_expand_prologue): Likewise.
7412 (arm_expand_epilogue): Likewise.
7414 2013-09-18 Richard Biener <rguenther@suse.de>
7416 PR tree-optimization/58417
7417 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
7418 have chrecs with symbols defined in the loop as operands.
7419 (chrec_fold_multiply): Likewise.
7420 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
7421 parameters before folding binary operations.
7422 (struct instantiate_cache_entry_hasher): Remove.
7423 (struct instantiate_cache_type): Use a pointer-map.
7424 (instantiate_cache_type::instantiate_cache_type): New function.
7425 (instantiate_cache_type::get): Likewise.
7426 (instantiate_cache_type::set): Likewise.
7427 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7428 (get_instantiated_value_entry): Likewise.
7429 (global_cache): New global.
7430 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
7431 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
7432 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
7433 (instantiate_scev_name): Adjust.
7434 (instantiate_scev): Construct global instead of local cache.
7435 (resolve_mixers): Likewise.
7437 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
7438 Paolo Carlini <paolo.carlini@oracle.com>
7441 * doc/implement-cxx.texi: Fix references to the C++ standards.
7443 2013-09-18 Jakub Jelinek <jakub@redhat.com>
7445 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
7446 * tree-vect-data-refs.c (vect_analyze_data_refs): For
7447 simd_lane_access drs, update also DR_ALIGNED_TO.
7449 2013-09-18 Marek Polacek <polacek@redhat.com>
7452 * doc/extend.texi: Document no_sanitize_undefined attribute.
7453 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
7454 no_sanitize_undefined attribute.
7456 2013-09-18 Nick Clifton <nickc@redhat.com>
7458 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
7459 (ASM_DECLARE_FUNCTION_NAME): Define.
7461 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
7463 * compare-elim.c (find_comparison_dom_walker): New class
7464 (find_comparisons_in_bb): Rename to
7465 find_comparison_dom_walker::before_dom_children
7466 (find_comparisons): Adjust
7467 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
7469 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
7470 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
7471 (init_walk_dominator_tree): Remove declaration.
7472 (fini_walk_dominator_tree): Remove declaration.
7473 * fwprop.c (single_def_use_dom_walker): New class
7474 (single_def_use_enter_block): Convert to
7475 single_def_use_dom_walker::before_dom_children.
7476 (single_def_use_leave_block): Convert to
7477 single_def_use_dom_walker::after_dom_children.
7478 (build_single_def_use_links): Adjust.
7479 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
7481 (find_candidates_in_block): Convert to
7482 find_candidates_dom_walker::before_dom_children.
7483 (execute_strength_reduction): Adjust.
7484 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
7485 (sese_dom_walker): New class.
7486 (sese_dom_walker::sese_dom_walker): New constructor.
7487 (sese_dom_walker::~sese_dom_walker): New destructor.
7488 (build_sese_conditions_before): Convert to
7489 sese_dom_walker::before_dom_children.
7490 (build_sese_conditions_after): Convert to
7491 sese_dom_walker::after_dom_children.
7492 (build_poly_scop): Adjust
7493 * tree-into-ssa.c (rewrite_dom_walker): New class
7494 (rewrite_enter_block): Convert to
7495 rewrite_dom_walker::before_dom_children.
7496 (rewrite_leave_block): Convert to
7497 rewrite_dom_walker::after_dom_children.
7498 (rewrite_update_dom_walker): New class.
7499 (rewrite_update_enter_block): Convert to
7500 rewrite_update_dom_walker::before_dom_children.
7501 (rewrite_update_leave_block): Convert to
7502 rewrite_update_dom_walker::after_dom_children.
7503 (rewrite_blocks, rewrite_into_ssa): Adjust.
7504 (mark_def_dom_walker): New class.
7505 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
7506 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
7507 (mark_def_sites_blocks): Convert to
7508 mark_def_dom_walker::before_dom_children.
7509 (mark_def_site_blocks): Remove.
7510 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
7511 (tree_ssa_dominator_optimize): Adjust.
7512 (dom_thread_across_edge): Convert to method
7513 dom_opt_dom_walker::thread_across_edge.
7514 (dom_opt_enter_block): Convert to member function
7515 dom_opt_dom_walker::before_dom_children.
7516 (dom_opt_leave_block): Convert to member function
7517 dom_opt_dom_walker::after_dom_children.
7518 * tree-ssa-dse.c (dse_dom_walker): New class.
7519 (dse_enter_block): Convert to member function
7520 dse_dom_walker::before_dom_children.
7521 (tree_ssa_dse): Adjust.
7522 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
7523 (determine_invariantness_stmt): Convert to method
7524 invariantness_dom_walker::before_dom_children.
7525 (determine_invariantness): Remove
7526 (move_computations_dom_walker): New class.
7527 (move_computations_stmt): Convert to method
7528 move_computations_dom_walker::before_dom_children.
7529 (move_computations, tree_ssa_lim): Adjust.
7530 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
7531 (nt_init_block): Convert to method
7532 notrappping_dom_walker::before_dom_children.
7533 (nt_fini_block): Convert to method
7534 method nontrapping_dom_walker::after_dom_children.
7535 (get_non_trapping): Adjust.
7536 * tree-ssa-pre.c (eliminate_dom_walker): New class.
7537 (eliminate_bb): Convert to method
7538 eliminate_dom_walker::before_dom_children.
7539 (eliminate_leave_block): Convert to method
7540 eliminate_dom_walker::after_dom_children.
7541 (eliminate): Adjust.
7542 * tree-ssa-strlen.c (strlen_dom_walker): New class.
7543 (strlen_enter_block): Convert to method
7544 strlen_dom_walker::before_dom_children.
7545 (strlen_leave_block): Convert to method
7546 method strlen_dom_walker::after_dom_children.
7547 (tree_ssa_strlen): Adjust.
7548 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
7549 (tree_ssa_uncprop): Adjust.
7550 (uncprop_leave_block): Convert to method
7551 uncprop_dom_walker::after_dom_children.
7552 (uncprop_leave_block): Convert to method
7553 uncprop_dom_walker::before_dom_children.
7555 2013-09-18 Bin Cheng <bin.cheng@arm.com>
7557 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
7558 branch in same basic block. Check both src and dest of the move insn.
7560 2013-09-17 Nick Clifton <nickc@redhat.com>
7562 * config/rl78/rl78-real.md (bf): New pattern.
7564 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
7565 (rl78_print_operand): Do not put a # before a %B.
7566 * config/rl78/rl78.opt: Tweak doc strings.
7568 2013-09-17 DJ Delorie <dj@redhat.com>
7570 * config/rl78/constraints.md (Wcv): Allow up to $r31.
7571 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
7572 (rl78_option_override): Likewise, if -mallregs.
7573 (is_virtual_register): Likewise.
7574 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
7575 (REGNO_OK_FOR_BASE_P): Likewise.
7576 * config/rl78/rl78.opt (-mallregs): New.
7578 2013-09-17 Nick Clifton <nickc@redhat.com>
7580 * config/rl78/rl78.c (need_to_save): Change return type to bool.
7581 For interrupt functions: save all call clobbered registers if the
7582 interrupt handler is not a leaf function.
7583 (rl78_expand_prologue): Always recompute the frame information.
7584 For interrupt functions: only select bank 0 if one of the bank 0
7585 registers is going to be psuhed.
7587 2013-09-17 DJ Delorie <dj@redhat.com>
7589 * config/rl78/constraints.md: For each W* constraint, rename to C*
7590 and create a W* constraint that checks for an optional ES: prefix
7592 * config/rl78/rl78.md (UNS_ES_ADDR): New.
7593 (es_addr): New. Used to wrap far addresses.
7594 * config/rl78/rl78-protos.h (rl78_es_addr): New.
7595 (rl78_es_base): New.
7596 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
7597 wrapped far addresses.
7598 (rl78_print_operand_1): Unwrap far addresses before processing.
7599 (rl78_lo16): Wrap far addresses in unspecs.
7600 (rl78_es_addr): New.
7601 (rl78_es_base): New.
7602 (insn_ok_now): Check for not-yet-wrapped far addresses.
7603 (transcode_memory_rtx): Properly re-wrap far addresses.
7605 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
7607 * config/sparc/t-rtems: Add leon3 multilibs.
7609 2013-09-17 Cong Hou <congh@google.com>
7611 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
7612 when checking the dot production pattern. The type of rhs operand
7613 of multiply is now checked correctly.
7615 2013-09-17 Jeff Law <law@redhat.com>
7617 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
7618 edge implied equivalences into successor phis.
7619 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
7620 here from tree-ssa-threadedge.c.
7621 (mark_threaded_blocks): When threading through a joiner, if both
7622 successors of the joiner's clone reach the same block, verify the
7623 PHI arguments are equal. If not, cancel the jump threading request.
7624 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
7625 tree-ssa-threadupdate.c
7626 (thread_across_edge): Don't check PHI argument equality when
7627 threading through joiner block here.
7629 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
7631 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
7632 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
7633 (warn_uninit, warn_uninitialized_vars,
7634 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7635 Move to tree-ssa-uninit.c.
7636 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
7637 (has_undefined_value_p): New. Pass dependant parts of
7638 ssa_undefined_value_p.
7639 (uninit_undefined_value_p): Use has_undefined_value_p.
7640 (warn_uninit, warn_uninitialized_vars,
7641 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7642 Move from tree-ssa.c.
7643 * tree-ssa.h: Adjust prototypes.
7645 2013-09-17 Jan Hubicka <jh@suse.cz>
7648 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
7649 * ipa-inline.c (can_inline_edge_p): Do not downgrade
7650 FUNCTION_NOT_OPTIMIZED.
7651 * ipa-inline-analysis.c (compute_inline_parameters): Function
7652 not optimized is not inlinable unless it is alwaysinline.
7653 (inline_analyze_function): Force calls in not optimized
7654 function not inlinable.
7656 2013-09-17 Jan Hubicka <jh@suse.cz>
7659 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
7661 * ipa.c (function_and_variable_visibility): Likewise.
7662 * ipa-profile.c (ipa_profile): Likewise.
7664 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
7667 * cgraph.c (cgraph_function_body_availability): Check for ifunc
7668 attribute, and don't inline the resolver in this case.
7670 2013-09-17 Teresa Johnson <tejohnson@google.com>
7672 * coverage.c (get_coverage_counts): Add missing newline.
7674 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7676 PR tree-optimization/58088
7677 * fold-const.c (mask_with_trailing_zeros): New function.
7678 (fold_binary_loc): Make sure we don't recurse infinitely
7679 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
7680 Use mask_with_trailing_zeros where appropriate.
7682 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
7684 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
7685 of 'prev' var to get better distance estimation.
7687 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
7689 * tree-inline.h (struct copy_body_data): Add transform_parameter.
7690 * tree-inline.c (is_parameter_of): New predicate.
7691 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
7692 a parameter has been remapped.
7693 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
7694 (optimize_inline_calls): Initialize transform_parameter.
7695 (copy_gimple_seq_and_replace_locals): Likewise.
7696 (tree_function_versioning): Likewise.
7697 (maybe_inline_call_in_expr): Likewise.
7699 2013-09-17 Nick Clifton <nickc@redhat.com>
7701 * config/msp430/msp430-protos.h: Add prototypes for new functions.
7702 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
7704 (is_attr_func): New function.
7705 (msp430_is_interrupt_func): New function.
7706 (is_naked_func): New function.
7707 (is_reentrant_func): New function.
7708 (is_critical_func): New function.
7709 (msp430_start_function): Add annotations for function attributes.
7710 (msp430_attr): New function.
7711 (msp430_attribute_table): New.
7712 (msp430_function_section): New function.
7713 (TARGET_ASM_FUNCTION_SECTION): Define.
7714 (msp430_builtin): New enum.
7715 (msp430_init_builtins): New function.
7716 (msp430_builtin_devl): New function.
7717 (msp430_expand_builtin): New function.
7718 (TARGET_INIT_BUILTINS): Define.
7719 (TARGET_EXPAND_BUILTINS): Define.
7720 (TARGET_BUILTIN_DECL): Define.
7721 (msp430_expand_prologue): Add support for naked, interrupt,
7722 critical and reentrant functions.
7723 (msp430_expand_epilogue): Likewise.
7724 (msp430_print_operand): Handle 'O' character.
7725 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
7727 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
7728 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
7729 (pushm): Use a 'n' rather than an 'i' constraint.
7730 (msp_return): Add generation of the interrupt return instruction.
7731 (disable_interrupts): New pattern.
7732 (enable_interrupts): New pattern.
7733 (push_intr_state): New pattern.
7734 (pop_intr_state): New pattern.
7735 (bic_SR): New pattern.
7736 (bis_SR): New pattern.
7737 * doc/extend.texi: Document MSP430 function attributes and builtin
7740 2013-09-17 Richard Biener <rguenther@suse.de>
7742 PR tree-optimization/58432
7743 * tree-loop-distribution.c (tree_loop_distribution): Also
7744 scan PHIs for outside loop uses and seed a partition from them.
7746 2013-09-17 Bin Cheng <bin.cheng@arm.com>
7748 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
7749 (restructure_reference): Call backtrace_base_for_ref.
7751 2013-09-17 Alan Modra <amodra@gmail.com>
7754 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
7757 2013-09-16 DJ Delorie <dj@redhat.com>
7759 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
7760 vregs location for RL78/G10.
7761 (rl78_expand_prologue): Avoid SEL on G10.
7762 (rl78_expand_epilogue): Likewise.
7763 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
7764 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
7765 __RL78_G10__ when appropriate.
7766 (ASM_SPEC): Pass -mg10 along to the assembler.
7767 * config/rl78/rl78.md (sel_rb): Disable for G10.
7768 * config/rl78/rl78.opt: Add -mg10 option.
7769 * config/rl78/t-rl78: Add -mg10 multilib.
7771 2013-09-16 Xinliang David Li <davidxl@google.com>
7773 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
7774 * omp-low.c (omp_max_vf): Ditto.
7775 (expand_omp_simd): Ditto.
7776 * tree-vectorizer.c (vectorize_loops): Ditto.
7777 (gate_vect_slp): Ditto.
7778 (gate_increase_alignment): Ditto.
7779 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
7780 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
7781 (gate_tree_vectorize): Name change.
7782 (tree_vectorize): Ditto.
7783 (pass_vectorize::gate): Call new function.
7784 (pass_vectorize::execute): Ditto.
7785 * opts.c: O3 default setting change.
7786 (finish_options): Check new flag.
7787 * doc/invoke.texi: Document new flags.
7788 * common.opt: New flags.
7790 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
7792 * doc/tm.texi.in (Cond Exec Macros): Remove node.
7793 (Condition Code): Don't reference it.
7794 * doc/tm.texi: Regenerate.
7796 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
7799 * lra-constraints.c (undo_optional_reloads): Consider all optional
7800 reload even if it did not get a hard reg.
7802 2013-09-16 Teresa Johnson <tejohnson@google.com>
7804 * dumpfile.c (dump_loc): Remove newline emission.
7805 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
7806 emission to dump_printf_loc calls where missing.
7807 (vect_mark_for_runtime_alias_test): Ditto.
7808 (vect_analyze_data_ref_dependence): Ditto.
7809 (vect_analyze_data_ref_dependences): Ditto.
7810 (vect_slp_analyze_data_ref_dependence): Ditto.
7811 (vect_slp_analyze_data_ref_dependences): Ditto.
7812 (vect_compute_data_ref_alignment): Ditto.
7813 (vect_update_misalignment_for_peel): Ditto.
7814 (vect_verify_datarefs_alignment): Ditto.
7815 (vector_alignment_reachable_p): Ditto.
7816 (vect_get_data_access_cost): Ditto.
7817 (vect_enhance_data_refs_alignment): Ditto.
7818 (vect_find_same_alignment_drs): Ditto.
7819 (vect_analyze_data_refs_alignment): Ditto.
7820 (vect_analyze_group_access): Ditto.
7821 (vect_analyze_data_ref_access): Ditto.
7822 (vect_analyze_data_ref_accesses): Ditto.
7823 (vect_prune_runtime_alias_test_list): Ditto.
7824 (vect_analyze_data_refs): Ditto.
7825 (vect_create_addr_base_for_vector_ref): Ditto.
7826 (vect_create_data_ref_ptr): Ditto.
7827 (vect_grouped_store_supported): Ditto.
7828 (vect_grouped_load_supported): Ditto.
7829 * value-prof.c (check_counter): Ditto.
7830 (check_ic_target): Ditto.
7831 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
7832 (vect_recog_widen_mult_pattern): Ditto.
7833 (vect_recog_widen_sum_pattern): Ditto.
7834 (vect_recog_over_widening_pattern): Ditto.
7835 (vect_recog_widen_shift_pattern): Ditto.
7836 (vect_recog_rotate_pattern): Ditto.
7837 (vect_recog_vector_vector_shift_pattern): Ditto.
7838 (vect_recog_divmod_pattern): Ditto.
7839 (vect_recog_mixed_size_cond_pattern): Ditto.
7840 (vect_recog_bool_pattern): Ditto.
7841 (vect_pattern_recog_1): Ditto.
7842 (vect_pattern_recog): Ditto.
7843 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
7844 (vect_is_simple_iv_evolution): Ditto.
7845 (vect_analyze_scalar_cycles_1): Ditto.
7846 (vect_get_loop_niters): Ditto.
7847 (vect_analyze_loop_1): Ditto.
7848 (vect_analyze_loop_form): Ditto.
7849 (vect_analyze_loop_operations): Ditto.
7850 (vect_analyze_loop_2): Ditto.
7851 (vect_analyze_loop): Ditto.
7852 (report_vect_op): Ditto.
7853 (vect_is_slp_reduction): Ditto.
7854 (vect_is_simple_reduction_1): Ditto.
7855 (vect_get_known_peeling_cost): Ditto.
7856 (vect_estimate_min_profitable_iters): Ditto.
7857 (vect_model_reduction_cost): Ditto.
7858 (vect_model_induction_cost): Ditto.
7859 (get_initial_def_for_induction): Ditto.
7860 (vect_create_epilog_for_reduction): Ditto.
7861 (vectorizable_reduction): Ditto.
7862 (vectorizable_induction): Ditto.
7863 (vectorizable_live_operation): Ditto.
7864 (vect_loop_kill_debug_uses): Ditto.
7865 (vect_transform_loop): Ditto.
7866 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
7867 (vect_stmt_relevant_p): Ditto.
7868 (process_use): Ditto.
7869 (vect_mark_stmts_to_be_vectorized): Ditto.
7870 (vect_model_simple_cost): Ditto.
7871 (vect_model_promotion_demotion_cost): Ditto.
7872 (vect_model_store_cost): Ditto.
7873 (vect_get_store_cost): Ditto.
7874 (vect_model_load_cost): Ditto.
7875 (vect_get_load_cost): Ditto.
7876 (vect_init_vector_1): Ditto.
7877 (vect_get_vec_def_for_operand): Ditto.
7878 (vect_finish_stmt_generation): Ditto.
7879 (vectorizable_call): Ditto.
7880 (vectorizable_conversion): Ditto.
7881 (vectorizable_assignment): Ditto.
7882 (vectorizable_shift): Ditto.
7883 (vectorizable_operation): Ditto.
7884 (vectorizable_store): Ditto.
7885 (vectorizable_load): Ditto.
7886 (vectorizable_condition): Ditto.
7887 (vect_analyze_stmt): Ditto.
7888 (vect_transform_stmt): Ditto.
7889 (vect_is_simple_use): Ditto.
7890 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
7891 (vect_can_advance_ivs_p): Ditto.
7892 (vect_update_ivs_after_vectorizer): Ditto.
7893 (vect_do_peeling_for_loop_bound): Ditto.
7894 (vect_gen_niters_for_prolog_loop): Ditto.
7895 (vect_update_inits_of_drs): Ditto.
7896 (vect_create_cond_for_alias_checks): Ditto.
7897 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
7898 (vect_build_slp_tree_1): Ditto.
7899 (vect_supported_load_permutation_p): Ditto.
7900 (vect_analyze_slp_instance): Ditto.
7901 (vect_analyze_slp): Ditto.
7902 (vect_make_slp_decision): Ditto.
7903 (vect_detect_hybrid_slp): Ditto.
7904 (vect_bb_vectorization_profitable_p): Ditto.
7905 (vect_slp_analyze_bb_1): Ditto.
7906 (vect_update_slp_costs_according_to_vf): Ditto.
7907 (vect_get_mask_element): Ditto.
7908 (vect_transform_slp_perm_load): Ditto.
7909 (vect_schedule_slp_instance): Ditto.
7910 (vect_schedule_slp): Ditto.
7911 (vect_slp_transform_bb): Ditto.
7912 * profile.c (read_profile_edge_counts): Ditto.
7913 (compute_branch_probabilities): Ditto.
7914 * coverage.c (get_coverage_counts): Ditto.
7916 2013-09-16 Diego Novillo <dnovillo@google.com>
7918 * tree-core.h: Add missing comment lines from refactoring of tree.h.
7920 2013-09-16 Jan Hubicka <jh@suse.cz>
7922 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
7923 abstract functions; for static functions check the presence of body.
7925 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7927 * config/aarch64/aarch64-simd-builtins.def (fma): New.
7928 * config/aarch64/aarch64-simd.md
7929 (aarch64_mla_elt<mode>): New.
7930 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7931 (aarch64_mls_elt<mode>): Likewise.
7932 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7933 (aarch64_fma4_elt<mode>): Likewise.
7934 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7935 (aarch64_fma4_elt_to_128v2df): Likewise.
7936 (aarch64_fma4_elt_to_64df): Likewise.
7937 (fnma<mode>4): Likewise.
7938 (aarch64_fnma4_elt<mode>): Likewise.
7939 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7940 (aarch64_fnma4_elt_to_128v2df): Likewise.
7941 (aarch64_fnma4_elt_to_64df): Likewise.
7942 * config/aarch64/iterators.md (VDQSF): New.
7943 * config/aarch64/arm_neon.h
7944 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
7945 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
7947 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7949 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
7950 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
7951 (aarch64_mul3_elt_to_128df): Likewise.
7952 (aarch64_mul3_elt_to_64v2df): Likewise.
7953 * config/aarch64/iterators.md (VEL): Also handle DFmode.
7955 (VMUL_CHANGE_NLANES) Likewise.
7958 * config/aarch64/arm_neon.h
7959 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
7961 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7963 * config/aarch64/arm_neon.h
7964 (vcvtx_high_f32_f64): Fix parameters.
7966 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7967 Uros Bizjak <ubizjak@gmail.com>
7969 * config/alpha.c: Include tree-ssanames.h.
7971 2013-09-16 Richard Biener <rguenther@suse.de>
7973 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
7974 (dot_rdg_1): Handle control_dd.
7975 (create_edge_for_control_dependence): New function.
7976 (create_rdg_edges): Add control dependences if asked for.
7977 (build_rdg): Likewise.
7978 (generate_loops_for_partition): If there are not necessary
7979 control stmts remove all their dependencies.
7980 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
7981 (distribute_loop): Pass on control dependences.
7982 (tree_loop_distribution): Compute control dependences and remove
7983 restriction on number of loop nodes.
7985 2013-09-16 Jakub Jelinek <jakub@redhat.com>
7987 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
7990 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
7992 * cse.c (try_const_anchors): Punt on CC modes.
7994 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7996 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
7998 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
8001 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
8004 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8007 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
8009 (darwin_objc1_section): Likewise.
8010 (darwin_file_end): Emit Image Info section when required.
8012 2013-09-14 Jan Hubicka <jh@suse.cz>
8014 * tree-into-ssa.c (gate_into_ssa): New.
8015 (pass_data_build_ssa): Use it.
8016 * cgraph.h (expand_thunk): Update prototype.
8017 * cgraphunit.c (analyze_function): Expand thunks early.
8018 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
8019 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
8020 set lowered flag; do not add new function.
8021 (assemble_thunks_and_aliases): Update.
8022 * tree-ssa.c (gate_init_datastructures): New gate.
8023 (pass_data_init_datastructures): Use it.
8025 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8028 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
8029 xmm register set described in the psABI.
8031 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
8033 * dwarf2out.c (should_emit_struct_debug): Add check
8034 for type_decl variable is not NULL.
8036 2013-09-13 Jacek Caban <jacek@codeweavers.com>
8038 * config.gcc: Use new winnt-c.c target hooks
8039 * config/t-winnt: New file
8040 * config/winnt-c.c: New file
8041 * doc/tm.texi.in: Document new hook
8042 * doc/tm.texi: Regenerated
8044 2013-09-13 Jan Hubicka <jh@suse.cz>
8047 * ipa-inline.c (check_callers): New function.
8048 (check_caller_edge): Remove.
8049 (want_inline_function_to_all_callers_p): Also permit alises that are
8051 (inline_to_all_callers): Terminate the walk when devirtualization
8052 introduce new calls.
8054 2013-09-13 Jan Hubicka <jh@suse.cz>
8056 * ipa-inline-analysis.c (struct growth_data): Add node.
8057 (do_estimate_growth_1): Fix detection of recursion.
8059 2013-09-13 Jakub Jelinek <jakub@redhat.com>
8061 PR tree-optimization/58392
8062 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
8063 to avoid shadowing of outer loop variable. If
8064 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
8065 replace_by_duplicate_decl simduid of loops that have it set and
8066 set dest_cfun->has_simduid_loops and/or
8067 dest_cfun->has_force_vect_loops.
8068 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
8069 instead of maybe_lookup_decl.
8070 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
8071 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
8072 call. Copy over force_vect and copy and remap simduid. Set
8073 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
8074 (copy_cfg_body): Remove blocks_to_copy argument. Use
8075 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
8076 caller. Don't set cfun->has_simduid_loops and/or
8077 cfun->has_force_vect_loops here.
8078 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
8080 (expand_call_inline, tree_function_versioning): Adjust copy_body
8083 2013-09-13 Martin Jambor <mjambor@suse.cz>
8086 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
8087 the assert if the edge was a speculative one.
8089 2013-09-13 Richard Biener <rguenther@suse.de>
8091 * tree-data-ref.h (known_dependences_p): Move here ...
8092 * tree-loop-distribution.c (known_dependences_p): ... from here.
8093 (dump_rdg_component, debug_rdg_component): Remove.
8095 (generate_loops_for_partition): Use gimple_uid instead of
8096 relying on matching stmt visit order.
8097 (rdg_build_partitions): Take starting stmt vector.
8098 (ldist_gen): Merge into ...
8099 (distribute_loop): ... this function. Do not compute starting
8101 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
8103 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8105 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
8106 Set type attribute correctly. Set predicable_short_it attribute.
8107 (cmpsi_shiftsi): Remove %? from output template.
8109 2013-09-13 Richard Biener <rguenther@suse.de>
8111 * tree-loop-distribution.c (struct rdg_component,
8112 rdg_defs_used_in_other_loops_p, free_rdg_components,
8113 rdg_build_components): Remove.
8114 (stmts_from_loop): Do not record virtual PHIs.
8115 (generate_loops_for_partition): Skip virtual PHIs.
8116 (build_rdg_partition_for_component): Rename to ...
8117 (build_rdg_partition_for_vertex): ... this and adjust.
8118 (rdg_build_partitions): Take a vector of starting vertices
8119 instead of components. Remove unnecessary leftover handling.
8120 (ldist_gen): Do not build components or record other stores.
8121 (distribute_loop): Do not distribute loops containing stmts
8124 2013-09-13 Christian Bruel <christian.bruel@st.com>
8127 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
8129 2013-09-13 Kai Tietz <ktietz@redhat.com>
8131 * config.gcc: Separate cases for mingw and cygwin targets,
8132 and add 64-bit cygwin target case.
8134 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
8135 dll-export inline-functions.
8136 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
8139 2013-09-13 Jeff Law <law@redhat.com>
8143 2013-09-06 Jeff Law <law@redhat.com>
8145 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8146 edge implied equivalences into successor phis.
8148 2013-09-12 DJ Delorie <dj@redhat.com>
8150 * config/rl78/rl78-virt.md: Change from | to \; for asm line
8153 2013-09-12 Brooks Moses <bmoses@google.com>
8156 * Makefile.in: Do not install driver binaries in $(target)/bin.
8158 2013-09-12 DJ Delorie <dj@redhat.com>
8160 * config/rl78/rl78.opt (mrelax): New.
8161 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
8162 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
8164 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
8166 (rl78_expand_epilogue): Likewise.
8168 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
8171 * lra-eliminations.c (remove_reg_equal_offset_note): New.
8172 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
8173 pointer elimination with using remove_reg_equal_offset_note.
8175 2013-09-12 DJ Delorie <dj@redhat.com>
8177 * config/msp430/: New port.
8178 * config.gcc (msp430): Added.
8179 * doc/invoke.texi: Document MSP430 options.
8180 * doc/install.texi: Document msp430-elf
8181 * doc/md.texi: Document msp430-elf
8182 * doc/contrib.texi: Document msp430-elf
8184 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
8187 2013-09-12 Martin Jambor <mjambor@suse.cz>
8190 * ipa-prop.c (remove_described_reference): Give up if the edge in the
8191 reference descriptor is NULL.
8192 (ipa_edge_removal_hook): If owning a reference descriptor, set its
8195 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8197 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
8198 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
8199 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
8200 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
8201 make_temp_ssa_name): move to tree-ssanames.h
8202 * tree-ssa-alias.h: Move prototype.
8203 * tree-ssa.h: Include tree-ssanames.h.
8204 * tree-ssanames.c (FREE_SSANAMES): Move to here.
8205 * tree-ssanames.h: New. Move items from tree-flow*.h
8206 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
8208 2013-09-12 Richard Biener <rguenther@suse.de>
8210 PR tree-optimization/58404
8211 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
8212 propagate non-invariant addresses into dereferences wrapped
8213 in component references.
8215 2013-09-12 Richard Biener <rguenther@suse.de>
8217 PR tree-optimization/58402
8218 * passes.def: Move pass_late_warn_uninitialized later.
8220 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8222 * tree-ssa.h: New. Move content from tree-flow.h and
8224 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
8225 Move prototypes belonging to tree-ssa.c.
8226 * tree-flow-inline.h (redirect_edge_var_map_def,
8227 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
8229 * gimple.h: Adjust prototypes.
8230 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
8232 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
8233 * tree.h: Move prototype to tree-ssa.h.
8234 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
8235 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
8236 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
8237 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
8238 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
8239 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
8240 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
8241 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
8242 graphite-blocking.c, graphite-clast-to-gimple.c,
8243 graphite-dependences.c, graphite-interchange.c,
8244 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
8245 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
8246 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
8247 ipa-reference.c, ipa-split.c, ipa-utils.c,
8248 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
8249 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
8250 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
8251 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
8252 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
8253 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
8254 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
8255 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
8256 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
8257 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
8258 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
8259 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
8260 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
8261 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
8262 value-prof.c, var-tracking.c,
8263 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
8265 2013-09-12 Richard Biener <rguenther@suse.de>
8267 PR tree-optimization/58396
8268 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
8269 (build_rdg): Take a loop-nest parameter, fix memleaks.
8270 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
8272 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8274 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
8277 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
8279 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
8280 The FMA instruction names should have a 'v' prefix.
8282 2013-09-12 Richard Biener <rguenther@suse.de>
8284 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
8285 (dot_rdg): Use popen instead of system in optional code.
8286 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
8287 (already_processed_vertex_p): Adjust.
8288 (has_anti_or_output_dependence, predecessor_has_mem_write,
8289 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
8290 rdg_flag_uses): Remove.
8291 (rdg_flag_vertex): Simplify.
8292 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
8294 (build_rdg_partition_for_component): Process the first vertex
8295 of a component only.
8296 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
8298 2013-09-12 Alan Modra <amodra@gmail.com>
8300 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
8302 2013-09-11 DJ Delorie <dj@redhat.com>
8303 Nick Clifton <nickc@redhat.com>
8305 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
8306 (rl78_stack_based_mem): New.
8307 * config/rl78/constraints.md (Iv08): New.
8316 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
8318 (movsi): Change from expand to insn-and-split.
8319 (ashrsi3): Clobber AX.
8323 * config/rl78/rl78.md (CC_REG): Fix.
8324 (addsi3): Allow memory and immediate operands.
8325 (addsi3_internal): Split into...
8326 (addsi3_internal_virt): ...new, and ...
8327 (addsi3_internal_real): ...new.
8329 (subsi3_internal_virt): New.
8330 (subsi3_internal_real): New.
8331 (mulsi3): Add memory operand.
8332 (mulsi3_rl78): Likewise.
8333 (mulsi3_g13): Likewise.
8334 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
8335 (cbranchqi4_real): Add more constraint options.
8336 (cbranchhi4_real): Expand pattern.
8337 (cbranchhi4_real_signed): New.
8338 (cbranchhi4_real_inverted): New.
8339 (cbranchsi4_real_lt): New.
8340 (cbranchsi4_real_ge): New.
8341 (cbranchsi4_real_signed): New.
8342 (cbranchsi4_real): New.
8344 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
8346 (lshrsi3_virt): Likewise.
8347 (ashlsi3_virt): Likewise.
8348 (cbranchqi4_virt_signed): New.
8349 (cbranchhi4_virt_signed): New.
8350 (cbranchsi4_virt): New.
8351 * config/rl78/rl78.c: Whitespace fixes throughout.
8352 (move_elim_pass): New.
8353 (pass_data_rl78_move_elim): New.
8354 (pass_rl78_move_elim): New.
8355 (make_pass_rl78_move_elim): New.
8356 (rl78_devirt_info): Run devirt earlier.
8357 (rl78_move_elim_info): New.
8358 (rl78_asm_file_start): Register it.
8359 (rl78_split_movsi): New.
8360 (rl78_as_legitimate_address): Allow virtual base registers when
8362 (rl78_addr_space_convert): Remove spurious debug stuff.
8363 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
8364 (rl78_print_operand): More cases for not printing '#'.
8365 (rl78_expand_compare): Remove most of the logic.
8366 (content_memory): New.
8367 (clear_content_memory): New.
8368 (get_content_index): New.
8369 (get_content_name): New.
8370 (display_content_memory): New.
8371 (update_content): New.
8372 (record_content): New.
8373 (already_contains): New.
8374 (insn_ok_now): Re-recog insns with virtual registers.
8375 (add_postponed_content_update): New.
8376 (process_postponed_content_update): New.
8377 (gen_and_emit_move): New.
8378 (transcode_memory_rtx): Record new location content.
8379 Use gen_and_emit_move.
8380 (force_into_acc): New.
8381 (move_to_acc): Use gen_and_emit_move.
8382 (move_from_acc): Likewise.
8383 (move_acc_to_reg): Likewise.
8384 (move_to_x): Likewise.
8385 (move_to_hl): Likewise.
8386 (move_to_de): Likewise.
8387 (rl78_alloc_physical_registers_op1): Record location content.
8388 (has_constraint): New.
8389 (rl78_alloc_physical_registers_op2): Record location content.
8391 (rl78_alloc_physical_registers_ro1): Likewise.
8392 (rl78_alloc_physical_registers_cmp): Likewise.
8393 (rl78_alloc_physical_registers_umul): Likewise.
8394 (rl78_alloc_address_registers_macax): New.
8395 (rl78_alloc_physical_registers): Initialize and set location
8396 content memory as needed.
8397 (rl78_reorg): Make sure split2 is called.
8398 (rl78_rtx_costs): New.
8400 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
8402 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
8403 for (not (neg ...)) and (neg (not ...)) cases.
8405 2013-09-11 Richard Biener <rguenther@suse.de>
8408 * passes.def: Split critical edges before late uninit warning passes.
8409 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
8411 2013-09-11 Jakub Jelinek <jakub@redhat.com>
8413 PR tree-optimization/58385
8414 * fold-const.c (build_range_check): If both low and high are NULL,
8415 use omit_one_operand_loc to preserve exp side-effects.
8417 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8419 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
8421 2013-09-11 Richard Biener <rguenther@suse.de>
8423 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
8424 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
8425 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
8426 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
8427 stmts_from_loop, known_dependences_p, build_empty_rdg,
8428 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
8429 * tree-loop-distribution.c: ... here.
8430 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
8431 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
8432 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
8433 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
8434 * tree-loop-distribution.c: ... here.
8435 * tree-loop-distribution.c: Include gimple-pretty-print.h.
8436 (struct partition_s): Add loops member.
8437 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
8438 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
8439 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
8441 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8442 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8443 Sergey Lega <sergey.s.lega@intel.com>
8444 Anna Tikhonova <anna.tikhonova@intel.com>
8445 Ilya Tocar <ilya.tocar@intel.com>
8446 Andrey Turetskiy <andrey.turetskiy@intel.com>
8447 Ilya Verbin <ilya.verbin@intel.com>
8448 Kirill Yukhin <kirill.yukhin@intel.com>
8449 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8451 * config/i386/constraints.md (k): New.
8453 * config/i386/i386.c (const regclass_map): Add new mask registers.
8454 (dbx_register_map): Ditto.
8455 (dbx64_register_map): Ditto.
8456 (svr4_dbx_register_map): Ditto.
8457 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
8459 (ix86_preferred_reload_class): Disable constants for mask registers.
8460 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
8461 (ix86_hard_regno_mode_ok): Support new mask registers.
8462 (x86_order_regs_for_local_alloc): Ditto.
8463 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
8464 (FIXED_REGISTERS): Add new mask registers.
8465 (CALL_USED_REGISTERS): Ditto.
8466 (REG_ALLOC_ORDER): Ditto.
8467 (VALID_MASK_REG_MODE): New.
8468 (FIRST_MASK_REG): Ditto.
8469 (LAST_MASK_REG): Ditto.
8470 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
8471 (MAYBE_MASK_CLASS_P): New.
8472 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
8473 (REG_CLASS_CONTENTS): Ditto.
8474 (MASK_REGNO_P): New.
8475 (ANY_MASK_REG_P): Ditto.
8476 (HI_REGISTER_NAMES): Add new mask registers.
8477 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
8478 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
8480 (attribute "type"): Add mskmov, msklog.
8481 (attribute "length_immediate"): Support them.
8482 (attribute "memory"): Ditto.
8483 (attribute "prefix_0f"): Ditto.
8484 (*movhi_internal): Support new mask registers.
8485 (*movqi_internal): Ditto.
8486 (define_split): Split out clobber pattern is a logic
8487 insn on mask registers.
8488 (*k<logic><mode>): New.
8489 (*andhi_1): Extend to support mask regs.
8490 (*andqi_1): Extend to support mask regs.
8492 (define_split): Split and-not to and and not if operands
8494 (*<code><mode>_1): Separate HI mode to new pattern...
8495 (*<code>hi_1): This.
8496 (*<code>qi_1): Extend to support mask regs.
8498 (kortestzhi): Ditto.
8499 (kortestchi): Ditto.
8501 (*one_cmpl<mode>2_1): Remove HImode and handle it...
8502 (*one_cmplhi2_1): ...Here, now with mask registers support.
8503 (*one_cmplqi2_1): Support new mask registers.
8504 (HI/QImode arithmetics splitter): Don't split if mask registers
8506 (HI/QImode not splitter): Ditto.
8507 * config/i386/predicated.md (mask_reg_operand): New.
8508 (general_reg_operand): Ditto.
8510 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8512 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
8513 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
8515 2013-09-10 Jeff Law <law@redhat.com>
8517 PR tree-optimization/58380
8518 * tree-ssa-threadupdate.c (thread_block): Recognize another case
8519 of threading through a buried loop header.
8521 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
8522 return value for single successor case.
8524 2013-09-10 Jan Hubicka <jh@suse.cz>
8526 * ipa-devirt.c (ipa_devirt): Enable with LTO.
8528 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
8531 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
8532 support conditional execution.
8533 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
8535 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
8537 * lra.c (lra): Clear lra_optional_reload_pseudos before every
8539 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
8540 Check destination too to check move insn.
8541 (undo_optional_reloads): Add check that the original peudo did not
8542 changed its allocation and the optional reload was inherited on last
8543 inheritance pass. Break loop after deciding to keep optional reload.
8544 (lra_undo_inheritance): Add check that inherited pseudo still in
8547 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
8549 * config/aarch64/aarch64.md (generic_sched): New.
8550 * config/aarch64/aarch64-generic.md (load): Make conditional
8551 on generic_sched attribute.
8552 (nonload): Likewise.
8554 2013-09-10 Jan Hubicka <jh@suse.cz>
8556 * lto-cgraph.c: Include ipa-utils.h.
8557 (compute_ltrans_boundary): Also add possible targets into the boundary.
8559 2013-09-10 Jan Hubicka <jh@suse.cz>
8561 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
8562 VAR_DECL of vtable rather than full expression.
8564 2013-09-10 Jan Hubicka <jh@suse.cz>
8565 Paolo Carlini <paolo.carlini@oracle.com>
8567 * cgraphunit.c (analyze_functions): Save input_location, set it
8568 to UNKNOWN_LOCATION and restore it at the end.
8570 2013-09-10 Martin Jambor <mjambor@suse.cz>
8572 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
8573 represented by a thunk.
8575 2013-09-10 Jeff Law <law@redhat.com>
8577 PR tree-optimization/58343
8578 * tree-ssa-threadupdate.c (thread_block): Identify and disable
8579 jump threading requests through loop headers buried in the middle
8580 of a jump threading path.
8582 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
8583 in return value/type.
8585 2013-09-10 Jakub Jelinek <jakub@redhat.com>
8587 PR rtl-optimization/58365
8588 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
8589 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
8592 2013-09-10 Richard Biener <rguenther@suse.de>
8594 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
8595 * tree-data-ref.c (create_rdg_vertices): Collect all data
8596 references, signal failure to the caller, use data-ref API.
8597 (build_rdg): Compute data references only once. Maintain lifetime
8598 of data references and data dependences from within RDG.
8599 (free_rdg): Free dependence relations.
8600 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
8601 inventing extra dependences.
8602 (distribute_loop): Update for RDG API changes.
8604 2013-09-10 Kai Tietz <ktietz@redhat.com>
8606 * doc/invoke.texi (fms-extensions): Document changed
8607 behavior for ms-abi targets.
8608 * config/i386/i386.c (ix86_option_override_internal):
8609 Set default value of option -fms-extension for ms-abi targets.
8611 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
8613 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
8615 2013-09-10 Alan Modra <amodra@gmail.com>
8618 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
8620 2013-09-10 Alan Modra <amodra@gmail.com>
8622 * config/rs6000/predicates.md (add_cint_operand): New.
8623 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
8624 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
8625 using add_cint_operand.
8626 (largetoc_high_plus_aix): Likewise.
8628 2013-09-09 Jakub Jelinek <jakub@redhat.com>
8630 PR tree-optimization/58364
8631 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
8632 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
8633 the current range can't be an unconditional true or false.
8635 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
8637 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
8639 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
8641 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
8643 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
8646 * doc/invoke.texi (-Wdelete-incomplete): Document it.
8648 2013-09-09 Ian Bolton <ian.bolton@arm.com>
8650 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
8651 NO_REGS for immediate that can't be moved directly into FP_REGS.
8653 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8655 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
8656 comparison with negated operand.
8657 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
8660 2013-09-09 Richard Biener <rguenther@suse.de>
8663 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
8664 of a subloop record all its block as affecting loop-closed SSA form.
8666 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8668 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
8669 of an rtx/bitpos pair.
8670 (store_fixed_bit_field): Update accordingly.
8672 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8674 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
8676 * builtins.c (expand_errno_check): Likewise.
8677 * dwarf2cfi.c (init_return_column_size): Likewise.
8678 * except.c (sjlj_mark_call_sites): Likewise.
8679 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
8680 * lra-constraints.c (emit_inc): Likewise.
8681 * ree.c (combine_set_extension): Likewise.
8682 * regmove.c (fixup_match_2): Likewise.
8683 * reload1.c (inc_for_reload): Likewise.
8685 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8687 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
8688 (force_to_mode, simplify_shift_const_1, simplify_comparison):
8689 Use gen_int_mode with the mode of the associated simplify_* call.
8690 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
8691 * expmed.c (expand_shift_1): Likewise.
8692 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8693 * loop-iv.c (iv_number_of_iterations): Likewise.
8694 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8695 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
8697 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8699 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
8700 of the associated expand_* call.
8701 (asan_emit_stack_protection): Likewise.
8702 * builtins.c (round_trampoline_addr): Likewise.
8703 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
8704 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
8705 (emit_store_flag): Likewise.
8706 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
8707 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
8709 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8710 * ifcvt.c (noce_try_store_flag_constants): Likewise.
8711 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8712 * modulo-sched.c (generate_prolog_epilog): Likewise.
8713 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
8714 (expand_ctz, expand_ffs, expand_unop): Likewise.
8716 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8718 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
8719 of the associated gen_rtx_* call.
8720 * caller-save.c (init_caller_save): Likewise.
8721 * combine.c (find_split_point, make_extraction): Likewise.
8722 (make_compound_operation): Likewise.
8723 * dwarf2out.c (mem_loc_descriptor): Likewise.
8724 * explow.c (plus_constant, probe_stack_range): Likewise.
8725 * expmed.c (expand_mult_const): Likewise.
8726 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
8727 * reload1.c (init_reload): Likewise.
8728 * valtrack.c (cleanup_auto_inc_dec): Likewise.
8729 * var-tracking.c (adjust_mems): Likewise.
8730 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
8731 rather than gen_rtx_fmt_ee.
8733 2013-09-09 Jan Hubicka <jh@suse.cz>
8736 * value-prof.c (gimple_ic): Copy also abnormal edges.
8738 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8740 * asan.c (asan_shadow_cst): Use gen_int_mode.
8742 2013-09-08 Jan Hubicka <jh@suse.cz>
8744 * ipa-profile.c: Add toplevel comment.
8745 (ipa_propagate_frequency_1): Be more conservative when profile is read.
8746 (contains_hot_call_p): New function.
8747 (ipa_propagate_frequency): Set frequencies based on counts when
8749 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
8750 profile; do not tamper with profile after inlining if it is read.
8752 2013-09-08 Jan Hubicka <jh@suse.cz>
8754 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
8757 2013-09-08 Jan Hubicka <jh@suse.cz>
8759 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
8762 2013-09-08 Jeff Law <law@redhat.com>
8765 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
8768 2013-09-08 Andi Kleen <ak@linux.intel.com>
8770 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
8772 2013-09-08 Jan Hubicka <jh@suse.cz>
8774 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
8776 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
8778 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
8779 for non-debug insns.
8780 * lra.c (new_insn_reg): Take the containing insn as a parameter.
8781 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
8782 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
8785 2013-09-08 Jan Hubicka <jh@suse.cz>
8787 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
8788 targets and devirtualize to BUILT_IN_UNREACHABLE.
8789 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
8790 * ipa.c (walk_polymorphic_call_targets): New function.
8791 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
8792 functions; use the new timevar.
8793 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
8794 was removed from the program.
8795 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
8796 not consider it in the walk when its vtable is dead.
8797 (possible_polymorphic_call_targets_1): Pass anonymous flag to
8799 (devirt_variable_node_removal_hook): New function.
8800 (possible_polymorphic_call_targets): Also register
8801 devirt_variable_node_removal_hook.
8802 (ipa_devirt): Do not do non-speculative devirtualization.
8803 (gate_ipa_devirt): One execute if devirtualizing speculatively.
8805 2013-09-08 Jan Hubicka <jh@suse.cz>
8807 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
8808 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
8809 varpool_remove_variable_insertion_hook): Declare.
8810 * varpool.c (varpool_node_hook_list): New structure.
8811 (first_varpool_node_removal_hook,
8812 first_varpool_variable_insertion_hook): New variables.
8813 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
8814 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
8815 varpool_remove_variable_insertion_hook,
8816 varpool_call_variable_insertion_hooks): New functions.
8817 (varpool_remove_node): Use it.
8819 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
8822 * diagnostic.c (diagnostic_build_prefix): When s.file is
8823 "<built-in>" don't output line and column numbers.
8825 2013-09-06 Jan Hubicka <jh@suse.cz>
8827 * cgraphunit.c (expand_thunk): Get body before touching arguments.
8828 * lto-streamer-out.c: Stream thunks, too.
8829 * lto-streamer-in.c (input_function): Pop cfun here
8830 (lto_read_body): Instead of here.
8832 2013-09-06 Caroline Tice <cmtice@google.com>
8834 * doc/install.texi: Add documentation for the --enable-vtable-verify
8835 and the --disable-libvtv configure options.
8837 2013-09-06 Jeff Law <law@redhat.com>
8839 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8840 edge implied equivalences into successor phis.
8842 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
8844 * resource.c (mark_referenced_resources): Handle COND_EXEC.
8846 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
8848 * resource.c (mark_target_live_regs): Compute resources taking
8849 into account if a call is predicated or not.
8851 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
8853 * toplev.c (output_stack_usage): Be prepared for suffixes created by
8854 the compiler in the function names.
8856 2013-09-06 Jan Hubicka <jh@suse.cz>
8859 * ipa-inline.c (has_caller_p): New function.
8860 (want_inline_function_to_all_callers_p): Use it.
8861 (sum_callers, inline_to_all_callers): Break out from ...
8862 (ipa_inline): ... here.
8864 2013-09-06 Jan Hubicka <jh@suse.cz>
8866 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
8867 only when AVX is enabled.
8869 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8871 * config/aarch64/aarch64.md
8872 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
8873 is fpsimd_<load/store>2.
8874 (load_pair<mode>): Likewise.
8875 (store_pair<mode>): Likewise.
8877 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8879 * config/arm/types.md (type): Add "mrs" type.
8880 * config/aarch64/aarch64.md
8881 (aarch64_load_tp_hard): Make type "mrs".
8883 (load_tp_hard): Make type "mrs".
8884 * config/arm/cortex-a15.md: Update with new attributes.
8885 * config/arm/cortex-a5.md: Update with new attributes.
8886 * config/arm/cortex-a53.md: Update with new attributes.
8887 * config/arm/cortex-a7.md: Update with new attributes.
8888 * config/arm/cortex-a8.md: Update with new attributes.
8889 * config/arm/cortex-a9.md: Update with new attributes.
8890 * config/arm/cortex-m4.md: Update with new attributes.
8891 * config/arm/cortex-r4.md: Update with new attributes.
8892 * config/arm/fa526.md: Update with new attributes.
8893 * config/arm/fa606te.md: Update with new attributes.
8894 * config/arm/fa626te.md: Update with new attributes.
8895 * config/arm/fa726te.md: Update with new attributes.
8897 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8899 * config/aarch64/aarch64.md
8900 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
8901 (*movtf_aarch64): Likewise.
8903 (thumb1_movdi_insn): Use "multiple" for type where more than one
8904 instruction is used for a move.
8905 (*arm32_movhf): Likewise.
8906 (*thumb_movdf_insn): Likewise.
8908 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8910 * config/arm/types.md (type): Rename fcpys to fmov.
8912 (*arm_movsi_vfp): Rename type fcpys as fmov.
8913 (*thumb2_movsi_vfp): Likewise
8914 (*movhf_vfp_neon): Likewise
8915 (*movhf_vfp): Likewise
8916 (*movsf_vfp): Likewise
8917 (*thumb2_movsf_vfp): Likewise
8918 (*movsfcc_vfp): Likewise
8919 (*thumb2_movsfcc_vfp): Likewise
8920 * config/aarch64/aarch64-simd.md
8921 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
8922 * config/aarch64/aarch64.md
8923 (*movsi_aarch64): Replace type mov_reg with fmovs.
8924 (*movdi_aarch64): Likewise
8925 (*movsf_aarch64): Likewise
8926 (*movdf_aarch64): Likewise
8928 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
8929 * config/arm/iwmmxt.md
8930 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
8931 * config/arm/arm1020e.md: Update with new attributes.
8932 * config/arm/cortex-a15-neon.md: Update with new attributes.
8933 * config/arm/cortex-a5.md: Update with new attributes.
8934 * config/arm/cortex-a53.md: Update with new attributes.
8935 * config/arm/cortex-a7.md: Update with new attributes.
8936 * config/arm/cortex-a8-neon.md: Update with new attributes.
8937 * config/arm/cortex-a9.md: Update with new attributes.
8938 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8939 * config/arm/cortex-r4f.md: Update with new attributes.
8940 * config/arm/marvell-pj4.md: Update with new attributes.
8941 * config/arm/vfp11.md: Update with new attributes.
8943 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8945 * config/aarch64/aarch64.md
8946 (*madd<mode>): Fix type attribute.
8947 (*maddsi_uxtw): Likewise.
8948 (*msub<mode>): Likewise.
8949 (*msubsi_uxtw): Likewise.
8950 (<su_optab>maddsidi4): Likewise.
8951 (<su_optab>msubsidi4): Likewise.
8953 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8955 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
8956 * config/arm/arm.md (core_cycles): Remove fdiv.
8957 * config/arm/vfp.md:
8958 (*sqrtsf2_vfp): Update for attribute changes.
8959 (*sqrtdf2_vfp): Likewise.
8960 * config/aarch64/aarch64.md:
8961 (sqrt<mode>2): Update for attribute changes.
8962 * config/arm/arm1020e.md: Update with new attributes.
8963 * config/arm/cortex-a15-neon.md: Update with new attributes.
8964 * config/arm/cortex-a5.md: Update with new attributes.
8965 * config/arm/cortex-a53.md: Update with new attributes.
8966 * config/arm/cortex-a7.md: Update with new attributes.
8967 * config/arm/cortex-a8-neon.md: Update with new attributes.
8968 * config/arm/cortex-a9.md: Update with new attributes.
8969 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8970 * config/arm/cortex-r4f.md: Update with new attributes.
8971 * config/arm/marvell-pj4.md: Update with new attributes.
8972 * config/arm/vfp11.md: Update with new attributes.
8974 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8976 * config/arm/types.md
8977 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
8978 * config/aarch64/aarch64.md
8979 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
8981 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
8982 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
8983 (float<GPI:mode><GPF:mode>2): Likewise.
8985 (*truncsisf2_vfp): Update with new attributes.
8986 (*truncsidf2_vfp): Likewise.
8987 (fixuns_truncsfsi2): Likewise.
8988 (fixuns_truncdfsi2): Likewise.
8989 (*floatsisf2_vfp): Likewise.
8990 (*floatsidf2_vfp): Likewise.
8991 (floatunssisf2): Likewise.
8992 (floatunssidf2): Likewise.
8993 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
8994 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
8995 * config/arm/arm1020e.md: Update with new attributes.
8996 * config/arm/cortex-a15-neon.md: Update with new attributes.
8997 * config/arm/cortex-a5.md: Update with new attributes.
8998 * config/arm/cortex-a53.md: Update with new attributes.
8999 * config/arm/cortex-a7.md: Update with new attributes.
9000 * config/arm/cortex-a8-neon.md: Update with new attributes.
9001 * config/arm/cortex-a9.md: Update with new attributes.
9002 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9003 * config/arm/cortex-r4f.md: Update with new attributes.
9004 * config/arm/marvell-pj4.md: Update with new attributes.
9005 * config/arm/vfp11.md: Update with new attributes.
9007 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9009 * config/aarch64/arm_neon.h
9010 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
9011 (vqtbx<1,2,3,4><q>_s8): Likewise.
9013 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9015 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
9016 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
9017 (add<mode>3): Add type attribute.
9018 (add<mode>3): Likewise.
9019 (usadd<mode>3): Likewise.
9020 (ssadd<mode>3): Likewise.
9021 (sub<mode>3): Likewise.
9022 (sub<mode>3): Likewise.
9023 (ussub<mode>3): Likewise.
9024 (sssub<mode>3): Likewise.
9025 (ssmulsa3): Likewise.
9026 (usmulusa3): Likewise.
9027 (arm_usatsihi): Likewise.
9029 (*movdi_vfp): Add types for all instructions.
9030 (*movdi_vfp_cortexa8): Likewise.
9031 (*movhf_vfp_neon): Likewise.
9032 (*movhf_vfp): Likewise.
9033 (*movdf_vfp): Likewise.
9034 (*thumb2_movdf_vfp): Likewise.
9035 (*thumb2_movdfcc_vfp): Likewise.
9036 * config/arm/arm.md: Add type attribute to all insn patterns.
9037 (*thumb1_adddi3): Add type attribute.
9038 (*arm_adddi3): Likewise.
9039 (*adddi_sesidi_di): Likewise.
9040 (*adddi_zesidi_di): Likewise.
9041 (*thumb1_addsi3): Likewise.
9042 (addsi3_compare0): Likewise.
9043 (*addsi3_compare0_scratch): Likewise.
9044 (*compare_negsi_si): Likewise.
9045 (cmpsi2_addneg): Likewise.
9046 (*addsi3_carryin_<optab>): Likewise.
9047 (*addsi3_carryin_alt2_<optab>): Likewise.
9048 (*addsi3_carryin_clobercc_<optab>): Likewise.
9049 (*subsi3_carryin): Likewise.
9050 (*subsi3_carryin_const): Likewise.
9051 (*subsi3_carryin_compare): Likewise.
9052 (*subsi3_carryin_compare_const): Likewise.
9053 (*arm_subdi3): Likewise.
9054 (*thumb_subdi3): Likewise.
9055 (*subdi_di_zesidi): Likewise.
9056 (*subdi_di_sesidi): Likewise.
9057 (*subdi_zesidi_di): Likewise.
9058 (*subdi_sesidi_di): Likewise.
9059 (*subdi_zesidi_ze): Likewise.
9060 (thumb1_subsi3_insn): Likewise.
9061 (*arm_subsi3_insn): Likewise.
9062 (*anddi3_insn): Likewise.
9063 (*anddi_zesidi_di): Likewise.
9064 (*anddi_sesdi_di): Likewise.
9065 (*ne_zeroextracts): Likewise.
9066 (*ne_zeroextracts): Likewise.
9067 (*ite_ne_zeroextr): Likewise.
9068 (*ite_ne_zeroextr): Likewise.
9069 (*anddi_notdi_di): Likewise.
9070 (*anddi_notzesidi): Likewise.
9071 (*anddi_notsesidi): Likewise.
9072 (andsi_notsi_si): Likewise.
9073 (thumb1_bicsi3): Likewise.
9074 (*iordi3_insn): Likewise.
9075 (*iordi_zesidi_di): Likewise.
9076 (*iordi_sesidi_di): Likewise.
9077 (*thumb1_iorsi3_insn): Likewise.
9078 (*xordi3_insn): Likewise.
9079 (*xordi_zesidi_di): Likewise.
9080 (*xordi_sesidi_di): Likewise.
9081 (*arm_xorsi3): Likewise.
9082 (*andsi_iorsi3_no): Likewise.
9083 (*smax_0): Likewise.
9084 (*smax_m1): Likewise.
9085 (*arm_smax_insn): Likewise.
9086 (*smin_0): Likewise.
9087 (*arm_smin_insn): Likewise.
9088 (*arm_umaxsi3): Likewise.
9089 (*arm_uminsi3): Likewise.
9090 (*minmax_arithsi): Likewise.
9091 (*minmax_arithsi_): Likewise.
9092 (*satsi_<SAT:code>): Likewise.
9093 (arm_ashldi3_1bit): Likewise.
9094 (arm_ashrdi3_1bit): Likewise.
9095 (arm_lshrdi3_1bit): Likewise.
9096 (*arm_negdi2): Likewise.
9097 (*thumb1_negdi2): Likewise.
9098 (*arm_negsi2): Likewise.
9099 (*thumb1_negsi2): Likewise.
9100 (*negdi_extendsid): Likewise.
9101 (*negdi_zero_extend): Likewise.
9102 (*arm_abssi2): Likewise.
9103 (*thumb1_abssi2): Likewise.
9104 (*arm_neg_abssi2): Likewise.
9105 (*thumb1_neg_abss): Likewise.
9106 (one_cmpldi2): Likewise.
9107 (extend<mode>di2): Likewise.
9108 (*compareqi_eq0): Likewise.
9109 (*arm_extendhisi2addsi): Likewise.
9110 (*arm_movdi): Likewise.
9111 (*thumb1_movdi_insn): Likewise.
9112 (*arm_movt): Likewise.
9113 (*thumb1_movsi_insn): Likewise.
9114 (pic_add_dot_plus_four): Likewise.
9115 (pic_add_dot_plus_eight): Likewise.
9116 (tls_load_dot_plus_eight): Likewise.
9117 (*thumb1_movhi_insn): Likewise.
9118 (*thumb1_movsf_insn): Likewise.
9119 (*movdf_soft_insn): Likewise.
9120 (*thumb_movdf_insn): Likewise.
9121 (cbranchsi4_insn): Likewise.
9122 (cbranchsi4_scratch): Likewise.
9123 (*negated_cbranchsi4): Likewise.
9124 (*tbit_cbranch): Likewise.
9125 (*tlobits_cbranch): Likewise.
9126 (*tstsi3_cbranch): Likewise.
9127 (*cbranchne_decr1): Likewise.
9128 (*addsi3_cbranch): Likewise.
9129 (*addsi3_cbranch_scratch): Likewise.
9130 (*arm_cmpdi_insn): Likewise.
9131 (*arm_cmpdi_unsig): Likewise.
9132 (*arm_cmpdi_zero): Likewise.
9133 (*thumb_cmpdi_zero): Likewise.
9134 (*deleted_compare): Likewise.
9135 (*mov_scc): Likewise.
9136 (*mov_negscc): Likewise.
9137 (*mov_notscc): Likewise.
9138 (*cstoresi_eq0_thumb1_insn): Likewise.
9139 (cstoresi_nltu_thumb1): Likewise.
9140 (cstoresi_ltu_thu): Likewise.
9141 (thumb1_addsi3_addgeu): Likewise.
9142 (*arm_jump): Likewise.
9143 (*thumb_jump): Likewise.
9144 (*check_arch2): Likewise.
9145 (arm_casesi_internal): Likewise.
9146 (thumb1_casesi_dispatch): Likewise.
9147 (*arm_indirect_jump): Likewise.
9148 (*thumb1_indirect_jump): Likewise.
9150 (*and_scc): Likewise.
9151 (*ior_scc): Likewise.
9152 (*compare_scc): Likewise.
9153 (*cond_move): Likewise.
9154 (*cond_arith): Likewise.
9155 (*cond_sub): Likewise.
9156 (*cmp_ite0): Likewise.
9157 (*cmp_ite1): Likewise.
9158 (*cmp_and): Likewise.
9159 (*cmp_ior): Likewise.
9160 (*ior_scc_scc): Likewise.
9161 (*ior_scc_scc_cmp): Likewise.
9162 (*and_scc_scc): Likewise.
9163 (*and_scc_scc_cmp): Likewise.
9164 (*and_scc_scc_nod): Likewise.
9165 (*negscc): Likewise.
9166 (movcond_addsi): Likewise.
9167 (movcond): Likewise.
9168 (*ifcompare_plus_move): Likewise.
9169 (*if_plus_move): Likewise.
9170 (*ifcompare_move_plus): Likewise.
9171 (*if_move_plus): Likewise.
9172 (*ifcompare_arith_arith): Likewise.
9173 (*if_arith_arith): Likewise.
9174 (*ifcompare_arith_move): Likewise.
9175 (*if_arith_move): Likewise.
9176 (*ifcompare_move_arith): Likewise.
9177 (*if_move_arith): Likewise.
9178 (*ifcompare_move_not): Likewise.
9179 (*if_move_not): Likewise.
9180 (*ifcompare_not_move): Likewise.
9181 (*if_not_move): Likewise.
9182 (*ifcompare_shift_move): Likewise.
9183 (*if_shift_move): Likewise.
9184 (*ifcompare_move_shift): Likewise.
9185 (*if_move_shift): Likewise.
9186 (*ifcompare_shift_shift): Likewise.
9187 (*ifcompare_not_arith): Likewise.
9188 (*ifcompare_arith_not): Likewise.
9189 (*if_arith_not): Likewise.
9190 (*ifcompare_neg_move): Likewise.
9191 (*if_neg_move): Likewise.
9192 (*ifcompare_move_neg): Likewise.
9193 (*if_move_neg): Likewise.
9194 (prologue_thumb1_interwork): Likewise.
9195 (*cond_move_not): Likewise.
9196 (*sign_extract_onebit): Likewise.
9197 (*not_signextract_onebit): Likewise.
9198 (stack_tie): Likewise.
9199 (align_4): Likewise.
9200 (align_8): Likewise.
9201 (consttable_end): Likewise.
9202 (consttable_1): Likewise.
9203 (consttable_2): Likewise.
9204 (consttable_4): Likewise.
9205 (consttable_8): Likewise.
9206 (consttable_16): Likewise.
9207 (*thumb1_tablejump): Likewise.
9208 (prefetch): Likewise.
9209 (force_register_use): Likewise.
9210 (thumb_eh_return): Likewise.
9211 (load_tp_hard): Likewise.
9212 (load_tp_soft): Likewise.
9213 (tlscall): Likewise.
9214 (*arm_movtas_ze): Likewise.
9215 (*arm_rev): Likewise.
9216 (*arm_revsh): Likewise.
9217 (*arm_rev16): Likewise.
9218 * config/arm/thumb2.md
9219 (*thumb2_smaxsi3): Likewise.
9220 (*thumb2_sminsi3): Likewise.
9221 (*thumb32_umaxsi3): Likewise.
9222 (*thumb2_uminsi3): Likewise.
9223 (*thumb2_negdi2): Likewise.
9224 (*thumb2_abssi2): Likewise.
9225 (*thumb2_neg_abss): Likewise.
9226 (*thumb2_movsi_insn): Likewise.
9227 (tls_load_dot_plus_four): Likewise.
9228 (*thumb2_movhi_insn): Likewise.
9229 (*thumb2_mov_scc): Likewise.
9230 (*thumb2_mov_negs): Likewise.
9231 (*thumb2_mov_negs): Likewise.
9232 (*thumb2_mov_nots): Likewise.
9233 (*thumb2_mov_nots): Likewise.
9234 (*thumb2_movsicc_): Likewise.
9235 (*thumb2_movsfcc_soft_insn): Likewise.
9236 (*thumb2_indirect_jump): Likewise.
9237 (*thumb2_and_scc): Likewise.
9238 (*thumb2_ior_scc): Likewise.
9239 (*thumb2_ior_scc_strict_it): Likewise.
9240 (*thumb2_cond_move): Likewise.
9241 (*thumb2_cond_arith): Likewise.
9242 (*thumb2_cond_ari): Likewise.
9243 (*thumb2_cond_sub): Likewise.
9244 (*thumb2_negscc): Likewise.
9245 (*thumb2_movcond): Likewise.
9246 (thumb2_casesi_internal): Likewise.
9247 (thumb2_casesi_internal_pic): Likewise.
9248 (*thumb2_alusi3_short): Likewise.
9249 (*thumb2_mov<mode>_shortim): Likewise.
9250 (*thumb2_addsi_short): Likewise.
9251 (*thumb2_subsi_short): Likewise.
9252 (thumb2_addsi3_compare0): Likewise.
9253 (*thumb2_cbz): Likewise.
9254 (*thumb2_cbnz): Likewise.
9255 (*thumb2_one_cmplsi2_short): Likewise.
9256 (*thumb2_negsi2_short): Likewise.
9257 (*orsi_notsi_si): Likewise.
9258 * config/arm/arm1020e.md: Update with new attributes.
9259 * config/arm/arm1026ejs.md: Update with new attributes.
9260 * config/arm/arm1136jfs.md: Update with new attributes.
9261 * config/arm/arm926ejs.md: Update with new attributes.
9262 * config/arm/cortex-a15.md: Update with new attributes.
9263 * config/arm/cortex-a5.md: Update with new attributes.
9264 * config/arm/cortex-a53.md: Update with new attributes.
9265 * config/arm/cortex-a7.md: Update with new attributes.
9266 * config/arm/cortex-a8.md: Update with new attributes.
9267 * config/arm/cortex-a9.md: Update with new attributes.
9268 * config/arm/cortex-m4.md: Update with new attributes.
9269 * config/arm/cortex-r4.md: Update with new attributes.
9270 * config/arm/fa526.md: Update with new attributes.
9271 * config/arm/fa606te.md: Update with new attributes.
9272 * config/arm/fa626te.md: Update with new attributes.
9273 * config/arm/fa726te.md: Update with new attributes.
9275 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9277 * config/aarch64/aarch64-simd.md
9278 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
9279 <vwx> iterator to ensure correct register choice.
9280 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
9281 (aarch64_sqdmull_n<mode>): Likewise.
9282 (aarch64_sqdmull2_n<mode>_internal): Likewise.
9283 * config/aarch64/arm_neon.h
9284 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
9285 (vml<as><q>_n_<su>16): Likewise.
9286 (vml<as>l_high_lane<q>_<su>16): Likewise.
9287 (vml<as>l_high_n_<su>16): Likewise.
9288 (vml<as>l_lane<q>_<su>16): Likewise.
9289 (vml<as>l_n_<su>16): Likewise.
9290 (vmul<q>_lane<q>_<su>16): Likewise.
9291 (vmul<q>_n_<su>16): Likewise.
9292 (vmull_lane<q>_<su>16): Likewise.
9293 (vmull_n_<su>16): Likewise.
9294 (vmull_high_lane<q>_<su>16): Likewise.
9295 (vmull_high_n_<su>16): Likewise.
9296 (vqrdmulh<q>_n_s16): Likewise.
9298 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
9300 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
9301 have the correct lane parameter.
9303 2013-09-06 Richard Biener <rguenther@suse.de>
9305 * cfganal.c (control_dependences::~control_dependences):
9306 Properly free all of the vector.
9308 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
9311 * config/i386/i386.c (ix86_conditional_register_usage):
9312 Proper initialize extended SSE registers.
9314 2013-09-06 Jan Hubicka <jh@suse.cz>
9316 PR tree-optimization/58311
9317 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
9319 2013-09-06 Jan Hubicka <jh@suse.cz>
9321 * Makefile.in (tree-sra.o): Update dependencies.
9322 * tree-sra.c: Include ipa-utils.h
9323 (scan_function): Use recursive_call_p.
9324 (has_caller_p): New function.
9325 (cgraph_for_node_and_aliases): Count also callers of aliases.
9327 2013-09-06 Jan Hubicka <jh@suse.cz>
9330 * cgraph.h (symtab_semantically_equivalent_p): Declare.
9331 * tree-tailcall.c: Include ipa-utils.h.
9332 (find_tail_calls): Use it.
9333 * ipa-pure-const.c (check_call): Likewise.
9334 * ipa-utils.c (recursive_call_p): New function.
9335 * ipa-utils.h (recursive_call_p): Dclare.
9336 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
9337 (symtab_semantically_equivalent_p): New function.
9338 * Makefile.in (tree-tailcall.o): Update dependencies.
9340 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9342 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
9345 2013-09-06 Richard Biener <rguenther@suse.de>
9347 * lto-streamer.h (lto_global_var_decls): Remove.
9348 * Makefile.in (OBJS): Remove lto-symtab.o.
9349 (lto-symtab.o): Remove.
9350 (GTFILES): Remove lto-symtab.c
9351 * lto-symtab.c: Move to lto/
9353 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9355 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
9356 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
9357 (UNSPEC_FPINT_RINT): New constant definitions.
9358 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
9359 definition with 2 attributes.
9360 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
9361 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
9364 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9366 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
9367 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
9368 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
9369 attribute to "z196_alone".
9370 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
9373 2013-09-06 Richard Biener <rguenther@suse.de>
9375 * basic-block.h (class control_dependences): New.
9376 * tree-ssa-dce.c (control_dependence_map): Remove.
9378 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
9379 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
9380 find_pdom, find_control_dependence, find_all_control_dependences):
9382 (mark_control_dependent_edges_necessary,
9383 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
9384 tree_dce_done, perform_tree_ssa_dce): Adjust.
9385 * cfganal.c (set_control_dependence_map_bit,
9386 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
9387 find_all_control_dependences): Move from tree-ssa-dce.c and
9388 implement as methods of control_dependences class.
9389 (control_dependences::control_dependences): New.
9390 (control_dependences::~control_dependences): Likewise.
9391 (control_dependences::get_edges_dependent_on): Likewise.
9392 (control_dependences::get_edge): Likewise.
9394 2013-09-04 Jan Hubicka <jh@suse.cz>
9396 * tree.c (types_same_for_odr): Drop overactive check.
9397 * ipa-devirt.c (hash_type_name): Likewise.
9399 2013-09-04 Jan Hubicka <jh@suse.cz>
9401 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
9402 (analyze_functions): ... here.
9404 2013-09-04 Jan Hubicka <jh@suse.cz>
9407 * cgraphunit.c (analyze_functions): Clear AUX fields
9408 after processing; initialize assembler name has.
9410 2013-09-05 Jeff Law <law@redhat.com>
9412 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
9413 from thread_around_empty_block. Record threading path into PATH.
9414 Recurse if threading through the initial block is successful.
9415 (thread_across_edge): Corresponding changes to slightly simplify.
9417 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9419 * config/aarch64/aarch64.md
9420 (type): Remove frecpe, frecps, frecpx.
9421 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
9422 fix to be a TARGET_SIMD instruction.
9423 (aarch64_frecps): Remove.
9424 * config/aarch64/aarch64-simd.md
9425 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
9426 (aarch64_frecps<mode>): Handle all float/vector of float modes.
9428 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9429 Sofiane Naci <sofiane.naci@arm.com>
9431 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
9432 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
9433 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
9434 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
9435 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
9436 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
9437 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
9438 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
9439 into "clz, "rbit". Rename "shift" to "shift_imm".
9440 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
9441 changes. Update for attribute changes all occurrences of arlo_* and
9443 * config/arm/arm-fixed.md: Update for attribute changes
9444 all occurrences of arlo_* types.
9445 * config/arm/thumb2.md: Update for attribute changes all occurrences
9447 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
9448 (cortexa7_older_only): Likewise.
9449 (cortexa7_younger): Likewise.
9450 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9451 (1020alu_shift_op): Likewise.
9452 (1020alu_shift_reg_op): Likewise.
9453 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9454 (alu_shift_op): Likewise.
9455 (alu_shift_reg_op): Likewise.
9456 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9457 (11_alu_shift_op): Likewise.
9458 (11_alu_shift_reg_op): Likewise.
9459 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9460 (9_alu_shift_reg_op): Likewise.
9461 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
9463 (cortex_a15_alu_shift): Likewise.
9464 (cortex_a15_alu_shift_reg): Likewise.
9465 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
9467 (cortex_a5_alu_shift): Likewise.
9468 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
9470 (cortex_a53_alu_shift): Likewise.
9471 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
9473 (cortex_a7_alu_reg): Likewise.
9474 (cortex_a7_alu_shift): Likewise.
9475 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
9477 (cortex_a8_alu_shift): Likewise.
9478 (cortex_a8_alu_shift_reg): Likewise.
9479 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9480 (cortex_a9_dp_shift): Likewise.
9481 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
9483 * config/arm/cortex-r4.md
9484 (cortex_r4_alu): Update for attribute changes.
9485 (cortex_r4_mov): Likewise.
9486 (cortex_r4_alu_shift_reg): Likewise.
9487 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9488 (526_alu_shift_op): Likewise.
9489 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9490 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9491 (626te_alu_shift_op): Likewise.
9492 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
9493 (726te_alu_shift_op): Likewise.
9494 (726te_alu_shift_reg_op): Likewise.
9495 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9496 (mp626_alu_shift_op): Likewise.
9497 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
9498 (pj4_alu_conds): Likewise.
9499 (pj4_shift): Likewise.
9500 (pj4_shift_conds): Likewise.
9501 (pj4_alu_shift): Likewise.
9502 (pj4_alu_shift_conds): Likewise.
9503 * config/aarch64/aarch64.md: Update for attribute change
9504 all occurrences of arlo_* and shift* types.
9506 2013-09-05 Mike Stump <mikestump@comcast.net>
9508 * tree.h: Move documentation for tree_function_decl to tree-core.h
9509 with the declaration.
9511 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
9514 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
9515 looking for widest mode.
9517 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
9519 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
9521 2013-09-05 Richard Biener <rguenther@suse.de>
9523 PR tree-optimization/58137
9524 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
9525 Do not create vectors of pointers.
9526 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
9527 types for the components of the vector initializer.
9528 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
9529 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
9531 2013-09-05 Martin Jambor <mjambor@suse.cz>
9533 * ipa-prop.c (remove_described_reference): Accept missing references,
9534 return false if that hppens, otherwise return true.
9535 (cgraph_node_for_jfunc): New function.
9536 (try_decrement_rdesc_refcount): Likewise.
9537 (try_make_edge_direct_simple_call): Use them.
9538 (ipa_edge_removal_hook): Remove references from rdescs.
9539 (ipa_edge_duplication_hook): Clone rdescs and their references
9540 when the new edge has the same caller as the old one.
9541 * cgraph.c (cgraph_resolve_speculation): Remove speculative
9542 reference before removing any edges.
9544 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
9546 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
9548 * thumb2.md (thumb2_storewb_parisi): New pattern.
9550 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
9552 * config/aarch64/aarch64-option-extensions.def: Add
9553 AARCH64_OPT_EXTENSION of 'crc'.
9554 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
9555 (AARCH64_ISA_CRC): Ditto.
9556 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
9557 description of the CRC extension.
9559 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
9561 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
9562 * config/rs6000/linux.h: Ditto.
9563 * alpha/linux.h: Ditto.
9564 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
9565 no_c99_libc_has_function.
9566 * config/c6x/uclinux-elf.h: Ditto.
9567 * config/lm32/uclinux-elf.h: Ditto.
9568 * config/m68k/uclinux.h: Ditto.
9569 * config/moxie/uclinux.h: Ditto.
9570 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
9571 (crisv32-*-linux*, cris-*-linux*): Ditto.
9572 * config/bfin/bfin.c: Include "tm_p.h".
9574 2013-09-05 Richard Biener <rguenther@suse.de>
9576 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
9577 check for a definition without a basic-block.
9579 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9580 Sofiane Naci <sofiane.naci@arm.com>
9582 * config/aarch64/aarch64.md
9583 (*movti_aarch64): Rename r_2_f and f_2_r.
9584 (*movsf_aarch64): Likewise.
9585 (*movdf_aarch64): Likewise.
9586 (*movtf_aarch64): Likewise.
9587 (aarch64_movdi_<mode>low): Likewise.
9588 (aarch64_movdi_<mode>high): Likewise.
9589 (aarch64_mov<mode>high_di): Likewise.
9590 (aarch64_mov<mode>low_di): Likewise.
9591 (aarch64_movtilow_tilow): Likewise.
9592 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
9593 values to config/arm/types.md
9594 (attribute "conds"): Update for attribute change.
9595 (anddi3_insn): Likewise.
9596 (iordi3_insn): Likewise.
9597 (xordi3_insn): Likewise.
9598 (one_cmpldi2): Likewise.
9599 * config/arm/types.md (type): Add Neon types.
9600 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
9601 use "type" attribute.
9602 (movmisalign<mode>_neon_store): Likewise.
9603 (movmisalign<mode>_neon_load): Likewise.
9604 (vec_set<mode>_internal): Likewise.
9605 (vec_setv2di_internal): Likewise.
9606 (vec_extract<mode>): Likewise.
9607 (vec_extractv2di): Likewise.
9608 (add<mode>3_neon): Likewise.
9609 (adddi3_neon): Likewise.
9610 (sub<mode>3_neon): Likewise.
9611 (subdi3_neon): Likewise.
9612 (mul<mode>3_neon): Likewise.
9613 (mul<mode>3add<mode>_neon): Likewise.
9614 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
9615 (fma<VCVTF:mode>4)): Likewise.
9616 (fma<VCVTF:mode>4_intrinsic): Likewise.
9617 (fmsub<VCVTF:mode>4)): Likewise.
9618 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
9619 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
9620 (ior<mode>3): Likewise.
9621 (and<mode>3): Likewise.
9622 (anddi3_neon): Likewise.
9623 (orn<mode>3_neon): Likewise.
9624 (orndi3_neon): Likewise.
9625 (bic<mode>3_neon): Likewise.
9626 (bicdi3_neon): Likewise.
9627 (xor<mode>3): Likewise.
9628 (one_cmpl<mode>2): Likewise.
9629 (abs<mode>2): Likewise.
9630 (neg<mode>2): Likewise.
9631 (umin<mode>3_neon): Likewise.
9632 (umax<mode>3_neon): Likewise.
9633 (smin<mode>3_neon): Likewise.
9634 (smax<mode>3_neon): Likewise.
9635 (vashl<mode>3): Likewise.
9636 (vashr<mode>3_imm): Likewise.
9637 (vlshr<mode>3_imm): Likewise.
9638 (ashl<mode>3_signed): Likewise.
9639 (ashl<mode>3_unsigned): Likewise.
9640 (neon_load_count): Likewise.
9641 (ashldi3_neon_noclobber): Likewise.
9642 (signed_shift_di3_neon): Likewise.
9643 (unsigned_shift_di3_neon): Likewise.
9644 (ashrdi3_neon_imm_noclobber): Likewise.
9645 (lshrdi3_neon_imm_noclobber): Likewise.
9646 (widen_ssum<mode>3): Likewise.
9647 (widen_usum<mode>3): Likewise.
9648 (quad_halves_<code>v4si): Likewise.
9649 (quad_halves_<code>v4sf): Likewise.
9650 (quad_halves_<code>v8hi): Likewise.
9651 (quad_halves_<code>v16qi): Likewise.
9652 (reduc_splus_v2di): Likewise.
9653 (neon_vpadd_internal<mode>): Likewise.
9654 (neon_vpsmin<mode>): Likewise.
9655 (neon_vpsmax<mode>): Likewise.
9656 (neon_vpumin<mode>): Likewise.
9657 (neon_vpumax<mode>): Likewise.
9658 (ss_add<mode>_neon): Likewise.
9659 (us_add<mode>_neon): Likewise.
9660 (ss_sub<mode>_neon): Likewise.
9661 (us_sub<mode>_neon): Likewise.
9662 (neon_vadd<mode>_unspec): Likewise.
9663 (neon_vaddl<mode>): Likewise.
9664 (neon_vaddw<mode>): Likewise.
9665 (neon_vhadd<mode>): Likewise.
9666 (neon_vqadd<mode>): Likewise.
9667 (neon_vaddhn<mode>): Likewise.
9668 (neon_vmul<mode>): Likewise.
9669 (neon_vmla<mode>): Likewise.
9670 (neon_vmlal<mode>): Likewise.
9671 (neon_vmls<mode>): Likewise.
9672 (neon_vmlsl<mode>): Likewise.
9673 (neon_vqdmulh<mode>): Likewise.
9674 (neon_vqdmlal<mode>): Likewise.
9675 (neon_vqdmlsl<mode>): Likewise.
9676 (neon_vmull<mode>): Likewise.
9677 (neon_vqdmull<mode>): Likewise.
9678 (neon_vsub<mode>_unspec): Likewise.
9679 (neon_vsubl<mode>): Likewise.
9680 (neon_vsubw<mode>): Likewise.
9681 (neon_vqsub<mode>): Likewise.
9682 (neon_vhsub<mode>): Likewise.
9683 (neon_vsubhn<mode>): Likewise.
9684 (neon_vceq<mode>): Likewise.
9685 (neon_vcge<mode>): Likewise.
9686 (neon_vcgeu<mode>): Likewise.
9687 (neon_vcgt<mode>): Likewise.
9688 (neon_vcgtu<mode>): Likewise.
9689 (neon_vcle<mode>): Likewise.
9690 (neon_vclt<mode>): Likewise.
9691 (neon_vcage<mode>): Likewise.
9692 (neon_vcagt<mode>): Likewise.
9693 (neon_vtst<mode>): Likewise.
9694 (neon_vabd<mode>): Likewise.
9695 (neon_vabdl<mode>): Likewise.
9696 (neon_vaba<mode>): Likewise.
9697 (neon_vabal<mode>): Likewise.
9698 (neon_vmax<mode>): Likewise.
9699 (neon_vmin<mode>): Likewise.
9700 (neon_vpaddl<mode>): Likewise.
9701 (neon_vpadal<mode>): Likewise.
9702 (neon_vpmax<mode>): Likewise.
9703 (neon_vpmin<mode>): Likewise.
9704 (neon_vrecps<mode>): Likewise.
9705 (neon_vrsqrts<mode>): Likewise.
9706 (neon_vqabs<mode>): Likewise.
9707 (neon_vqneg<mode>): Likewise.
9708 (neon_vcls<mode>): Likewise.
9709 (clz<mode>2): Likewise.
9710 (popcount<mode>2): Likewise.
9711 (neon_vrecpe): Likewise.
9712 (neon_vrsqrte): Likewise.
9713 (neon_vget_lane<mode>_sext_internal): Likewise.
9714 (neon_vget_lane<mode>_zext_internal): Likewise.
9715 (neon_vdup_n<mode>): Likewise.
9716 (neon_vdup_nv2di): Likewise.
9717 (neon_vdpu_lane<mode>_internal): Likewise.
9718 (neon_vswp<mode>): Likewise.
9719 (float<mode><V_cvtto>2): Likewise.
9720 (floatuns<mode><V_cvtto>2): Likewise.
9721 (fix_trunc<mode><V_cvtto>)2): Likewise
9722 (fixuns_trunc<mode><V_cvtto)2): Likewise.
9723 (neon_vcvt<mode>): Likewise.
9724 (neon_vcvtv4sfv4hf): Likewise.
9725 (neon_vcvtv4hfv4sf): Likewise.
9726 (neon_vcvt_n<mode>): Likewise.
9727 (neon_vmovn<mode>): Likewise.
9728 (neon_vqmovn<mode>): Likewise.
9729 (neon_vqmovun<mode>): Likewise.
9730 (neon_vmovl<mode>): Likewise.
9731 (neon_vmul_lane<mode>): Likewise.
9732 (neon_vmull_lane<mode>): Likewise.
9733 (neon_vqdmull_lane<mode>): Likewise.
9734 (neon_vqdmulh_lane<mode>): Likewise.
9735 (neon_vmla_lane<mode>): Likewise.
9736 (neon_vmlal_lane<mode>): Likewise.
9737 (neon_vqdmlal_lane<mode>): Likewise.
9738 (neon_vmls_lane<mode>): Likewise.
9739 (neon_vmlsl_lane<mode>): Likewise.
9740 (neon_vqdmlsl_lane<mode>): Likewise.
9741 (neon_vext<mode>): Likewise.
9742 (neon_vrev64<mode>): Likewise.
9743 (neon_vrev32<mode>): Likewise.
9744 (neon_vrev16<mode>): Likewise.
9745 (neon_vbsl<mode>_internal): Likewise.
9746 (neon_vshl<mode>): Likewise.
9747 (neon_vqshl<mode>): Likewise.
9748 (neon_vshr_n<mode>): Likewise.
9749 (neon_vshrn_n<mode>): Likewise.
9750 (neon_vqshrn_n<mode>): Likewise.
9751 (neon_vqshrun_n<mode>): Likewise.
9752 (neon_vshl_n<mode>): Likewise.
9753 (neon_vqshl_n<mode>): Likewise.
9754 (neon_vqshlu_n<mode>): Likewise.
9755 (neon_vshll_n<mode>): Likewise.
9756 (neon_vsra_n<mode>): Likewise.
9757 (neon_vsri_n<mode>): Likewise.
9758 (neon_vsli_n<mode>): Likewise.
9759 (neon_vtbl1v8qi): Likewise.
9760 (neon_vtbl2v8qi): Likewise.
9761 (neon_vtbl3v8qi): Likewise.
9762 (neon_vtbl4v8qi): Likewise.
9763 (neon_vtbx1v8qi): Likewise.
9764 (neon_vtbx2v8qi): Likewise.
9765 (neon_vtbx3v8qi): Likewise.
9766 (neon_vtbx4v8qi): Likewise.
9767 (neon_vtrn<mode>_internal): Likewise.
9768 (neon_vzip<mode>_internal): Likewise.
9769 (neon_vuzp<mode>_internal): Likewise.
9770 (neon_vld1<mode>): Likewise.
9771 (neon_vld1_lane<mode>): Likewise.
9772 (neon_vld1_dup<mode>): Likewise.
9773 (neon_vld1_dupv2di): Likewise.
9774 (neon_vst1<mode>): Likewise.
9775 (neon_vst1_lane<mode>): Likewise.
9776 (neon_vld2<mode>): Likewise.
9777 (neon_vld2_lane<mode>): Likewise.
9778 (neon_vld2_dup<mode>): Likewise.
9779 (neon_vst2<mode>): Likewise.
9780 (neon_vst2_lane<mode>): Likewise.
9781 (neon_vld3<mode>): Likewise.
9782 (neon_vld3qa<mode>): Likewise.
9783 (neon_vld3qb<mode>): Likewise.
9784 (neon_vld3_lane<mode>): Likewise.
9785 (neon_vld3_dup<mode>): Likewise.
9786 (neon_vst3<mode>): Likewise.
9787 (neon_vst3qa<mode>): Likewise.
9788 (neon_vst3qb<mode>): Likewise.
9789 (neon_vst3_lane<mode>): Likewise.
9790 (neon_vld4<mode>): Likewise.
9791 (neon_vld4qa<mode>): Likewise.
9792 (neon_vld4qb<mode>): Likewise.
9793 (neon_vld4_lane<mode>): Likewise.
9794 (neon_vld4_dup<mode>): Likewise.
9795 (neon_vst4<mode>): Likewise.
9796 (neon_vst4qa<mode>): Likewise.
9797 (neon_vst4qb<mode>): Likewise.
9798 (neon_vst4_lane<mode>): Likewise.
9799 (neon_vec_unpack<US>_lo_<mode>): Likewise.
9800 (neon_vec_unpack<US>_hi_<mode>): Likewise.
9801 (neon_vec_<US>mult_lo_<mode>): Likewise.
9802 (neon_vec_<US>mult_hi_<mode>): Likewise.
9803 (neon_vec_<US>shiftl_<mode>): Likewise.
9804 (neon_unpack<US>_<mode>): Likewise.
9805 (neon_vec_<US>mult_<mode>): Likewise.
9806 (vec_pack_trunc_<mode>): Likewise.
9807 (neon_vec_pack_trunk_<mode>): Likewise.
9808 (neon_vabd<mode>_2): Likewise.
9809 (neon_vabd<mode>_3): Likewise.
9810 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
9811 (thumb2_movsi_vfp): Likewise.
9812 (movdi_vfp): Likewise.
9813 (movdi_vfp_cortexa8): Likewise.
9814 (movhf_vfp_neon): Likewise.
9815 (movhf_vfp): Likewiwse.
9816 (movsf_vfp): Likewiwse.
9817 (thumb2_movsf_vfp): Likewiwse.
9818 (movdf_vfp): Likewise.
9819 (thumb2_movdf_vfp): Likewise.
9820 (movsfcc_vfp): Likewise.
9821 (thumb2_movsfcc_vfp): Likewise.
9822 (movdfcc_vfp): Likewise.
9823 (thumb2_movdfcc_vfp): Likewise.
9824 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
9825 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
9826 (v10_v2c): Likewise.
9827 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
9829 (cortex_a15_neon_int_2): Likewise.
9830 (cortex_a15_neon_int_3): Likewise.
9831 (cortex_a15_neon_int_4): Likewise.
9832 (cortex_a15_neon_int_5): Likewise.
9833 (cortex_a15_neon_vqneg_vqabs): Likewise.
9834 (cortex_a15_neon_vmov): Likewise.
9835 (cortex_a15_neon_vaba): Likewise.
9836 (cortex_a15_neon_vaba_qqq): Likewise.
9837 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9838 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9839 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
9840 scalar_64_32_long_scalar): Likewise.
9841 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9842 (cortex_a15_neon_mla_qqq_8_16): Likewise.
9843 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9844 lotype_qdd_64_32_long): Likewise.
9845 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9846 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9847 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
9848 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9849 (cortex_a15_neon_shift_1): Likewise.
9850 (cortex_a15_neon_shift_2): Likewise.
9851 (cortex_a15_neon_shift_3): Likewise.
9852 (cortex_a15_neon_vshl_ddd): Likewise.
9853 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9854 (cortex_a15_neon_vsra_vrsra): Likewise.
9855 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
9856 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
9857 (cortex_a15_neon_fp_vmul_ddd): Likewise.
9858 (cortex_a15_neon_fp_vmul_qqd): Likewise.
9859 (cortex_a15_neon_fp_vmla_ddd): Likewise.
9860 (cortex_a15_neon_fp_vmla_qqq): Likewise.
9861 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
9862 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
9863 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9864 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9865 (cortex_a15_neon_bp_simple): Likewise.
9866 (cortex_a15_neon_bp_2cycle): Likewise.
9867 (cortex_a15_neon_bp_3cycle): Likewise.
9868 (cortex_a15_neon_vld1_1_2_regs): Likewise.
9869 (cortex_a15_neon_vld1_3_4_regs): Likewise.
9870 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9871 (cortex_a15_neon_vld2_4_regs): Likewise.
9872 (cortex_a15_neon_vld3_vld4): Likewise.
9873 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9874 (cortex_a15_neon_vst1_3_4_regs): Likewise.
9875 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
9876 (cortex_a15_neon_vst3_vst4): Likewise.
9877 (cortex_a15_neon_vld1_vld2_lane): Likewise.
9878 (cortex_a15_neon_vld3_vld4_lane" 10
9879 (cortex_a15_neon_vst1_vst2_lane): Likewise.
9880 (cortex_a15_neon_vst3_vst4_lane): Likewise.
9881 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
9882 (cortex_a15_neon_ldm_2): Likewise.0
9883 (cortex_a15_neon_stm_2): Likewise.
9884 (cortex_a15_neon_mcr): Likewise.
9885 (cortex_a15_neon_mcr_2_mcrr): Likewise.
9886 (cortex_a15_neon_mrc): Likewise.
9887 (cortex_a15_neon_mrrc): Likewise.
9888 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9890 (cortex_a15_alu_shift): Likewise.
9891 (cortex_a15_alu_shift_reg): Likewise.
9892 (cortex_a15_mult32): Likewise.
9893 (cortex_a15_mult64): Likewise.
9894 (cortex_a15_block): Likewise.
9895 (cortex_a15_branch): Likewise.
9896 (cortex_a15_load1): Likewise.
9897 (cortex_a15_load3): Likewise.
9898 (cortex_a15_store1): Likewise.
9899 (cortex_a15_store3): Likewise.
9900 (cortex_a15_call): Likewise.
9901 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
9902 (cortex_a5_f2r): Likewise.
9903 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
9905 (cortex_a53_f2r): Likewise.
9906 * config/arm/cortex-a7.md
9907 (cortex_a7_branch): Update for attribute change.
9908 (cortex_a7_call): Likewise.
9909 (cortex_a7_alu_imm): Likewise.
9910 (cortex_a7_alu_reg): Likewise.
9911 (cortex_a7_alu_shift): Likewise.
9912 (cortex_a7_mul): Likewise.
9913 (cortex_a7_load1): Likewise.
9914 (cortex_a7_store1): Likewise.
9915 (cortex_a7_load2): Likewise.
9916 (cortex_a7_store2): Likewise.
9917 (cortex_a7_load3): Likewise.
9918 (cortex_a7_store3): Likewise.
9919 (cortex_a7_load4): Likewise.
9920 (cortex_a7_store4): Likewise.
9921 (cortex_a7_fpalu): Likewise.
9922 (cortex_a7_fconst): Likewise.
9923 (cortex_a7_fpmuls): Likewise.
9924 (cortex_a7_neon_mul): Likewise.
9925 (cortex_a7_fpmacs): Likewise.
9926 (cortex_a7_neon_mla: Likewise.
9927 (cortex_a7_fpmuld: Likewise.
9928 (cortex_a7_fpmacd: Likewise.
9929 (cortex_a7_fpfmad: Likewise.
9930 (cortex_a7_fdivs: Likewise.
9931 (cortex_a7_fdivd: Likewise.
9932 (cortex_a7_r2f: Likewise.
9933 (cortex_a7_f2r: Likewise.
9934 (cortex_a7_f_flags: Likewise.
9935 (cortex_a7_f_loads: Likewise.
9936 (cortex_a7_f_loadd: Likewise.
9937 (cortex_a7_f_stores: Likewise.
9938 (cortex_a7_f_stored: Likewise.
9939 (cortex_a7_neon): Likewise.
9940 * config/arm/cortex-a8-neon.md
9941 (cortex_a8_neon_mrc): Update for attribute change.
9942 (cortex_a8_neon_mrrc): Likewise.
9943 (cortex_a8_neon_int_1): Likewise.
9944 (cortex_a8_neon_int_2): Likewise.
9945 (cortex_a8_neon_int_3): Likewise.
9946 (cortex_a8_neon_int_4): Likewise.
9947 (cortex_a8_neon_int_5): Likewise.
9948 (cortex_a8_neon_vqneg_vqabs): Likewise.
9949 (cortex_a8_neon_vmov): Likewise.
9950 (cortex_a8_neon_vaba): Likewise.
9951 (cortex_a8_neon_vaba_qqq): Likewise.
9952 (cortex_a8_neon_vsma): Likewise.
9953 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9954 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9955 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
9956 long_scalar): Likewise.
9957 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9958 (cortex_a8_neon_mla_qqq_8_16): Likewise.
9959 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9960 long_scalar_qdd_64_32_long): Likewise.
9961 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9962 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9963 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
9964 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9965 (cortex_a8_neon_shift_1): Likewise.
9966 (cortex_a8_neon_shift_2): Likewise.
9967 (cortex_a8_neon_shift_3): Likewise.
9968 (cortex_a8_neon_vshl_ddd): Likewise.
9969 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9970 (cortex_a8_neon_vsra_vrsra): Likewise.
9971 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
9972 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
9973 (cortex_a8_neon_fp_vsum): Likewise.
9974 (cortex_a8_neon_fp_vmul_ddd): Likewise.
9975 (cortex_a8_neon_fp_vmul_qqd): Likewise.
9976 (cortex_a8_neon_fp_vmla_ddd): Likewise.
9977 (cortex_a8_neon_fp_vmla_qqq): Likewise.
9978 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
9979 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
9980 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9981 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9982 (cortex_a8_neon_bp_simple): Likewise.
9983 (cortex_a8_neon_bp_2cycle): Likewise.
9984 (cortex_a8_neon_bp_3cycle): Likewise.
9985 (cortex_a8_neon_ldr): Likewise.
9986 (cortex_a8_neon_str): Likewise.
9987 (cortex_a8_neon_vld1_1_2_regs): Likewise.
9988 (cortex_a8_neon_vld1_3_4_regs): Likewise.
9989 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9990 (cortex_a8_neon_vld2_4_regs): Likewise.
9991 (cortex_a8_neon_vld3_vld4): Likewise.
9992 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9993 (cortex_a8_neon_vst1_3_4_regs): Likewise.
9994 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
9995 (cortex_a8_neon_vst3_vst4): Likewise.
9996 (cortex_a8_neon_vld1_vld2_lane): Likewise.
9997 (cortex_a8_neon_vld3_vld4_lane): Likewise.
9998 (cortex_a8_neon_vst1_vst2_lane): Likewise.
9999 (cortex_a8_neon_vst3_vst4_lane): Likewise.
10000 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
10001 (cortex_a8_neon_mcr): Likewise.
10002 (cortex_a8_neon_mcr_2_mcrr): Likewise.
10003 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10004 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
10006 (ca9_neon_mrrc): Likewise.
10007 (cortex_a9_neon_int_1): Likewise.
10008 (cortex_a9_neon_int_2): Likewise.
10009 (cortex_a9_neon_int_3): Likewise.
10010 (cortex_a9_neon_int_4): Likewise.
10011 (cortex_a9_neon_int_5): Likewise.
10012 (cortex_a9_neon_vqneg_vqabs): Likewise.
10013 (cortex_a9_neon_vmov): Likewise.
10014 (cortex_a9_neon_vaba): Likewise.
10015 (cortex_a9_neon_vaba_qqq): Likewise.
10016 (cortex_a9_neon_vsma): Likewise.
10017 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10018 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10019 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10020 long_scalar): Likewise.
10021 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10022 (cortex_a9_neon_mla_qqq_8_16): Likewise.
10023 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10024 long_scalar_qdd_64_32_long): Likewise.
10025 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10026 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10027 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
10028 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10029 (cortex_a9_neon_shift_1): Likewise.
10030 (cortex_a9_neon_shift_2): Likewise.
10031 (cortex_a9_neon_shift_3): Likewise.
10032 (cortex_a9_neon_vshl_ddd): Likewise.
10033 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10034 (cortex_a9_neon_vsra_vrsra): Likewise.
10035 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
10036 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
10037 (cortex_a9_neon_fp_vsum): Likewise.
10038 (cortex_a9_neon_fp_vmul_ddd): Likewise.
10039 (cortex_a9_neon_fp_vmul_qqd): Likewise.
10040 (cortex_a9_neon_fp_vmla_ddd): Likewise.
10041 (cortex_a9_neon_fp_vmla_qqq): Likewise.
10042 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
10043 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
10044 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10045 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10046 (cortex_a9_neon_bp_simple): Likewise.
10047 (cortex_a9_neon_bp_2cycle): Likewise.
10048 (cortex_a9_neon_bp_3cycle): Likewise.
10049 (cortex_a9_neon_ldr): Likewise.
10050 (cortex_a9_neon_str): Likewise.
10051 (cortex_a9_neon_vld1_1_2_regs): Likewise.
10052 (cortex_a9_neon_vld1_3_4_regs): Likewise.
10053 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10054 (cortex_a9_neon_vld2_4_regs): Likewise.
10055 (cortex_a9_neon_vld3_vld4): Likewise.
10056 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10057 (cortex_a9_neon_vst1_3_4_regs): Likewise.
10058 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
10059 (cortex_a9_neon_vst3_vst4): Likewise.
10060 (cortex_a9_neon_vld1_vld2_lane): Likewise.
10061 (cortex_a9_neon_vld3_vld4_lane): Likewise.
10062 (cortex_a9_neon_vst1_vst2_lane): Likewise.
10063 (cortex_a9_neon_vst3_vst4_lane): Likewise.
10064 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
10065 (cortex_a9_neon_mcr): Likewise.
10066 (cortex_a9_neon_mcr_2_mcrr): Likewise.
10067 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
10068 (cortex_a9_fps): Likewise.
10069 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
10071 (cortex_m4_fmuls): Likewise.
10072 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
10074 (cortex_r4_mrc): Likewise.
10075 * config/arm/iterators.md: Update comment referring to neon_type.
10076 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
10077 (iwmmxt_movsi_insn): Likewise.
10078 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
10080 (pj4_core_to_vfp): Likewise.
10081 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
10083 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
10084 (vfp_fstore): Likewise.
10085 * doc/md.texi: Change references to neon_type to refer to type.
10087 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10089 * tree.h (DECL_BUILT_IN): Fix typo in comment.
10091 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
10093 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
10094 lglobl if not weak.
10096 2013-09-04 Easwaran Raman <eraman@google.com>
10098 PR middle-end/57370
10099 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
10100 (build_and_add_sum): Use it.
10101 (appears_later_in_bb): Simplify code.
10103 2013-09-04 Teresa Johnson <tejohnson@google.com>
10105 * dumpfile.c (dump_finish): Don't close stderr/stdout.
10107 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
10109 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
10111 2013-09-04 Jan Hubicka <jh@suse.cz>
10113 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
10114 * ipa-devirt.c: Include diganostic.h
10115 (odr_type_d): Add types and types_set.
10116 (hash_type_name): Work for types with vtables during LTO.
10117 (odr_hasher::remove): Fix comment; destroy types_set.
10118 (add_type_duplicate): New function,
10119 (get_odr_type): Use it.
10120 (dump_type_inheritance_graph): Dump type duplicates.
10121 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
10123 * tree.c (types_same_for_odr): Give exact answers on types with
10126 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10128 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
10129 explaining their differences.
10131 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
10133 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
10135 2013-09-03 Jeff Law <law@redhat.com>
10137 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
10138 when not threading through a joiner block. Pass joiner/no joiner
10139 state to register_jump_thread.
10140 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
10141 state from argument rather than implying on path length.
10142 Dump the entire jump thread path into debugging dump.
10143 * tree-flow.h (register_jump_thread): Update prototype.
10145 2013-08-29 Xinliang David Li <davidxl@google.com>
10147 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10148 Remove a trivial gcc_assert.
10150 2013-08-29 Xinliang David Li <davidxl@google.com>
10152 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
10153 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
10154 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10155 Delay base decl alignment adjustment.
10156 * tree-vectorizer.c (vect_destroy_datarefs): New function.
10157 * tree-vectorizer.h: New data structure.
10158 (set_dr_misalignment): New function.
10159 (dr_misalignment): Ditto.
10160 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
10161 (vectorizable_load): Ditto.
10162 (ensure_base_align): New function.
10163 (vectorize_loops): Add dbg_cnt support.
10164 (execute_vect_slp): Ditto.
10165 * dbgcnt.def: New debug counter.
10166 * Makefile: New dependency.
10168 2013-09-03 Meador Inge <meadori@codesourcery.com>
10172 2013-08-30 Meador Inge <meadori@codesourcery.com>
10174 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10176 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
10178 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
10179 function descriptor.
10181 2013-09-03 Richard Biener <rguenther@suse.de>
10183 * tree-affine.c (add_elt_to_tree): Fix association issue,
10184 avoid useless converts and make sure to always return a
10185 properly typed result.
10187 2013-09-03 Richard Biener <rguenther@suse.de>
10189 PR middle-end/57656
10190 * fold-const.c (negate_expr_p): Fix division case.
10191 (negate_expr): Likewise.
10193 2013-09-03 Richard Biener <rguenther@suse.de>
10196 * tree-streamer-out.c: Include tm.h.
10197 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
10199 2013-09-03 Jan Hubicka <jh@suse.cz>
10201 * tree-profile.c (tree_profiling): Cleanup CFG when done.
10203 2013-09-03 Alan Modra <amodra@gmail.com>
10205 * config.gcc (powerpc*-*-linux*): Add support for little-endian
10206 multilibs to big-endian target and vice versa.
10207 * config/rs6000/t-linux64: Use := assignment on all vars.
10208 (MULTILIB_EXTRA_OPTS): Remove fPIC.
10209 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
10210 * config/rs6000/t-linux64le: New file.
10211 * config/rs6000/t-linux64bele: New file.
10212 * config/rs6000/t-linux64lebe: New file.
10214 2013-09-02 Jan Hubicka <jh@suse.cz>
10216 * ipa-inline-transform.c (inline_transform): Do not
10217 optimize_inline_calls when not optimizing.
10219 2013-09-02 Jan Hubicka <jh@suse.cz>
10221 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
10222 duplicated nodes for assembler names.
10223 * symtab.c (symtab_unregister_node): Do not attempt to unlink
10224 hard registers from assembler name hash.
10226 2013-09-02 Jan Hubicka <jh@suse.cz>
10228 * ipa-split.c (execute_split_functions): Split externally visible
10229 functions called once.
10231 2013-09-02 Martin Jambor <mjambor@suse.cz>
10234 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
10235 linked list. When finding the correct duplicate, also consider also
10236 the caller in additon to its inlined_to node.
10238 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
10240 * config/aarch64/aarch64-simd-builtins.def
10241 (dup_lane_scalar): Remove.
10242 * config/aarch64/aarch64-simd.md
10243 (aarch64_simd_dup): Add 'w->w' alternative.
10244 (aarch64_dup_lane<mode>): Allow for VALL.
10245 (aarch64_dup_lane_scalar<mode>): Remove.
10246 (aarch64_dup_lane_<vswap_width_name><mode>): New.
10247 (aarch64_get_lane_signed<mode>): Add w->w altenative.
10248 (aarch64_get_lane_unsigned<mode>): Likewise.
10249 (aarch64_get_lane<mode>): Likewise.
10250 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
10251 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
10252 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
10253 (VCON): Change container of V2SF.
10254 (vswap_width_name): Likewise.
10255 * config/aarch64/arm_neon.h
10256 (__aarch64_vdup_lane_any): New.
10257 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10258 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
10259 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10261 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
10263 PR middle-end/56382
10264 * expr.c (emit_move_complex): Do not move complex FP values as parts if
10265 the source or the destination is a single hard register.
10267 2013-09-02 Richard Biener <rguenther@suse.de>
10269 PR middle-end/57511
10270 * tree-scalar-evolution.c (instantiate_scev_name): Allow
10273 2013-09-02 Richard Biener <rguenther@suse.de>
10275 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
10276 arithmetic to sizetype.
10278 2013-09-02 Bin Cheng <bin.cheng@arm.com>
10280 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
10281 Find auto-increment use both before and after candidate.
10283 2013-09-02 Marek Polacek <polacek@redhat.com>
10285 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
10287 2013-09-01 Jan Hubicka <jh@suse.cz>
10289 * Makefile.in: Add ipa-profile.o
10290 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
10291 * cgraph.c (struct cgraph_propagate_frequency_data,
10292 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
10293 ipa-profile.c; replace cgraph_ by ipa_ prefix.
10294 * cgraph.h (cgraph_propagate_frequency): Remove.
10295 * ipa-inline-analysis.c: Include ipa-utils.h;
10296 drop duplicated cfgloop.h.
10297 (inline_update_callee_summaries): Update.
10298 * ipa-profile.c: New file.
10299 * ipa-utils.h (ipa_propagate_frequency): Declare.
10300 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
10301 data-streamer.h, value-prof.h.
10302 (symtab_remove_unreachable_nodes): Update profile.
10303 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
10304 account_time_size, cmp_counts, dump_histogram,
10305 ipa_profile_generate_summary, ipa_profile_write_summary,
10306 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
10307 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
10308 Move to ipa-profile.c.
10310 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
10312 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
10314 2013-09-01 Jan Hubicka <jh@suse.cz>
10316 * common.opt (fdevirtualize-speculatively): New function.
10317 * invoke.texi (fdevirtualize-speculatively): Document.
10318 * ipa-devirt.c: Include ipa-inline.h
10319 (likely_target_p): New function.
10320 (ipa_devirt): New function.
10321 (gate_ipa_devirt): New function.
10322 (pass_data_ipa_devirt): New static var.
10323 (pass_ipa_devirt): Likewise.
10324 (make_pass_ipa_devirt): New function.
10325 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
10326 (common_handle_option): Disable devirtualization when
10327 value range profiling is available.
10328 * passes.def (pass_ipa_devirt): Add.
10329 * timever.def (TV_IPA_DEVIRT): New timevar.
10330 * tree-pass.h (make_pass_ipa_devirt):
10332 2013-09-01 Iain Sandoe <iain@codesourcery.com>
10334 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
10335 include sanitize(undefined).
10337 2013-08-31 Diego Novillo <dnovillo@google.com>
10339 * Makefile.in (TREE_CORE_H): Define.
10341 (GTFILES): Add tree-core.h.
10342 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
10344 * tree-core.h: New file.
10345 Move all data structures, enum, typedefs, global
10346 declarations and constants from ...
10347 * tree.h: ... here.
10349 2013-08-31 Jan Hubicka <jh@suse.cz>
10351 * bulitins.c (expand_builtin): Do not early exit for gcov
10352 instrumented functions.
10354 2013-08-31 Marek Polacek <polacek@redhat.com>
10356 * ubsan.c: Include tm_p.h.
10358 2013-08-31 Jan Hubicka <jh@suse.cz>
10360 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
10363 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
10364 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
10365 * tree-cfg.c (verify_gimple_label): ... here.
10366 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
10367 (ipa_merge_profiles): New function.
10368 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
10369 (lto_input_function_body): Likewise.
10370 * ipa-utils.h (ipa_merge_profiles): Declare.
10371 * lto-streamer.h (lto_input_function_body): Update prototype.
10372 (emit_label_in_global_context_p): Remove.
10373 * lto-symtab.c: Include ipa-utils.h
10374 (lto_cgraph_replace_node): Use ipa_merge_profiles.
10376 2013-08-31 Jan Hubicka <jh@suse.cz>
10378 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
10380 2013-08-31 Jan Hubicka <jh@suse.cz>
10382 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
10384 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
10386 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
10387 "cmp" RTX before signed_comparison_operator check to account
10388 for "code" changes.
10390 2013-08-30 Jan Hubicka <jh@suse.cz>
10392 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
10393 (detect_type_change_1): Rename to ...
10394 (detect_type_change): ... this one; early return on non-polymorphic
10396 (detect_type_change_ssa): Add comp_type parameter; update
10397 use of detect_type_change.
10398 (compute_complex_assign_jump_func): Add param_type parameter;
10399 update use of detect_type_change_ssa.
10400 (compute_complex_ancestor_jump_func): Likewise.
10401 (ipa_get_callee_param_type): New function.
10402 (ipa_compute_jump_functions_for_edge): Compute parameter type;
10403 update calls to the jump function computation functions.
10405 2013-08-30 Teresa Johnson <tejohnson@google.com>
10406 Steven Bosscher <steven@gcc.gnu.org>
10408 * cfgrtl.c (fixup_new_cold_bb): New routine.
10409 (commit_edge_insertions): Invoke fixup_partitions.
10410 (find_partition_fixes): New routine.
10411 (fixup_partitions): Ditto.
10412 (verify_hot_cold_block_grouping): Update comments.
10413 (rtl_verify_edges): Invoke find_partition_fixes.
10414 (rtl_verify_bb_pointers): Update comments.
10415 (rtl_verify_bb_layout): Ditto.
10416 * basic-block.h (probably_never_executed_edge_p): Declare.
10417 (fixup_partitions): Ditto.
10418 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
10419 * bb-reorder.c (sanitize_hot_paths): New function.
10420 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
10421 sanitize_hot_paths.
10422 * predict.c (probably_never_executed_edge_p): New routine.
10423 * cfg.c (check_bb_profile): Add partition insanity warnings.
10425 2013-08-30 Meador Inge <meadori@codesourcery.com>
10427 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10429 2013-08-30 Marek Polacek <polacek@redhat.com>
10431 * Makefile.in (ubsan.o): Add.
10432 (c-family/c-ubsan.o): Add.
10433 (builtins.o): Add ubsan.h dependency.
10434 * ubsan.h: New file.
10435 * ubsan.c: New file.
10436 * common.opt: Add -fsanitize=undefined option.
10437 (flag_sanitize): Add variable.
10438 (fsanitize=): Add option. Add Driver.
10439 (fsanitize=thread): Remove option.
10440 (fsanitize=address): Likewise.
10441 (static-libubsan): New option.
10442 * doc/invoke.texi: Document the new flag and -static-libubsan.
10443 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
10444 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
10445 * builtin-attrs.def (ATTR_COLD): Define.
10446 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10447 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
10448 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
10449 * flag-types.h (sanitize_code): New enum.
10450 * opts.c (common_handle_option): Parse command line arguments
10451 of -fsanitize=. Add -fsanitize=unreachable option.
10452 * varasm.c (get_variable_section): Adjust.
10453 (assemble_noswitch_variable): Likewise.
10454 (assemble_variable): Likewise.
10455 (output_constant_def_contents): Likewise.
10456 (categorize_decl_for_section): Likewise.
10457 (place_block_symbol): Likewise.
10458 (output_object_block): Likewise.
10459 * builtins.def: Likewise.
10460 * toplev.c (compile_file): Likewise.
10461 (process_options): Likewise.
10462 * cppbuiltin.c: Likewise.
10463 * tsan.c (tsan_pass): Likewise.
10464 (tsan_gate): Likewise.
10465 (tsan_gate_O0): Likewise.
10466 * cfgexpand.c (partition_stack_vars): Likewise.
10467 (expand_stack_vars): Likewise.
10468 (defer_stack_allocation): Likewise.
10469 (expand_used_vars): Likewise.
10470 * cfgcleanup.c (old_insns_match_p): Likewise.
10471 * asan.c (asan_finish_file): Likewise.
10472 (asan_instrument): Likewise.
10473 (gate_asan): Likewise.
10474 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
10475 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10476 (asan_global_struct): Use pointer_sized_int_node instead
10477 calling build_nonstandard_integer_type.
10478 (initialize_sanitizer_builtins): Likewise.
10479 (asan_finish_file): Likewise.
10480 * gcc.c: Document %{%:function(args):X}.
10481 (static_spec_functions): Add sanitize.
10482 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
10483 store funcval != NULL there.
10484 (do_spec_1): Adjust handle_spec_function caller.
10485 (handle_braces): Allow %:function(args) as condition.
10486 (sanitize_spec_function): New function.
10487 (ADD_STATIC_LIBUBSAN_LIBS): Define.
10488 (LIBUBSAN_SPEC): Likewise.
10489 (LIBUBSAN_EARLY_SPEC): Likewise.
10490 (SANITIZER_SPEC): Handle libubsan.
10491 (SANITIZER_EARLY_SPEC): Likewise.
10492 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
10493 instead of fsanitize=address.
10494 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
10495 instead of fsanitize=address*.
10496 * builtins.c: Include ubsan.h.
10497 (fold_builtin_0): Instrument __builtin_unreachable.
10498 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
10499 instead of flag_asan.
10500 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
10501 (pointer_sized_int_node): Define.
10502 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
10504 2013-08-30 Mike Stump <mikestump@comcast.net>
10506 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
10509 2013-08-29 Jan Hubicka <jh@suse.cz>
10511 * cgraph.c (cgraph_function_body_availability): Handle weakref
10513 * passes.def: Remove pass_fixup_cfg.
10514 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
10515 track when we need to remove functions.
10516 (gate_ipa_inline): Execute inlining always; add comment why.
10517 (pass_data_ipa_inline): Remove TODO_remove_functions.
10518 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
10519 do not produce summaries.
10520 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
10521 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
10522 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
10523 constant pool and vtable.
10525 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
10527 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
10528 New arm_neon.h's internal macros to specify 64-bit constants.
10529 Avoid using stdint.h's macros.
10531 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
10533 * recog.c (verify_changes): Verify that changes[i].old is non-zero
10534 before applying REG_P.
10536 2013-08-30 Jakub Jelinek <jakub@redhat.com>
10538 PR tree-optimization/58277
10539 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
10540 after seeing too many stmts with vdef in between dombb and current
10541 bb, invalidate everything.
10543 2013-08-30 Richard Biener <rguenther@suse.de>
10545 * fold-const.c (fold_single_bit_test): Fix overflow test.
10547 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
10549 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
10550 and which can live in a register, always retrieve the value on entry.
10551 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
10552 passed by invisible reference specially.
10553 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
10554 (vt_add_function_parameter): Correctly deal with a parameter passed by
10555 invisible reference.
10557 2013-08-30 Jan Hubicka <jh@suse.cz>
10559 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
10561 2013-08-30 Richard Biener <rguenther@suse.de>
10563 PR tree-optimization/58228
10564 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
10565 allow invariant loads in nested loop vectorization.
10567 2013-08-30 Richard Biener <rguenther@suse.de>
10569 PR tree-optimization/58223
10570 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
10571 (has_anti_or_output_dependence): ... this and adjust to also
10572 look for output dependences.
10573 (mark_nodes_having_upstream_mem_writes): Adjust.
10574 (rdg_flag_uses): Likewise.
10576 2013-08-30 Richard Biener <rguenther@suse.de>
10578 PR tree-optimization/58010
10579 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
10580 assert that we have a loop-closed PHI.
10582 2013-08-29 Jan Hubicka <jh@suse.cz>
10584 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
10585 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
10586 * lto-section-in.c (lto_free_function_in_decl_state): New function.
10587 (lto_free_function_in_decl_state_for_node): New function.
10589 2013-08-29 Xinliang David Li <davidxl@google.com>
10591 * loop-unroll.c (report_unroll_peel): Minor message change.
10592 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
10593 Emit alignment peeling message with default -fopt-info.
10594 (vect_loop_versioning): Emit loop version info message.
10595 * tree-vectorizer.c (vectorize_loops): Minor message change.
10596 (execute_vect_slp): Ditto.
10598 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
10600 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
10601 of the clone from the DECL_NAME of the original function.
10603 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
10605 * passes.c (register_pass): Add overload.
10606 * tree-pass.h (register_pass): Forward declare it. Add comment.
10608 2013-08-29 Jan Hubicka <jh@suse.cz>
10610 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
10611 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
10612 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
10613 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10614 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
10615 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
10616 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10617 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
10619 2013-08-29 Teresa Johnson <tejohnson@google.com>
10621 * dumpfile.c (dump_loc): Output column number.
10622 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
10623 * doc/invoke.texi: Document optall -fopt-info flag.
10624 * profile.c (read_profile_edge_counts): Use new dump framework.
10625 (compute_branch_probabilities): Ditto.
10626 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
10627 when pass not in any opt group.
10628 * pass_manager.h (pass_manager::get_pass_profile): New method.
10629 * value-prof.c (check_counter): Use new dump framework.
10630 (check_ic_target): Ditto.
10631 * coverage.c (get_coverage_counts): Ditto.
10632 (coverage_init): Setup new dump framework.
10634 2013-08-29 Richard Biener <rguenther@suse.de>
10636 PR tree-optimization/58246
10637 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
10638 handle the dominance check inside a basic-block.
10640 2013-08-29 Richard Biener <rguenther@suse.de>
10642 PR middle-end/57287
10643 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
10644 of default defs that appear in abnormal PHI nodes.
10646 2013-08-29 Richard Biener <rguenther@suse.de>
10648 PR tree-optimization/57685
10649 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
10650 single-use operands to avoid exponential complexity.
10652 2013-08-28 Dehao Chen <dehao@google.com>
10654 * ipa-inline.c (edge_badness): Fix integer underflow.
10656 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
10658 * gtm-builtins.def (_ITM_free): Declare leaf.
10660 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10663 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
10664 (*tls_local_dynamic_base_64_largepic): Likewise.
10665 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
10666 Remove predicate from call operand.
10667 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
10668 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
10670 2013-08-28 Jeff Law <law@redhat.com>
10672 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
10673 checks for the number of predecessors and successors allowed.
10674 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
10675 which require copying a joiner block if there is a request which
10676 is a subpath that requires no joiner block copying.
10678 2013-08-28 Jan Hubicka <jh@suse.cz>
10680 * lto-streamer-out.c (DFS_write_tree_body): Drop
10681 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
10682 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
10683 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
10684 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
10685 Do not read DECL_ERROR_ISSUED.
10686 (unpack_ts_decl_with_vis_value_fields): Do not read
10688 (lto_input_ts_binfo_tree_pointers): Do not read
10689 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
10690 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
10691 write DECL_ERROR_ISSUED..
10692 (pack_ts_decl_with_vis_value_fields): Do not write
10694 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
10695 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
10696 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
10697 * tree.h (tree_decl_common): Update comment.
10698 (DECL_ERROR_ISSUED): Remove.
10700 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10702 PR middle-end/58257
10703 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
10705 2013-08-28 Jan Hubicka <jh@suse.cz>
10707 * builtins.def (free): Declare leaf.
10709 2013-08-27 David Malcolm <dmalcolm@redhat.com>
10711 * gdbhooks.py: New.
10712 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
10713 * configure: Regenerate.
10715 2013-08-27 Martin Jambor <mjambor@suse.cz>
10717 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
10718 (ipa_ancestor_jf_data): Likewise.
10719 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
10720 (ipa_get_jf_pass_through_type_preserved): New function.
10721 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
10722 (ipa_get_jf_ancestor_type_preserved): New function.
10723 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
10724 (ipa_get_jf_ancestor_result): Likewise.
10725 (propagate_vals_accross_pass_through): Use
10726 ipa_get_jf_pass_through_result to do all the value mappings.
10727 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
10728 type_preserved flag.
10729 (ipa_set_jf_cst_copy): New function.
10730 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
10731 (ipa_set_jf_arith_pass_through): Likewise.
10732 (ipa_set_ancestor_jf): Likewise.
10733 (compute_complex_assign_jump_func): Set type_preserved instead of
10735 (ipa_compute_jump_functions_for_edge): Likewise.
10736 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
10737 (update_jump_functions_after_inlining): Update type_preserved.
10738 Explicitely create jump functions when combining one with pass_through.
10739 (ipa_write_jump_function): Stream the type_preserved flags.
10740 (ipa_read_jump_function): Likewise.
10742 2013-08-27 Jakub Jelinek <jakub@redhat.com>
10743 Aldy Hernandez <aldyh@redhat.com>
10745 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
10746 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
10747 * function.h (struct function): Add has_force_vect_loops and
10749 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
10750 * gimple.c (gimple_build_omp_critical): Add KIND argument and
10752 * gimple.def: Update CLAUSES comments.
10753 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
10754 (gimple_build_omp_for): Add argument to prototype.
10755 (gimple_omp_for_kind): New.
10756 (gimple_omp_for_set_kind): New.
10757 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
10758 GOVD_DATA_SHARE_CLASS.
10759 (enum omp_region_type): Add ORT_SIMD.
10760 (gimple_add_tmp_var): Handle ORT_SIMD.
10761 (gimplify_var_or_parm_decl): Same.
10762 (is_gimple_stmt): Same.
10763 (omp_firstprivatize_variable): Same.
10764 (omp_add_variable): Only use splay_tree_insert if lookup failed.
10765 (omp_notice_variable): Handle ORT_SIMD.
10766 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
10767 (omp_check_private): Handle ORT_SIMD.
10768 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10769 OMP_CLAUSE_SAFELEN.
10770 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
10771 Handle OMP_CLAUSE_LASTPRIVATE.
10772 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10773 OMP_CLAUSE_SAFELEN.
10774 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
10775 (gimplify_expr): Handle OMP_SIMD.
10776 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
10777 (expand_GOMP_SIMD_VF): New.
10778 (expand_GOMP_SIMD_LAST_LANE): New.
10779 * internal-fn.def (GOMP_SIMD_LANE): New.
10780 (GOMP_SIMD_VF): New.
10781 (GOMP_SIMD_LAST_LANE): New.
10782 * omp-low.c: Include target.h.
10783 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
10784 OMP_CLAUSE_SAFELEN.
10785 (check_omp_nesting_restrictions): Same.
10787 (lower_rec_simd_input_clauses): New.
10788 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
10790 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
10791 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
10792 (expand_omp_build_assign): New.
10793 (expand_omp_for_init_counts): New.
10794 (expand_omp_for_init_vars): New.
10795 (extract_omp_for_update_vars): New.
10796 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
10797 and rewrite accordingly.
10798 (expand_omp_simd): New.
10799 (expand_omp_for): Use expand_omp_simd.
10800 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
10801 (lower_omp_for): Do not lower the body.
10802 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
10803 in their own loops.
10804 * tree-flow.h (find_omp_clause): Remove prototype.
10805 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
10806 forcing vectorization of the loop, or if flag_tree_vectorize.
10807 (gate_tree_if_conversion): Similarly.
10808 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
10809 gimple_build_omp_for.
10810 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
10811 * tree-parloops (create_parallel_loop): Pass kind argument to
10812 gimple_build_omp_for.
10813 * tree-pretty-print.c (dump_omp_clause): Add cases for
10814 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
10815 OMP_CLAUSE__SIMDUID_.
10816 (dump_generic_node): Handle OMP_SIMD.
10817 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
10818 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
10819 unroll OMP_SIMD loops here.
10820 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
10821 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
10823 (vect_analyze_data_refs): Handle simd loops.
10824 * tree-vect-loop.c (vectorizable_live_operation): Handle
10826 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
10827 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
10828 (vectorizable_load): Same.
10829 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
10830 (struct simduid_to_vf): New.
10831 (simduid_to_vf::hash): New.
10832 (simduid_to-vf::equal): New.
10833 (struct simd_array_to_simduid): New.
10834 (simd_array_to_simduid::hash): New.
10835 (simd_array_to_simduid::equal): New.
10836 (adjust_simduid_builtins): New.
10837 (struct note_simd_array_uses_struct): New.
10838 (note_simd_array_uses_cb): New.
10839 (note_simd_array_uses): New.
10840 (vectorize_loops): Handle simd hints and adjust simd builtins
10842 * tree-vectorizer.h (struct _stmt_vec_info): Add
10843 simd_lane_access_p field.
10844 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
10845 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
10846 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
10847 (omp_clause_code_name): Same.
10848 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
10849 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
10850 * tree.def (OMP_SIMD): New entry.
10851 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
10852 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
10853 (OMP_CLAUSE_DECL): Adjust range for new clauses.
10854 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
10855 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
10856 (OMP_CLAUSE_LINEAR_STEP): New.
10857 (OMP_CLAUSE_SAFELEN_EXPR): New.
10858 (OMP_CLAUSE__SIMDUID__DECL): New.
10859 (find_omp_clause): New prototype.
10861 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
10863 * config/i386/driver-i386.c (host_detect_local_cpu): Update
10864 Haswell processor detection.
10866 2013-08-27 Christian Widmer <shadow@umbrox.de>
10869 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
10870 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
10871 AVX2 capable processors.
10873 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
10875 * config/aarch64/arm_neon.h: Replace all inline asm implementations
10876 of vget_low_* with implementations in terms of other intrinsics.
10878 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
10880 PR middle-end/57219
10881 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
10882 values to -1, 0 and 1.
10884 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
10886 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
10887 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
10888 (<optab><mode>3_insn): Remove.
10889 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
10890 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
10891 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
10892 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
10893 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
10894 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
10895 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
10896 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
10897 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
10898 (ror<mode>3_insn): Likewise.
10899 * config/aarch64/predicates.md (aarch64_simd_register): New.
10901 2013-08-27 Richard Biener <rguenther@suse.de>
10903 PR tree-optimization/57521
10904 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
10905 one edge is non-critical.
10906 (find_phi_replacement_condition): Make sure to use a non-critical
10907 edge. Cleanup and remove old bug workarounds.
10908 (bb_postdominates_preds): Remove.
10909 (if_convertible_loop_p_1): Do not compute post-dominators.
10910 (combine_blocks): Do not free post-dominators.
10911 (main_tree_if_conversion): Likewise.
10912 (pass_data_if_conversion): Add TODO_verify_ssa.
10914 2013-08-27 DJ Delorie <dj@redhat.com>
10916 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
10918 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
10920 * function.c (assign_parm_find_data_types): Set passed_mode and
10921 nominal_mode to the TYPE_MODE of nominal_type for the built
10922 pointer type in case of the struct-pass-by-reference.
10924 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10926 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
10927 (UINT16_TYPE): Change default to "unsigned int".
10929 * config/avr/avr.opt (mfract-convert-truncate): New option.
10930 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
10931 is set, round negative fractional integers according to n1169
10932 when converting to integer types.
10934 2013-08-26 Jan Hubicka <jh@suse.cz>
10936 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
10937 methods can not be called indirectly when their address is not taken.
10939 2013-08-26 Jan Hubicka <jh@suse.cz>
10941 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
10944 2013-08-26 Jan Hubicka <jh@suse.cz>
10946 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
10949 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10951 * reload.c (find_valid_class): Allow classes that do not include
10952 FIRST_PSEUDO_REGISTER - 1.
10954 2013-08-26 Jan Hubicka <jh@suse.cz>
10956 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
10957 fix edge count/frequency when speculation failed; fix type check
10958 for the direct call.
10960 2013-08-26 Jan Hubicka <jh@suse.cz>
10962 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
10964 2013-08-26 Jan Hubicka <jh@suse.cz>
10966 * ipa-inline-transform.c (inline_transform): Be ready for basic block
10967 to be changed by edge redirection.
10969 2013-08-26 Jan Hubicka <jh@suse.cz>
10971 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
10972 formating; add sanity check.
10973 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
10974 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
10975 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
10976 (resolve_noninline_speculation): Update callee keys, too.
10978 2013-08-26 Jan Hubicka <jh@suse.cz>
10980 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
10981 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
10983 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10985 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
10988 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
10990 * config/i386/i386.c (ix86_debug_options): Remove prototype.
10991 (x86_64_elf_select_section): Ditto.
10992 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
10994 (ix86_pass_by_reference): Ditto.
10995 (output_set_got): Ditto.
10996 (ix86_unary_operator_ok): Ditto.
10997 (ix86_expand_builtin): Ditto.
10999 2013-08-23 Jan Hubicka <jh@suse.cz>
11001 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
11003 2013-08-23 Jan Hubicka <jh@suse.cz>
11005 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
11006 (tree_decl_with_vis): Add FINAL field.
11008 2013-08-23 Jeff Law <law@redhat.com>
11010 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
11011 the debugging dump when the expression is fully redundant.
11013 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11015 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
11016 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
11017 * pretty-print.c (pp_formatted_text_data): Likewise.
11018 (pp_write_text_to_stream): Likewise.
11019 (pp_write_text_as_dot_label_to_stream): Likewise.
11020 (pp_append_r): Likewise.
11021 (pp_format): Likewise.
11022 (pp_flush): Likewise.
11023 (pp_clear_output_area): Likewise.
11024 (pp_append_text): Likewise.
11025 (pp_formatted_text): Likewise.
11026 (pp_remaining_character_count_for_line): Likewise.
11027 (pp_newline): Likewise.
11028 (pp_character): Likewise.
11029 (output_buffer::~output_buffer): Define.
11030 (pretty_printer::~pretty_printer): Destruct output buffer.
11031 * pretty-print.h (output_buffer::~output_buffer): Declare.
11032 (pretty_printer::~pretty_printer): Declare virtual.
11034 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
11037 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
11038 HOST_WIDE_INT_M1U): New macros.
11039 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
11040 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
11041 unsigned -1 for lshift.
11042 * cse.c (cse_insn): Likewise.
11043 * double-int.c (rshift_double, lshift_double): Likewise.
11044 * builtins.c (fold_builtin_bitop): Likewise.
11045 * combine.c (force_to_mode): Likewise.
11046 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
11047 * simplify-rtx.c (simplify_const_unary_operation,
11048 simplify_const_binary_operation): Likewise.
11049 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
11050 check_va_list_escapes): Likewise.
11051 * rtlanal.c (nonzero_bits1): Likewise.
11052 * expmed.c (expand_smod_pow2): Likewise.
11053 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
11055 2013-08-23 Jan Hubicka <jh@suse.cz>
11057 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
11058 as having address taken.
11060 2013-08-23 Jan Hubicka <jh@suse.cz>
11062 * ipa-utils.h (method_class_type): Declare.
11063 * ipa-devirt.c (method_class_type): Export.
11065 * cgraphunit.c (analyze_functions): Do basic devirtualization;
11066 do not walk base classes of anonymous types.
11068 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
11070 PR rtl-optimization/58220
11071 PR regression/58221
11072 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
11073 handle SEQUENCE insns properly.
11075 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11077 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
11079 (pp_newline_and_indent): Likewise.
11080 (pp_separate_with): Likewise.
11081 * pretty-print.c (pp_newline_and_flush): Define.
11082 (pp_newline_and_indent): Likewise.
11083 (pp_separate_with): Likewise.
11085 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11088 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
11089 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
11091 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
11093 * config/i386/predicates.md (ext_sse_reg_operand): New.
11094 * config/i386/i386.md (*movti_internal): Use
11095 predicate to determine if EVEX is needed.
11096 (*movsi_internal): Ditto.
11097 (*movdf_internal): Ditto.
11098 (*movsf_internal): Ditto.
11099 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
11101 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11103 PR tree-optimization/58209
11104 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
11105 (find_tail_calls): Give up for pointer result types if m is non-NULL.
11106 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
11107 emit POINTER_PLUS_EXPR.
11108 (create_tailcall_accumulator): For pointer result type accumulate in
11111 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
11113 * configure.ac: Add backslashes missing from the last change.
11114 * configure: Regenerate.
11116 2013-08-22 Jan Hubicka <jh@susue.cz>
11118 * ipa.c (function_and_variable_visibility): First remember function
11119 was global and then make it local.
11121 2013-08-22 Julian Brown <julian@codesourcery.com>
11123 * configure.ac: Add aarch64 to list of arches which use "nop" in
11125 * configure: Regenerate.
11127 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11129 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
11130 gnu_libc_has_function.
11131 * config/s390/tpf.h: Likewise.
11133 2013-08-22 Jan Hubicka <jh@susue.cz>
11135 * timevar.c (validate_phases): Add cast.
11137 2013-08-22 Jan Hubicka <jh@susue.cz>
11139 * timevar.c (validate_phases): Use size_t for memory.
11140 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
11142 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
11144 * pretty-print.h (output_buffer::output_buffer): Declare.
11145 (pretty_printer::pretty_printer): Likewise.
11146 (pp_construct): Remove.
11147 * pretty-print.c (output_buffer::output_buffer): Define.
11148 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
11149 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
11150 (print_gimple_expr): Likewise.
11151 (print_gimple_seq): Likewise.
11152 (gimple_dump_bb): Likewise.
11153 * sched-vis.c (dump_value_slim): Likewise.
11154 (dump_insn_slim): Likewise.
11155 (dump_rtl_slim): Likewise.
11156 (str_pattern_slim): Likewise.
11157 * tree-mudflap.c (mf_varname_tree): Likewise.
11158 * graph.c (print_graph_cfg): Likewise.
11159 (start_graph_dump): Likewise.
11160 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
11162 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
11163 pretty printer initialization.
11164 * coretypes.h (diagnostic_context): Remove superflous type alias
11166 (pretty_printer): Likewise. Declare directly as a class.
11167 (pretty_print_info): Remove declaration as class.
11168 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
11169 and pp_clear_output_area.
11170 (asan_add_global): Likewise.
11172 2013-08-22 Jan Hubicka <jh@suse.cz>
11174 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
11175 * ipa-utils.h (update_type_inheritance_graph): Declare.
11176 (possible_polymorphic_call_target_p): Declare.
11177 (possible_polymorphic_call_target_p): New.
11178 * ipa-devirt.c: Update toplevel comments.
11179 (cached_polymorphic_call_targets): Move up.
11180 (odr_type_d): Move ID down.
11181 (polymorphic_type_binfo_p): Update comment.
11182 (odr_hasher::remove): Likewise;
11183 (get_odr_type): Set anonymous_namespace.
11184 (dump_odr_type): Dump it.
11185 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
11186 (maybe_record_node): Record node in cached_polymorphic_call_targets.
11187 (record_binfo): Add comment.
11188 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
11190 (devirt_node_removal_hook): Do not iCE when cache is freed.
11191 (possible_polymorphic_call_target_p): New predicate.
11192 (update_type_inheritance_graph): New function.
11194 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
11195 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
11196 Sergey Lega <sergey.s.lega@intel.com>
11197 Anna Tikhonova <anna.tikhonova@intel.com>
11198 Ilya Tocar <ilya.tocar@intel.com>
11199 Andrey Turetskiy <andrey.turetskiy@intel.com>
11200 Ilya Verbin <ilya.verbin@intel.com>
11201 Kirill Yukhin <kirill.yukhin@intel.com>
11202 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11204 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
11205 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
11206 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
11207 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
11208 (OPTION_MASK_ISA_AVX2_UNSET): Update.
11209 (OPTION_MASK_ISA_AVX512F_UNSET): New.
11210 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
11211 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
11212 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
11213 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
11214 OPT_mavx512pf, OPT_mavx512er cases.
11215 * config/i386/constraints.md (v): New constraint.
11216 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
11217 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
11218 (bit_AVX512CD): New.
11219 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11220 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
11221 * config/i386/i386-c.c (ix86_target_macros_internal):
11222 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
11224 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
11225 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
11226 * config/i386/i386.c (regclass_map, dbx_register_map)
11227 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
11228 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
11229 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
11230 -mavx512pf options.
11231 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
11232 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
11233 -mavx512pf options. Fix formatting.
11234 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
11235 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
11236 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
11237 -mavx512cd, -mavx512pf options.
11238 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
11239 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
11240 (ix86_preferred_output_reload_class): Replace SSE_REGS with
11242 (ix86_hard_regno_mode_ok): Support 512-bit registers.
11243 (ix86_set_reg_reg_cost): Ditto.
11244 (x86_order_regs_for_local_alloc): Ditto.
11245 (MAX_VECT_LEN): Extend to 64-byte.
11246 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
11247 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
11248 (TARGET_AVX512ER, TARGET_AVX512CD): New.
11249 (BIGGEST_ALIGNMENT): Extend to 512-bits.
11250 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
11251 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
11252 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
11253 (SSE_REG_MODE_P): Support new modes.
11254 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
11255 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
11256 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
11257 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
11258 (REG_CLASS_CONTENTS): Add new registers.
11259 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
11260 (EXT_REX_SSE_REGNO_P): New.
11261 (HI_REGISTER_NAMES): Add new registers.
11262 * config/i386/i386.md: Define constants for new registers.
11263 (mode): Add new 512-bit modes.
11264 (prefix): Support evex prefix.
11265 (isa): Support avx512f, noavx512f, fma_avx512f.
11266 (ssemodesuffix): Add new 512-bit modes.
11268 (*movxi_internal_avx512f): Ditto.
11269 (*movdi_internal): Replace constraint "x" with the new constraint "v".
11271 (*movsi_internal): Likewise.
11272 (*movdf_internal): Likewise.
11273 (*movsf_internal): Likewise.
11274 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
11275 (<code><mode>3): Likewise.
11276 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
11278 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
11279 with the new constraint "v".
11280 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
11282 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
11283 with the new constraint "v".
11284 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
11285 (<sse2>_storedqu<avxsizesuffix>): Likewise.
11286 (*<plusminus_insn><mode>3): Likewise.
11287 (<sse>_vm<plusminus_insn><mode>3): Likewise.
11288 (*mul<mode>3): Likewise.
11289 (<sse>_vmmul<mode>3): Likewise.
11290 (<sse>_div<mode>3): Likewise.
11291 (<sse>_vmdiv<mode>3): Likewise.
11292 (<sse>_sqrt<mode>2): Likewise.
11293 (<sse>_vmsqrt<mode>2): Likewise.
11294 (*<code><mode>3_finite): Likewise.
11295 (*<code><mode>3) <smaxmin>: Likewise.
11296 (<sse>_vm<code><mode>3): Likewise.
11297 (*<code><mode>3) <any_logic>: Likewise.
11298 (*fma_fmadd_<mode>): Likewise.
11299 (*fma_fmsub_<mode>): Likewise.
11300 (*fma_fnmadd_<mode>): Likewise.
11301 (*fma_fnmsub_<mode>): Likewise.
11302 (*fma_fmaddsub_<mode>): Likewise.
11303 (*fma_fmsubadd_<mode>): Likewise.
11304 (*fmai_fmadd_<mode>): Likewise.
11305 (*fmai_fmsub_<mode>): Likewise.
11306 (*fmai_fnmadd_<mode>): Likewise.
11307 (*fmai_fnmsub_<mode>): Likewise.
11308 (sse_cvtsi2ss): Likewise.
11309 (sse_cvtsi2ssq): Likewise.
11310 (sse_cvtss2si): Likewise.
11311 (sse_cvtss2si_2): Likewise.
11312 (sse_cvtss2siq): Likewise.
11313 (sse_cvtss2siq_2): Likewise.
11314 (sse_cvttss2si): Likewise.
11315 (sse_cvtss2siq_2): Likewise.
11316 (float<sseintvecmodelower><mode>2): Likewise.
11317 (sse2_cvtsd2si_2): Likewise.
11318 (sse2_cvtsd2siq_2): Likewise.
11319 (*<plusminus_insn><mode>3): Likewise.
11320 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
11321 (*<sse4_1_avx2>_mul<mode>3): Likewise.
11322 (ashr<mode>3): Likewise.
11323 (<shift_insn><mode>3): Likewise.
11324 (avx2_<code><mode>3): Likewise.
11325 (*avx2_<code><mode>3): Likewise.
11326 (*andnot<mode>3): Likewise.
11327 (*<code><mode>3) <any_logic>: Likewise.
11328 (abs<mode>2): Likewise.
11329 (avx2_permvar<mode>): Likewise.
11330 (avx2_perm<mode>_1): Likewise.
11331 (*avx_vpermilp<mode>): Likewise.
11332 (avx_vpermilvar<mode>3): Likewise.
11333 (avx2_ashrv<mode>): Likewise.
11334 (avx2_<shift_insn>v<mode>): Likewise.
11335 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
11337 * doc/rtl.texi: Document XImode.
11339 2013-08-21 Jeff Law <law@redhat.com>
11341 * tree-flow.h (register_jump_thread): Pass vector of edges
11342 instead of each important edge.
11343 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
11344 thread path into a vector and pass that to register_jump_thread.
11345 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
11346 passed in edge vector to the current 3-edge form.
11349 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11351 * dce.c (fini_dce): Call df_analyze again just in case
11352 delete_unmarked_insns removed anything.
11354 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
11356 * reload.h (struct reg_equivs): Rename to ..
11357 (struct reg_equivs_s): .. this.
11359 2013-08-20 Martin Liska <marxin.liska@gmail.com>
11361 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
11363 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11365 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
11367 2013-08-21 Jeff Law <law@redhat.com>
11369 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
11370 simplify assignments too. If the RHS collapses to a singleton
11371 range, then return the value for the range.
11373 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
11375 * config/i386/sse.md (V16): Rename to...
11377 (mov<mode>): Update iterator name.
11378 (*mov<mode>_internal): Ditto.
11379 (push<mode>1): Ditto.
11380 (movmisalign<mode>): Ditto.
11382 2013-08-20 Jan Hubicka <jh@suse.cz>
11385 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
11386 entry for direct edges.
11387 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
11389 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11391 Revert my last two changes, r201865 and r201864:
11394 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11396 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11397 instances can own GC refs.
11399 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11400 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11401 (gcc::context::gt_pch_nx): Likewise.
11402 (gcc::context::gt_pch_nx): Likewise.
11403 * ggc.h (gt_ggc_mx <T>): New.
11404 (gt_pch_nx_with_op <T>): New.
11405 (gt_pch_nx <T>): New.
11406 * passes.c (opt_pass::gt_ggc_mx): New.
11407 (opt_pass::gt_pch_nx): New.
11408 (opt_pass::gt_pch_nx_with_op): New.
11409 (pass_manager::gt_ggc_mx): New.
11410 (pass_manager::gt_pch_nx): New.
11411 (pass_manager::gt_pch_nx_with_op): New.
11412 (pass_manager::operator new): Use
11413 ggc_internal_cleared_alloc_stat rather than xcalloc.
11414 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11415 (pass_manager::gt_ggc_mx): New.
11416 (pass_manager::gt_pch_nx): New.
11417 (pass_manager::gt_pch_nx_with_op): New.
11418 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11419 (opt_pass::operator new): New.
11420 (opt_pass::gt_ggc_mx): New.
11421 (opt_pass::gt_pch_nx): New.
11422 (opt_pass::gt_pch_nx_with_op): New.
11425 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11427 * Makefile.in (GTFILES): Add context.h.
11428 * context.c (gcc::context::operator new): New.
11429 (gcc::context::gt_ggc_mx): New.
11430 (gcc::context::gt_pch_nx): New.
11431 (gcc::context::gt_pch_nx): New.
11432 * context.h (gcc::context): Add GTY((user)) marking.
11433 (gcc::context::operator new): New.
11434 (gcc::context::gt_ggc_mx): New.
11435 (gcc::context::gt_pch_nx): New.
11436 (gcc::context::gt_pch_nx): New.
11437 (g): Add GTY marking.
11438 (gt_ggc_mx (gcc::context *)): New.
11439 (gt_pch_nx (gcc::context *)): New.
11440 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11441 void *cookie)): New.
11442 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11444 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11446 * dce.c (fini_dce): Call df_analyze again just in case
11447 delete_unmarked_insns removed anything.
11449 2013-08-20 Teresa Johnson <tejohnson@google.com>
11451 PR rtl-optimizations/57451
11452 * final.c (reemit_insn_block_notes): Prevent lexical blocks
11453 from crossing split section boundaries.
11455 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11457 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
11458 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
11459 with MULTLIB_DEFAULTS.
11461 2013-08-20 Nick Clifton <nickc@redhat.com>
11463 * target.def (narrow_volatile_bitfield): Note that the default
11464 value is false, not !TARGET_STRICT_ALIGN.
11465 * doc/tm.texi: Regenerate.
11467 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
11469 Fix LIB_SPEC for systems without libpthread.
11471 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
11472 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
11474 * config/i386/linux-common.h: Likewise.
11475 * config/mips/linux-common.h: Likewise.
11477 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11479 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
11482 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11484 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11485 instances can own GC refs.
11487 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11488 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11489 (gcc::context::gt_pch_nx): Likewise.
11490 (gcc::context::gt_pch_nx): Likewise.
11491 * ggc.h (gt_ggc_mx <T>): New.
11492 (gt_pch_nx_with_op <T>): New.
11493 (gt_pch_nx <T>): New.
11494 * passes.c (opt_pass::gt_ggc_mx): New.
11495 (opt_pass::gt_pch_nx): New.
11496 (opt_pass::gt_pch_nx_with_op): New.
11497 (pass_manager::gt_ggc_mx): New.
11498 (pass_manager::gt_pch_nx): New.
11499 (pass_manager::gt_pch_nx_with_op): New.
11500 (pass_manager::operator new): Use
11501 ggc_internal_cleared_alloc_stat rather than xcalloc.
11502 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11503 (pass_manager::gt_ggc_mx): New.
11504 (pass_manager::gt_pch_nx): New.
11505 (pass_manager::gt_pch_nx_with_op): New.
11506 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11507 (opt_pass::operator new): New.
11508 (opt_pass::gt_ggc_mx): New.
11509 (opt_pass::gt_pch_nx): New.
11510 (opt_pass::gt_pch_nx_with_op): New.
11512 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11514 * Makefile.in (GTFILES): Add context.h.
11515 * context.c (gcc::context::operator new): New.
11516 (gcc::context::gt_ggc_mx): New.
11517 (gcc::context::gt_pch_nx): New.
11518 (gcc::context::gt_pch_nx): New.
11519 * context.h (gcc::context): Add GTY((user)) marking.
11520 (gcc::context::operator new): New.
11521 (gcc::context::gt_ggc_mx): New.
11522 (gcc::context::gt_pch_nx): New.
11523 (gcc::context::gt_pch_nx): New.
11524 (g): Add GTY marking.
11525 (gt_ggc_mx (gcc::context *)): New.
11526 (gt_pch_nx (gcc::context *)): New.
11527 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11528 void *cookie)): New.
11529 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11531 2013-08-20 Alan Modra <amodra@gmail.com>
11534 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
11535 (rs6000_emit_epilogue): Likewise.
11537 2013-08-19 Dehao Chen <dehao@google.com>
11539 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
11541 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
11542 Jakub Jelinek <jakub@redhat.com>
11544 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
11545 (BUILT_IN_FABSD64): Likewise.
11546 (BUILT_IN_FABSD128): Likewise.
11547 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
11548 (fold_builtin_1): Likewise.
11549 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
11550 destination and source operands.
11551 (*abstd2_fpr): Likewise.
11552 (*nabstd2_fpr): Likewise.
11554 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
11556 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
11559 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
11561 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
11564 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
11566 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
11567 * builtins.c (default_libc_has_function): New.
11568 (gnu_libc_has_function): Ditto.
11569 (no_c99_libc_has_function): Ditto.
11570 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
11571 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
11572 (fold_builtin_sincos): Likewise.
11573 (fold_builtin_cexp): Likewise.
11574 * builtins.def (DEF_C94_BUILTIN): Likewise.
11575 (DEF_C99_BUILTIN): Likewise.
11576 (DEF_C99_C90RES_BUILTIN): Likewise.
11577 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
11578 definitions to using this define.
11579 * config/darwin-protos.h (darwin_libc_has_function): New.
11580 * config/darwin.c (darwin_libc_has_function): Ditto.
11581 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
11582 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
11583 * config/darwin.h: Ditto.
11584 * config/elfos.h: Ditto.
11585 * config/freebsd.h: Ditto.
11586 * config/i386/cygming.h: Ditto.
11587 * config/i386/djgpp.h: Ditto.
11588 * config/i386/i386-interix.h: Ditto.
11589 * config/microblaze/microblaze.h: Ditto.
11590 * config/mmix/mmix.h: Ditto.
11591 * config/gnu-user.h: Ditto.
11592 * config/ia64/hpux.h: Ditto.
11593 * config/pa/pa-hpux.h: Ditto.
11594 * config/pdp11/pdp11.h: Ditto.
11595 * config/picochip/picochip.h: Ditto.
11596 * config/linux.h: Ditto.
11597 * config/netbsd.h: Ditto.
11598 * config/openbsd.h: Ditto.
11599 * config/rs6000/aix43.h: Ditto.
11600 * config/rs6000/aix51.h: Ditto.
11601 * config/rs6000/aix52.h: Ditto.
11602 * config/rs6000/aix53.h: Ditto.
11603 * config/rs6000/aix61.h: Ditto.
11604 * config/rs6000/darwin.h: Ditto.
11605 * config/rs6000/linux.h: Ditto.
11606 * config/rs6000/linux64.h: Ditto.
11607 * config/s390/tpf.h: Ditto.
11608 * config/sol2-10.h: Ditto.
11609 * config/sol2.h: Ditto.
11610 * config/vms/vms.h: Ditto.
11611 * config/vxworks.h: Ditto.
11612 * config/linux-android.c (linux_android_libc_has_function):
11613 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
11614 * config/linux-protos.h (linux_android_libc_has_function):
11616 * config/i386/i386.c (ix86_libc_has_function): New.
11617 * config/i386/i386-protos.h
11618 (ix86_libc_has_function): New declaration.
11619 * config/i386/i386.md
11620 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
11621 ("isinf<mode>2): Likewise.
11622 * convert.c (convert_to_integer): Using new target hook
11623 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
11624 TARGET_C99_FUNCTIONS.
11625 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
11626 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
11627 * coretypes.h (function_class): New enum for different
11628 classes of functions.
11629 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
11630 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
11631 (TARGET_HAS_SINCOS): Likewise.
11632 (TARGET_LIBC_HAS_FUNCTION): New.
11633 * doc/tm.texi: Regenerated.
11634 * targhooks.h (default_libc_has_function): New declaration.
11635 (no_c99_libc_has_function): Ditto.
11636 (gnu_libc_has_function): Ditto.
11637 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
11638 and TARGET_HAS_SINCOS.
11640 2013-08-18 Jan Hubicka <jh@suse.cz>
11642 * Makeifle-in (ipa-devirt.o): New.
11643 (GTFILES): Add ipa-utils.h and ipa-devirt.c
11644 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
11645 (analyze_functions): Look into possible targets of polymorphic call.
11646 * dumpfile.c (dump_files): Add type-inheritance dump.
11647 * dumpfile.h (TDI_inheritance): New.
11648 * ipa-devirt.c: New file.
11649 * ipa-utils.h (odr_type_d): Forward declare.
11650 (odr_type): New type.
11651 (build_type_inheritance_graph): Declare.
11652 (possible_polymorphic_call_targets): Declare and introduce inline
11653 variant when only edge is pased.
11654 (dump_possible_polymorphic_call_targets): Likewise.
11655 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
11656 * tree.c (type_in_anonymous_namespace_p): Break out from ...
11657 (types_same_for_odr): ... here.
11658 * tree.h (type_in_anonymous_namespace_p): Declare.
11660 2013-08-18 Jakub Jelinek <jakub@redhat.com>
11662 PR tree-optimization/58006
11663 * tree-parloops.c (take_address_of): Don't ICE if get_name
11665 (eliminate_local_variables_stmt): Remove clobber stmts.
11667 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
11669 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
11670 error message is issued for an alias to undefined symbol.
11672 2013-08-18 Jan Hubicka <jh@suse.cz>
11674 * cgraph.c (cgraph_create_indirect_edge): Discover
11675 polymorphic calls and record basic info into indirect_info.
11676 * gimple-fold.c (gimple_fold_call): When doing BINFO based
11677 devirtualization, ignore objc function calls.
11678 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
11679 call with no parm index info.
11680 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
11681 * tree.c (virtual_method_call_p): New function.
11682 * tree.h (virtual_method_call_p): Declare.
11684 2013-08-16 Jan Hubicka <jh@suse.cz>
11686 PR middle-end/58179
11687 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
11689 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
11691 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
11694 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11696 * gengtype.c (type_for_name): Add special-case support for
11697 locating types within the "gcc::" namespace.
11698 (open_base_files): Emit a "using namespace gcc" directive.
11700 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11703 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
11704 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
11706 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
11707 array instead of each individual operand as a separate argument.
11708 (emit_fusion_gpr_load): Likewise.
11709 (expand_fusion_gpr_load): Add new function declaration.
11711 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
11712 signature to have the operands passed as an array, instead of as
11713 separate arguments. Allow ZERO_EXTEND to be in the memory
11714 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
11715 depend on the register live/dead flags when peepholes are run.
11716 (expand_fusion_gpr_load): New function to be called from the
11717 peephole2 pass, to change the register that addis sets to be the
11719 (emit_fusion_gpr_load): Change the calling signature to have the
11720 operands passed as an array, instead of as separate arguments.
11721 Allow ZERO_EXTEND to be in the memory address, and also
11722 SIGN_EXTEND if -mpower8-fusion-sign.
11724 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
11725 unspec enumeration.
11726 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
11727 adjust the register addis loads up in the peephole2 pass. Do not
11728 depend on the register live/dead state when the peephole pass is done.
11730 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11732 * gengtype.c (create_user_defined_type): Ensure that the kind
11733 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
11734 declaration is seen before the GTY((user)) marking.
11736 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
11739 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
11741 2013-08-16 Jan Hubicka <jh@suse.cz>
11743 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
11744 arugment expected_type.
11745 (gimple_fold_call): Use it.
11746 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
11747 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
11748 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
11749 (try_make_edge_direct_virtual_call): Likewise.
11750 * tree.c (obj_type_ref_class): New.
11751 * tree.h (obj_type_ref_class): Use it.
11753 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
11755 * sched-vis.c (rtl_slim_pp_initialized): Remove.
11756 (rtl_slim_pp): Likewise.
11757 (init_rtl_slim_pretty_print): Likewise.
11758 (dump_value_slim): Don't call it. Use local pretty printer.
11759 (dump_insn_slim): Likewise.
11760 (dump_rtl_slim): Likewise.
11761 (str_pattern_slim): Likewise.
11762 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
11765 2013-08-16 Jakub Jelinek <jakub@redhat.com>
11767 PR tree-optimization/58164
11768 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
11769 walk gimple_goto_dest of GIMPLE_GOTO.
11771 PR tree-optimization/58165
11772 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
11773 bi_call must be the last stmt in a bb, don't split_block, instead
11774 use fallthru edge from it and give up if there is none.
11775 Release conds vector when returning early.
11777 2013-08-14 Xinliang David Li <davidxl@google.com>
11779 * config/i386/i386.c (ix86_option_override_internal):
11780 Remove unused variable and field.
11782 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11785 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
11786 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
11787 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
11788 and Linux, correct BLKmode alignment when 128-bit alignment is
11789 required and compatibility flag is not set.
11790 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
11791 for zero-size arguments when compatibility flag is not set.
11793 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11795 PR tree-optimization/58145
11796 * tree-sra.c (build_ref_for_offset): If prev_base has
11797 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
11799 2013-08-14 Xinliang David Li <davidxl@google.com>
11801 * config/i386/i386.c (ix86_option_override_internal):
11802 Fix uninitialized variable error.
11804 2013-08-14 Xinliang David Li <davidxl@google.com>
11806 * config/i386/i386.opt: Define two new options.
11807 * config/i386/x86-tune.def: Add arch selector field in macros.
11808 * config/i386/i386.h: Adjust macro definition.
11809 * config/i386/i386.c (ix86_option_override_internal):
11811 (parse_mtune_ctrl_str): New function.
11812 (set_ix86_tune_features): New function.
11813 (ix86_function_specific_restore): Call the new helper function.
11815 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
11817 PR rtl-optimization/57662
11818 * sel-sched.c (code_motion_process_successors): When the current insn
11819 is removed after the recursive traversal, break from the loop.
11820 Add comments and debug printouts.
11822 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11823 Alexandre Oliva <aoliva@redhat.com>
11826 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
11827 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
11828 there also UNSPEC_PLTOFF.
11830 2013-08-14 Marek Polacek <polacek@redhat.com>
11832 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
11833 NUM_CONDITIONS bit positions.
11835 2013-08-13 Cary Coutant <ccoutant@google.com>
11837 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
11838 (attr_checksum): Hash vector contents instead of pointer.
11839 (attr_checksum_ordered): Likewise.
11841 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
11843 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
11844 when Pmode != word_mode. Add length_address attribute.
11845 (sse3_monitor_<mode>): Merge from sse3_monitor and
11846 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
11847 Pmode != word_mode. Update insn length attribute.
11848 * config/i386/i386.c (ix86_option_override_internal): Update
11849 ix86_gen_monitor selection for merged sse3_monitor insn.
11851 2013-08-13 Julian Brown <julian@codesourcery.com>
11853 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
11854 perform invalid legitimization on greater-than-word-size modes for
11855 TARGET_E500_DOUBLE.
11857 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
11859 * ira.c (setup_class_translate_array): Use aclass instead of cl
11860 for classes not fully covered by allocno classes.
11862 2013-08-13 Jakub Jelinek <jakub@redhat.com>
11864 PR tree-optimization/57661
11865 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
11866 * tree-inline.c (tree_function_versioning): Initialize it.
11867 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
11868 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
11869 that is not being copied.
11872 * asan.c (instrument_strlen_call): Fix typo in comment.
11873 Use char * type even for the lhs of POINTER_PLUS_EXPR.
11875 2013-08-13 Steve Ellcey <sellcey@mips.com>
11877 * config/mips/mips.md (prefetch): Use lw instead of ld on
11878 loongson in 32bit mode.
11880 2013-08-13 Nick Clifton <nickc@redhat.com>
11882 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
11884 2013-08-13 Jan Hubicka <jh@suse.cz>
11886 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
11887 introduced edge; fix typo in sanity check.
11888 (cgraph_resolve_speculation): Export; improve diagnostic.
11889 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
11890 speculation at type mismatch.
11891 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
11892 (cgraph_resolve_speculation): Declare.
11893 (symtab_can_be_discarded): New function.
11894 * value-prof.c (gimple_ic_transform): Remove actual transform code.
11895 * ipa-inline-transform.c (speculation_removed): New global var.
11896 (clone_inlined_nodes): See if speculation can be removed.
11897 (inline_call): If speculations was removed, we growths may not match.
11898 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
11899 (speculation_useful_p): New function.
11900 (resolve_noninline_speculation): New function.
11901 (inline_small_functions): Resolve useless speculations.
11902 * ipa-inline.h (speculation_useful_p): Declare
11903 * ipa.c (can_replace_by_local_alias): Simplify.
11904 (ipa_profile): Produce speculative calls in non-lto, too;
11905 add simple cost model; produce local aliases.
11907 2013-08-13 David Malcolm <dmalcolm@redhat.com>
11909 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
11912 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
11914 * config/i386/i386.c (ix86_function_versions): Use error + inform.
11916 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
11918 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
11919 iterator instead of X87MODEF.
11921 2013-08-12 Perez Read <netfirewall@gmail.com>
11924 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
11925 operand 0 for intel asm alternative.
11926 (*movabs<mode>_2): Ditto for operand 1.
11928 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
11930 * config/aarch64/arm_none.h
11931 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
11933 2013-08-12 Nick Clifton <nickc@redhat.com>
11935 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
11937 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
11939 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
11940 expand for QI/HImode operand to produce more effictive code for
11941 unsigned char(short) --> float(double) conversion.
11943 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
11945 * doc/invoke.texi: Mention that -ftls-model does not force the final
11948 2013-08-12 Marek Polacek <polacek@redhat.com>
11949 Marc Glisse <marc.glisse@inria.fr>
11951 PR tree-optimization/57980
11952 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
11953 when creating -1 constant.
11955 2013-08-10 Jan Hubicka <jh@suse.cz>
11957 Workaround binutils PR14342.
11958 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
11959 (gimple_init_edge_profiler): Likewise.
11960 (gimple_gen_ic_func_profiler): Likewise.
11962 2013-08-09 Jan Hubicka <jh@suse.cz>
11964 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
11966 2013-08-09 Xinliang David Li <davidxl@google.com>
11968 * config/i386/stringop.def: New file.
11969 * config/i386/stringop.opt: New file.
11970 * config/i386/i386-opts.h: Include stringopt.def.
11971 * config/i386/i386.opt: Include stringopt.opt.
11972 * config/i386/i386.c (ix86_option_override_internal):
11973 Override default size based stringop inline strategies with options.
11974 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
11977 2013-08-09 Jan Hubicka <jh@suse.cz>
11979 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
11981 2013-08-09 Jan Hubicka <jh@suse.cz>
11983 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
11985 (dump_cgraph_node): Dump profile-id.
11986 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
11987 and common_target_probability.
11988 * lto-cgraph.c (lto_output_edge): Stream common targets.
11989 (lto_output_node): Stream profile ids.
11990 (input_node): Stream profile ids.
11991 (input_edge): Stream common targets.
11992 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
11993 * ipa.c: Include value-prof.h
11994 (ipa_profile_generate_summary): Turn indirect call statement histograms
11995 into common targets.
11996 (ipa_profile): Turn common targets into speculative edges.
11998 2013-08-09 Jan Hubicka <jh@suse.cz>
12000 * cgraph.h (cgraph_node): Add profile_id.
12001 * value-prof.c (cgraph_node_map): Turn into pointer_map.
12002 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
12003 (del_node_map): Update.
12004 (find_func_by_funcdef_no): Replace by ...
12005 (find_func_by_profile_id): ... this one.
12006 (gimple_ic_transform): Do not remove useful histograms when
12007 speculation is not done; dump info when indirect call removal
12009 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
12010 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
12011 (__gcov_indirect_call_profiler_v2): .. this one.
12012 * profile.h (init_node_map): Update.
12013 * coverage.c (coverage_compute_profile_id): New function.
12014 * coverage.h (coverage_compute_profile_id): Declare.
12015 * tree-profile.c (init_ic_make_global_vars): Make
12016 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
12017 (gimple_init_edge_profiler): Update prototype of
12018 __gcov_indirect_call_profiler.
12019 (gimple_gen_ic_func_profiler): Simplify.
12020 (tree_profiling): Use init_node_map
12022 2013-08-09 Jan Hubicka <jh@suse.cz>
12024 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
12025 non-speculative refs.
12026 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
12027 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
12028 (cgraph_set_call_stmt): Likewise.
12029 (cgraph_create_edge_1): Fix release checking compilatoin;
12030 clear lto_stmt_uid.
12031 (cgraph_free_edge): Free indirect info.
12032 (cgraph_turn_edge_to_speculative): New function.
12033 (cgraph_speculative_call_info): New function.
12034 (cgraph_make_edge_direct): Return direct edge; handle speculation.
12035 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
12036 (dump_cgraph_node): Dump speculation.
12037 (verify_edge_count_and_frequency): Accept speculative edges.
12038 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
12039 (verify_cgraph_node): Handle speculation.
12040 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
12041 (cgraph_set_call_stmt): Update prototype.
12042 (cgraph_make_edge_direct): Update prototype.
12043 (cgraph_speculative_call_info): Declare.
12044 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
12045 to change; update call of ipa_find_references.
12046 * ipa-ref.c (ipa_record_reference): Fix return value; clear
12047 lto_stmt_uid and speculative flags.
12048 (ipa_dump_references): Dump speculation.
12049 (ipa_clone_references): Clone speculative flag.
12050 (ipa_clone_referring): Likewise.
12051 (ipa_clone_ref): New function.
12052 (ipa_find_reference): Look into lto_stmt_uids
12053 (ipa_clear_stmts_in_references): Do not clear speculative calls.
12054 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
12055 (ipa_find_reference): Update declaration.
12056 (ipa_clone_ref): Declare.
12057 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
12058 stream speculative flag.
12059 (lto_output_ref): Stream statements uids and speculation.
12060 (input_ref): Likewise.
12061 (input_edge): Stream speuclation.
12062 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
12063 (cgraph_set_call_stmt_including_clones): Handle speculation.
12064 * ipa-inline.c (heap_edge_removal_hook): New function.
12065 (inline_small_functions): Register it.
12066 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
12067 also initialize refs.
12068 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
12070 (try_make_edge_direct_simple_call): Likewise.
12071 (try_make_edge_direct_simple_call): Likewise.
12072 (update_indirect_edges_after_inlining): Likewise.
12073 (remove_described_reference): Look proper lto_stmt_uid.
12074 (propagate_controlled_uses): Likewise.
12075 (propagate_controlled_uses): Liekwise.
12076 * tree-inline.c (copy_bb): Copy speculative edges.
12077 (redirect_all_calls): New function.
12078 (copy_cfg_body): Do redirection after loop info is updated.
12079 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
12081 2013-08-09 Jan Hubicka <jh@suse.cz>
12083 * lto-streamer-out.c (output_function): Renumber PHIs.
12084 * lto-streamer-in.c (input_function): Likewise.
12086 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
12088 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
12089 (get_lane_unsigned): Likewise.
12090 (dup_lane_scalar): Likewise.
12091 (get_lane): enable for VALL.
12092 * config/aarch64/aarch64-simd.md
12093 (aarch64_dup_lane_scalar<mode>): Remove.
12094 (aarch64_get_lane_signed<mode>): Likewise.
12095 (aarch64_get_lane_unsigned<mode>): Likewise.
12096 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
12097 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
12098 (aarch64_get_lane<mode>): Enable for all vector modes.
12099 (aarch64_get_lanedi): Remove misleading constraints.
12100 * config/aarch64/arm_neon.h
12101 (__aarch64_vget_lane_any): Define.
12102 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
12103 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
12104 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
12105 * config/aarch64/iterators.md (VDQQH): New.
12106 (VDQQHS): Likewise.
12107 (vwcore): Likewise.
12109 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
12111 * configure.ac: Add GAS check for LEON instructions on SPARC.
12112 * configure: Regenerate.
12113 * config.in: Likewise.
12114 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
12116 * config/sparc/sparc.opt (LEON, LEON3): New masks.
12117 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
12119 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
12120 (AS_LEON_FLAG): New macro.
12121 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
12122 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
12123 Deal with LEON and LEON3 for the memory model.
12124 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
12125 (atomic_compare_and_swap<mode>_1): Likewise.
12126 (*atomic_compare_and_swap<mode>_1): Likewise.
12128 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12130 * config/arm/neon.md (vcond): Fix floating-point vector
12131 comparisons against 0.
12133 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
12135 * lra-constraints.c (emit_spill_move): Remove assert.
12136 (process_alt_operands): Add more debugging
12137 output. Increase reject for spilling into memory. Decrease
12138 reject for reloading scratch.
12139 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
12141 2013-08-08 Steve Ellcey <sellcey@mips.com>
12143 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
12144 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
12145 micromips incompatible. Add nan2008.
12146 (MULTILIB_DIRNAMES): Add nan2008.
12147 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12148 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
12149 and micromips incompatible. Add nan2008.
12150 (MULTILIB_DIRNAMES): Add nan2008.
12151 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12153 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
12155 PR rtl-optimization/58079
12156 * combine.c (combine_simplify_rtx): Avoid using SUBST if
12157 simplify_comparison has widened a comparison with an integer.
12159 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12161 * config/arm/neon.md (movmisalign<mode>): Disable when we
12162 don't allow unaligned accesses.
12163 (*movmisalign<mode>_neon_store): Likewise.
12164 (*movmisalign<mode>_neon_load): Likewise.
12165 (*movmisalign<mode>_neon_store): Likewise.
12166 (*movmisalign<mode>_neon_load): Likewise.
12168 2013-08-08 Jan Hubicka <jh@suse.cz>
12170 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
12171 (make_pass_rebuild_cgraph_edges): Also clear references.
12172 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
12173 * ipa-inline-transform.c (inline_transform): Remove all references
12175 * cgraphunit.c (expand_function): Remove all references after
12177 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
12178 (ipa_find_reference): Rewrite to iterator.
12179 (remove_stmt_references): Likewise.
12180 (ipa_clear_stmts_in_references): New function.
12181 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
12182 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
12184 * ipa-split.c (split_function): Remove references in split function.
12186 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12189 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
12190 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
12192 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12195 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
12196 suggested mode for the assignment isn't compatible with the
12197 registers required.
12199 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12202 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
12204 2013-08-07 Xinliang David Li <davidxl@google.com>
12206 * config/i386/i386.opt: New option -mtune-ctrl=.
12207 * config/i386/x86-tune.def: New file.
12208 * config/i386/i386.h: include x86-tune.def.
12209 * config/i386/i386.c (ix86_option_override_internal):
12210 Parsing -mtune-ctrl= option and set tune features.
12212 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
12215 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
12216 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
12219 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12221 * rtl.h (update_alignments): Declare.
12222 * final.c (grow_label_align): New function extracted from...
12223 (shorten_branches): ...here. Call it.
12224 (update_alignments): New function.
12225 * reorg.c (sibling_labels): New variable.
12226 (get_label_before): Add SIBLING parameter. If it is non-zero, push
12227 the new label along with it onto the sibling_labels vector.
12228 (fill_simple_delay_slots): Adjust call to get_label_before.
12229 (fill_slots_from_thread): Likewise.
12230 (relax_delay_slots): Likewise.
12231 (make_return_insns): Likewise.
12232 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
12234 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12236 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
12237 document its semantics.
12238 (diagnostic_report_diagnostic): Adjust accordingly.
12240 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12242 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
12243 (sparc_option_override): ...and port to new C++ pass API.
12244 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
12246 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
12248 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
12250 2013-08-06 Caroline Tice <cmtice@google.com>
12252 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
12253 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
12254 * tree-pass.h: Add pass_vtable_verify.
12255 * varasm.c (assemble_variable): Add code to properly set the comdat
12256 section and name for the .vtable_map_vars section.
12257 (assemble_vtyv_preinit_initializer): New function.
12258 (default_sectin_type_flags): Make sure .vtable_map_vars section has
12260 * output.h: Add function decl for assemble_vtv_preinit_initializer.
12261 * vtable-verify.c: New file.
12262 * vtable-verify.h: New file.
12263 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
12264 initialiation levels.
12265 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
12266 * passes.def: Insert pass_vtable_verify.
12267 * aclocal.m4: Reorder includes.
12268 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
12269 -fvtv-counts options.
12270 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
12271 as appropriate, if -fvtable-verify=... is used.
12272 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
12273 -fvtable-verify=... is used.
12274 * Makefile.in (OBJS): Add vtable-verify.o to list.
12275 (vtable-verify.o): Add new build rule.
12276 (GTFILES): Add vtable-verify.c to list.
12277 * common.opt (fvtable-verify=): New flag.
12278 (vtv_priority): Values for fvtable-verify= flag.
12279 (fvtv-counts): New flag.
12280 (fvtv-debug): New flag.
12281 * tree.h (save_vtable_map_decl): New extern function decl.
12283 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12285 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
12286 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
12287 (pass_data_rl78_devirt): ...new pass_data instance and...
12288 (make_pass_rl78_devirt): ...new function.
12289 (rl78_asm_file_start): Port pass registration to new C++ API.
12291 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12293 * coretypes.h (rtl_opt_pass): Add.
12294 (gcc::context): Add.
12295 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
12296 (epiphany_init): Port to new C++ pass API.
12297 (epiphany_optimize_mode_switching): Likewise.
12298 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
12299 (pass_manager::get_pass_mode_switching): New.
12300 (pass_manager::get_pass_peephole2): New.
12301 * mode-switching.c (pass_mode_switching): Add clone method.
12302 * recog.c (pass_peephole2): Add clone method.
12303 (pass_split_all_insns): Add clone method.
12305 2013-08-06 David Malcolm <dmalcolm@redhat.com>
12307 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
12308 (mips_option_override): ...here, porting to new C++ API for passes.
12310 2013-08-06 Jan Hubicka <jh@suse.cz>
12312 * cgraph.c (cgraph_get_body): New function based on lto.c
12314 * cgraph.h (cgraph_get_body): Declare.
12315 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
12317 * cgraphunit.c (expand_function): Get body prior expanding.
12318 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
12319 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
12321 * passes.c (do_per_function_toporder): Get body.
12322 * tree-inline.c (expand_call_inline): Get body prior inlining it.
12323 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
12325 2013-08-06 Martin Jambor <mjambor@suse.cz>
12328 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
12329 is not re-finalized. Rename second parameter to no_collect.
12331 2013-08-06 Martin Jambor <mjambor@suse.cz>
12333 PR middle-end/58041
12334 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
12335 MEM_REF has proper alignment information.
12337 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
12340 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
12342 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
12343 rtx (*) (rtx, ...) with insn_gen_fn.
12344 * genoutput.c (output_insn_data): Cast gen_? function pointers to
12345 insn_gen_fn::stored_funcptr. Add initializer braces.
12347 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12349 Rewrite how instances of passes are cloned to remove assumptions
12350 about their sizes (thus allowing pass subclasses to have
12351 additional data fields, albeit non-GC-managed ones at this point).
12353 * passes.c (make_pass_instance): Now that passes have clone
12354 methods, rewrite this function to eliminate XNEW and memcpy
12355 calls that used hardcoded sizes. Since this function no longer
12356 creates pass instances, rename it to...
12357 (add_pass_instance): ...this. Document the old way that passes were
12358 numbered and flagged, and rework this function to continue using it.
12359 (next_pass_1): Add an initial_pass argument for use by
12361 (position_pass): When adding multiple instances of a pass, use
12362 the pass's clone method, rather than relying on the XNEW/memcpy
12363 within the former make_pass_instance (now add_pass_instance).
12364 (pass_manager::pass_manager): When invoking next_pass_1, also supply
12365 the initial instance of the current pass within the pass manager.
12367 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12369 This is the automated part of the conversion of passes from C
12370 structs to C++ classes.
12372 Patch autogenerated by refactor_passes.py from
12373 https://github.com/davidmalcolm/gcc-refactoring-scripts
12374 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
12376 * asan.c (pass_asan): Convert from a global struct to a subclass of
12377 gimple_opt_pass along with...
12378 (pass_data_asan): ...new pass_data instance and...
12379 (make_pass_asan): ...new function.
12380 (pass_asan_O0): Convert from a global struct to a subclass of
12381 gimple_opt_pass along with...
12382 (pass_data_asan_O0): ...new pass_data instance and...
12383 (make_pass_asan_O0): ...new function.
12384 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
12385 subclass of rtl_opt_pass along with...
12386 (pass_data_inc_dec): ...new pass_data instance and...
12387 (make_pass_inc_dec): ...new function.
12388 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
12389 a subclass of rtl_opt_pass along with...
12390 (pass_data_reorder_blocks): ...new pass_data instance and...
12391 (make_pass_reorder_blocks): ...new function.
12392 (pass_duplicate_computed_gotos): Convert from a global struct to a
12393 subclass of rtl_opt_pass along with...
12394 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
12395 (make_pass_duplicate_computed_gotos): ...new function.
12396 (pass_partition_blocks): Convert from a global struct to a subclass of
12397 rtl_opt_pass along with...
12398 (pass_data_partition_blocks): ...new pass_data instance and...
12399 (make_pass_partition_blocks): ...new function.
12400 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
12401 struct to a subclass of rtl_opt_pass along with...
12402 (pass_data_branch_target_load_optimize1): ...new pass_data instance
12404 (make_pass_branch_target_load_optimize1): ...new function.
12405 (pass_branch_target_load_optimize2): Convert from a global struct to a
12406 subclass of rtl_opt_pass along with...
12407 (pass_data_branch_target_load_optimize2): ...new pass_data instance
12409 (make_pass_branch_target_load_optimize2): ...new function.
12410 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
12411 of rtl_opt_pass along with...
12412 (pass_data_jump): ...new pass_data instance and...
12413 (make_pass_jump): ...new function.
12414 (pass_jump2): Convert from a global struct to a subclass of
12415 rtl_opt_pass along with...
12416 (pass_data_jump2): ...new pass_data instance and...
12417 (make_pass_jump2): ...new function.
12418 * cfgexpand.c (pass_expand): Convert from a global struct to a
12419 subclass of rtl_opt_pass along with...
12420 (pass_data_expand): ...new pass_data instance and...
12421 (make_pass_expand): ...new function.
12422 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
12423 of rtl_opt_pass along with...
12424 (pass_data_free_cfg): ...new pass_data instance and...
12425 (make_pass_free_cfg): ...new function.
12426 (pass_into_cfg_layout_mode): Convert from a global struct to a
12427 subclass of rtl_opt_pass along with...
12428 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
12429 (make_pass_into_cfg_layout_mode): ...new function.
12430 (pass_outof_cfg_layout_mode): Convert from a global struct to a
12431 subclass of rtl_opt_pass along with...
12432 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
12433 (make_pass_outof_cfg_layout_mode): ...new function.
12434 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
12435 struct to a subclass of gimple_opt_pass along with...
12436 (pass_data_build_cgraph_edges): ...new pass_data instance and...
12437 (make_pass_build_cgraph_edges): ...new function.
12438 (pass_rebuild_cgraph_edges): Convert from a global struct to a
12439 subclass of gimple_opt_pass along with...
12440 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
12441 (make_pass_rebuild_cgraph_edges): ...new function.
12442 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
12443 subclass of gimple_opt_pass along with...
12444 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
12446 (make_pass_remove_cgraph_callee_edges): ...new function.
12447 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
12448 struct to a subclass of rtl_opt_pass along with...
12449 (pass_data_stack_adjustments): ...new pass_data instance and...
12450 (make_pass_stack_adjustments): ...new function.
12451 * combine.c (pass_combine): Convert from a global struct to a subclass
12452 of rtl_opt_pass along with...
12453 (pass_data_combine): ...new pass_data instance and...
12454 (make_pass_combine): ...new function.
12455 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
12456 global struct to a subclass of rtl_opt_pass along with...
12457 (pass_data_compare_elim_after_reload): ...new pass_data instance
12459 (make_pass_compare_elim_after_reload): ...new function.
12460 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
12461 of rtl_opt_pass along with...
12462 (pass_data_rtl_cprop): ...new pass_data instance and...
12463 (make_pass_rtl_cprop): ...new function.
12464 * cse.c (pass_cse): Convert from a global struct to a subclass of
12465 rtl_opt_pass along with...
12466 (pass_data_cse): ...new pass_data instance and...
12467 (make_pass_cse): ...new function.
12468 (pass_cse2): Convert from a global struct to a subclass of
12469 rtl_opt_pass along with...
12470 (pass_data_cse2): ...new pass_data instance and...
12471 (make_pass_cse2): ...new function.
12472 (pass_cse_after_global_opts): Convert from a global struct to a
12473 subclass of rtl_opt_pass along with...
12474 (pass_data_cse_after_global_opts): ...new pass_data instance and...
12475 (make_pass_cse_after_global_opts): ...new function.
12476 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
12477 of rtl_opt_pass along with...
12478 (pass_data_ud_rtl_dce): ...new pass_data instance and...
12479 (make_pass_ud_rtl_dce): ...new function.
12480 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
12481 rtl_opt_pass along with...
12482 (pass_data_fast_rtl_dce): ...new pass_data instance and...
12483 (make_pass_fast_rtl_dce): ...new function.
12484 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
12485 a subclass of rtl_opt_pass along with...
12486 (pass_data_df_initialize_opt): ...new pass_data instance and...
12487 (make_pass_df_initialize_opt): ...new function.
12488 (pass_df_initialize_no_opt): Convert from a global struct to a
12489 subclass of rtl_opt_pass along with...
12490 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
12491 (make_pass_df_initialize_no_opt): ...new function.
12492 (pass_df_finish): Convert from a global struct to a subclass of
12493 rtl_opt_pass along with...
12494 (pass_data_df_finish): ...new pass_data instance and...
12495 (make_pass_df_finish): ...new function.
12496 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
12497 rtl_opt_pass along with...
12498 (pass_data_rtl_dse1): ...new pass_data instance and...
12499 (make_pass_rtl_dse1): ...new function.
12500 (pass_rtl_dse2): Convert from a global struct to a subclass of
12501 rtl_opt_pass along with...
12502 (pass_data_rtl_dse2): ...new pass_data instance and...
12503 (make_pass_rtl_dse2): ...new function.
12504 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
12505 subclass of rtl_opt_pass along with...
12506 (pass_data_dwarf2_frame): ...new pass_data instance and...
12507 (make_pass_dwarf2_frame): ...new function.
12508 * except.c (pass_set_nothrow_function_flags): Convert from a global
12509 struct to a subclass of rtl_opt_pass along with...
12510 (pass_data_set_nothrow_function_flags): ...new pass_data instance
12512 (make_pass_set_nothrow_function_flags): ...new function.
12513 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
12514 subclass of rtl_opt_pass along with...
12515 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
12517 (make_pass_convert_to_eh_region_ranges): ...new function.
12518 * final.c (pass_compute_alignments): Convert from a global struct to a
12519 subclass of rtl_opt_pass along with...
12520 (pass_data_compute_alignments): ...new pass_data instance and...
12521 (make_pass_compute_alignments): ...new function.
12522 (pass_final): Convert from a global struct to a subclass of
12523 rtl_opt_pass along with...
12524 (pass_data_final): ...new pass_data instance and...
12525 (make_pass_final): ...new function.
12526 (pass_shorten_branches): Convert from a global struct to a subclass of
12527 rtl_opt_pass along with...
12528 (pass_data_shorten_branches): ...new pass_data instance and...
12529 (make_pass_shorten_branches): ...new function.
12530 (pass_clean_state): Convert from a global struct to a subclass of
12531 rtl_opt_pass along with...
12532 (pass_data_clean_state): ...new pass_data instance and...
12533 (make_pass_clean_state): ...new function.
12534 * function.c (pass_instantiate_virtual_regs): Convert from a global
12535 struct to a subclass of rtl_opt_pass along with...
12536 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
12537 (make_pass_instantiate_virtual_regs): ...new function.
12538 (pass_leaf_regs): Convert from a global struct to a subclass of
12539 rtl_opt_pass along with...
12540 (pass_data_leaf_regs): ...new pass_data instance and...
12541 (make_pass_leaf_regs): ...new function.
12542 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
12543 subclass of rtl_opt_pass along with...
12544 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
12546 (make_pass_thread_prologue_and_epilogue): ...new function.
12547 (pass_match_asm_constraints): Convert from a global struct to a
12548 subclass of rtl_opt_pass along with...
12549 (pass_data_match_asm_constraints): ...new pass_data instance and...
12550 (make_pass_match_asm_constraints): ...new function.
12551 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
12552 subclass of rtl_opt_pass along with...
12553 (pass_data_rtl_fwprop): ...new pass_data instance and...
12554 (make_pass_rtl_fwprop): ...new function.
12555 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
12556 rtl_opt_pass along with...
12557 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
12558 (make_pass_rtl_fwprop_addr): ...new function.
12559 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
12560 rtl_opt_pass along with...
12561 (pass_data_rtl_pre): ...new pass_data instance and...
12562 (make_pass_rtl_pre): ...new function.
12563 (pass_rtl_hoist): Convert from a global struct to a subclass of
12564 rtl_opt_pass along with...
12565 (pass_data_rtl_hoist): ...new pass_data instance and...
12566 (make_pass_rtl_hoist): ...new function.
12567 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
12568 subclass of gimple_opt_pass along with...
12569 (pass_data_lower_cf): ...new pass_data instance and...
12570 (make_pass_lower_cf): ...new function.
12571 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
12572 from a global struct to a subclass of gimple_opt_pass along with...
12573 (pass_data_strength_reduction): ...new pass_data instance and...
12574 (make_pass_strength_reduction): ...new function.
12575 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
12576 of rtl_opt_pass along with...
12577 (pass_data_rtl_ifcvt): ...new pass_data instance and...
12578 (make_pass_rtl_ifcvt): ...new function.
12579 (pass_if_after_combine): Convert from a global struct to a subclass of
12580 rtl_opt_pass along with...
12581 (pass_data_if_after_combine): ...new pass_data instance and...
12582 (make_pass_if_after_combine): ...new function.
12583 (pass_if_after_reload): Convert from a global struct to a subclass of
12584 rtl_opt_pass along with...
12585 (pass_data_if_after_reload): ...new pass_data instance and...
12586 (make_pass_if_after_reload): ...new function.
12587 * init-regs.c (pass_initialize_regs): Convert from a global struct to
12588 a subclass of rtl_opt_pass along with...
12589 (pass_data_initialize_regs): ...new pass_data instance and...
12590 (make_pass_initialize_regs): ...new function.
12591 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
12592 of ipa_opt_pass_d along with...
12593 (pass_data_ipa_cp): ...new pass_data instance and...
12594 (make_pass_ipa_cp): ...new function.
12595 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
12596 global struct to a subclass of gimple_opt_pass along with...
12597 (pass_data_inline_parameters): ...new pass_data instance and...
12598 (make_pass_inline_parameters): ...new function.
12599 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
12600 subclass of gimple_opt_pass along with...
12601 (pass_data_early_inline): ...new pass_data instance and...
12602 (make_pass_early_inline): ...new function.
12603 (pass_ipa_inline): Convert from a global struct to a subclass of
12604 ipa_opt_pass_d along with...
12605 (pass_data_ipa_inline): ...new pass_data instance and...
12606 (make_pass_ipa_inline): ...new function.
12607 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
12608 struct to a subclass of gimple_opt_pass along with...
12609 (pass_data_local_pure_const): ...new pass_data instance and...
12610 (make_pass_local_pure_const): ...new function.
12611 (pass_ipa_pure_const): Convert from a global struct to a subclass of
12612 ipa_opt_pass_d along with...
12613 (pass_data_ipa_pure_const): ...new pass_data instance and...
12614 (make_pass_ipa_pure_const): ...new function.
12615 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
12616 to a subclass of ipa_opt_pass_d along with...
12617 (pass_data_ipa_reference): ...new pass_data instance and...
12618 (make_pass_ipa_reference): ...new function.
12619 * ipa-split.c (pass_split_functions): Convert from a global struct to
12620 a subclass of gimple_opt_pass along with...
12621 (pass_data_split_functions): ...new pass_data instance and...
12622 (make_pass_split_functions): ...new function.
12623 (pass_feedback_split_functions): Convert from a global struct to a
12624 subclass of gimple_opt_pass along with...
12625 (pass_data_feedback_split_functions): ...new pass_data instance and...
12626 (make_pass_feedback_split_functions): ...new function.
12627 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
12628 global struct to a subclass of simple_ipa_opt_pass along with...
12629 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
12631 (make_pass_ipa_function_and_variable_visibility): ...new function.
12632 (pass_ipa_free_inline_summary): Convert from a global struct to a
12633 subclass of simple_ipa_opt_pass along with...
12634 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
12635 (make_pass_ipa_free_inline_summary): ...new function.
12636 (pass_ipa_whole_program_visibility): Convert from a global struct to a
12637 subclass of ipa_opt_pass_d along with...
12638 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
12640 (make_pass_ipa_whole_program_visibility): ...new function.
12641 (pass_ipa_profile): Convert from a global struct to a subclass of
12642 ipa_opt_pass_d along with...
12643 (pass_data_ipa_profile): ...new pass_data instance and...
12644 (make_pass_ipa_profile): ...new function.
12645 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
12646 ipa_opt_pass_d along with...
12647 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
12648 (make_pass_ipa_cdtor_merge): ...new function.
12649 * ira.c (pass_ira): Convert from a global struct to a subclass of
12650 rtl_opt_pass along with...
12651 (pass_data_ira): ...new pass_data instance and...
12652 (make_pass_ira): ...new function.
12653 (pass_reload): Convert from a global struct to a subclass of
12654 rtl_opt_pass along with...
12655 (pass_data_reload): ...new pass_data instance and...
12656 (make_pass_reload): ...new function.
12657 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
12658 subclass of rtl_opt_pass along with...
12659 (pass_data_cleanup_barriers): ...new pass_data instance and...
12660 (make_pass_cleanup_barriers): ...new function.
12661 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
12662 of rtl_opt_pass along with...
12663 (pass_data_loop2): ...new pass_data instance and...
12664 (make_pass_loop2): ...new function.
12665 (pass_rtl_loop_init): Convert from a global struct to a subclass of
12666 rtl_opt_pass along with...
12667 (pass_data_rtl_loop_init): ...new pass_data instance and...
12668 (make_pass_rtl_loop_init): ...new function.
12669 (pass_rtl_loop_done): Convert from a global struct to a subclass of
12670 rtl_opt_pass along with...
12671 (pass_data_rtl_loop_done): ...new pass_data instance and...
12672 (make_pass_rtl_loop_done): ...new function.
12673 (pass_rtl_move_loop_invariants): Convert from a global struct to a
12674 subclass of rtl_opt_pass along with...
12675 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
12676 (make_pass_rtl_move_loop_invariants): ...new function.
12677 (pass_rtl_unswitch): Convert from a global struct to a subclass of
12678 rtl_opt_pass along with...
12679 (pass_data_rtl_unswitch): ...new pass_data instance and...
12680 (make_pass_rtl_unswitch): ...new function.
12681 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
12682 subclass of rtl_opt_pass along with...
12683 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
12685 (make_pass_rtl_unroll_and_peel_loops): ...new function.
12686 (pass_rtl_doloop): Convert from a global struct to a subclass of
12687 rtl_opt_pass along with...
12688 (pass_data_rtl_doloop): ...new pass_data instance and...
12689 (make_pass_rtl_doloop): ...new function.
12690 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
12691 a subclass of rtl_opt_pass along with...
12692 (pass_data_lower_subreg): ...new pass_data instance and...
12693 (make_pass_lower_subreg): ...new function.
12694 (pass_lower_subreg2): Convert from a global struct to a subclass of
12695 rtl_opt_pass along with...
12696 (pass_data_lower_subreg2): ...new pass_data instance and...
12697 (make_pass_lower_subreg2): ...new function.
12698 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
12699 struct to a subclass of ipa_opt_pass_d along with...
12700 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
12701 (make_pass_ipa_lto_gimple_out): ...new function.
12702 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
12703 of ipa_opt_pass_d along with...
12704 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
12705 (make_pass_ipa_lto_finish_out): ...new function.
12706 * mode-switching.c (pass_mode_switching): Convert from a global struct
12707 to a subclass of rtl_opt_pass along with...
12708 (pass_data_mode_switching): ...new pass_data instance and...
12709 (make_pass_mode_switching): ...new function.
12710 * modulo-sched.c (pass_sms): Convert from a global struct to a
12711 subclass of rtl_opt_pass along with...
12712 (pass_data_sms): ...new pass_data instance and...
12713 (make_pass_sms): ...new function.
12714 * omp-low.c (pass_expand_omp): Convert from a global struct to a
12715 subclass of gimple_opt_pass along with...
12716 (pass_data_expand_omp): ...new pass_data instance and...
12717 (make_pass_expand_omp): ...new function.
12718 (pass_lower_omp): Convert from a global struct to a subclass of
12719 gimple_opt_pass along with...
12720 (pass_data_lower_omp): ...new pass_data instance and...
12721 (make_pass_lower_omp): ...new function.
12722 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
12723 of gimple_opt_pass along with...
12724 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
12725 (make_pass_diagnose_omp_blocks): ...new function.
12726 * passes.c (pass_early_local_passes): Convert from a global struct to
12727 a subclass of simple_ipa_opt_pass along with...
12728 (pass_data_early_local_passes): ...new pass_data instance and...
12729 (make_pass_early_local_passes): ...new function.
12730 (pass_all_early_optimizations): Convert from a global struct to a
12731 subclass of gimple_opt_pass along with...
12732 (pass_data_all_early_optimizations): ...new pass_data instance and...
12733 (make_pass_all_early_optimizations): ...new function.
12734 (pass_all_optimizations): Convert from a global struct to a subclass
12735 of gimple_opt_pass along with...
12736 (pass_data_all_optimizations): ...new pass_data instance and...
12737 (make_pass_all_optimizations): ...new function.
12738 (pass_all_optimizations_g): Convert from a global struct to a subclass
12739 of gimple_opt_pass along with...
12740 (pass_data_all_optimizations_g): ...new pass_data instance and...
12741 (make_pass_all_optimizations_g): ...new function.
12742 (pass_rest_of_compilation): Convert from a global struct to a subclass
12743 of rtl_opt_pass along with...
12744 (pass_data_rest_of_compilation): ...new pass_data instance and...
12745 (make_pass_rest_of_compilation): ...new function.
12746 (pass_postreload): Convert from a global struct to a subclass of
12747 rtl_opt_pass along with...
12748 (pass_data_postreload): ...new pass_data instance and...
12749 (make_pass_postreload): ...new function.
12750 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
12751 subclass of rtl_opt_pass along with...
12752 (pass_data_gcse2): ...new pass_data instance and...
12753 (make_pass_gcse2): ...new function.
12754 * postreload.c (pass_postreload_cse): Convert from a global struct to
12755 a subclass of rtl_opt_pass along with...
12756 (pass_data_postreload_cse): ...new pass_data instance and...
12757 (make_pass_postreload_cse): ...new function.
12758 * predict.c (pass_profile): Convert from a global struct to a subclass
12759 of gimple_opt_pass along with...
12760 (pass_data_profile): ...new pass_data instance and...
12761 (make_pass_profile): ...new function.
12762 (pass_strip_predict_hints): Convert from a global struct to a subclass
12763 of gimple_opt_pass along with...
12764 (pass_data_strip_predict_hints): ...new pass_data instance and...
12765 (make_pass_strip_predict_hints): ...new function.
12766 * recog.c (pass_peephole2): Convert from a global struct to a subclass
12767 of rtl_opt_pass along with...
12768 (pass_data_peephole2): ...new pass_data instance and...
12769 (make_pass_peephole2): ...new function.
12770 (pass_split_all_insns): Convert from a global struct to a subclass of
12771 rtl_opt_pass along with...
12772 (pass_data_split_all_insns): ...new pass_data instance and...
12773 (make_pass_split_all_insns): ...new function.
12774 (pass_split_after_reload): Convert from a global struct to a subclass
12775 of rtl_opt_pass along with...
12776 (pass_data_split_after_reload): ...new pass_data instance and...
12777 (make_pass_split_after_reload): ...new function.
12778 (pass_split_before_regstack): Convert from a global struct to a
12779 subclass of rtl_opt_pass along with...
12780 (pass_data_split_before_regstack): ...new pass_data instance and...
12781 (make_pass_split_before_regstack): ...new function.
12782 (pass_split_before_sched2): Convert from a global struct to a subclass
12783 of rtl_opt_pass along with...
12784 (pass_data_split_before_sched2): ...new pass_data instance and...
12785 (make_pass_split_before_sched2): ...new function.
12786 (pass_split_for_shorten_branches): Convert from a global struct to a
12787 subclass of rtl_opt_pass along with...
12788 (pass_data_split_for_shorten_branches): ...new pass_data instance
12790 (make_pass_split_for_shorten_branches): ...new function.
12791 * ree.c (pass_ree): Convert from a global struct to a subclass of
12792 rtl_opt_pass along with...
12793 (pass_data_ree): ...new pass_data instance and...
12794 (make_pass_ree): ...new function.
12795 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
12796 subclass of rtl_opt_pass along with...
12797 (pass_data_stack_regs): ...new pass_data instance and...
12798 (make_pass_stack_regs): ...new function.
12799 (pass_stack_regs_run): Convert from a global struct to a subclass of
12800 rtl_opt_pass along with...
12801 (pass_data_stack_regs_run): ...new pass_data instance and...
12802 (make_pass_stack_regs_run): ...new function.
12803 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
12804 subclass of rtl_opt_pass along with...
12805 (pass_data_cprop_hardreg): ...new pass_data instance and...
12806 (make_pass_cprop_hardreg): ...new function.
12807 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
12808 subclass of rtl_opt_pass along with...
12809 (pass_data_reginfo_init): ...new pass_data instance and...
12810 (make_pass_reginfo_init): ...new function.
12811 * regmove.c (pass_regmove): Convert from a global struct to a subclass
12812 of rtl_opt_pass along with...
12813 (pass_data_regmove): ...new pass_data instance and...
12814 (make_pass_regmove): ...new function.
12815 * regrename.c (pass_regrename): Convert from a global struct to a
12816 subclass of rtl_opt_pass along with...
12817 (pass_data_regrename): ...new pass_data instance and...
12818 (make_pass_regrename): ...new function.
12819 * reorg.c (pass_delay_slots): Convert from a global struct to a
12820 subclass of rtl_opt_pass along with...
12821 (pass_data_delay_slots): ...new pass_data instance and...
12822 (make_pass_delay_slots): ...new function.
12823 (pass_machine_reorg): Convert from a global struct to a subclass of
12824 rtl_opt_pass along with...
12825 (pass_data_machine_reorg): ...new pass_data instance and...
12826 (make_pass_machine_reorg): ...new function.
12827 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
12828 of rtl_opt_pass along with...
12829 (pass_data_sched): ...new pass_data instance and...
12830 (make_pass_sched): ...new function.
12831 (pass_sched2): Convert from a global struct to a subclass of
12832 rtl_opt_pass along with...
12833 (pass_data_sched2): ...new pass_data instance and...
12834 (make_pass_sched2): ...new function.
12835 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
12836 to a subclass of rtl_opt_pass along with...
12837 (pass_data_stack_ptr_mod): ...new pass_data instance and...
12838 (make_pass_stack_ptr_mod): ...new function.
12839 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
12840 to a subclass of rtl_opt_pass along with...
12841 (pass_data_rtl_store_motion): ...new pass_data instance and...
12842 (make_pass_rtl_store_motion): ...new function.
12843 * tracer.c (pass_tracer): Convert from a global struct to a subclass
12844 of gimple_opt_pass along with...
12845 (pass_data_tracer): ...new pass_data instance and...
12846 (make_pass_tracer): ...new function.
12847 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
12848 to a subclass of gimple_opt_pass along with...
12849 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
12850 (make_pass_diagnose_tm_blocks): ...new function.
12851 (pass_lower_tm): Convert from a global struct to a subclass of
12852 gimple_opt_pass along with...
12853 (pass_data_lower_tm): ...new pass_data instance and...
12854 (make_pass_lower_tm): ...new function.
12855 (pass_tm_init): Convert from a global struct to a subclass of
12856 gimple_opt_pass along with...
12857 (pass_data_tm_init): ...new pass_data instance and...
12858 (make_pass_tm_init): ...new function.
12859 (pass_tm_mark): Convert from a global struct to a subclass of
12860 gimple_opt_pass along with...
12861 (pass_data_tm_mark): ...new pass_data instance and...
12862 (make_pass_tm_mark): ...new function.
12863 (pass_tm_edges): Convert from a global struct to a subclass of
12864 gimple_opt_pass along with...
12865 (pass_data_tm_edges): ...new pass_data instance and...
12866 (make_pass_tm_edges): ...new function.
12867 (pass_tm_memopt): Convert from a global struct to a subclass of
12868 gimple_opt_pass along with...
12869 (pass_data_tm_memopt): ...new pass_data instance and...
12870 (make_pass_tm_memopt): ...new function.
12871 (pass_ipa_tm): Convert from a global struct to a subclass of
12872 simple_ipa_opt_pass along with...
12873 (pass_data_ipa_tm): ...new pass_data instance and...
12874 (make_pass_ipa_tm): ...new function.
12875 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
12876 subclass of gimple_opt_pass along with...
12877 (pass_data_call_cdce): ...new pass_data instance and...
12878 (make_pass_call_cdce): ...new function.
12879 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
12880 subclass of gimple_opt_pass along with...
12881 (pass_data_build_cfg): ...new pass_data instance and...
12882 (make_pass_build_cfg): ...new function.
12883 (pass_split_crit_edges): Convert from a global struct to a subclass of
12884 gimple_opt_pass along with...
12885 (pass_data_split_crit_edges): ...new pass_data instance and...
12886 (make_pass_split_crit_edges): ...new function.
12887 (pass_warn_function_return): Convert from a global struct to a
12888 subclass of gimple_opt_pass along with...
12889 (pass_data_warn_function_return): ...new pass_data instance and...
12890 (make_pass_warn_function_return): ...new function.
12891 (pass_warn_function_noreturn): Convert from a global struct to a
12892 subclass of gimple_opt_pass along with...
12893 (pass_data_warn_function_noreturn): ...new pass_data instance and...
12894 (make_pass_warn_function_noreturn): ...new function.
12895 (pass_warn_unused_result): Convert from a global struct to a subclass
12896 of gimple_opt_pass along with...
12897 (pass_data_warn_unused_result): ...new pass_data instance and...
12898 (make_pass_warn_unused_result): ...new function.
12899 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
12900 a subclass of gimple_opt_pass along with...
12901 (pass_data_merge_phi): ...new pass_data instance and...
12902 (make_pass_merge_phi): ...new function.
12903 * tree-complex.c (pass_lower_complex): Convert from a global struct to
12904 a subclass of gimple_opt_pass along with...
12905 (pass_data_lower_complex): ...new pass_data instance and...
12906 (make_pass_lower_complex): ...new function.
12907 (pass_lower_complex_O0): Convert from a global struct to a subclass of
12908 gimple_opt_pass along with...
12909 (pass_data_lower_complex_O0): ...new pass_data instance and...
12910 (make_pass_lower_complex_O0): ...new function.
12911 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
12912 subclass of gimple_opt_pass along with...
12913 (pass_data_lower_eh): ...new pass_data instance and...
12914 (make_pass_lower_eh): ...new function.
12915 (pass_refactor_eh): Convert from a global struct to a subclass of
12916 gimple_opt_pass along with...
12917 (pass_data_refactor_eh): ...new pass_data instance and...
12918 (make_pass_refactor_eh): ...new function.
12919 (pass_lower_resx): Convert from a global struct to a subclass of
12920 gimple_opt_pass along with...
12921 (pass_data_lower_resx): ...new pass_data instance and...
12922 (make_pass_lower_resx): ...new function.
12923 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
12924 of gimple_opt_pass along with...
12925 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
12926 (make_pass_lower_eh_dispatch): ...new function.
12927 (pass_cleanup_eh): Convert from a global struct to a subclass of
12928 gimple_opt_pass along with...
12929 (pass_data_cleanup_eh): ...new pass_data instance and...
12930 (make_pass_cleanup_eh): ...new function.
12931 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
12932 to a subclass of simple_ipa_opt_pass along with...
12933 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
12934 (make_pass_ipa_lower_emutls): ...new function.
12935 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
12936 a subclass of gimple_opt_pass along with...
12937 (pass_data_if_conversion): ...new pass_data instance and...
12938 (make_pass_if_conversion): ...new function.
12939 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
12940 subclass of gimple_opt_pass along with...
12941 (pass_data_build_ssa): ...new pass_data instance and...
12942 (make_pass_build_ssa): ...new function.
12943 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
12944 global struct to a subclass of gimple_opt_pass along with...
12945 (pass_data_loop_distribution): ...new pass_data instance and...
12946 (make_pass_loop_distribution): ...new function.
12947 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
12948 subclass of gimple_opt_pass along with...
12949 (pass_data_mudflap_1): ...new pass_data instance and...
12950 (make_pass_mudflap_1): ...new function.
12951 (pass_mudflap_2): Convert from a global struct to a subclass of
12952 gimple_opt_pass along with...
12953 (pass_data_mudflap_2): ...new pass_data instance and...
12954 (make_pass_mudflap_2): ...new function.
12955 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
12956 subclass of gimple_opt_pass along with...
12957 (pass_data_mudflap_1): ...new pass_data instance and...
12958 (make_pass_mudflap_1): ...new function.
12959 (pass_mudflap_2): Convert from a global struct to a subclass of
12960 gimple_opt_pass along with...
12961 (pass_data_mudflap_2): ...new pass_data instance and...
12962 (make_pass_mudflap_2): ...new function.
12963 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
12964 gimple_opt_pass along with...
12965 (pass_data_nrv): ...new pass_data instance and...
12966 (make_pass_nrv): ...new function.
12967 (pass_return_slot): Convert from a global struct to a subclass of
12968 gimple_opt_pass along with...
12969 (pass_data_return_slot): ...new pass_data instance and...
12970 (make_pass_return_slot): ...new function.
12971 * tree-object-size.c (pass_object_sizes): Convert from a global struct
12972 to a subclass of gimple_opt_pass along with...
12973 (pass_data_object_sizes): ...new pass_data instance and...
12974 (make_pass_object_sizes): ...new function.
12975 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
12976 global struct to a subclass of gimple_opt_pass along with...
12977 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
12979 (make_pass_cleanup_cfg_post_optimizing): ...new function.
12980 (pass_fixup_cfg): Convert from a global struct to a subclass of
12981 gimple_opt_pass along with...
12982 (pass_data_fixup_cfg): ...new pass_data instance and...
12983 (make_pass_fixup_cfg): ...new function.
12984 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
12985 (make_pass_mudflap_1): ...new function.
12986 (pass_mudflap_2): Replace declaration with that of...
12987 (make_pass_mudflap_2): ...new function.
12988 (pass_asan): Replace declaration with that of...
12989 (make_pass_asan): ...new function.
12990 (pass_asan_O0): Replace declaration with that of...
12991 (make_pass_asan_O0): ...new function.
12992 (pass_tsan): Replace declaration with that of...
12993 (make_pass_tsan): ...new function.
12994 (pass_tsan_O0): Replace declaration with that of...
12995 (make_pass_tsan_O0): ...new function.
12996 (pass_lower_cf): Replace declaration with that of...
12997 (make_pass_lower_cf): ...new function.
12998 (pass_refactor_eh): Replace declaration with that of...
12999 (make_pass_refactor_eh): ...new function.
13000 (pass_lower_eh): Replace declaration with that of...
13001 (make_pass_lower_eh): ...new function.
13002 (pass_lower_eh_dispatch): Replace declaration with that of...
13003 (make_pass_lower_eh_dispatch): ...new function.
13004 (pass_lower_resx): Replace declaration with that of...
13005 (make_pass_lower_resx): ...new function.
13006 (pass_build_cfg): Replace declaration with that of...
13007 (make_pass_build_cfg): ...new function.
13008 (pass_early_tree_profile): Replace declaration with that of...
13009 (make_pass_early_tree_profile): ...new function.
13010 (pass_cleanup_eh): Replace declaration with that of...
13011 (make_pass_cleanup_eh): ...new function.
13012 (pass_sra): Replace declaration with that of...
13013 (make_pass_sra): ...new function.
13014 (pass_sra_early): Replace declaration with that of...
13015 (make_pass_sra_early): ...new function.
13016 (pass_early_ipa_sra): Replace declaration with that of...
13017 (make_pass_early_ipa_sra): ...new function.
13018 (pass_tail_recursion): Replace declaration with that of...
13019 (make_pass_tail_recursion): ...new function.
13020 (pass_tail_calls): Replace declaration with that of...
13021 (make_pass_tail_calls): ...new function.
13022 (pass_tree_loop): Replace declaration with that of...
13023 (make_pass_tree_loop): ...new function.
13024 (pass_tree_loop_init): Replace declaration with that of...
13025 (make_pass_tree_loop_init): ...new function.
13026 (pass_lim): Replace declaration with that of...
13027 (make_pass_lim): ...new function.
13028 (pass_tree_unswitch): Replace declaration with that of...
13029 (make_pass_tree_unswitch): ...new function.
13030 (pass_predcom): Replace declaration with that of...
13031 (make_pass_predcom): ...new function.
13032 (pass_iv_canon): Replace declaration with that of...
13033 (make_pass_iv_canon): ...new function.
13034 (pass_scev_cprop): Replace declaration with that of...
13035 (make_pass_scev_cprop): ...new function.
13036 (pass_empty_loop): Replace declaration with that of...
13037 (make_pass_empty_loop): ...new function.
13038 (pass_record_bounds): Replace declaration with that of...
13039 (make_pass_record_bounds): ...new function.
13040 (pass_graphite): Replace declaration with that of...
13041 (make_pass_graphite): ...new function.
13042 (pass_graphite_transforms): Replace declaration with that of...
13043 (make_pass_graphite_transforms): ...new function.
13044 (pass_if_conversion): Replace declaration with that of...
13045 (make_pass_if_conversion): ...new function.
13046 (pass_loop_distribution): Replace declaration with that of...
13047 (make_pass_loop_distribution): ...new function.
13048 (pass_vectorize): Replace declaration with that of...
13049 (make_pass_vectorize): ...new function.
13050 (pass_slp_vectorize): Replace declaration with that of...
13051 (make_pass_slp_vectorize): ...new function.
13052 (pass_complete_unroll): Replace declaration with that of...
13053 (make_pass_complete_unroll): ...new function.
13054 (pass_complete_unrolli): Replace declaration with that of...
13055 (make_pass_complete_unrolli): ...new function.
13056 (pass_parallelize_loops): Replace declaration with that of...
13057 (make_pass_parallelize_loops): ...new function.
13058 (pass_loop_prefetch): Replace declaration with that of...
13059 (make_pass_loop_prefetch): ...new function.
13060 (pass_iv_optimize): Replace declaration with that of...
13061 (make_pass_iv_optimize): ...new function.
13062 (pass_tree_loop_done): Replace declaration with that of...
13063 (make_pass_tree_loop_done): ...new function.
13064 (pass_ch): Replace declaration with that of...
13065 (make_pass_ch): ...new function.
13066 (pass_ccp): Replace declaration with that of...
13067 (make_pass_ccp): ...new function.
13068 (pass_phi_only_cprop): Replace declaration with that of...
13069 (make_pass_phi_only_cprop): ...new function.
13070 (pass_build_ssa): Replace declaration with that of...
13071 (make_pass_build_ssa): ...new function.
13072 (pass_build_alias): Replace declaration with that of...
13073 (make_pass_build_alias): ...new function.
13074 (pass_build_ealias): Replace declaration with that of...
13075 (make_pass_build_ealias): ...new function.
13076 (pass_dominator): Replace declaration with that of...
13077 (make_pass_dominator): ...new function.
13078 (pass_dce): Replace declaration with that of...
13079 (make_pass_dce): ...new function.
13080 (pass_dce_loop): Replace declaration with that of...
13081 (make_pass_dce_loop): ...new function.
13082 (pass_cd_dce): Replace declaration with that of...
13083 (make_pass_cd_dce): ...new function.
13084 (pass_call_cdce): Replace declaration with that of...
13085 (make_pass_call_cdce): ...new function.
13086 (pass_merge_phi): Replace declaration with that of...
13087 (make_pass_merge_phi): ...new function.
13088 (pass_split_crit_edges): Replace declaration with that of...
13089 (make_pass_split_crit_edges): ...new function.
13090 (pass_pre): Replace declaration with that of...
13091 (make_pass_pre): ...new function.
13092 (pass_profile): Replace declaration with that of...
13093 (make_pass_profile): ...new function.
13094 (pass_strip_predict_hints): Replace declaration with that of...
13095 (make_pass_strip_predict_hints): ...new function.
13096 (pass_lower_complex_O0): Replace declaration with that of...
13097 (make_pass_lower_complex_O0): ...new function.
13098 (pass_lower_complex): Replace declaration with that of...
13099 (make_pass_lower_complex): ...new function.
13100 (pass_lower_vector): Replace declaration with that of...
13101 (make_pass_lower_vector): ...new function.
13102 (pass_lower_vector_ssa): Replace declaration with that of...
13103 (make_pass_lower_vector_ssa): ...new function.
13104 (pass_lower_omp): Replace declaration with that of...
13105 (make_pass_lower_omp): ...new function.
13106 (pass_diagnose_omp_blocks): Replace declaration with that of...
13107 (make_pass_diagnose_omp_blocks): ...new function.
13108 (pass_expand_omp): Replace declaration with that of...
13109 (make_pass_expand_omp): ...new function.
13110 (pass_expand_omp_ssa): Replace declaration with that of...
13111 (make_pass_expand_omp_ssa): ...new function.
13112 (pass_object_sizes): Replace declaration with that of...
13113 (make_pass_object_sizes): ...new function.
13114 (pass_strlen): Replace declaration with that of...
13115 (make_pass_strlen): ...new function.
13116 (pass_fold_builtins): Replace declaration with that of...
13117 (make_pass_fold_builtins): ...new function.
13118 (pass_stdarg): Replace declaration with that of...
13119 (make_pass_stdarg): ...new function.
13120 (pass_early_warn_uninitialized): Replace declaration with that of...
13121 (make_pass_early_warn_uninitialized): ...new function.
13122 (pass_late_warn_uninitialized): Replace declaration with that of...
13123 (make_pass_late_warn_uninitialized): ...new function.
13124 (pass_cse_reciprocals): Replace declaration with that of...
13125 (make_pass_cse_reciprocals): ...new function.
13126 (pass_cse_sincos): Replace declaration with that of...
13127 (make_pass_cse_sincos): ...new function.
13128 (pass_optimize_bswap): Replace declaration with that of...
13129 (make_pass_optimize_bswap): ...new function.
13130 (pass_optimize_widening_mul): Replace declaration with that of...
13131 (make_pass_optimize_widening_mul): ...new function.
13132 (pass_warn_function_return): Replace declaration with that of...
13133 (make_pass_warn_function_return): ...new function.
13134 (pass_warn_function_noreturn): Replace declaration with that of...
13135 (make_pass_warn_function_noreturn): ...new function.
13136 (pass_cselim): Replace declaration with that of...
13137 (make_pass_cselim): ...new function.
13138 (pass_phiopt): Replace declaration with that of...
13139 (make_pass_phiopt): ...new function.
13140 (pass_forwprop): Replace declaration with that of...
13141 (make_pass_forwprop): ...new function.
13142 (pass_phiprop): Replace declaration with that of...
13143 (make_pass_phiprop): ...new function.
13144 (pass_tree_ifcombine): Replace declaration with that of...
13145 (make_pass_tree_ifcombine): ...new function.
13146 (pass_dse): Replace declaration with that of...
13147 (make_pass_dse): ...new function.
13148 (pass_nrv): Replace declaration with that of...
13149 (make_pass_nrv): ...new function.
13150 (pass_rename_ssa_copies): Replace declaration with that of...
13151 (make_pass_rename_ssa_copies): ...new function.
13152 (pass_sink_code): Replace declaration with that of...
13153 (make_pass_sink_code): ...new function.
13154 (pass_fre): Replace declaration with that of...
13155 (make_pass_fre): ...new function.
13156 (pass_check_data_deps): Replace declaration with that of...
13157 (make_pass_check_data_deps): ...new function.
13158 (pass_copy_prop): Replace declaration with that of...
13159 (make_pass_copy_prop): ...new function.
13160 (pass_vrp): Replace declaration with that of...
13161 (make_pass_vrp): ...new function.
13162 (pass_uncprop): Replace declaration with that of...
13163 (make_pass_uncprop): ...new function.
13164 (pass_return_slot): Replace declaration with that of...
13165 (make_pass_return_slot): ...new function.
13166 (pass_reassoc): Replace declaration with that of...
13167 (make_pass_reassoc): ...new function.
13168 (pass_rebuild_cgraph_edges): Replace declaration with that of...
13169 (make_pass_rebuild_cgraph_edges): ...new function.
13170 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
13171 (make_pass_remove_cgraph_callee_edges): ...new function.
13172 (pass_build_cgraph_edges): Replace declaration with that of...
13173 (make_pass_build_cgraph_edges): ...new function.
13174 (pass_local_pure_const): Replace declaration with that of...
13175 (make_pass_local_pure_const): ...new function.
13176 (pass_tracer): Replace declaration with that of...
13177 (make_pass_tracer): ...new function.
13178 (pass_warn_unused_result): Replace declaration with that of...
13179 (make_pass_warn_unused_result): ...new function.
13180 (pass_diagnose_tm_blocks): Replace declaration with that of...
13181 (make_pass_diagnose_tm_blocks): ...new function.
13182 (pass_lower_tm): Replace declaration with that of...
13183 (make_pass_lower_tm): ...new function.
13184 (pass_tm_init): Replace declaration with that of...
13185 (make_pass_tm_init): ...new function.
13186 (pass_tm_mark): Replace declaration with that of...
13187 (make_pass_tm_mark): ...new function.
13188 (pass_tm_memopt): Replace declaration with that of...
13189 (make_pass_tm_memopt): ...new function.
13190 (pass_tm_edges): Replace declaration with that of...
13191 (make_pass_tm_edges): ...new function.
13192 (pass_split_functions): Replace declaration with that of...
13193 (make_pass_split_functions): ...new function.
13194 (pass_feedback_split_functions): Replace declaration with that of...
13195 (make_pass_feedback_split_functions): ...new function.
13196 (pass_strength_reduction): Replace declaration with that of...
13197 (make_pass_strength_reduction): ...new function.
13198 (pass_ipa_lower_emutls): Replace declaration with that of...
13199 (make_pass_ipa_lower_emutls): ...new function.
13200 (pass_ipa_function_and_variable_visibility): Replace declaration with
13202 (make_pass_ipa_function_and_variable_visibility): ...new function.
13203 (pass_ipa_tree_profile): Replace declaration with that of...
13204 (make_pass_ipa_tree_profile): ...new function.
13205 (pass_early_local_passes): Replace declaration with that of...
13206 (make_pass_early_local_passes): ...new function.
13207 (pass_ipa_whole_program_visibility): Replace declaration with that
13209 (make_pass_ipa_whole_program_visibility): ...new function.
13210 (pass_ipa_lto_gimple_out): Replace declaration with that of...
13211 (make_pass_ipa_lto_gimple_out): ...new function.
13212 (pass_ipa_increase_alignment): Replace declaration with that of...
13213 (make_pass_ipa_increase_alignment): ...new function.
13214 (pass_ipa_inline): Replace declaration with that of...
13215 (make_pass_ipa_inline): ...new function.
13216 (pass_ipa_free_lang_data): Replace declaration with that of...
13217 (make_pass_ipa_free_lang_data): ...new function.
13218 (pass_ipa_free_inline_summary): Replace declaration with that of...
13219 (make_pass_ipa_free_inline_summary): ...new function.
13220 (pass_ipa_cp): Replace declaration with that of...
13221 (make_pass_ipa_cp): ...new function.
13222 (pass_ipa_reference): Replace declaration with that of...
13223 (make_pass_ipa_reference): ...new function.
13224 (pass_ipa_pure_const): Replace declaration with that of...
13225 (make_pass_ipa_pure_const): ...new function.
13226 (pass_ipa_pta): Replace declaration with that of...
13227 (make_pass_ipa_pta): ...new function.
13228 (pass_ipa_lto_finish_out): Replace declaration with that of...
13229 (make_pass_ipa_lto_finish_out): ...new function.
13230 (pass_ipa_tm): Replace declaration with that of...
13231 (make_pass_ipa_tm): ...new function.
13232 (pass_ipa_profile): Replace declaration with that of...
13233 (make_pass_ipa_profile): ...new function.
13234 (pass_ipa_cdtor_merge): Replace declaration with that of...
13235 (make_pass_ipa_cdtor_merge): ...new function.
13236 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
13238 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13239 (pass_init_datastructures): Replace declaration with that of...
13240 (make_pass_init_datastructures): ...new function.
13241 (pass_fixup_cfg): Replace declaration with that of...
13242 (make_pass_fixup_cfg): ...new function.
13243 (pass_expand): Replace declaration with that of...
13244 (make_pass_expand): ...new function.
13245 (pass_instantiate_virtual_regs): Replace declaration with that of...
13246 (make_pass_instantiate_virtual_regs): ...new function.
13247 (pass_rtl_fwprop): Replace declaration with that of...
13248 (make_pass_rtl_fwprop): ...new function.
13249 (pass_rtl_fwprop_addr): Replace declaration with that of...
13250 (make_pass_rtl_fwprop_addr): ...new function.
13251 (pass_jump): Replace declaration with that of...
13252 (make_pass_jump): ...new function.
13253 (pass_jump2): Replace declaration with that of...
13254 (make_pass_jump2): ...new function.
13255 (pass_lower_subreg): Replace declaration with that of...
13256 (make_pass_lower_subreg): ...new function.
13257 (pass_cse): Replace declaration with that of...
13258 (make_pass_cse): ...new function.
13259 (pass_fast_rtl_dce): Replace declaration with that of...
13260 (make_pass_fast_rtl_dce): ...new function.
13261 (pass_ud_rtl_dce): Replace declaration with that of...
13262 (make_pass_ud_rtl_dce): ...new function.
13263 (pass_rtl_dce): Replace declaration with that of...
13264 (make_pass_rtl_dce): ...new function.
13265 (pass_rtl_dse1): Replace declaration with that of...
13266 (make_pass_rtl_dse1): ...new function.
13267 (pass_rtl_dse2): Replace declaration with that of...
13268 (make_pass_rtl_dse2): ...new function.
13269 (pass_rtl_dse3): Replace declaration with that of...
13270 (make_pass_rtl_dse3): ...new function.
13271 (pass_rtl_cprop): Replace declaration with that of...
13272 (make_pass_rtl_cprop): ...new function.
13273 (pass_rtl_pre): Replace declaration with that of...
13274 (make_pass_rtl_pre): ...new function.
13275 (pass_rtl_hoist): Replace declaration with that of...
13276 (make_pass_rtl_hoist): ...new function.
13277 (pass_rtl_store_motion): Replace declaration with that of...
13278 (make_pass_rtl_store_motion): ...new function.
13279 (pass_cse_after_global_opts): Replace declaration with that of...
13280 (make_pass_cse_after_global_opts): ...new function.
13281 (pass_rtl_ifcvt): Replace declaration with that of...
13282 (make_pass_rtl_ifcvt): ...new function.
13283 (pass_into_cfg_layout_mode): Replace declaration with that of...
13284 (make_pass_into_cfg_layout_mode): ...new function.
13285 (pass_outof_cfg_layout_mode): Replace declaration with that of...
13286 (make_pass_outof_cfg_layout_mode): ...new function.
13287 (pass_loop2): Replace declaration with that of...
13288 (make_pass_loop2): ...new function.
13289 (pass_rtl_loop_init): Replace declaration with that of...
13290 (make_pass_rtl_loop_init): ...new function.
13291 (pass_rtl_move_loop_invariants): Replace declaration with that of...
13292 (make_pass_rtl_move_loop_invariants): ...new function.
13293 (pass_rtl_unswitch): Replace declaration with that of...
13294 (make_pass_rtl_unswitch): ...new function.
13295 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
13296 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13297 (pass_rtl_doloop): Replace declaration with that of...
13298 (make_pass_rtl_doloop): ...new function.
13299 (pass_rtl_loop_done): Replace declaration with that of...
13300 (make_pass_rtl_loop_done): ...new function.
13301 (pass_web): Replace declaration with that of...
13302 (make_pass_web): ...new function.
13303 (pass_cse2): Replace declaration with that of...
13304 (make_pass_cse2): ...new function.
13305 (pass_df_initialize_opt): Replace declaration with that of...
13306 (make_pass_df_initialize_opt): ...new function.
13307 (pass_df_initialize_no_opt): Replace declaration with that of...
13308 (make_pass_df_initialize_no_opt): ...new function.
13309 (pass_reginfo_init): Replace declaration with that of...
13310 (make_pass_reginfo_init): ...new function.
13311 (pass_inc_dec): Replace declaration with that of...
13312 (make_pass_inc_dec): ...new function.
13313 (pass_stack_ptr_mod): Replace declaration with that of...
13314 (make_pass_stack_ptr_mod): ...new function.
13315 (pass_initialize_regs): Replace declaration with that of...
13316 (make_pass_initialize_regs): ...new function.
13317 (pass_combine): Replace declaration with that of...
13318 (make_pass_combine): ...new function.
13319 (pass_if_after_combine): Replace declaration with that of...
13320 (make_pass_if_after_combine): ...new function.
13321 (pass_ree): Replace declaration with that of...
13322 (make_pass_ree): ...new function.
13323 (pass_partition_blocks): Replace declaration with that of...
13324 (make_pass_partition_blocks): ...new function.
13325 (pass_match_asm_constraints): Replace declaration with that of...
13326 (make_pass_match_asm_constraints): ...new function.
13327 (pass_regmove): Replace declaration with that of...
13328 (make_pass_regmove): ...new function.
13329 (pass_split_all_insns): Replace declaration with that of...
13330 (make_pass_split_all_insns): ...new function.
13331 (pass_fast_rtl_byte_dce): Replace declaration with that of...
13332 (make_pass_fast_rtl_byte_dce): ...new function.
13333 (pass_lower_subreg2): Replace declaration with that of...
13334 (make_pass_lower_subreg2): ...new function.
13335 (pass_mode_switching): Replace declaration with that of...
13336 (make_pass_mode_switching): ...new function.
13337 (pass_sms): Replace declaration with that of...
13338 (make_pass_sms): ...new function.
13339 (pass_sched): Replace declaration with that of...
13340 (make_pass_sched): ...new function.
13341 (pass_ira): Replace declaration with that of...
13342 (make_pass_ira): ...new function.
13343 (pass_reload): Replace declaration with that of...
13344 (make_pass_reload): ...new function.
13345 (pass_clean_state): Replace declaration with that of...
13346 (make_pass_clean_state): ...new function.
13347 (pass_branch_prob): Replace declaration with that of...
13348 (make_pass_branch_prob): ...new function.
13349 (pass_value_profile_transformations): Replace declaration with that
13351 (make_pass_value_profile_transformations): ...new function.
13352 (pass_postreload_cse): Replace declaration with that of...
13353 (make_pass_postreload_cse): ...new function.
13354 (pass_gcse2): Replace declaration with that of...
13355 (make_pass_gcse2): ...new function.
13356 (pass_split_after_reload): Replace declaration with that of...
13357 (make_pass_split_after_reload): ...new function.
13358 (pass_branch_target_load_optimize1): Replace declaration with that
13360 (make_pass_branch_target_load_optimize1): ...new function.
13361 (pass_thread_prologue_and_epilogue): Replace declaration with that
13363 (make_pass_thread_prologue_and_epilogue): ...new function.
13364 (pass_stack_adjustments): Replace declaration with that of...
13365 (make_pass_stack_adjustments): ...new function.
13366 (pass_peephole2): Replace declaration with that of...
13367 (make_pass_peephole2): ...new function.
13368 (pass_if_after_reload): Replace declaration with that of...
13369 (make_pass_if_after_reload): ...new function.
13370 (pass_regrename): Replace declaration with that of...
13371 (make_pass_regrename): ...new function.
13372 (pass_cprop_hardreg): Replace declaration with that of...
13373 (make_pass_cprop_hardreg): ...new function.
13374 (pass_reorder_blocks): Replace declaration with that of...
13375 (make_pass_reorder_blocks): ...new function.
13376 (pass_branch_target_load_optimize2): Replace declaration with that
13378 (make_pass_branch_target_load_optimize2): ...new function.
13379 (pass_leaf_regs): Replace declaration with that of...
13380 (make_pass_leaf_regs): ...new function.
13381 (pass_split_before_sched2): Replace declaration with that of...
13382 (make_pass_split_before_sched2): ...new function.
13383 (pass_compare_elim_after_reload): Replace declaration with that of...
13384 (make_pass_compare_elim_after_reload): ...new function.
13385 (pass_sched2): Replace declaration with that of...
13386 (make_pass_sched2): ...new function.
13387 (pass_stack_regs): Replace declaration with that of...
13388 (make_pass_stack_regs): ...new function.
13389 (pass_stack_regs_run): Replace declaration with that of...
13390 (make_pass_stack_regs_run): ...new function.
13391 (pass_df_finish): Replace declaration with that of...
13392 (make_pass_df_finish): ...new function.
13393 (pass_compute_alignments): Replace declaration with that of...
13394 (make_pass_compute_alignments): ...new function.
13395 (pass_duplicate_computed_gotos): Replace declaration with that of...
13396 (make_pass_duplicate_computed_gotos): ...new function.
13397 (pass_variable_tracking): Replace declaration with that of...
13398 (make_pass_variable_tracking): ...new function.
13399 (pass_free_cfg): Replace declaration with that of...
13400 (make_pass_free_cfg): ...new function.
13401 (pass_machine_reorg): Replace declaration with that of...
13402 (make_pass_machine_reorg): ...new function.
13403 (pass_cleanup_barriers): Replace declaration with that of...
13404 (make_pass_cleanup_barriers): ...new function.
13405 (pass_delay_slots): Replace declaration with that of...
13406 (make_pass_delay_slots): ...new function.
13407 (pass_split_for_shorten_branches): Replace declaration with that of...
13408 (make_pass_split_for_shorten_branches): ...new function.
13409 (pass_split_before_regstack): Replace declaration with that of...
13410 (make_pass_split_before_regstack): ...new function.
13411 (pass_convert_to_eh_region_ranges): Replace declaration with that
13413 (make_pass_convert_to_eh_region_ranges): ...new function.
13414 (pass_shorten_branches): Replace declaration with that of...
13415 (make_pass_shorten_branches): ...new function.
13416 (pass_set_nothrow_function_flags): Replace declaration with that of...
13417 (make_pass_set_nothrow_function_flags): ...new function.
13418 (pass_dwarf2_frame): Replace declaration with that of...
13419 (make_pass_dwarf2_frame): ...new function.
13420 (pass_final): Replace declaration with that of...
13421 (make_pass_final): ...new function.
13422 (pass_rtl_seqabstr): Replace declaration with that of...
13423 (make_pass_rtl_seqabstr): ...new function.
13424 (pass_release_ssa_names): Replace declaration with that of...
13425 (make_pass_release_ssa_names): ...new function.
13426 (pass_early_inline): Replace declaration with that of...
13427 (make_pass_early_inline): ...new function.
13428 (pass_inline_parameters): Replace declaration with that of...
13429 (make_pass_inline_parameters): ...new function.
13430 (pass_update_address_taken): Replace declaration with that of...
13431 (make_pass_update_address_taken): ...new function.
13432 (pass_convert_switch): Replace declaration with that of...
13433 (make_pass_convert_switch): ...new function.
13434 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
13435 to a subclass of simple_ipa_opt_pass along with...
13436 (pass_data_ipa_tree_profile): ...new pass_data instance and...
13437 (make_pass_ipa_tree_profile): ...new function.
13438 * tree-sra.c (pass_sra_early): Convert from a global struct to a
13439 subclass of gimple_opt_pass along with...
13440 (pass_data_sra_early): ...new pass_data instance and...
13441 (make_pass_sra_early): ...new function.
13442 (pass_sra): Convert from a global struct to a subclass of
13443 gimple_opt_pass along with...
13444 (pass_data_sra): ...new pass_data instance and...
13445 (make_pass_sra): ...new function.
13446 (pass_early_ipa_sra): Convert from a global struct to a subclass of
13447 gimple_opt_pass along with...
13448 (pass_data_early_ipa_sra): ...new pass_data instance and...
13449 (make_pass_early_ipa_sra): ...new function.
13450 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
13451 subclass of gimple_opt_pass along with...
13452 (pass_data_ccp): ...new pass_data instance and...
13453 (make_pass_ccp): ...new function.
13454 (pass_fold_builtins): Convert from a global struct to a subclass of
13455 gimple_opt_pass along with...
13456 (pass_data_fold_builtins): ...new pass_data instance and...
13457 (make_pass_fold_builtins): ...new function.
13458 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
13459 subclass of gimple_opt_pass along with...
13460 (pass_data_copy_prop): ...new pass_data instance and...
13461 (make_pass_copy_prop): ...new function.
13462 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
13463 global struct to a subclass of gimple_opt_pass along with...
13464 (pass_data_rename_ssa_copies): ...new pass_data instance and...
13465 (make_pass_rename_ssa_copies): ...new function.
13466 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
13467 subclass of gimple_opt_pass along with...
13468 (pass_data_dce): ...new pass_data instance and...
13469 (make_pass_dce): ...new function.
13470 (pass_dce_loop): Convert from a global struct to a subclass of
13471 gimple_opt_pass along with...
13472 (pass_data_dce_loop): ...new pass_data instance and...
13473 (make_pass_dce_loop): ...new function.
13474 (pass_cd_dce): Convert from a global struct to a subclass of
13475 gimple_opt_pass along with...
13476 (pass_data_cd_dce): ...new pass_data instance and...
13477 (make_pass_cd_dce): ...new function.
13478 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
13479 subclass of gimple_opt_pass along with...
13480 (pass_data_dominator): ...new pass_data instance and...
13481 (make_pass_dominator): ...new function.
13482 (pass_phi_only_cprop): Convert from a global struct to a subclass of
13483 gimple_opt_pass along with...
13484 (pass_data_phi_only_cprop): ...new pass_data instance and...
13485 (make_pass_phi_only_cprop): ...new function.
13486 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
13487 subclass of gimple_opt_pass along with...
13488 (pass_data_dse): ...new pass_data instance and...
13489 (make_pass_dse): ...new function.
13490 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
13491 a subclass of gimple_opt_pass along with...
13492 (pass_data_forwprop): ...new pass_data instance and...
13493 (make_pass_forwprop): ...new function.
13494 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
13495 struct to a subclass of gimple_opt_pass along with...
13496 (pass_data_tree_ifcombine): ...new pass_data instance and...
13497 (make_pass_tree_ifcombine): ...new function.
13498 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
13499 subclass of gimple_opt_pass along with...
13500 (pass_data_ch): ...new pass_data instance and...
13501 (make_pass_ch): ...new function.
13502 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
13503 subclass of gimple_opt_pass along with...
13504 (pass_data_tree_loop): ...new pass_data instance and...
13505 (make_pass_tree_loop): ...new function.
13506 (pass_tree_loop_init): Convert from a global struct to a subclass of
13507 gimple_opt_pass along with...
13508 (pass_data_tree_loop_init): ...new pass_data instance and...
13509 (make_pass_tree_loop_init): ...new function.
13510 (pass_lim): Convert from a global struct to a subclass of
13511 gimple_opt_pass along with...
13512 (pass_data_lim): ...new pass_data instance and...
13513 (make_pass_lim): ...new function.
13514 (pass_tree_unswitch): Convert from a global struct to a subclass of
13515 gimple_opt_pass along with...
13516 (pass_data_tree_unswitch): ...new pass_data instance and...
13517 (make_pass_tree_unswitch): ...new function.
13518 (pass_predcom): Convert from a global struct to a subclass of
13519 gimple_opt_pass along with...
13520 (pass_data_predcom): ...new pass_data instance and...
13521 (make_pass_predcom): ...new function.
13522 (pass_vectorize): Convert from a global struct to a subclass of
13523 gimple_opt_pass along with...
13524 (pass_data_vectorize): ...new pass_data instance and...
13525 (make_pass_vectorize): ...new function.
13526 (pass_graphite): Convert from a global struct to a subclass of
13527 gimple_opt_pass along with...
13528 (pass_data_graphite): ...new pass_data instance and...
13529 (make_pass_graphite): ...new function.
13530 (pass_graphite_transforms): Convert from a global struct to a subclass
13531 of gimple_opt_pass along with...
13532 (pass_data_graphite_transforms): ...new pass_data instance and...
13533 (make_pass_graphite_transforms): ...new function.
13534 (pass_check_data_deps): Convert from a global struct to a subclass of
13535 gimple_opt_pass along with...
13536 (pass_data_check_data_deps): ...new pass_data instance and...
13537 (make_pass_check_data_deps): ...new function.
13538 (pass_iv_canon): Convert from a global struct to a subclass of
13539 gimple_opt_pass along with...
13540 (pass_data_iv_canon): ...new pass_data instance and...
13541 (make_pass_iv_canon): ...new function.
13542 (pass_scev_cprop): Convert from a global struct to a subclass of
13543 gimple_opt_pass along with...
13544 (pass_data_scev_cprop): ...new pass_data instance and...
13545 (make_pass_scev_cprop): ...new function.
13546 (pass_record_bounds): Convert from a global struct to a subclass of
13547 gimple_opt_pass along with...
13548 (pass_data_record_bounds): ...new pass_data instance and...
13549 (make_pass_record_bounds): ...new function.
13550 (pass_complete_unroll): Convert from a global struct to a subclass of
13551 gimple_opt_pass along with...
13552 (pass_data_complete_unroll): ...new pass_data instance and...
13553 (make_pass_complete_unroll): ...new function.
13554 (pass_complete_unrolli): Convert from a global struct to a subclass of
13555 gimple_opt_pass along with...
13556 (pass_data_complete_unrolli): ...new pass_data instance and...
13557 (make_pass_complete_unrolli): ...new function.
13558 (pass_parallelize_loops): Convert from a global struct to a subclass
13559 of gimple_opt_pass along with...
13560 (pass_data_parallelize_loops): ...new pass_data instance and...
13561 (make_pass_parallelize_loops): ...new function.
13562 (pass_loop_prefetch): Convert from a global struct to a subclass of
13563 gimple_opt_pass along with...
13564 (pass_data_loop_prefetch): ...new pass_data instance and...
13565 (make_pass_loop_prefetch): ...new function.
13566 (pass_iv_optimize): Convert from a global struct to a subclass of
13567 gimple_opt_pass along with...
13568 (pass_data_iv_optimize): ...new pass_data instance and...
13569 (make_pass_iv_optimize): ...new function.
13570 (pass_tree_loop_done): Convert from a global struct to a subclass of
13571 gimple_opt_pass along with...
13572 (pass_data_tree_loop_done): ...new pass_data instance and...
13573 (make_pass_tree_loop_done): ...new function.
13574 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
13575 struct to a subclass of gimple_opt_pass along with...
13576 (pass_data_cse_reciprocals): ...new pass_data instance and...
13577 (make_pass_cse_reciprocals): ...new function.
13578 (pass_cse_sincos): Convert from a global struct to a subclass of
13579 gimple_opt_pass along with...
13580 (pass_data_cse_sincos): ...new pass_data instance and...
13581 (make_pass_cse_sincos): ...new function.
13582 (pass_optimize_bswap): Convert from a global struct to a subclass of
13583 gimple_opt_pass along with...
13584 (pass_data_optimize_bswap): ...new pass_data instance and...
13585 (make_pass_optimize_bswap): ...new function.
13586 (pass_optimize_widening_mul): Convert from a global struct to a
13587 subclass of gimple_opt_pass along with...
13588 (pass_data_optimize_widening_mul): ...new pass_data instance and...
13589 (make_pass_optimize_widening_mul): ...new function.
13590 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
13591 subclass of gimple_opt_pass along with...
13592 (pass_data_phiopt): ...new pass_data instance and...
13593 (make_pass_phiopt): ...new function.
13594 (pass_cselim): Convert from a global struct to a subclass of
13595 gimple_opt_pass along with...
13596 (pass_data_cselim): ...new pass_data instance and...
13597 (make_pass_cselim): ...new function.
13598 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
13599 subclass of gimple_opt_pass along with...
13600 (pass_data_phiprop): ...new pass_data instance and...
13601 (make_pass_phiprop): ...new function.
13602 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
13603 subclass of gimple_opt_pass along with...
13604 (pass_data_pre): ...new pass_data instance and...
13605 (make_pass_pre): ...new function.
13606 (pass_fre): Convert from a global struct to a subclass of
13607 gimple_opt_pass along with...
13608 (pass_data_fre): ...new pass_data instance and...
13609 (make_pass_fre): ...new function.
13610 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
13611 subclass of gimple_opt_pass along with...
13612 (pass_data_reassoc): ...new pass_data instance and...
13613 (make_pass_reassoc): ...new function.
13614 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
13615 subclass of gimple_opt_pass along with...
13616 (pass_data_sink_code): ...new pass_data instance and...
13617 (make_pass_sink_code): ...new function.
13618 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
13619 subclass of gimple_opt_pass along with...
13620 (pass_data_strlen): ...new pass_data instance and...
13621 (make_pass_strlen): ...new function.
13622 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
13623 struct to a subclass of gimple_opt_pass along with...
13624 (pass_data_build_alias): ...new pass_data instance and...
13625 (make_pass_build_alias): ...new function.
13626 (pass_build_ealias): Convert from a global struct to a subclass of
13627 gimple_opt_pass along with...
13628 (pass_data_build_ealias): ...new pass_data instance and...
13629 (make_pass_build_ealias): ...new function.
13630 (pass_ipa_pta): Convert from a global struct to a subclass of
13631 simple_ipa_opt_pass along with...
13632 (pass_data_ipa_pta): ...new pass_data instance and...
13633 (make_pass_ipa_pta): ...new function.
13634 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
13635 subclass of gimple_opt_pass along with...
13636 (pass_data_uncprop): ...new pass_data instance and...
13637 (make_pass_uncprop): ...new function.
13638 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
13639 global struct to a subclass of gimple_opt_pass along with...
13640 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
13641 (make_pass_late_warn_uninitialized): ...new function.
13642 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
13643 to a subclass of gimple_opt_pass along with...
13644 (pass_data_init_datastructures): ...new pass_data instance and...
13645 (make_pass_init_datastructures): ...new function.
13646 (pass_early_warn_uninitialized): Convert from a global struct to a
13647 subclass of gimple_opt_pass along with...
13648 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
13649 (make_pass_early_warn_uninitialized): ...new function.
13650 (pass_update_address_taken): Convert from a global struct to a
13651 subclass of gimple_opt_pass along with...
13652 (pass_data_update_address_taken): ...new pass_data instance and...
13653 (make_pass_update_address_taken): ...new function.
13654 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
13655 struct to a subclass of gimple_opt_pass along with...
13656 (pass_data_release_ssa_names): ...new pass_data instance and...
13657 (make_pass_release_ssa_names): ...new function.
13658 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
13659 subclass of gimple_opt_pass along with...
13660 (pass_data_stdarg): ...new pass_data instance and...
13661 (make_pass_stdarg): ...new function.
13662 * tree-switch-conversion.c (pass_convert_switch): Convert from a
13663 global struct to a subclass of gimple_opt_pass along with...
13664 (pass_data_convert_switch): ...new pass_data instance and...
13665 (make_pass_convert_switch): ...new function.
13666 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
13667 to a subclass of gimple_opt_pass along with...
13668 (pass_data_tail_recursion): ...new pass_data instance and...
13669 (make_pass_tail_recursion): ...new function.
13670 (pass_tail_calls): Convert from a global struct to a subclass of
13671 gimple_opt_pass along with...
13672 (pass_data_tail_calls): ...new pass_data instance and...
13673 (make_pass_tail_calls): ...new function.
13674 * tree-vect-generic.c (pass_lower_vector): Convert from a global
13675 struct to a subclass of gimple_opt_pass along with...
13676 (pass_data_lower_vector): ...new pass_data instance and...
13677 (make_pass_lower_vector): ...new function.
13678 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
13679 gimple_opt_pass along with...
13680 (pass_data_lower_vector_ssa): ...new pass_data instance and...
13681 (make_pass_lower_vector_ssa): ...new function.
13682 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
13683 to a subclass of gimple_opt_pass along with...
13684 (pass_data_slp_vectorize): ...new pass_data instance and...
13685 (make_pass_slp_vectorize): ...new function.
13686 (pass_ipa_increase_alignment): Convert from a global struct to a
13687 subclass of simple_ipa_opt_pass along with...
13688 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
13689 (make_pass_ipa_increase_alignment): ...new function.
13690 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
13691 gimple_opt_pass along with...
13692 (pass_data_vrp): ...new pass_data instance and...
13693 (make_pass_vrp): ...new function.
13694 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
13695 subclass of simple_ipa_opt_pass along with...
13696 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
13697 (make_pass_ipa_free_lang_data): ...new function.
13698 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
13699 gimple_opt_pass along with...
13700 (pass_data_tsan): ...new pass_data instance and...
13701 (make_pass_tsan): ...new function.
13702 (pass_tsan_O0): Convert from a global struct to a subclass of
13703 gimple_opt_pass along with...
13704 (pass_data_tsan_O0): ...new pass_data instance and...
13705 (make_pass_tsan_O0): ...new function.
13706 * var-tracking.c (pass_variable_tracking): Convert from a global
13707 struct to a subclass of rtl_opt_pass along with...
13708 (pass_data_variable_tracking): ...new pass_data instance and...
13709 (make_pass_variable_tracking): ...new function.
13710 * web.c (pass_web): Convert from a global struct to a subclass of
13711 rtl_opt_pass along with...
13712 (pass_data_web): ...new pass_data instance and...
13713 (make_pass_web): ...new function.
13714 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
13715 declaration with that of...
13716 (make_pass_mode_switch_use): ...new function.
13717 (pass_resolve_sw_modes): Replace declaration with that of...
13718 (make_pass_resolve_sw_modes): ...new function.
13719 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
13720 from a global struct to a subclass of rtl_opt_pass along with...
13721 (pass_data_mode_switch_use): ...new pass_data instance and...
13722 (make_pass_mode_switch_use): ...new function.
13723 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
13724 from a global struct to a subclass of rtl_opt_pass along with...
13725 (pass_data_resolve_sw_modes): ...new pass_data instance and...
13726 (make_pass_resolve_sw_modes): ...new function.
13727 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
13728 struct to a subclass of rtl_opt_pass along with...
13729 (pass_data_insert_vzeroupper): ...new pass_data instance and...
13730 (make_pass_insert_vzeroupper): ...new function.
13731 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
13732 global struct to a subclass of rtl_opt_pass along with...
13733 (pass_data_work_around_errata): ...new pass_data instance and...
13734 (make_pass_work_around_errata): ...new function.
13735 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
13736 struct to a subclass of rtl_opt_pass along with...
13737 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
13738 (make_pass_mips_machine_reorg2): ...new function.
13740 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13742 * passes.c (pass_manager::operator new): New.
13744 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13746 Handwritten part of conversion of passes to C++ classes.
13748 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
13749 (toplev.o): Add dep on PASS_MANAGER_H.
13750 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
13751 of early local pases to reflect this moving from a global to a
13752 member of gcc::pass_manager.
13753 (cgraph_add_new_function): Likewise.
13754 * lto-cgraph.c (lto_output_node): Update for conversion of
13755 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
13756 * passes.c (opt_pass::clone): New.
13757 (opt_pass::gate): New.
13758 (opt_pass::execute): New.
13759 (opt_pass::opt_pass): New.
13760 (pass_manager::execute_early_local_passes): New.
13761 (pass_manager::execute_pass_mode_switching): new.
13762 (finish_optimization_passes): Convert to...
13763 (pass_manager::finish_optimization_passes): ...this.
13764 (finish_optimization_passes): Update for conversion of passes to
13766 (register_dump_files_1): Use has_gate since we cannot portably
13767 check a vtable entry against NULL.
13768 (dump_one_pass): Likewise.
13769 (ipa_write_summaries_2): Likewise.
13770 (ipa_write_optimization_summaries_1): Likewise.
13771 (ipa_read_summaries_1): Likewise.
13772 (ipa_read_optimization_summaries_1): Likewise.
13773 (execute_ipa_stmt_fixups): Likewise.
13774 (pass_manager::pass_manager): Rewrite pass-creation, invoking
13775 pass-creation functions rather than wiring up globals, and
13776 storing the results in fields of pass_manager generated using
13777 pass-instances.def.
13778 (pass_manager::dump_profile_report): Update for conversion of
13779 passes to C++ classes.
13780 (pass_manager::execute_ipa_summary_passes): Likewise.
13781 (execute_one_ipa_transform_pass): Likewise.
13782 (execute_one_pass): Use has_gate and has_execute since we cannot
13783 portably check a vtable entry against NULL.
13784 * pass_manager.h (pass_manager::finish_optimization_passes): New.
13785 (pass_manager): Use pass-instances.def to add fields for the
13786 various pass instances.
13787 * toplev.c (finalize): Update for move of
13788 finish_optimization_passes to a method of gcc::pass_manager.
13789 * toplev.h (finish_optimization_passes): Move to method of class
13791 * tree-pass.h (struct pass_data): New.
13792 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
13793 (opt_pass::gate): Convert to virtual function.
13794 (opt_pass::~opt_pass): New.
13795 (opt_pass::clone): New.
13796 (opt_pass::execute): Convert to virtual function.
13797 (opt_pass::opt_pass): New.
13798 (opt_pass::ctxt_): new.
13799 (gimple_opt_pass): Convert to subclass of opt_pass.
13800 (gimple_opt_pass::gimple_opt_pass): New.
13801 (rtl_opt_pass): Convert to subclass of opt_pass.
13802 (rtl_opt_pass::rtl_opt_pass): New.
13803 (ipa_opt_pass_d): Convert to subclass of opt_pass.
13804 (ipa_opt_pass_d::ipa_opt_pass_d): New.
13805 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
13806 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
13807 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
13808 invocation of pass_mode_switching to reflect this moving from a
13809 global to a member of gcc::pass_manager.
13810 (ix86_option_override): Rework how pass_insert_vzeroupper is
13811 added to the pass_manager to reflect autogenerated changes.
13812 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
13814 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
13816 PR rtl-optimization/57708
13817 * recog.c (peep2_find_free_register): Validate all regs in a
13820 2013-08-05 Jan Hubicka <jh@suse.cz>
13823 * cgraph.c (verify_cgraph_node): Accept local flags from other
13825 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
13826 (function_and_variable_visibility): Likewise.
13827 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
13829 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13831 * graph.c (init_graph_slim_pretty_print): Remove.
13832 (print_graph_cfg): Do not call it. Use local pretty printer.
13833 (start_graph_dump): Likewise.
13835 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13837 * gimple-pretty-print.c (buffer): Remove.
13838 (initialized): Likewise.
13839 (maybe_init_pretty_print): Likewise.
13840 (print_gimple_stmt): Do not call it. Use non-static local
13841 pretty_printer variable.
13842 (print_gimple_expr): Likewise.
13843 (print_gimple_seq): Likewise.
13844 (gimple_dump_bb): Likewise.
13846 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13848 * asan.c (asan_pp): Remove.
13849 (asan_pp_initialized): Likewise.
13850 (asan_pp_initialize): Likewise.
13851 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
13852 (asan_emit_stack_protection): Tidy. Use local pretty printer.
13853 (asan_add_global): Likewise.
13855 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
13857 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
13858 * diagnostic.h (diagnostic_flush_buffer): Adjust.
13859 * pretty-print.c (pp_formatted_text_data): Likewise.
13860 (pp_indent): Rename from pp_base_indent.
13861 (pp_format): Rename from pp_base_format.
13862 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
13863 (pp_format_verbatim): Rename from pp_base_format_verbatim.
13864 (pp_flush): Rename from pp_base_flush.
13865 (pp_set_line_maximum_length): Rename from
13866 pp_base_set_line_maximum_length.
13867 (pp_clear_output_area): Rename from pp_base_clear_output_area.
13868 (pp_set_prefix): Rename from pp_base_set_prefix.
13869 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
13870 (pp_emit_prefix): Rename from pp_base_emit_prefix.
13871 (pp_append_text): Rename from pp_base_append_text.
13872 (pp_formatted_text): Rename from pp_base_formatted_text.
13873 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
13874 (pp_remaining_character_count_for_line): Rename from
13875 pp_base_remaining_character_count_for_line.
13876 (pp_newline): Rename from pp_base_newline.
13877 (pp_character): Rename from pp_base_character.
13878 (pp_string): Rename from pp_base_string.
13879 (pp_maybe_space): Rename from pp_base_maybe_space.
13880 * asan.c (asan_pp_string): Adjust.
13881 (asan_emit_stack_protection): Likewise.
13882 (asan_add_global): Likewise.
13883 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
13884 * tree-mudflap.c (mf_varname_tree): Likewise.
13885 * tree-pretty-print.c (pp_tree_identifier): Rename from
13886 pp_base_tree_identifier.
13887 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
13888 Declare as function.
13890 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13892 * pretty-print.h (pp_bar_bar): New.
13893 (pp_ampersand_ampersand): Likewise.
13894 (pp_less_equal): Likewise.
13895 (pp_greater_equal): Likewise.
13896 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
13897 printer functions instead of pp_string or operators and punctuators.
13898 (dump_gimple_call): Likewise.
13899 (dump_gimple_omp_for): Likewise.
13900 (dump_gimple_transaction): Likewise.
13901 (dump_gimple_phi): Likewise.
13902 (pp_gimple_stmt_1): Likewise.
13903 * sched-vis.c (print_insn): Likewise.
13904 * tree-mudflap.c (mf_varname_tree): Likewise.
13905 * tree-pretty-print.c (dump_block_node): Likewise.
13906 (dump_generic_node): Likewise.
13908 2013-08-02 Jan Hubicka <jh@suse.cz>
13910 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
13912 * lto-streamer-out.c (tree_is_indexable): Results decls and
13913 parm decls are not indexable.
13914 (DFS_write_tree_body): Do not follow args and results.
13915 (hash_tree): Likewise.
13916 (output_functions): Rearrange so struct function is needed
13917 only when real body is output; be able to also ouptut abstract
13918 functions; output DECL_ARGUMENTS and DECL_RESULT.
13919 (lto_output): When not in WPA, ale store abstract functions.
13920 (write_symbol): Do not care about RESULT_DECL.
13921 (output_symbol_p): Handle correctly sbtract decls.
13922 * lto-streamer-in.c (input_function): Rearrange so struct
13923 function can be NULL at entry; allow streaming of
13924 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
13925 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
13926 sanity check during LTO.
13927 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
13928 RESULT_DECl and DECL_ARGUMENTS.
13929 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
13932 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13934 * pretty-print.h (pp_underscore): New.
13936 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
13937 printer functions instead of pp_character.
13938 (dump_binary_rhs): Likewise.
13939 (dump_ternary_rhs): Likewise.
13940 (dump_gimple_call_args): Likewise.
13941 (pp_points_to_solution): Likewise.
13942 (dump_gimple_call): Likewise.
13943 (dump_gimple_switch): Likewise.
13944 (dump_gimple_cond): Likewise.
13945 (dump_gimple_bind): Likewise.
13946 (dump_gimple_try): Likewise.
13947 (dump_gimple_omp_for): Likewise.
13948 (dump_gimple_omp_continue): Likewise.
13949 (dump_gimple_omp_single): Likewise.
13950 (dump_gimple_omp_sections): Likewise.
13951 (dump_gimple_omp_block): Likewise.
13952 (dump_gimple_omp_critical): Likewise.
13953 (dump_gimple_transaction): Likewise.
13954 (dump_gimple_asm): Likewise.
13955 (dump_gimple_phi): Likewise.
13956 (dump_gimple_omp_parallel): Likewise.
13957 (dump_gimple_omp_task): Likewise.
13958 (dump_gimple_omp_atomic_load): Likewise.
13959 (dump_gimple_omp_atomic_store): Likewise.
13960 (dump_gimple_mem_ops): Likewise.
13961 (pp_gimple_stmt_1): Likewise.
13962 (pp_cfg_jump): Likewise.
13963 (dump_implicit_edges): Likewise.
13964 (gimple_dump_bb_for_graph): Likewise.
13965 * graph.c (draw_cfg_node): Likewise.
13966 * langhooks.c (lhd_print_error_function): Likewise.
13967 * sched-vis.c (print_exp): Likewise.
13968 (print_value): Likewise.
13969 (print_pattern): Likewise.
13970 (print_insn): Likewise.
13971 (rtl_dump_bb_for_graph): Likewise.
13972 * tree-pretty-print.c (dump_function_declaration): Likewise.
13973 (dump_array_domain): Likewise.
13974 (dump_omp_clause): Likewise.
13975 (dump_location): Likewise.
13976 (dump_generic_node): Likewise.
13977 (print_struct_decl): Likewise.
13978 * diagnostic.c (diagnostic_show_locus): Use pp_space.
13980 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
13982 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
13983 candidate table when replacing a candidate statement.
13984 (replace_rhs_if_not_dup): Likewise.
13985 (replace_one_candidate): Likewise.
13987 2013-08-02 Jan Hubicka <jh@suse.cz>
13988 Martin Liska <marxin.liska@gmail.com>
13990 * cgraphunit.c (add_new_function): Fix logic when adding from
13992 (assemble_thunk): Rename to ...
13993 (expand_thunk); .. this one; export; get it working with
13994 general functions; make produced gimple valid.
13995 * cgraph.h (expand_thunk): Declare.
13997 2013-08-02 Jan Hubicka <jh@suse.cz>
13999 * ipa-cp.c (gather_context_independent_values): Use
14000 ipa_get_param_move_cost.
14001 (get_replacement_map): Remove PARAM; move parameter folding
14003 (create_specialized_node): Update.
14004 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
14005 assert that we have gimple body; update move_cost.
14006 (count_formal_params): Assert that we have gimple body.
14007 (ipa_dump_param): New function.
14008 (ipa_alloc_node_params): Break out from ...
14009 (ipa_initialize_node_params): ... here.
14010 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
14011 (ipa_write_node_info): Stream move costs.
14012 (ipa_read_node_info): Read move costs.
14013 (ipa_update_after_lto_read): Do not recompute node params.
14014 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
14015 (ipa_get_param): Check we are not in WPA.
14016 (ipa_get_param_move_cost): New.
14017 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
14018 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
14019 parm numbers to be present.
14021 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14023 PR rtl-optimization/58048
14024 * lra-constraints.c (process_alt_operands): Don't check asm
14025 operand on register.
14027 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
14029 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
14030 the implied StoreLoad barrier for atomic operations if before.
14032 2013-08-02 Jan Hubicka <jh@suse.cz>
14033 Martin Liska <marxin.liska@gmail.com>
14035 * cgraph.c (cgraph_function_body_availability): Do not check
14037 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
14038 symtab_node_availability): Declare.
14039 * ipa.c (can_replace_by_local_alias): New.
14040 (function_and_variable_visibility): Use it.
14041 * symtab.c (symtab_for_node_and_aliases,
14042 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
14044 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14046 PR rtl-optimization/57963
14047 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
14048 (lra_constraints): Use them.
14050 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
14052 * config/arm/types.md (define_attr "type"): Add "load_acq"
14054 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
14056 (cortex_a53_store1): Likewise.
14058 2013-08-01 Jan Hubicka <jh@suse.cz>
14060 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
14061 partitions are not needed.
14063 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
14065 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
14066 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
14067 MAYBE_NON_Q_CLASS_P where appropriate.
14069 2013-08-01 Jan Hubicka <jh@suse.cz>
14071 * cgraph.h (release_function_body): Declare.
14072 * tree.c (free_lang_data_in_decl): Free, parameters and return values
14073 of unused delcarations.
14075 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14077 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
14078 RTL form when subtracting a constant.
14080 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14082 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
14083 Generate canonical plus rtx with negated immediate instead of minus
14085 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
14087 2013-08-01 Jan Hubicka <jh@suse.cz>
14089 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
14090 (cgraph_release_function_body): Likewise.
14091 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
14092 * cgraph.h (cgrpah_node): Rename abstract_and_needed
14093 to used_as_abstract_origin.
14094 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
14095 symbols used as abstract origins.
14096 * cgraphunit.c (analyze_functions): Update.
14097 * ipa.c (symtab_remove_unreachable_nodes): Recompute
14098 used_as_abstract_origin.
14099 * tree-inline.c (tree_function_versioning): Update
14100 used_as_abstract_origin; be ready for DECL_RESULT and
14101 DECL_ARGUMENTS to be NULL.
14103 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
14104 for abstract functions.
14105 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
14108 2013-08-01 Jan Hubicka <jh@suse.cz>
14110 * profile.c (compute_value_histograms): Fix thinko.
14112 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14114 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
14115 aarch-common-protos.h to extra_headers.
14116 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14117 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
14118 * config/aarch64/t-aarch64 (aarch-common.o): Define.
14120 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14122 * config/aarch64/aarch64.md (define_attr "type"): Delete.
14123 Include "../arm/types.md". Define "type" attribute for all patterns.
14124 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
14127 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
14129 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
14130 to support power8 load fusion.
14131 (fusion_gpr_mem_load): Likewise.
14133 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
14135 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
14136 declarations for power8 load fusion.
14137 (emit_fusion_gpr_load): Likewise.
14139 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
14140 tuning for power8, turn on fusion mode by default. Turn on sign
14141 extending fusion mode if normal fusion mode is on, and we are at
14143 (fusion_gpr_load_p): New function, return true if we can fuse an
14144 addis instruction with a dependent load to a GPR.
14145 (emit_fusion_gpr_load): Emit the instructions for power8 load
14148 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
14149 (VSX load fusion peepholes): New peepholes to fuse together an
14150 addi instruction with a VSX load instruction.
14152 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
14153 peepholes to fuse an addis instruction with a load to a GPR base
14154 register. If we are supporting sign extending fusions, convert
14155 sign extending loads to zero extending loads and add an explicit
14158 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14160 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
14161 aarch-common-protos.h to extra_headers.
14162 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14163 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
14164 (arm_early_store_addr_dep): Likewise.
14165 (arm_no_early_alu_shift_dep): Likewise.
14166 (arm_no_early_alu_shift_value_dep): Likewise.
14167 (arm_no_early_mul_dep): Likewise.
14168 (arm_no_early_store_addr_dep): Likewise.
14169 (arm_mac_accumulator_is_mul_result): Likewise.
14170 (arm_mac_accumulator_is_result): Likewise.
14171 * config/arm/aarch-common.c: ... here. New file.
14172 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
14174 (arm_early_store_addr_dep): Likewise.
14175 (arm_no_early_alu_shift_dep): Likewise.
14176 (arm_no_early_alu_shift_value_dep): Likewise.
14177 (arm_no_early_mul_dep): Likewise.
14178 (arm_no_early_store_addr_dep): Likewise.
14179 (arm_mac_accumulator_is_mul_result): Likewise.
14180 (arm_mac_accumulator_is_result): Likewise.
14181 * config/arm/aarch-common-protos.h: ... here. New file.
14182 * config/arm/t-arm (aarch-common.o): Define.
14184 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14186 * config/arm/arm.md: Include new file "types.md".
14187 (define_attr "type"): Move from here to ...
14188 (define_attr "mul32"): Likewise.
14189 (define_attr "mul64"): Likewise.
14190 * config/arm/types.md: ... here. New file.
14192 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
14194 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
14195 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
14197 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
14199 * gen-pass-instances.awk: Fix offset of substr().
14201 2013-07-31 David Malcolm <dmalcolm@redhat.com>
14203 * Makefile.in (pass-instances.def): New.
14204 (passes.o): Replace dependency on passes.def with one on
14207 * gen-pass-instances.awk: New.
14209 * passes.c (pass_manager::pass_manager): Use pass-instances.def
14210 rather than passes.def, updating local definition of NEXT_PASS
14211 macro to add an extra NUM parameter (currently unused).
14213 2013-07-30 David Malcolm <dmalcolm@redhat.com>
14215 * Makefile.in (PASS_MANAGER_H): New.
14216 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
14217 (passes.o): Likewise.
14218 (statistics.o): Likewise.
14219 (cgraphunit.o): Likewise.
14220 (context.o): Depend on PASS_MANAGER_H.
14222 * pass_manager.h: New.
14224 * cgraphunit.c (cgraph_add_new_function): Update for moves
14225 of globals to fields of pass_manager.
14226 (analyze_function): Likewise.
14227 (expand_function): Likewise.
14228 (ipa_passes): Likewise.
14229 (compile): Likewise.
14231 * context.c (context::context): New.
14232 * context.h (context::context): New.
14233 (context::get_passes): New.
14234 (context::passes_): New.
14236 * lto-cgraph.c (input_node): Update for moves of globals to
14237 fields of pass_manager.
14239 * passes.c (all_passes): Remove, in favor of a field of the
14240 same name within the new class pass_manager.
14241 (all_small_ipa_passes): Likewise.
14242 (all_lowering_passes): Likewise.
14243 (all_regular_ipa_passes): Likewise.
14244 (all_late_ipa_passes): Likewise.
14245 (all_lto_gen_passes): Likewise.
14246 (passes_by_id): Likewise.
14247 (passes_by_id_size): Likewise.
14248 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14249 the new class pass_manager.
14250 (set_pass_for_id): Convert to...
14251 (pass_manager::set_pass_for_id): ...method.
14252 (get_pass_for_id): Convert to...
14253 (pass_manager::get_pass_for_id): ...method.
14254 (register_one_dump_file): Move body of implementation into...
14255 (pass_manager::register_one_dump_file): ...here.
14256 (register_dump_files_1): Convert to...
14257 (pass_manager::register_dump_files_1): ...method.
14258 (register_dump_files): Convert to...
14259 (pass_manager::register_dump_files): ...method.
14260 (create_pass_tab): Update for moves of globals to fields of
14262 (dump_passes): Move body of implementation into...
14263 (pass_manager::dump_passes): ...here.
14264 (register_pass): Move body of implementation into...
14265 (pass_manager::register_pass): ...here.
14266 (init_optimization_passes): Convert into...
14267 (pass_manager::pass_manager): ...constructor for new
14268 pass_manager class, and initialize the pass_lists array.
14269 (check_profile_consistency): Update for moves of globals to
14270 fields of pass_manager.
14271 (dump_profile_report): Move body of implementation into...
14272 (pass_manager::dump_profile_report): ...here.
14273 (ipa_write_summaries_1): Update for moves of pass lists from
14274 being globals to fields of pass_manager.
14275 (ipa_write_optimization_summaries): Likewise.
14276 (ipa_read_summaries): Likewise.
14277 (ipa_read_optimization_summaries): Likewise.
14278 (execute_all_ipa_stmt_fixups): Likewise.
14280 * statistics.c (statistics_fini): Update for moves of globals to
14281 fields of pass_manager.
14283 * toplev.c (general_init): Replace call to
14284 init_optimization_passes with construction of the pass_manager
14287 * tree-pass.h (all_passes): Remove, in favor of a field of the
14288 same name within the new class pass_manager.
14289 (all_small_ipa_passes): Likewise.
14290 (all_lowering_passes): Likewise.
14291 (all_regular_ipa_passes): Likewise.
14292 (all_lto_gen_passes): Likewise.
14293 (all_late_ipa_passes): Likewise.
14294 (passes_by_id): Likewise.
14295 (passes_by_id_size): Likewise.
14296 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14297 the new class pass_manager.
14298 (get_pass_for_id): Remove.
14300 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14302 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
14305 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14307 * arm.md (mulhi3): New expand pattern.
14309 2013-07-30 Jan Hubicka <jh@suse.cz>
14310 Martin Liska <marxin.liska@gmail.com>
14312 * profile.c (compute_value_histograms): Do not ICE when
14313 there is mismatch only on some counters.
14315 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14317 PR rtl-optimization/57637
14318 * function.c (move_insn_for_shrink_wrap): Also check the
14319 GEN set of the LIVE problem for the liveness analysis
14320 if it exists, otherwise give up.
14322 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14324 PR tree-optimization/57993
14325 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
14326 replaced statement in the candidate table.
14327 (phi_add_costs): Return infinite cost when the hidden basis does
14328 not dominate all phis on which the candidate is dependent.
14329 (replace_one_candidate): Record replaced statement in the
14332 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
14334 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
14335 (ashlv2si3): New expander.
14336 (*ashlv2si3_i): New define_insn_and_split.
14337 * predicates.md (float_operation): Allow patterns with three
14338 basic sub-patterns.
14340 PR rtl-optimization/58021
14341 * mode-switching.c (create_pre_exit): Always split off preceding
14342 insns if we are not at the basic block head.
14344 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
14346 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
14347 (UCLIBC_DYNAMIC_LINKER): New macro.
14348 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
14350 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
14351 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
14353 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
14354 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
14355 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
14356 for SF and DF modes. Use ieee_quad_format for TF mode.
14357 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
14358 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
14359 (mips_option_override): Handle `-mnan=legacy'.
14360 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
14361 `-mabs=2008' and `-mnan=2008'.
14362 (OPTION_DEFAULT_SPECS): Add "nan" default.
14363 (ASM_SPEC): Handle `-mnan='.
14364 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
14365 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
14366 comment accordingly.
14367 (neg<mode>2): Likewise.
14368 * config/mips/mips.opt (mabs, mnan): New options.
14369 * doc/install.texi (Configuration): Document `--with-nan=' option.
14370 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
14372 (MIPS Options): Document them.
14373 * config.gcc <mips*-*-*>: Handle `--with-nan='.
14374 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
14375 * configure: Regenerate.
14376 * config.in: Regenerate.
14378 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14380 * config/i386/i386.md (float post-reload splitters): Do not check
14381 for subregs of SSE registers.
14383 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14384 H.J. Lu <hongjiu.lu@intel.com>
14388 * config/i386/i386.md (post-reload splitter
14389 to avoid partial SSE reg dependency stalls): New pattern.
14391 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
14393 * config/s390/s390.md ("movcc"): Swap load and store instructions.
14395 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
14397 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14398 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
14400 2013-07-26 Cary Coutant <ccoutant@google.com>
14402 * dwarf2out.c (die_checksum_ordered): Don't include template
14403 instantiations in signature.
14404 (is_template_parameter): New function.
14405 (is_template_instantiation): New function.
14406 (generate_skeleton_bottom_up): Don't include template instantiations
14408 (generate_skeleton): Likewise.
14409 (break_out_comdat_types): Move recursive call to break out nested
14411 (prune_unused_types_mark_generic_parms_dies): Call
14412 is_template_parameter.
14414 2013-07-26 Ian Bolton <ian.bolton@arm.com>
14416 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
14417 uses vector registers.
14418 * config/aarch64/iterators.md: Add attributes rtn and vas.
14420 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14421 Richard Earnshaw <richard.earnshaw@arm.com>
14423 * combine.c (simplify_comparison): Re-canonicalize operands
14425 * config/arm/arm.md (movcond_addsi): New splitter.
14427 2013-07-25 Sterling Augustine <saugustine@google.com>
14429 * dwarf2out.c (size_of_pubnames): Move code to...
14430 (include_pubname_in_output): ...here. New.
14431 (want_pubnames): Rearrange.
14432 (output_pubnames): Call include_pubname_in_output. Move assertion.
14434 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14436 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
14438 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14441 * doc/extend.texi: Remove obsolete builtins. Fix
14442 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
14444 2013-07-25 Jan Hubicka <jh@suse.cz>
14446 * cgraph.c (release_function_body): Break out from ...
14447 (cgraph_release_function_body): ... this one; also release DECL_RESULT
14448 and DECL_ARGUMENTS.
14449 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
14450 old_tree in the map.
14451 (create_specialized_node): Update.
14452 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
14454 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
14455 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
14456 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
14457 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
14460 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14462 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
14463 addsi3_carryin_alt2_<optab>): Correct output template.
14465 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14467 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
14468 Adjust for arm_restrict_it.
14469 Remove trailing whitespace.
14471 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
14473 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
14474 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
14476 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
14478 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
14480 PR rtl-optimization/57960
14481 * lra-constraints.c (process_alt_operands): Use the right mode
14482 when checking strict_low.
14484 2013-07-25 Jan Hubicka <jh@suse.cz>
14486 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
14487 * cgraph.c (cgraph_remove_node): Do not release function body
14488 when in cgraph streaming.
14489 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
14490 in other partitions are not considered reachable; fix handling of
14493 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14495 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
14497 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14502 * config/arm/arm.md ("*sibcall_insn): Replace use of
14503 Ss with US. Adjust output for v5 and v4t.
14504 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
14506 * config/arm/constraints.md ("Ss"): Rename to US.
14508 2013-07-25 Terry Guo <terry.guo@arm.com>
14510 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
14511 shift_add/shift_sub0/shift_sub1 RTXs.
14513 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
14514 Anton Blanchard <anton@au1.ibm.com>
14516 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
14517 (altivec_vpks<VI_char>ss): Likewise.
14518 (altivec_vpks<VI_char>us): Likewise.
14519 (altivec_vpku<VI_char>us): Likewise.
14520 (altivec_vpku<VI_char>um): Likewise.
14522 2013-07-24 David Malcolm <dmalcolm@redhat.com>
14524 Introduce context class.
14526 * Makefile.in (CONTEXT_H): New.
14527 (OBJS): Add context.o.
14528 (toplev.o): Add CONTEXT_H to dependencies.
14531 * toplev.c (general_init): Create the singleton gcc::context instance.
14537 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
14539 PR rtl-optimization/57968
14540 * mode-switching.c (create_pre_exit): Allow instructions that
14541 don't set a return register to need a non-exit mode.
14543 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14544 Anton Blanchard <anton@au1.ibm.com>
14546 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
14547 operands to vperm for little endian.
14548 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
14549 of lvsl to create the control mask for a vperm for little endian.
14551 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14552 Anton Blanchard <anton@au1.ibm.com>
14554 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
14555 two operands for little-endian.
14557 2013-07-23 Steve Ellcey <sellcey@mips.com>
14559 * config/mips/mips.c (mips_case_values_threshold): New.
14560 (TARGET_CASE_VALUES_THRESHOLD): Define.
14562 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14563 Anton Blanchard <anton@au1.ibm.com>
14565 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
14566 selection of field for vector splat in little endian mode.
14568 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
14570 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
14571 expanders to rs6000.md.
14572 (ior<mode>3): Likewise.
14573 (and<mode>3): Likewise.
14574 (one_cmpl<mode>2): Likewise.
14575 (nor<mode>3): Likewise.
14576 (andc<mode>3): Likewise.
14577 (eqv<mode>3): Likewise.
14578 (nand<mode>3): Likewise.
14579 (orc<mode>3): Likewise.
14581 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
14584 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
14585 to split multi-word logical operations.
14586 (rs6000_split_logical_di): Likewise.
14587 (rs6000_split_logical): Likewise.
14589 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
14590 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
14591 and allow TImode operations in 32-bit.
14592 (vsx_and<mode>3_64bit): Likewise.
14593 (vsx_ior<mode>3_32bit): Likewise.
14594 (vsx_ior<mode>3_64bit): Likewise.
14595 (vsx_xor<mode>3_32bit): Likewise.
14596 (vsx_xor<mode>3_64bit): Likewise.
14597 (vsx_one_cmpl<mode>2_32bit): Likewise.
14598 (vsx_one_cmpl<mode>2_64bit): Likewise.
14599 (vsx_nor<mode>3_32bit): Likewise.
14600 (vsx_nor<mode>3_64bit): Likewise.
14601 (vsx_andc<mode>3_32bit): Likewise.
14602 (vsx_andc<mode>3_64bit): Likewise.
14603 (vsx_eqv<mode>3_32bit): Likewise.
14604 (vsx_eqv<mode>3_64bit): Likewise.
14605 (vsx_nand<mode>3_32bit): Likewise.
14606 (vsx_nand<mode>3_64bit): Likewise.
14607 (vsx_orc<mode>3_32bit): Likewise.
14608 (vsx_orc<mode>3_64bit): Likewise.
14610 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
14611 logical types in GPRs.
14613 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
14614 logical insns to rs6000.md, and allow TImode operations in
14616 (altivec_ior<mode>3): Likewise.
14617 (altivec_xor<mode>3): Likewise.
14618 (altivec_one_cmpl<mode>2): Likewise.
14619 (altivec_nor<mode>3): Likewise.
14620 (altivec_andc<mode>3): Likewise.
14622 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
14623 attributes for moving the 128-bit logical operations into
14625 (BOOL_REGS_OUTPUT): Likewise.
14626 (BOOL_REGS_OP1): Likewise.
14627 (BOOL_REGS_OP2): Likewise.
14628 (BOOL_REGS_UNARY): Likewise.
14629 (BOOL_REGS_AND_CR0): Likewise.
14630 (one_cmpl<mode>2): Add support for DI logical operations on
14631 32-bit, splitting the operations to 32-bit.
14632 (anddi3): Likewise.
14633 (iordi3): Likewise.
14634 (xordi3): Likewise.
14635 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
14636 changes to combine the 32/64-bit code, allow logical operations on
14637 TI mode in 32-bit, and to use similar match_operator patterns like
14638 scalar mode uses. Combine the Altivec and VSX code for logical
14639 operations, and move it here.
14640 (ior<mode>3, 128-bit types): Likewise.
14641 (xor<mode>3, 128-bit types): Likewise.
14642 (one_cmpl<mode>3, 128-bit types): Likewise.
14643 (nor<mode>3, 128-bit types): Likewise.
14644 (andc<mode>3, 128-bit types): Likewise.
14645 (eqv<mode>3, 128-bit types): Likewise.
14646 (nand<mode>3, 128-bit types): Likewise.
14647 (orc<mode>3, 128-bit types): Likewise.
14648 (and<mode>3_internal): Likewise.
14649 (bool<mode>3_internal): Likewise.
14650 (boolc<mode>3_internal1): Likewise.
14651 (boolc<mode>3_internal2): Likewise.
14652 (boolcc<mode>3_internal1): Likewise.
14653 (boolcc<mode>3_internal2): Likewise.
14654 (eqv<mode>3_internal1): Likewise.
14655 (eqv<mode>3_internal2): Likewise.
14656 (one_cmpl1<mode>3_internal): Likewise.
14658 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14660 * config/microblaze/microblaze.c (microblaze_expand_prologue):
14661 Rename flag_stack_usage to flag_stack_usage_info.
14663 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14665 * config/microblaze/sync.md: New file.
14666 * config/microblaze/microblaze.md: Include sync.md
14667 * config/microblaze/microblaze.c: Add print_operand 'y'.
14668 * config/microblaze/constraints.md: Add memory_contraint
14669 'Q' which is a single register.
14671 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
14673 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
14675 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14677 * reload.c (find_reloads): Exit loop once we find this operand
14678 cannot be reloaded somehow for this alternative.
14680 * reload.c (find_reloads): Exit loop once we find a hard register.
14682 * rtlanal.c (computed_jump_p): Exit loop once we find label
14685 * i386.c (ix86_pad_returns): Exit loop after setting replace.
14687 * cfgloopmanip.c (remove_path): Exit loop after setting
14690 * gensupport.c (subst_dup): Avoid loop if code is not
14691 MATCH_DUP nor MATCH_OP_DUP.
14693 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
14695 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
14697 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14699 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
14700 true for SP_REGNUM if mode == ptr_mode.
14701 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
14702 with value R0_REGNUM + 31.
14704 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14706 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
14707 pad pointer-typed argument downward.
14709 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14711 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
14712 and __ILP32__ when the ILP32 model is in use.
14714 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14716 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
14717 (aarch64_load_symref_appropriately): In the case of
14718 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
14719 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
14720 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
14721 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
14722 if 'mode' doesn't equal to 'ptr_mode'.
14723 (aarch64_output_mi_thunk): Add an assertion on the alignment of
14724 'vcall_offset'; change to call aarch64_emit_move differently depending
14725 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
14726 to calculate the upper bound of 'vcall_offset'.
14727 (aarch64_cannot_force_const_mem): Change to also return true if
14729 (aarch64_legitimize_reload_address): In the case of large
14730 displacements, add new local variable 'xmode' and an assertion
14731 based on it; change to use 'xmode' to generate the new rtx and
14733 (aarch64_asm_trampoline_template): Change to generate the template
14734 differently depending on TARGET_ILP32 or not; change to use
14735 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
14736 (aarch64_trampoline_size): Removed.
14737 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
14738 and replace immediate literals with it. Change to use 'ptr_mode'
14739 instead of 'DImode' and call convert_memory_address if the mode
14740 of 'fnaddr' doesn't equal to 'ptr_mode'.
14741 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
14743 (aarch64_elf_asm_destructor): Likewise.
14744 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
14745 on TARGET_ILP32 instead of aarch64_trampoline_size.
14746 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
14747 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
14748 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
14749 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
14750 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
14751 (storewb_pair<GPI:mode>_<P:mode>): ... this.
14752 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
14753 depending on the value of 'mode'.
14754 (add_losym_<mode>): New.
14755 (ldr_got_small_<mode>): New, based on ldr_got_small.
14756 (ldr_got_small): Remove.
14757 (ldr_got_small_sidi): New.
14758 * config/aarch64/iterators.md (P): New.
14759 (PTR): Change to 'ptr_mode' in the condition.
14761 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14763 * config.gcc (aarch64*-*-*): Support --with-abi.
14764 (aarch64*-*-elf): Support --with-multilib-list.
14765 (aarch64*-*-linux*): Likewise.
14766 (supported_defaults): Add abi to aarch64*-*-*.
14767 * configure.ac: Mention AArch64 for --with-multilib-list.
14768 * configure: Re-generated.
14769 * config/aarch64/biarchilp32.h: New file.
14770 * config/aarch64/biarchlp64.h: New file.
14771 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
14773 (MULTILIB_DEFAULTS): Ditto.
14774 (DRIVER_SELF_SPECS): Ditto.
14775 (ASM_SPEC): Update to also substitute -mabi.
14776 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
14777 file whose name depends on -mabi= and -mbig-endian.
14778 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
14780 (POINTER_SIZE): New define.
14781 (POINTERS_EXTEND_UNSIGNED): Ditto.
14782 (enum aarch64_abi_type): New enumeration tag.
14783 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
14784 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
14785 (TARGET_ILP32): New define.
14786 * config/aarch64/aarch64.opt (mabi): New.
14787 (aarch64_abi): New.
14788 (ilp32, lp64): New values for -mabi.
14789 * config/aarch64/t-aarch64 (comma): New define.
14790 (MULTILIB_OPTIONS): Ditto.
14791 (MULTILIB_DIRNAMES): Ditto.
14792 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
14793 * doc/invoke.texi: Document -mabi for AArch64.
14795 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
14797 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
14799 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14800 Anton Blanchard <anton@au1.ibm.com>
14802 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
14803 endianness when selecting field to splat.
14805 2013-07-22 Eric Christopher <echristo@gmail.com>
14807 * dwarf2out.c (die_odr_checksum): New function to use
14808 CHECKSUM_ macros and ULEB128 for DIE tag.
14809 (generate_type_signature): Use.
14811 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
14813 * config.gcc (sparc*-*-*): Accept leon3 processor.
14814 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
14815 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
14816 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
14817 * config/sparc/sparc.opt (enum processor_type): Add leon3.
14818 (mfix-ut699): Adjust comment.
14819 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
14820 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
14821 (CPP_CPU_SPEC): Likewise.
14822 (ASM_CPU_SPEC): Likewise.
14823 * config/sparc/sparc.c (leon3_cost): New constant.
14824 (sparc_option_override): Add leon3 support.
14825 (mem_ref): New function.
14826 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
14827 (sparc_do_work_around_errata): Look into the instruction in the delay
14828 slot and adjust accordingly. Add fix for the data cache nullify issues
14829 of the UT699. Change insertion position for the NOP.
14830 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
14831 (leon3_load): New reservation.
14832 (leon_store): Bump latency to 2.
14833 (grfpu): New automaton.
14834 (grfpu_alu): New unit.
14835 (grfpu_ds): Likewise.
14836 (leon_fp_alu): Adjust.
14837 (leon_fp_mult): Delete.
14838 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
14839 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
14840 * config/sparc/sparc.md (cpu): Add leon3.
14841 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
14842 (swapsi): Likewise.
14843 (atomic_test_and_set): Likewise.
14844 (ldstub): Likewise.
14846 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
14848 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
14849 default for R5900 targets.
14850 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
14851 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
14852 * config/mips/mips.c (mips_option_override): Report an error for
14853 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
14854 for -march=r5900 -mhard-float.
14856 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14858 * df-problems.c (can_move_insns_across): Exit loop once we
14859 find a non-fixed, non-global register.
14861 * ipa-pure-const.c (propagate_nothrow): Exit loop after
14864 * omega.c (omega_eliminate_red): Break after setting red_found.
14865 (omega_problem_has_red_equations): Similarly after setting found.
14866 (omega_query_variable): Similarly after setting coupled.
14868 2013-07-22 Marek Polacek <polacek@redhat.com>
14870 * gimplify.c: Don't include gimple.h twice.
14872 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14874 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
14875 instead of TARGET_THUMB1.
14876 (Pz): New constraint.
14877 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
14879 (compare_negsi_si): Likewise.
14880 (compare_addsi2_op0): Likewise.
14881 (compare_addsi2_op1): Likewise.
14882 (addsi3_carryin_<optab>): Likewise.
14883 (addsi3_carryin_alt2_<optab>): Likewise.
14884 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
14885 for arm_restrict_it.
14886 (subsi3_carryin): Likewise.
14887 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
14888 (minmax_arithsi): Disable for arm_restrict_it.
14889 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
14890 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
14891 (satsi_<SAT:code>_shift): Likewise.
14892 (arm_shiftsi3): Add alternative for 16-bit encoding.
14893 (arm32_movhf): Disable for arm_restrict_it.
14894 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
14895 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
14897 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
14899 * config/arm/arm.md (attribute "insn"): Delete.
14900 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
14901 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
14902 (not_shiftsi): Update for attribute change.
14903 (not_shiftsi_compare0): Likewise.
14904 (not_shiftsi_compare0_scratch): Likewise.
14905 (arm_one_cmplsi2): Likewise.
14906 (thumb1_one_cmplsi2): Likewise.
14907 (notsi_compare0): Likewise.
14908 (notsi_compare0_scratch): Likewise.
14909 (thumb1_movdi_insn): Likewise.
14910 (arm_movsi_insn): Likewise.
14911 (movhi_insn_arch4): Likewise.
14912 (movhi_bytes): Likewise.
14913 (arm_movqi_insn): Likewise.
14914 (thumb1_movqi_insn): Likewise.
14915 (arm32_movhf): Likewise.
14916 (thumb1_movhf): Likewise.
14917 (arm_movsf_soft_insn): Likewise.
14918 (thumb1_movsf_insn): Likewise.
14919 (thumb_movdf_insn): Likewise.
14920 (movsicc_insn): Likewise.
14921 (movsfcc_soft_insn): Likewise.
14922 (and_scc): Likewise.
14923 (cond_move): Likewise.
14924 (if_move_not): Likewise.
14925 (if_not_move): Likewise.
14926 (if_shift_move): Likewise.
14927 (if_move_shift): Likewise.
14928 (if_shift_shift): Likewise.
14929 (if_not_arith): Likewise.
14930 (if_arith_not): Likewise.
14931 (cond_move_not): Likewise.
14932 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
14933 (neon_mov<mode>): Likewise.
14934 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
14935 (thumb2_movsi_vfp): Likewise.
14936 (movsf_vfp): Likewise.
14937 (thumb2_movsf_vfp): Likewise.
14938 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
14940 (cortexa7_older_only): Likewise.
14941 (cortexa7_younger): Likewise.
14942 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
14943 (1020alu_shift_op): Likewise.
14944 (1020alu_shift_reg_op): Likewise.
14945 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
14946 (alu_shift_op): Likewise.
14947 (alu_shift_reg_op): Likewise.
14948 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
14949 (11_alu_shift_op): Likewise.
14950 (11_alu_shift_reg_op): Likewise.
14951 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
14952 (9_alu_shift_reg_op): Likewise.
14953 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
14955 (cortex_a15_alu_shift): Likewise.
14956 (cortex_a15_alu_shift_reg): Likewise.
14957 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
14958 (cortex_a5_alu_shift): Likewise.
14959 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
14961 (cortex_a53_alu_shift): Likewise.
14962 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
14964 (cortex_a7_alu_reg): Likewise.
14965 (cortex_a7_alu_shift): Likewise.
14966 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
14967 (cortex_a8_alu_shift): Likewise.
14968 (cortex_a8_alu_shift_reg): Likewise.
14969 (cortex_a8_mov): Likewise.
14970 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
14971 (cortex_a9_dp_shift): Likewise.
14972 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
14973 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
14974 (cortex_r4_mov): Likewise.
14975 (cortex_r4_alu_shift): Likewise.
14976 (cortex_r4_alu_shift_reg): Likewise.
14977 * config/arm/fa526.md (526_alu_op): Update for attribute change.
14978 (526_alu_shift_op): Likewise.
14979 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
14980 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
14981 (626te_alu_shift_op): Likewise.
14982 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
14983 (726te_alu_op): Likewise.
14984 (726te_alu_shift_op): Likewise.
14985 (726te_alu_shift_reg_op): Likewise.
14986 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
14987 (mp626_alu_shift_op): Likewise.
14988 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
14989 (pj4_alu_e1_conds): Likewise.
14990 (pj4_alu): Likewise.
14991 (pj4_alu_conds): Likewise.
14992 (pj4_shift): Likewise.
14993 (pj4_shift_conds): Likewise.
14994 (pj4_alu_shift): Likewise.
14995 (pj4_alu_shift_conds): Likewise.
14997 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14999 * config/arm/predicates.md (shiftable_operator_strict_it):
15001 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
15002 Disable cond_exec version for arm_restrict_it.
15003 (thumb2_smaxsi3): Convert to generate cond_exec.
15004 (thumb2_sminsi3): Likewise.
15005 (thumb32_umaxsi3): Likewise.
15006 (thumb2_uminsi3): Likewise.
15007 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
15008 (thumb2_neg_abssi2): Likewise.
15009 (thumb2_mov_scc): Add alternative for 16-bit encoding.
15010 (thumb2_movsicc_insn): Adjust alternatives.
15011 (thumb2_mov_negscc): Disable for arm_restrict_it.
15012 (thumb2_mov_negscc_strict_it): New pattern.
15013 (thumb2_mov_notscc_strict_it): New pattern.
15014 (thumb2_mov_notscc): Disable for arm_restrict_it.
15015 (thumb2_ior_scc): Likewise.
15016 (thumb2_ior_scc_strict_it): New pattern.
15017 (thumb2_cond_move): Adjust for arm_restrict_it.
15018 (thumb2_cond_arith): Disable for arm_restrict_it.
15019 (thumb2_cond_arith_strict_it): New pattern.
15020 (thumb2_cond_sub): Adjust for arm_restrict_it.
15021 (thumb2_movcond): Likewise.
15022 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
15023 (thumb2_zero_extendhisi2_v6): Likewise.
15024 (thumb2_zero_extendqisi2_v6): Likewise.
15025 (orsi_notsi_si): Likewise.
15026 (orsi_not_shiftsi_si): Likewise.
15028 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
15030 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
15031 instruction sequence is 1 byte shorter.
15033 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
15035 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
15036 it is not needed after split.
15038 2013-07-20 Iain Sandoe <iain@codesourcery.com>
15041 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
15042 second label for nonlocal goto receivers. Don't output pic base labels
15043 unless we're producing PIC; mark that action unreachable().
15044 (ix86_save_reg): If the function contains a nonlocal label, save the
15046 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
15047 * config/darwin.c (emitted_pic_label_num): New GTY.
15048 (update_pic_label_number_if_needed): New.
15049 (machopic_output_function_base_name): Adjust for nonlocal receiver
15051 (machopic_should_output_picbase_label): New.
15052 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
15053 (nonlocal_goto_receiver): New insn and split.
15055 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
15057 * config/aarch64/aarch64-builtins.c
15058 (aarch64_fold_builtin): Fold abs in all modes.
15059 * config/aarch64/aarch64-simd-builtins.def
15060 (abs): Enable for all modes.
15061 * config/aarch64/arm_neon.h
15062 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
15063 (vabs_f64): Add missing intrinsic.
15065 2013-07-19 Ian Bolton <ian.bolton@arm.com>
15067 * config/aarch64/arm_neon.h (vabs_s64): New function
15069 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
15072 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
15073 * config/avr/avr.md (adjust_len): Add `round'.
15074 * config/avr/avr-protos.h (avr_out_round): New prototype.
15075 (avr_out_plus): Add `out_label' argument.
15076 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
15077 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
15078 Handle the case where `insn' is just a pattern.
15079 (avr_out_bitop): Handle the case where `insn' is just a pattern.
15080 (avr_out_round): New function.
15081 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
15083 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
15085 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15086 Add check for flag_stack_usage to handle -fstack-usage support
15088 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
15090 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
15091 interaction for new Power8 flags and VSX.
15093 2013-07-18 Sriraman Tallam <tmsriram@google.com>
15095 PR middle-end/57698
15096 * tree-inline.c (expand_call_inline): Emit errors during
15097 early_inlining only if optimization is not turned on.
15099 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15103 * passes.c (init_optimization_passes): Move the construction of
15104 the pass hierarchy into a new passes.def file.
15106 * Makefile.in (passes.o): Add dependency on passes.def.
15108 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15110 * passes.c (init_optimization_passes): Introduce macros for
15111 constructing the tree of passes (INSERT_PASSES_AFTER,
15112 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
15113 TERMINATE_PASS_LIST).
15115 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
15116 Wei Mi <wmi@google.com>
15118 PR rtl-optimization/57878
15119 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
15121 (reload_pseudo_compare_func): Check nregs first for reload
15124 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15126 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
15128 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
15130 * read-rtl.c (validate_const_int): Once an invalid character is
15131 seen, quit the loop.
15133 * gengtype.c (write_roots): Similarly once we find the "deletable"
15134 or "if_marked" option.
15136 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15138 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
15139 "xtab" and "sat". Move value "clz" from here to ...
15140 (attriubte "type"): ... here.
15141 (satsi_<SAT:code>): Delete "insn" attribute.
15142 (satsi_<SAT:code>_shift): Likewise.
15143 (arm_zero_extendqisi2addsi): Likewise.
15144 (arm_extendqisi2addsi): Likewise.
15145 (clzsi2): Update for attribute changes.
15146 (rbitsi2): Likewise.
15147 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
15149 (arm_usatsihi): Likewise.
15150 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15152 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15154 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
15155 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
15156 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
15157 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
15158 in alphabetical order.
15159 (attribute "core_cycles"): Update for attribute changes.
15160 (arm_addsi3): Likewise.
15161 (addsi3_compare0): Likewise.
15162 (addsi3_compare0_scratch): Likewise.
15163 (addsi3_compare_op1): Likewise.
15164 (addsi3_compare_op2): Likewise.
15165 (compare_addsi2_op0): Likewise.
15166 (compare_addsi2_op1): Likewise.
15167 (addsi3_carryin_shift_<optab>): Likewise.
15168 (subsi3_carryin_shift): Likewise.
15169 (rsbsi3_carryin_shift): Likewise.
15170 (arm_subsi3_insn): Likewise.
15171 (subsi3_compare0): Likewise.
15172 (subsi3_compare): Likewise.
15173 (arm_andsi3_insn): Likewise.
15174 (thumb1_andsi3_insn): Likewise.
15175 (andsi3_compare0): Likewise.
15176 (andsi3_compare0_scratch): Likewise.
15177 (zeroextractsi_compare0_scratch
15178 (andsi_not_shiftsi_si): Likewise.
15179 (iorsi3_insn): Likewise.
15180 (iorsi3_compare0): Likewise.
15181 (iorsi3_compare0_scratch): Likewise.
15182 (arm_xorsi3): Likewise.
15183 (thumb1_xorsi3_insn): Likewise.
15184 (xorsi3_compare0): Likewise.
15185 (xorsi3_compare0_scratch): Likewise.
15186 (satsi_<SAT:code>_shift): Likewise.
15188 (arm_shiftsi3): Likewise.
15189 (shiftsi3_compare0): Likewise.
15190 (not_shiftsi): Likewise.
15191 (not_shiftsi_compare0): Likewise.
15192 (not_shiftsi_compare0_scratch): Likewise.
15193 (arm_one_cmplsi2): Likewise.
15194 (thumb_one_complsi2): Likewise.
15195 (notsi_compare0): Likewise.
15196 (notsi_compare0_scratch): Likewise.
15197 (thumb1_zero_extendhisi2): Likewise.
15198 (arm_zero_extendhisi2): Likewise.
15199 (arm_zero_extendhisi2_v6): Likewise.
15200 (arm_zero_extendhisi2addsi): Likewise.
15201 (thumb1_zero_extendqisi2): Likewise.
15202 (thumb1_zero_extendqisi2_v6): Likewise.
15203 (arm_zero_extendqisi2): Likewise.
15204 (arm_zero_extendqisi2_v6): Likewise.
15205 (arm_zero_extendqisi2addsi): Likewise.
15206 (thumb1_extendhisi2): Likewise.
15207 (arm_extendhisi2): Likewise.
15208 (arm_extendhisi2_v6): Likewise.
15209 (arm_extendqisi): Likewise.
15210 (arm_extendqisi_v6): Likewise.
15211 (arm_extendqisi2addsi): Likewise.
15212 (thumb1_extendqisi2): Likewise.
15213 (thumb1_movdi_insn): Likewise.
15214 (arm_movsi_insn): Likewise.
15215 (movsi_compare0): Likewise.
15216 (movhi_insn_arch4): Likewise.
15217 (movhi_bytes): Likewise.
15218 (arm_movqi_insn): Likewise.
15219 (thumb1_movqi_insn): Likewise.
15220 (arm32_movhf): Likewise.
15221 (thumb1_movhf): Likewise.
15222 (arm_movsf_soft_insn): Likewise.
15223 (thumb1_movsf_insn): Likewise.
15224 (movdf_soft_insn): Likewise.
15225 (thumb_movdf_insn): Likewise.
15226 (arm_cmpsi_insn): Likewise.
15227 (cmpsi_shiftsi): Likewise.
15228 (cmpsi_shiftsi_swp): Likewise.
15229 (arm_cmpsi_negshiftsi_si): Likewise.
15230 (movsicc_insn): Likewise.
15231 (movsfcc_soft_insn): Likewise.
15232 (arith_shiftsi): Likewise.
15233 (arith_shiftsi_compare0
15234 (arith_shiftsi_compare0_scratch
15235 (sub_shiftsi): Likewise.
15236 (sub_shiftsi_compare0
15237 (sub_shiftsi_compare0_scratch
15238 (and_scc): Likewise.
15239 (cond_move): Likewise.
15240 (if_plus_move): Likewise.
15241 (if_move_plus): Likewise.
15242 (if_move_not): Likewise.
15243 (if_not_move): Likewise.
15244 (if_shift_move): Likewise.
15245 (if_move_shift): Likewise.
15246 (if_shift_shift): Likewise.
15247 (if_not_arith): Likewise.
15248 (if_arith_not): Likewise.
15249 (cond_move_not): Likewise.
15250 (thumb1_ashlsi3): Set type attribute.
15251 (thumb1_ashrsi3): Likewise.
15252 (thumb1_lshrsi3): Likewise.
15253 (thumb1_rotrsi3): Likewise.
15254 (shiftsi3_compare0_scratch): Likewise.
15255 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
15256 (neon_mov<mode>): Likewise.
15257 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
15259 (thumb2_movsi_insn): Likewise.
15260 (thumb2_cmpsi_neg_shiftsi): Likewise.
15261 (thumb2_extendqisi_v6): Likewise.
15262 (thumb2_zero_extendhisi2_v6): Likewise.
15263 (thumb2_zero_extendqisi2_v6): Likewise.
15264 (thumb2_shiftsi3_short): Likewise.
15265 (thumb2_addsi3_compare0_scratch): Likewise.
15266 (orsi_not_shiftsi_si): Likewise.
15267 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
15268 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
15270 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
15271 (1020alu_shift_op): Likewise.
15272 (1020alu_shift_reg_op): Likewise.
15273 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
15274 (alu_shift_op): Likewise.
15275 (alu_shift_reg_op): Likewise.
15276 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
15277 (11_alu_shift_op): Likewise.
15278 (11_alu_shift_reg_op): Likewise.
15279 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
15280 (9_alu_shift_reg_op): Likewise.
15281 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15283 (cortex_a15_alu_shift): Likewise.
15284 (cortex_a15_alu_shift_reg): Likewise.
15285 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
15287 (cortex_a5_alu_shift): Likewise.
15288 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
15290 (cortex_a53_alu_shift): Likewise.
15291 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15293 (cortex_a7_alu_reg): Likewise.
15294 (cortex_a7_alu_shift): Likewise.
15295 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
15297 (cortex_a8_alu_shift): Likewise.
15298 (cortex_a8_alu_shift_reg): Likewise.
15299 (cortex_a8_mov): Likewise.
15300 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
15301 (cortex_a9_dp_shift): Likewise.
15302 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
15304 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
15306 (cortex_r4_mov): Likewise.
15307 (cortex_r4_alu_shift): Likewise.
15308 (cortex_r4_alu_shift_reg): Likewise.
15309 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
15310 (526_alu_shift_op): Likewise.
15311 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
15312 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
15313 (626te_alu_shift_op): Likewise.
15314 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
15315 (726te_alu_op): Likewise.
15316 (726te_alu_shift_op): Likewise.
15317 (726te_alu_shift_reg_op): Likewise.
15318 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
15319 (mp626_alu_shift_op): Likewise.
15320 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
15321 (pj4_alu_e1_conds): Likewise.
15322 (pj4_alu): Likewise.
15323 (pj4_alu_conds): Likewise.
15324 (pj4_shift): Likewise.
15325 (pj4_shift_conds): Likewise.
15326 (pj4_alu_shift): Likewise.
15327 (pj4_alu_shift_conds): Likewise.
15328 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15330 (cortexa7_older_only): Likewise.
15331 (cortexa7_younger): Likewise.
15333 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15335 * ipa-pure-const.c (generate_summary): Rename to...
15336 (pure_const_generate_summary): ... this.
15338 2013-07-17 Iain Sandoe <iain@codesourcery.com>
15340 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
15342 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
15345 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
15348 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15350 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
15351 enabled without -march=zEC12.
15352 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
15355 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15357 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
15358 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
15359 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
15360 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
15361 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
15362 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
15363 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
15365 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
15366 and ISA_HAS_NMADD3_NMSUB3.
15367 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
15368 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
15369 (nmsub4<mode>, nmsub3<mode>): Likewise.
15370 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
15372 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15374 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
15375 TARGET_MIPS5400 checking.
15377 2013-07-16 Jakub Jelinek <jakub@redhat.com>
15378 Peter Bergner <bergner@vnet.ibm.com>
15380 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
15381 registers in the comment.
15382 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
15383 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
15384 rather than FIRST_PSEUDO_REGISTERS.
15386 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
15388 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
15389 enable extra ISA flags with TARGET_HTM.
15391 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15393 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
15396 2013-07-15 Cong Hou <congh@google.com>
15398 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
15399 in compare function for sorting.
15401 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
15403 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
15404 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
15405 * config/rs6000/rs6000.opt: Add -mhtm option.
15406 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
15407 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
15408 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
15409 __HTM__ if the HTM instructions are available.
15410 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
15411 htm_spr_reg_operand): New define_predicates.
15412 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
15413 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
15415 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
15416 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
15417 HTM builtin functions.
15418 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
15419 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
15420 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
15421 (rs6000_builtin_mask_calculate): Likewise.
15422 (rs6000_option_override_internal): Likewise.
15423 (bdesc_htm): Add new HTM builtin support.
15424 (htm_spr_num): New function.
15425 (htm_spr_regno): Likewise.
15426 (rs6000_htm_spr_icode): Likewise.
15427 (htm_expand_builtin): Likewise.
15428 (htm_init_builtins): Likewise.
15429 (rs6000_expand_builtin): Add support for HTM builtin functions.
15430 (rs6000_init_builtins): Likewise.
15431 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
15433 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
15434 (TARGET_HTM, MASK_HTM): Define macros.
15435 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
15436 (FIXED_REGISTERS): Likewise.
15437 (CALL_USED_REGISTERS): Likewise.
15438 (CALL_REALLY_USED_REGISTERS): Likewise.
15439 (REG_ALLOC_ORDER): Likewise.
15440 (enum reg_class): Likewise.
15441 (REG_CLASS_NAMES): Likewise.
15442 (REG_CLASS_CONTENTS): Likewise.
15443 (REGISTER_NAMES): Likewise.
15444 (ADDITIONAL_REGISTER_NAMES): Likewise.
15445 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
15446 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
15447 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
15448 * config/rs6000/htm.md: New file.
15449 * config/rs6000/htmintrin.h: New file.
15450 * config/rs6000/htmxlintrin.h: New file.
15452 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
15454 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15455 Define SYMBOL_TINY_GOT, update comment.
15456 * config/aarch64/aarch64.c
15457 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
15458 (aarch64_expand_mov_immediate): Likewise.
15459 (aarch64_print_operand): Likewise.
15460 (aarch64_classify_symbol): Likewise.
15461 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
15462 (ldr_got_tiny): Define.
15464 2013-07-13 Tobias Grosser <tobias@grosser.es>
15466 PR tree-optimization/54094
15467 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
15468 scheduling dimension for the parallelism check from the polyhedral
15469 information in the AST.
15470 * graphite-dependences.c (carries_deps): Do not assume the schedule is
15473 2013-07-13 Jason Merrill <jason@redhat.com>
15475 * print-tree.c (debug_vec_tree): Use debug_raw.
15476 (debug_raw (vec<tree, va_gc> &)): New.
15477 (debug_raw (vec<tree, va_gc> *)): New.
15478 * tree.h: Declare them.
15480 2013-07-13 Bin Cheng <bin.cheng@arm.com>
15482 * ifcvt.c (ifcvt_after_combine): New static variable.
15483 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
15485 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
15486 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
15487 rest_of_handle_if_after_reload): Pass new argument for if_convert.
15489 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
15491 * config/mips/mips.c (mips_expand_call): Remove empty statement.
15493 2013-07-12 Michael Matz <matz@suse.de>
15495 PR middle-end/55771
15496 * convert.c (convert_to_real): Reject non-float inner types.
15498 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
15500 * config/aarch64/aarch64-protos.h
15501 (aarch64_simd_immediate_valid_for_move): Remove.
15502 * config/aarch64/aarch64.c (simd_immediate_info): New member.
15503 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
15505 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
15507 2013-07-11 Steve Ellcey <sellcey@mips.com>
15509 * config/mips/mips.c (mips_conditional_register_usage): Do not
15510 use t[0-7] registers in MIPS16 mode when optimizing for size.
15512 2013-07-11 Sriraman Tallam <tmsriram@google.com>
15514 * config/i386/i386.c (dispatch_function_versions): Fix array
15515 indexing of function_version_info to match actual_versions.
15517 2013-07-11 Teresa Johnson <tejohnson@google.com>
15519 * vec.h (struct va_gc): Move release out-of-line.
15520 (va_gc::release): Call ggc_free on released vec.
15522 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15524 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
15525 Require GOT register as additional operand in UNSPEC.
15526 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
15527 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15528 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15529 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
15530 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
15531 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
15532 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
15534 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
15537 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
15538 name seen by assembler/linker rather if available.
15540 2013-07-11 Andreas Schwab <schwab@suse.de>
15542 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
15544 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
15546 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
15548 2013-07-10 Joseph Myers <joseph@codesourcery.com>
15550 * doc/tm.texi.in: Move hook documentation to ....
15551 * target.def: ... here.
15553 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
15554 text on @hook line.
15555 * doc/tm.texi: Regenerate.
15557 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
15560 * doc/invoke.texi: Document -Wconditionally-supported.
15562 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15565 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
15568 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15571 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
15572 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
15573 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
15574 Remove duplicate devices.
15575 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
15576 * config/avr/t-multilib: Regenerate.
15577 * config/avr/avr-tables.opt: Regenerate.
15578 * doc/avr-mmcu.texi: Regenerate.
15580 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15583 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
15585 2013-07-10 Graham Stott <graham.stott@btinternet.com>
15587 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
15588 the cost of MULT when optimizing for size.
15590 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
15592 * config/cr16/cr16-protos.h: Don't include target.h.
15594 2013-07-09 Joseph Myers <joseph@codesourcery.com>
15596 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
15597 adjust register size for TDmode and TFmode for VSX registers.
15599 2013-07-08 Kai Tietz <ktietz@redhat.com>
15602 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
15603 hook_bool_const_tree_true.
15605 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15607 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
15608 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
15609 * config/s390/s390.md: Define FPR*_REGNUM constants.
15610 Fix FPR2_REGNUM constant (18 -> 17).
15611 ("*trunc<BFP:mode><DFP_ALL:mode>2")
15612 ("*trunc<DFP_ALL:mode><BFP:mode>2")
15613 ("trunc<BFP:mode><DFP_ALL:mode>2")
15614 ("trunc<DFP_ALL:mode><BFP:mode>2")
15615 ("*extend<BFP:mode><DFP_ALL:mode>2")
15616 ("*extend<DFP_ALL:mode><BFP:mode>2")
15617 ("extend<BFP:mode><DFP_ALL:mode>2")
15618 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
15621 2013-07-08 Graham Stott <graham.stott@btinternet.com>
15623 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
15625 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15627 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
15628 and cfun_fpr_bit_p to cfun_fpr_save_p.
15629 (s390_frame_area, s390_register_info, s390_frame_info)
15630 (s390_emit_prologue, s390_emit_epilogue)
15631 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
15633 * config/s390/s390.h: Define *_REGNUM macros for floating point
15636 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
15638 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
15640 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
15642 PR rtl-optimization/57786
15643 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
15644 and break out of the loop when it is set to false.
15646 2013-07-08 Jakub Jelinek <jakub@redhat.com>
15649 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
15650 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
15651 (const_int 63)) 0)).
15652 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
15653 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
15654 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
15656 PR rtl-optimization/57829
15657 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
15658 mask bits outside of mode are just sign-extension from mode to HWI.
15660 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
15662 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
15663 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
15664 adjust_address instead of change_address to keep info about alignment.
15665 (emit_strmov): Remove.
15666 (emit_memmov): New function.
15667 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
15668 (expand_movmem_epilogue): Likewise and return updated rtx for
15670 (expand_constant_movmem_prologue): Likewise and return updated rtx for
15671 destination and source.
15672 (decide_alignment): Refactor, handle vector_loop.
15673 (ix86_expand_movmem): Likewise.
15674 (ix86_expand_setmem): Likewise.
15675 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
15677 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
15679 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
15680 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
15682 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15684 * config/i386/sse.md (sse_movlhps): Change alternative 3
15685 of operand 2 to "m".
15687 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15690 * config/i386/sse.md (iptr): New mode attribute.
15691 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
15692 (<sse>_vm<plusminus_insn><mode>3): Ditto.
15693 (<sse>_vmmul<mode>3): Ditto.
15694 (<sse>_vmdiv<mode>3): Ditto.
15695 (sse_vmrcpv4sf2): Ditto.
15696 (<sse>_vmsqrt<mode>2): Ditto.
15697 (sse_vmrsqrtv4sf2): Ditto.
15698 (<sse>_vm<code><mode>3): Ditto.
15699 (avx_vmcmp<mode>3): Ditto.
15700 (<sse>_vmmaskcmp<mode>3): Ditto.
15701 (<sse>_comi): Ditto.
15702 (<sse>_ucomi): Ditto.
15703 (*xop_vmfrcz_<mode>): Ditto.
15704 (*fmai_fmadd_<mode>): Ditto.
15705 (*fmai_fmsub_<mode>): Ditto.
15706 (*fmai_fnmadd_<mode>): Ditto.
15707 (*fmai_fnmsub_<mode>): Ditto.
15708 (*fma4i_vmfmadd_<mode>): Ditto.
15709 (*fma4i_vmfmsub_<mode>): Ditto.
15710 (*fma4i_vmfnmadd_<mode>): Ditto.
15711 (*fma4i_vmfnmsub_<mode>): Ditto.
15712 (*xop_vmfrcz_<mode>): Ditto.
15713 (sse_cvtps2pi): Ditto.
15714 (sse_cvttps2pi): Ditto.
15715 (sse_cvtss2si): Ditto.
15716 (sse_cvtss2si_2): Ditto.
15717 (sse_cvtss2siq_2): Ditto.
15718 (sse_cvttss2si): Ditto.
15719 (sse_cvttss2siq): Ditto.
15720 (sse_cvtsd2si): Ditto.
15721 (sse_cvtsd2si_2): Ditto.
15722 (sse_cvtsd2siq_2): Ditto.
15723 (sse_cvttsd2si): Ditto.
15724 (sse_cvttsd2siq): Ditto.
15725 (sse_cvtsd2ss): Ditto.
15726 (sse_cvtss2sd): Ditto.
15727 (avx2_pbroadcast<mode>): Ditto.
15728 (avx2_pbroadcast<mode>_1): Ditto.
15729 (*avx_vperm_broadcast_v4sf): Ditto.
15731 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
15732 (sse_movlhps): Ditto.
15733 (sse_storehps): Ditto.
15734 (sse_loadhps): Ditto.
15735 (sse_storelps): Ditto.
15736 (sse_loadlps): Ditto.
15737 (*vec_concatv4sf): Ditto.
15738 (*vec_interleave_highv2df): Ditto.
15739 (*vec_interleave_lowv2df): Ditto.
15740 (*vec_extractv2df_1_sse): Ditto.
15741 (*vec_extractv2df_0_sse): Ditto.
15742 (sse2_storelpd): Ditto.
15743 (sse2_loadlpd): Ditto.
15744 (sse2_movsd): Ditto.
15745 (*vec_concatv4si): Ditto.
15746 (vec_concatv2di): Ditto.
15748 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
15749 for Intel asm dialect.
15750 (mmx_punpcklwd): Ditto.
15751 (mmx_punpckldq): Ditto.
15753 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
15754 for intel assembler dialect.
15756 2013-07-06 Jakub Jelinek <jakub@redhat.com>
15759 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
15760 for BUILT_IN_C{LZ,LRSB}*.
15761 * tree.h (CASE_INT_FN): Add FN##IMAX case.
15762 * tree-vrp.c (extract_range_basic): Handle
15763 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
15764 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
15765 fall thru to code calling set_value*.
15766 * builtins.c (expand_builtin): Remove *IMAX cases.
15767 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
15768 if width is bigger than 2*HWI.
15770 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
15772 PR rtl-optimization/55342
15773 * lra-int.h (lra_subreg_reload_pseudos): New.
15774 * lra.c: Add undoing optional reloads to the block diagram.
15775 (lra_subreg_reload_pseudos): New.
15776 (lra_optional_reload_pseudos): Change comments.
15777 (lra): Init and clear lra_subreg_reload_pseudos. Clear
15778 lra_optional_reload_pseudos after undo transformations.
15779 * lra-assigns.c (pseudo_prefix_title): New.
15780 (lra_setup_reg_renumber): Use it.
15781 (spill_for): Ditto. Check subreg reload pseudos too.
15782 (assign_by_spills): Consider subreg reload pseudos too.
15783 * lra-constraints.c (simplify_operand_subreg): Use
15784 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
15785 (curr_insn_transform): Recognize and do optional reloads.
15786 (undo_optional_reloads): New.
15787 (lra_undo_inheritance): Call undo_optional_reloads.
15789 2013-07-05 Thomas Quinot <quinot@adacore.com>
15791 * tree-complex.c (expand_complex_operations_1): Fix typo.
15793 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
15795 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
15796 (tune_params): New member 'const vec_costs'.
15797 * config/aarch64/aarch64.c (generic_vector_cost): New.
15798 (generic_tunings): New member 'generic_vector_cost'.
15799 (aarch64_builtin_vectorization_cost): New.
15800 (aarch64_add_stmt_cost): New.
15801 (TARGET_VECTORIZE_ADD_STMT_COST): New.
15802 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
15804 2013-07-03 Jakub Jelinek <jakub@redhat.com>
15807 * config/i386/predicates.md (vsib_address_operand): Disallow
15808 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
15810 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
15812 PR middle-end/55030
15813 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
15814 expand_builtin_setjmp_receiver.
15815 (expand_label): Adjust, call expand_builtin_setjmp_receiver
15816 with NULL for the label parameter.
15817 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
15818 the frame-pointer. Adjust comments.
15819 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
15820 only if LABEL is non-NULL.
15822 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
15824 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
15825 (ARM_ABI_AAPCS64): Ditto.
15827 (ARM_DEFAULT_ABI): Ditto.
15829 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
15831 * config/aarch64/aarch64-builtins.c
15832 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
15833 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
15835 * config/aarch64/aarch64-simd.md
15836 (aarch64_ld1<VALL:mode>): New.
15837 (aarch64_st1<VALL:mode>): Likewise.
15838 * config/aarch64/arm_neon.h
15839 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
15841 2013-07-02 Sriraman Tallam <tmsriram@google.com>
15843 * config/i386/i386.c (gate_insert_vzeroupper): Check if
15845 (ix86_option_override_internal):Turn on all -mavx target flags by
15846 default as they are dependent on AVX anyway.
15848 2013-07-02 Cary Coutant <ccoutant@google.com>
15850 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
15851 deterministic hash.
15852 (loc_checksum_ordered): Likewise.
15853 (hash_loc_operands): Remove inline keyword.
15855 2013-07-02 Jakub Jelinek <jakub@redhat.com>
15857 PR tree-optimization/57741
15858 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
15859 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
15860 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
15861 Allow REAL_CST step_exprs if flag_associative_math.
15862 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
15864 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15866 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
15868 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15870 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
15872 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15874 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
15876 (iorsi3_insn): Likewise.
15877 (arm_xorsi3): Likewise.
15879 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
15881 * arm.md (attribute "wtype"): Delete. Move attribute values from here
15883 (attribute "type"): ... here, and prefix with "wmmx_".
15884 (attribute "core_cycles"): Update for attribute changes.
15885 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
15886 (tbcstv4hi): Likewise.
15887 (tbcstv2si): Likewise.
15888 (iwmmxt_iordi3): Likewise.
15889 (iwmmxt_xordi3): Likewise.
15890 (iwmmxt_anddi3): Likewise.
15891 (iwmmxt_nanddi3): Likewise.
15892 (iwmmxt_arm_movdi): Likewise.
15893 (iwmmxt_movsi_insn): Likewise.
15894 (mov<mode>_internal): Likewise.
15895 (and<mode>3_iwmmxt): Likewise.
15896 (ior<mode>3_iwmmxt): Likewise.
15897 (xor<mode>3_iwmmxt): Likewise.
15898 (add<mode>3_iwmmxt): Likewise.
15899 (ssaddv8qi3): Likewise.
15900 (ssaddv4hi3): Likewise.
15901 (ssaddv2si3): Likewise.
15902 (usaddv8qi3): Likewise.
15903 (usaddv4hi3): Likewise.
15904 (usaddv2si3): Likewise.
15905 (sub<mode>3_iwmmxt): Likewise.
15906 (sssubv8qi3): Likewise.
15907 (sssubv4hi3): Likewise.
15908 (sssubv2si3): Likewise.
15909 (ussubv8qi3): Likewise.
15910 (ussubv4hi3): Likewise.
15911 (ussubv2si3): Likewise.
15912 (mulv4hi3_iwmmxt): Likewise.
15913 (smulv4hi3_highpart): Likewise.
15914 (umulv4hi3_highpart): Likewise.
15915 (iwmmxt_wmacs): Likewise.
15916 (iwmmxt_wmacsz): Likewise.
15917 (iwmmxt_wmacu): Likewise.
15918 (iwmmxt_wmacuz): Likewise.
15919 (iwmmxt_clrdi): Likewise.
15920 (iwmmxt_clrv8qi): Likewise.
15921 (iwmmxt_clr4hi): Likewise.
15922 (iwmmxt_clr2si): Likewise.
15923 (iwmmxt_uavgrndv8qi3): Likewise.
15924 (iwmmxt_uavgrndv4hi3): Likewise.
15925 (iwmmxt_uavgv8qi3): Likewise.
15926 (iwmmxt_uavgv4hi3): Likewise.
15927 (iwmmxt_tinsrb): Likewise.
15928 (iwmmxt_tinsrh): Likewise.
15929 (iwmmxt_tinsrw): Likewise.
15930 (iwmmxt_textrmub): Likewise.
15931 (iwmmxt_textrmsb): Likewise.
15932 (iwmmxt_textrmuh): Likewise.
15933 (iwmmxt_textrmsh): Likewise.
15934 (iwmmxt_textrmw): Likewise.
15935 (iwmxxt_wshufh): Likewise.
15936 (eqv8qi3): Likewise.
15937 (eqv4hi3): Likewise.
15938 (eqv2si3): Likewise.
15939 (gtuv8qi3): Likewise.
15940 (gtuv4hi3): Likewise.
15941 (gtuv2si3): Likewise.
15942 (gtv8qi3): Likewise.
15943 (gtv4hi3): Likewise.
15944 (gtv2si3): Likewise.
15945 (smax<mode>3_iwmmxt): Likewise.
15946 (umax<mode>3_iwmmxt): Likewise.
15947 (smin<mode>3_iwmmxt): Likewise.
15948 (umin<mode>3_iwmmxt): Likewise.
15949 (iwmmxt_wpackhss): Likewise.
15950 (iwmmxt_wpackwss): Likewise.
15951 (iwmmxt_wpackdss): Likewise.
15952 (iwmmxt_wpackhus): Likewise.
15953 (iwmmxt_wpackwus): Likewise.
15954 (iwmmxt_wpackdus): Likewise.
15955 (iwmmxt_wunpckihb): Likewise.
15956 (iwmmxt_wunpckihh): Likewise.
15957 (iwmmxt_wunpckihw): Likewise.
15958 (iwmmxt_wunpckilb): Likewise.
15959 (iwmmxt_wunpckilh): Likewise.
15960 (iwmmxt_wunpckilw): Likewise.
15961 (iwmmxt_wunpckehub): Likewise.
15962 (iwmmxt_wunpckehuh): Likewise.
15963 (iwmmxt_wunpckehuw): Likewise.
15964 (iwmmxt_wunpckehsb): Likewise.
15965 (iwmmxt_wunpckehsh): Likewise.
15966 (iwmmxt_wunpckehsw): Likewise.
15967 (iwmmxt_wunpckelub): Likewise.
15968 (iwmmxt_wunpckeluh): Likewise.
15969 (iwmmxt_wunpckeluw): Likewise.
15970 (iwmmxt_wunpckelsb): Likewise.
15971 (iwmmxt_wunpckelsh): Likewise.
15972 (iwmmxt_wunpckelsw): Likewise.
15973 (ror<mode>3): Likewise.
15974 (ashr<mode>3_iwmmxt): Likewise.
15975 (lshr<mode>3_iwmmxt): Likewise.
15976 (ashl<mode>3_iwmmxt): Likewise.
15977 (ror<mode>3_di): Likewise.
15978 (ashr<mode>3_di): Likewise.
15979 (lshr<mode>3_di): Likewise.
15980 (ashl<mode>3_di): Likewise.
15981 (iwmmxt_wmadds): Likewise.
15982 (iwmmxt_wmaddu): Likewise.
15983 (iwmmxt_tmia): Likewise.
15984 (iwmmxt_tmiaph): Likewise.
15985 (iwmmxt_tmiabb): Likewise.
15986 (iwmmxt_tmiatb): Likewise.
15987 (iwmmxt_tmiabt): Likewise.
15988 (iwmmxt_tmiatt): Likewise.
15989 (iwmmxt_tmovmskb): Likewise.
15990 (iwmmxt_tmovmskh): Likewise.
15991 (iwmmxt_tmovmskw): Likewise.
15992 (iwmmxt_waccb): Likewise.
15993 (iwmmxt_wacch): Likewise.
15994 (iwmmxt_waccw): Likewise.
15995 (iwmmxt_waligni): Likewise.
15996 (iwmmxt_walignr): Likewise.
15997 (iwmmxt_walignr0): Likewise.
15998 (iwmmxt_walignr1): Likewise.
15999 (iwmmxt_walignr2): Likewise.
16000 (iwmmxt_walignr3): Likewise.
16001 (iwmmxt_wsadb): Likewise.
16002 (iwmmxt_wsadh): Likewise.
16003 (iwmmxt_wsadbz): Likewise.
16004 (iwmmxt_wsadhz): Likewise.
16005 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
16006 (iwmmxt_wabsdiffb): Likewise.
16007 (iwmmxt_wabsdiffh): Likewise.
16008 (iwmmxt_wabsdiffw): Likewise.
16009 (iwmmxt_waddsubhx): Likewise
16010 (iwmmxt_wsubaddhx): Likewise.
16011 (addc<mode>3): Likewise.
16012 (iwmmxt_avg4): Likewise.
16013 (iwmmxt_avg4r): Likewise.
16014 (iwmmxt_wmaddsx): Likewise.
16015 (iwmmxt_wmaddux): Likewise.
16016 (iwmmxt_wmaddsn): Likewise.
16017 (iwmmxt_wmaddun): Likewise.
16018 (iwmmxt_wmulwsm): Likewise.
16019 (iwmmxt_wmulwum): Likewise.
16020 (iwmmxt_wmulsmr): Likewise.
16021 (iwmmxt_wmulumr): Likewise.
16022 (iwmmxt_wmulwsmr): Likewise.
16023 (iwmmxt_wmulwumr): Likewise.
16024 (iwmmxt_wmulwl): Likewise.
16025 (iwmmxt_wqmulm): Likewise.
16026 (iwmmxt_wqmulwm): Likewise.
16027 (iwmmxt_wqmulmr): Likewise.
16028 (iwmmxt_wqmulwmr): Likewise.
16029 (iwmmxt_waddbhusm): Likewise.
16030 (iwmmxt_waddbhusl): Likewise.
16031 (iwmmxt_wqmiabb): Likewise.
16032 (iwmmxt_wqmiabt): Likewise.
16033 (iwmmxt_wqmiatb): Likewise.
16034 (iwmmxt_wqmiatt): Likewise.
16035 (iwmmxt_wqmiabbn): Likewise.
16036 (iwmmxt_wqmiabtn): Likewise.
16037 (iwmmxt_wqmiatbn): Likewise.
16038 (iwmmxt_wqmiattn): Likewise.
16039 (iwmmxt_wmiabb): Likewise.
16040 (iwmmxt_wmiabt): Likewise.
16041 (iwmmxt_wmiatb): Likewise.
16042 (iwmmxt_wmiatt): Likewise.
16043 (iwmmxt_wmiabbn): Likewise.
16044 (iwmmxt_wmiabtn): Likewise.
16045 (iwmmxt_wmiatbn): Likewise.
16046 (iwmmxt_wmiattn): Likewise.
16047 (iwmmxt_wmiawbb): Likewise.
16048 (iwmmxt_wmiawbt): Likewise.
16049 (iwmmxt_wmiawtb): Likewise.
16050 (iwmmxt_wmiawtt): Likewise.
16051 (iwmmxt_wmiawbbn): Likewise.
16052 (iwmmxt_wmiawbtn): Likewise.
16053 (iwmmxt_wmiawtbn): Likewise.
16054 (iwmmxt_wmiawttn): Likewise.
16055 (iwmmxt_wmerge): Likewise.
16056 (iwmmxt_tandc<mode>3): Likewise.
16057 (iwmmxt_torc<mode>3): Likewise.
16058 (iwmmxt_torvsc<mode>3): Likewise.
16059 (iwmmxt_textrc<mode>3): Likewise.
16060 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
16061 (wmmxt_pack): Likewise.
16062 (wmmxt_mult_c1): Likewise.
16063 (wmmxt_mult_c2): Likewise.
16064 (wmmxt_alu_c1): Likewise.
16065 (wmmxt_alu_c2): Likewise.
16066 (wmmxt_alu_c3): Likewise.
16067 (wmmxt_transfer_c1): Likewise.
16068 (wmmxt_transfer_c2): Likewise.
16069 (wmmxt_transfer_c3): Likewise.
16070 (marvell_f_iwmmxt_wstr): Likewise.
16071 (marvell_f_iwmmxt_wldr): Likewise.
16073 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
16075 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
16077 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16080 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16081 * lra-constraints.c (need_for_split_p): Check call used hard regs
16082 living through calls.
16084 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
16085 call used regs for call insn.
16087 2013-06-28 Jakub Jelinek <jakub@redhat.com>
16090 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
16091 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
16093 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16095 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
16097 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16099 * lra-constraints.c (need_for_split_p): Check call used hard regs
16100 living through calls.
16102 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
16105 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
16106 to tie with any other modes. Eliminate Altivec vector mode tests,
16107 since these are a subset of ALTIVEC or VSX vector modes. Simplify
16108 code, to return 0 if testing MODE2 for a condition, if we've
16109 already tested MODE1 for the same condition.
16111 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16113 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
16116 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16118 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
16119 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
16121 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16123 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
16125 (aarch64_symbolic_constant_p): Remove.
16126 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
16127 static. Fix line length and white space.
16128 (aarch64_symbolic_constant_p): Remove.
16129 * config/aarch64/predicates.md (aarch64_valid_symref):
16130 Use aarch64_classify_symbol_expression.
16132 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16134 * config/arm/constraints.md (Ts): New constraint.
16135 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
16137 (compare_scc): Use "Ts" constraint for operand 0.
16138 (ior_scc_scc): Likewise.
16139 (and_scc_scc): Likewise.
16140 (and_scc_scc_nodom): Likewise.
16141 (ior_scc_scc_cmp): Likewise for operand 7.
16142 (and_scc_scc_cmp): Likewise.
16143 * config/arm/thumb2.md (thumb2_movsi_insn):
16144 Add alternatives for 16-bit encodings.
16145 (thumb2_movhi_insn): Likewise.
16146 (thumb2_movsicc_insn): Likewise.
16147 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
16148 (thumb2_negscc): Use "Ts" constraint.
16149 Move mvn instruction outside cond_exec block.
16150 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
16151 for 16-bit encodings.
16153 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16155 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
16157 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
16158 (mulsi3subsi): Likewise.
16159 (mulsidi3adddi): Likewise.
16160 (mulsidi3_v6): Likewise.
16161 (umulsidi3_v6): Likewise.
16162 (umulsidi3adddi_v6): Likewise.
16163 (smulsi3_highpart_v6): Likewise.
16164 (umulsi3_highpart_v6): Likewise.
16165 (mulhisi3tb): Likewise.
16166 (mulhisi3bt): Likewise.
16167 (mulhisi3tt): Likewise.
16168 (maddhisi4): Likewise.
16169 (maddhisi4tb): Likewise.
16170 (maddhisi4tt): Likewise.
16171 (maddhidi4): Likewise.
16172 (maddhidi4tb): Likewise.
16173 (maddhidi4tt): Likewise.
16174 (zeroextractsi_compare0_scratch): Likewise.
16175 (insv_zero): Likewise.
16176 (insv_t2): Likewise.
16177 (anddi_notzesidi_di): Likewise.
16178 (anddi_notsesidi_di): Likewise.
16179 (andsi_notsi_si): Likewise.
16180 (iordi_zesidi_di): Likewise.
16181 (xordi_zesidi_di): Likewise.
16182 (andsi_iorsi3_notsi): Likewise.
16183 (smax_0): Likewise.
16184 (smax_m1): Likewise.
16185 (smin_0): Likewise.
16186 (not_shiftsi): Likewise.
16187 (unaligned_loadsi): Likewise.
16188 (unaligned_loadhis): Likewise.
16189 (unaligned_loadhiu): Likewise.
16190 (unaligned_storesi): Likewise.
16191 (unaligned_storehi): Likewise.
16192 (extv_reg): Likewise.
16193 (extzv_t2): Likewise.
16194 (divsi3): Likewise.
16195 (udivsi3): Likewise.
16196 (arm_zero_extendhisi2addsi): Likewise.
16197 (arm_zero_extendqisi2addsi): Likewise.
16198 (compareqi_eq0): Likewise.
16199 (arm_extendhisi2_v6): Likewise.
16200 (arm_extendqisi2addsi): Likewise.
16201 (arm_movt): Likewise.
16202 (thumb2_ldrd): Likewise.
16203 (thumb2_ldrd_base): Likewise.
16204 (thumb2_ldrd_base_neg): Likewise.
16205 (thumb2_strd): Likewise.
16206 (thumb2_strd_base): Likewise.
16207 (thumb2_strd_base_neg): Likewise.
16208 (arm_negsi2): Add alternative for 16-bit encoding.
16209 (arm_one_cmplsi2): Likewise.
16211 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16213 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
16214 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
16215 (movdfcc): Likewise.
16216 * config/arm/vfp.md (*thumb2_movsf_vfp):
16217 Disable predication for arm_restrict_it.
16218 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
16219 (*thumb2_movdfcc_vfp): Likewise.
16220 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
16221 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
16222 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
16223 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
16224 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
16225 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
16226 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
16227 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
16228 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
16229 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
16230 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
16231 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
16232 Disable predication for arm_restrict_it.
16234 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
16236 * config/i386/bmiintrin.h (_bextr_u32): New.
16237 (_bextr_u64): Ditto.
16239 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
16241 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
16242 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
16243 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
16244 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
16245 * config/mips/n32-elf.h: ...this new file.
16247 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
16250 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
16251 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
16253 2013-06-27 Catherine Moore <clm@codesourcery.com>
16255 * config/mips/mips-tables.opt: Regenerate.
16256 * config/mips/mips-cpus.def: Add m14ke and m14kec.
16257 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
16258 * doc/invoke.texi: Add -m14kc.
16260 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16263 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
16264 constraints of operand 1 and 2.
16267 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
16268 to match RTL canonicalization. Swap predicates and
16269 constraints of operand 1 and 2.
16271 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
16273 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
16274 Process OP_INOUT regs for splitting too.
16276 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16278 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
16279 decl before the loop, initialize to NULL.
16280 (vectorizable_load): Initialize ptr_incr to NULL.
16282 2013-06-27 Martin Jambor <mjambor@suse.cz>
16285 * ipa-ref.h (ipa_maybe_record_reference): Declare.
16286 * ipa-ref.c (ipa_maybe_record_reference): New function.
16287 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
16288 * ipa-cp.c (create_specialized_node): Record potential references from
16290 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
16292 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16294 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
16295 parameter 'mode' of type 'enum machine_mode mode'; change to pass
16296 'mode' to force_reg.
16297 (aarch64_add_offset): Update calls to aarch64_force_temporary.
16298 (aarch64_expand_mov_immediate): Likewise.
16300 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16302 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
16303 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
16305 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16307 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
16308 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16309 (struct machine_function): Add tbegin_p.
16310 (s390_canonicalize_comparison): Fold CC mode compares to
16311 conditional jump if possible.
16312 (s390_emit_jump): Return the emitted jump.
16313 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
16314 Handle CCRAWmode compares.
16315 (s390_option_override): Default to -mhtm if available.
16316 (s390_reg_clobbered_rtx): Handle floating point regs as well.
16317 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
16318 FPRs instead of df_regs_ever_live_p.
16319 (s390_optimize_nonescaping_tx): New function.
16320 (s390_init_frame_layout): Extend clobbered_regs array to cover
16322 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
16323 (s390_expand_tbegin): New function.
16324 (enum s390_builtin): New enum definition.
16325 (code_for_builtin): New array definition.
16326 (s390_init_builtins): New function.
16327 (s390_expand_builtin): New function.
16328 (TARGET_INIT_BUILTINS): Define.
16329 (TARGET_EXPAND_BUILTIN): Define.
16330 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
16331 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
16332 (s390_alc_comparison): Likewise.
16333 * config/s390/s390-modes.def: Add CCRAWmode.
16334 * config/s390/s390.h (processor_flags): Add PF_TX.
16335 (TARGET_CPU_HTM): Define macro.
16336 (TARGET_HTM): Define macro.
16337 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
16338 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
16339 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16340 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
16341 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
16342 (TBEGIN_MASK, TBEGINC_MASK): New constants.
16343 ("*cc_to_int"): Move up.
16344 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
16345 constants other than 0.
16346 ("*ccraw_to_int"): New insn and splitter definition.
16347 ("tbegin", "tbegin_nofloat", "tbegin_retry")
16348 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
16349 ("tx_assist"): New expander.
16350 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
16351 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
16352 * config/s390/s390.opt: Add -mhtm option.
16353 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
16354 * config/s390/htmxlintrin.h: New file.
16355 * config/s390/htmintrin.h: New file.
16356 * config/s390/s390intrin.h: New file.
16357 * doc/extend.texi: Document htm builtins.
16358 * config.gcc: Add the new header files to extra_headers.
16360 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
16362 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
16363 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
16365 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
16366 Pat Haugen <pthaugen@us.ibm.com>
16367 Peter Bergner <bergner@vnet.ibm.com>
16369 * config/rs6000/power8.md: New.
16370 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
16371 setting for power8 entry.
16372 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
16373 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
16374 test for Power4/Power5 only.
16375 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
16377 (force_new_group): Adjust comment.
16378 * config/rs6000/rs6000.md: Include power8.md.
16380 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
16382 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
16383 * config/arm/arm-protos.h (arm_max_conditional_execute): New
16385 (tune_params): Update comment.
16386 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
16387 (arm_max_conditional_execute): New function.
16388 (thumb2_final_prescan_insn): Use max_insn_skipped and
16389 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
16391 2013-06-25 Jakub Jelinek <jakub@redhat.com>
16393 PR tree-optimization/57705
16394 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
16395 SSA_NAME step, provided that it is not defined inside the loop.
16396 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
16397 (get_initial_def_for_induction): Handle SSA_NAME IV step.
16399 2013-06-25 Martin Jambor <mjambor@suse.cz>
16401 PR middle-end/57670
16402 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
16403 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
16405 (ipa_note_param_call): Initialize member_ptr flag.
16406 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
16407 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
16408 (ipa_write_indirect_edge_info): Stream member_ptr flag.
16409 (ipa_read_indirect_edge_info): Likewise.
16411 2013-06-25 Richard Biener <rguenther@suse.de>
16413 PR middle-end/56977
16414 * passes.c (init_optimization_passes): Move pass_fold_builtins
16415 and pass_dce earlier with -Og.
16417 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
16419 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
16420 <BIT_FIELD_REF>: Remove trailing TAB.
16421 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
16424 2013-06-24 Martin Jambor <mjambor@suse.cz>
16426 PR tree-optimization/57358
16427 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
16428 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
16429 (ipa_analyze_params_uses): Generate pessimistic info when true.
16431 2013-06-24 Martin Jambor <mjambor@suse.cz>
16433 PR tree-optimization/57539
16434 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
16435 global.inlined_to of the new node to it. All callers changed.
16436 * ipa-inline-transform.c (clone_inlined_nodes): New variable
16437 inlining_into, pass it to cgraph_clone_node.
16438 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
16439 ipa_free_edge_args_substructures.
16440 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
16441 rdesc linked list. Do not assert rdesc edges have inlined caller.
16442 Assert we have found an rdesc in the rdesc list.
16444 2013-06-24 Richard Biener <rguenther@suse.de>
16446 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
16447 (pointer_set_lookup): Declare.
16448 (class pointer_map): New template class implementing a
16449 generic pointer to T map.
16450 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
16451 pointer_map<T>::contains, pointer_map<T>::insert,
16452 pointer_map<T>::traverse): New functions.
16453 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
16454 (pointer_set_lookup): New function.
16455 (pointer_set_contains): Use pointer_set_lookup.
16456 (pointer_set_insert): Likewise.
16457 (insert_aux): Remove.
16458 (struct pointer_map_t): Embed a pointer_set_t.
16459 (pointer_map_create): Adjust.
16460 (pointer_map_destroy): Likewise.
16461 (pointer_map_contains): Likewise.
16462 (pointer_map_insert): Likewise.
16463 (pointer_map_traverse): Likewise.
16464 * tree-streamer.h (struct streamer_tree_cache_d): Use a
16465 pointer_map<unsigned> instead of a pointer_map_t.
16466 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
16467 (streamer_tree_cache_lookup): Likewise.
16468 (streamer_tree_cache_create): Likewise.
16469 (streamer_tree_cache_delete): Likewise.
16470 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
16471 pointer_map<unsigned> instead of a pointer_map_t.
16472 (lto_init_tree_ref_encoder): Adjust.
16473 (lto_destroy_tree_ref_encoder): Likewise.
16474 * lto-section-out.c (lto_output_decl_index): Likewise.
16475 (lto_record_function_out_decl_state): Likewise.
16476 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
16478 2013-06-24 Richard Biener <rguenther@suse.de>
16480 PR tree-optimization/57488
16481 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
16483 2013-06-24 Alan Modra <amodra@gmail.com>
16485 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
16486 (gen_easy_altivec_constant): Likewise.
16487 * config/rs6000/predicates.md (easy_vector_constant_add_self,
16488 easy_vector_constant_msb): Likewise.
16490 2013-06-23 Jakub Jelinek <jakub@redhat.com>
16493 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
16494 add missing return true.
16496 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
16499 * config/sh/predicates.md (general_extend_operand): Invoke
16500 general_movsrc_operand for memory operands.
16501 (general_movsrc_operand): Allow reg+reg addressing, do not use
16502 general_operand for memory operands.
16504 2013-06-23 Sriraman Tallam <tmsriram@google.com>
16506 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
16507 when current target options does not apply.
16508 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
16509 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
16510 * config/i386/bmiintrin.h: Pass appropriate target
16511 attributes to header.
16512 * config/i386/mmintrin.h: Ditto.
16513 * config/i386/nmmintrin.h: Ditto.
16514 * config/i386/avx2intrin.h: Ditto.
16515 * config/i386/fxsrintrin.h: Ditto.
16516 * config/i386/tbmintrin.h: Ditto.
16517 * config/i386/xsaveintrin.h: Ditto.
16518 * config/i386/f16cintrin.h: Ditto.
16519 * config/i386/xtestintrin.h: Ditto.
16520 * config/i386/xsaveoptintrin.h: Ditto.
16521 * config/i386/bmi2intrin.h: Ditto.
16522 * config/i386/lzcntintrin.h: Ditto.
16523 * config/i386/smmintrin.h: Ditto.
16524 * config/i386/wmmintrin.h: Ditto.
16525 * config/i386/x86intrin.h: Remove all header include guards.
16526 * config/i386/prfchwintrin.h: Ditto.
16527 * config/i386/pmmintrin.h: Ditto.
16528 * config/i386/tmmintrin.h: Ditto.
16529 * config/i386/xmmintrin.h: Ditto.
16530 * config/i386/popcntintrin.h: Ditto.
16531 * config/i386/rdseedintrin.h: Ditto.
16532 * config/i386/ammintrin.h: Ditto.
16533 * config/i386/emmintrin.h: Ditto.
16534 * config/i386/immintrin.h: Remove all header include guards.
16535 * config/i386/fma4intrin.h: Ditto.
16536 * config/i386/lwpintrin.h: Ditto.
16537 * config/i386/xopintrin.h: Ditto.
16538 * config/i386/ia32intrin.h: Ditto.
16539 * config/i386/avxintrin.h: Ditto.
16540 * config/i386/rtmintrin.h: Ditto.
16541 * config/i386/fmaintrin.h: Ditto.
16542 * config/i386/mm3dnow.h: Ditto.
16544 2013-06-22 Sriraman Tallam <tmsriram@google.com>
16546 * common/config/i386/i386-common.c: Handle LZCNT.
16548 2013-06-22 Andi Kleen <ak@linux.intel.com>
16550 * doc/extend.texi: Use __atomic_store_n instead of
16551 __atomic_store in HLE example.
16553 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
16555 * config/sh/sh.c: Remove <cstdlib> workaround.
16557 2013-06-21 Andi Kleen <ak@linux.intel.com>
16559 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
16561 2013-06-21 Andi Kleen <ak@linux.intel.com>
16563 * doc/extend.texi: Document that __atomic_clear and
16564 __atomic_test_and_set should only be used with bool.
16566 2013-06-20 Jan Hubicka <jh@suse.cz>
16568 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
16569 types_same_for_odr.
16570 * tree.c (decls_same_for_odr): New function.
16571 (same_for_edr): New function.
16572 (types_same_for_odr): New function.
16573 (get_binfo_at_offset): Use it.
16574 * tree.h (types_same_for_odr): Declare.
16576 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
16577 Jason Merrill <jason@redhat.com>
16579 * system.h: Include <cstdlib> as well as <stdlib.h>.
16581 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
16584 * config/i386/i386.c (construct_container): Report error if
16585 long double is used with disabled x87 float returns.
16587 2013-06-20 Jan Hubicka <jh@suse.cz>
16589 * lto-cgraph.c (input_symtab): Do not set cgraph state.
16591 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
16593 PR rtl-optimization/57425
16594 PR rtl-optimization/57569
16595 * alias.c (write_dependence_p): Remove parameters mem_mode and
16596 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
16597 Changed all callers.
16598 (canon_anti_dependence): Get comments and semantics in sync.
16599 Add parameter mem_canonicalized. Changed all callers.
16600 * rtl.h (canon_anti_dependence): Update prototype.
16602 2013-06-20 Richard Biener <rguenther@suse.de>
16604 * data-streamer-in.c (streamer_read_uhwi): Optimize single
16605 byte case, inline streamer_read_uchar and defer section
16608 2013-06-20 Richard Biener <rguenther@suse.de>
16610 PR tree-optimization/57584
16611 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
16612 SSA names into the expanded expression that take part in
16613 abnormal coalescing.
16615 2013-06-19 Sharad Singhai <singhai@google.com>
16617 * gcov.c (print_usage): Handle new option.
16618 (process_args): Ditto.
16619 (get_gcov_intermediate_filename): New function.
16620 (output_intermediate_file): New function.
16621 (output_gcov_file): New function
16622 (generate_results): Handle new option.
16623 (release_function): Relase demangled name.
16624 (read_graph_file): Handle demangled name.
16625 (output_lines): Ditto.
16626 * doc/gcov.texi: Document gcov intermediate format.
16628 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
16631 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
16632 (lra_emit_add): Use the functions. Add comment about Y as an
16635 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
16638 * collect2.c (collect_atexit): New.
16639 (collect_exit): Delete.
16640 (main): Register collect_atexit with atexit.
16641 (collect_wait): Change collect_exit to exit.
16643 * collect2.h (collect_exit): Delete.
16644 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
16646 2013-06-19 Wei Mi <wmi@google.com>
16648 PR rtl-optimization/57518
16649 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
16650 if regno is used in paradoxical subreg.
16651 (update_equiv_regs): Check pdx_subregs[regno] before
16652 set a reg to be equivalent with a mem.
16654 2013-06-19 Matthias Klose <doko@ubuntu.com>
16657 * file-find.h (find_a_file): Add a mode parameter.
16658 * file-find.c (find_a_file): Likewise.
16659 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
16660 with X_OK for the executables.
16661 * collect2.c (main): Call find_a_file with X_OK.
16663 2013-06-19 Steve Ellcey <sellcey@mips.com>
16666 * config/mips/mips.md (casesi_internal_mips16_<mode>):
16667 Use NEXT_INSN instead of next_real_insn.
16669 2013-06-19 Jan Hubicka <jh@suse.cz>
16671 * cgraph.h (const_value_known_p): Replace by ...
16672 (ctor_for_folding): .. this one.
16673 * cgraphunit.c (process_function_and_variable_attributes): Use it.
16674 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
16675 * expr.c (expand_expr_real_1): Likewise.
16676 (string_constant): Likewise.
16677 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
16678 * ipa.c (process_references): Likewise.
16679 (symtab_remove_unreachable_nodes): Likewise.
16680 * ipa-inline-analysis.c (param_change_prob): Likewise.
16681 * gimple-fold.c (canonicalize_constructor_val): Likewise.
16682 (get_base_constructor): Likwise.
16683 * varpool.c (varpool_remove_node): Likewise.
16684 (varpool_remove_initializer): LIkewise.
16685 (dump_varpool_node): LIkwise.
16686 (const_value_known_p): Rewrite to ...
16687 (ctor_for_folding): ... this one.
16689 2013-06-19 Jakub Jelinek <jakub@redhat.com>
16692 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
16693 PERSONALITY in $PATH derived prefixes.
16695 2013-06-19 Jeff Law <law@redhat.com>
16697 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
16700 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
16701 (simplify_bitwise_binary): Use it to simpify certain binary ops on
16704 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
16706 * config/arm/vfp.md: Move VFP instruction classification documentation
16708 * config/arm/arm.md: ... here. Update instruction classification
16711 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
16713 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
16714 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
16715 pattern. Use more efficient sequences on ARMv5 and Thumb2.
16717 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
16720 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16721 with NEXT_INSN. Use tablejump_p to check for jump table data
16724 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
16727 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
16728 that now in C++ the value is correct per the C++ standards.
16730 2013-06-19 Richard Biener <rguenther@suse.de>
16732 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
16733 for global context.
16735 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16738 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16741 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16742 with next_active_insn.
16744 2013-06-18 Sriraman Tallam <tmsriram@google.com>
16746 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
16747 functions are inlined during failures to flag an error.
16748 * tree-inline.c (expand_call_inline): Allow the error to be flagged
16749 in early inline pass.
16751 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
16753 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
16756 2013-06-18 Julian Brown <julian@codesourcery.com>
16758 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
16759 Permit virtual register pre-reload if !strict.
16760 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
16762 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
16764 * config/arm/neon.md (movmisalign<mode>): Use
16765 neon_perm_struct_or_reg_operand instead of
16766 neon_struct_or_register_operand.
16767 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
16768 neon_permissive_struct_operand instead of neon_struct_operand.
16769 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
16770 neon_vector_mem_operand.
16771 * config/arm/predicates.md (neon_struct_operand): Adjust call to
16772 neon_vector_mem_operand.
16773 (neon_permissive_struct_operand): New.
16774 (neon_struct_or_register_operand): Rename to...
16775 (neon_perm_struct_or_reg_operand): This. Adjust call to
16776 neon_vector_mem_operand.
16778 2013-06-18 Richard Biener <rguenther@suse.de>
16780 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
16781 * lto-streamer.h: Include pointer-set.h.
16782 (struct lto_decl_slot): Remove.
16783 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
16784 Remove next_index entry.
16785 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16786 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16787 (lto_init_tree_ref_encoder): Adjust.
16788 (lto_destroy_tree_ref_encoder): Likewise.
16789 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16790 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16791 (lto_output_decl_index): Adjust.
16792 (lto_new_out_decl_state): Likewise.
16793 (lto_record_function_out_decl_state): Likewise.
16794 * lto-streamer-out.c (copy_function): Likewise.
16796 2013-06-18 Richard Biener <rguenther@suse.de>
16798 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
16799 * cgraphunit.c: Include cfgloop.h.
16800 (init_lowered_empty_function): Initialize the loop tree.
16801 (assemble_thunk): Insert new BBs into loops.
16803 2013-06-18 Richard Biener <rguenther@suse.de>
16805 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
16806 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
16807 the map from cache entry to cache index optional.
16808 (streamer_tree_cache_replace_tree): Adjust accordingly.
16809 (streamer_tree_cache_append): Likewise.
16810 (streamer_tree_cache_delete): Likewise.
16811 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
16812 streamer cache map from cache entry to cache index.
16813 * lto-streamer-out.c (create_output_block): Adjust.
16815 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
16817 * config/arm/arm.md (attribute "insn"): Move multiplication and
16818 division attributes to...
16819 (attribute "type"): ... here. Remove mult.
16820 (attribute "mul32"): New attribute.
16821 (attribute "mul64"): Add umaal.
16822 (*arm_mulsi3): Update attributes.
16823 (*arm_mulsi3_v6): Likewise.
16824 (*thumb_mulsi3): Likewise.
16825 (*thumb_mulsi3_v6): Likewise.
16826 (*mulsi3_compare0): Likewise.
16827 (*mulsi3_compare0_v6): Likewise.
16828 (*mulsi_compare0_scratch): Likewise.
16829 (*mulsi_compare0_scratch_v6): Likewise.
16830 (*mulsi3addsi): Likewise.
16831 (*mulsi3addsi_v6): Likewise.
16832 (*mulsi3addsi_compare0): Likewise.
16833 (*mulsi3addsi_compare0_v6): Likewise.
16834 (*mulsi3addsi_compare0_scratch): Likewise.
16835 (*mulsi3addsi_compare0_scratch_v6): Likewise.
16836 (*mulsi3subsi): Likewise.
16837 (*mulsidi3adddi): Likewise.
16838 (*mulsi3addsi_v6): Likewise.
16839 (*mulsidi3adddi_v6): Likewise.
16840 (*mulsidi3_nov6): Likewise.
16841 (*mulsidi3_v6): Likewise.
16842 (*umulsidi3_nov6): Likewise.
16843 (*umulsidi3_v6): Likewise.
16844 (*umulsidi3adddi): Likewise.
16845 (*umulsidi3adddi_v6): Likewise.
16846 (*smulsi3_highpart_nov6): Likewise.
16847 (*smulsi3_highpart_v6): Likewise.
16848 (*umulsi3_highpart_nov6): Likewise.
16849 (*umulsi3_highpart_v6): Likewise.
16850 (mulhisi3): Likewise.
16851 (*mulhisi3tb): Likewise.
16852 (*mulhisi3bt): Likewise.
16853 (*mulhisi3tt): Likewise.
16854 (maddhisi4): Likewise.
16855 (*maddhisi4tb): Likewise.
16856 (*maddhisi4tt): Likewise.
16857 (maddhidi4): Likewise.
16858 (*maddhidi4tb): Likewise.
16859 (*maddhidi4tt): Likewise.
16860 (divsi3): Likewise.
16861 (udivsi3): Likewise.
16862 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
16863 (thumb2_mulsi_short_compare0): Likewise.
16864 (thumb2_mulsi_short_compare0_scratch): Likewise.
16865 * config/arm/arm1020e.md (1020mult1): Update attribute change.
16866 (1020mult2): Likewise.
16867 (1020mult3): Likewise.
16868 (1020mult4): Likewise.
16869 (1020mult5): Likewise.
16870 (1020mult6): Likewise.
16871 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
16873 (cortex_a15_mult64): Likewise.
16874 (cortex_a15_sdiv): Likewise.
16875 (cortex_a15_udiv): Likewise.
16876 * config/arm/arm1026ejs.md (mult1): Update attribute change.
16882 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
16883 (pj4_ir_div): Likewise.
16884 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
16885 (11_mult2): Likewise.
16886 (11_mult3): Likewise.
16887 (11_mult4): Likewise.
16888 (11_mult5): Likewise.
16889 (11_mult6): Likewise.
16890 (11_mult7): Likewise.
16891 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
16892 (cortex_a8_mla): Likewise.
16893 (cortex_a8_mull): Likewise.
16894 (cortex_a8_smulwy): Likewise.
16895 (cortex_a8_smlald): Likewise.
16896 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
16897 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
16898 (cortex_r4_mul_3): Likewise.
16899 (cortex_r4_mla_4): Likewise.
16900 (cortex_r4_mla_3): Likewise.
16901 (cortex_r4_smlald): Likewise.
16902 (cortex_r4_mull): Likewise.
16903 (cortex_r4_sdiv): Likewise.
16904 (cortex_r4_udiv): Likewise.
16905 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
16906 (cortex_a7_idiv): Likewise.
16907 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
16908 (9_mult2): Likewise.
16909 (9_mult3): Likewise.
16910 (9_mult4): Likewise.
16911 (9_mult5): Likewise.
16912 (9_mult6): Likewise.
16913 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
16914 (cortex_a53_sdiv): Likewise.
16915 (cortex_a53_udiv): Likewise.
16916 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
16917 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
16918 (mp626_mult2): Likewise.
16919 (mp626_mult3): Likewise.
16920 (mp626_mult4): Likewise.
16921 * config/arm/fa526.md (526_mult1): Update attribute change.
16922 (526_mult2): Likewise.
16923 * config/arm/arm-generic.md (mult): Update attribute change.
16924 (mult_ldsched_strongarm): Likewise.
16925 (mult_ldsched): Likewise.
16926 (multi_cycle): Likewise.
16927 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
16928 * config/arm/fa606te.md (606te_mult1): Update attribute change.
16929 (606te_mult2): Likewise.
16930 (606te_mult3): Likewise.
16931 (606te_mult4): Likewise.
16932 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
16933 (cortex_a9_mac16): Likewise.
16934 (cortex_a9_multiply): Likewise.
16935 (cortex_a9_mac): Likewise.
16936 (cortex_a9_multiply_long): Likewise.
16937 * config/arm/fa626te.md (626te_mult1): Update attribute change.
16938 (626te_mult2): Likewise.
16939 (626te_mult3): Likewise.
16940 (626te_mult4): Likewise.
16942 2013-06-18 Richard Biener <rguenther@suse.de>
16945 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
16947 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16950 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16951 with next_active_insn.
16953 2013-06-18 Alan Modra <amodra@gmail.com>
16955 * config/rs6000/rs6000.h (enum data_align): New.
16956 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
16957 (DATA_ABI_ALIGNMENT): Define.
16958 (CONSTANT_ALIGNMENT): Correct comment.
16959 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
16960 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
16962 2013-06-17 David Malcolm <dmalcolm@redhat.com>
16964 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
16965 ATTRIBUTE_UNUSED marking.
16967 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
16969 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
16970 alternative and update.
16971 (aarch64_dup_lanedi): Delete.
16972 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
16973 * config/aarch64/aarch64-simd-builtins.def: Update.
16975 2013-06-17 Richard Biener <rguenther@suse.de>
16977 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
16978 (lto_input_scc): Declare.
16979 (lto_input_tree_1): Likewise.
16980 (struct lto_stats_d): Add num_tree_bodies_output and
16981 num_pickle_refs_output.
16982 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
16983 (lto_read_tree_1): Split out from ...
16984 (lto_read_tree): ... this.
16985 (lto_input_scc): New function.
16986 (lto_input_tree_1): Split out from ...
16987 (lto_input_tree): ... this. Handle LTO_tree_scc.
16988 (lto_data_in_create): Create the streamer cache without hashes.
16989 * lto-streamer-out.c (create_output_block): Create the streamer
16990 cache with hashes when not doing WPA.
16991 (lto_write_tree_1): Split out from ...
16992 (lto_write_tree): ... this.
16993 (get_symbol_initial_value): New function.
16994 (lto_output_tree_1): Split out from ...
16995 (lto_output_tree): ... this. Write trees as series of SCCs
16996 using a DFS walk via DFS_write_tree.
16997 (struct sccs, struct scc_entry): New types.
16998 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
16999 (DFS_write_tree_body): New function.
17000 (DFS_write_tree): Likewise.
17001 (hash_tree): Likewise.
17002 (scc_entry_compare): Likewise.
17003 (hash_scc): Likewise.
17004 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
17005 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
17006 TREE_CHAIN as regular reference.
17007 (streamer_read_integer_cst): Remove.
17008 (streamer_get_pickled_tree): Adjust.
17009 * tree-streamer-out.c (streamer_write_chain): Disable streaming
17010 of DECL_EXTERNALs in BLOCK_VARS for now.
17011 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
17013 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
17014 Add hash value argument and record that if hashes are recorded
17016 (streamer_tree_cache_insert_1): Adjust.
17017 (streamer_tree_cache_insert): Likewise.
17018 (streamer_tree_cache_insert_at): Rename to ...
17019 (streamer_tree_cache_replace_tree): ... this and adjust.
17020 (streamer_tree_cache_append): Adjust.
17021 (record_common_node): Likewise.
17022 (streamer_tree_cache_create): Add argument whether to
17023 record hash values together with trees.
17024 (streamer_tree_cache_delete): Adjust.
17025 * tree-streamer.h (struct streamer_tree_cache_d): Add
17027 (streamer_read_integer_cst): Remove.
17028 (streamer_tree_cache_insert): Adjust.
17029 (streamer_tree_cache_append): Likewise.
17030 (streamer_tree_cache_insert_at): Rename to ...
17031 (streamer_tree_cache_replace_tree): ... this and adjust.
17032 (streamer_tree_cache_create): Add argument whether to record hashes.
17033 (streamer_tree_cache_get): Rename to ...
17034 (streamer_tree_cache_get_tree): ... this.
17035 (streamer_tree_cache_get_hash): New function.
17036 * tree.c (cache_integer_cst): New function.
17037 * tree.h (cache_integer_cst): Declare.
17038 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
17039 * lto-symtab.c (lto_varpool_replace_node): Only release
17040 DECL_INITIAL of non-prevailing decls.
17041 * varpool.c (varpool_remove_initializer): Do not release
17042 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
17044 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17046 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
17047 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
17048 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
17049 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
17050 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
17051 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
17052 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
17053 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
17054 instead of TARGET_64BIT.
17055 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
17056 Require ISA_HAS_<D>DIV.
17058 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
17060 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
17061 (mips*-*-linux*): Move default with_llsc setting to where other
17063 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
17065 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
17066 Likewise. Remove default with_tune setting. Move default float
17067 setting to its own block. Handle with_llsc in the same block as above.
17069 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
17071 PR rtl-optimization/57425
17072 PR rtl-optimization/57569
17073 * alias.c (write_dependence_p): Add new parameters mem_mode,
17074 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
17075 Changed all callers.
17076 (canon_anti_dependence): New function.
17077 * cse.c (check_dependence): Use canon_anti_dependence.
17078 * cselib.c (cselib_invalidate_mem): Likewise.
17079 * rtl.h (canon_anti_dependence): Declare.
17081 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17083 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
17084 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
17085 ".set mips3" for 64-bit targets.
17087 2013-06-15 Dehao Chen <dehao@google.com>
17089 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
17090 * gimple-low.c (gimple_check_call_matching_types): Likewise.
17091 (gimple_check_call_args): Likewise.
17092 * value-prof.c (check_ic_target): Likewise.
17093 * ipa-inline.c (early_inliner): Likewise.
17094 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
17095 * cgraph.c (cgraph_create_edge_1): Likewise.
17096 (cgraph_make_edge_direct): Likewise.
17098 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
17101 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
17102 rs6000_output_move_128bit to handle emitting quad memory
17103 operations. Set attribute length to 8 bytes.
17105 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
17107 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
17109 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
17110 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
17111 (aarch64_<su>mlsl<mode>): Likewise.
17113 2013-06-14 Mike Stump <mikestump@comcast.net>
17115 * Makefile.in (TARGET_H): Add insn-codes.h.
17117 2013-06-14 Alan Modra <amodra@gmail.com>
17119 PR middle-end/57134
17120 PR middle-end/57586
17121 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
17122 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
17123 bitfield expansion when EXPAND_MEMORY.
17124 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
17126 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
17128 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
17129 test for clearing quad memory on 32-bit later.
17131 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17133 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
17134 (fold_negate_expr): Likewise.
17135 (fold_real_zero_addition_p): Handle vectors.
17136 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
17138 2013-06-14 Alan Modra <amodra@gmail.com>
17140 * varasm.c (force_const_mem): Revert 2013-06-07 change.
17142 2013-06-13 Jan Hubicka <jh@suse.cz>
17144 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17145 Local comdats are not externally visible.
17146 * symtab.c (dump_symtab_base): Dump externally visible.
17147 (verify_symtab_base): Verify back links in the symtab hash.
17149 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17151 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
17152 CONVERT_EXPR as equal nodes.
17154 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17156 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
17158 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17160 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
17161 Generalize to complex and vector.
17162 * tree.c (build_all_ones_cst): New function.
17163 * tree.h (build_all_ones_cst): Declare it.
17165 2013-06-13 Alan Modra <amodra@gmail.com>
17167 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
17168 * config/rs6000/rs6000.md (signbittf2): New insn.
17169 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
17170 (abstf2_internal, cmptf_internal2): Likewise.
17171 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
17173 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
17174 Pat Haugen <pthaugen@us.ibm.com>
17175 Peter Bergner <bergner@vnet.ibm.com>
17177 * config/rs6000/rs6000.c (emit_load_locked): Add support for
17178 power8 byte, half-word, and quad-word atomic instructions.
17179 (emit_store_conditional): Likewise.
17180 (rs6000_expand_atomic_compare_and_swap): Likewise.
17181 (rs6000_expand_atomic_op): Likewise.
17183 * config/rs6000/sync.md (larx): Add new modes for power8.
17185 (AINT): New mode iterator to include TImode as well as normal
17186 integer modes on power8.
17187 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
17188 that VSX registers are not considered. Use AINT mode iterator
17189 instead of INT1 to allow inclusion of quad word atomic operations
17191 (load_locked<mode>): Likewise.
17192 (store_conditional<mode>): Likewise.
17193 (atomic_compare_and_swap<mode>): Likewise.
17194 (atomic_exchange<mode>): Likewise.
17195 (atomic_nand<mode>): Likewise.
17196 (atomic_fetch_<fetchop_name><mode>): Likewise.
17197 (atomic_nand_fetch<mode>): Likewise.
17198 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
17200 (ATOMIC): On power8, add QImode, HImode modes.
17201 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
17202 modes that promote to SImode.
17203 (load_lockedti): Convert TImode arguments to PTImode, so that we
17204 get a guaranteed even/odd register pair.
17205 (load_lockedpti): Likewise.
17206 (store_conditionalti): Likewise.
17207 (store_conditionalpti): Likewise.
17209 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
17210 atomic load/store instructions.
17213 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
17215 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
17217 (insn_count): New attribute, with most cases extracted from...
17218 (length): ...here. Redefine most cases in terms of insn_count.
17219 (single_insn): Delete.
17220 (can_delay): Use insn_count to check for single instructions.
17221 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
17222 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
17223 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
17224 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
17225 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
17226 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
17227 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
17228 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
17229 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
17230 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
17231 rather than "length".
17232 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
17233 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
17234 Use "insn_count" rather than "length".
17235 * config/mips/mips-dsp.md
17236 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
17237 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
17240 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
17242 PR tree-optimization/57361
17243 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
17245 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
17247 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
17249 (aarch64_simd_combine<mode>): New instruction expansion.
17250 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
17251 function prototype.
17252 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
17253 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
17255 2013-06-12 Jan Hubicka <jh@suse.cz>
17257 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
17258 decl has when in streaming stage.
17259 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
17260 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
17262 2013-06-12 Roland Stigge <stigge@antcom.de>
17265 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
17267 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17269 PR tree-optimization/57537
17270 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
17271 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
17273 2013-06-12 Richard Biener <rguenther@suse.de>
17275 * data-streamer.h (streamer_write_char_stream): CSE
17276 obs->current_pointer.
17277 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
17278 streamer_write_char_stream manually and optimize the resulting loop.
17279 (streamer_write_hwi_stream): Likewise.
17281 2013-06-12 Jan Hubicka <jh@suse.cz>
17283 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
17284 * cgraph.h (varpool_create_empty_node): Declare.
17285 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
17287 * symtab.c (symtab_unregister_node): Be lax about missin entries
17289 (symtab_get_node): Update comment.
17290 * varpool.c (varpool_create_empty_node): Break out from ...
17291 (varpool_node_for_decl): ... here.
17292 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
17294 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
17296 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
17297 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
17298 part. Use straight-line flow at the end.
17299 <COMPONENT_REF>: Remove superfluous else.
17300 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
17302 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17305 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
17306 target hook even for !TREE_PUBLIC decls. If no resolution info
17307 is available, return false for common and external decls.
17309 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17311 * config/rl78/constraints.md (U): New constraint.
17312 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
17315 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17318 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
17319 to allow returning address to AT_PLATFORM name.
17321 2013-06-11 Jan Hubicka <jh@suse.cz>
17323 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
17324 * cgraph.h (symtab_node_base): Add weakref flag.
17325 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
17326 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
17327 (output_weakrefs): Use weakref flag.
17328 * fold-const.c (simple_operand_p): Handle WEAK.
17329 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
17330 * ipa.c (varpool_externally_visible_p): Drop weakref.
17331 (function_and_variable_visibility): Update comment; fix weakref
17332 sanity checks; do not clear DECL_WEAK on them.
17333 * lto-cgraph.c (lto_output_node): update.
17334 (lto_output_varpool_node): Update.
17335 (input_overwrite_node): Update.
17336 (input_node): Update.
17337 (input_varpool_node): Update.
17338 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
17339 (lto_symtab_merge_symbols): Add sanity check.
17340 (lto_symtab_prevailing_decl): Do not special case weakrefs.
17341 * passes.c (rest_of_decl_compilation): Set static flag, too.
17342 * symtab.c (dump_symtab_base): Dump weakref.
17343 (verify_symtab_base): Sanity check weakrefs.
17344 (symtab_make_decl_local): Remove duplicated code.
17345 (symtab_alias_ultimate_target): Simplify.
17346 * varpool.c (varpool_create_variable_alias): Set weakref flag.
17348 2013-06-11 Tom de Vries <tom@codesourcery.com>
17350 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
17351 sequence_vect == NULL.
17353 2013-06-11 DJ Delorie <dj@redhat.com>
17355 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
17356 (rl78_unwind_word_mode): New.
17358 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17360 * final.c (debug_prefix_maps): Make static.
17362 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17364 * function.c (initial_trampoline): Remove stray copy.
17366 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
17368 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
17370 2013-06-11 Martin Jambor <mjambor@suse.cz>
17372 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
17373 within bounds at the beginning of the function.
17375 2013-06-11 Alan Modra <amodra@gmail.com>
17377 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
17379 (get_named_section): Don't NULL !DECL_P decl.
17381 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
17383 * doc/invoke.texi (core-avx2): Document.
17385 (atom): Updated with MOVBE.
17387 2013-06-11 Richard Biener <rguenther@suse.de>
17389 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
17391 2013-06-11 Anton Blanchard <anton@samba.org>
17393 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
17394 correct shift value in little-endian mode.
17396 2013-06-11 Jakub Jelinek <jakub@redhat.com>
17399 * varasm.c (get_variable_align): Move #endif to the right place.
17401 2013-06-10 Cary Coutant <ccoutant@google.com>
17403 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
17404 for hash so that hash table traversal order is deterministic.
17406 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
17407 Pat Haugen <pthaugen@us.ibm.com>
17408 Peter Bergner <bergner@vnet.ibm.com>
17410 * config/rs6000/vector.md (GPR move splitter): Do not split moves
17411 of vectors in GPRS if they are direct moves or quad word load or
17414 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
17416 (direct_move_p): Likewise.
17417 (quad_load_store_p): Likewise.
17419 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
17420 classes into bins based on the physical register type.
17421 (reg_class_to_reg_type): Likewise.
17422 (IS_STD_REG_TYPE): Likewise.
17423 (IS_FP_VECT_REG_TYPE): Likewise.
17424 (reload_fpr_gpr): Arrays to determine what insn to use if we can
17425 use direct move instructions.
17426 (reload_gpr_vsx): Likewise.
17427 (reload_vsx_gpr): Likewise.
17428 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
17429 information that is a simplification of register classes. Also
17430 precalculate direct move reload helpers.
17431 (direct_move_p): New function to return true if the operation can
17432 be done as a direct move instruciton.
17433 (quad_load_store_p): New function to return true if the operation
17434 is a quad memory operation.
17435 (rs6000_legitimize_address): If quad memory, only allow register
17436 indirect for TImode addresses.
17437 (rs6000_legitimate_address_p): Likewise.
17438 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
17439 (rs6000_reload_register_type): Likewise.
17440 (register_to_reg_type): Return register type.
17441 (rs6000_secondary_reload_simple_move): New helper function for
17442 secondary reload and secondary memory needed to identify anything
17443 that is a simple move, and does not need reloading.
17444 (rs6000_secondary_reload_direct_move): New helper function for
17445 secondary reload to identify cases that can be done with several
17446 instructions via the direct move instructions.
17447 (rs6000_secondary_reload_move): New helper function for secondary
17448 reload to identify moves between register types that can be done.
17449 (rs6000_secondary_reload): Add support for quad memory operations
17450 and for direct move.
17451 (rs6000_secondary_memory_needed): Likewise.
17452 (rs6000_debug_secondary_memory_needed): Change argument names.
17453 (rs6000_output_move_128bit): New function to return the move to
17454 use for 128-bit moves, including knowing about the various
17455 limitations of quad memory operations.
17457 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
17458 memory operations. call rs6000_output_move_128bit for the actual
17459 instruciton(s) to generate.
17460 (vsx_movti_64bit): Likewise.
17462 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
17463 (UNSPEC_P8V_MTVSRWZ): Likewise.
17464 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
17465 (UNSPEC_P8V_MTVSRD): Likewise.
17466 (UNSPEC_P8V_XXPERMDI): Likewise.
17467 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
17468 (UNSPEC_FUSION_GPR): Likewise.
17469 (FMOVE128_GPR): New iterator for direct move.
17470 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
17471 (f32_sv): Likewise.
17472 (f32_dm): Likewise.
17473 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
17474 loads and direct move instructions.
17475 (zero_extendsidi2_lfiwzx): Likewise.
17476 (extendsidi2_lfiwax): Likewise.
17477 (extendsidi2_nocell): Likewise.
17478 (floatsi<mode>2_lfiwax): Likewise.
17479 (lfiwax): Likewise.
17480 (floatunssi<mode>2_lfiwzx): Likewise.
17481 (lfiwzx): Likewise.
17482 (fix_trunc<mode>_stfiwx): Likewise.
17483 (fixuns_trunc<mode>_stfiwx): Likewise.
17484 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
17485 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
17486 (parity<mode>2_cmpb): Set length/type attr.
17487 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
17488 for 'mr.' to fast_compare.
17489 (bpermd_<mode>): Change type attr to popcnt.
17490 (p8_fmrgow_<mode>): New insns for power8 direct move support.
17491 (p8_mtvsrwz_1): Likewise.
17492 (p8_mtvsrwz_2): Likewise.
17493 (reload_fpr_from_gpr<mode>): Likewise.
17494 (p8_mtvsrd_1): Likewise.
17495 (p8_mtvsrd_2): Likewise.
17496 (p8_xxpermdi_<mode>): Likewise.
17497 (reload_vsx_from_gpr<mode>): Likewise.
17498 (reload_vsx_from_gprsf): Likewise.
17499 (p8_mfvsrd_3_<mode>): LIkewise.
17500 (reload_gpr_from_vsx<mode>): Likewise.
17501 (reload_gpr_from_vsxsf): Likewise.
17502 (p8_mfvsrd_4_disf): Likewise.
17503 (multi-word GPR splits): Do not split direct moves or quad memory
17506 2013-06-10 David Malcolm <dmalcolm@redhat.com>
17508 * tree-into-ssa.c (interesting_blocks): Make static.
17510 2013-06-10 Jakub Jelinek <jakub@redhat.com>
17513 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
17514 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
17515 Use DATA_ABI_ALIGNMENT for that case instead if defined.
17516 (get_variable_align): New function.
17517 (get_variable_section, emit_bss, emit_common,
17518 assemble_variable_contents, place_block_symbol): Use
17519 get_variable_align instead of DECL_ALIGN.
17520 (assemble_noswitch_variable): Add align argument, use it
17521 instead of DECL_ALIGN.
17522 (assemble_variable): Adjust caller. Use get_variable_align
17523 instead of DECL_ALIGN.
17524 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
17526 (DATA_ABI_ALIGNMENT): Define.
17527 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
17528 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
17529 opt is false, only return the psABI mandated alignment increase.
17530 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
17531 (DATA_ABI_ALIGNMENT): ... this.
17532 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
17533 (DATA_ABI_ALIGNMENT): ... this.
17534 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
17535 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
17536 (DATA_ABI_ALIGNMENT): ... this.
17537 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
17538 * doc/tm.texi: Regenerated.
17540 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
17542 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
17543 cmp_code to construct REG_EQUAL note.
17545 2013-06-09 Jakub Jelinek <jakub@redhat.com>
17548 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
17549 that operands[2] doesn't overlap with operands[0].
17551 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
17552 Jan Hubicka <jh@suse.cz>
17554 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
17555 hack to mark symbols as used.
17557 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
17559 PR rtl-optimization/57559
17560 * lra-constraints.c (process_alt_operands): Don't discourage
17561 memory with known offset for offsetable memory constraint.
17562 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
17564 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
17566 * varasm.c (struct oc_local_state): Reorder fields.
17567 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
17568 and adjust accordingly.
17569 (output_constructor): Reorder initialization code and adjust call to
17570 output_constructor_bitfield.
17572 2013-06-07 Jan Hubicka <jh@suse.cz>
17574 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
17576 2013-06-07 David Malcolm <dmalcolm@redhat.com>
17578 * tree-object-size.c (unknown): Make const.
17580 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17582 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
17583 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
17584 for last alternative in the cpu_facility attribute.
17586 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17589 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
17590 (xordi3): Change operand 2 constraint to arm_xordi_operand.
17591 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
17592 * config/arm/constraints.md (Dg): New constraint.
17593 * config/arm/neon.md (xordi3_neon): Remove.
17594 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
17595 * config/arm/predicates.md (arm_xordi_operand): New predicate.
17597 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17599 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
17600 Clean up alternatives.
17602 2013-06-07 Alan Modra <amodra@gmail.com>
17604 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
17607 2013-06-07 Alan Modra <amodra@gmail.com>
17609 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
17611 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17613 * config/arm/constraints.md (Df): New constraint.
17614 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
17615 Correct length attribute for last two alternatives.
17617 2013-06-07 Alan Modra <amodra@gmail.com>
17619 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
17620 override user -mfp-in-toc.
17621 (offsettable_ok_by_alignment): Consider just the current access
17622 rather than the whole object, unless BLKmode. Handle
17623 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
17624 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
17625 for -mcmodel=medium.
17626 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
17627 override user -mfp-in-toc or -msum-in-toc. Default to
17628 -mno-fp-in-toc for -mcmodel=medium.
17630 2013-06-06 DJ Delorie <dj@redhat.com>
17632 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
17633 TARGET_VALID_POINTER_MODE.
17635 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
17636 Pat Haugen <pthaugen@us.ibm.com>
17637 Peter Bergner <bergner@vnet.ibm.com>
17639 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
17640 Document new power8 builtins.
17642 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
17643 condition code register, to allow 128-bit logical operations to be
17644 done in the VSX or GPR registers.
17645 (nor<mode>3): Use the canonical form for nor.
17646 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
17647 vclz*, and vpopcnt* vector instructions.
17648 (nand<mode>3): Likewise.
17649 (orc<mode>3): Likewise.
17650 (clz<mode>2): LIkewise.
17651 (popcount<mode>2): Likewise.
17653 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
17654 that only the GPRs are recognized.
17656 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
17657 support for new power8 builtins.
17659 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
17661 (xscvdpspn): Likewise.
17667 (vpopcnt): Likewise.
17668 (vpopcntb): Likewise.
17669 (vpopcnth): Likewise.
17670 (vpopcntw): Likewise.
17671 (vpopcntd): Likewise.
17673 (vmrgew): Likewise.
17674 (vmrgow): Likewise.
17676 (eqv_v16qi3): Likewise.
17677 (eqv_v8hi3): Likewise.
17678 (eqv_v4si3): Likewise.
17679 (eqv_v2di3): Likewise.
17680 (eqv_v4sf3): Likewise.
17681 (eqv_v2df3): Likewise.
17683 (nand_v16qi3): Likewise.
17684 (nand_v8hi3): Likewise.
17685 (nand_v4si3): Likewise.
17686 (nand_v2di3): Likewise.
17687 (nand_v4sf3): Likewise.
17688 (nand_v2df3): Likewise.
17690 (orc_v16qi3): Likewise.
17691 (orc_v8hi3): Likewise.
17692 (orc_v4si3): Likewise.
17693 (orc_v2di3): Likewise.
17694 (orc_v4sf3): Likewise.
17695 (orc_v2df3): Likewise.
17697 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
17698 allow power8 quad mode in 64-bit.
17699 (rs6000_builtin_vectorized_function): Add support to vectorize
17700 ISA 2.07 count leading zeros, population count builtins.
17701 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
17702 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
17703 (builtin_function_type): Add vgbbd builtin function which takes an
17705 (altivec_expand_vec_perm_const): Add support for new power8 merge
17708 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
17709 that does not include TImdoe for use with 32-bit.
17710 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
17712 (UNSPEC_VSX_CVDPSPN): Likewise.
17713 (vsx_xscvdpspn): Likewise.
17714 (vsx_xscvspdpn): Likewise.
17715 (vsx_xscvdpspn_scalar): Likewise.
17716 (vsx_xscvspdpn_directmove): Likewise.
17717 (vsx_and<mode>3): Split logical operations into 32-bit and
17718 64-bit. Add support to do logical operations on TImode as well as
17719 VSX vector types. Allow logical operations to be done in either
17720 VSX registers or in general purpose registers in 64-bit mode. Add
17721 splitters if GPRs were used. For AND, add clobber of CCmode to
17722 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
17724 (vsx_and<mode>3_32bit): Likewise.
17725 (vsx_and<mode>3_64bit): Likewise.
17726 (vsx_ior<mode>3): Likewise.
17727 (vsx_ior<mode>3_32bit): Likewise.
17728 (vsx_ior<mode>3_64bit): Likewise.
17729 (vsx_xor<mode>3): Likewise.
17730 (vsx_xor<mode>3_32bit): Likewise.
17731 (vsx_xor<mode>3_64bit): Likewise.
17732 (vsx_one_cmpl<mode>2): Likewise.
17733 (vsx_one_cmpl<mode>2_32bit): Likewise.
17734 (vsx_one_cmpl<mode>2_64bit): Likewise.
17735 (vsx_nor<mode>3): Likewise.
17736 (vsx_nor<mode>3_32bit): Likewise.
17737 (vsx_nor<mode>3_64bit): Likewise.
17738 (vsx_andc<mode>3): Likewise.
17739 (vsx_andc<mode>3_32bit): Likewise.
17740 (vsx_andc<mode>3_64bit): Likewise.
17741 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
17742 and xxlorc instructions.
17743 (vsx_eqv<mode>3_64bit): Likewise.
17744 (vsx_nand<mode>3_32bit): Likewise.
17745 (vsx_nand<mode>3_64bit): Likewise.
17746 (vsx_orc<mode>3_32bit): Likewise.
17747 (vsx_orc<mode>3_64bit): Likewise.
17749 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
17751 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
17753 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
17754 (p8_vmrgow): Likewise.
17755 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
17756 GPRs to be split under VSX.
17757 (p8v_clz<mode>2): Add power8 count leading zero support.
17758 (p8v_popcount<mode>2): Add power8 population count support.
17759 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
17762 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
17765 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
17767 (vec_nand): Likewise.
17768 (vec_vclz): Likewise.
17769 (vec_vclzb): Likewise.
17770 (vec_vclzd): Likewise.
17771 (vec_vclzh): Likewise.
17772 (vec_vclzw): Likewise.
17773 (vec_vgbbd): Likewise.
17774 (vec_vmrgew): Likewise.
17775 (vec_vmrgow): Likewise.
17776 (vec_vpopcnt): Likewise.
17777 (vec_vpopcntb): Likewise.
17778 (vec_vpopcntd): Likewise.
17779 (vec_vpopcnth): Likewise.
17780 (vec_vpopcntw): Likewise.
17782 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17784 PR rtl-optimization/57468
17785 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
17788 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17790 PR rtl-optimization/57459
17791 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
17792 type when setting live regs.
17794 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17796 * config/s390/s390.opt (mlra): New option.
17797 * config/s390/s390.c (s390_decompose_address): Check displacement
17798 for all registers for LRA.
17799 (s390_secondary_reload): Don't used secondary reloads for LRA.
17800 (s390_lra_p): New function.
17801 (TARGET_LRA_P): Define.
17802 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
17803 of attribute cpu_facility to zarch for the last alternative.
17804 (*cmpmem_short): Ditto.
17806 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
17808 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
17809 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
17810 (arm_expand_prologue): Likewise.
17812 2013-06-06 Teresa Johnson <tejohnson@google.com>
17815 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
17816 as this is now done by redirect_edge_and_branch_force.
17817 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
17818 barriers, and fix interaction with splitting.
17819 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
17820 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
17821 reflect changes made in the routine.
17822 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
17823 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
17824 since this is called in cfglayout mode, and replace partition fixup
17825 with assert as that is now done by force_nonfallthru_and_redirect.
17826 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
17827 already be marked with region crossing note.
17828 (insert_section_boundary_note): Make non-static, gate on flag
17829 has_bb_partition, rewrite to also check for multiple partitions.
17830 (rest_of_handle_reorder_blocks): Remove call to
17831 insert_section_boundary_note, now done later during free_cfg.
17832 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
17833 * bb-reorder.h (insert_section_boundary_note): Declare.
17834 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
17835 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
17836 invoke insert_section_boundary_note.
17837 (try_redirect_by_replacing_jump): Remove unnecessary
17838 check for region crossing note.
17839 (fixup_partition_crossing): New function.
17840 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
17841 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
17842 in non-cfglayout mode.
17843 (force_nonfallthru_and_redirect): Fixup partition boundaries,
17844 remove old code that tried to do this. Emit barrier correctly
17845 when we are in cfglayout mode.
17846 (last_bb_in_partition): New function.
17847 (rtl_split_edge): Correctly fixup partition boundaries.
17848 (commit_one_edge_insertion): Remove old code that tried to
17849 fixup region crossing edge since this is now handled in
17850 split_block, and set up insertion point correctly since
17851 block may now end in a jump.
17852 (verify_hot_cold_block_grouping): Guard against checking when not in
17853 linearized RTL mode.
17854 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
17856 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
17857 rtl_verify_flow_info, so not called in cfglayout mode.
17858 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
17859 (fixup_reorder_chain): Remove old code that attempted to fixup region
17860 crossing note as this is now handled in force_nonfallthru_and_redirect.
17861 (duplicate_insn_chain): Don't duplicate switch section notes.
17862 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
17864 * basic-block.h (emit_barrier_after_bb): Declare.
17866 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17868 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
17869 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
17870 arm_usatsihi): Adjust alternatives for arm_restrict_it.
17872 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17874 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
17876 * config/arm/ldmstm.md: Regenerate.
17878 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17880 * config/arm/sync.md (atomic_loaddi_1):
17881 Disable predication for arm_restrict_it.
17882 (arm_load_exclusive<mode>): Likewise.
17883 (arm_load_exclusivesi): Likewise.
17884 (arm_load_exclusivedi): Likewise.
17885 (arm_load_acquire_exclusive<mode>): Likewise.
17886 (arm_load_acquire_exclusivesi): Likewise.
17887 (arm_load_acquire_exclusivedi): Likewise.
17888 (arm_store_exclusive<mode>): Likewise.
17889 (arm_store_exclusive<mode>): Likewise.
17890 (arm_store_release_exclusivedi): Likewise.
17891 (arm_store_release_exclusive<mode>): Likewise.
17893 2013-06-06 Richard Biener <rguenther@suse.de>
17895 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
17897 (lto_tag_is_tree_code_p): Adjust.
17898 (lto_tag_is_gimple_code_p): Likewise.
17899 (lto_gimple_code_to_tag): Likewise.
17900 (lto_tag_to_gimple_code): Likewise.
17901 (lto_tree_code_to_tag): Likewise.
17902 (lto_tag_to_tree_code): Likewise.
17903 * data-streamer.h (streamer_write_hwi_in_range): Use
17904 uhwi streaming to stream the normalized range.
17905 (streamer_read_hwi_in_range): Likewise.
17907 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17909 * config/arm/arm.md (enabled_for_depr_it): New attribute.
17910 (predicable_short_it): Likewise.
17911 (predicated): Likewise.
17912 (enabled): Handle above.
17913 (define_cond_exec): Set predicated attribute to yes.
17915 2013-06-05 Mike Stump <mikestump@comcast.net>
17917 * gdbinit.in (__FUNCTION__): Add.
17919 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
17921 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
17922 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
17924 2013-06-05 Jan Hubicka <jh@suse.cz>
17926 * varasm.c (mark_decl_referenced): Revert the removal until targets
17929 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
17931 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
17932 instead of mark_decl_referenced.
17934 2013-06-05 Jan Hubicka <jh@suse.cz>
17936 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
17937 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
17938 and symtab_used_from_object_file_p.
17939 (cgraph_make_node_local_1): Clear forced_by_abi.
17940 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
17941 * cgraph.h (symtab_node_base): Add forced_by_abi.
17942 (decide_is_variable_needed): Remove.
17943 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
17944 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
17945 (decide_is_symbol_needed): ... this one; handle symbols in general;
17946 always analyze virtuals; honnor forced_by_abi.
17947 (cgraph_finalize_function): Update.
17948 (varpool_finalize_decl): Update.
17949 (symbol_defined_and_needed): Remove.
17950 (analyze_functions): Update.
17951 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17952 output_refs, input_overwrite_node): Handle forced_by_abi.
17953 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
17954 (address_taken_from_non_vtable_p): ... this one.
17955 (comdat_can_be_unshared_p_1): New function.
17956 (cgraph_comdat_can_be_unshared_p): Rename to ...
17957 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
17958 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
17959 (function_and_variable_visibility): Clear forced_by_abi as needed.
17960 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
17961 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
17962 * symtab.c (dump_symtab_base): Dump forced_by_abi.
17963 * varpool.c (decide_is_variable_needed): Remove.
17965 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17967 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
17968 (arm_option_override): Override arm_restrict_it where appropriate.
17969 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
17970 * config/arm/arm.opt (mrestrict-it): New command-line option.
17971 * doc/invoke.texi: Document -mrestrict-it.
17973 2013-06-05 David Malcolm <dmalcolm@redhat.com>
17975 * tsan.c (tsan_atomic_table): Make const.
17977 2013-06-05 Richard Biener <rguenther@suse.de>
17979 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
17980 index associated with the tree we are supposed to replace.
17981 * tree-streamer-out.c (pack_ts_base_value_fields): Output
17982 TREE_ASM_WRITTEN as zero for everything but SSA names.
17984 2013-06-05 David Malcolm <dmalcolm@redhat.com>
17986 * tree-ssa-structalias.c (call_stmt_vars): Make static.
17988 2013-06-04 Jan Hubicka <jh@suse.cz>
17990 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
17991 (input_node, input_varpool_node): Handle correctly external same
17993 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
17994 nodes at ltrans stage.
17996 2013-06-04 Jan Hubicka <jh@suse.cz>
17998 * ipa-inline.c (update_caller_keys): Fix availability test.
17999 (update_callee_keys): Likewise.
18000 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
18001 to follow ELF standard.
18003 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
18005 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
18006 (mips64r5900el-*-elf*): New configurations.
18007 * config/mips/mips-cpus.def (r5900): New processor.
18008 * config/mips/mips-tables.opt: Regenerate.
18009 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
18010 (mips_issue_rate): Handle PROCESSOR_R5900.
18011 (mips_reorg_process_insns): Force reorder mode for the R5900.
18012 * config/mips/mips.h (TARGET_MIPS5900): Define.
18013 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
18015 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
18017 * config/mips/mips.md (processor): Add r5900.
18018 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
18020 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18022 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
18023 into function to generate MOVI instruction.
18024 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
18025 (aarch64_preferred_simd_mode): Turn into wrapper.
18026 (aarch64_output_scalar_simd_mov_immediate): New function.
18027 * config/aarch64/aarch64-protos.h: Add prototype for above.
18029 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18031 * config/aarch64/aarch64.c (simd_immediate_info): Remove
18032 element_char member.
18033 (sizetochar): Return signed char.
18034 (aarch64_simd_valid_immediate): Remove elchar and other
18035 unnecessary variables.
18036 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
18037 Calculate element_char as required.
18038 * config/aarch64/aarch64-protos.h: Update and move prototype
18039 for aarch64_output_simd_mov_immediate.
18040 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
18043 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18045 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
18046 information completed by aarch64_simd_valid_immediate.
18047 (aarch64_legitimate_constant_p): Update arguments.
18048 (aarch64_simd_valid_immediate): Work with struct rather than many
18050 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
18051 (aarch64_simd_make_constant): Update arguments.
18052 (aarch64_output_simd_mov_immediate): Work with struct rather than
18053 many pointers. Output immediate directly rather than as operand.
18054 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18056 * config/aarch64/constraints.md (Dn): Update arguments.
18058 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18060 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
18062 (aarch64_simd_immediate_valid_for_move): Remove.
18063 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
18064 (aarch64_simd_make_constant): Update call.
18065 (aarch64_output_simd_mov_immediate): Update call.
18066 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18068 * config/aarch64/constraints.md (Dn): Update call.
18070 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18072 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
18073 return type to bool for prototype.
18074 (aarch64_legitimate_constant_p): Check for true instead of not -1.
18075 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
18076 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
18078 2013-06-04 Catherine Moore <clm@codesourcery.com>
18080 * config/mips/mips.opt (meva): New.
18081 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
18082 (ASM_SPEC): Handle -meva.
18083 * doc/invoke.texi (meva): Document.
18085 2013-06-04 Alan Modra <amodra@gmail.com>
18087 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
18090 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18092 * rtl.def: Add extra fourth optional field to define_cond_exec.
18093 * gensupport.c (process_one_cond_exec): Process attributes from
18095 * doc/md.texi: Document fourth field in define_cond_exec.
18097 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
18099 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
18100 out the processing order as in store_bit_field_1.
18102 2013-06-04 Jan Hubicka <jh@suse.cz>
18104 PR middle-end/57500
18105 * cgraphunit.c (cgraph_process_same_body_aliases): Create
18106 non-VAR_DECL node if it does not exist yet.
18108 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
18110 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
18111 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
18112 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
18113 target_cpu_default setting.
18115 2013-06-03 Teresa Johnson <tejohnson@google.com>
18117 * dumpfile.c (opt_info_switch_p): Change -fopt-info
18118 default to -fopt-info=optimized instead of all.
18119 * doc/invoke.texi: Ditto.
18120 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
18121 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
18122 (execute_vect_slp): Emit BB vectorization success under
18123 MSG_OPTIMIZED_LOCATIONS.
18124 * tree-vect-slp.c (vect_slp_transform_bb): Change
18125 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
18126 * tree-vect-loop.c (vect_transform_loop): Ditto.
18128 2013-06-03 Jason Merrill <jason@redhat.com>
18131 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
18132 Use TARGET_EXPR for C++.
18134 2013-06-03 Jakub Jelinek <jakub@redhat.com>
18136 PR rtl-optimization/57268
18137 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
18138 if DEBUG_INSN_P (insn).
18141 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18143 PR rtl-optimization/57268
18144 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18145 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18147 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18149 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
18150 (ix86_avoid_lea_for_addr): Likewise.
18151 (exact_dependency_1): Likewise.
18152 (ix86_adjust_cost): Likewise.
18153 (swap_top_of_ready_list): Fix formatting and !reload_completed check
18155 (do_reorder_for_imul): Fix typo, formatting and
18156 !reload_completed check removed.
18157 (ix86_sched_reorder): Fix typo and formatting.
18158 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
18161 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
18163 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
18165 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
18167 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
18168 <STRING_CST>: Likewise.
18169 <VECTOR_CST>: Likewise.
18171 2013-06-01 Janus Weil <janus@gcc.gnu.org>
18172 Mikael Morin <mikael@gcc.gnu.org>
18174 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
18175 * config.in: Regenerated.
18176 * configure: Regenerated.
18178 2013-06-01 Jan Hubicka <jh@suse.cz>
18180 PR middle-end/57366
18181 * cgraphunit.c (compile): When weakref is not supported,
18182 set up transparent aliases before final output pass.
18183 * varasm.c (assemble_alias): Do not try to do it here.
18185 2013-06-01 Jan Hubicka <jh@suse.cz>
18187 PR middle-end/57467
18188 * passes.c (for_per_function): Skip unanalyzed functions.
18190 2013-06-01 Jan Hubicka <jh@suse.cz>
18192 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
18193 (lto_symtab_merge_symbols_1): ... this one.
18194 (lto_symtab_merge_cgraph_nodes): Rename to ...
18195 (lto_symtab_merge_symbols): ... this one; simplify.
18196 * cgraph.c (same_body_aliases_done): Rename to ...
18197 (cpp_implicit_aliases_done): ... this one.
18198 (cgraph_create_function_alias): Update.
18199 (cgraph_same_body_alias): Update.
18200 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
18201 (verify_edge_corresponds_to_fndecl): Simplify.
18202 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
18203 (cgraph_node): Remove same_body_alias.
18204 (varpool_node): Remove alias_of and extra_name_alias.
18205 (same_body_aliases_done): Rename to ..
18206 (cpp_implicit_aliases_done): ... this one.
18207 (symtab_alias_ultimate_target): Add default parameter.
18208 (symtab_resolve_alias): New function.
18209 (fixup_same_cpp_alias_visibility): Declare.
18210 (cgraph_function_node): Add default parameter.
18211 (cgraph_node_asm_name): Likewise.
18212 (cgraph_function_or_thunk_node): Add default parameter; do
18213 not ICE when it is NULL.
18214 (varpool_variable_node): Likewise.
18215 * tree-emutls.c (create_emultls_var): Update.
18216 (ipa_lower_emutls): Update.
18217 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
18218 (cgraph_reset_node): Reset alias info.
18219 (cgraph_finalize_function): Update.
18220 (fixup_same_cpp_alias_visibility): Move to symtab.c.
18221 (analyze_function): Simplify.
18222 (cgraph_process_same_body_aliases): Simplify.
18223 (analyze_functions): Fixup same body aliases.
18224 (handle_alias_pairs): Simplify.
18225 (assemble_thunk): Update.
18226 (assemble_thunks_and_aliases): Update.
18227 (output_weakrefs): Rewrite.
18228 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
18229 (lto_output_varpool_node): Likewise.
18230 (compute_ltrans_boundary): Remve assert.
18231 (get_alias_symbol): New functoin.
18232 (input_node): Rewrite alias handling.
18233 (input_varpool_node): Likewise.
18234 * ipa-pure-const.c (propagate_pure_const): Fix formating.
18235 * ipa.c (process_references): Handle weakrefs correctly.
18236 (symtab_remove_unreachable_nodes): Likewise.
18237 * trans-mem.c (get_cg_data): Update.
18238 (ipa_tm_create_version_alias): Update.
18239 (ipa_tm_execute): Update.
18240 * symtab.c (dump_symtab_base): Dump aliases.
18241 (verify_symtab_base): Verify aliases.
18242 (symtab_node_availability): New function.
18243 (symtab_alias_ultimate_target): Simplify.
18244 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
18245 handle all the fixup cases.
18246 (symtab_resolve_alias): New function.
18247 * passes.c (ipa_write_summaries): Handle weakrefs.
18248 * varpool.c (varpool_analyze_node): Simplify.
18249 (assemble_aliases): Update.
18250 (varpool_create_variable_alias): Simplify.
18251 (varpool_extra_name_alias): Simplify.
18252 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
18253 (lto_symtab_merge_symbols): ... this one.
18255 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
18258 PR rtl-optimization/57268
18259 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18260 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18262 2013-06-01 Tobias Burnus <burnus@net-b.de>
18264 Partially reverted:
18265 2013-05-31 Tobias Burnus <burnus@net-b.de>
18267 PR middle-end/57073
18268 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
18271 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18273 PR rtl-optimization/57268
18274 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18275 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18277 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18279 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
18280 unordered comparison operators when -fno-trapping-math is in effect
18282 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
18283 and implement unordered comparison operators properly on the e500.
18285 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18287 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
18288 for constant scalar integers.
18289 (simplify_relational_operation_1): Likewise.
18291 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
18293 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
18294 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
18297 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18298 Igor Zamyatin <igor.zamyatin@intel.com>
18300 Silvermont (SLM) architecture performance tuning.
18301 * config/i386/i386.h (enum ix86_tune_indices): Add
18302 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
18303 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
18305 * config/i386/i386.c (initial_ix86_tune_features)
18306 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
18307 (ix86_lea_outperforms): Handle Silvermont tuning.
18308 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
18310 (ix86_use_lea_for_mov): Likewise.
18311 (ix86_avoid_lea_for_addr): Likewise.
18312 (ix86_lea_for_add_ok): Likewise.
18313 (exact_dependency_1): New function.
18314 (exact_store_load_dependency): Likewise.
18315 (ix86_adjust_cost): Handle Silvermont tuning.
18316 (do_reoder_for_imul): Likewise.
18317 (swap_top_of_ready_list): New function.
18318 (ix86_sched_reorder): Changed to handle Silvermont tuning.
18320 * config/i386/i386.md (peepholes that split memory operand in fp
18323 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18325 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
18326 Remove un-necessary braces.
18328 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18330 * config/aarch64/aarch64.c (aarch64_classify_symbol):
18331 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
18333 2013-05-31 Tobias Burnus <burnus@net-b.de>
18335 PR middle-end/57073
18336 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
18338 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18341 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
18342 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
18343 * config/arm/neon.md (iordi3_neon): Remove.
18344 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
18345 * config/arm/predicates.md (imm_for_neon_logic_operand):
18346 Move to earlier in the file.
18347 (neon_logic_op2): Likewise.
18348 (arm_iordi_operand_neon): New predicate.
18350 2013-05-31 Richard Biener <rguenther@suse.de>
18352 PR tree-optimization/57478
18353 PR tree-optimization/57453
18354 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
18357 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18359 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
18360 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
18362 2013-05-30 Tobias Burnus <burnus@net-b.de>
18363 Thomas Koenig <tkoenig@gcc.gnu.org>
18365 PR middle-end/57073
18366 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
18367 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
18369 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
18371 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
18373 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
18375 * target.def (register_usage_leveling_p): New hook.
18376 * targhooks.c (default_register_usage_leveling_p): New.
18377 * targhooks.h (default_register_usage_leveling_p): New prototype.
18378 * lra-assigns.c (register_usage_leveling_p): Use the hook.
18379 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
18380 * doc/tm.texi: Update.
18381 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
18383 2013-05-30 Ian Bolton <ian.bolton@arm.com>
18385 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
18386 (*insv_reg<mode>): New define_insn.
18388 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
18390 PR rtl-optimization/57439
18391 * postreload.c (move2add_valid_value_p): Check that we have
18392 a zero subreg_regno_offset when accessing the register in
18393 the requested mode.
18395 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18396 Igor Zamyatin <igor.zamyatin@intel.com>
18398 Silvermont (SLM) architecture pipeline model, tuning and
18400 * config.gcc: Add slm config options and target.
18402 * config/i386/slm.md: New.
18404 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
18406 * config/i386/i386-c.c (ix86_target_macros_internal): New case
18408 (ix86_target_macros_internal): Likewise.
18410 * config/i386/i386.c (slm_cost): New cost.
18411 (m_SLM): New macro flag.
18412 (initial_ix86_tune_features): Set m_SLM.
18413 (x86_accumulate_outgoing_args): Likewise.
18414 (x86_arch_always_fancy_math_387): Likewise.
18415 (processor_target_table): Add slm cost.
18416 (cpu_names): Add slm cpu name.
18417 (x86_option_override_internal): Set SLM ISA.
18418 (ix86_issue_rate): New case PROCESSOR_SLM.
18419 (ia32_multipass_dfa_lookahead): Likewise.
18420 (fold_builtin_cpu): Add slm.
18422 * config/i386/i386.h (TARGET_SLM): New target macro.
18423 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
18424 (processor_type): Add PROCESSOR_SLM.
18426 * config/i386/i386.md (cpu): Add new value "slm".
18427 (slm.md): Include slm.md.
18429 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
18430 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18432 * config/arm/arm-protos.h: Add and update function protos.
18433 * config/arm/arm.c (use_simple_return_p): New added.
18434 (thumb2_expand_return): Check simple_return flag.
18435 * config/arm/arm.md: Add simple_return and conditional simple_return.
18436 * config/arm/iterators.md: Add iterator for return and simple_return.
18438 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18440 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
18441 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
18442 (arm_emit_vfp_multi_reg_pop): Likewise.
18443 (thumb2_emit_ldrd_pop): Likewise.
18444 (arm_expand_epilogue): Add misc REG_CFA notes.
18445 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
18447 2013-05-29 Lawrence Crowl <crowl@google.com>
18449 * config/arm/t-arm: Update for below.
18451 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
18452 Change type to hash_table. Update dependent calls and types.
18454 * config/i386/t-cygming: Update for below.
18456 * config/i386/t-interix: Update for below.
18458 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
18459 Change type to hash_table. Update dependent calls and types.
18460 (i386_find_on_wrapper_list::wrappers): Likewise.
18462 * config/ia64/t-ia64: Update for below.
18464 * config/ia64/ia64.c (bundle_state_table):
18465 Change type to hash_table. Update dependent calls and types.
18467 * config/mips/mips.c (mips_reorg_process_insns::htab):
18468 Change type to hash_table. Update dependent calls and types.
18470 * config/sol2.c (solaris_comdat_htab):
18471 Change type to hash_table. Update dependent calls and types.
18473 * config/t-sol2: Update for above.
18475 2013-05-29 Teresa Johnson <tejohnson@google.com>
18477 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
18478 functions are not yet marked as defined.
18480 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
18481 Pat Haugen <pthaugen@us.ibm.com>
18482 Peter Bergner <bergner@vnet.ibm.com>
18484 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
18488 (vrotl<mode>3): Likewise.
18489 (vashl<mode>3): Likewise.
18490 (vlshr<mode>3): Likewise.
18491 (vashr<mode>3): Likewise.
18493 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18494 support for power8 V2DI builtins.
18496 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
18497 power8 V2DI builtins.
18498 (vupkhsw): Likewise.
18499 (vupklsw): Likewise.
18500 (vaddudm): Likewise.
18501 (vminsd): Likewise.
18502 (vmaxsd): Likewise.
18503 (vminud): Likewise.
18504 (vmaxud): Likewise.
18505 (vpkudum): Likewise.
18506 (vpksdss): Likewise.
18507 (vpkudus): Likewise.
18508 (vpksdus): Likewise.
18513 (vsubudm): Likewise.
18514 (vcmpequd): Likewise.
18515 (vcmpgtsd): Likewise.
18516 (vcmpgtud): Likewise.
18517 (vcmpequd_p): Likewise.
18518 (vcmpgtsd_p): Likewise.
18519 (vcmpgtud_p): Likewise.
18520 (vupkhsw): Likewise.
18521 (vupklsw): Likewise.
18522 (vaddudm): Likewise.
18523 (vmaxsd): Likewise.
18524 (vmaxud): Likewise.
18525 (vminsd): Likewise.
18526 (vminud): Likewise.
18527 (vpksdss): Likewise.
18528 (vpksdus): Likewise.
18529 (vpkudum): Likewise.
18530 (vpkudus): Likewise.
18535 (vsubudm): Likewise.
18537 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
18538 support for power8 V2DI instructions.
18540 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
18541 power8 V2DI instructions. Combine pack and unpack insns to use an
18542 iterator for each mode. Check whether a particular mode supports
18543 Altivec instructions instead of just checking TARGET_ALTIVEC.
18544 (UNSPEC_VPKUWUM): Likewise.
18545 (UNSPEC_VPKSHSS): Likewise.
18546 (UNSPEC_VPKSWSS): Likewise.
18547 (UNSPEC_VPKUHUS): Likewise.
18548 (UNSPEC_VPKSHUS): Likewise.
18549 (UNSPEC_VPKUWUS): Likewise.
18550 (UNSPEC_VPKSWUS): Likewise.
18551 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
18552 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
18553 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
18554 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
18555 (UNSPEC_VUPKHSB): Likewise.
18556 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
18557 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
18558 (UNSPEC_VUPKHSH): Likewise.
18559 (UNSPEC_VUPKLSB): Likewise.
18560 (UNSPEC_VUPKLSH): Likewise.
18562 (VI_char): Likewise.
18563 (VI_scalar): Likewise.
18564 (VI_unit): Likewise.
18566 (VP_small): Likewise.
18567 (VP_small_lc): Likewise.
18568 (VU_char): Likewise.
18569 (add<mode>3): Likewise.
18570 (altivec_vaddcuw): Likewise.
18571 (altivec_vaddu<VI_char>s): Likewise.
18572 (altivec_vadds<VI_char>s): Likewise.
18573 (sub<mode>3): Likewise.
18574 (altivec_vsubcuw): Likewise.
18575 (altivec_vsubu<VI_char>s): Likewise.
18576 (altivec_vsubs<VI_char>s): Likewise.
18577 (altivec_vavgs<VI_char>): Likewise.
18578 (altivec_vcmpbfp): Likewise.
18579 (altivec_eq<mode>): Likewise.
18580 (altivec_gt<mode>): Likewise.
18581 (altivec_gtu<mode>): Likewise.
18582 (umax<mode>3): Likewise.
18583 (smax<mode>3): Likewise.
18584 (umin<mode>3): Likewise.
18585 (smin<mode>3): Likewise.
18586 (altivec_vpkuhum): Likewise.
18587 (altivec_vpkuwum): Likewise.
18588 (altivec_vpkshss): Likewise.
18589 (altivec_vpkswss): Likewise.
18590 (altivec_vpkuhus): Likewise.
18591 (altivec_vpkshus): Likewise.
18592 (altivec_vpkuwus): Likewise.
18593 (altivec_vpkswus): Likewise.
18594 (altivec_vpks<VI_char>ss): Likewise.
18595 (altivec_vpks<VI_char>us): Likewise.
18596 (altivec_vpku<VI_char>us): Likewise.
18597 (altivec_vpku<VI_char>um): Likewise.
18598 (altivec_vrl<VI_char>): Likewise.
18599 (altivec_vsl<VI_char>): Likewise.
18600 (altivec_vsr<VI_char>): Likewise.
18601 (altivec_vsra<VI_char>): Likewise.
18602 (altivec_vsldoi_<mode>): Likewise.
18603 (altivec_vupkhsb): Likewise.
18604 (altivec_vupkhs<VU_char>): Likewise.
18605 (altivec_vupkls<VU_char>): Likewise.
18606 (altivec_vupkhsh): Likewise.
18607 (altivec_vupklsb): Likewise.
18608 (altivec_vupklsh): Likewise.
18609 (altivec_vcmpequ<VI_char>_p): Likewise.
18610 (altivec_vcmpgts<VI_char>_p): Likewise.
18611 (altivec_vcmpgtu<VI_char>_p): Likewise.
18612 (abs<mode>2): Likewise.
18613 (vec_unpacks_hi_v16qi): Likewise.
18614 (vec_unpacks_hi_v8hi): Likewise.
18615 (vec_unpacks_lo_v16qi): Likewise.
18616 (vec_unpacks_hi_<VP_small_lc>): Likewise.
18617 (vec_unpacks_lo_v8hi): Likewise.
18618 (vec_unpacks_lo_<VP_small_lc>): Likewise.
18619 (vec_pack_trunc_v8h): Likewise.
18620 (vec_pack_trunc_v4si): Likewise.
18621 (vec_pack_trunc_<mode>): Likewise.
18623 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
18625 (vec_vmaxsd): Likewise.
18626 (vec_vmaxud): Likewise.
18627 (vec_vminsd): Likewise.
18628 (vec_vminud): Likewise.
18629 (vec_vpksdss): Likewise.
18630 (vec_vpksdus): Likewise.
18631 (vec_vpkudum): Likewise.
18632 (vec_vpkudus): Likewise.
18633 (vec_vrld): Likewise.
18634 (vec_vsld): Likewise.
18635 (vec_vsrad): Likewise.
18636 (vec_vsrd): Likewise.
18637 (vec_vsubudm): Likewise.
18638 (vec_vupkhsw): Likewise.
18639 (vec_vupklsw): Likewise.
18641 2013-05-29 Jan Hubicka <jh@suse.cz>
18643 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
18644 flags; reorder rest of fields in more consistent way.
18645 (varpool_node): Remove analyzed, finalized and alias.
18646 (cgraph_ndoe): Likewise.
18647 (symtab_alias_ultimate_target): New function.
18648 (cgraph_function_node): Move offline.
18649 (cgraph_reset_node): Declare.
18650 (cgraph_comdat_can_be_unshared_p): Remove.
18651 (varpool_remove_initializer): Declare.
18652 (varpool_first_defined_variable, varpool_next_defined_variable
18653 cgraph_first_defined_function, cgraph_next_defined_function): Update.
18654 (cgraph_function_with_gimple_body_p): Update.
18655 (varpool_all_refs_explicit_p): Update.
18656 (symtab_alias_target): New function.
18657 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
18658 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
18659 (cgraph_function_or_thunk_node): Simplify using
18660 symtab_alias_ultimate_target.
18661 (varpool_variable_node): Likewise.
18662 * cgraph.c (cgraph_create_function_alias): Update.
18663 (cgraph_add_thunk): Update.
18664 (cgraph_remove_node): Update.
18665 (dump_cgraph_node): Do not dump removed flags.
18666 (cgraph_function_body_availability): Update.
18667 (cgraph_propagate_frequency): Update.
18668 (verify_cgraph_node): Check sanity of local flag.
18669 (cgraph_function_node): Move here from cgraph.h; revamp for
18670 cgraph_function_or_thunk_node.
18671 * lto-symtab.c (lto_varpool_replace_node): Update.
18672 (lto_symtab_resolve_can_prevail_p): Update.
18673 (lto_symtab_merge_cgraph_nodes): Update.
18674 * ipa-cp.c (determine_versionability, initialize_node_lattices,
18675 propagate_constants_accross_call, devirtualization_time_bonus,
18676 ipcp_propagate_stage): Update.
18677 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
18678 * ipa-inline-transform.c (clone_inlined_nodes,
18679 preserve_function_body_p): Update.
18680 * ipa-reference.c (propagate): Update.
18681 (write_node_summary_p): Update.
18682 * toplev.c (wrapup_global_declaration_2): Update.
18683 * cgraphunit.c (cgraph_analyze_function): Rename to ...
18684 (analyze_function) ... this one.
18685 (cgraph_process_new_functions): Update.
18686 (cgraph_reset_node): Export.
18687 (cgraph_finalize_function): Update.
18688 (cgraph_add_new_function): Update.
18689 (process_function_and_variable_attributes): Update.
18690 (varpool_finalize_decl): Update.
18691 (symbol_finalized): Remove.
18692 (symbol_finalized_and_needed): Rename to ...
18693 (symbol_defined_and_needed): ... update.
18694 (cgraph_analyze_functions): Update.
18695 (handle_alias_pairs): Update.
18696 (mark_functions_to_output): Update.
18697 (assemble_thunk): Update.
18698 (output_in_order): Update.
18699 (output_weakrefs): Update.
18700 (finalize_compilation_unit): Update.
18701 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
18702 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
18703 input_node, input_varpool_node): Update.
18704 * dbxout.c (dbxout_expand_expr): Update.
18705 * cgraphclones.c (cgraph_clone_node): Update.
18706 (cgraph_copy_node_for_versioning): Update.
18707 (cgraph_materialize_clone): Update.
18708 (cgraph_materialize_all_clones): Update.
18709 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
18710 propagate_pure_const, propagate_nothrow): Update.
18711 * lto-streamer-out.c (lto_output, write_symbol): Update.
18712 * ipa-utils.c (ipa_reverse_postorder): Update.
18713 * ipa-inline.c (can_inline_edge_p): Update.
18714 (update_caller_keys, ipa_inline): Update.
18715 * dwarf2out.c (reference_to_unused,
18716 premark_types_used_by_global_vars_helper): Update.
18717 * tree-eh.c (tree_could_trap_p): Update.
18718 * ipa-split.c (consider_split, execute_split_functions): Update.
18719 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
18720 has_addr_references_p): Update; move ahead in file for better
18722 (process_references): Simplify.
18723 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
18724 bodies are removed.
18725 (cgraph_comdat_can_be_unshared_p): Make static.
18726 (cgraph_externally_visible_p): Update.
18727 (varpool_externally_visible_p): Update.
18728 (function_and_variable_visibility): Update.
18729 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
18730 ipa_tm_mark_force_output_node): Update.
18731 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
18732 estimate_edge_devirt_benefit, inline_generate_summary,
18733 inline_write_summary): Update.
18734 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
18735 * ipa-prop.c (ipa_compute_jump_functions): Update.
18736 (ipa_print_node_params, ipa_prop_read_section,
18737 ipa_update_after_lto_read, read_replacements_section): Update.
18738 * varasm.c (mark_decl_referenced): Update.
18739 (assemble_alias, dump_tm_clone_pairs): Update.
18740 * tree-inline.c (copy_bb): Update.
18741 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
18743 * symtab.c (dump_symtab_base): Print new flags.
18744 (verify_symtab_base): Verify new flags.
18745 (symtab_alias_ultimate_target): New function.
18746 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
18747 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
18749 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
18751 * i386.c (ix86_get_function_versions_dispatcher,
18752 ix86_generate_version_dispatcher_body): Update.
18753 (fold_builtin_cpu): Use varpool_add_new_variable.
18754 * varpool.c (varpool_remove_initializer): Break out from ...
18755 (varpool_remove_node): ... this one.
18756 (dump_varpool_node, varpool_node_for_asm,
18757 cgraph_variable_initializer_availability, varpool_analyze_node,
18758 varpool_assemble_decl, varpool_remove_unreferenced_decls,
18759 varpool_finalize_named_section_flags, varpool_create_variable_alias):
18762 2013-05-29 Jan Hubicka <jh@suse.cz>
18764 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
18766 2013-05-29 Easwaran Raman <eraman@google.com>
18768 PR tree-optimization/57442
18769 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
18770 when control exits the main loop.
18772 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
18774 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
18776 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
18777 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
18778 * rx/t-rx: Add rx100 under multi library matches option for nofpu
18781 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18783 PR tree-optimization/57441
18784 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
18785 Don't limit size of incr_vec to number of candidates.
18787 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
18789 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
18790 and mips16 directories.
18791 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
18792 (MULTILIB_DIRNAMES): Ditto.
18793 (MULTILIB_EXCEPTIONS): Add new exceptions.
18794 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
18795 (MULTILIB_DIRNAMES): Ditto.
18796 (MULTILIB_EXCEPTIONS): Add new exceptions.
18798 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18799 Marcus Shawcroft <marcus.shawcroft@arm.com>
18801 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
18802 SYMBOL_TINY_ABSOLUTE.
18803 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
18804 SYMBOL_TINY_ABSOLUTE.
18805 (aarch64_expand_mov_immediate): Likewise.
18806 (aarch64_classify_symbol): Likewise.
18807 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
18808 Permit SYMBOL_TINY_ABSOLUTE.
18809 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
18811 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18812 Marcus Shawcroft <marcus.shawcroft@arm.com>
18814 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
18815 Refactor if/switch. Replace gcc_assert with if.
18817 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
18819 * config/i386/i386.c (initial_ix86_tune_features): Enable
18820 FP Reassociation for AMD bdver1 and bdver2.
18822 2013-05-29 Martin Jambor <mjambor@suse.cz>
18824 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
18825 and IMAGPART_EXPR do not occur within other handled_components.
18827 2013-05-29 Richard Biener <rguenther@suse.de>
18829 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
18830 access on whether the use is in the BB we currently try to
18832 (vect_bb_vectorization_profitable_p): Pass the BB we currently
18833 vectorize to vect_bb_slp_scalar_cost.
18835 2013-05-29 Richard Biener <rguenther@suse.de>
18837 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
18838 computing scalar cost offsetted by stmts that are kept live
18840 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
18841 for computation of scalar cost.
18843 2013-05-28 Steve Ellcey <sellcey@mips.com>
18845 * config/mips/mips-cpus.def (mips32r2): Change processor type.
18847 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
18849 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
18850 array notation built-in reduction functions.
18851 * doc/passes.texi (Passes): Added documentation about changes done
18853 * doc/invoke.texi (C Dialect Options): Added documentation about
18854 the -fcilkplus flag.
18855 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
18856 (BUILTINS_DEF): Depend on cilkplus.def.
18857 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
18858 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
18859 * cilkplus.def: New file.
18861 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
18863 PR rtl-optimization/57439
18864 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
18866 2013-05-28 Easwaran Raman <eraman@google.com>
18868 PR tree-optimization/57337
18869 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
18870 (find_insert_point): Correctly identify the insertion point
18871 when two statements with the same UID is compared.
18873 2013-05-28 Richard Biener <rguenther@suse.de>
18875 PR tree-optimization/56787
18876 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
18877 from the list of data references.
18878 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
18880 (vect_analyze_loop_operations): Likewise.
18881 (vect_transform_loop): Remove clobbers.
18883 2013-05-28 Martin Jambor <mjambor@suse.cz>
18885 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
18886 and REALPART_EXPRs have scalar type.
18888 2013-05-28 Richard Biener <rguenther@suse.de>
18890 PR tree-optimization/57411
18891 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
18893 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
18894 virtual operand propagation.
18896 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18898 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
18899 destination register for bmasksi_vis.
18900 (vector_init_bshuffle): Likewise.
18901 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
18903 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18905 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
18906 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
18907 mode if the instruction isn't available in the original mode.
18908 * config/sparc/sparc.opt (mfix-ut699): New option.
18909 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
18910 (divdf3): Turn into expander.
18911 (divdf3_nofix): New insn.
18912 (divdf3_fix): Likewise.
18913 (divsf3): Disable if -mfix-ut699.
18914 (sqrtdf2): Turn into expander.
18915 (sqrtdf2_nofix): New insn.
18916 (sqrtdf2_fix): Likewise.
18917 (sqrtsf2): Disable if -mfix-ut699.
18919 2013-05-27 Richard Biener <rguenther@suse.de>
18921 PR middle-end/57412
18922 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
18923 block for the new loop.
18925 2013-05-27 Richard Biener <rguenther@suse.de>
18927 PR tree-optimization/57343
18928 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
18929 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
18930 (number_of_iterations_cond): Do not build the folded tree.
18932 2013-05-27 Richard Biener <rguenther@suse.de>
18935 PR middle-end/57381
18936 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
18937 OEP_CONSTANT_ADDRESS_OF retained.
18939 PR tree-optimization/57417
18940 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
18941 for unchanged base.
18942 (set_ssa_val_to): Compare addresses using
18943 get_addr_base_and_unit_offset.
18945 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
18947 PR rtl-optimization/56833
18948 * postreload.c (move2add_record_mode): New function.
18949 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
18950 (move2add_use_add2_insn): Use move2add_record_sym_value.
18951 (move2add_use_add3_insn): Likewise.
18952 (reload_cse_move2add): Use move2add_valid_value_p and
18953 move2add_record_mode. Invalidate call-clobbered and REG_INC
18954 affected regs by setting reg_mode to VOIDmode.
18955 (move2add_note_store): Don't pretend the inside of a SUBREG is
18956 the actual destination. Invalidate single/leading registers by
18957 setting reg_mode to VOIDmode.
18958 Use move2add_record_sym_value, move2add_valid_value_p and
18959 move2add_record_mode.
18961 2013-05-27 Richard Biener <rguenther@suse.de>
18963 PR tree-optimization/57396
18964 * tree-affine.c (double_int_constant_multiple_p): Properly
18965 return false for val == 0 and div != 0.
18967 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18969 * config/mips/mips.h: Use #elif in preprocessor conditions.
18971 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18974 * config/mips/constraints.md (kl): New constraint.
18975 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
18976 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
18977 constraint for operand 0. Split after CSE for MIPS16. Emit a move
18978 from LO for MIPS16.
18979 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
18981 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18984 * config/mips/mips.c (mips_can_inline_p): New function.
18985 (TARGET_CAN_INLINE_P): Define.
18987 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
18989 * sched-int.h (ds_t, dw_t): Make unsigned int.
18990 Fix documentation that describes how all the ds_t bits are used.
18991 Reserve the last bit for delayed-branch scheduling.
18992 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
18993 (BITS_PER_DEP_WEAK): Fix definition and documentation.
18994 (gen_dep_weak_1): Remove prototype.
18995 * sched-deps.c (get_dep_weak_1): Make static.
18996 * target.def (speculate_insn, needs_block_p, gen_spec_check,
18997 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
18998 * doc/tm.texi: Regenerate.
18999 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
19001 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
19004 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
19006 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
19007 Sandra Loosemore <sandra@codesourcery.com>
19009 * config.gcc (powerpc-*): Allow native for with-cpu.
19011 2013-05-24 Jeff Law <law@redhat.com>
19013 PR tree-optimization/57124
19014 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
19015 conversion feeding a condition if the range has an overflow
19016 if -fstrict-overflow. Add warnings for when we do make the
19019 2013-05-24 Dehao Chen <dehao@google.com>
19021 * tree-cfg.c (locus_discrim_map): Fix the typo.
19022 (locus_discrim_hasher): Likewise.
19023 (locus_discrim_hasher::hash): Likewise.
19024 (locus_discrim_hasher::equal): Likewise.
19026 2013-05-24 Martin Jambor <mjambor@suse.cz>
19028 PR tree-optimization/57294
19029 * cgraph.h (ipa_record_stmt_references): Declare.
19030 * cgraphbuild.c (ipa_record_stmt_references): New function.
19031 (build_cgraph_edges): Use ipa_record_stmt_references.
19032 (rebuild_cgraph_edges): Likewise.
19033 (cgraph_rebuild_references): Likewise.
19034 * ipa-prop.c (ipa_modify_call_arguments): Discard references
19035 associated with the old statement and build references from the
19036 newly built statements.
19037 * ipa-ref.c (ipa_remove_stmt_references): New function.
19038 * ipa-ref.h (ipa_remove_stmt_references): Declare.
19040 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
19042 * lra-constraints.c (emit_spill_move): Use smaller mode for
19044 (check_and_process_move): Consider mem-reg moves for secondary
19046 (curr_insn_transform): Don't lose insns emitted before for
19047 secondary memory moves.
19048 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
19049 reg set up in the current insn.
19051 2013-05-24 Dehao Chen <dehao@google.com>
19053 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
19055 (locus_descrim_hasher::equal): Likewise.
19056 (build_gimple_cfg): New discriminator assignment algorithm.
19057 (make_edges): Likewise.
19058 (next_discriminator_for_locus): Likewise.
19059 (same_line_p): Likewise.
19060 (assign_discriminators): Likewise.
19061 (make_cond_expr_edges): Likewise.
19062 (make_gimple_switch_edges): Likewise.
19063 (make_goto_expr_edges): Likewise.
19064 (make_gimple_asm_edges): Likewise.
19066 2013-05-24 Ian Bolton <ian.bolton@arm.com>
19068 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
19069 X format specifier to only display bottom 16 bits.
19070 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
19071 immediate to match for operand 2, since it will be masked.
19073 2013-05-24 Richard Biener <rguenther@suse.de>
19075 PR tree-optimization/57287
19076 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
19077 all SSA names that occur in abnormal PHIs.
19079 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
19082 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
19083 that index is not negative.
19085 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
19087 PR rtl-optimization/55177
19088 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
19089 (simplify_byte_swapping_operation): New.
19090 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
19091 (simplify_relational_operation_1): Deal with BSWAP.
19093 2013-05-23 Richard Henderson <rth@redhat.com>
19096 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
19097 (ix86_reorg): Call it.
19099 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
19102 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
19103 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
19104 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
19106 2013-05-23 Christian Bruel <christian.bruel@st.com>
19109 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
19111 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19112 Marcus Shawcroft <marcus.shawcroft@arm.com>
19114 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
19115 * config/aarch64/constraints.md (Usa): Remove.
19116 * doc/md.texi (AArch64 Usa): Remove.
19118 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19119 Marcus Shawcroft <marcus.shawcroft@arm.com>
19121 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
19122 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
19123 * config/aarch64/predicates.md (aarch64_const_address): Remove.
19124 (aarch64_mov_operand): Use aarch64_mov_operand_p.
19126 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
19128 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
19129 instruction (AdvSIMD).
19130 * config/aarch64/aarch64-builtins.c
19131 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
19132 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
19134 2013-05-23 Martin Jambor <mjambor@suse.cz>
19136 PR middle-end/57347
19137 * tree.h (contains_bitfld_component_ref_p): Declare.
19138 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
19139 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
19141 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
19142 not access a bit-field. Assert all final offsets are byte-aligned.
19144 2013-05-23 Richard Biener <rguenther@suse.de>
19146 PR tree-optimization/57380
19147 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
19148 least one invariant or re-used load.
19149 * passes.c (init_optimization_passes): Move pass_phiprop before
19152 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
19154 * config/aarch64/aarch64-simd.md
19155 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
19157 2013-05-23 Richard Biener <rguenther@suse.de>
19159 PR middle-end/57381
19160 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19161 OEP_CONSTANT_ADDRESS_OF retained.
19163 2013-05-23 Jakub Jelinek <jakub@redhat.com>
19165 PR middle-end/57344
19166 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
19167 don't lower unit. Handle unit not being always BITS_PER_WORD.
19169 2013-05-23 Richard Biener <rguenther@suse.de>
19171 PR rtl-optimization/57341
19172 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
19173 instead of true_dependence.
19175 2013-05-22 David Malcolm <dmalcolm@redhat.com>
19177 * bb-reorder.c (branch_threshold): Make const.
19178 (exec_threshold): Ditto.
19180 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19181 Pat Haugen <pthaugen@us.ibm.com>
19182 Peter Bergner <bergner@vnet.ibm.com>
19184 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
19185 documentation for the power8 crypto builtins.
19187 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
19189 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
19190 macros for defining power8 builtin functions.
19191 (BU_P8V_AV_2): Likewise.
19192 (BU_P8V_AV_P): Likewise.
19193 (BU_P8V_VSX_1): Likewise.
19194 (BU_P8V_OVERLOAD_1): Likewise.
19195 (BU_P8V_OVERLOAD_2): Likewise.
19196 (BU_CRYPTO_1): Likewise.
19197 (BU_CRYPTO_2): Likewise.
19198 (BU_CRYPTO_3): Likewise.
19199 (BU_CRYPTO_OVERLOAD_1): Likewise.
19200 (BU_CRYPTO_OVERLOAD_2): Likewise.
19201 (XSCVSPDP): Fix typo, point to the correct instruction.
19202 (VCIPHER): Add power8 crypto builtins.
19203 (VCIPHERLAST): Likewise.
19204 (VNCIPHER): Likewise.
19205 (VNCIPHERLAST): Likewise.
19206 (VPMSUMB): Likewise.
19207 (VPMSUMH): Likewise.
19208 (VPMSUMW): Likewise.
19209 (VPERMXOR_V2DI): Likewise.
19210 (VPERMXOR_V4SI: Likewise.
19211 (VPERMXOR_V8HI: Likewise.
19212 (VPERMXOR_V16QI: Likewise.
19213 (VSHASIGMAW): Likewise.
19214 (VSHASIGMAD): Likewise.
19215 (VPMSUM): Likewise.
19216 (VPERMXOR): Likewise.
19217 (VSHASIGMA): Likewise.
19219 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
19220 __CRYPTO__ if the crypto instructions are available.
19221 (altivec_overloaded_builtins): Add support for overloaded power8
19224 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
19225 support for power8 crypto builtins.
19226 (builtin_function_type): Likewise.
19227 (altivec_init_builtins): Add support for builtins that take vector
19228 long long (V2DI) arguments.
19230 * config/rs6000/crypto.md: New file, define power8 crypto
19233 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19234 Pat Haugen <pthaugen@us.ibm.com>
19235 Peter Bergner <bergner@vnet.ibm.com>
19237 * doc/invoke.texi (Option Summary): Add power8 options.
19238 (RS/6000 and PowerPC Options): Likewise.
19240 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
19241 constraints.md instead of rs6000.h. Reorder w* constraints. Add
19242 wm, wn, wr documentation.
19244 * config/rs6000/constraints.md (wm): New constraint for VSX
19245 registers if direct move instructions are enabled.
19246 (wn): New constraint for no registers.
19247 (wq): New constraint for quad word even GPR registers.
19248 (wr): New constraint if 64-bit instructions are enabled.
19249 (wv): New constraint if power8 vector instructions are enabled.
19250 (wQ): New constraint for quad word memory locations.
19252 * config/rs6000/predicates.md (const_0_to_15_operand): New
19253 constraint for 0..15 for crypto instructions.
19254 (gpc_reg_operand): If VSX allow registers in VSX registers as well
19255 as GPR and floating point registers.
19256 (int_reg_operand): New predicate to match only GPR registers.
19257 (base_reg_operand): New predicate to match base registers.
19258 (quad_int_reg_operand): New predicate to match even GPR registers
19259 for quad memory operations.
19260 (vsx_reg_or_cint_operand): New predicate to allow vector logical
19261 operations in both GPR and VSX registers.
19262 (quad_memory_operand): New predicate for quad memory operations.
19263 (reg_or_indexed_operand): New predicate for direct move support.
19265 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
19266 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
19267 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
19268 (POWERPC_MASKS): Add power8 options.
19269 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
19272 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
19273 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
19275 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
19276 (-mpower8-fusion): New power8 options.
19277 (-mpower8-fusion-sign): Likewise.
19278 (-mpower8-vector): Likewise.
19279 (-mcrypto): Likewise.
19280 (-mdirect-move): Likewise.
19281 (-mquad-memory): Likewise.
19283 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
19284 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
19286 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
19287 (rs6000_debug_vector_unit): Add p8_vector.
19288 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
19289 definitions. Also print fusion state.
19290 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
19291 (rs6000_builtin_mask_calculate): Add power8 builtin support.
19292 (rs6000_option_override_internal): Add support for power8.
19293 (rs6000_common_init_builtins): Add debugging for skipped builtins
19294 if -mdebug=builtin.
19295 (rs6000_adjust_cost): Add power8 support.
19296 (rs6000_issue_rate): Likewise.
19297 (insn_must_be_first_in_group): Likewise.
19298 (insn_must_be_last_in_group): Likewise.
19299 (force_new_group): Likewise.
19300 (rs6000_register_move_cost): Likewise.
19301 (rs6000_opt_masks): Likewise.
19303 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
19304 power8 capable assembler, default to power7 options.
19305 (TARGET_DIRECT_MOVE): Likewise.
19306 (TARGET_CRYPTO): Likewise.
19307 (TARGET_P8_VECTOR): Likewise.
19308 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
19309 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
19310 (VECTOR_MEM_P8_VECTOR_P): Likewise.
19311 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
19312 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
19313 (TARGET_XSCVDPSPN): Likewise.
19314 (TARGET_XSCVSPDPN): Likewsie.
19315 (TARGET_SYNC_HI_QI): Likewise.
19316 (TARGET_SYNC_TI): Likewise.
19317 (MASK_CRYPTO): Likewise.
19318 (MASK_DIRECT_MOVE): Likewise.
19319 (MASK_P8_FUSION): Likewise.
19320 (MASK_P8_VECTOR): Likewise.
19321 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
19322 temporary used by some of the direct move instructions to get two FP
19323 temporary registers does not force creation of a stack frame.
19324 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
19325 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
19326 that any VSX registers are tieable, even if they are also an
19327 Altivec vector mode.
19328 (r6000_reg_class_enum): Add wm, wr, wv constraints.
19329 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
19330 (RS6000_BTM_CRYPTO): Likewise.
19331 (RS6000_BTM_COMMON): Likewise.
19333 * config/rs6000/rs6000.md (cpu attribute): Add power8.
19334 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
19335 (enum rs6000_vector): Add power8 vector support.
19337 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19341 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
19342 (any_sibcall_could_use_r3): this and handle indirect calls.
19343 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
19345 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19347 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
19349 2013-05-22 Richard Biener <rguenther@suse.de>
19351 PR middle-end/57349
19352 * profile.c (branch_prob): Do not split blocks that are
19353 abnormally receiving from ECF_RETURNS_TWICE functions.
19355 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19357 * recog.c (offsettable_address_addr_space_p): Fix calculation of
19358 address mode. Move pointer mode initialization to the same place.
19360 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
19362 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
19363 while it has any effect.
19365 2013-05-21 Easwaran Raman <eraman@google.com>
19367 PR tree-optimization/57322
19368 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
19369 UID of the statement added to the BB to be 1.
19371 2013-05-21 Jakub Jelinek <jakub@redhat.com>
19373 PR tree-optimization/57331
19374 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
19375 of conversion from pointer type to integral type with integer.
19377 2013-05-21 Martin Jambor <mjambor@suse.cz>
19380 * ipa-prop.c (ipa_read_node_info): Process param_used and
19381 controlled_uses in the same order as when writing.
19383 2013-05-21 Magnus Granberg <baldrick@free.fr>
19386 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
19388 2013-05-21 Richard Biener <rguenther@suse.de>
19390 PR tree-optimization/57318
19391 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
19392 estimate stmts with side-effects as likely eliminated.
19394 2013-05-21 Richard Biener <rguenther@suse.de>
19396 PR tree-optimization/57330
19397 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
19398 preserve the call stmts fntype.
19400 2013-05-21 Richard Biener <rguenther@suse.de>
19402 PR tree-optimization/57303
19403 * tree-ssa-sink.c (statement_sink_location): Improve killing
19404 stmt detection and properly handle self-assignments.
19406 2013-05-21 Christian Bruel <christian.bruel@st.com>
19408 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
19409 spanning registers. LEAF_REG_REMAP is supported only for contiguous
19410 registers. Set register size out of the PARALLEL loop.
19412 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
19415 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
19416 (*fmasf4, *fmasf4_media): New insns.
19418 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19420 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
19421 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
19422 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
19423 (mips_idiv_insns): Update the comments to say that the returned
19424 instruction counts are in units of BASE_INSN_LENGTH.
19425 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
19426 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
19427 using 2 rather than 4 as the length of indirect MIPS16 and
19428 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
19429 length of a NOP. Don't divide MIPS16 lengths by 2.
19430 (mips16_split_long_branches): Assume a branch is long if the
19431 length is greater than 4 rather than 8.
19432 * config/mips/mips.md (length): Give MIPS16 lengths directly,
19433 rather than multiplying them by 2. Multiply instruction counts
19434 by BASE_INSN_LENGTH rather than 4.
19435 (*jump_mips16, tls_get_tp_mips16_<mode>)
19436 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
19438 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19440 * config/mips/mips.md (extended_mips16): Remove branch case.
19441 (length): Remove duplicated extended_mips16 test.
19443 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19445 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
19447 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
19449 * recog.h (Recog_data): Rename to...
19450 (recog_data_d): ...this.
19451 (recog_data): Update accordingly.
19452 * recog.c (recog_data): Likewise.
19453 * reload.c (save_recog_data): Likewise.
19454 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
19455 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
19457 2013-05-17 Julian Brown <julian@codesourcery.com>
19459 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
19460 found in a REG_EQUAL note, invalidate it.
19462 2013-05-17 Easwaran Raman <eraman@google.com>
19464 * tree-ssa-reassoc.c (find_insert_point): New function.
19465 (insert_stmt_after): Likewise.
19466 (get_def_stmt): Likewise.
19467 (ensure_ops_are_available): Likewise.
19468 (not_dominated_by): Likewise.
19469 (rewrite_expr_tree): Do not move statements beyond what is
19470 necessary. Remove call to swap_ops_for_binary_stmt...
19471 (reassociate_bb): ... and move it here.
19472 (build_and_add_sum): Assign UIDs for new statements.
19473 (linearize_expr): Likewise.
19474 (do_reassoc): Renumber gimple statement UIDs.
19476 2013-05-17 Jan Hubicka <jh@suse.cz>
19478 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
19480 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
19481 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
19483 (output_weakrefs): Update.
19485 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
19486 Martin Jambor <mjambor@suse.cz>
19488 PR middle-end/57276
19489 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
19490 value that corresponds to the given aggval is found in values vector.
19492 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
19494 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
19495 sse, sse2, sse3, ssse3 and sse4a flags to options.
19497 2013-05-17 David Malcolm <dmalcolm@redhat.com>
19499 * gengtype-state.c: (s_expr_writer): New class, to handle
19500 prettifying of output layout of s-expressions.
19501 (state_writer): New class, to write out gtype.state.
19502 (state_written_type_count): Move this variable into member data of
19504 (s_expr_writer::s_expr_writer): New code: constructor for new class
19505 (state_writer::state_writer(): ditto
19506 (s_expr_writer::write_new_line): New function
19507 (s_expr_writer::write_any_indent): ditto
19508 (s_expr_writer::begin_s_expr): ditto
19509 (s_expr_writer::end_s_expr): ditto
19510 (write_state_fileloc): convert to method of state_writer...
19511 (state_writer:: write_state_fileloc): ...and use methods of
19512 s_expr_writer to write indentation into the gtype.state output file
19513 to visually represent the hierarchical structure of the list
19515 (write_state_fields): ditto, renaming to...
19516 (state_writer::write_state_fields)
19517 (write_state_a_string): ditto, renaming to...
19518 (state_writer::write_state_a_string)
19519 (write_state_string_option): ditto, renaming to...
19520 (state_writer::write_state_string_option)
19521 (write_state_type_option): ditto, renaming to...
19522 (state_writer::write_state_type_option)
19523 (write_state_nested_option): ditto, renaming to...
19524 (state_writer::write_state_nested_option)
19525 (write_state_option): ditto, renaming to...
19526 (state_writer::write_state_option)
19527 (write_state_options): ditto, renaming to...
19528 (state_writer::write_state_options)
19529 (write_state_lang_bitmap): ditto, renaming to...
19530 (state_writer::write_state_lang_bitmap)
19531 (write_state_version): ditto, renaming to...
19532 (state_writer::write_state_version)
19533 (write_state_scalar_type): ditto, renaming to...
19534 (state_writer::write_state_scalar_type)
19535 (write_state_string_type): ditto, renaming to...
19536 (state_writer::write_state_string_type)
19537 (write_state_undefined_type): ditto, renaming to...
19538 (state_writer::write_state_undefined_type)
19539 (write_state_struct_union_type): ditto, renaming to...
19540 (state_writer::write_state_struct_union_type)
19541 (write_state_struct_type): ditto, renaming to...
19542 (state_writer::write_state_struct_type)
19543 (write_state_user_struct_type): ditto, renaming to...
19544 (state_writer::write_state_user_struct_type)
19545 (write_state_lang_struct_type): ditto, renaming to...
19546 (state_writer::write_state_lang_struct_type)
19547 (write_state_param_struct_type): ditto, renaming to...
19548 (state_writer::write_state_param_struct_type)
19549 (write_state_pointer_type): ditto, renaming to...
19550 (state_writer::write_state_pointer_type)
19551 (write_state_array_type): ditto, renaming to...
19552 (state_writer::write_state_array_type)
19553 (write_state_gc_used): ditto, renaming to...
19554 (state_writer::write_state_gc_used)
19555 (write_state_common_type_content): ditto, renaming to...
19556 (state_writer::write_state_common_type_content)
19557 (write_state_type): ditto, renaming to...
19558 (state_writer::write_state_type)
19559 (write_state_pair_list): ditto, renaming to...
19560 (state_writer::write_state_pair_list)
19561 (write_state_pair): ditto, renaming to...
19562 (state_writer::write_state_pair)
19563 (write_state_typedefs): ditto, renaming to...
19564 (state_writer::write_state_typedefs)
19565 (write_state_structures): ditto, renaming to...
19566 (state_writer::write_state_structures)
19567 (write_state_param_structs): ditto, renaming to...
19568 (state_writer::write_state_param_structs)
19569 (write_state_variables): ditto, renaming to...
19570 (state_writer::write_state_variables)
19571 (write_state_srcdir): ditto, renaming to...
19572 (state_writer::write_state_srcdir)
19573 (write_state_files_list): ditto, renaming to...
19574 (state_writer::write_state_files_list)
19575 (write_state_languages): ditto, renaming to...
19576 (state_writer::write_state_languages)
19577 (write_state): create a state_writer instance and use it when
19578 writing out the state file
19580 2013-05-17 Mike Stump <mikestump@comcast.net>
19582 PR rtl-optimization/57304
19583 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
19584 accessing DF_REF_REAL_LOC.
19586 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19588 PR rtl-optimization/57281
19589 PR rtl-optimization/57300
19590 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
19591 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
19592 what the other splitter did if the registers are dead.
19594 2013-05-17 Richard Biener <rguenther@suse.de>
19596 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
19599 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19601 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
19604 2013-05-17 Marek Polacek <polacek@redhat.com>
19606 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
19607 length when doing non-zero store of storing '\0' to '\0'.
19609 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19611 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
19612 vect_external_def oprnd1 with loop_vinfo, try to emit
19613 optional cast, negation and and stmts on the loop preheader
19614 edge instead of into the pattern def seq.
19616 PR tree-optimization/57051
19617 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
19618 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
19620 2013-05-16 Nick Clifton <nickc@redhat.com>
19622 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
19623 (rl78_is_naked_func): New function.
19624 (rl78_expand_prologue): Skip prologue generation for naked functions.
19625 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
19626 * doc/extend.texi (naked): Add RL78 to the list of processors
19627 that supports this attribute.
19629 2013-05-16 Jeff Law <law@redhat.com>
19631 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
19633 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19635 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
19636 cache parameters using detect_caches_amd also for CYRIX,
19637 NSC and TM2 signatures.
19639 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19640 Dzianis Kahanovich <mahatma@eu.by>
19644 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
19645 VIA/Centaur processors and determine their cache parameters
19646 using detect_caches_amd.
19648 2013-05-16 Teresa Johnson <tejohnson@google.com>
19650 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
19651 (rtl_verify_edges): New function.
19652 (rtl_verify_bb_insns): Ditto.
19653 (rtl_verify_bb_pointers): Ditto.
19654 (rtl_verify_bb_insn_chain): Ditto.
19655 (rtl_verify_fallthru): Ditto.
19656 (rtl_verify_bb_layout): Ditto.
19657 (rtl_verify_flow_info_1): Outline checks into new functions.
19658 (rtl_verify_flow_info): Ditto.
19660 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
19662 * cfghooks.c (copy_bbs): Add update_dominance argument.
19663 * cfghooks.h (copy_bbs): Update prototype.
19664 * tree-cfg.c (gimple_duplicate_sese_region):
19665 Add update_dominance argument.
19666 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
19667 * tree-ssa-loop-ch.c (copy_loop_headers): Update
19668 gimple_duplicate_sese_region call.
19669 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
19670 Update copy_bbs call.
19671 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
19672 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
19674 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19676 * tree-vectorizer.h (NUM_PATTERNS): Increment.
19677 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
19678 vect_recog_rotate_pattern.
19679 (vect_recog_rotate_pattern): New function.
19681 2013-05-16 Jason Merrill <jason@redhat.com>
19683 * Makefile.in (LLINKER): New variable.
19684 (mostlyclean): Remove link mutex.
19685 * configure.ac: Handle --enable-link-mutex.
19686 * lock-and-run.sh: New script.
19688 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19691 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
19694 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19696 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
19698 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
19700 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
19701 * config/arm/arm.c (next_consecutive_mem): New function.
19702 (gen_movmem_ldrd_strd): Likewise.
19703 * config/arm/arm.md (movmemqi): Update condition and code.
19704 (unaligned_loaddi, unaligned_storedi): New patterns.
19706 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19708 * config.gcc: Obsolete *-*-solaris2.9*.
19709 * doc/install.texi (Specific, *-*-solaris2*): Document it.
19711 2013-05-16 Richard Biener <rguenther@suse.de>
19713 * passes.c (init_optimization_passes): Move pass_parallelize_loops
19714 earlier, after GRAPHITE transforms and IV canonicalization.
19716 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19718 * omp-low.c (extract_omp_for_data): For collapsed loops,
19719 if at least one of the loops is known at compile time to
19720 iterate zero times, set count to 0.
19721 (expand_omp_regimplify_p): New function.
19722 (expand_omp_for_generic): For collapsed loops, if at least
19723 one of the loops isn't known to iterate at least once,
19724 add runtime check with setting count to 0.
19725 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
19726 For unsigned types if it isn't known at compile time that
19727 the loop will iterate at least once, add runtime check to bypass
19728 the whole loop if initial condition isn't true.
19730 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
19732 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
19734 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
19736 PR middle-end/57286
19737 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
19738 transformations to avoid an infinite loop.
19740 2013-05-16 Marek Polacek <polacek@redhat.com>
19742 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
19744 2013-05-15 Leif Ekblad <leif@rdos.net>
19746 * config/i386/i386.c (ix86_decompose_address): Use
19747 DEFAULT_TLS_SEG_REG to access TLS segment register.
19748 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
19749 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
19750 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
19752 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
19755 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
19756 sibling calls to functions that would normally be lazily bound,
19757 unless $gp is call-clobbered.
19759 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
19761 * config/i386/i386.c (ix86_option_override_internal): Update
19762 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
19763 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
19764 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
19765 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
19767 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
19769 2013-05-15 Andreas Schwab <schwab@suse.de>
19771 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
19772 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
19775 2013-05-15 Teresa Johnson <tejohnson@google.com>
19777 * loop-unroll.c (report_unroll_peel): Check decision before
19778 emitting unroll/peel message.
19780 2013-05-15 Teresa Johnson <tejohnson@google.com>
19782 * function.h (has_bb_partition): New rtl_data flag.
19783 (bb_reorder_complete): Ditto.
19784 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
19785 instead of flag_reorder_blocks_and_partition.
19786 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
19787 with some enhancements.
19788 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
19789 * bb-reorder.c (connect_traces): Check for has_bb_partition
19790 instead of flag_reorder_blocks_and_partition.
19791 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
19792 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
19793 verify_hot_cold_block_grouping.
19794 (partition_hot_cold_basic_blocks): Set has_bb_partition.
19796 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19799 * config/arm/predicates.md (call_insn_operand): New predicate.
19800 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
19801 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
19802 if insn is not a tail call.
19803 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
19805 * config/arm/arm.h (enum reg_class): New caller save register class.
19806 (REG_CLASS_NAMES): Likewise.
19807 (REG_CLASS_CONTENTS): Likewise.
19808 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
19811 2013-05-15 Richard Biener <rguenther@suse.de>
19813 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
19814 of MSG_OPTIMIZED_LOCATIONS.
19815 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
19816 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
19818 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
19819 of MSG_OPTIMIZED_LOCATIONS.
19820 (execute_vect_slp): Likewise.
19821 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
19822 (vect_create_cond_for_alias_checks): Likewise.
19823 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
19824 (vect_recog_widen_mult_pattern): Likewise.
19825 (vect_recog_widen_sum_pattern): Likewise.
19826 (vect_recog_over_widening_pattern): Likewise.
19827 (vect_recog_widen_shift_pattern): Likewise.
19828 (vect_recog_vector_vector_shift_pattern): Likewise.
19829 (vect_recog_divmod_pattern): Likewise.
19830 (vect_recog_mixed_size_cond_pattern): Likewise.
19831 (vect_recog_bool_pattern): Likewise.
19832 (vect_pattern_recog_1): Likewise.
19834 2013-05-15 Martin Jambor <mjambor@suse.cz>
19836 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
19837 non-functions to builtin_unreachable.
19838 * ipa-inline-transform.c (inline_call): Do not assert estimates were
19839 correct when new direct edges were discovered.
19841 2013-05-15 Martin Jambor <mjambor@suse.cz>
19843 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
19844 header, print symbol order instead of node uid, print more information
19845 about indirect edge targets.
19846 (ipa_make_edge_direct_to_target): Print symbol order instead of node
19848 (ipa_make_edge_direct_to_target): Likewise.
19849 (remove_described_reference): Likewise.
19850 (propagate_controlled_uses): Likewise.
19851 (ipa_print_node_params): Also print symbol order.
19852 (ipcp_transform_function): Print symbol order instead of node uids.
19853 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
19854 (cgraph_get_create_real_symbol_node): Likewise.
19855 * ipa-cp.c (print_lattice): Likewise.
19856 (print_all_lattices): Likewise.
19857 (determine_versionability): Likewise.
19858 (initialize_node_lattices): Likewise.
19859 (estimate_local_effects): Likewise.
19860 (update_profiling_info): Likewise.
19861 (create_specialized_node): Likewise.
19862 (perhaps_add_new_callers): Likewise.
19863 (decide_about_value): Likewise.
19864 (decide_whether_version_node): Likewise.
19865 (identify_dead_nodes): Likewise.
19866 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
19867 (dump_inline_summary): Likewise.
19868 (estimate_node_size_and_time): Likewise.
19869 (inline_analyze_function): Likewise.
19870 * ipa-inline.c (report_inline_failed_reason): Likewise.
19871 (want_early_inline_function_p): Likewise.
19872 (edge_badness): Likewise.
19873 (update_edge_key): Likewise.
19874 (inline_small_functions): Likewise. Add dumping of order to two other
19876 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
19877 instead of node uids.
19878 (propagate_pure_const): Likewise.
19879 (propagate_pure_const): Likewise.
19880 * ipa-utils.c (dump_cgraph_node_set): Likewise.
19881 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
19882 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
19884 * tree-pretty-print.c (dump_function_header): Likewise.
19885 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
19886 Print symbol order instead of node uids.
19888 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19890 * config/s390/s390.c (s390_register_move_cost): Don't impose the
19891 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
19893 2013-05-15 Richard Biener <rguenther@suse.de>
19895 PR tree-optimization/57275
19896 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
19897 return value for fail to do runtime alias checks for gather loads.
19899 2013-05-15 Jan Hubicka <jh@suse.cz>
19903 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
19904 weakrefs are not external.
19905 (lto_symtab_merge_decls): Fix thinko when dealing with
19906 non-lto_symtab decls.
19907 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
19908 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
19909 * varpool.c (dump_varpool_node): Dump more flags.
19911 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19913 * config/i386/i386.c (processor_alias_table): Add instruction
19914 FSGSBASE for AMD bdver3 architecture.
19916 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19918 * tree.c (warn_deprecated_use): Print file:line using locus color.
19919 * diagnostic.c (diagnostic_report_current_module): Print file:line
19920 and file:line:column using locus color.
19922 2013-05-14 Mike Stump <mikestump@comcast.net>
19924 * gdbinit.in: Add __null.
19926 2013-05-14 Mike Stump <mikestump@comcast.net>
19928 * recog.h: Rename struct recog_data to Recog_data.
19929 * recog.c: Likewise.
19930 * reload.c (can_reload_into): Likewise.
19931 * config/picochip/picochip.c: Likewise.
19933 2013-05-14 Mike Stump <mikestump@comcast.net>
19935 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
19937 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
19939 * resource.h (struct resources): Remove unch_memory member.
19940 (CLEAR_RESOURCE): Don't clear unch_memory.
19941 * resource.c (mark_referenced_resources): Don't set it.
19942 (mark_set_resources): Likewise.
19943 (mark_target_live_regs): Don't clear it.
19944 (init_resource_info): Likewise.
19945 * reorg.c (resource_conflicts_p): Don't compare it.
19946 (redundant_insn): Don't set it.
19948 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
19950 * emit-rtl.c (next_label): Remove unused function.
19951 (skip_consecutive_labels, link_cc0_insns): Move to ...
19952 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
19953 only place where these functions are used, and make them static.
19955 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
19957 * fold-const.c (fold_negate_expr): Handle vectors.
19958 (fold_truth_not_expr): Make it static.
19959 (fold_invert_truthvalue): New static function.
19960 (invert_truthvalue_loc): Handle vectors. Do not call
19961 fold_truth_not_expr directly.
19962 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
19963 <TRUTH_NOT_EXPR>: Do not cast to boolean.
19964 (fold_comparison): Handle vector constants.
19965 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
19966 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
19967 * tree.h (fold_truth_not_expr): Remove declaration.
19969 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
19971 * config/aarch64/aarch64-simd.md
19972 (aarch64_vcond_internal<mode>): Rename to...
19973 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
19974 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
19975 float modes. Clarify all iterator modes.
19976 (vcond<mode><mode>): Use new name for vcond expanders.
19977 (vcond<v_cmp_result><mode>): Likewise.
19978 (vcondu<mode><mode>: Likewise.
19979 * config/aarch64/iterators.md (VDQF_COND): New.
19981 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
19984 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
19985 variable for the shift amount. Check that we shift by non-negative
19988 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
19991 * config/arm/arm.h (EPILOGUE_USES): Only return true
19992 for LR_REGNUM after epilogue_completed.
19994 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
19996 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
19997 is error_mark_node.
19999 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20002 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
20003 and Solaris 11+/x86 with gld.
20004 * configure: Regenerate.
20006 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20008 * expmed.c (expand_shift_1): Canonicalize rotates by
20009 constant bitsize / 2 to bitsize - 1.
20010 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
20011 case ROTATERT>: Likewise.
20014 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20016 * config/i386/i386.md (rotateinv): New code attr.
20017 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20018 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20019 roll $31, %eax, etc.
20021 2013-05-14 Richard Biener <rguenther@suse.de>
20023 PR middle-end/57235
20024 * tree-eh.c (sink_clobbers): Give up for successors with
20025 multiple predecessors and no virtual uses.
20027 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
20029 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
20030 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
20032 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20034 PR middle-end/57251
20035 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
20036 the case when both op0 and op1 have VOIDmode.
20038 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20040 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
20041 in multiply-accumulate mode.
20043 2013-05-13 Guozhi Wei <carrot@google.com>
20045 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
20047 2013-05-13 Kai Tietz <ktietz@redhat.com>
20050 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
20051 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
20052 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
20053 * config/i386/i386.c (ix86_option_override_internal): Likewise.
20054 (ix86_expand_prologue): Likewise.
20055 (ix86_expand_split_stack_prologue): Likewise.
20056 (legitimate_pic_address_disp_p): Likewise.
20057 (legitimize_pic_address): Likewise.
20058 (legitimize_tls_address): Likewise.
20059 (legitimize_pe_coff_symbol): Likewise.
20060 (output_pic_addr_const): Likewise.
20061 (construct_plt_address): Likewise.
20062 (ix86_expand_call): Likewise.
20063 (x86_output_mi_thunk): Likewise.
20064 (x86_function_profiler): Likewise.
20066 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
20068 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
20069 similar switch cases.
20070 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
20071 (aarch64_simd_mov_to_<mode>low): Delete.
20072 (aarch64_simd_mov_to_<mode>high): Delete.
20073 (move_lo_quad_<mode>): Add w<-r alternative.
20074 (aarch64_simd_move_hi_quad_<mode>): Likewise.
20075 (aarch64_simd_mov_from_*): Update type attribute.
20076 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
20079 2013-05-13 Jan Hubicka <jh@suse.cz>
20081 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
20082 * config/i386/i386.c (ix86_compute_frame_layout,
20083 ix86_expand_epilogue, emit_i387_cw_initialization,
20084 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
20085 ix86_local_alignment): Fix use of size/speed predicates.
20087 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20089 PR tree-optimization/45216
20090 PR tree-optimization/57157
20091 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
20092 the (-Y) & (B - 1) variant if OP is |.
20093 * expmed.c (expand_shift_1): For rotations by const0_rtx just
20094 return shifted. Use (-op1) & (prec - 1) as other_amount
20095 instead of prec - op1.
20097 2013-05-13 Martin Jambor <mjambor@suse.cz>
20099 PR middle-end/42371
20100 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
20101 (ipa_constant_data): New type.
20102 (ipa_jump_func): Use ipa_constant_data to hold information about
20103 constant jump functions.
20104 (ipa_get_jf_constant): Adjust to jump function type changes.
20105 (ipa_get_jf_constant_rdesc): New function.
20106 (ipa_param_descriptor): New field controlled_uses.
20107 (ipa_get_controlled_uses): New function.
20108 (ipa_set_controlled_uses): Likewise.
20109 * ipa-ref.h (ipa_find_reference): Declare.
20110 * ipa-prop.c (ipa_cst_ref_desc): New type.
20111 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
20113 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
20114 New parameter cs. Adjust all callers.
20115 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
20116 (remove_described_reference): New function.
20117 (jfunc_rdesc_usable): Likewise.
20118 (try_make_edge_direct_simple_call): Decrement controlled use count,
20119 attempt to remove reference if it hits zero.
20120 (combine_controlled_uses_counters): New function.
20121 (propagate_controlled_uses): Likewise.
20122 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
20123 (ipa_edge_duplication_hook): Duplicate reference descriptions.
20124 (ipa_print_node_params): Print described use counter.
20125 (ipa_write_jump_function): Adjust to jump function type changes.
20126 (ipa_read_jump_function): New parameter CS, pass it to
20127 ipa_set_jf_constant. Adjust caller.
20128 (ipa_write_node_info): Stream controlled use count
20129 (ipa_read_node_info): Likewise.
20130 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
20132 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
20133 count. Remove cloning-added reference if it reaches zero.
20134 * ipa-ref.c (ipa_find_reference): New function.
20136 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20138 * config/i386/i386.c (processor_target_table): Modified default
20139 alignment values for AMD BD and BT architectures.
20141 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
20143 * tree-vect-generic.c (uniform_vector_p): Move ...
20144 * tree.c (uniform_vector_p): ... here.
20145 * tree.h (uniform_vector_p): Declare it.
20146 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
20149 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20151 PR tree-optimization/57230
20152 * tree-ssa-strlen.c (handle_char_store): Record length for
20153 array store from STRING_CST.
20155 PR tree-optimization/57230
20156 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
20159 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
20161 * config/epiphany/epiphany.c (epiphany_init): Check size of
20162 NUM_MODES_FOR_MODE_SWITCHING.
20163 (epiphany_expand_prologue):
20164 Remove CONFIG_REGNUM initial value handling code.
20165 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
20166 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
20167 (emit_set_fp_mode, epiphany_mode_after): Likewise.
20168 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
20169 Don't return 1 for FP_MODE_NONE.
20170 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
20171 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
20172 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
20173 * config/epiphany/epiphany.md (save_config): New pattern.
20175 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
20177 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
20179 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
20181 * config/i386/i386.md (memory): Handle sseishft1.
20182 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
20183 (*vec_extractv2di_1): Ditto.
20185 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
20187 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
20190 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
20192 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
20193 Add mthumb/march=armv7-a multilib.
20194 Add mthumb/march=armv7-r multilib.
20195 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
20197 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
20199 * config/v850/t-rtems: Add more multilibs.
20201 2013-05-10 Richard Biener <rguenther@suse.de>
20203 PR tree-optimization/57214
20204 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
20205 not propagate from SSA names that occur in abnormal PHI nodes.
20207 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
20209 * stor-layout.c (element_precision): New function.
20210 * machmode.h (element_precision): Declare it.
20211 * tree.c (build_minus_one_cst): New function.
20212 (element_precision): Likewise.
20213 * tree.h (build_minus_one_cst): Declare new function.
20214 (element_precision): Likewise.
20215 * fold-const.c (operand_equal_p): Use element_precision.
20216 (fold_binary_loc): Handle vector types.
20217 * convert.c (convert_to_integer): Use element_precision.
20218 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
20221 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
20223 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
20224 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
20225 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
20226 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20227 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
20228 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
20229 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
20230 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20231 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
20232 (Uuw8): New constraints.
20233 (Usb4): Move into alphabetical order.
20234 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
20235 (sd8_operand, ub8_operand, uw8_operand): New predicates.
20236 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
20237 previously unnamed patterns.
20238 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
20239 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
20240 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
20241 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
20242 of set_attr_alternative/if_then_else. Use extended_mips16 instead
20243 of specific lengths.
20245 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20247 * config/i386/i386.md (rotateinv): New code attr.
20248 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20249 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20250 roll $31, %eax, etc.
20252 PR tree-optimization/45216
20253 PR tree-optimization/57157
20254 * tree-ssa-forwprop.c (simplify_rotate): New function.
20255 (ssa_forward_propagate_and_combine): Call it.
20257 2013-05-10 Richard Biener <rguenther@suse.de>
20259 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
20260 disable peeling when we version for aliasing.
20261 (vector_alignment_reachable_p): Honor explicit user alignment.
20262 (vect_supportable_dr_alignment): Likewise.
20263 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
20264 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
20265 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
20266 then peeling to arrange for the cost-model check to come first.
20268 2013-05-10 Alan Modra <amodra@gmail.com>
20270 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
20271 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
20272 * configure: Regenerate.
20274 2013-05-10 Alan Modra <amodra@gmail.com>
20277 * varasm.c (default_elf_select_section): Move !DECL_P check..
20278 (get_named_section): ..to here before calling get_section_name.
20280 (default_section_type_flags): Add DECL_P check.
20281 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
20282 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
20284 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
20286 * config/epiphany/epiphany.c (epiphany_expand_prologue):
20287 When using gen_stack_adjust_str with a register offset, add a
20288 REG_FRAME_RELATED_EXPR note.
20290 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
20292 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
20293 (*vec_extractv4si_zext_mem): Ditto.
20294 (*vec_extractv2di): Add 0->x and x->x alternatives.
20295 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
20296 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
20298 2013-05-09 Jason Merrill <jason@redhat.com>
20300 N3639 C++1y VLA support
20301 * gimplify.c (gimplify_vla_decl): Don't touch an existing
20304 * tree.c (build_constructor_va): New.
20305 * tree.h: Declare it.
20307 2013-05-09 Martin Jambor <mjambor@suse.cz>
20310 * gimple-fold.c (canonicalize_constructor_val): Call
20311 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
20313 2013-05-09 Jan Hubicka <jh@suse.cz>
20314 Richard Biener <rguenther@suse.de>
20317 * symtab.c (symtab_make_decl_local): Do not add private names.
20319 2013-05-09 Jan Hubicka <jh@suse.cz>
20322 * symtab.c (insert_to_assembler_name_hash): Handle clones.
20323 (unlink_from_assembler_name_hash): Likewise.
20324 (symtab_prevail_in_asm_name_hash, symtab_register_node,
20325 symtab_unregister_node, symtab_initialize_asm_name_hash,
20326 change_decl_assembler_name): Update.
20328 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
20330 * config/aarch64/aarch64.md: New movtf split.
20331 (*movtf_aarch64): Update.
20332 (aarch64_movdi_tilow): Handle TF modes and rename to
20333 aarch64_movdi_<mode>low.
20334 (aarch64_movdi_tihigh): Handle TF modes and rename to
20335 aarch64_movdi_<mode>high
20336 (aarch64_movtihigh_di): Handle TF modes and rename to
20337 aarch64_mov<mode>high_di
20338 (aarch64_movtilow_di): Handle TF modes and rename to
20339 aarch64_mov<mode>low_di
20340 (aarch64_movtilow_tilow): Remove spurious whitespace.
20341 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
20343 (aarch64_print_operand): Update.
20345 2013-05-09 Alan Modra <amodra@gmail.com>
20347 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
20349 * configure: Regenerate.
20351 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20353 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
20354 splitter preparation statements.
20355 * config/i386/sse.md (*vec_extract* splitters): Ditto.
20356 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
20359 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20361 * gimple-ssa-strength-reduction.c (count_candidates): Change
20362 return value to int.
20363 (analyze_candidates_and_replace): Change type of length to int.
20365 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20367 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
20368 (*vec_extract<mode>): Use VI12_128 mode iterator.
20369 (*vec_extract<mode>_mem): Ditto.
20370 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
20373 2013-05-08 Diego Novillo <dnovillo@google.com>
20378 2012-08-17 Diego Novillo <dnovillo@google.com>
20381 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
20382 * config.in: Regenerate.
20383 * configure: Regenerate.
20384 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
20386 2013-05-08 Jan Hubicka <jh@suse.cz>
20389 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
20390 * cgraph.h (symtab_node_base): Add unique_name.
20391 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20392 input_overwrite_node, input_varpool_node): Stream unique_name.
20393 * cgraphclones.c (cgraph_create_virtual_clone,
20394 cgraph_function_versioning): Set unique_name.
20395 * ipa.c (function_and_variable_visibility): Set unique_name.
20397 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20399 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
20400 (alloc_cand_and_find_basis): Restrict conditional candidate
20401 processing to CAND_MULTs.
20403 2013-05-08 Jan Hubicka <jh@suse.cz>
20406 lto-symtab.c (lto_symtab_symbol_p): New function.
20407 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
20408 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
20409 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
20410 Skip static symbols.
20412 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
20414 PR tree-optimization/57200
20415 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
20416 Only call inform if the preceding warning_at returns true.
20418 2013-05-07 Han Shen <shenhan@google.com>
20420 * cfgexpand.c (record_or_union_type_has_array_p): New function.
20421 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
20422 * common.opt (fstack-protector-strong): New option.
20423 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
20424 * doc/invoke.texi (Optimization Options): Document
20425 "-fstack-protector-strong".
20426 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
20428 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20430 * config/mips/mips.c (mips_machine_reorg2): Return 0.
20432 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
20434 * ira.c (update_equiv_regs): Add insn having equiv memory even if
20435 it is not lhs of the insn.
20436 (setup_reg_equiv): Remove insn having equiv memory which it is not
20438 * lra-constraints.c (process_address): Try to improve generation
20439 code for address base + disp.
20440 (lra_constraints): Make correct the code for checking insn setting
20441 up backward equivalence. Remove insn only if it is in the init
20443 * lra-eliminations.c (update_reg_eliminate): Change return value.
20444 (lra_eliminate): Use the result.
20446 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
20448 * config/i386/sse.md (ssescalarnummask): New mode attribute.
20449 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
20450 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
20451 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
20452 register target operands.
20453 (*vec_extractv8hi_sse2): New pattern.
20454 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
20455 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
20456 (*vec_extract<mode>_mem): New insn and split pattern.
20458 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
20460 * config/arm/arm.c (arm_asan_shadow_offset): New function.
20461 (TARGET_ASAN_SHADOW_OFFSET): Define.
20462 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
20463 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
20465 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20467 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
20468 (incr_vec_index): Return -1 if increment not found.
20469 (create_add_on_incoming_edge): Assert if increment not found.
20470 (record_increment): Limit number of increments recorded.
20471 (all_phi_incrs_profitable): Return false if an increment not found.
20472 (replace_profitable_candidates): Don't process increments that were
20474 (analyze_candidates_and_replace): Limit size of incr_vec.
20476 2013-05-07 Richard Biener <rguenther@suse.de>
20478 * calls.c (special_function_p): setjmp-like functions are leaf.
20479 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
20480 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
20482 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
20484 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
20485 (aarch64_simd_mov<mode>): New expander.
20486 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
20487 (aarch64_simd_mov_to_<mode>high): Likewise.
20488 (aarch64_simd_mov_from_<mode>low): Likewise.
20489 (aarch64_simd_mov_from_<mode>high): Likewise.
20490 (aarch64_dup_lane<mode>): Update.
20491 (aarch64_dup_lanedi): New instruction pattern.
20492 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
20493 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
20495 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20497 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
20498 (replace_mult_candidate): Remove unnecessary argument; remove
20499 unnecessary parameter from call to introduce_cast_before_cand.
20500 (replace_unconditional_candidate): Remove unnecessary parameter
20501 from call to replace_mult_candidate.
20502 (replace_conditional_candidate): Likewise.
20503 (insert_initializers): Use make_temp_ssa_name.
20504 (introduce_cast_before_cand): Remove unnecessary argument; use
20505 make_temp_ssa_name.
20506 (replace_one_candidate): Remove unnecessary argument; remove
20507 unnecessary parameter from calls to introduce_cast_before_cand.
20508 (replace_profitable_candidates): Remove unnecessary parameters
20509 from calls to replace_one_candidate.
20511 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20513 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
20514 phi def as possibly hiding a basis for a CAND_ADD whose operands
20515 have been commuted in the analysis.
20516 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
20518 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20520 * config/aarch64/aarch64.md
20521 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
20522 shift value between 0-4.
20524 2013-05-07 Richard Biener <rguenther@suse.de>
20526 * double-int.h (rshift): New overload.
20527 * double-int.c (rshift): New function.
20528 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
20529 (create_reference_ops_from_ref): Remove.
20530 (vn_reference_insert): Use shared ops for constructing the
20531 reference and copy it.
20533 2013-05-07 Richard Biener <rguenther@suse.de>
20535 PR middle-end/57190
20536 * tree-eh.c (sink_clobbers): Properly propagate
20537 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
20539 2013-05-07 Jakub Jelinek <jakub@redhat.com>
20541 PR tree-optimization/57149
20542 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
20543 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
20544 collect_phi_def_edges, execute_late_warn_uninitialized): Use
20545 uninit_undefined_value_p instead of ssa_undefined_value_p.
20548 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
20549 for modifier == EXPAND_INITIALIZER.
20551 2013-05-07 Anton Blanchard <anton@samba.org>
20553 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
20554 for powerpc64 little endian.
20555 * configure: Regenerate.
20557 2013-05-06 Graham Stott <grahams@btinternet.com>
20559 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
20560 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
20561 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
20562 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
20564 2013-05-06 Graham Stott <grahams@btinternet.com>
20566 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
20567 codes which allow non-lvalues.
20569 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
20571 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
20572 components are all 1s.
20573 (integer_minus_onep): New function.
20574 * tree.h (integer_minus_onep): Declare it.
20575 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
20576 integer_minus_onep instead of integer_all_onesp.
20578 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20581 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
20582 variations of these patterns.
20584 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20586 * config/i386/i386.md (isa): Add x64_sse4 member.
20587 (enabled): Handle x64_sse4.
20588 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
20589 instruction for 64bit SSE4_1 targets. Update insn attributes.
20590 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
20591 instruction for SSE4_1 targets. Update insn attributes.
20592 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20593 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
20594 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
20596 (*vec_extractv4si): Rename from *sse4_1_pextrd.
20597 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
20598 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
20600 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20603 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
20605 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
20607 * final.c (do_assembler_dialects): Don't handle curly braces and
20608 vertical bar escaped by % as dialect delimiters.
20609 (output_asm_insn): Print curly braces and vertical bar if escaped
20610 by % and ASSEMBLER_DIALECT defined.
20611 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
20612 * doc/tm.texi: Regenerated.
20614 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20616 * config/mips/mips.c: Include tree-pass.h.
20617 (mips_reorg): Split in pre- and post-dbr_schedule parts.
20618 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
20619 (pass_mips_machine_reorg2): New machine specific pass.
20620 (insert_pass_mips_machine_reorg2): New pass plugin definition.
20621 (mips_option_override): Register the new pass.
20622 * rtl.h (cleanup_barriers): Remove prototype.
20623 (dbr_schedule): Likewise.
20624 * jump.c (cleanup_barriers): Make static.
20625 * reorg.c (dbr_schedule): Likewise.
20627 2013-05-06 Richard Biener <rguenther@suse.de>
20629 PR tree-optimization/57185
20630 * tree-parloops.c (add_field_for_reduction): Handle anonymous
20631 SSA names properly.
20633 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20636 * config/i386/i386.c (add_parameter_dependencies): Add dependence
20637 between "first_arg" and "insn", not "last" and "insn".
20639 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20641 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
20642 (find_candidates_in_block): Re-enable slsr_process_phi.
20643 (create_phi_basis): Fix double counting of candidate adjustment.
20645 2013-05-06 Richard Biener <rguenther@suse.de>
20647 PR middle-end/57147
20648 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
20649 the edge is also fallthru, preserve it and just clear the
20651 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
20652 also complex, preserve that and just clear the fallthru flag.
20653 * tree-inline.c (update_ssa_across_abnormal_edges): Also
20654 update virtual operands.
20656 2013-05-06 Alan Modra <amodra@gmail.com>
20658 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
20659 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
20660 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
20661 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
20662 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20663 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
20665 2013-05-06 Alan Modra <amodra@gmail.com>
20667 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
20668 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
20669 (DEFAULT_ASM_ENDIAN): Define.
20670 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20671 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
20672 Update -K PIC clause from sysv4.h.
20673 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
20674 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
20676 2013-05-06 Alan Modra <amodra@gmail.com>
20678 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
20679 twice for little-endian.
20680 (ashrdi3_no_power, ashrdi3): Support little-endian.
20682 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20685 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
20686 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
20688 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
20689 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
20690 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
20693 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
20694 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20696 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
20697 * config/pa/pa.opt: Make mbig-switch a no-op.
20698 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
20699 (CASE_VECTOR_MODE): Always return SImode.
20700 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
20701 for the !TARGET_BIG_SWITCH case.
20702 * config/pa/pa-linux.h: Likewise.
20703 * config/pa/pa-openbsd.h: Likewise.
20704 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
20705 * config/pa/pa.md (short_jump): Remove define_insn.
20706 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
20707 (casesi0): Remove define_insn.
20708 (type): Remove btable_branch.
20709 (pa_combine_type): Likewise.
20710 (in_nullified_branch_delay): Likewise.
20711 (in_call_delay): Likewise.
20712 (define_delay): Likewise.
20713 (define_insn_reservation "Z3"): Likewise.
20714 (define_insn_reservation "Z4"): Likewise.
20715 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
20716 (pa_adjust_insn_length): Remove adjustment for btable branches.
20717 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
20720 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
20722 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20723 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
20724 Add m->r,x alternatives.
20725 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
20726 splitters using SWI48x mode iterator.
20727 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
20728 TARGET_64BIT. Add m->x alternative.
20729 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
20730 Add o->x alternative. Enable for TARGET_SSE.
20731 (sse_storeq): Remove expander.
20732 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
20733 with memory input operand.
20734 (*vec_extractv2di_1 splitter): New.
20735 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
20736 * config/i386/i386.md (ssevecmodelower): New mode attribute.
20738 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20740 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
20741 (INT_LOWPART): Delete.
20742 (extract_MB): Adjust.
20743 (extract_ME): Adjust.
20744 (print_operand): Adjust.
20746 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20748 * config/rs6000/predicates.md (reg_or_add_cint_operand,
20749 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20750 (reg_or_logical_cint_operand, easy_fp_constant,
20751 logical_const_operand): Delete "CONST_DOUBLE" case.
20752 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
20753 "HOST_BITS_PER_WIDE_INT == 64" test.
20754 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
20755 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
20756 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
20757 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
20759 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
20760 CONST_DOUBLE DImode/VOIDmode case.
20761 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
20762 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20763 CONST_DOUBLE VOIDmode case.
20764 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20765 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
20766 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
20767 Delete CONST_DOUBLE case.
20768 (splitters for mov FMOVE64 const_double): Delete
20769 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20770 "HOST_BITS_PER_WIDE_INT >= 64" test.
20771 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
20773 (mov DI const_double): Delete.
20775 2013-05-04 Jakub Jelinek <jakub@redhat.com>
20777 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
20778 on op shows all bits zero in mode of a lowpart subreg, return zero.
20780 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
20783 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
20784 to save TFmode registers and DImode to save TImode registers for
20785 caller save operations.
20786 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
20787 mark being partially clobbered since they only use the first
20790 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
20791 and TDmode only use the upper 64-bits of each VSX register.
20793 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20795 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
20796 (find_candidates_in_block): Disable slsr_process_phi.
20798 2013-05-03 Guozhi Wei <carrot@google.com>
20800 * coverage.c (coverage_obj_init): Move the construction of gcov
20802 (build_init_ctor): ... here.
20804 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20806 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
20807 (slsr_cand_d): Redefine def_phi.
20808 (stride_status, phi_adjust_status, count_phis_status): New enums.
20809 (find_phi_def): New.
20810 (find_basis_for_base_expr): New.
20811 (find_basis_for_candidate): Handle hidden bases.
20812 (alloc_cand_and_find_basis): Handle phi candidates.
20813 (slsr_process_phi): New.
20814 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
20815 (create_mul_imm_cand): Likewise.
20816 (create_add_ssa_cand): Exclude phi base candidates.
20817 (create_add_imm_cand): Likewise.
20818 (slsr_process_cast): Likewise.
20819 (slsr_process_copy): Likewise.
20820 (find_candidates_in_block): Handle phi candidates.
20821 (dump_candidate): Likewise.
20822 (unconditional_cands): Delete.
20823 (unconditional_cands_with_known_stride_p): Delete.
20824 (phi_dependent_cand_p): New.
20825 (cand_increment): Handle phi-dependent candidates.
20826 (replace_dependent): Delete.
20827 (replace_mult_candidate): New.
20828 (replace_unconditional_candidate): New.
20829 (incr_vec_index): Move to avoid forward reference.
20830 (create_add_on_incoming_edge): New.
20831 (create_phi_basis): New.
20832 (replace_dependents): Delete.
20833 (replace_conditional_candidate): New.
20834 (phi_add_costs): New.
20835 (replace_uncond_cands_and_profitable_phis): New.
20836 (record_increment): Handle phi adjustments.
20837 (record_phi_increments): New.
20838 (record_increments): Handle phi adjustments.
20839 (phi_incr_cost): New.
20840 (lowest_cost_path): Handle phis.
20841 (total_savings): Likewise.
20842 (analyze_increments): Likewise.
20843 (ncd_with_phi): New.
20844 (ncd_of_cand_and_phis): New.
20845 (nearest_common_dominator_for_cands): Handle phi increments.
20846 (all_phi_incrs_profitable): New.
20847 (replace_profitable_candidates): Handle phi-dependent candidates.
20848 (analyze_candidates_and_replace): Likewise.
20850 2013-05-03 Teresa Johnson <tejohnson@google.com>
20853 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
20854 do not exceed REG_BR_PROB_BASE.
20856 2013-05-03 Jeff Law <law@redhat.com>
20858 PR tree-optimization/57144
20859 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
20860 operand of the condition will bit into the new type when eliminating
20861 a cast feeding a condition.
20863 2013-05-03 Jakub Jelinek <jakub@redhat.com>
20865 PR rtl-optimization/57130
20866 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
20867 of COMPARE as in_code to the recursive call if needed.
20869 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
20871 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
20872 (enabled): Handle new members.
20873 * config/i386/sse.md (*vec_concatv2si): Merge from
20874 *vec_concatv2si_sse2 and vec_concatv2si_sse.
20875 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
20877 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
20879 PR tree-optimization/57027
20880 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
20881 for fnms opportunity, check we got the prerequisite kind
20882 of tree / gimple before using accessor functions.
20884 2013-05-03 Richard Biener <rguenther@suse.de>
20886 * double-int.h (lshift): New overload without precision
20887 and arith argument.
20888 (operator *=, operator +=, operator -=): Move ...
20889 * double-int.c (operator *=, operator +=, operator -=): ... here
20890 and implement more efficiently.
20891 (mul_double_with_sign): Remove.
20892 (lshift_double): Adjust to take unsinged shift argument, push
20893 dispatching code to callers.
20894 (mul_double_wide_with_sign): Add early out for callers that
20895 are not interested in high parts or overflow.
20896 (lshift): New function.
20897 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
20898 dispatch code here.
20899 (lrotate, rrotate): Use logical shifts.
20900 * expr.c (get_inner_reference): Use lshift.
20901 * fixed-value.c (do_fixed_divide): Likewise.
20902 * tree-dfa.c (get_ref_base_and_extent): Likewise.
20903 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
20904 (indirect_refs_may_alias_p): Likewise.
20905 (stmt_kills_ref_p_1): Likewise.
20907 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20909 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
20911 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20913 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
20914 scalar form of FABD instruction.
20916 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
20918 * lra-constraints.c (process_alt_operands): Add checking alt
20919 number to choose the best alternative.
20921 2013-05-02 Richard Biener <rguenther@suse.de>
20923 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
20924 bitmap and its handling.
20925 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
20927 2013-05-02 Richard Biener <rguenther@suse.de>
20929 PR middle-end/57140
20930 * tree-inline.c (copy_loops): Properly handle removed loops.
20931 (copy_cfg_body): Mark destination loops for fixup if source
20932 loops needed fixup.
20934 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
20937 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
20938 generating simple_return for naked functions.
20940 2013-05-02 Martin Jambor <mjambor@suse.cz>
20942 PR middle-end/56988
20943 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
20944 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
20946 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
20947 ipa_agg_replacement_value structures.
20948 (known_aggs_to_agg_replacement_list): Likewise.
20949 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
20950 (read_agg_replacement_chain): Likewise.
20951 (ipcp_transform_function): Also check that by_ref flags match.
20953 2013-05-02 Richard Biener <rguenther@suse.de>
20955 * graphds.h (struct graph): Add obstack member.
20956 * graphds.c (new_graph): Initialize obstack and allocate
20958 (add_edge): Allocate edge from the obstack.
20959 (free_graph): Free the obstack instead of all edges and vertices.
20961 2013-05-02 Teresa Johnson <tejohnson@google.com>
20963 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
20965 * cfg.c (update_bb_profile_for_threading): Ditto.
20966 * tree-inline.c (copy_bb): Ditto.
20967 (copy_edges_for_bb): Ditto.
20968 (initialize_cfun): Ditto.
20969 (copy_cfg_body): Ditto.
20970 (expand_call_inline): Ditto.
20971 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
20972 (estimate_node_size_and_time): Ditto.
20973 (inline_merge_summary): Ditto.
20974 * cgraphclones.c (cgraph_clone_edge): Ditto.
20975 (cgraph_clone_node): Ditto.
20976 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20977 (compute_trg_info): Ditto.
20979 2013-05-02 Ian Bolton <ian.bolton@arm.com>
20981 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
20982 S reg when fp attribute set.
20983 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
20985 2013-05-02 Ian Bolton <ian.bolton@arm.com>
20987 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
20989 (*and_one_cmplsi3_compare0_uxtw): Likewise.
20990 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
20991 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
20993 2013-05-02 Richard Biener <rguenther@suse.de>
20995 * tree-scalar-evolution.c (scev_info_hasher): Remove.
20996 (struct instantiate_cache_entry): New type.
20997 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
20998 (struct instantiate_cache_type): New type.
20999 (set_instantiated_value, get_instantiated_value): Remove.
21000 (get_instantiated_value_entry): New function.
21001 (instantiate_scev_name): Use the new cache and adjust.
21002 (instantiate_scev_poly): Adjust.
21003 (instantiate_scev_binary): Likewise.
21004 (instantiate_array_ref): Likewise.
21005 (instantiate_scev_convert): Likewise.
21006 (instantiate_scev_not): Likewise.
21007 (instantiate_scev_3): Likewise.
21008 (instantiate_scev_2): Likewise.
21009 (instantiate_scev_r): Likewise.
21010 (instantiate_scev): Likewise.
21011 (resolve_mixers): Likewise.
21013 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
21016 * lra-constraints.c (best_small_class_operands_num): Remove.
21017 (process_alt_operands): Remove small_class_operands_num. Take
21018 small classes operands into losers and only if the operand is not
21019 matched. Modify debugging output.
21020 (curr_insn_transform): Remove best_small_class_operands_num.
21023 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21025 * config/aarch64/aarch64-builtins.c
21026 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
21027 * config/aarch64/aarch64-simd-builtins.def
21028 (reduc_splus_): Add new modes.
21029 (reduc_uplus_): New.
21030 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
21031 (reduc_uplus_v4sf): Likewise.
21032 (reduc_splus_v4sf): Likewise.
21033 (aarch64_addv<mode>): Likewise.
21034 (reduc_uplus_<mode>): Likewise.
21035 (reduc_splus_<mode>): Likewise.
21036 (aarch64_addvv2di): Likewise.
21037 (reduc_uplus_v2di): Likewise.
21038 (reduc_splus_v2di): Likewise.
21039 (aarch64_addvv2si): Likewise.
21040 (reduc_uplus_v2si): Likewise.
21041 (reduc_splus_v2si): Likewise.
21042 (reduc_<sur>plus_<mode>): New.
21043 (reduc_<sur>plus_v2di): Likewise.
21044 (reduc_<sur>plus_v2si): Likewise.
21045 (reduc_<sur>plus_v4sf): Likewise.
21046 (aarch64_addpv4sf): Likewise.
21047 * config/aarch64/arm_neon.h
21048 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
21049 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
21050 add UNSPEC_SADDV, UNSPEC_UADDV.
21052 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
21054 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21056 * config/aarch64/arm_neon.h
21057 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
21059 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21061 * config/aarch64/aarch64-builtins
21062 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
21064 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21066 * config/aarch64/aarch64-simd-builtins.def
21067 (reduc_smax_): New.
21068 (reduc_smin_): Likewise.
21069 (reduc_umax_): Likewise.
21070 (reduc_umin_): Likewise.
21071 (reduc_smax_nan_): Likewise.
21072 (reduc_smin_nan_): Likewise.
21075 (smax): Update for V2SF, V4SF and V2DF modes.
21078 (smin_nan): Likewise.
21079 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
21080 (<su><maxmin><mode>3): ...This, refactor.
21081 (s<maxmin><mode>3): New.
21082 (<maxmin_uns><mode>3): Likewise.
21083 (reduc_<maxmin_uns>_<mode>): Refactor.
21084 (reduc_<maxmin_uns>_v4sf): Likewise.
21085 (reduc_<maxmin_uns>_v2si): Likewise.
21086 (aarch64_<fmaxmin><mode>: Remove.
21087 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
21089 (vmin<q>_f<32,64>): Likewise.
21090 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21092 (su): Add mappings for smax, smin, umax, umin.
21094 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21095 (FMAXMIN): Rename as...
21096 (FMAXMIN_UNS): ...This.
21098 (fmaxminv): Likewise.
21099 (fmaxmin): Likewise.
21101 (maxmin_uns_op): Likewise.
21103 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21105 * config/aarch64/arm_neon.h
21106 (vac<ge, gt><sd>_f<32, 64>): Rename to...
21107 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
21108 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
21110 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21112 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
21113 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
21115 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21117 * config/aarch64/aarch64-simd.md
21118 (vcond<mode>_internal): Handle special cases for constant masks.
21119 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
21120 (vcondu<mode><mode>): Likewise.
21121 (vcond<v_cmp_result><mode>): New.
21123 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21125 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
21126 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
21127 * config/aarch64/aarch64-simd-builtins.def
21128 (cmeq): Update to BUILTIN_VALLDI.
21133 * config/aarch64/arm_neon.h
21134 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
21135 to builtins or C as appropriate.
21137 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21139 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
21141 (cmhi): Rename to...
21143 * config/aarch64/aarch64-simd.md
21144 (simd_mode): Add SF.
21145 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
21146 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
21147 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
21148 (cstore<mode>_neg): ...This.
21149 * config/aarch64/iterators.md
21151 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
21152 (COMPARISONS): New.
21153 (UCOMPARISONS): Likewise.
21154 (optab): Add missing comparisons.
21160 (VCMP_S): Likewise.
21161 (VCMP_U): Likewise.
21162 (V_cmp_result): Add DF, SF modes.
21163 (v_cmp_result): Likewise.
21165 (vmtype): Likewise.
21166 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
21168 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
21170 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
21171 define_insn to define_insn_and_split.
21172 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
21173 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
21174 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
21175 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
21176 (thumb2_negscc): Likewise.
21178 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21180 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
21182 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21184 * config/arm/thumb2.md: Remove trailing whitespaces.
21186 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21188 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
21189 Use gen_int_mode rather than GEN_INT.
21191 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
21193 * value-prof.c (stream_in_histogram_value): Remove the strayed
21196 2013-04-30 Richard Biener <rguenther@suse.de>
21198 PR middle-end/57122
21199 * cfghooks.c (split_edge): Properly check for the loop latch edge.
21201 2013-04-30 Richard Biener <rguenther@suse.de>
21203 PR middle-end/57107
21204 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
21206 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
21208 PR rtl-optimization/56957
21209 PR rtl-optimization/57105
21210 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
21211 variable. Use just INSN_UID for determining whether an insn
21212 should be only disconnected from the insn stream.
21213 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
21215 2013-04-30 Jakub Jelinek <jakub@redhat.com>
21217 PR tree-optimization/57104
21218 * tsan.c (instrument_expr): Don't instrument accesses to
21219 DECL_HARD_REGISTER VAR_DECLs.
21221 2013-04-30 Richard Biener <rguenther@suse.de>
21223 * function.h (loops_for_fn): New inline function.
21224 (set_loops_for_fn): Likewise.
21225 * cfgloop.h (place_new_loop): Add struct function parameter.
21226 (get_loop): Likewise.
21227 (get_loops): Likewise.
21228 (number_of_loops): Likewise.
21229 (fel_next): Adjust.
21230 (fel_init): Likewise.
21231 * cfg.c (get_loop_copy): Adjust.
21232 * cfgloop.c (flow_loops_dump): Likewise.
21233 (record_loop_exits): Likewise.
21234 (verify_loop_structure): Likewise.
21235 * cfgloopanal.c (mark_irreducible_loops): Likewise.
21236 (estimate_reg_pressure_cost): Likewise.
21237 (mark_loop_exit_edges): Likewise.
21238 * cfgloopmanip.c (place_new_loop): Likewise.
21239 (add_loop): Likewise.
21240 (duplicate_loop): Likewise.
21241 * graph.c (draw_cfg_nodes): Likewise.
21242 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
21243 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
21244 (extract_affine_chrec): Likewise.
21245 (build_scop_iteration_domain): Likewise.
21246 * graphite.c (graphite_initialize): Likewise.
21247 * ira-build.c (create_loop_tree_nodes): Likewise.
21248 (more_one_region_p): Likewise.
21249 (rebuild_regno_allocno_maps): Likewise.
21250 (mark_loops_for_removal): Likewise.
21251 (mark_all_loops_for_removal): Likewise.
21252 (remove_unnecessary_regions): Likewise.
21253 (ira_build): Likewise.
21254 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21255 * loop-init.c (fix_loop_structure): Likewise.
21256 (gate_rtl_move_loop_invariants): Likewise.
21257 (gate_rtl_unswitch): Likewise.
21258 (gate_rtl_unroll_and_peel_loops): Likewise.
21259 (rtl_doloop): Likewise.
21260 * lto-streamer-in.c (input_cfg): Likewise.
21261 * lto-streamer-out.c (output_cfg): Likewise.
21262 * modulo-sched.c (sms_schedule): Likewise.
21263 * predict.c (tree_estimate_probability): Likewise.
21264 (tree_estimate_probability_driver): Likewise.
21265 (estimate_loops): Likewise.
21266 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
21267 (move_sese_region_to_fn): Likewise.
21268 (debug_loop_num): Likewise.
21269 * tree-chrec.c (chrec_evaluate): Likewise.
21270 (hide_evolution_in_other_loops_than_loop): Likewise.
21271 (chrec_component_in_loop_num): Likewise.
21272 (reset_evolution_in_loop): Likewise.
21273 (evolution_function_is_invariant_rec_p): Likewise.
21274 * tree-if-conv.c (main_tree_if_conversion): Likewise.
21275 * tree-inline.c (copy_loops): Likewise.
21276 (copy_cfg_body): Likewise.
21277 (tree_function_versioning): Likewise.
21278 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
21279 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
21281 (add_to_evolution_1): Likewise.
21282 (scev_const_prop): Likewise.
21283 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
21284 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
21285 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
21286 (tree_ssa_lim_initialize): Likewise.
21287 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
21288 (verify_loop_closed_ssa): Likewise.
21289 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
21290 (tree_ssa_loop_im): Likewise.
21291 (tree_ssa_loop_unswitch): Likewise.
21292 (tree_vectorize): Likewise.
21293 (check_data_deps): Likewise.
21294 (tree_ssa_loop_ivcanon): Likewise.
21295 (tree_ssa_loop_bounds): Likewise.
21296 (tree_complete_unroll): Likewise.
21297 (tree_complete_unroll_inner): Likewise.
21298 (tree_parallelize_loops): Likewise.
21299 (tree_ssa_loop_prefetch): Likewise.
21300 (tree_ssa_loop_ivopts): Likewise.
21301 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
21302 * tree-vectorizer.c (vectorize_loops): Likewise.
21304 2013-04-29 Mike Frysinger <vapier@gentoo.org>
21306 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
21307 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
21308 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
21309 with EABI_LINK_SPEC.
21311 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21314 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
21317 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
21320 * lra-constraints.c (process_alt_operands): Discourage a bit more
21321 using memory for pseudos. Print cost dump for alternatives.
21322 Modify cost values for conflicts with early clobbers.
21323 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
21325 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21328 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
21330 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21332 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
21333 from/to S register.
21334 (movdi_aarch64): Support LDR/STR from/to D register.
21336 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21338 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
21339 or higher by default.
21341 2013-04-29 Richard Biener <rguenther@suse.de>
21343 PR middle-end/57075
21344 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
21345 even if not adding abnormal edges for calls that can make
21348 2013-04-29 Richard Biener <rguenther@suse.de>
21350 PR middle-end/57103
21351 * tree-cfg.c (move_stmt_op): Fix condition under which to update
21353 (move_stmt_r): Remove redundant checking.
21355 2013-04-29 Teresa Johnson <tejohnson@google.com>
21358 * basic-block.h (apply_scale): New function.
21359 (apply_probability): Use apply_scale.
21360 * gimple-streamer-in.c (input_bb): Ditto.
21361 * lto-streamer-in.c (input_cfg): Ditto.
21362 * lto-cgraph.c (merge_profile_summaries): Ditto.
21363 * tree-optimize.c (execute_fixup_cfg): Ditto.
21364 * tree-inline.c (copy_bb): Update comment to use apply_scale.
21365 (copy_edges_for_bb): Ditto.
21366 (copy_cfg_body): Ditto.
21368 2013-04-29 Tom de Vries <tom@codesourcery.com>
21370 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
21371 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
21372 (tail_merge_optimize): Handle current_loops == NULL.
21374 2013-04-26 Jeff Law <law@redhat.com>
21376 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
21377 (simplify_cond_using_ranges): Generalize code to simplify
21378 COND_EXPRs where one argument is a constant and the other
21379 is an SSA_NAME created by an integral type conversion.
21381 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21383 * config/arm/arm.md (store_minmaxsi): Use only when
21384 optimize_insn_for_size_p.
21386 2013-04-29 Christian Bruel <christian.bruel@st.com>
21389 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
21391 2013-04-29 Richard Biener <rguenther@suse.de>
21393 PR middle-end/57089
21394 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
21395 loop tree make sure to schedule a fixup for the child as well.
21396 (expand_omp_for_generic): Properly add loops.
21397 (expand_omp_for_static_nochunk): Likewise.
21398 (expand_omp_for_static_chunk): Likewise.
21399 (expand_omp_for): For the degenerate case fixup loops.
21400 (expand_omp_sections): Fix default bb placement in loops.
21401 (expand_omp_atomic_pipeline): Properly add loops.
21403 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21405 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
21407 2013-04-29 Tom de Vries <tom@codesourcery.com>
21409 * tree-ssa-tail-merge.c: Update header comment.
21411 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21413 * config/aarch64/arm_neon.h
21414 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
21415 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
21416 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
21417 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
21418 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
21419 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
21420 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
21421 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
21423 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21425 * config/aarch64/aarch64-simd.md
21426 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
21427 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
21428 fix_trunc, fixuns_trunc.
21429 (ftrunc<VDQF:mode>2): New.
21430 * config/aarch64/iterators.md (optab): Add fix, fixuns.
21431 (fix_trunc_optab): New.
21433 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21435 * config/aarch64/aarch64-builtins.c
21436 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
21437 iceilf, lround, iroundf.
21439 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21442 * config/i386/i386.h (enum ix86_tune_indices)
21443 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
21444 New, split from X86_TUNE_INTER_UNIT_MOVES.
21445 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
21446 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
21447 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
21448 (TARGET_INTER_UNIT_MOVES): Remove.
21449 * config/i386/i386.c (initial_ix86_tune_features): Update.
21450 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
21451 (ix86_expand_convert_uns_didf_sse): Use
21452 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21453 (ix86_expand_vector_init_one_nonzero): Ditto.
21454 (ix86_expand_vector_init_interleave): Ditto.
21455 (inline_secondary_memory_needed): Return true for moves from SSE class
21456 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
21457 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
21458 * config/i386/constraints.md (Yi, Ym): Depend on
21459 TARGET_INTER_UNIT_MOVES_TO_VEC.
21460 (Yj, Yn): New constraints.
21461 * config/i386/i386.md (*movdi_internal): Change constraints of
21462 operand 1 from Yi to Yj and from Ym to Yn.
21463 (*movsi_internal): Ditto.
21464 (*movdf_internal): Ditto.
21465 (*movsf_internal): Ditto.
21466 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
21467 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21468 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
21469 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
21470 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
21471 * config/i386/sse.md (movdi_to_sse): Ditto.
21472 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
21473 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
21474 TARGET_INTER_UNIT_MOVES.
21475 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
21476 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
21477 instead of TARGET_INTER_UNIT_MOVES.
21478 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
21479 operand 1 from Yi to Yj and from Ym to Yn.
21481 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21483 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
21484 (float_truncate_hi_): Likewise.
21485 (float_extend_lo_): Likewise.
21486 (float_truncate_lo_): Likewise.
21487 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
21488 (aarch64_float_extend_lo_v2df): Likewise.
21489 (vec_unpacks_hi_v4sf): Likewise.
21490 (aarch64_float_truncate_lo_v2sf): Likewise.
21491 (aarch64_float_truncate_hi_v4sf): Likewise.
21492 (vec_pack_trunc_v2df): Likewise.
21493 (vec_pack_trunc_df): Likewise.
21495 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21497 * config/aarch64/aarch64-builtins.c
21498 (aarch64_fold_builtin): Fold float conversions.
21499 * config/aarch64/aarch64-simd-builtins.def
21500 (floatv2si, floatv4si, floatv2di): New.
21501 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
21502 * config/aarch64/aarch64-simd.md
21503 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
21504 * config/aarch64/iterators.md (FLOATUORS): New.
21505 (optab): Add float, floatuns.
21506 (su_optab): Likewise.
21508 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21510 * config/aarch64/aarch64-builtins.c
21511 (aarch64_builtin_vectorized_function): Use new names for
21513 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
21514 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
21515 (fcvtzu): Split as...
21516 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
21517 (fcvtas): Split as...
21518 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
21519 (fcvtau): Split as...
21520 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
21521 (fcvtps): Split as...
21522 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
21523 (fcvtpu): Split as...
21524 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
21525 (fcvtms): Split as...
21526 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
21527 (fcvtmu): Split as...
21528 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
21529 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
21530 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
21531 (lfrintnusf, lfrintnudf): Likewise.
21532 * config/aarch64/aarch64-simd.md
21533 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
21535 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
21536 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
21537 (fcvt_pattern): Likewise.
21539 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21541 * config/aarch64/aarch64-simd.md
21542 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
21543 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
21545 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21547 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
21548 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
21549 (vrnd<a,m,n,p>_f32): Implement using builtins.
21550 (vrnd<i,x><q>_f<32, 64>): New.
21552 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21554 * config/aarch64/aarch64-builtins.c
21555 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
21556 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
21557 (frintz): Rename to...
21559 (frintp): Rename to...
21561 (frintm): Rename to...
21563 (frinti): Rename to...
21564 (nearbyint): ...this.
21565 (frintx): Rename to...
21567 (frinta): Rename to...
21569 * config/aarch64/aarch64-simd.md
21570 (aarch64_frint<frint_suffix><mode>): Delete.
21571 (<frint_pattern><mode>2): Convert to insn.
21572 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
21573 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
21574 (frint_pattern): Likewise.
21575 (frint_suffix): Likewise.
21577 2013-04-29 Richard Biener <rguenther@suse.de>
21579 PR tree-optimization/57081
21580 * loop-init.c: Include tree-flow.h.
21581 (loop_optimizer_finalize): Free number of iteration estimates.
21582 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
21584 2013-04-29 Jakub Jelinek <jakub@redhat.com>
21586 PR tree-optimization/57083
21587 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
21588 non-singleton shift count range, zero extend low_bound for uns case.
21590 * config/i386/predicates.md (general_vector_operand): New predicate.
21591 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
21592 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
21593 if they aren't nonimmediate operands. If their original values
21594 satisfy const_vector_equal_evenodd_p, don't shift them.
21595 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
21596 predicates. For the SSE4.1 case force operands[{1,2}] into registers
21597 if not nonimmediate_operand.
21598 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
21599 instead of register_operand.
21600 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
21602 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
21604 * stor-layout.c (finalize_size_functions): Allocate a structure and
21605 reset cfun before dumping the functions.
21607 2013-04-27 Jakub Jelinek <jakub@redhat.com>
21609 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
21612 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
21613 use xop_pmacsdqh if uns_p.
21614 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
21615 the immediate rotate count.
21617 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
21619 * rtl.h (struct rtx_def): Add comment for field jump.
21620 (LRA_SUBREG_P): New macro.
21621 * recog.c (register_operand): Check LRA_SUBREG_P.
21622 * lra.c (lra): Add note at the end of RTL code. Align non-empty
21624 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
21625 (lra_final_code_change): Skip subreg change for operators.
21626 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
21627 if there are no operand changes.
21628 * lra-constraints.c (curr_insn_set): New.
21629 (match_reload): Set LRA_SUBREG_P.
21630 (emit_spill_move): Ditto.
21631 (check_and_process_move): Use curr_insn_set. Process only single
21632 set insns. Don't initialize sec_mem_p and change_p.
21633 (simplify_operand_subreg): Use LRA_SUBREG_P.
21634 (reg_in_class_p): New function.
21635 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
21636 of #ifdef. Add code to remove cycling.
21637 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
21638 non-null disp. Reload inner instead of disp when base and index
21639 are null. Try to put lo_sum into register.
21640 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
21641 (check_and_process_move): Move code for move cost check to
21642 simple_move_p. Remove equiv_substitution.
21643 (simple_move_p): New function.
21644 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
21645 curr_insn_set. Call check_and_process_move only for single set
21646 insns. Use the new function. Move call of check_and_process_move
21647 after operand equiv substitution and address process.
21649 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21652 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
21653 with nonlocal goto receivers or returns twice calls, ignore
21654 unininitialized values from abnormal edges to nl goto receiver
21655 or returns twice call.
21657 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21659 PR tree-optimization/57051
21660 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
21661 and VEC_RSHIFT_EXPR if shift count is a multiple of element
21664 2013-04-26 Richard Biener <rguenther@suse.de>
21666 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
21667 (expand_omp_taskreg): Likewise. Mark loops for fixup.
21668 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
21669 (fixup_loop_arrays_after_move): New function.
21670 (move_sese_region_to_fn): Properly outline the loop tree parts
21671 of the SESE region.
21673 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
21675 * config/i386/i386.md (type, unit): Fix long lines.
21677 2013-04-26 Richard Biener <rguenther@suse.de>
21679 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
21680 (lto-streamer-out.o): Likewise.
21681 * cfgloop.c (init_loops_structure): Export, add struct function
21682 argument and adjust.
21683 (flow_loops_find): Adjust.
21684 * cfgloop.h (enum loop_estimation): Add EST_LAST.
21685 (init_loops_structure): Declare.
21686 * lto-streamer-in.c: Include cfgloop.h.
21687 (input_cfg): Input the loop tree.
21688 * lto-streamer-out.c: Include cfgloop.h.
21689 (output_cfg): Output the loop tree.
21690 (output_struct_function_base): Do not drop PROP_loops.
21692 2013-03-26 Richard Biener <rguenther@suse.de>
21694 * tree-cfg.c (execute_build_cfg): Build the loop tree.
21695 (pass_build_cfg): Provide PROP_loops.
21696 (move_sese_region_to_fn): Remove loops that are outlined into fn
21698 * tree-inline.c: Include cfgloop.h.
21699 (initialize_cfun): Do not drop PROP_loops.
21700 (copy_loops): New function.
21701 (copy_cfg_body): Copy loop structure.
21702 (tree_function_versioning): Initialize destination loop tree.
21703 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
21704 (pass_parallelize_loops): Do IL verification.
21705 * loop-init.c (loop_optimizer_init): Fixup loops if required.
21706 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
21707 the CFG make sure we fixup loops as well.
21708 * tree-ssa-tail-merge.c: Include cfgloop.h.
21709 (replace_block_by): When merging loop latches mark loops for fixup.
21710 * lto-streamer-out.c (output_struct_function_base): Drop
21711 PROP_loops for now.
21712 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
21713 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
21714 * ipa-split.c: Include cfgloop.h.
21715 (split_function): Add the new return block to the loop tree root.
21716 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
21717 whether we have removed the forwarder block.
21718 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
21719 * cfgloop.h (place_new_loop): Declare.
21720 * cfgloopmanip.c (place_new_loop): Export.
21721 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
21722 (tree-switch-conversion.o): Likewise.
21723 (tree-complex.o): Likewise.
21724 (tree-inline.o): Likewise.
21725 (tree-ssa-tailmerge.o): Likewise.
21726 (ipa-split.o): Likewise.
21727 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
21728 (tree-ssa-copy.o): Likewise.
21729 * tree-switch-conversion.c: Include cfgloop.h
21730 (process_switch): If we emit a bit-test cascade, schedule loops
21732 * tree-complex.c: Include cfgloop.h.
21733 (expand_complex_div_wide): Properly add new basic-blocks to loops.
21734 * asan.c: Include cfgloop.h.
21735 (create_cond_insert_point): Properly add new basic-blocks to
21736 loops, schedule loop fixup.
21737 * cfgloop.c (verify_loop_structure): Check that looks are not
21739 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
21741 (expand_omp_for_generic): Likewise.
21742 (expand_omp_sections): Likewise.
21743 (expand_omp_atomic_pipeline): Schedule loops for fixup.
21744 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
21745 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
21746 is initialized, not when loops are present.
21747 * tree-parloops.c (parallelize_loops): Remove checking here.
21748 * passes.c (init_optimization_passes): Schedule a copy-propagation
21749 pass before complete unrolling of inner loops.
21751 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21753 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
21754 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
21755 (colorize_init): Add argument to _WIN32 version.
21756 * toplev.c: Include diagnostic-color.h.
21757 (process_options): Default to -fdiagnostics-color=auto if
21758 GCC_COLORS env var is in the environment.
21759 * common.opt (fdiagnostics-color=): Add Var and Init.
21760 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
21761 env var is in the environment, the default is auto rather than never.
21763 * diagnostic.h (file_name_as_prefix): Add context argument.
21764 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
21765 the string as locus.
21766 * langhooks.c (lhd_print_error_function): Adjust caller.
21768 2013-04-25 Lawrence Crowl <crowl@google.com>
21770 * var-tracking.c (shared_hash_def::htab):
21771 Change type to hash_table. Update dependent calls and types.
21773 2013-04-25 Lawrence Crowl <crowl@google.com>
21775 * Makefile.in: Update as needed below.
21777 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
21778 Move declaration to after the type's method definitons.
21780 * attribs.c (htab_t scoped_attributes::attribute_hash):
21781 Change type to hash_table. Update dependent calls and types.
21783 * bitmap.c (htab_t bitmap_desc_hash):
21784 Change type to hash_table. Update dependent calls and types.
21786 * cselib.c (htab_t cselib_hash_table):
21787 Change type to hash_table. Update dependent calls and types.
21789 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
21790 (hash_string_slot_node): Move implementation into lto-streamer.h
21791 struct string_slot_hasher.
21792 (eq_string_slot_node): Likewise.
21794 * data-streamer-out.c: Update output_block::string_hash_table
21795 dependent calls and types.
21797 * dwarf2cfi.c (htab_t trace_index):
21798 Change type to hash_table. Update dependent calls and types.
21800 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
21801 Change type to hash_table. Update dependent calls and types.
21802 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
21803 (htab_t optimize_external_refs::map): Likewise.
21804 (htab_t output_comp_unit::extern_map): Likewise.
21805 (htab_t output_comdat_type_unit::extern_map): Likewise.
21806 (htab_t output_macinfo::macinfo_htab): Likewise.
21807 (htab_t optimize_location_lists::htab): Likewise.
21808 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
21810 * except.c (htab_t ehspec_hash_type):
21811 Change type to hash_table. Update dependent calls and types.
21812 (assign_filter_values::ttypes): Likewise.
21813 (assign_filter_values::ehspec): Likewise.
21814 (sjlj_assign_call_site_values::ar_hash): Likewise.
21815 (convert_to_eh_region_ranges::ar_hash): Likewise.
21817 * gcse.c (htab_t pre_ldst_table):
21818 Change type to hash_table. Update dependent calls and types.
21820 * ggc-common.c (htab_t saving_htab):
21821 Change type to hash_table. Update dependent calls and types.
21822 (htab_t loc_hash): Likewise.
21823 (htab_t ptr_hash): Likewise.
21824 (call_count): Rename ggc_call_count.
21825 (call_alloc): Rename ggc_call_alloc.
21826 (loc_descriptor): Rename make_loc_descriptor.
21827 (add_statistics): Rename ggc_add_statistics.
21829 * ggc-common.c (saving_htab):
21830 Change type to hash_table. Update dependent calls and types.
21832 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
21833 (push_gimplify_context): Likewise.
21834 (pop_gimplify_context): Likewise.
21835 (struct gimple_temp_hash_elt): Added.
21836 (struct gimplify_hasher): Likewise.
21837 (struct gimplify_ctx.temp_htab):
21838 Change type to hash_table. Update dependent calls and types.
21840 * gimple-fold.c: Include gimplify-ctx.h.
21842 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
21843 Change type to hash_table. Update dependent calls and types.
21844 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
21845 avoid potential global name collision.
21847 * gimplify.c: Include gimplify-ctx.h.
21848 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
21849 (htab_t gimplify_ctx::temp_htab):
21850 Update dependent calls and types for new type hash_table.
21851 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
21852 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
21854 * gimplify-ctx.h: New.
21855 (struct gimple_temp_hash_elt): Move from gimplify.c.
21856 (class gimplify_hasher): New.
21857 (struct gimplify_ctx): Move from gimple.h.
21858 (htab_t gimplify_ctx::temp_htab):
21859 Change type to hash_table. Update dependent calls and types.
21861 * graphite-clast-to-gimple.c: Include graphite-htab.h.
21862 (htab_t ivs_params::newivs_index):
21863 Change type to hash_table. Update dependent calls and types.
21864 (htab_t ivs_params::params_index): Likewise.
21865 (htab_t print_generated_program::params_index): Likewise.
21866 (htab_t gloog::newivs_index): Likewise.
21867 (htab_t gloog::params_index): Likewise.
21869 * graphite.c: Include graphite-htab.h.
21870 4htab_t graphite_transform_loops::bb_pbb_mapping):
21871 Change type to hash_table. Update dependent calls and types.
21873 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
21874 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
21875 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
21877 * graphite-dependences.c: Include graphite-htab.h.
21878 (loop_is_parallel_p): Change hash table type of parameter.
21880 * graphite-htab.h: New.
21881 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
21882 (extern find_pbb_via_hash): Move from graphite-poly.h.
21883 (extern loop_is_parallel_p): Move from graphite-poly.h.
21884 (extern get_loop_body_pbbs): Move from graphite-poly.h.
21886 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
21887 (extern loop_is_parallel_p): Move to graphite-htab.h.
21888 (extern get_loop_body_pbbs): Move to graphite-htab.h.
21890 * haifa-sched.c (htab_t delay_htab):
21891 Change type to hash_table. Update dependent calls and types.
21892 (htab_t delay_htab_i2): Likewise.
21894 * ira-color.c (htab_t allocno_hard_regs_htab):
21895 Change type to hash_table. Update dependent calls and types.
21897 * ira-costs.c (htab_t cost_classes_htab):
21898 Change type to hash_table. Update dependent calls and types.
21900 * loop-invariant.c (htab_t merge_identical_invariants::eq):
21901 Change type to hash_table. Update dependent calls and types.
21903 * loop-iv.c (htab_t bivs):
21904 Change type to hash_table. Update dependent calls and types.
21906 * loop-unroll.c (htab_t opt_info::insns_to_split):
21907 Change type to hash_table. Update dependent calls and types.
21908 (htab_t opt_info::insns_with_var_to_expand): Likewise.
21910 * lto-streamer.h (struct string_slot): Move from data-streamer.h
21911 (struct string_slot_hasher): New.
21912 (htab_t output_block::string_hash_table):
21913 Change type to hash_table. Update dependent calls and types.
21915 * lto-streamer-in.c (freeing_string_slot_hasher): New.
21916 (htab_t file_name_hash_table):
21917 Change type to hash_table. Update dependent calls and types.
21919 * lto-streamer-out.c: Update output_block::string_hash_table dependent
21922 * lto-streamer.c (htab_t tree_htab):
21923 Change type to hash_table. Update dependent calls and types.
21925 * omp-low.c: Include gimplify-ctx.h.
21927 * passes.c (htab_t name_to_pass_map):
21928 Change type to hash_table. Update dependent calls and types.
21929 (pass_traverse): Rename to passes_pass_traverse.
21931 * plugin.c (htab_t event_tab):
21932 Change type to hash_table. Update dependent calls and types.
21934 * postreload-gcse.c (htab_t expr_table):
21935 Change type to hash_table. Update dependent calls and types.
21936 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
21938 * sese.c (debug_rename_map_1): Make extern.
21939 (htab_t copy_bb_and_scalar_dependences::rename_map):
21940 Change type to hash_table. Update dependent calls and types.
21942 * sese.h (extern debug_rename_map): Move to .c file.
21944 * store-motion.c (htab_t store_motion_mems_table):
21945 Change type to hash_table. Update dependent calls and types.
21947 * trans-mem.c (htab_t tm_new_mem_hash):
21948 Change type to hash_table. Update dependent calls and types.
21950 * tree-browser.c (htab_t TB_up_ht):
21951 Change type to hash_table. Update dependent calls and types.
21953 * tree-cfg.c (htab_t discriminator_per_locus):
21954 Change type to hash_table. Update dependent calls and types.
21956 * tree-complex.c: Include tree-hasher.h
21957 (htab_t complex_variable_components):
21958 Change type to hash_table. Update dependent calls and types.
21960 * tree-eh.c (htab_t finally_tree):
21961 Change type to hash_table. Update dependent calls and types.
21963 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
21964 struct int_tree_hasher.
21965 (extern int_tree_map_eq): Likewise.
21966 (uid_decl_map_hash): Removed.
21967 (extern decl_tree_map_eq): Likewise.
21969 * tree-hasher.h: New.
21970 (struct int_tree_hasher): New.
21971 (typedef int_tree_htab_type): New.
21973 * tree-inline.c: Include gimplify-ctx.h.
21975 * tree-mudflap.c: Include gimplify-ctx.h.
21977 * tree-parloops.c: Include tree-hasher.h.
21978 (htab_t eliminate_local_variables_stmt::decl_address):
21979 Change type to hash_table. Update dependent calls and types.
21980 (htab_t separate_decls_in_region::decl_copies): Likewise.
21982 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
21983 Change type to hash_table. Update dependent calls and types.
21985 * tree-sra.c (candidates):
21986 Change type to hash_table. Update dependent calls and types.
21988 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
21990 (int_tree_map_hash): Likewise.
21992 * tree-ssa-dom.c (htab_t avail_exprs):
21993 Change type to hash_table. Update dependent calls and types.
21995 * tree-ssa-live.c (var_map_base_init::tree_to_index):
21996 Change type to hash_table. Update dependent calls and types.
21998 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
21999 Change type to hash_table. Update dependent calls and types.
22001 * tree-ssa-phiopt.c (seen_ssa_names):
22002 Change type to hash_table. Update dependent calls and types.
22004 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
22005 Change type to hash_table. Update dependent calls and types.
22007 * tree-ssa-uncprop.c (equiv):
22008 Change type to hash_table. Update dependent calls and types.
22010 2013-04-25 Jakub Jelinek <jakub@redhat.com>
22012 PR rtl-optimization/57003
22013 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
22014 call note_stores with kill_clobbered_value callback again after
22015 killing regs_invalidated_by_call.
22017 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22019 * config/aarch64/aarch64-simd.md
22020 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
22021 (aarch64_simd_bsl<mode>): Likewise.
22022 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
22024 2013-04-25 Marek Polacek <polacek@redhat.com>
22026 PR tree-optimization/57066
22027 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
22029 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22031 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
22033 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22035 * config/aarch64/aarch64-builtins.c
22036 (aarch64_fold_builtin): New.
22037 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
22038 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
22039 * config/aarch64/aarch64-simd-builtins.def (abs): New.
22040 * config/aarch64/arm_neon.h
22041 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
22043 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22044 Tejas Belagod <tejas.belagod@arm.com>
22046 * config/aarch64/aarch64-builtins.c
22047 (aarch64_gimple_fold_builtin): New.
22048 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
22049 * config/aarch64/aarch64-simd-builtins.def (addv): New.
22050 * config/aarch64/aarch64-simd.md (addpv4sf): New.
22051 (addvv4sf): Update.
22052 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
22054 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22056 * config/aarch64/aarch64.md
22057 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
22059 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22061 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
22062 (*ngcsi_uxtw): New pattern.
22064 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22065 Julian Brown <julian@codesourcery.com>
22067 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
22068 (TB_DREG): Add T_V4HF.
22069 (v4hf_UP): New macro.
22070 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22071 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22072 Handle initialisation of V4HF. Adjust initialisation of reinterpret
22074 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22075 (arm_vector_mode_supported_p): Handle V4HF.
22076 (arm_mangle_map): Handle V4HFmode.
22077 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
22078 * config/arm/arm_neon_builtins.def: Add entries for
22079 vcvtv4hfv4sf, vcvtv4sfv4hf.
22080 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
22081 (neon_vcvtv4hfv4sf): Likewise.
22082 * config/arm/neon-gen.ml: Handle half-precision floating point
22084 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
22085 * config/arm/arm_neon.h: Regenerate.
22086 * config/arm/neon.ml (type elts): Add F16.
22087 (type vectype): Add T_float16x4, T_floatHF.
22088 (type vecmode): Add V4HF.
22089 (type features): Add Requires_FP_bit feature.
22090 (elt_width): Handle F16.
22091 (elt_class): Likewise.
22092 (elt_of_class_width): Likewise.
22093 (mode_of_elt): Refactor.
22094 (type_for_elt): Handle F16, fix error messages.
22095 (vectype_size): Handle T_float16x4.
22096 (vcvt_sh): New function.
22097 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
22098 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
22099 (string_of_mode): Handle V4HF.
22100 * doc/arm-neon-intrinsics.texi: Regenerate.
22102 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22104 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
22105 format specifier in 'X' case.
22107 2013-04-25 Alan Modra <amodra@gmail.com>
22110 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
22111 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
22112 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
22113 Repeat for many other rotate/shift and mask patterns using subregs.
22114 Name lshiftrt insns.
22115 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
22116 on WORDS_BIG_ENDIAN.
22118 2013-04-25 Alan Modra <amodra@gmail.com>
22120 * config.gcc: Support little-endian powerpc-linux targets.
22121 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
22122 (LINK_OS_LINUX_SPEC): Define.
22123 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
22124 Preserve MASK_LITTLE_ENDIAN.
22125 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
22126 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
22127 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
22128 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
22129 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
22130 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
22131 Correct fp word order for little-endian. Don't shift toc entries
22132 smaller than a word for little-endian.
22133 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
22134 (bswapdi2 splits): Correct low-part subreg for little-endian.
22135 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
22136 low/high where such is correct only for be.
22137 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
22138 little-endian for -mcall-aixdesc.
22140 2013-04-25 Alan Modra <amodra@gmail.com>
22142 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
22143 replace_equiv_address_nv.
22145 2013-04-25 Alan Modra <amodra@gmail.com>
22147 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
22149 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22152 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22153 * rtl.h (struct rtx_def): ...
22155 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22157 PR rtl-optimizations/57046
22158 * lra-constraints (split_reg): Set up lra_risky_transformations_p
22159 for multi-reg splits.
22161 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
22163 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
22165 2013-04-24 Sterling Augustine <saugustine@google.com>
22167 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
22168 (comp_dir_string, debug_str_dwo_section): New.
22169 (DEBUG_STR_DWO_SECTION): Rename to ...
22170 (DEBUG_DWO_STR_SECTION): ... this.
22171 (DEBUG_NORM_STR_SECTION): Delete.
22172 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
22173 (DEBUG_STR_DWO_SECTION_FLAGS): New.
22174 (find_AT_string): Move most logic to ...
22175 (find_AT_string_in_table): ... here. New.
22176 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
22177 add_skeleton_AT_string. Delete logic.
22178 (output_skeleton_debug_sections): Remove call to
22179 add_top_level_skeleton_die_attrs.
22180 (add_comp_dir_attribute): Move logic to comp_dir_string.
22181 (dwarf2out_init): Initialize debug_str_dwo_section.
22182 (output_indirect_string): Call find_string_form.
22183 (output_indirect_strings): Rewrite.
22184 (prune_unused_types): Empty skeleton_debug_str_hash.
22185 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
22186 (dwarf2out_finish): Call output_indirect_strings.
22188 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22190 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
22192 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22194 * rtl.h (struct rtx_def): Add comment for field jump.
22195 (LRA_SUBREG_P): New macro.
22196 * recog.c (register_operand): Check LRA_SUBREG_P.
22197 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22199 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22200 (lra_final_code_change): Skip subreg change for operators.
22201 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22202 if there are no operand changes.
22203 * lra-constraints.c (curr_insn_set): New.
22204 (match_reload): Set LRA_SUBREG_P.
22205 (emit_spill_move): Ditto.
22206 (check_and_process_move): Use curr_insn_set. Process only single
22207 set insns. Don't initialize sec_mem_p and change_p.
22208 (simplify_operand_subreg): Use LRA_SUBREG_P.
22209 (reg_in_class_p): New function.
22210 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22211 of #ifdef. Add code to remove cycling.
22212 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22213 non-null disp. Reload inner instead of disp when base and index
22214 are null. Try to put lo_sum into register.
22215 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22216 (check_and_process_move): Move code for move cost check to
22217 simple_move_p. Remove equiv_substitution.
22218 (simple_move_p): New function.
22219 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22220 curr_insn_set. Call check_and_process_move only for single set
22221 insns. Use the new function. Move call of check_and_process_move
22222 after operand equiv substitution and address process.
22224 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
22226 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
22227 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
22228 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
22230 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22232 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
22234 2013-04-24 Marek Polacek <polacek@redhat.com>
22236 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
22237 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
22238 (select_loops_exit_conditions): Likewise.
22239 (number_of_iterations_for_all_loops): Likewise.
22240 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
22241 (scev_analysis): Likewise.
22243 2013-04-02 Catherine Moore <clm@codesourcery.com>
22244 Chao-ying Fu <fu@mips.com>
22246 * config/mips/micromips.md (jraddiusp): New pattern.
22247 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
22248 instruction if possible.
22250 2013-04-24 Alan Modra <amodra@gmail.com>
22252 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
22254 2013-04-24 Julian Brown <julian@codesourcery.com>
22255 Chung-Lin Tang <cltang@codesourcery.com>
22257 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
22258 dependency behavior in enumeration type DIE generation. Add TODO note
22259 to comments about future DW_FORM_sdata/udata re-work of related code.
22261 2013-04-23 Lawrence Crowl <crowl@google.com>
22263 * Makefile.in: Update as needed below.
22265 * hash-table.h (class hash_table):
22266 Correct many methods with parameter types compare_type to the correct
22267 value_type. (Correct code was unlikely to notice the change.)
22268 (hash_table::elements_with_deleted) New.
22269 (class hashtable::iterator): New.
22270 (hashtable::begin()): New.
22271 (hashtable::end()): New.
22272 (FOR_EACH_HASH_TABLE_ELEMENT): New.
22274 * statistics.c (statistics_hashes):
22275 Change type to hash_table. Update dependent calls and types.
22277 * tree-into-ssa.c (var_infos):
22278 Change type to hash_table. Update dependent calls and types.
22280 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
22281 Change type to hash_table. Update dependent calls and types.
22283 * tree-ssa-loop-im.c (struct mem_ref.refs):
22284 Change type to hash_table. Update dependent calls and types.
22286 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
22287 Change type to hash_table. Update dependent calls and types.
22289 * tree-ssa-sccvn.c (vn_tables_s::nary):
22290 Change type to hash_table. Update dependent calls and types.
22291 (vn_tables_s::phis): Likewise.
22292 (vn_tables_s::references): Likewise.
22294 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
22295 (vn_reference_eq): Update parameter and return types.
22297 * tree-ssa-structalias.c (pointer_equiv_class_table):
22298 Change type to hash_table. Update dependent calls and types.
22299 (location_equiv_class_table): Likewise.
22301 * tree-vect-data-refs.c: Consequential changes for making
22302 peeling a hash_table.
22304 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
22305 (destroy_loop_vec_info): Dependent hash_table update.
22307 * tree-vectorizer.h (peeling_htab):
22308 Change type to hash_table. Update dependent calls and types.
22310 2013-04-23 Shiva Chen <shiva0217@gmail.com>
22312 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
22313 to check the register content is equal or not.
22314 * lra-constraints.c (match_reload): Use lra_assign_reg_val
22315 to assign register content record.
22316 * lra-eliminations.c (update_reg_eliminate): Use
22317 lra_update_reg_val_offset to update register content offset.
22318 * lra-int.h (struct lra_reg): Add offset member.
22319 (lra_reg_val_equal_p): New static inline function.
22320 (lra_update_reg_val_offset): New static inline function.
22321 (lra_assign_reg_val): New static inline function.
22322 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
22323 to assign register content record.
22324 (initialize_lra_reg_info_element): Initial offset to zero.
22326 2013-04-23 Catherine Moore <clm@codesourcery.com>
22328 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
22329 operands. Record compression.
22331 2013-04-23 Xinliang David Li <davidxl@google.com>
22333 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
22335 2013-04-23 Richard Biener <rguenther@suse.de>
22337 PR middle-end/57036
22338 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
22339 parameter, only add abnormal goto edges from the copied body
22340 if the call could perform abnormal gotos.
22341 (copy_cfg_body): Adjust.
22343 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
22345 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
22347 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
22349 * coretypes.h (gimple_stmt_iterator): Add struct to make
22352 2013-04-23 Richard Biener <rguenther@suse.de>
22354 PR tree-optimization/57026
22355 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
22356 from SSA names occuring in abnormal PHI nodes.
22358 2013-04-22 Andi Kleen <ak@linux.intel.com>
22360 * lto/lto.c (print_lto_report_1): Fix LTO report names.
22362 2013-04-22 Andi Kleen <ak@linux.intel.com>
22364 * lto/lto.c (print_lto_report_1): Declare early.
22365 (read_cgraph_and_symbols): Call print_lto_report_1 early.
22367 2013-04-22 Andi Kleen <ak@linux.intel.com>
22369 * common.opt (-flto-report-wpa): Add.
22370 * doc/invoke.texi (-flto-report-wpa): Add.
22371 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
22374 2013-04-22 Xinliang David Li <davidxl@google.com>
22376 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
22377 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
22378 * Makefile.in: New dependency
22380 David Daney <ddaney.cavm@gmail.com>
22382 * configure.ac (gcc_cv_as_micromips_support): Use the
22383 --fatal-warnings option.
22384 * configure: Regenerate.
22386 2013-04-22 Marek Polacek <polacek@redhat.com>
22389 * tsan.c (instrument_expr): Don't instrument expression
22390 in case its size is zero.
22392 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
22396 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22398 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22400 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22402 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
22403 (gimple_stmt_iterator): New typedef.
22404 * gimple.h (gimple_stmt_iterator): Rename to...
22405 (gimple_stmt_iterator_d): ... This.
22406 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
22407 trees be valid for GIMPLE and GENERIC.
22408 (TARGET_GIMPLE_FOLD_BUILTIN): New.
22409 * gimple-fold.c (gimple_fold_call): Call target hook
22410 gimple_fold_builtin.
22411 * hooks.c (hook_bool_gsiptr_false): New.
22412 * hooks.h (hook_bool_gsiptr_false): New.
22413 * target.def (fold_stmt): New.
22414 * doc/tm.texi: Regenerate.
22416 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
22419 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
22420 a set sp if no stack realignment.
22422 2013-04-22 Nick Clifton <nickc@redhat.com>
22424 * config.gcc (tilegx-linux): Extend extra_objs rather than
22426 (tilepro-linux): Likewise.
22428 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22430 * config/aarch64/aarch64-builtins.c
22432 (CF0, CF1, CF2, CF3, CF4, CF10): New.
22433 (VAR<1-12>): Add MAP parameter.
22434 (BUILTIN_*): Likewise.
22435 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
22436 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
22437 (aarch64_ushl_n<mode>): Likewise.
22438 (aarch64_sshr_n<mode>): Likewise.
22439 (aarch64_ushr_n<mode>): Likewise.
22440 (aarch64_<maxmin><mode>): Likewise.
22441 (aarch64_sqrt<mode>): Likewise.
22442 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
22443 (vshr<q>_n_*): Likewise.
22445 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22447 * config/aarch64/aarch64-builtins.c
22448 (aarch64_simd_builtin_type_mode): Handle SF types.
22450 (BUILTIN_GPF): Define.
22451 (aarch64_init_simd_builtins): Handle SF types.
22452 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
22453 (frecps): Likewise.
22454 (frecpx): Likewise.
22455 * config/aarch64/aarch64-simd.md
22456 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
22457 (aarch64_frecpe<mode>): New.
22458 (aarch64_frecps<mode>): Likewise.
22459 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
22460 (v8type): Add frecp<esx>.
22461 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
22462 (aarch64_frecps<mode>): Likewise.
22463 * config/aarch64/iterators.md (FRECP): New.
22464 (frecp_suffix): Likewise.
22465 * config/aarch64/arm_neon.h
22466 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
22468 2013-04-22 Christian Bruel <christian.bruel@st.com>
22471 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
22472 (REG_CLASS_NAMES): Idem.
22473 (REG_CLASS_CONTENTS): Idem.
22474 (REGCLASS_HAS_FP_REG): Idem.
22475 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
22476 (sh_conditional_register_usage): Idem.
22478 2013-04-21 Jeff Law <law@redhat.com>
22480 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
22481 (ssa_forward_propagate_and_combine): Use it.
22483 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22485 * lra.c: Update the flow chart diagram.
22487 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22489 PR rtl-optimization/56847
22490 * lra-constraints.c (process_alt_operands): Discourage alternative
22491 with non-matche doffsettable memory constraint fro memory with
22494 2013-04-19 Richard Biener <rguenther@suse.de>
22496 PR tree-optimization/56982
22497 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
22499 * gimplify.c (gimplify_call_expr): Notice special calls.
22500 (gimplify_modify_expr): Likewise.
22501 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
22502 abnormal control flow receivers.
22503 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
22504 in the same way as cfun->has_nonlocal_labels.
22505 (gimple_purge_dead_abnormal_call_edges): Likewise.
22506 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
22507 receivers start a basic-block.
22509 2013-04-19 Richard Biener <rguenther@suse.de>
22511 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
22513 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
22514 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
22515 (SLP_TREE_LOAD_PERMUTATION): Add.
22516 (vect_transform_slp_perm_load): Adjust prototype.
22517 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
22518 (vect_free_slp_instance): Likewise.
22519 (vect_create_new_slp_node): Likewise.
22520 (vect_supported_slp_permutation_p): Remove.
22521 (vect_slp_rearrange_stmts): Adjust.
22522 (vect_supported_load_permutation_p): Likewise. Inline
22523 vect_supported_slp_permutation_p here.
22524 (vect_analyze_slp_instance): Compute load permutations per
22525 slp node instead of per instance.
22526 (vect_get_slp_defs): Adjust.
22527 (vect_transform_slp_perm_load): Likewise.
22528 (vect_schedule_slp_instance): Remove redundant code.
22529 (vect_schedule_slp): Remove hack for PR56270, add it ...
22530 * tree-vect-stmts.c (vectorizable_load): ... here, do not
22531 CSE loads for SLP. Adjust.
22533 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22535 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
22536 spelling in two comments.
22538 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22541 * config/arm/arm.c (load_multiple_sequence): Require SP
22542 as base register for loads if SP is in the register list.
22544 2013-04-19 Martin Jambor <mjambor@suse.cz>
22546 PR tree-optimization/56718
22547 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
22548 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
22549 and made public. Adjusted all callers.
22550 (ipa_intraprocedural_devirtualization): New function.
22551 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
22552 (ipa_intraprocedural_devirtualization): Likewise.
22553 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
22555 2013-04-19 Richard Biener <rguenther@suse.de>
22557 PR tree-optimization/57000
22558 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
22560 2013-04-19 Terry Guo <terry.guo@arm.com>
22562 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
22564 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
22565 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
22566 (cortex_m4_fmacs): Use new reservations.
22567 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
22569 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
22571 PR rtl-optimization/56999
22572 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
22574 (lra_coalesce): Remove split_origin_bitmap and related code.
22575 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
22576 ranges if necessary.
22578 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
22580 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
22582 (ix86_expand_call): Remove clobbered_registers array and use
22583 x86_64_ms_sysv_extra_clobbered_registers instead.
22584 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
22586 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
22588 * config/i386/i386.md (*call_rex64_ms_sysv): Use
22589 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
22590 (*call_value_rex64_ms_sysv): Ditto.
22592 2013-04-18 Cary Coutant <ccoutant@google.com>
22594 * dwarf2out.c (output_pubnames): Check die_perennial_p of
22595 parent instead of die_mark.
22597 2013-04-18 Diego Novillo <dnovillo@google.com>
22599 * gimple.c (create_gimple_tmp): New.
22600 (get_expr_type): New.
22601 (build_assign): New.
22602 (build_type_cast): New.
22603 * gimple.h (enum ssa_mode): Define.
22604 (gimple_seq_set_location): New.
22605 * asan.c (build_check_stmt): Change some gimple_build_* calls
22606 to use build_assign and build_type_cast.
22608 2013-04-18 Richard Biener <rguenther@suse.de>
22610 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
22611 handle negative step. Remove redundant checks.
22612 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
22613 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
22614 for negative step and grouped loads fail to vectorize.
22616 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22618 * emit-rtl.c (reset_insn_used_flags): New function.
22619 (reset_all_used_flags): Use it.
22620 (verify_insn_sharing): New function.
22621 (verify_rtl_sharing): Fix verification for SEQUENCEs.
22623 2013-04-18 Jakub Jelinek <jakub@redhat.com>
22625 PR tree-optimization/56984
22626 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
22627 and (x >> M) >= N don't register any assertion if N << M is the
22630 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22632 * lower-subreg.c (resolve_simple_move): If called self-recursive,
22633 do not delete_insn insns that have not yet been emitted, only
22634 unlink them with remove_insn.
22635 * df-scan.c (df_insn_delete): Revert r197492.
22637 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
22639 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
22640 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
22642 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22644 * config/arm/arm.md (movsicc_insn): Convert define_insn into
22645 define_insn_and_split.
22646 (and_scc,ior_scc,negscc): Likewise.
22647 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
22649 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22651 * config/arm/arm.c (use_return_insn): Return 0 for targets that
22652 can benefit from using a sequence of LDRD instructions in epilogue
22653 instead of a single LDM instruction.
22655 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
22658 * doc/extend.texi: Fix typo.
22660 2013-04-17 Richard Biener <rguenther@suse.de>
22662 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
22663 (vect_build_slp_tree): ... here.
22664 (vect_build_slp_tree_1): Compute which stmts of the SLP group
22665 match. Remove special-casing of mismatched complex loads.
22666 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
22667 re-try the match with swapped commutative operands.
22668 (vect_supported_load_permutation_p): Remove special-casing of
22669 mismatched complex loads.
22670 (vect_analyze_slp_instance): Adjust.
22672 2013-04-17 Richard Biener <rguenther@suse.de>
22674 PR rtl-optimization/56921
22675 * cfgloop.h (struct loop): Add simple_loop_desc member.
22676 (struct niter_desc): Mark with GTY(()).
22677 (simple_loop_desc): Do not use aux field but simple_loop_desc.
22678 * loop-iv.c (get_simple_loop_desc): Likewise.
22679 (free_simple_loop_desc): Likewise.
22682 2013-04-16 Richard Biener <rguenther@suse.de>
22684 PR rtl-optimization/56921
22685 * loop-init.c (pass_rtl_move_loop_invariants): Add
22686 TODO_do_not_ggc_collect to todo_flags_finish.
22687 (pass_rtl_unswitch): Same.
22688 (pass_rtl_unroll_and_peel_loops): Same.
22689 (pass_rtl_doloop): Same.
22691 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
22693 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
22694 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
22695 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
22697 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
22698 * tree-streamer.c (record_common_node): Adjust reference in comment.
22700 2013-04-17 Terry Guo <terry.guo@arm.com>
22702 * config/arm/cortex-m4.md: Add a new bypass.
22704 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22706 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
22708 (*subs_<optab><mode>_multp2): New pattern.
22709 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22710 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22712 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22714 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
22715 (*subs_mul_imm_<mode>): New pattern.
22717 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
22720 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
22721 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
22722 (vsx_movti_32bit): Change j->wa to O->wa.
22724 2013-04-16 Richard Biener <rguenther@suse.de>
22726 PR rtl-optimization/56921
22727 * loop-init.c (pass_rtl_move_loop_invariants): Add
22728 TODO_do_not_ggc_collect to todo_flags_finish.
22729 (pass_rtl_unswitch): Same.
22730 (pass_rtl_unroll_and_peel_loops): Same.
22731 (pass_rtl_doloop): Same.
22733 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
22735 * config/arm/arm.c (emit_multi_reg_push): New declaration
22736 for an existing function.
22737 (arm_emit_strd_push): New function.
22738 (arm_expand_prologue): Used here.
22739 (arm_emit_ldrd_pop): New function.
22740 (arm_expand_epilogue): Used here.
22741 (arm_get_frame_offsets): Update condition.
22742 (arm_emit_multi_reg_pop): Add a special case for load of a single
22743 register with writeback.
22745 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
22747 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
22750 2013-04-16 Richard Biener <rguenther@suse.de>
22752 PR tree-optimization/56756
22753 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
22754 (first_mem_ref_loc): New.
22755 (execute_sm): Place the load temporarily before a previous
22756 access instead of in the latch edge to ensure its SSA dependencies
22757 are defined at points dominating the load.
22759 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
22761 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
22762 correct fix by moving header and footer insn to the footer of
22763 the merged basic block. Clear BB_END of the merged-away block.
22765 PR middle-end/43631
22766 * emit-rtl.c (make_note_raw): New function.
22767 (link_insn_into_chain): New static inline function.
22768 (add_insn): Use it.
22769 (add_insn_before, add_insn_after): Factor insn chain linking code...
22770 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
22771 using link_insn_into_chain.
22772 (note_outside_basic_block_p): New helper function for emit_note_after
22773 and emit_note_before.
22774 (emit_note_after): Use nobb variant of add_insn_after if the note
22775 should not be contained in a basic block.
22776 (emit_note_before): Use nobb variant of add_insn_before if the note
22777 should not be contained in a basic block.
22778 (emit_note_copy): Use make_note_raw.
22779 (emit_note): Likewise.
22780 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
22781 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
22782 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
22783 the moved barrier the tail of the basic block it follows.
22784 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
22786 2013-04-15 Jakub Jelinek <jakub@redhat.com>
22788 PR tree-optimization/56962
22789 * gimple-ssa-strength-reduction.c (record_increment): Only set
22790 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
22791 either rhs1 or rhs2 is equal to c->base_expr.
22793 2013-04-15 Richard Biener <rguenther@suse.de>
22795 PR tree-optimization/56933
22796 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
22798 (GROUP_READ_WRITE_DEPENDENCE): Remove.
22799 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
22800 * tree-vect-data-refs.c (vect_analyze_group_access): Move
22801 dependence check ...
22802 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
22804 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
22805 GROUP_READ_WRITE_DEPENDENCE.
22807 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22809 * emit-rtl.c (reset_all_used_flags): New function.
22810 (verify_rtl_sharing): Call reset_all_used_flags before and after
22811 performing the checks.
22813 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22815 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
22816 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
22817 * config/arm/constraints.md (De): New constraint.
22818 * config/arm/neon.md (anddi3_neon): Delete.
22819 (neon_vand<mode>): Expand to standard anddi3 pattern.
22820 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
22821 Move earlier in the file.
22822 (neon_inv_logic_op2): Likewise.
22823 (arm_anddi_operand_neon): New predicate.
22825 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22827 * configure.ac (gcc_cv_ld_as_needed): Set
22828 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
22829 Use -z ignore, -z record on *-*-solaris2*.
22830 (HAVE_LD_AS_NEEDED): Update comment.
22831 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
22832 * configure: Regenerate.
22833 * config.in: Regenerate.
22834 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
22835 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
22836 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
22837 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
22838 equivalents. Fix markup.
22839 * doc/tm.texi: Regenerate.
22841 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
22843 * config/i386/i386.opt: New option mstack-protector-guard=.
22844 * config/i386/i386-opts.h: Add enum stack_protector_guard.
22845 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
22846 TARGET_SSP_TLS_GUARD.
22847 * config/i386/i386.c (ix86_option_override_internal): Set
22848 ix86_stack_protector_guard.
22849 * config/i386/i386.md (stack_protect_set): Enable for
22850 TARGET_SSP_TLS_GUARD only.
22851 (stack_protect_set_<mode>): Ditto.
22852 (stack_protect_test): Ditto.
22853 (stack_protect_test_<mode>): Ditto.
22854 * doc/invoke.texi (i386 Option): Document.
22856 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
22859 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
22860 (S_MODES): Set H_MODE bit.
22861 (SF_MODES): Set only S_MODE and SF_MODE bits.
22862 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
22863 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
22864 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
22865 <MODE_FLOAT>: Likewise.
22867 2013-04-15 Joey Ye <joey.ye@arm.com>
22869 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
22871 2013-04-15 Joey Ye <joey.ye@arm.com>
22873 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
22875 (thumb_far_jump_used_p): Count instruction size and set
22878 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
22880 * reorg.c (fill_simple_delay_slots): Reindent block of code.
22881 * resource.c (mark_target_live_regs): Reformat conditional block.
22883 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22885 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
22886 notes, they are emitted only just before final.
22887 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
22889 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22891 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
22892 * cfgrtl.c (delete_insn): Call it here instead.
22893 * lra-spills.c (lra_final_code_change): Use delete_insn.
22894 * haifa-sched.c (sched_remove_insn): Likewise.
22895 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
22896 returning to the nop pool.
22897 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
22898 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
22900 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22902 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
22903 * doc/tm.texi: Regenerated.
22905 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
22907 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
22910 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22912 * df-core.c (df_find_def): Compare register numbers.
22913 (df_find_use): Likewise.
22915 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
22918 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
22919 lra_in_progress for return.
22921 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22923 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
22924 define_insn into define_insn_and_split and emit movsicc patterns.
22926 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22928 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
22930 2013-04-12 Richard Biener <rguenther@suse.de>
22932 * tree-pass.h (TODO_do_not_ggc_collect): New.
22933 * passes.c (execute_one_ipa_transform_pass): Honor
22934 TODO_do_not_ggc_collect.
22935 (execute_one_pass): Likewise.
22938 2013-04-10 Richard Biener <rguenther@suse.de>
22940 * passes.c (init_optimization_passes): Remove reload pass.
22941 * ira.c (do_reload): Merge into ...
22943 (rest_of_handle_reload): Remove.
22944 (pass_reload): Likewise.
22945 * config/i386/i386.c (ix86_option_override): Refer to ira instead
22946 of reload for vzeroupper pass placement.
22948 2013-04-12 Jakub Jelinek <jakub@redhat.com>
22950 PR tree-optimization/56918
22951 PR tree-optimization/56920
22952 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
22953 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
22954 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
22955 use wide_mul_with_sign method.
22957 2013-04-12 Richard Biener <rguenther@suse.de>
22959 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
22960 not be considered a gimple constant.
22962 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
22964 * fold-const.c (const_binop): Handle vector shifts by a scalar.
22965 (fold_binary_loc): Call const_binop also for mixed vector-scalar
22968 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
22969 Jakub Jelinek <jakub@redhat.com>
22971 * opts.c: Include diagnostic-color.h.
22972 (common_handle_option): Handle OPT_fdiagnostics_color_.
22973 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
22974 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
22975 (diagnostic-color.o): New.
22976 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
22977 (diagnostic_color_rule): New enum.
22978 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
22979 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
22980 the location string.
22981 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
22982 either NULL, or color kind.
22983 * diagnostic-color.c: New file.
22984 * diagnostic-color.h: New file.
22985 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
22987 * doc/invoke.texi (-fdiagnostics-color): Document.
22988 * pretty-print.h (pp_show_color): Define.
22989 (struct pretty_print_info): Add show_color field.
22990 * diagnostic.c: Include diagnostic-color.h.
22991 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
22992 macros. Colorize error:, warning: etc. strings and also the location
22994 (diagnostic_show_locus): Colorize the caret line.
22995 * pretty-print.c: Include diagnostic-color.h.
22996 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
22997 inside of %< %> quotes or quoted through q format modifier.
22999 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23001 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
23003 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23005 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
23006 code in CC_NZ mode.
23007 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
23010 2013-04-11 Marek Polacek <polacek@redhat.com>
23012 PR tree-optimization/48184
23013 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
23015 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
23017 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
23018 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
23019 (skip_simple_arithmetic): Tidy up.
23020 * tree.h (skip_simple_constant_arithmetic): Declare.
23022 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23024 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
23026 2013-04-11 Richard Biener <rguenther@suse.de>
23028 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23029 generate vector constants.
23031 2013-04-11 Richard Biener <rguenther@suse.de>
23033 PR tree-optimization/56878
23034 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
23035 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
23037 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
23038 Prefer to align the DR with the most invariant base address.
23040 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23042 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
23045 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
23047 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
23048 floating-point vector comparisons against 0.
23050 2013-04-11 Jakub Jelinek <jakub@redhat.com>
23052 PR tree-optimization/56899
23053 * fold-const.c (extract_muldiv_1): Apply distributive law
23054 only if TYPE_OVERFLOW_WRAPS (ctype).
23056 2013-04-11 Bin Cheng <bin.cheng@arm.com>
23059 * ira-costs.c (scan_one_insn): Check whether the source rtx of
23060 loading has side effect.
23062 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
23064 * config/sparc/sparc.c: Include tree-pass.h.
23065 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
23066 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
23067 head of file. Change return type. Split off gate function.
23068 (sparc_gate_work_around_errata): New function.
23069 (pass_work_around_errata): New pass definition.
23070 (insert_pass_work_around_errata) New pass insert definition to
23071 insert pass_work_around_errata just after delayed-branch scheduling.
23072 (sparc_option_override): Insert the pass.
23073 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
23075 2013-04-10 David S. Miller <davem@davemloft.net>
23077 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
23078 or -mcpu=hypersparc.
23080 * target.def (cstore_mode): New hook.
23081 * target.h: Include insn-codes.h
23082 * targhooks.c: Likewise.
23083 (default_cstore_mode): New function.
23084 * targhooks.h: Declare it.
23085 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
23086 * doc/tm.texi: Rebuild.
23087 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
23088 target hook, rather than inspecting the insn_data.
23089 * config/sparc/sparc.c (sparc_cstore_mode): New function.
23090 (TARGET_CSTORE_MODE): Redefine.
23091 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
23093 * config/sparc/predicates.md (cstore_result_operand): New special
23095 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
23096 Use it for operand 0.
23097 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
23098 (*snesi_special): Likewise.
23099 (*snesi_zero): Likewise.
23100 (*seqsi_zero): Likewise.
23101 (*sltu_insn): Likewise.
23102 (*sgeu_insn): Likewise.
23103 (*seqdi_special): Make operand 0 and comparison operation be of
23105 (*snedi_special): Likewise.
23106 (*snedi_special_vis3): Likewise.
23107 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
23108 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
23109 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
23110 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
23111 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
23112 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
23113 (*sltu_extend_sp64): Likewise.
23114 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
23115 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
23116 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
23117 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
23118 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
23120 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
23122 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
23123 (aarch64_start_file): Use the new function.
23125 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23126 Jason Merrill <jason@redhat.com>
23128 * common.opt: Add -gdwarf.
23129 * opts.c (common_handle_option): Handle it.
23130 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
23132 2013-04-10 Richard Biener <rguenther@suse.de>
23134 * passes.c (execute_todo): Do not call ggc_collect conditional here.
23135 (execute_one_ipa_transform_pass): But unconditionally here.
23136 (execute_one_pass): And here.
23137 (init_optimization_passes): Remove reload pass.
23138 * tree-pass.h (TODO_ggc_collect): Remove.
23139 (pass_reload): Likewise.
23140 * ira.c (do_reload): Merge into ...
23142 (rest_of_handle_reload): Remove.
23143 (pass_reload): Likewise.
23144 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23145 of reload for vzeroupper pass placement.
23146 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
23147 and todo_flags_finish of all passes.
23149 2013-04-10 Richard Biener <rguenther@suse.de>
23151 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
23152 first_const_oprnd field, rename first_def_type to first_op_type.
23153 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
23154 (vect_get_and_check_slp_defs): Always use the type of the
23155 operand. Allow mixed vect_external_def, vect_constant_def types.
23156 (vect_get_constant_vectors): Handle mixed vect_external_def,
23157 vect_constant_def types.
23159 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
23161 PR tree-optimization/55524
23162 * tree-ssa-math-opts.c
23163 (convert_mult_to_fma): Don't use an fms construct
23164 when we don't have an fms operation, but fnma, and it looks
23165 likely that we'll be able to use the latter.
23167 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
23169 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
23171 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
23172 inline fail caused by overwritable functions.
23174 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
23176 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
23177 unnecessary bits in the constant power of two case.
23179 2013-04-10 Richard Biener <rguenther@suse.de>
23181 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
23182 broken code swapping operands.
23183 (vect_build_slp_tree): Do not compute load permutations here.
23184 (vect_analyze_slp_instance): Compute load permutations here,
23185 after building the SLP tree.
23187 2013-04-09 Christian Bruel <christian.bruel@st.com>
23189 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
23190 of next/prev_real_insn.
23192 2013-04-09 Jan Hubicka <jh@suse.cz>
23194 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
23195 Drop aliased parameter.
23196 (function_and_variable_visibility): Do not handle alias pairs.
23197 * cgraph.c (varpool_externally_visible_p): Update prototype.
23198 * varpool.c (varpool_add_new_variable): Update.
23200 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23202 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
23204 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
23206 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
23208 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
23210 2013-04-09 Marek Polacek <polacek@redhat.com>
23212 PR tree-optimization/48762
23213 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
23215 2013-04-09 Richard Biener <rguenther@suse.de>
23217 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
23219 (vect_build_slp_tree): Likewise.
23220 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
23221 calculating the cost of a SLP instance.
23222 (vect_analyze_slp_instance): Use it from here, after building
23225 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23227 PR middle-end/56883
23228 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
23229 expand_omp_for_static_chunk): Use simple_p = true in
23230 force_gimple_operand_gsi calls when assigning to addressable decls.
23232 2013-04-09 Jeff Law <law@redhat.com>
23234 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
23235 when the boolean was created by converting a wider object which
23236 had a boolean range.
23238 2013-04-09 Richard Biener <rguenther@suse.de>
23240 * tree-vectorizer.h (slp_void_p): Remove.
23241 (slp_tree): Typedef before _slp_tree declaration.
23242 (struct _slp_tree): Use a vector of slp_tree as children.
23243 (vect_get_place_in_interleaving_chain): Remove.
23244 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
23246 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
23248 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
23249 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
23250 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
23251 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
23252 Use slp_node instead of slp_void_p and adjust.
23254 2013-04-09 Richard Biener <rguenther@suse.de>
23256 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
23257 work that is not necessary.
23259 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23261 PR tree-optimization/56854
23262 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
23263 forward into clobber stmts if it would change MEM_REF lhs into
23266 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
23268 * tree.c (type_hash_lookup, type_hash_add): Make static.
23269 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
23271 2013-04-09 Richard Biener <rguenther@suse.de>
23273 * tree.h (unsave_expr_now): Remove.
23274 * tree-inline.c (mark_local_for_remap_r): Remove.
23275 (unsave_expr_1): Likewise.
23276 (unsave_r): Likewise.
23277 (unsave_expr_now): Likewise.
23278 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
23279 (propagate_tree_value): Likewise.
23281 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
23283 * doc/rtl.texi (sequence): Rewrite documentation to match the
23284 current use of SEQUENCE rtl objects.
23285 * rtl.def (SEQUENCE): Likewise.
23287 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
23288 Update documentation.
23289 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
23290 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
23292 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
23294 2013-04-08 Teresa Johnson <tejohnson@google.com>
23296 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
23297 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
23299 (estimate_edge_size_and_time): Add comment to suggest using rounding
23301 (estimate_node_size_and_time): Ditto.
23302 (remap_edge_change_prob): Use helper rounding divide methods.
23303 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
23304 (gimple_mod_pow2_value_transform): Ditto.
23305 (gimple_mod_subtract_transform): Ditto.
23306 (gimple_ic_transform): Ditto.
23307 (gimple_stringops_transform): Ditto.
23308 * stmt.c (conditional_probability): Ditto.
23309 (emit_case_dispatch_table): Ditto.
23310 * lto-cgraph.c (merge_profile_summaries): Ditto.
23311 * tree-optimize.c (execute_fixup_cfg): Ditto.
23312 * cfgcleanup.c (try_forward_edges): Ditto.
23313 * cfgloopmanip.c (scale_loop_profile): Ditto.
23315 (duplicate_loop_to_header_edge): Ditto.
23316 (lv_adjust_loop_entry_edge): Ditto.
23317 * tree-vect-loop.c (vect_transform_loop): Ditto.
23318 * profile.c (compute_branch_probabilities): Ditto.
23319 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
23320 * lto-streamer-in.c (input_cfg): Ditto.
23321 * gimple-streamer-in.c (input_bb): Ditto.
23322 * ipa-cp.c (update_profiling_info): Ditto.
23323 (update_specialized_profile): Ditto.
23324 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
23325 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
23327 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23328 (compute_trg_info): Ditto.
23329 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
23330 (purge_dead_edges): Ditto.
23331 * loop-unswitch.c (unswitch_loop): Ditto.
23332 * cgraphclones.c (cgraph_clone_edge): Ditto.
23333 (cgraph_clone_node): Ditto.
23334 * tree-inline.c (copy_bb): Ditto.
23335 (copy_edges_for_bb): Ditto.
23336 (initialize_cfun): Ditto.
23337 (copy_cfg_body): Ditto.
23338 (expand_call_inline): Ditto.
23340 2013-04-08 Kai Tietz <ktietz@redhat.com>
23342 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
23343 TARGET_CYGWIN64 by TARGET_64BIT.
23345 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23347 * config/epiphany/epiphany.md (GPR_1): New constant.
23348 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
23349 * config/epiphany/epiphany.c (gen_compare_reg):
23350 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
23351 is already in place.
23352 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
23353 Don't require being called during rtl expansion; If y operlaps r0,
23355 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
23356 (epiphany_expand_epilogue): Likewise.
23358 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
23359 Don't use CC_FPmode for ORDERED / UNORDERED.
23360 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
23362 * config/epiphany/constraints.md (CnL): New constraint.
23363 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
23364 * config/epiphany/predicates.md (add_operand): Allow 1024.
23366 * config/epiphany/epiphany.md (logical_op): New code iterator.
23367 (op_mnc): New code attribute.
23368 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
23369 (mov_f+1, mov_f+2): New peephole2 patterns.
23371 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
23372 (cstoresi4): Also allow re-use of zero result when doing a NE
23373 comparison to a non-zero operand.
23374 Use (clobber (scratch)) for first insn if the gpr output is not needed.
23376 * config/epiphany/epiphany.md (<insn_opname>v2si3):
23377 Use gen_addsi3_i / gen_subsi3_i.
23379 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23383 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
23384 contain anything but clobbers, at most one __builtin_stack_restore,
23385 optionally debug stmts and final resx, and if it has at least one
23386 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
23387 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
23388 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
23389 which isn't defaut definition, remove them.
23390 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
23391 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
23392 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
23393 with MEM_REF LHS with SSA_NAME address.
23395 2013-04-08 Jeff Law <law@redhat.com>
23397 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
23399 2013-04-08 Richard Biener <rguenther@suse.de>
23401 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
23403 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
23404 determined vector type.
23405 (vect_analyze_data_refs): Likewise.
23406 (vect_get_new_vect_var): Adjust.
23407 (vect_create_destination_var): Preserve SSA name versions.
23408 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
23409 not dump anything here.
23411 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23413 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
23414 Add member lr_slot_known.
23415 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
23417 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
23418 Remove code that sets lr_slot_offset according to what a previous
23419 version of epiphany_emit_save_restore used to do.
23420 (epiphany_emit_save_restore): When doing an lr save or restore,
23421 set/verify lr_slot_known and lr_slot_offset.
23423 2013-04-08 Xinyu Qi <xyqi@marvell.com>
23426 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
23429 2013-04-08 Richard Biener <rguenther@suse.de>
23431 * alias.c (find_base_term): Fix thinko in previous change.
23433 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23435 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
23436 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
23437 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
23438 if possible to compute val.
23439 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
23440 For QImode integers don't require anything about precision. Use
23441 const_with_all_bytes_same to find out if the constant doesn't have
23442 repeated bytes in it.
23444 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23446 * config/s390/s390.c (s390_expand_insv): Only accept insertions
23449 2013-04-08 Marek Polacek <polacek@redhat.com>
23451 PR rtl-optimization/48182
23452 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
23455 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23458 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
23459 nuses, make sure we have a label.
23461 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23464 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
23465 (rs6000_emit_swdiv_low_precision): Remove.
23466 (rs6000_emit_swdiv): Rewrite to handle between one and four
23467 iterations of Newton-Raphson generally; modify required number of
23468 iterations for some cases.
23469 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
23471 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
23473 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
23474 set-but-unused variable.
23476 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
23477 basic blocks of released function bodies garbage-collectable.
23479 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
23480 (struct rtl_opt_pass): Add TODO_df_finish.
23482 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
23484 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23486 * config/arm/constraints.md (q): New constraint.
23487 * config/arm/ldrdstrd.md: New file.
23488 * config/arm/arm.md (ldrdstrd.md) New include.
23489 (arm_movdi): Use "q" instead of "r" constraint
23490 for double-word memory access.
23491 (movdf_soft_insn): Likewise.
23492 * config/arm/vfp.md (movdi_vfp): Likewise.
23493 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
23494 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
23495 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
23496 (mem_ok_for_ldrd_strd): Likewise.
23497 (output_move_double): Update assertion.
23499 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23501 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
23503 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23505 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
23506 define_insn_and_split.
23507 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
23509 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23511 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
23512 define_insn_and_split.
23513 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
23514 (shiftsi3_compare): New pattern.
23515 (rrx): New pattern.
23516 * config/arm/unspecs.md (UNSPEC_RRX): New.
23518 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23520 * config/arm/arm.md (negdi_extendsidi): New pattern.
23521 (negdi_zero_extendsidi): Likewise.
23523 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23525 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
23526 define_insn_and_split.
23527 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
23528 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
23530 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23532 * config/arm/arm.md (arm_subdi3): Convert define_insn into
23533 define_insn_and_split.
23534 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
23535 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
23537 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23539 * config/arm/arm.md (subsi3_carryin): New pattern.
23540 (subsi3_carryin_const): Likewise.
23541 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
23542 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
23544 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23546 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
23548 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23550 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
23551 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
23553 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23555 * config/arm/arm.c (arm_expand_builtin): Change fcode
23556 type to unsigned int.
23558 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23560 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
23562 2013-04-04 Ian Lance Taylor <iant@google.com>
23564 * doc/standards.texi (Standards): The Go frontend supports the Go 1
23567 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
23569 PR middle-end/56729
23570 * df-scan.c (df_insn_delete): Disable failing assert.
23572 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23574 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
23575 New function prototype.
23576 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
23577 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
23578 (arm_builtin_vectorized_function): New function.
23580 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23582 * config/arm/arm_neon_builtins.def: New file.
23583 * config/arm/arm.c (neon_builtin_data): Move contents to
23584 arm_neon_builtins.def.
23585 (enum arm_builtins): Include neon builtin definitions.
23586 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
23587 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
23589 2013-04-04 Marek Polacek <polacek@redhat.com>
23591 PR tree-optimization/48186
23592 * predict.c (maybe_hot_frequency_p): Return false if
23593 HOT_BB_FREQUENCY_FRACTION is 0.
23594 (cgraph_maybe_hot_edge_p): Likewise.
23596 2013-04-04 Richard Biener <rguenther@suse.de>
23598 PR tree-optimization/56826
23599 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
23602 2013-04-04 Richard Biener <rguenther@suse.de>
23604 PR tree-optimization/56213
23605 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
23606 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
23608 2013-04-04 Richard Biener <rguenther@suse.de>
23610 PR tree-optimization/56837
23611 * tree-loop-distribution.c (classify_partition): For non-zero
23612 values require that the value has the same precision as its
23613 mode to be useful as memset value.
23615 2013-04-03 Nick Clifton <nickc@redhat.com>
23617 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
23618 (fmssf4): Use fmsf.s on E3V5 architectures.
23619 (fnmasf4): Use fnmaf.s on E3V5 architectures.
23620 (fnmssf4): Use fnmsf.s on E3V5 architectures.
23622 2013-04-03 Jeff Law <law@redhat.com>
23624 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
23625 (lra-eliminations.o): Likewise.
23627 2013-04-03 Teresa Johnson <tejohnson@google.com>
23629 * gcov-io.c (compute_working_sets): Moved most of body of old
23630 compute_working_sets here from profile.c.
23631 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
23632 (gcov_working_set_t): Moved typedef here from basic-block.h
23633 (compute_working_set): Declare.
23634 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
23635 (get_working_sets): Renamed from compute_working_set,
23636 replace most of body with call to new compute_working_sets.
23637 (get_exec_counts): Replace call to compute_working_sets
23638 to get_working_sets.
23639 * profile.h (get_working_sets): Renamed from compute_working_set.
23640 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
23641 to get_working_sets.
23642 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
23643 * gcov-dump.c (dump_working_sets): New function.
23645 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
23647 * hwint.c (sext_hwi, zext_hwi): New functions.
23648 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
23649 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
23650 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
23651 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
23652 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
23653 (sext_hwi, zext_hwi): New functions.
23655 2013-04-03 Jeff Law <law@redhat.com>
23657 PR tree-optimization/56799
23658 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
23659 back test for widening conversion erroneously dropped in prior change.
23661 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23664 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
23665 instead of next_real_insn.
23667 2013-04-03 Marek Polacek <polacek@redhat.com>
23670 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
23672 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23675 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
23677 (thumb1_output_casesi): Likewise.
23678 (thumb2_output_casesi): Likewise.
23680 2013-04-03 Richard Biener <rguenther@suse.de>
23682 PR tree-optimization/56817
23683 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
23685 (tree_unroll_loops_completely_1): ... new function to manually
23686 walk the loop tree, properly defering outer loops of unrolled
23687 loops to later iterations.
23689 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23691 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
23692 (vectorizable_load): Likewise.
23693 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
23694 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
23696 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23698 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
23701 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
23703 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
23705 2013-04-03 Bin Cheng <bin.cheng@arm.com>
23707 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
23709 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23711 PR tree-optimization/56790
23712 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
23715 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23717 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
23719 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
23720 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
23723 2013-04-03 Jakub Jelinek <jakub@redhat.com>
23726 * tree.h (force_folding_builtin_constant_p): New decl.
23727 * builtins.c (force_folding_builtin_constant_p): New variable.
23728 (fold_builtin_constant_p): Fold immediately also if
23729 force_folding_builtin_constant_p.
23731 2013-04-03 Richard Biener <rguenther@suse.de>
23733 PR tree-optimization/56812
23734 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
23735 DRs of the same interleaving chain are independent.
23737 2013-04-02 Jason Merrill <jason@redhat.com>
23739 * gdbinit.in (pbb): Use debug fn.
23741 2013-04-02 Lawrence Crowl <crowl@google.com>
23743 * sese.h (struct ivtype_map_elt_s): Remove unused.
23744 (extern debug_ivtype_map): Remove unused.
23745 (extern eq_ivtype_map_elts): Remove unused.
23746 * sese.c (debug_ivtype_map): Removed unused.
23747 (debug_ivtype_map_1): Removed unused.
23748 (debug_ivtype_elt): Remove unused.
23749 (eq_ivtype_map_elts): Remove unused.
23752 2013-04-02 Kai Tietz <ktietz@redhat.com>
23755 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
23756 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
23757 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
23759 (legitimize_pe_coff_symbol): Likewise.
23760 (is_imported_p): New helper-function.
23761 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
23762 for Windows x64 targets.
23763 (ix86_expand_prologue): Optimize for pe-coff targets.
23764 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
23765 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
23766 medium/large code-model.
23767 (legitimize_pic_address): Likewise.
23768 (legitimize_tls_address): Likewise.
23769 (ix86_expand_call): Likewise.
23770 (x86_output_mi_thunk): Likewise.
23771 (get_dllimport_decl): Add new beimport argument.
23772 (construct_plt_address): Don't assert for x64 pe-coff targets.
23773 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
23775 (SYMBOL_FLAG_STUBVAR): New macro.
23776 (SYMBOL_REF_STUBVAR_P): Likewise.
23777 * config/i386/winnt.c (stub_list): New structure.
23778 (stub_head): New local variable.
23779 (i386_pe_record_stub): New function.
23780 (i386_pe_file_end): Emit refptr-stubs.
23782 2013-04-02 Jakub Jelinek <jakub@redhat.com>
23784 PR rtl-optimization/56745
23785 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
23786 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
23789 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
23790 and both of them are MEM_REFs, just compare first argument for
23791 equality and attempt to deal even with differing offsets.
23794 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
23795 of gimple_clobber_p to be MEM_REF.
23796 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
23797 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
23798 after gimplification.
23799 * asan.c (get_mem_ref_of_assignment): Don't instrument
23800 gimple_clobber_p stmts.
23801 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
23802 gimple_clobber_p stmt if they have MEM_REF lhs and
23803 are dead because of another gimple_clobber_p stmt.
23804 * tree-ssa-live.c (clear_unused_block_pointer): Treat
23805 gimple_clobber_p stmts like debug stmts.
23806 (remove_unused_locals): Remove clobbers with MEM_REF lhs
23807 that refer to unused VAR_DECLs or uninitialized values.
23808 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
23809 gimple_clobber_p stmts if they refer to removed parameters.
23810 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
23813 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
23815 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
23816 using SWI48 mode attribute.
23818 2013-04-02 Wei Mi <wmi@google.com>
23820 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
23821 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
23822 *<rotate_insn><mode>3_mask in i386.md.
23824 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
23826 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
23828 2013-04-02 Richard Biener <rguenther@suse.de>
23830 PR tree-optimization/56778
23831 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
23832 Runtime alias tests are not supported for gather loads.
23833 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
23834 stmts referenced from SSA operands before updating SSA form.
23836 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
23837 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23839 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
23840 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
23841 * config/arm/cortex-a53.md: New file.
23842 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
23843 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
23844 * config/arm/arm.c (arm_issue_rate): Likewise.
23845 * config/arm/arm-tune.md: Regenerate
23846 * config/arm/arm-tables.opt: Regenerate.
23847 * config/arm/arm-cores.def: Add cortex-a53.
23849 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
23851 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
23854 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23856 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
23857 scalar load/store operations using B/H registers.
23858 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
23860 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23862 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
23864 * config/aarch64/aarch64.c
23865 (aarch64_simd_scalar_immediate_valid_for_move): New.
23866 * config/aarch64/aarch64-protos.h
23867 (aarch64_simd_scalar_immediate_valid_for_move): New.
23868 * config/aarch64/constraints.md (Dh, Dq): New.
23869 * config/aarch64/iterators.md (hq): New.
23871 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
23873 * reorg.c (get_branch_condition): Deal with conditional returns.
23874 (fill_simple_delay_slots): Remove dead code dealing with jumps.
23876 2013-04-01 Wei Mi <wmi@google.com>
23878 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
23879 Truncate operand 2 using %b asm operand modifier.
23880 (*<shift_insn><mode>3_mask): Ditto.
23881 (*<rotate_insn><mode>3_mask): Ditto.
23883 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
23885 PR middle-end/56798
23886 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
23888 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
23890 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
23892 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
23894 2013-03-30 Lawrence Crowl <crowl@google.com>
23896 * dse.c (clear_alias_sets): Remove never set.
23897 (disqualified_clear_alias_sets): Remove never set.
23898 (clear_alias_mode_pool): Remove never set.
23899 (dse_step0): Remove condition that is never true.
23900 (canon_address): Remove condition that is never true.
23901 (dse_step7): Remove condition that is never true.
23902 (rest_of_handle_dse): Remove condition that is never true.
23903 (rest_of_handle_dse::did_global): Remove never read from above.
23904 (dse_step2_spill): Remove never called from above.
23905 (dse_step5_spill): Remove never called from above.
23907 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
23909 * doc/md.texi (Standard Names) <casesi>: Update documentation for
23910 JUMP_TABLE_DATA changes.
23911 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
23912 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
23913 (Insns) <jump_table_data>: New entry.
23914 * doc/tm.texi: Regenerate.
23916 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
23918 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
23919 for table jump at the end of a basic block using tablejump_p.
23920 * targhooks.c (default_invalid_within_doloop): Likewise.
23921 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
23922 target hook implementation that is identical to the default hook.
23923 (rs6000_invalid_within_doloop): Remove.
23925 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
23926 unused variable from tablejump_p call.
23928 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
23929 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
23930 (INSN_DELETED_P): Likewise.
23931 (emit_jump_table_data): New prototype.
23932 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
23933 after 4th as unused.
23934 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
23935 * sched-vis.c (print_insn): Likewise.
23936 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
23937 insn for compatibility with back ends that use next_active_insn to
23938 identify jump table data.
23939 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
23940 (remove_insn): Likewise.
23941 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
23943 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
23944 (emit_jump_table_data): New function.
23946 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
23947 basic block, a JUMP_TABLE_DATA never is.
23948 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
23949 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
23950 off from code handling real insns.
23951 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
23952 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
23953 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
23954 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
23955 is not a NONDEBUG_INSN_P.
23956 * ira-costs.c (scan_one_insn): Likewise.
23957 * jump.c (mark_all_labels): Likewise.
23958 (mark_jump_label_1): Likewise.
23959 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
23960 * lra.c (get_insn_freq): Expect all insns reaching here to be in
23962 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
23963 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
23964 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
23965 JUMP_TABLE_DATA_P insns.
23966 (calculate_elim_costs_all_insns): Likewise.
23967 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
23968 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
23969 (delete_output_reload): Code style fixups.
23970 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
23971 insn flags on this non-insn.
23972 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
23973 as scheduling barriers, for pre-change compatibility.
23974 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
23975 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
23977 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
23978 redundant JUMP_TABLE_DATA_P test.
23979 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
23980 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
23981 (frv_for_each_packet): Likewise.
23982 * config/i386/i386.c (min_insn_size): Likewise.
23983 (ix86_avoid_jump_mispredicts): Likewise.
23984 * config/m32r/m32r.c (m32r_is_insn): Likewise.
23985 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
23986 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
23987 (mips16_insn_length): Robustify.
23988 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
23989 (mips16_split_long_branches): Likewise.
23990 * config/pa/pa.c (pa_combine_instructions): Likewise.
23991 * config/rs6000/rs6000.c (get_next_active_insn): Treat
23992 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
23993 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
23994 as contributing to pool range lengths.
23995 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
23996 Remove redundant JUMP_TABLE_DATA_P test.
23997 (sh_loop_align): Likewise.
23998 (split_branches): Likewise.
23999 (sh_insn_length_adjustment): Likewise.
24000 * config/spu/spu.c (get_branch_target): Likewise.
24002 2013-03-29 Jan Hubicka <jh@suse.cz>
24004 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
24005 gcov streaming; stream hot bb threshold to ltrans.
24006 * predict.c (get_hot_bb_threshold): Break out from ....
24007 (maybe_hot_count_p): ... here.
24008 (set_hot_bb_threshold): New function.
24009 * lto-section-in.c (lto_section_name): Add profile.
24010 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
24011 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
24012 and data-streamer.h
24013 (histogram_entry): New structure.
24014 (histogram, histogram_pool): New global vars.
24015 (histogram_hash): New structure.
24016 (histogram_hash::hash): New method.
24017 (histogram_hash::equal): Likewise.
24018 (account_time_size): New function.
24019 (cmp_counts): New function.
24020 (dump_histogram): New function.
24021 (ipa_profile_generate_summary): New function.
24022 (ipa_profile_write_summary): New function.
24023 (ipa_profile_read_summary): New function.
24024 (ipa_profile): Decide on threshold.
24025 (pass_ipa_profile): Add ipa_profile_write_summary and
24026 ipa_profile_read_summary.
24027 * Makefile.in (ipa.o): Update dependencies.
24028 * lto-streamer.h (LTO_section_ipa_profile): New section.
24030 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
24032 * tree.h (VAR_P): New.
24034 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
24037 * doc/invoke.texi ([-fwhole-program]): Fix typo.
24039 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
24041 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
24042 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24043 (control_flow_insn_p): Likewise.
24044 * cfgrtl.c (duplicate_insn_chain): Likewise.
24045 * final.c (get_attr_length_1): Likewise.
24046 (shorten_branches): Likewise.
24047 (final_scan_insn): Likewise.
24048 * function.c (instantiate_virtual_regs): Likewise.
24049 * gcse.c (insert_insn_end_basic_block): Likewise.
24050 * ira-costs.c (scan_one_insn): Likewise.
24051 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24052 * lra.c (check_rtl): Likewise.
24053 * reload1.c (elimination_costs_in_insn): Likewise.
24054 * reorg.c (follow_jumps): Likewise.
24056 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
24057 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24058 (thumb_far_jump_used_p): Likewise.
24059 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
24060 (workaround_speculation): Likewise.
24061 (add_sched_insns_for_speculation): Likewise.
24062 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
24063 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24064 (frv_for_each_packet): Likewise.
24065 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
24066 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
24067 (final_emit_insn_group_barriers): Likewise.
24068 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24069 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24070 (mips16_insn_length): Likewise.
24071 * config/pa/pa.c (pa_reorg): Likewise.
24072 (pa_combine_instructions): Likewise.
24073 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
24074 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
24075 (sh_reorg): Likewise.
24076 (split_branches): Likewise.
24077 * config/spu/spu.c (get_branch_target): Likewise.
24079 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
24082 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
24084 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
24085 Fix declaration name.
24087 2013-03-28 Lawrence Crowl <crowl@google.com>
24089 * graphds.h (struct graph.indicies): Remove unused.
24090 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
24091 (SCOP_ORIGINAL_PDDRS): Remove unused.
24092 * sese.h (extern insert_loop_close_phis): Removed unused.
24093 (extern insert_guard_phis): Removed unused.
24094 (extern ivtype_map_elt_info): Removed unused.
24095 (new_ivtype_map_elt): Removed unused.
24096 * sese.c (ivtype_map_elt_info): Removed unused.
24098 2013-03-28 Lawrence Crowl <crowl@google.com>
24100 * Makefile.in: Add several missing include dependences.
24102 (test-dump.o): New. This object is not added to any executable,
24103 but is present for ad-hoc testing.
24105 (debug (const bitmap_head_def &)): New.
24106 (debug (const bitmap_head_def *)): New.
24108 (extern debug (const bitmap_head_def &)): New.
24109 (extern debug (const bitmap_head_def *)): New.
24111 (debug (edge_def &)): New.
24112 (debug (edge_def *)): New.
24114 (debug (basic_block_def &)): New.
24115 (debug (basic_block_def *)): New.
24117 (dump_node (const_tree, int, FILE *)): Correct source file.
24119 (debug (die_struct &)): New.
24120 (debug (die_struct *)): New.
24122 (extern debug (die_struct &)): New.
24123 (extern debug (die_struct *)): New.
24124 * gimple-pretty-print.c
24125 (debug (gimple_statement_d &)): New.
24126 (debug (gimple_statement_d *)): New.
24127 * gimple-pretty-print.h
24128 (extern debug (gimple_statement_d &)): New.
24129 (extern debug (gimple_statement_d *)): New.
24131 (debug (ira_allocno_copy &)): New.
24132 (debug (ira_allocno_copy *)): New.
24133 (debug (ira_allocno &)): New.
24134 (debug (ira_allocno *)): New.
24136 (extern debug (ira_allocno_copy &)): New.
24137 (extern debug (ira_allocno_copy *)): New.
24138 (extern debug (ira_allocno &)): New.
24139 (extern debug (ira_allocno *)): New.
24141 (debug (live_range &)): New.
24142 (debug (live_range *)): New.
24144 (debug (lra_live_range &)): New.
24145 (debug (lra_live_range *)): New.
24147 (debug (lra_live_range &)): New.
24148 (debug (lra_live_range *)): New.
24150 (debug (omega_pb_d &)): New.
24151 (debug (omega_pb_d *)): New.
24153 (extern debug (omega_pb_d &)): New.
24154 (extern debug (omega_pb_d *)): New.
24156 (debug (const rtx_def &)): New.
24157 (debug (const rtx_def *)): New.
24159 (debug_tree (tree): Move within file.
24160 (debug_raw (const tree_node &)): New.
24161 (debug_raw (const tree_node *)): New.
24162 (dump_tree_via_hooks (const tree_node *, int)): New.
24163 (debug (const tree_node &)): New.
24164 (debug (const tree_node *)): New.
24165 (debug_verbose (const tree_node &)): New.
24166 (debug_verbose (const tree_node *)): New.
24167 (debug_head (const tree_node &)): New.
24168 (debug_head (const tree_node *)): New.
24169 (debug_body (const tree_node &)): New.
24170 (debug_body (const tree_node *)): New.
24171 (debug_vec_tree (tree): Move and reimplement in terms of dump.
24172 (debug (vec<tree, va_gc> &)): New.
24173 (debug (vec<tree, va_gc> *)): New.
24175 (extern debug (const rtx_def &)): New.
24176 (extern debug (const rtx_def *)): New.
24178 (debug_raw (simple_bitmap_def &)): New.
24179 (debug_raw (simple_bitmap_def *)): New.
24180 (debug (simple_bitmap_def &)): New.
24181 (debug (simple_bitmap_def *)): New.
24183 (extern debug (simple_bitmap_def &)): New.
24184 (extern debug (simple_bitmap_def *)): New.
24185 (extern debug_raw (simple_bitmap_def &)): New.
24186 (extern debug_raw (simple_bitmap_def *)): New.
24188 (debug (vinsn_def &)): New.
24189 (debug (vinsn_def *)): New.
24190 (debug_verbose (vinsn_def &)): New.
24191 (debug_verbose (vinsn_def *)): New.
24192 (debug (expr_def &)): New.
24193 (debug (expr_def *)): New.
24194 (debug_verbose (expr_def &)): New.
24195 (debug_verbose (expr_def *)): New.
24196 (debug (vec<rtx> &)): New.
24197 (debug (vec<rtx> *)): New.
24199 (extern debug (vinsn_def &)): New.
24200 (extern debug (vinsn_def *)): New.
24201 (extern debug_verbose (vinsn_def &)): New.
24202 (extern debug_verbose (vinsn_def *)): New.
24203 (extern debug (expr_def &)): New.
24204 (extern debug (expr_def *)): New.
24205 (extern debug_verbose (expr_def &)): New.
24206 (extern debug_verbose (expr_def *)): New.
24207 (extern debug (vec<rtx> &)): New.
24208 (extern debug (vec<rtx> *)): New.
24210 (_list_iter_cond_expr): Make inline instead of static.
24212 (debug (sreal &)): New.
24213 (debug (sreal *)): New.
24215 (extern debug (sreal &)): New.
24216 (extern debug (sreal *)): New.
24218 (extern debug_raw (const tree_node &)): New.
24219 (extern debug_raw (const tree_node *)): New.
24220 (extern debug (const tree_node &)): New.
24221 (extern debug (const tree_node *)): New.
24222 (extern debug_verbose (const tree_node &)): New.
24223 (extern debug_verbose (const tree_node *)): New.
24224 (extern debug_head (const tree_node &)): New.
24225 (extern debug_head (const tree_node *)): New.
24226 (extern debug_body (const tree_node &)): New.
24227 (extern debug_body (const tree_node *)): New.
24228 (extern debug (vec<tree, va_gc> &)): New.
24229 (extern debug (vec<tree, va_gc> *)): New.
24231 (debug (struct loop &)): New.
24232 (debug (struct loop *)): New.
24233 (debug_verbose (struct loop &)): New.
24234 (debug_verbose (struct loop *)): New.
24235 * tree-dump.c: Add header dependence.
24237 (extern debug (struct loop &)): New.
24238 (extern debug (struct loop *)): New.
24239 (extern debug_verbose (struct loop &)): New.
24240 (extern debug_verbose (struct loop *)): New.
24242 (debug (data_reference &)): New.
24243 (debug (data_reference *)): New.
24244 (debug (vec<data_reference_p> &)): New.
24245 (debug (vec<data_reference_p> *)): New.
24246 (debug (vec<ddr_p> &)): New.
24247 (debug (vec<ddr_p> *)): New.
24249 (extern debug (data_reference &)): New.
24250 (extern debug (data_reference *)): New.
24251 (extern debug (vec<data_reference_p> &)): New.
24252 (extern debug (vec<data_reference_p> *)): New.
24253 (extern debug (vec<ddr_p> &)): New.
24254 (extern debug (vec<ddr_p> *)): New.
24256 (debug (pt_solution &)): New.
24257 (debug (pt_solution *)): New.
24259 (extern debug (pt_solution &)): New.
24260 (extern debug (pt_solution *)): New.
24262 (debug (_var_map &)): New.
24263 (debug (_var_map *)): New.
24264 (debug (tree_live_info_d &)): New.
24265 (debug (tree_live_info_d *)): New.
24267 (extern debug (_var_map &)): New.
24268 (extern debug (_var_map *)): New.
24269 (extern debug (tree_live_info_d &)): New.
24270 (extern debug (tree_live_info_d *)): New.
24272 2013-03-28 Jan Hubicka <jh@suse.cz>
24274 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
24276 2013-03-28 Ian Bolton <ian.bolton@arm.com>
24278 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
24279 record only when desired or required.
24281 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
24283 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
24284 *vec_extractv2di_1_rex64. Use x64 isa attribute.
24286 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24288 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
24289 (*andsi3_compare0_uxtw): New pattern.
24290 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
24291 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
24293 2013-03-28 Jan Hubicka <jh@suse.cz>
24295 * data-streamer-in.c (streamer_read_gcov_count): New function.
24296 * gimple-streamer-out.c: Include value-prof.h.
24297 (output_gimple_stmt): Output histogram.
24298 (output_bb): Use streamer_write_gcov_count.
24299 * value-prof.c: Include data-streamer.h
24300 (dump_histogram_value): Add HIST_TYPE_MAX.
24301 (stream_out_histogram_value): New function.
24302 (stream_in_histogram_value): New function.
24303 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
24304 (stream_out_histogram_value, stream_in_histogram_value): Declare.
24305 * data-streamer-out.c (streamer_write_gcov_count): New function.
24306 (streamer_write_gcov_count_stream): New function.
24307 * lto-cgraph.c (lto_output_edge): Update counter streaming.
24308 (lto_output_node): Likewise.
24309 (input_node, input_edge): Likewise.
24310 * lto-streamer-out.c (output_cfg): Update streaming.
24311 * lto-streamer-in.c (input_cfg): Likewise.
24312 * data-streamer.h (streamer_write_gcov_count,
24313 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
24314 * gimple-streamer-in.c: Include value-prof.h
24315 (input_gimple_stmt): Input histograms.
24316 (input_bb): Update profile streaming.
24318 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24320 * genmodes.c (emit_max_int): New function.
24321 (emit_insn_modes_h): Added call to emit_max_function.
24322 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
24324 * machmode.def: Fixed comment.
24326 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24328 * combine.c (try_combine): Removed useless assert.
24329 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
24331 2013-03-28 Marek Polacek <polacek@redhat.com>
24332 Richard Biener <rguenther@suse.de>
24334 PR tree-optimization/56695
24335 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
24336 build signed result of a vector comparison.
24337 * tree-cfg.c (verify_gimple_comparison): Check that a result
24338 of a vector comparison has signed type.
24340 2013-03-28 Richard Biener <rguenther@suse.de>
24342 PR tree-optimization/37021
24343 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
24344 do not restrict gaps between groups.
24345 * tree-vect-stmts.c (vectorizable_load): Properly account for
24346 a gap between groups.
24348 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
24350 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
24351 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
24354 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
24356 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
24357 * doc/extend.texi (Named Address Spaces): Ditto.
24358 (Variable Attributes): Ditto.
24360 2013-03-27 Kai Tietz <ktietz@redhat.com>
24362 * config.build: Add support for cygwin x64 target.
24363 * config.gcc: Likewise.
24364 * config.host: Likewise.
24365 * configure.ac: Likewise
24366 * configure: Regenerated.
24368 2013-03-27 Kai Tietz <ktietz@redhat.com>
24370 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
24371 * config/i386/t-cygwin-w64: New file.
24372 * config/i386/cygwin-w64.h: New file.
24373 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
24374 and add support for x64-cygwin target.
24375 (CPP_SPEC): Likewise.
24376 (CXX_WRAP_SPEC_LIST): Undefine before define.
24377 (LIBGCJ_SONAME): Use 15 as version.
24379 2013-03-27 Richard Biener <rguenther@suse.de>
24381 PR tree-optimization/56716
24382 * tree-ssa-structalias.c (perform_var_substitution): Adjust
24383 dumping for ref nodes.
24385 2013-03-27 Martin Jambor <mjambor@suse.cz>
24387 PR tree-optimization/55334
24388 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
24389 restricted pointers to arrays.
24391 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
24393 * Makefile.in (.SUFFIXES): Add .cc.
24394 (.c.o): Apply same recipe for implicit rule .cc.o.
24396 2013-03-27 Richard Biener <rguenther@suse.de>
24398 PR tree-optimization/37021
24399 * tree-vect-data-refs.c (vect_check_strided_load): Allow
24400 REALPART/IMAGPART_EXPRs around the supported refs.
24401 * tree-ssa-structalias.c (find_func_aliases): Assume that
24402 floating-point values are not used to transfer pointers.
24404 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
24406 * target.def (TARGET_HAS_IFUNC_P): New target hook.
24407 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
24408 * doc/tm.texi: Regenerate.
24409 * targhooks.h (default_has_ifunc_p): New.
24410 * targhooks.c (default_has_ifunc_p): Ditto.
24411 * config/linux-protos.h: New file.
24412 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
24413 hook for linux which disables support of indirect functions in android.
24414 * config/linux-android.c: New file.
24415 * config/t-linux-android.c: Ditto.
24416 * config.gcc: Added new object file linux-android.o.
24417 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
24418 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
24419 * varasm.c (do_assemble_alias): Likewise.
24420 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
24421 doesn't support indirect functions.
24422 * configure: Regenerate.
24424 2013-03-27 Bin Cheng <bin.cheng@arm.com>
24427 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
24428 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
24431 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24433 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
24435 2013-03-27 Terry Guo <terry.guo@arm.com>
24437 * config/arm/arm-cores.def: Added core cortex-r7.
24438 * config/arm/arm-tune.md: Regenerated.
24439 * config/arm/arm-tables.opt: Regenerated.
24440 * doc/invoke.texi: Added entry for core cortex-r7.
24442 2013-03-27 Walter Lee <walt@tilera.com>
24444 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
24445 double-decrement of next_scratch_regno.
24447 2013-03-27 Walter Lee <walt@tilera.com>
24449 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
24451 (insn_v1mulus): Ditto.
24452 (insn_v2muls): Ditto.
24454 2013-03-27 Walter Lee <walt@tilera.com>
24456 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
24457 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
24459 2013-03-27 Walter Lee <walt@tilera.com>
24461 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
24462 (*sibcall_value): Ditto.
24464 2013-03-27 Walter Lee <walt@tilera.com>
24466 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
24467 (insn_mnz_v8qi): ... this ...
24468 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
24470 (insn_v<n>mnz): Replaced by ...
24471 (insn_v1mnz): ... this ...
24472 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
24474 (insn_mz_<mode>): Replaced by ...
24475 (insn_mz_v8qi): ... this ...
24476 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
24478 (insn_v<n>mz): Replaced by ...
24479 (insn_v1mz): ... this ...
24480 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
24483 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
24485 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
24487 2013-03-26 Roland McGrath <mcgrathr@google.com>
24489 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
24490 than fprintf with a non-constant, non-format string.
24492 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24494 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
24495 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
24496 operand 0 predicate.
24497 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
24498 attribute. Use general_x64nomem_operand as operand 1 predicate.
24499 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
24500 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24501 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
24502 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24503 (mov<mode>_insv_1): Remove expander. Merge insn with
24504 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
24505 Use general_x64nomem_operand as operand 1 predicate.
24506 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
24507 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
24508 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24509 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
24510 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
24511 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
24512 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24513 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
24514 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
24515 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
24516 (general_x64nomem_operand): Ditto.
24518 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
24520 * config/rtems.opt: Add -pthread option.
24522 2013-03-26 Richard Biener <rguenther@suse.de>
24524 * alias.c (find_base_term): Avoid redundant and not used recursion.
24525 (base_alias_check): Get the initial base term from the caller.
24526 (true_dependence_1): Compute and pass base terms to base_alias_check.
24527 (write_dependence_p): Likewise.
24528 (may_alias_p): Likewise.
24530 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
24532 * config/aarch64/aarch64.c (aarch64_classify_address): Support
24533 PC-relative load in SI modes and above only.
24535 2013-03-26 Xinyu Qi <xyqi@marvell.com>
24537 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
24538 * config/arm/iwmmxt.md (WCGR0): Update.
24539 (WCGR1, WCGR2, WCGR3): Likewise.
24541 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24543 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
24544 Use x64 and nox64 isa attributes.
24546 2013-03-26 Richard Biener <rguenther@suse.de>
24548 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
24549 alignment computations and rely on get_object_alignment_1
24550 for the !TYPE_P case.
24551 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
24553 2013-03-26 Walter Lee <walt@tilera.com>
24555 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
24556 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
24558 2013-03-25 Jeff Law <law@redhat.com>
24560 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
24561 check for INTEGRAL_TYPE_P that was missing due to checking in
24562 wrong version of prior patch.
24564 2013-03-25 Walter Lee <walt@tilera.com>
24566 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
24567 TILEGX_INSN_SHUFFLEBYTES1.
24568 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
24570 (tilegx_builtins): Ditto.
24571 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
24573 2013-03-25 Walter Lee <walt@tilera.com>
24575 * config/tilegx/tilegx.md (floatsisf2): New pattern.
24576 (floatunssisf2): New pattern.
24577 (floatsidf2): New pattern.
24578 (floatunssidf2): New pattern.
24580 2013-03-25 Walter Lee <walt@tilera.com>
24582 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
24583 tests for constraint J, K, N, P.
24585 2013-03-25 Walter Lee <walt@tilera.com>
24587 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
24588 Use indirect/pcrel encoding.
24589 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
24592 2013-03-25 Steve Ellcey <sellcey@mips.com>
24594 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
24595 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
24596 * config/mips/mips.c (mips_option_override): Set IMADD default.
24597 * config/mips/mips.h (PTF_AVOID_IMADD): New.
24598 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
24599 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
24600 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
24601 * doc/invoke.texi (-mimadd/-mno-imadd): New.
24603 2013-03-25 Jeff Law <law@redhat.com>
24605 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
24606 slightly to avoid creating and folding useless trees. Simplify
24607 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
24609 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
24611 * config/i386/i386.md (*zero_extendsidi2): Merge with
24612 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
24613 * config/i386/predicates.md (x86_64_zext_operand): Rename from
24614 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
24615 targets. Clarify comment.
24617 2013-03-25 Martin Jambor <mjambor@suse.cz>
24619 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
24620 pass-through jump functions differently.
24621 (ipa_read_jump_function): Likewise. Also use setter functions to set
24624 2013-03-25 Martin Jambor <mjambor@suse.cz>
24626 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
24627 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
24629 (ipa_get_indirect_edge_target): New function.
24630 (devirtualization_time_bonus): New parameter known_aggs, pass it to
24631 ipa_get_indirect_edge_target. Update all callers.
24632 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
24633 ipa_get_indirect_edge_target_1 instead of calling
24634 ipa_get_indirect_edge_target.
24635 (create_specialized_node): Pass aggvlas to
24636 ipcp_discover_new_direct_edges.
24638 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24640 * config/arm/arm.md (f_sels, f_seld): New types.
24641 (*cmov<mode>): New pattern.
24642 * config/arm/predicates.md (arm_vsel_comparison_operator): New
24645 2013-03-25 Kai Tietz <ktietz@redhat.com>
24647 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
24648 POSIX-printf for mingw-hosted builds.
24650 2013-03-25 Richard Biener <rguenther@suse.de>
24652 PR middle-end/56694
24653 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
24654 must-not-throw stmt location.
24656 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24658 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
24659 Emit load-acquire versions when acq is true.
24660 (arm_emit_store_exclusive): Add rel parameter.
24661 Emit store-release versions when rel is true.
24662 (arm_split_compare_and_swap): Use acquire-release instructions
24664 of barriers when appropriate.
24665 (arm_split_atomic_op): Likewise.
24666 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
24667 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
24668 (VUNSPEC_SLX): Likewise.
24669 (VUNSPEC_LDA): Likewise.
24670 (VUNSPEC_STL): Likewise.
24671 * config/arm/sync.md (atomic_load<mode>): New pattern.
24672 (atomic_store<mode>): Likewise.
24673 (arm_load_acquire_exclusive<mode>): Likewise.
24674 (arm_load_acquire_exclusivesi): Likewise.
24675 (arm_load_acquire_exclusivedi): Likewise.
24676 (arm_store_release_exclusive<mode>): Likewise.
24678 2013-03-25 Catherine Moore <clm@codesourcery.com>
24680 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
24681 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
24682 * config/mip/predicates.md (lwsp_swsp_operand,
24683 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
24684 sb16_operand, db4_operand, db7_operand, ib3_operand,
24685 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
24686 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
24687 andi16_operand): New predicates.
24688 * config/mips/mips.md (compression): New attribute.
24689 (enabled): New attribute.
24690 (length): Consider compression in computing length.
24691 (shift_compression): New code attribute.
24692 (*add<mode>3): New operands. Record compression.
24693 (sub<mode>3): Likewise.
24694 (one_cmpl<mode>2): Likewise.
24695 (*and<mode>3): Likewise.
24696 (*ior<mode>3): Likewise.
24697 (unnamed pattern for xor): Likewise.
24698 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
24699 (*<optab><mode>3): Likewise.
24700 (*mov<mode>_internal: Likewise.
24701 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
24702 (mips_unsigned_immediate_p): New.
24703 (umips_lwsp_swsp_address_p): New.
24704 (m16_based_address_p): New.
24705 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
24706 (mips_unsigned_immediate_p): New prototype.
24707 (lwsp_swsp_address_p): New prototype.
24708 (m16_based_address_p): New prototype.
24709 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
24710 (mips_signed_immediate_p): New function.
24711 (m16_based_address_p): New function.
24712 (lwsp_swsp_address_p): New function.
24713 (mips_print_operand_punctuation): Recognize short delay slot insns
24714 for microMIPS.add<mode>3"
24716 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24719 * config/arm/iterators.md (v_cmp_result): New mode attribute.
24720 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
24722 2013-03-25 Richard Biener <rguenther@suse.de>
24724 PR tree-optimization/56689
24725 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
24728 2013-03-25 Richard Biener <rguenther@suse.de>
24730 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
24732 (memory_references): Likewise.
24733 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
24734 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
24735 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
24736 (gather_mem_refs_in_loops): Fold into ...
24737 (analyze_memory_references): ... this. Move initialization
24738 to tree_ssa_lim_initialize.
24739 (fill_always_executed_in): Rename to ...
24740 (fill_always_executed_in_1): ... this.
24741 (fill_always_executed_in): Move contains_call computation to
24742 this new function from ...
24743 (tree_ssa_lim_initialize): ... here.
24744 (tree_ssa_lim): Call fill_always_executed_in.
24746 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24748 * postreload.c (reload_combine): Fix code detecting returns.
24750 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24752 * function.c (emit_use_return_register_into_block): On cc0 targets,
24753 do not emit the sequence between cc0 setter and user.
24755 2013-03-25 Kai Tietz <ktietz@redhat.com>
24757 * config/i386/predicates.md (local_symbolic_operand): Interpret
24758 dll-imported symbols as none-local.
24760 2013-03-25 Richard Biener <rguenther@suse.de>
24762 * tree-ssa-loop-im.c (struct depend): Remove.
24763 (struct lim_aux_data): Make depends a vec of gimples.
24764 (free_lim_aux_data): Adjust.
24765 (add_dependency): Likewise.
24766 (set_level): Likewise.
24768 2013-03-25 Richard Biener <rguenther@suse.de>
24770 PR middle-end/56434
24771 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
24772 the pointer returned by calls with ECF_MALLOC set.
24774 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24776 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
24778 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24780 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
24781 using MMXMODE mode iterator.
24782 (*move<mode>_internal): Merge with *movv2sf_internal and
24783 *movv2sf_internal_rex64 using MMXMODE mode iterator.
24785 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24787 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
24788 (record_last_mem_set_info): Likewise.
24790 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
24791 of XNEWVEC followed by memset.
24792 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
24794 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24796 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
24797 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
24798 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
24799 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
24800 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
24801 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
24802 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
24803 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
24804 BARRIER_P instead of GET_CODE.
24806 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24808 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
24809 inaccuracy in the probing code.
24811 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
24812 (ctrapdi4): Likewise.
24814 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24816 * calls.c (expand_call): Add missing guard to code handling return
24817 of non-BLKmode structures in MSB.
24818 * function.c (expand_function_end): Likewise.
24820 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24822 * combine.c (try_combine): Adjust comment. Do not add the set of
24823 insn #0 if the destination indirectly is set or dies in insn #2.
24824 Tidy up code to distribute a new note.
24826 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24828 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
24829 also for alternatives 16 and 17.
24831 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24833 * config/i386/sse.md (*mov<mode>_internal): Merge with
24834 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
24835 Emit insn template depending on type attribute. Use
24836 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24837 movd instead of movq mnemonic for interunit moves. Rewrite mode
24838 attribute calculation. Remove unit attribute calculation.
24839 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
24840 Set prefix_data16 attribute for DImode ssemov types.
24841 Use Ym instead of y for SSE-MMX conversion alternatives.
24842 Reorder operand constraints.
24844 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
24846 * df.h (df_insn_delete): Adjust prototype.
24847 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
24848 and let it decide whether mark the basic block dirty.
24849 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
24850 * df-scan.c (df_insn_info_delete): New helper function, split
24851 off from df_insn_delete.
24852 (df_scan_free_bb_info): Use it.
24853 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
24855 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
24856 that the insn is actually an insn and it has a non-NULL basic block.
24857 Do not mark basic block dirty if only deleting a DEBUG_INSN.
24859 2013-03-22 Richard Biener <rguenther@suse.de>
24861 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
24863 (mem_ref_alloc): Do not allocate them.
24864 (refs_independent_p): Do not query or maintain a cache.
24866 2013-03-22 Richard Biener <rguenther@suse.de>
24868 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
24869 (gather_mem_refs_in_loops): Do not compute it.
24870 (analyze_memory_references): Do not allocate it.
24871 (tree_ssa_lim_finalize): Do not free it.
24872 (for_all_locs_in_loop): Do not query all_refs_in_loop.
24874 2013-03-22 Richard Biener <rguenther@suse.de>
24876 * is-a.h (as_a): Use gcc_checking_assert.
24878 2013-03-22 Ian Bolton <ian.bolton@arm.com>
24880 * config/aarch64/aarch64.c (aarch64_print_operand): New
24881 format specifier for printing a constant in hex.
24882 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
24883 format specifier for printing second operand.
24885 2013-03-22 Richard Biener <rguenther@suse.de>
24887 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
24889 (gather_mem_refs_in_loops): Perform store accumulation here.
24890 (create_vop_ref_mapping_loop): Remove.
24891 (create_vop_ref_mapping): Likewise.
24892 (analyze_memory_references): Initialize refs_stored_in_loop.
24893 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
24894 (record_indep_loop): Remove.
24895 (record_dep_loop): New function.
24896 (ref_indep_loop_p_1): Adjust to only walk over references
24897 in the loop, not its subloops.
24898 (ref_indep_loop_p): Rename to ...
24899 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
24900 maintaining a more fine-grained cache.
24901 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
24902 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
24904 2013-03-22 Richard Biener <rguenther@suse.de>
24906 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
24907 (struct mem_ref): Make accesses_in_loop a vec of a vec of
24908 aggregate mem_ref_loc.
24909 (free_mem_ref_locs): Inline into ...
24910 (memref_free): ... this and adjust.
24911 (mem_ref_alloc): Adjust.
24912 (mem_ref_locs_alloc): Remove.
24913 (record_mem_ref_loc): Adjust.
24914 (get_all_locs_in_loop): Rewrite into ...
24915 (for_all_locs_in_loop): ... this iterator.
24916 (rewrite_mem_ref_loc): New functor.
24917 (rewrite_mem_refs): Use for_all_locs_in_loop.
24918 (sm_set_flag_if_changed): New functor.
24919 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
24920 (ref_always_accessed): New functor.
24921 (ref_always_accessed_p): Use for_all_locs_in_loop.
24923 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24925 * tree-pass.h (PROP_gimple_lvec): New.
24926 * passes.c (dump_properties): Handle PROP_gimple_lvec.
24927 (init_optimization_passes): Move pass_lower_vector.
24928 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
24930 (pass_lower_vector): Provide PROP_gimple_lvec.
24931 (pass_lower_vector_ssa): Likewise.
24932 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
24934 2013-03-21 Mark Wielaard <mjw@redhat.com>
24936 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
24938 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
24940 * config/i386/i386.md (*movdi_internal): Disparage slightly
24941 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
24942 conversion alternatives.
24944 2013-03-21 Jakub Jelinek <jakub@redhat.com>
24946 PR middle-end/48087
24947 * diagnostic.def (DK_WERROR): New kind.
24948 * diagnostic.h (werrorcount): Define.
24949 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
24950 promoted to DK_ERROR, increment DK_WERROR counter instead of
24952 * toplev.c (toplev_main): Call print_ignored_options even if
24953 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
24954 even if just werrorcount is non-zero.
24957 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
24959 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
24960 (string_cst_pool_decl): New function.
24961 (optimize_one_addr_into_implicit_ptr): New function.
24962 (resolve_addr_in_expr): Optimize DWARF location expression
24963 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
24964 which doesn't live in memory, but has DW_AT_location or
24965 DW_AT_const_value, or refers to a string literal, into
24966 DW_OP_GNU_implicit_pointer.
24967 (optimize_location_into_implicit_ptr): New function.
24968 (resolve_addr): If removing DW_AT_location of a variable because
24969 it was DW_OP_addr of address of the variable, but the variable doesn't
24970 live in memory, try to emit const value attribute for the initializer.
24972 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24974 * tree.h (VECTOR_TYPE_P): New macro.
24975 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
24976 TYPE_MODE): Use it.
24977 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
24978 VEC_COND_EXPR cannot be lvalues.
24979 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
24981 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24983 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
24984 Restrict the transformation to equal modes.
24986 2013-03-21 Richard Biener <rguenther@suse.de>
24988 PR tree-optimization/39326
24989 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
24990 (MEM_ANALYZABLE): Adjust.
24991 (record_mem_ref_loc): Move bitmap ops ...
24992 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
24993 unanalyzable refs, do not record locations for it.
24994 (analyze_memory_references): Allocate ref zero as shared
24996 (refs_independent_p): Do not test for unanalyzed mems here.
24997 (ref_indep_loop_p_1): Special-case disambiguation against
24998 the unanalyzed ref.
24999 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
25001 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
25003 * config/arm/arm-protos.h (tune_params): Add
25004 prefer_neon_for_64bits field.
25005 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
25006 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
25007 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
25008 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
25009 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
25010 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
25011 (arm_option_override): Handle -mneon-for-64bits new option.
25012 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
25013 (prefer_neon_for_64bits): Declare new variable.
25014 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
25015 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
25016 (arch_enabled): Handle new arch types. Remove support for onlya8
25018 (one_cmpldi2): Use new arch names.
25019 (zero_extend<mode>di2, extend<mode>di2): Ditto.
25020 * config/arm/arm.opt (mneon-for-64bits): Add option.
25021 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
25022 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
25023 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
25025 * doc/invoke.texi (-mneon-for-64bits): Document.
25027 2013-03-21 Richard Biener <rguenther@suse.de>
25029 PR tree-optimization/39326
25030 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
25031 (sort_bbs_in_loop_postorder_cmp): New function.
25032 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
25034 2013-03-21 Richard Biener <rguenther@suse.de>
25036 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
25037 (vect_insert_into_interleaving_chain): Likewise.
25038 (vect_drs_dependent_in_basic_block): Inline ...
25039 (vect_slp_analyze_data_ref_dependence): ... here. New function,
25041 (vect_analyze_data_ref_dependence): ... here. Simplify.
25042 (vect_check_interleaving): Simplify.
25043 (vect_analyze_data_ref_dependences): Likewise. Split out ...
25044 (vect_slp_analyze_data_ref_dependences): ... this new function.
25045 (dr_group_sort_cmp): New function.
25046 (vect_analyze_data_ref_accesses): Compute data-reference groups
25047 here instead of in vect_analyze_data_ref_dependence. Use
25048 a more efficient algorithm.
25049 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
25050 vect_slp_analyze_data_ref_dependences. Call
25051 vect_analyze_data_ref_accesses earlier.
25052 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
25053 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
25054 (vect_slp_analyze_data_ref_dependences): New prototype.
25056 2013-03-21 Richard Biener <rguenther@suse.de>
25058 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
25059 ref is stored in the loop.
25060 (find_refs_for_sm): Walk only over all stores.
25061 (store_motion_loop): Allocate from lim_bitmap_obstack.
25062 (store_motion): Likewise.
25064 2013-03-21 Richard Biener <rguenther@suse.de>
25066 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
25067 Update virtual SSA form.
25069 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25071 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
25072 * configure: Regenerate.
25073 * config.in: Regenerate.
25074 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
25075 if !HAVE_LD_EH_FRAME_CIEV3.
25077 2013-03-21 Richard Biener <rguenther@suse.de>
25079 * tree-cfg.c (verify_expr_no_block): New function.
25080 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
25081 nor DECL_VALUE_EXPR have locations with associated blocks.
25082 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
25083 (clear_unused_block_pointer): Remove code dealing with
25084 blocks in DECL_DEBUG_EXPR locations.
25086 2013-03-21 Richard Biener <rguenther@suse.de>
25088 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
25089 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
25090 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
25091 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
25092 instead of DECL_DEBUG_EXPR_IS_FROM.
25093 * gimplify.c (gimplify_modify_expr): Likewise.
25094 * tree-cfg.c (verify_expr_location_1): Likewise.
25095 * tree-complex.c (create_one_component_var): Likewise.
25096 * tree-sra.c (create_access_replacement): Likewise.
25097 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
25098 (clear_unused_block_pointer): Likewise.
25099 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
25100 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
25101 * var-tracking.c (var_debug_decl): Likewise.
25102 (track_expr_p): Likewise.
25103 * tree-inline.c (add_local_variables): Likewise. Set
25104 DECL_HAS_DEBUG_EXPR_P after copying it.
25105 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
25106 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
25108 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25111 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
25112 * configure: Regenerate.
25113 * config.in: Regenerate.
25114 * config/i386/i386.md (*movdf_internal): Use
25115 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25116 movd instead of movq mnemonic for interunit moves.
25117 (*movdi_internal): Ditto.
25119 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25121 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
25122 (abd<mode>_3): New pattern.
25123 (aba<mode>_3): New pattern.
25124 (fabd<mode>_3): New pattern.
25126 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25128 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
25129 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
25130 occurrence of REGISTER_PREFIX as its empty string.
25132 2013-03-20 Jeff Law <law@redhat.com>
25134 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
25135 addititional equivalences for equality comparisons between an SSA_NAME
25136 and a constant where the SSA_NAME was set from a widening conversion.
25138 2013-03-20 Walter Lee <walt@tilera.com>
25140 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
25142 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25144 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
25145 depending on type attribute.
25146 (*movti_internal): Ditto.
25147 (*movtf_internal): Ditto.
25148 (*movxf_internal): Ditto.
25149 (*movdf_internal): Ditto.
25150 (*movsf_internal): Ditto.
25152 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25154 * config/i386/i386.md (*movti_internal): Set prefix attribute to
25155 maybe_vex for sselog1 and ssemov types.
25156 (*movdi_internal): Reorder operand constraints.
25157 (*movsi_internal): Ditto. Set prefix attribute to
25158 maybe_vex for sselog1 and ssemov types.
25159 (*movtf_internal): Set prefix attribute to maybe_vex
25160 for sselog1 and ssemov types.
25161 (*movdf_internal): Ditto. Set prefix_data16 attribute for
25162 DImode ssemov types. Reorder operand constraints.
25163 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
25164 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
25165 attribute for SImode ssemov types. Reorder operand constraints.
25167 2013-03-20 Martin Jambor <mjambor@suse.cz>
25169 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
25170 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
25172 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
25174 * config/rs6000/predicates.md (indexed_address, update_address_mem
25175 update_indexed_address_mem): New predicates.
25176 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
25177 attribute for load/store instructions.
25178 * config/rs6000/dfp.md (movsd_store): Likewise.
25179 (movsd_load): Likewise.
25180 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
25181 (unnamed HI->DI extend define_insn): Likewise.
25182 (unnamed SI->DI extend define_insn): Likewise.
25183 (unnamed QI->SI extend define_insn): Likewise.
25184 (unnamed QI->HI extend define_insn): Likewise.
25185 (unnamed HI->SI extend define_insn): Likewise.
25186 (unnamed HI->SI extend define_insn): Likewise.
25187 (extendsfdf2_fpr): Likewise.
25188 (movsi_internal1): Likewise.
25189 (movsi_internal1_single): Likewise.
25190 (movhi_internal): Likewise.
25191 (movqi_internal): Likewise.
25192 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
25193 attribute for load/store instructions.
25194 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
25196 (mov<mode>_softfloat): Likewise.
25197 (mov<mode>_hardfloat32): Likewise.
25198 (mov<mode>_hardfloat64): Likewise.
25199 (mov<mode>_softfloat64): Likewise.
25200 (movdi_internal32): Likewise.
25201 (movdi_internal64): Likewise.
25202 (probe_stack_<mode>): Likewise.
25204 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
25206 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
25207 floating point, and decimal floating point to reload iterator.
25209 * config/rs6000/constraints.md (wl constraint): New constraints to
25210 return FLOAT_REGS if certain options are used to reduce the number
25211 of separate patterns that exist in the file.
25212 (wx constraint): Likewise.
25213 (wz constraint): Likewise.
25215 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25216 -mdebug=reg, print wg, wl, wx, and wz constraints.
25217 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25218 Initialize the reload functions for 64-bit binary/decimal floating
25220 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25221 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25222 create the buffer on the stack to overcome not having a 32-bit
25224 (rs6000_emit_move): Likewise.
25225 (rs6000_secondary_memory_needed_rtx): Likewise.
25226 (rs6000_alloc_sdmode_stack_slot): Likewise.
25227 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25228 via xxlxor, just like DFmode 0.0.
25230 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
25231 define as 1 if we are running on a power7 or newer.
25232 (enum r6000_reg_class_enum): Add new constraints.
25234 * config/rs6000/dfp.md (movsd): Delete, combine with binary
25235 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25236 with other moves by using conditional constraits (wg). Use LFIWZX
25237 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
25238 (movsd splitter): Likewise.
25239 (movsd_hardfloat): Likewise.
25240 (movsd_softfloat): Likewise.
25242 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
25243 binary and decimal floating point moves.
25244 (fmove_ok): New attributes to combine binary and decimal floating
25245 point moves, and to combine power6x (mfpgpr) moves along normal
25247 (real_value_to_target): Likewise.
25248 (f32_lr): Likewise.
25249 (f32_lm): Likewise.
25250 (f32_li): Likewise.
25251 (f32_sr): Likewise.
25252 (f32_sm): Likewise.
25253 (f32_si): Likewise.
25254 (movsf): Combine binary and decimal floating point moves. Combine
25255 power6x (mfpgpr) moves with other moves by using conditional
25256 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
25257 (mov<mode> for SFmode/SDmode); Likewise.
25258 (SFmode/SDmode splitters): Likewise.
25259 (movsf_hardfloat): Likewise.
25260 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
25261 (movsf_softfloat): Likewise.
25262 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
25264 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
25265 wx and wz constraints.
25267 * config/rs6000/constraints.md (wg constraint): New constraint to
25268 return FLOAT_REGS if -mmfpgpr (power6x) was used.
25270 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
25273 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25274 -mdebug=reg, print wg, wl, wx, and wz constraints.
25275 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25276 Initialize the reload functions for 64-bit binary/decimal floating
25278 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25279 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25280 create the buffer on the stack to overcome not having a 32-bit
25282 (rs6000_emit_move): Likewise.
25283 (rs6000_secondary_memory_needed_rtx): Likewise.
25284 (rs6000_alloc_sdmode_stack_slot): Likewise.
25285 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25286 via xxlxor, just like DFmode 0.0.
25288 * config/rs6000/dfp.md (movdd): Delete, combine with binary
25289 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25290 with other moves by using conditional constraits (wg). Use LFIWZX
25291 and STFIWX for loading SDmode on power7.
25292 (movdd splitters): Likewise.
25293 (movdd_hardfloat32): Likewise.
25294 (movdd_softfloat32): Likewise.
25295 (movdd_hardfloat64_mfpgpr): Likewise.
25296 (movdd_hardfloat64): Likewise.
25297 (movdd_softfloat64): Likewise.
25299 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
25300 64-bit binary and decimal floating point moves.
25301 (FMOVE64X): Likewise.
25302 (movdf): Combine 64-bit binary and decimal floating point moves.
25303 Combine power6x (mfpgpr) moves with other moves by using
25304 conditional constraits (wg).
25305 (mov<mode> for DFmode/DDmode): Likewise.
25306 (DFmode/DDmode splitters): Likewise.
25307 (movdf_hardfloat32): Likewise.
25308 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
25309 (movdf_softfloat32): Likewise.
25310 (movdf_hardfloat64_mfpgpr): Likewise.
25311 (movdf_hardfloat64): Likewise.
25312 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
25313 (movdf_softfloat64): Likewise.
25314 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
25315 (reload_<mode>_load): Move to later in the file so they aren't in
25316 the middle of the floating point move insns.
25317 (reload_<mode>_store): Likewise.
25319 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
25322 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
25323 constraint if -mdebug=reg.
25324 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
25325 Enable using dd reload support if needed.
25327 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
25328 binary and decimal floating point moves in rs6000.md.
25329 (movtd_internal): Likewise.
25331 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
25332 decimal floating point moves.
25334 (movtf_internal): Likewise.
25335 (mov<mode>_internal, TDmode/TFmode): Likewise.
25336 (movtf_softfloat): Likewise.
25337 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
25339 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
25340 movdi_internal64, using wg constraint for move direct operations.
25341 (movdi_internal64): Likewise.
25343 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
25344 MODES_TIEABLE_P for selected modes. Print the numerical value of
25345 the various virtual registers. Use GPR/FPR first/last values,
25346 instead of hard coding the register numbers. Print which modes
25347 have reload functions registered.
25348 (rs6000_option_override_internal): If -mdebug=reg, trace the options
25349 settings before/after setting cpu, target and subtarget settings.
25350 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
25351 and for secondary reload failures in rs6000_secondary_reload_inner.
25352 (rs6000_secondary_reload_fail): Likewise.
25353 (rs6000_secondary_reload_inner): Likewise.
25355 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
25356 macros for first/last GPR and FPR registers.
25357 (LAST_GPR_REGNO): Likewise.
25358 (FIRST_FPR_REGNO): Likewise.
25359 (LAST_FPR_REGNO): Likewise.
25361 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
25362 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
25363 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
25364 (vcond<mode><mode>): Likewise.
25365 (vcondu<mode><mode>): Likewise.
25366 (vector_gtu<mode>): Likewise.
25367 (vector_gte<mode>): Likewise.
25368 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
25369 to prevent the compiler from converting DImode operations to TImode.
25370 (ior<mode>3): Likewise.
25371 (and<mode>3): Likewise.
25372 (one_cmpl<mode>2): Likewise.
25373 (nor<mode>3): Likewise.
25374 (andc<mode>3): Likewise.
25376 * config/rs6000/constraints.md (wt constraint): New constraint
25377 that returns VSX_REGS if TImode is allowed in VSX registers.
25379 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
25380 constant under VSX.
25382 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
25383 similar to TImode, but it is restricted to being in the GPRs.
25385 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
25386 TImode to occupy a single VSX register.
25388 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
25389 -mvsx-timode for power7/power8.
25390 (power7 cpu): Likewise.
25391 (power8 cpu): Likewise.
25393 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
25394 sure that TFmode/TDmode take up two registers if they are ever
25395 allowed in the upper VSX registers.
25396 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
25398 (rs6000_init_hard_regno_mode_ok): Likewise.
25399 (rs6000_debug_reg_global): Add debugging for PTImode and wt
25400 constraint. Print if LRA is turned on.
25401 (rs6000_option_override_internal): Give an error if -mvsx-timode
25402 and VSX is not enabled.
25403 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
25404 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
25405 to reg+offset addressing. Use PTImode when checking offset
25406 addresses for validity.
25407 (reg_offset_addressing_ok_p): Likewise.
25408 (rs6000_legitimate_offset_address_p): Likewise.
25409 (rs6000_legitimize_address): Likewise.
25410 (rs6000_legitimize_reload_address): Likewise.
25411 (rs6000_legitimate_address_p): Likewise.
25412 (rs6000_eliminate_indexed_memrefs): Likewise.
25413 (rs6000_emit_move): Likewise.
25414 (rs6000_secondary_reload): Likewise.
25415 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
25416 reloads to fpr registers to continue to use reg+offset addressing,
25417 but 64-bit reloads to altivec registers need reg+reg addressing.
25418 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
25419 it. Treat LO_SUM like a PLUS operation.
25420 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
25421 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
25422 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
25423 registers to share a register with a smaller sized type, since VSX
25424 puts scalars in the upper 64-bits.
25425 (print_operand): Add support for PTImode.
25426 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
25427 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
25428 registers, but don't have arithmetic support.
25429 (rs6000_memory_move_cost): Add test for VSX.
25430 (rs6000_opt_masks): Add -mvsx-timode.
25432 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
25435 (VSr): Use wt constraint for TImode.
25436 (VSv): Drop TImode support.
25437 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
25438 (vsx_movti_64bit): Likewise.
25439 (vsx_movti_32bit): Likewise.
25440 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
25441 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
25442 one '?' on the appropriate output constraint. Do not allow TImode
25443 logical operations on 32-bit systems.
25444 (vsx_ior<mode>3): Likewise.
25445 (vsx_xor<mode>3): Likewise.
25446 (vsx_one_cmpl<mode>2): Likewise.
25447 (vsx_nor<mode>3): Likewise.
25448 (vsx_andc<mode>3): Likewise.
25449 (vsx_concat_<mode>): Likewise.
25450 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
25452 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
25453 OPTION_MASK_VSX_TIMODE.
25454 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
25455 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
25457 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
25458 (TI2 iterator): New iterator for TImode, PTImode.
25459 (wd mode attribute): Add values for vector types.
25460 (movti_string): Replace TI move operations with operations for TImode
25461 and PTImode. Add support for TImode being allowed in VSX registers.
25462 (mov<mode>_string, TImode/PTImode): Likewise.
25463 (movti_ppc64): Likewise.
25464 (mov<mode>_ppc64, TImode/PTImode): Likewise.
25465 (TI mode splitters): Likewise.
25467 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
25470 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
25472 PR tree-optimization/56355
25473 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
25474 Also handle integers with undefined overflow.
25476 2013-03-20 Catherine Moore <clm@codesourcery.com>
25477 Maciej W. Rozycki <macro@codesourcery.com>
25478 Tom de Vries <tom@codesourcery.com>
25479 Nathan Sidwell <nathan@codesourcery.com>
25480 Iain Sandoe <iain@codesourcery.com>
25481 Nathan Froyd <froydnj@codesourcery.com>
25482 Chao-ying Fu <fu@mips.com>
25484 * doc/extend.texi: (micromips, nomicromips, nocompression):
25485 Document new function attributes.
25486 * doc/invoke.texi (minterlink-compressed, mmicromips,
25487 m14k, m14ke, m14kec): Document new options.
25488 (minterlink-mips16): Update documentation.
25489 * doc/md.texi (ZC, ZD): Document new constraints.
25490 * configure.ac (gcc_cv_as_micromips): Check if linker
25491 supports the .set micromips directive.
25492 * configure: Regenerate.
25493 * config.in: Regenerate.
25494 * config/mips/mips-tables.opt: Regenerate.
25495 * config/mips/micromips.md: New file.
25496 * constraints.md (ZC, ZD): New constraints.
25497 * config/mips/predicates.md (movep_src_register): New predicate.
25498 (movep_src_operand): New predicate.
25499 (non_volatile_mem_operand): New predicate.
25500 * config/mips/mips.md (multimem): New type.
25501 (length): Differentiate between 17-bit and 18-bit branch offsets.
25502 (MOVEP1, MOVEP2): New mode iterator.
25503 (mov_<load>l): Use ZC constraint.
25504 (mov_<load>r): Likewise.
25505 (mov_<store>l): Likewise.
25506 (mov_<store>r): Likewise.
25507 (*branch_equality<mode>_inverted): Add microMIPS support.
25508 (*branch_equality<mode>): Likewise.
25509 (*jump_absolute): Likewise.
25510 (indirect_jump_<mode>): Likewise.
25511 (tablejump_<mode>): Likewise.
25512 (<optab>_internal): Likewise.
25513 (sibcall_internal): Likewise.
25514 (sibcall_value_internal): Likewise.
25515 (prefetch): Use constraint ZD.
25516 * config/mips/mips.opt (minterlink-compressed): New option.
25517 (minterlink-mips16): Now an alias for minterlink-compressed.
25518 (mmicromips): New option.
25519 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
25520 (compare_and_swap_12): Likewise.
25521 (sync_add<mode>): Likewise.
25522 (sync_<optab>_12): Likewise.
25523 (sync_old_<optab>_12): Likewise.
25524 (sync_new_<optab>_12): Likewise.
25525 (sync_nand_12): Likewise.
25526 (sync_old_nand_12): Likewise.
25527 (sync_new_nand_12): Likewise.
25528 (sync_sub<mode>): Likewise.
25529 (sync_old_add<mode>): Likewise.
25530 (sync_old_sub<mode>): Likewise.
25531 (sync_new_add<mode>): Likewise.
25532 (sync_new_sub<mode>): Likewise.
25533 (sync_<optab><mode>): Likewise.
25534 (sync_old_<optab><mode>): Likewise.
25535 (sync_new_<optab><mode>): Likewise.
25536 (sync_nand<mode>): Likewise.
25537 (sync_old_nand<mode>): Likewise.
25538 (sync_new_nand<mode>): Likewise.
25539 (sync_lock_test_and_set<mode>): Likewise.
25540 (test_and_set_12): Likewise.
25541 (atomic_compare_and_swap<mode>): Likewise.
25542 (atomic_exchange<mode>_llsc): Likewise.
25543 (atomic_fetch_add<mode>_llsc): Likewise.
25544 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
25545 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
25546 (umips_save_restore_pattern_p): Likewise.
25547 (umips_load_store_pair_p): Likewise.
25548 (umips_output_load_store_pair): Likewise.
25549 (umips_movep_target_p): Likewise.
25550 (umips_12bit_offset_address_p): Likewise.
25551 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
25552 (mips_base_mips16): Rename this...
25553 (mips_base_compression_flags): ...to this. Update all uses.
25554 (mips_attribute_table): Add micromips, nomicromips and nocompression.
25555 (mips_mips16_decl_p): Delete.
25556 (mips_nomips16_decl_p): Delete.
25557 (mips_get_compress_on_flags): New function.
25558 (mips_get_compress_off_flags): New function.
25559 (mips_get_compress_mode): New function.
25560 (mips_get_compress_on_name): New function.
25561 (mips_get_compress_off_name): New function.
25562 (mips_insert_attributes): Support multiple compression types.
25563 (mips_merge_decl_attributes): Likewise.
25564 (umips_12bit_offset_address_p): New function.
25565 (mips_start_function_definition): Emit .set micromips directive.
25566 (mips_call_may_need_jalx_p): New function.
25567 (mips_function_ok_for_sibcall): Add microMIPS support.
25568 (mips_print_operand_punctuation): Support short delay slots and
25570 (umips_swm_mask, umips_swm_encoding): New.
25571 (umips_build_save_restore): New function.
25572 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
25573 (was_mips16_p): Remove.
25574 (old_compression_mode): New.
25575 (mips_set_compression_mode): New function.
25576 (mips_set_current_function): Add microMIPS support.
25577 (mips_option_override): Likewise.
25578 (umips_save_restore_pattern_p): New function.
25579 (umips_output_save_restore): New function.
25580 (umips_load_store_pair_p_1): New function.
25581 (umips_load_store_pair_p): New function.
25582 (umips_output_load_store_pair_1): New function.
25583 (umips_output_load_store_pair): New function.
25584 (umips_movep_target_p) New function.
25585 (mips_prepare_pch_save): Add microMIPS support.
25586 * config/mips/mips.h (TARGET_COMPRESSION): New.
25587 (TARGET_CPU_CPP_BUILTINS): Update macro
25588 to use new compression flags and to support microMIPS.
25589 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
25590 (MIPS_ARCH_FLOAT_SPEC): Likewise.
25591 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
25592 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
25593 (ASM_SPEC): Support mmicromips and mno-micromips.
25594 (M16STORE_REG_P): New macro.
25595 (MIPS_CALL): Support TARGET_MICROMIPS.
25596 (MICROMIPS_J): New macro.
25597 (mips_base_mips16): Rename this...
25598 (mips_base_compression_flags): ...to this.
25599 (UMIPS_12BIT_OFFSET_P): New macro.
25600 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
25601 (MULTILIB_DIRNAMES): Likewise.
25602 2013-03-20 Richard Biener <rguenther@suse.de>
25604 PR tree-optimization/56661
25605 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
25606 the result does not have to be distinct.
25608 2013-03-20 Richard Biener <rguenther@suse.de>
25610 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
25613 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
25614 Steven Bosscher <steven@gcc.gnu.org>
25616 PR rtl-optimization/56605
25617 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
25619 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25622 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
25623 that require movd instead of movq.
25625 2013-03-20 Richard Biener <rguenther@suse.de>
25627 * tree-ssa-structalias.c (struct variable_info): Add pointer
25628 to the first field of an aggregate with sub-vars. Make
25629 this and the pointer to the next subfield its ID.
25630 (vi_next): New function.
25631 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
25632 storedanything_id, integer_id): Increment by one.
25633 (new_var_info, get_call_vi, lookup_call_clobber_vi,
25634 get_call_clobber_vi): Adjust.
25635 (solution_set_expand): Simplify and speedup.
25636 (solution_set_add): Inline into ...
25637 (set_union_with_increment): ... this. Adjust accordingly.
25638 (do_sd_constraint): Likewise.
25639 (do_ds_constraint): Likewise.
25640 (do_complex_constraint): Simplify.
25641 (build_pred_graph): Adjust.
25642 (solve_graph): Likewise. Simplify and speedup.
25643 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
25644 get_constraint_for_component_ref, get_constraint_for_1,
25645 first_vi_for_offset, first_or_preceding_vi_for_offset,
25646 create_function_info_for, create_variable_info_for_1,
25647 create_variable_info_for, intra_create_variable_infos): Adjust.
25648 (init_base_vars): Push NULL for ID zero.
25649 (compute_points_to_sets): Adjust.
25651 2013-03-20 Richard Biener <rguenther@suse.de>
25653 * cfgloop.c (verify_loop_structure): Streamline and avoid
25654 ICEing on corrupt loop tree.
25655 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
25658 2013-03-20 Richard Biener <rguenther@suse.de>
25660 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
25661 check whether an SSA update is needed.
25663 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
25665 * config/mips/constraints.md (T): Rename to...
25669 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
25670 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
25672 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25674 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
25675 (*subsi3_carryin_uxtw): Likewise.
25677 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25679 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
25680 (*rorsi3_insn_uxtw): Likewise.
25682 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25684 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
25685 (*extrsi5_insn_uxtw): Likewise.
25687 2013-03-19 Richard Biener <rguenther@suse.de>
25689 PR tree-optimization/56273
25690 * passes.c (init_optimization_passes): Move second VRP after DOM.
25692 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
25694 * config/i386/i386.md (*movti_internal): Merge from
25695 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
25696 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
25697 nox64 isa attributes.
25699 2013-03-18 Richard Biener <rguenther@suse.de>
25701 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
25703 (merge_node_constraints): Likewise.
25704 (build_succ_graph): Likewise.
25705 (valid_graph_edge): Inline into single caller.
25706 (unify_nodes): Likewise. Use bitmap_set_bit return value
25708 (scc_visit): Fix formatting and variable use.
25709 (do_sd_constraint): Use gcc_checking_assert.
25710 (do_ds_constraint): Likewise.
25711 (do_complex_constraint): Likewise.
25712 (condense_visit): Likewise. Cleanup.
25713 (dump_pred_graph): New function.
25714 (perform_var_substitution): Dump the pred-graph before
25715 variable substitution.
25716 (find_equivalent_node): Use gcc_checking_assert.
25717 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
25719 2013-03-18 Richard Biener <rguenther@suse.de>
25721 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
25722 Remove cond_expr_stmt_list argument and do not gimplify the
25724 (vect_loop_versioning): Adjust.
25725 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
25726 Cleanup to use less temporaries.
25727 (vect_create_data_ref_ptr): Cleanup.
25729 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25731 PR tree-optimization/56635
25732 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
25733 require types_compatible_p types.
25735 2013-03-18 Nick Clifton <nickc@redhat.com>
25737 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
25738 spurious backslash.
25740 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
25741 Add missing line to comment describing function.
25743 2013-03-18 Richard Biener <rguenther@suse.de>
25745 PR tree-optimization/56210
25746 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
25747 Handle string / character search functions.
25748 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
25750 2013-03-18 Richard Biener <rguenther@suse.de>
25752 PR middle-end/56483
25753 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
25754 and implement properly.
25755 * gimple.h (gimple_cond_single_var_p): Remove.
25757 2013-03-18 Richard Biener <rguenther@suse.de>
25759 * tree-data-ref.h (find_data_references_in_loop): Declare.
25760 * tree-data-ref.c (get_references_in_stmt): Use a stack
25761 vector pre-allocated in the callers.
25762 (find_data_references_in_stmt): Adjust.
25763 (graphite_find_data_references_in_stmt): Likewise.
25764 (create_rdg_vertices): Likewise.
25765 (find_data_references_in_loop): Export.
25766 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
25767 Compute dependences here...
25768 (vect_analyze_data_refs): ...not here. When we encounter
25769 a non-vectorizable data reference in basic-block vectorization
25770 truncate the data reference vector. Do not bother to
25771 fixup data-dependence information for gather loads.
25772 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
25773 of data references, as reported.
25775 2013-03-18 Richard Biener <rguenther@suse.de>
25777 PR tree-optimization/3713
25778 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
25779 has_constants and expr.
25780 (stmt_has_constants): Properly valueize SSA names when deciding
25781 whether the stmt has constants.
25783 2013-03-18 Richard Biener <rguenther@suse.de>
25785 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
25786 whole function when there is nothing to do.
25787 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
25788 * tree-vectorizer.c (vectorize_loops): Update virtual and
25789 loop-closed SSA once.
25790 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
25792 2013-03-18 Richard Biener <rguenther@suse.de>
25794 PR middle-end/56113
25795 * domwalk.c (bb_postorder): New global static.
25796 (cmp_bb_postorder): New function.
25797 (walk_dominator_tree): Replace scheme imposing an order for
25798 visiting dominator sons by one sorting them at the time they
25799 are pushed on the stack.
25801 2013-03-18 Richard Biener <rguenther@suse.de>
25803 PR tree-optimization/39326
25804 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
25805 (struct mem_ref): Replace mem member with ao_ref typed member.
25806 (MEM_ANALYZABLE): Adjust.
25807 (memref_eq): Likewise.
25808 (mem_ref_alloc): Likewise.
25809 (gather_mem_refs_stmt): Likewise.
25810 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
25811 (execute_sm_if_changed_flag_set): Adjust.
25812 (execute_sm): Likewise.
25813 (ref_always_accessed_p): Likewise.
25814 (refs_independent_p): Likewise.
25815 (can_sm_ref_p): Likewise.
25817 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25820 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
25821 return 1 even for !unsignedp.
25823 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25825 * config/i386/i386.md (isa): Add x64 and nox64.
25826 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
25827 (*pushtf): Enable *roF alternative for x64 isa only.
25828 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
25829 mode attribute of integer alternatives to DImode for TARGET_64BIT.
25830 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
25831 (*movtf_internal): Merge from *movtf_internal_rex64 and
25832 *movtf_internal_sse. Use x64 and nox64 isa attributes.
25833 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
25834 nox64 isa attributes.
25835 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
25836 nox64 isa attributes.
25837 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
25839 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25841 * config/alpha/alpha.c (TARGET_LRA_P): New define.
25843 2013-03-17 Jakub Jelinek <jakub@redhat.com>
25846 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
25847 class names. Remove trailing comma after "ALL_REGS".
25849 2013-03-16 Jan Hubicka <jh@suse.cz>
25851 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
25852 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
25853 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
25854 of cgraph_get_create_node.
25855 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
25857 2013-03-16 Jason Merrill <jason@redhat.com>
25860 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
25861 with DW_AT_default_value.
25863 2013-03-16 Jakub Jelinek <jakub@redhat.com>
25865 * BASE-VER: Set to 4.9.0.
25867 2013-03-14 Andi Kleen <ak@linux.intel.com>
25870 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
25871 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
25872 Document _x* TSX intrinsics.
25874 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
25875 David Holsgrove <david.holsgrove@xilinx.com>
25877 * configure.ac: Add MicroBlaze TLS support detection.
25878 * configure: Regenerate.
25879 * config/microblaze/microblaze-protos.h
25880 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
25881 symbol_mentioned_p, label_mentioned_p): Add prototypes.
25882 * config/microblaze/microblaze.c (microblaze_address_type): Add
25883 ADDRESS_TLS and tls_reloc address types.
25884 (microblaze_address_info): Add tls_reloc.
25885 (TARGET_HAVE_TLS): Define.
25886 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
25887 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
25888 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
25889 load_tls_operand, microblaze_call_tls_get_addr,
25890 microblaze_legitimize_tls_address): New functions.
25891 (microblaze_classify_unspec): Handle UNSPEC_TLS.
25892 (get_base_reg): Use microblaze_tls_symbol_p.
25893 (microblaze_classify_address): Handle TLS.
25894 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
25895 label_mentioned_p and microblaze_tls_referenced_p.
25896 (microblaze_legitimize_address): Handle TLS.
25897 (microblaze_address_insns): Handle ADDRESS_TLS.
25898 (pic_address_needs_scratch): Handle TLS.
25899 (print_operand_address): Handle TLS.
25900 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
25901 (microblaze_expand_move): Handle TLS.
25902 (microblaze_legitimate_constant_p): Check
25903 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
25904 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
25905 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
25906 (PIC_OFFSET_TABLE_REGNUM): Set.
25907 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
25908 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
25909 (addsi3, movsi_internal2, movdf_internal): Update constraints
25910 * config/microblaze/predicates.md (arith_plus_operand): Define
25911 (move_operand): Redefine as move_src_operand,
25912 check microblaze_tls_referenced_p.
25914 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25916 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
25917 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
25919 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25921 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
25924 2013-03-14 Jakub Jelinek <jakub@redhat.com>
25926 PR tree-optimization/53265
25927 * common.opt (Waggressive-loop-optimizations): New option.
25928 * tree-ssa-loop-niter.c: Include tree-pass.h.
25929 (do_warn_aggressive_loop_optimizations): New function.
25930 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
25931 if number_of_latch_executions returned constant.
25932 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
25933 early. If number_of_latch_executions returned constant, set
25934 nb_iterations_upper_bound back to it.
25935 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
25937 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
25938 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
25940 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
25941 (MULTILIB_OSDIRNAMES): Set.
25942 * genmultilib: If defaultosdirname doesn't start with :: , set
25943 defaultosdirname2 instead, clear it and emit two . multilib_raw
25944 entries instead of just one.
25946 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
25948 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
25949 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
25950 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
25951 (SUBTARGET_OVERRIDE_OPTIONS): New.
25953 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
25956 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
25957 (musermode): Convert to Var(TARGET_USERMODE).
25958 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
25959 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
25960 * config/sh/sh.c (sh_option_override): Use
25961 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
25962 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
25964 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
25966 (udivsi3_i4_single, divsi3_i4_single): Use
25967 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
25969 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
25971 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
25974 2013-03-13 Richard Biener <rguenther@suse.de>
25976 PR tree-optimization/56608
25977 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
25978 calls when vectorizing basic-blocks.
25980 2013-03-13 Jakub Jelinek <jakub@redhat.com>
25983 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
25986 2013-03-12 Jakub Jelinek <jakub@redhat.com>
25988 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
25990 2013-03-11 Jan Hubicka <jh@suse.cz>
25993 * lto-streamer-out.c (output_symbol_p): Skip references from
25994 constructors of external variables.
25996 2013-03-11 Jan Hubicka <jh@suse.cz>
25998 PR middle-end/56571
25999 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
26001 * emit-rtl.c (verify_rtx_sharing): Likewise.
26002 (copy_insn_1): Likewise.
26003 * rtl.c (copy_rtx): Likewise.
26005 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
26008 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
26009 output_operand_lossage message.
26011 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
26014 * arm.c (shift_op): Validate RTL pattern on the fly.
26015 (arm_print_operand, case 'S'): Don't use shift_operator to validate
26018 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26021 * config/pa/pa.md (call_value): Check for calls to powf and direct to
26022 new call patterns that clobber %fr12.
26023 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
26024 split and postreload patterns.
26025 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
26026 registers %fr12 and %fr12R as call used.
26028 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
26030 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
26031 (canon_address, record_store, replace_read, check_mem_read_rtx,
26032 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
26033 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
26034 rest_of_handle_dse): Likewise.
26036 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
26038 PR middle-end/56524
26039 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
26041 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
26042 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
26043 (save_optabs_if_changed): Replace with...
26044 (init_tree_optimization_optabs): ...this.
26045 * optabs.c (save_optabs_if_changed): Rename to...
26046 (init_tree_optimization_optabs): ...this. Take the optimization node
26047 as argument. Do nothing if the base optabs are already correct.
26048 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
26049 to recompute optabs.
26050 * function.h (function): Remove optabs field.
26051 * function.c (invoke_set_current_function_hook): Call
26052 init_tree_optimization_optabs. Use the result to initialize
26055 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
26057 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
26058 if GTMA_HAS_NO_INSTRUMENTATION.
26059 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
26060 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
26061 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
26062 * gimple-pretty-print.c (dump_gimple_transaction): Handle
26063 GTMA_HAS_NO_INSTRUMENTATION.
26065 2013-03-08 Jakub Jelinek <jakub@redhat.com>
26067 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
26070 2013-03-08 Marek Polacek <polacek@redhat.com>
26071 Jakub Jelinek <jakub@redhat.com>
26073 PR tree-optimization/56478
26074 * predict.c (is_comparison_with_loop_invariant_p): Change the
26075 type of loop_step to tree.
26076 (predict_loops): Adjust.
26077 (predict_iv_comparison): Perform the computations on double_ints.
26079 2013-03-08 Richard Biener <rguenther@suse.de>
26081 PR tree-optimization/56570
26082 * tree-cfg.c (verify_expr_location_1): Verify locations for
26084 * tree-sra.c (create_access_replacement): Strip locations
26085 from DECL_DEBUG_EXPRs.
26087 2013-03-08 Richard Biener <rguenther@suse.de>
26089 * tree-inline.c (expand_call_inline): Do not associate
26090 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
26091 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
26093 2013-03-08 Richard Biener <rguenther@suse.de>
26095 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
26096 or block changes with -Og. Fix for location / block encoding
26097 changes and PHI arguments with locations.
26099 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
26101 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
26103 (struct output_info): Likewise.
26104 (register_overhead): Remove bad gcc_assert.
26105 (bitmap_find_bit): If there is only a single bitmap element, do not
26106 count a miss as a search.
26107 (print_statistics): Update for counter type changes.
26108 (dump_bitmap_statistics): Likewise. Print headers such that they
26109 are properly lined up with the printed counters.
26111 2013-03-07 Jakub Jelinek <jakub@redhat.com>
26113 PR tree-optimization/56559
26114 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
26115 check that it has only a single use.
26117 2013-03-07 Richard Biener <rguenther@suse.de>
26119 * doc/invoke.texi (fwhole-program): Discourage use in combination
26122 2013-03-06 Jakub Jelinek <jakub@redhat.com>
26124 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
26126 PR tree-optimization/56539
26127 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
26128 instead of GSI_CONTINUE_LINKING as last argument to
26129 force_gimple_operand_gsi. Adjust function comment.
26131 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
26134 PR middle-end/56548
26135 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
26136 promoted mode, convert the result back to the original mode.
26138 2013-03-06 Richard Biener <rguenther@suse.de>
26140 PR middle-end/56294
26141 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
26142 (insert_updated_phi_nodes_compare_uids): New function.
26143 (update_ssa): Sort symbols_to_rename after UID before
26144 traversing it to insert PHI nodes.
26146 2013-03-06 Richard Biener <rguenther@suse.de>
26148 PR middle-end/50494
26149 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
26150 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
26153 2013-02-13 Richard Biener <rguenther@suse.de>
26156 * varasm.c (output_constant_def_1): Get the decl representing
26157 the constant as argument.
26158 (output_constant_def): Wrap output_constant_def_1.
26159 (make_decl_rtl): Use output_constant_def_1 with the decl
26160 representing the constant.
26161 (build_constant_desc): Optionally re-use a decl already
26162 representing the constant.
26163 (tree_output_constant_def): Adjust.
26165 2013-03-06 Joey Ye <joey.ye@arm.com>
26168 * gcc.c (convert_white_space): New function.
26169 (main): Handles white space in function name.
26171 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
26174 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
26175 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
26176 to SH_DIV_CALL_TABLE for TARGET_SH2.
26177 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
26179 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
26180 call-table options.
26182 2013-03-05 Sterling Augustine <saugustine@google.com>
26183 Cary Coutant <ccoutant@google.com>
26186 * dwarf2out.c (resolve_addr): Don't call
26187 remove_loc_list_addr_table_entries a second time for the same
26190 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26193 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
26194 (avoid_complex_debug_insns): New function.
26195 (expand_debug_locations): Call it.
26197 PR rtl-optimization/56484
26198 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
26199 lifetimes of hard registers on small register class machines.
26201 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26203 * config/microblaze/microblaze-protos.h: Rename
26204 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
26205 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
26207 (microblaze_fast_interrupt_function_p): New function.
26208 (microblaze_is_interrupt_handler): Rename to
26209 microblaze_is_interrupt_variant and add fast_interrupt check.
26210 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
26211 (save_restore_insns): Likewise.
26212 (compute_frame_size): Likewise.
26213 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
26214 (microblaze_globalize_label): Likewise.
26215 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
26216 * config/microblaze/microblaze.md: Use wrapper
26217 microblaze_is_interrupt_variant.
26219 2013-03-05 Kai Tietz <ktietz@redhat.com>
26221 * sdbout.c (sdbout_one_type): Switch to current function's section
26222 supporting cold/hot.
26224 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26226 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
26229 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26231 PR middle-end/56461
26232 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
26233 if VALGRIND_GET_VBITS is defined, temporarily make object
26234 memory all defined, and restore previous valgrind addressability
26235 and definability afterwards. Free this_object at the end.
26237 PR middle-end/56461
26238 * lra.c (lra): Call lra_clear_live_ranges if live_p,
26239 right before calling lra_create_live_ranges, also call it
26240 when clearing live_p. Only call lra_clear_live_ranges
26241 at the end if live_p.
26243 PR middle-end/56461
26244 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
26246 2013-03-05 Richard Biener <rguenther@suse.de>
26248 PR tree-optimization/56521
26249 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
26252 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
26255 * except.h (remove_unreachable_eh_regions): New prototype.
26256 * except.c (remove_eh_handler_splicer): New function, split out
26257 of remove_eh_handler.
26258 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
26259 warning about running it on many EH regions one at a time.
26260 (remove_unreachable_eh_regions_worker): New function, walk the
26261 EH tree in depth-first order and remove non-marked regions.
26262 (remove_unreachable_eh_regions): New function.
26263 * tree-eh.c (mark_reachable_handlers): New function, split out
26264 from remove_unreachable_handlers.
26265 (remove_unreachable_handlers): Use mark_reachable_handlers and
26266 remove_unreachable_eh_regions.
26267 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
26268 and remove_unreachable_eh_regions.
26270 2013-03-05 Richard Biener <rguenther@suse.de>
26272 PR middle-end/56525
26273 * loop-init.c (fix_loop_structure): Remove loops in two stages,
26274 not freeing them until the end.
26276 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26278 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
26280 2013-03-05 Richard Biener <rguenther@suse.de>
26282 PR tree-optimization/56270
26283 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
26284 of loads after scheduling an SLP instance.
26286 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26288 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
26290 (check_gcc_parallelize): Run guality.exp as a separate job from
26291 vect.exp with unsorted.exp and $(dg_target_exps) separately from
26292 struct-layout-1.exp with stackalign.exp.
26294 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
26296 PR middle-end/56461
26297 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
26298 load_index sbitmap even if some bit in it isn't set.
26300 PR middle-end/56461
26301 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
26302 (discover_iteration_bound_by_body_walk): Change queues to
26303 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
26304 spelling in comment. Call safe_push on queues[bound_index] directly.
26305 Release queues[queue_index] in every iteration unconditionally.
26306 Release bounds vector.
26308 PR middle-end/56461
26309 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
26310 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
26311 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
26314 2013-03-05 Richard Biener <rguenther@suse.de>
26317 * tree-inline.c (remap_blocks_to_null): New function.
26318 (expand_call_inline): When expanding a call stmt without
26319 an associated block inline remap all callee blocks to NULL.
26321 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26323 PR rtl-optimization/56494
26324 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
26325 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
26326 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
26328 PR middle-end/56461
26329 * sel-sched-ir.c (free_sched_pools): Release
26330 succs_info_pool.stack[succs_info_pool.max_top] vectors too
26331 if succs_info_pool.max_top isn't -1.
26334 * opts.c (opts_obstack, opts_concat): Moved to...
26335 * opts-common.c (opts_obstack, opts_concat): ... here.
26337 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26339 PR middle-end/56461
26340 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
26342 2013-03-04 Martin Jambor <mjambor@suse.cz>
26344 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
26345 all appropriate places.
26347 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
26349 PR tree-optimization/56424
26350 * ipa-split.c (split_function): Do not set the RSO flag if result is
26351 not by reference and its type is a register type.
26353 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
26355 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
26356 (microblaze_legitimate_pic_operand): Likewise
26357 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
26358 new function microblaze_legitimate_pic_operand
26359 * config/microblaze/microblaze-protos.h
26360 (microblaze_legitimate_pic_operand): Declare.
26362 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26364 * config/microblaze/predicates.md (call_insn_simple_operand):
26365 New predicate for supported rtx code types.
26366 * config/microblaze/microblaze.md (call_internal1): Use
26367 call_insn_simple_operand predicate.
26369 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26371 PR middle-end/56461
26372 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
26373 partitions.ordered_remove.
26375 PR middle-end/56461
26376 * tree-vect-stmts.c (vectorizable_conversion): Don't call
26377 vec_oprnds0.create (1) for modifier == NONE.
26379 PR middle-end/56461
26380 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
26381 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
26382 vec_oprnds1 right before pushing anything to it for
26385 PR middle-end/56461
26386 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
26387 set nbbs to 0 instead of having separate code path.
26388 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
26389 instead of false as last argument if returning NULL.
26391 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
26393 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
26394 the attribute is now called "target" instead of "option".
26395 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
26396 * doc/tm.texi.in (Target Attributes): Likewise document the correct
26397 attribute/pragma name for TARGET_OPTION_VALID_P and
26398 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
26399 * doc/tm.texi: Regenerated.
26401 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
26403 * config/microblaze/microblaze.c:
26404 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
26405 * config/microblaze/microblaze.h: Add -mxl-reorder to
26407 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
26408 instructions emitted if TARGET_REORDER.
26409 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
26410 or 0 for -m/-mno case, but initialises as 2 to detect default use case
26413 2013-03-01 Xinliang David Li <davidxl@google.com>
26415 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
26418 2013-03-01 Jakub Jelinek <jakub@redhat.com>
26420 PR middle-end/56461
26421 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
26422 vector even when returning true. Fix up function comment formatting.
26424 PR middle-end/56461
26425 * ira-build.c (ira_loop_nodes_count): New variable.
26426 (create_loop_tree_nodes): Initialize it.
26427 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
26429 PR middle-end/56461
26430 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
26431 method on dr_chain and result_chain.
26432 * tree-vect-stmts.c (vectorizable_store): Only call
26433 result_chain.create if j == 0.
26435 PR middle-end/56461
26436 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
26437 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
26438 before overwriting it.
26440 2013-03-01 Tobias Burnus <burnus@net-b.de>
26442 * doc/extended.texi (C Extensions): Change order in @menu
26444 (Other MIPS Built-in Functions): Move last MIPS entry before
26445 "picoChip Built-in Functions".
26446 (SH Built-in Functions): Move after RX Built-in Functions.
26447 * doc/gcc.texi (Introduction): Change order in @menu
26449 * doc/md.texi (Constraints): Ditto.
26450 * gty.texi (Type Information): Ditto.
26451 (User-provided marking routines for template types): Make
26453 * doc/invoke.texi (AArch64 Options): Move before
26454 "Adapteva Epiphany Options".
26456 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
26457 Jakub Jelinek <jakub@redhat.com>
26460 * asan.c (gate_asan): Lookup no_sanitize_address instead of
26461 no_address_safety_analysis attribute.
26462 * doc/extend.texi (no_address_safety_attribute): Rename to
26463 no_sanitize_address attribute, mention no_address_safety_analysis
26464 attribute as deprecated alias.
26466 2013-02-28 Jakub Jelinek <jakub@redhat.com>
26468 PR middle-end/56461
26469 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
26470 type to vec<vec<tree> > *.
26471 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
26472 to be vec<tree> instead of vec<tree> *, set vec_defs
26473 to vNULL and call vec_defs.create (number_of_vects), adjust other
26475 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
26476 vectorizable_condition): Adjust vect_get_slp_defs callers.
26478 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26480 * config/aarch64/aarch64.c
26481 (aarch64_float_const_representable): Remove unused variable.
26483 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26485 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
26487 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26489 * config/aarch64/aarch64-builtins.c
26490 (aarch64_init_simd_builtins): Make static.
26492 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26494 * config/aarch64/aarch64.c
26495 (aarch64_simd_make_constant): Make static.
26497 2013-02-28 Martin Jambor <mjambor@suse.cz>
26499 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
26500 with no initialization to the RHS of debug statements.
26502 2013-02-28 Martin Jambor <mjambor@suse.cz>
26504 PR tree-optimization/56294
26505 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
26507 (get_access_replacement): Do not call create_access_replacement.
26508 Assert a replacement exists.
26509 (get_repl_default_def_ssa_name): Create the replacement declaration
26512 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26514 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
26515 final_end_function.
26517 2013-02-28 Marek Polacek <polacek@redhat.com>
26519 PR rtl-optimization/56466
26520 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
26521 if we're changing a loop.
26522 (peel_loops_completely): Likewise.
26524 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
26527 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
26529 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26532 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
26533 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
26534 INTX_FTYPE_FX, FX_FTYPE_INTX.
26535 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
26537 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26539 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
26540 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
26541 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
26542 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
26543 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
26544 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
26545 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
26546 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
26547 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
26548 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
26549 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
26550 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
26551 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
26552 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
26553 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
26554 (avrxmega6): Increase max flash segments from 5 to 6.
26555 * config/avr/t-multilib: Regenerate.
26556 * config/avr/avr-tables.opt: Regenerate.
26557 * doc/avr-mmcu.texi: Regenerate.
26559 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26561 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
26562 (avr_device_to_arch): Rename to avr_device_to_ld.
26563 (avr_device_to_as): New prototype.
26564 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
26565 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
26566 * config/avr/driver-avr.c (avr_device_to_as): New.
26567 (avr_device_to_arch): Rename to avr_device_to_ld.
26569 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26571 PR middle-end/56461
26572 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
26573 method on dr_chain and result_chain.
26575 PR middle-end/56461
26576 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
26577 pointer_set_destroy on not_executed_last_iteration.
26579 PR middle-end/56461
26580 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
26582 PR middle-end/56461
26583 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
26584 FOR_EACH_DEFINED_FUNCTION when freeing state.
26586 PR middle-end/56461
26587 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
26589 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
26592 PR middle-end/56461
26593 * ipa-cp.c (decide_whether_version_node): Call vec_free on
26594 known_aggs[i].items and release known_aggs vector.
26596 PR middle-end/56461
26597 * ipa-reference.c (propagate): Free node_info even for alias nodes.
26599 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26601 * config/microblaze/microblaze.c (microblaze_emit_compare):
26602 Use xor for EQ/NE comparisions.
26603 * config/microblaze/microblaze.md (cstoresf4): Add constraints
26604 (cbranchsf4): Adjust operator to comparison_operator.
26606 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26608 PR middle-end/56461
26609 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
26611 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
26612 vec_safe_push, always update *slot.
26613 (redirect_edge_var_map_clear): Use vec_free.
26614 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
26615 (free_var_map_entry): Use vec_free.
26616 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
26617 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
26619 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
26621 PR middle-end/45472
26622 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
26623 when the may_trap_p bit of the exprs being merged differs.
26624 Reorder tests for speculativeness in the logical and operator.
26626 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26628 * incpath.c (add_standard_paths): Use reconcat instead of concat
26629 where appropriate and avoid leaking memory.
26631 * opts.h: Include obstack.h.
26632 (opts_concat): New prototype.
26633 (opts_obstack): New declaration.
26634 * opts.c (opts_concat): New function.
26635 (opts_obstack): New variable.
26636 (init_options_struct): Call gcc_init_obstack on opts_obstack.
26637 (finish_options): Use opts_concat instead of concat
26638 and XOBNEWVEC instead of XNEWVEC.
26639 * opts-common.c (generate_canonical_option, decode_cmdline_option,
26640 generate_option): Likewise.
26641 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
26642 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
26645 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
26646 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
26648 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26650 PR middle-end/56461
26651 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
26653 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26655 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
26656 (arm_block_move_unaligned_straight): Likewise.
26657 (arm_adjust_block_mem): Likewise.
26659 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26662 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
26663 temp, cond and label.
26664 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
26667 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
26668 * config/c6x/c6x.h (dbx_register_map): Update declaration.
26671 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
26672 of prologue/epilogue functions.
26675 * config/tilegx/tilegx.c (tilegx_expand_prologue):
26676 Remove unused variable cfa_offset.
26677 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
26680 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
26681 type promotion to unsigned.
26684 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
26685 for HOST_WIDE_INT of 32 bit / same size as int.
26686 (arm_block_move_unaligned_straight): Likewise.
26687 (arm_adjust_block_mem): Likewise.
26690 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
26693 2013-02-26 Marek Polacek <polacek@redhat.com>
26695 PR tree-optimization/56426
26696 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
26698 2013-02-26 Richard Biener <rguenther@suse.de>
26701 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
26702 unused variable loops.
26704 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26706 PR tree-optimization/56448
26707 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
26708 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
26709 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
26710 later operands of the references, or even first operand for
26711 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
26713 PR tree-optimization/56443
26714 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
26715 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
26716 to type_for_mode langhook.
26718 2013-02-25 Matt Turner <mattst88@gmail.com>
26720 * doc/invoke.texi: Document r4700.
26722 2013-02-25 Richard Biener <rguenther@suse.de>
26724 PR tree-optimization/56175
26725 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
26727 (simplify_bitwise_binary): ... here. Also guard the conversion
26728 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
26730 2013-02-25 Catherine Moore <clm@codesourcery.com>
26733 2013-02-24 Catherine Moore <clm@codesourcery.com>
26734 Maciej W. Rozycki <macro@codesourcery.com>
26735 Tom de Vries <tom@codesourcery.com>
26736 Nathan Sidwell <nathan@codesourcery.com>
26737 Iain Sandoe <iain@codesourcery.com>
26738 Nathan Froyd <froydnj@codesourcery.com>
26739 Chao-ying Fu <fu@mips.com>
26741 * doc/extend.texi: (micromips, nomicromips, nocompression):
26742 Document new function attributes.
26743 * doc/invoke.texi (minterlink-compressed, mmicromips,
26744 m14k, m14ke, m14kec): Document new options.
26745 (minterlink-mips16): Update documentation.
26746 * doc/md.texi (ZC, ZD): Document new constraints.
26747 * configure.ac (gcc_cv_as_micromips): Check if linker
26748 supports the .set micromips directive.
26749 * configure: Regenerate.
26750 * config.in: Regenerate.
26751 * config/mips/mips-tables.opt: Regenerate.
26752 * config/mips/micromips.md: New file.
26753 * constraints.md (ZC, AD): New constraints.
26754 * config/mips/predicates.md (movep_src_register): New predicate.
26755 (movep_src_operand): New predicate.
26756 (non_volatile_mem_operand): New predicate.
26757 * config/mips/mips.md (multimem): New type.
26758 (length): Differentiate between 17-bit and 18-bit branch offsets.
26759 (MOVEP1, MOVEP2): New mode iterator.
26760 (mov_<load>l): Use ZC constraint.
26761 (mov_<load>r): Likewise.
26762 (mov_<store>l): Likewise.
26763 (mov_<store>r): Likewise.
26764 (*branch_equality<mode>_inverted): Add microMIPS support.
26765 (*branch_equality<mode>): Likewise.
26766 (*jump_absolute): Likewise.
26767 (indirect_jump_<mode>): Likewise.
26768 (tablejump_<mode>): Likewise.
26769 (<optab>_internal): Likewise.
26770 (sibcall_internal): Likewise.
26771 (sibcall_value_internal): Likewise.
26772 (prefetch): Use constraint ZD.
26773 * config/mips/mips.opt (minterlink-compressed): New option.
26774 (minterlink-mips16): Now an alias for minterlink-compressed.
26775 (mmicromips): New option.
26776 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26777 (compare_and_swap_12): Likewise.
26778 (sync_add<mode>): Likewise.
26779 (sync_<optab>_12): Likewise.
26780 (sync_old_<optab>_12): Likewise.
26781 (sync_new_<optab>_12): Likewise.
26782 (sync_nand_12): Likewise.
26783 (sync_old_nand_12): Likewise.
26784 (sync_new_nand_12): Likewise.
26785 (sync_sub<mode>): Likewise.
26786 (sync_old_add<mode>): Likewise.
26787 (sync_old_sub<mode>): Likewise.
26788 (sync_new_add<mode>): Likewise.
26789 (sync_new_sub<mode>): Likewise.
26790 (sync_<optab><mode>): Likewise.
26791 (sync_old_<optab><mode>): Likewise.
26792 (sync_new_<optab><mode>): Likewise.
26793 (sync_nand<mode>): Likewise.
26794 (sync_old_nand<mode>): Likewise.
26795 (sync_new_nand<mode>): Likewise.
26796 (sync_lock_test_and_set<mode>): Likewise.
26797 (test_and_set_12): Likewise.
26798 (atomic_compare_and_swap<mode>): Likewise.
26799 (atomic_exchange<mode>_llsc): Likewise.
26800 (atomic_fetch_add<mode>_llsc): Likewise.
26801 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26802 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26803 (umips_save_restore_pattern_p): Likewise.
26804 (umips_load_store_pair_p): Likewise.
26805 (umips_output_load_store_pair): Likewise.
26806 (umips_movep_target_p): Likewise.
26807 (umips_12bit_offset_address_p): Likewise.
26808 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26809 (mips_base_mips16): Rename this...
26810 (mips_base_compression_flags): ...to this. Update all uses.
26811 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26812 (mips_mips16_decl_p): Delete.
26813 (mips_nomips16_decl_p): Delete.
26814 (mips_get_compress_on_flags): New function.
26815 (mips_get_compress_off_flags): New function.
26816 (mips_get_compress_mode): New function.
26817 (mips_get_compress_on_name): New function.
26818 (mips_get_compress_off_name): New function.
26819 (mips_insert_attributes): Support multiple compression types.
26820 (mips_merge_decl_attributes): Likewise.
26821 (umips_12bit_offset_address_p): New function.
26822 (mips_start_function_definition): Emit .set micromips directive.
26823 (mips_call_may_need_jalx_p): New function.
26824 (mips_function_ok_for_sibcall): Add microMIPS support.
26825 (mips_print_operand_punctuation): Support short delay slots and
26827 (umips_swm_mask, umips_swm_encoding): New.
26828 (umips_build_save_restore): New function.
26829 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26830 (was_mips16_p): Remove.
26831 (old_compression_mode): New.
26832 (mips_set_compression_mode): New function.
26833 (mips_set_current_function): Add microMIPS support.
26834 (mips_option_override): Likewise.
26835 (umips_save_restore_pattern_p): New function.
26836 (umips_output_save_restore): New function.
26837 (umips_load_store_pair_p_1): New function.
26838 (umips_load_store_pair_p): New function.
26839 (umips_output_load_store_pair_1): New function.
26840 (umips_output_load_store_pair): New function.
26841 (umips_movep_target_p) New function.
26842 (mips_prepare_pch_save): Add microMIPS support.
26843 * config/mips/mips.h (TARGET_COMPRESSION): New.
26844 (TARGET_CPU_CPP_BUILTINS): Update macro
26845 to use new compression flags and to support microMIPS.
26846 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26847 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26848 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26849 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26850 (ASM_SPEC): Support mmicromips and mno-micromips.
26851 (M16STORE_REG_P): New macro.
26852 (MIPS_CALL): Support TARGET_MICROMIPS.
26853 (MICROMIPS_J): New macro.
26854 (mips_base_mips16): Rename this...
26855 (mips_base_compression_flags): ...to this.
26856 (UMIPS_12BIT_OFFSET_P): New macro.
26857 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26858 (MULTILIB_DIRNAMES): Likewise.
26860 2013-02-25 Tom de Vries <tom@codesourcery.com>
26862 PR rtl-optimization/56131
26863 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
26864 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
26865 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
26867 2013-02-25 Tobias Burnus <burnus@net-b.de>
26869 * doc/invoke.texi (-fsanitize=): Move from optimization
26870 to debugging options.
26872 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26874 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
26876 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26877 Alexander Monakov <amonakov@ispras.ru>
26879 PR middle-end/56077
26880 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
26881 flush pending lists also on non-jumps. Adjust comment.
26883 2013-02-24 Catherine Moore <clm@codesourcery.com>
26884 Maciej W. Rozycki <macro@codesourcery.com>
26885 Tom de Vries <tom@codesourcery.com>
26886 Nathan Sidwell <nathan@codesourcery.com>
26887 Iain Sandoe <iain@codesourcery.com>
26888 Nathan Froyd <froydnj@codesourcery.com>
26889 Chao-ying Fu <fu@mips.com>
26891 * doc/extend.texi: (micromips, nomicromips, nocompression):
26892 Document new function attributes.
26893 * doc/invoke.texi (minterlink-compressed, mmicromips,
26894 m14k, m14ke, m14kec): Document new options.
26895 (minterlink-mips16): Update documentation.
26896 * doc/md.texi (ZC, ZD): Document new constraints.
26897 * configure.ac (gcc_cv_as_micromips): Check if linker
26898 supports the .set micromips directive.
26899 * configure: Regenerate.
26900 * config.in: Regenerate.
26901 * config/mips/mips-tables.opt: Regenerate.
26902 * config/mips/micromips.md: New file.
26903 * constraints.md (ZC, AD): New constraints.
26904 * config/mips/predicates.md (movep_src_register): New predicate.
26905 (movep_src_operand): New predicate.
26906 (non_volatile_mem_operand): New predicate.
26907 * config/mips/mips.md (multimem): New type.
26908 (length): Differentiate between 17-bit and 18-bit branch offsets.
26909 (MOVEP1, MOVEP2): New mode iterator.
26910 (mov_<load>l): Use ZC constraint.
26911 (mov_<load>r): Likewise.
26912 (mov_<store>l): Likewise.
26913 (mov_<store>r): Likewise.
26914 (*branch_equality<mode>_inverted): Add microMIPS support.
26915 (*branch_equality<mode>): Likewise.
26916 (*jump_absolute): Likewise.
26917 (indirect_jump_<mode>): Likewise.
26918 (tablejump_<mode>): Likewise.
26919 (<optab>_internal): Likewise.
26920 (sibcall_internal): Likewise.
26921 (sibcall_value_internal): Likewise.
26922 (prefetch): Use constraint ZD.
26923 * config/mips/mips.opt (minterlink-compressed): New option.
26924 (minterlink-mips16): Now an alias for minterlink-compressed.
26925 (mmicromips): New option.
26926 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26927 (compare_and_swap_12): Likewise.
26928 (sync_add<mode>): Likewise.
26929 (sync_<optab>_12): Likewise.
26930 (sync_old_<optab>_12): Likewise.
26931 (sync_new_<optab>_12): Likewise.
26932 (sync_nand_12): Likewise.
26933 (sync_old_nand_12): Likewise.
26934 (sync_new_nand_12): Likewise.
26935 (sync_sub<mode>): Likewise.
26936 (sync_old_add<mode>): Likewise.
26937 (sync_old_sub<mode>): Likewise.
26938 (sync_new_add<mode>): Likewise.
26939 (sync_new_sub<mode>): Likewise.
26940 (sync_<optab><mode>): Likewise.
26941 (sync_old_<optab><mode>): Likewise.
26942 (sync_new_<optab><mode>): Likewise.
26943 (sync_nand<mode>): Likewise.
26944 (sync_old_nand<mode>): Likewise.
26945 (sync_new_nand<mode>): Likewise.
26946 (sync_lock_test_and_set<mode>): Likewise.
26947 (test_and_set_12): Likewise.
26948 (atomic_compare_and_swap<mode>): Likewise.
26949 (atomic_exchange<mode>_llsc): Likewise.
26950 (atomic_fetch_add<mode>_llsc): Likewise.
26951 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26952 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26953 (umips_save_restore_pattern_p): Likewise.
26954 (umips_load_store_pair_p): Likewise.
26955 (umips_output_load_store_pair): Likewise.
26956 (umips_movep_target_p): Likewise.
26957 (umips_12bit_offset_address_p): Likewise.
26958 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26959 (mips_base_mips16): Rename this...
26960 (mips_base_compression_flags): ...to this. Update all uses.
26961 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26962 (mips_mips16_decl_p): Delete.
26963 (mips_nomips16_decl_p): Delete.
26964 (mips_get_compress_on_flags): New function.
26965 (mips_get_compress_off_flags): New function.
26966 (mips_get_compress_mode): New function.
26967 (mips_get_compress_on_name): New function.
26968 (mips_get_compress_off_name): New function.
26969 (mips_insert_attributes): Support multiple compression types.
26970 (mips_merge_decl_attributes): Likewise.
26971 (umips_12bit_offset_address_p): New function.
26972 (mips_start_function_definition): Emit .set micromips directive.
26973 (mips_call_may_need_jalx_p): New function.
26974 (mips_function_ok_for_sibcall): Add microMIPS support.
26975 (mips_print_operand_punctuation): Support short delay slots and
26977 (umips_swm_mask, umips_swm_encoding): New.
26978 (umips_build_save_restore): New function.
26979 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26980 (was_mips16_p): Remove.
26981 (old_compression_mode): New.
26982 (mips_set_compression_mode): New function.
26983 (mips_set_current_function): Add microMIPS support.
26984 (mips_option_override): Likewise.
26985 (umips_save_restore_pattern_p): New function.
26986 (umips_output_save_restore): New function.
26987 (umips_load_store_pair_p_1): New function.
26988 (umips_load_store_pair_p): New function.
26989 (umips_output_load_store_pair_1): New function.
26990 (umips_output_load_store_pair): New function.
26991 (umips_movep_target_p) New function.
26992 (mips_prepare_pch_save): Add microMIPS support.
26993 * config/mips/mips.h (TARGET_COMPRESSION): New.
26994 (TARGET_CPU_CPP_BUILTINS): Update macro
26995 to use new compression flags and to support microMIPS.
26996 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26997 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26998 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26999 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27000 (ASM_SPEC): Support mmicromips and mno-micromips.
27001 (M16STORE_REG_P): New macro.
27002 (MIPS_CALL): Support TARGET_MICROMIPS.
27003 (MICROMIPS_J): New macro.
27004 (mips_base_mips16): Rename this...
27005 (mips_base_compression_flags): ...to this.
27006 (UMIPS_12BIT_OFFSET_P): New macro.
27007 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27008 (MULTILIB_DIRNAMES): Likewise.
27010 2013-02-24 Jakub Jelinek <jakub@redhat.com>
27013 * target-globals.c (save_target_globals): For init_reg_sets and
27014 target_reinit remporarily set this_fn_optabs to this_target_optabs.
27016 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
27018 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
27019 * config/aarch64/t-aarch64
27020 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
27022 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
27024 PR inline-asm/56148
27025 * lra-constraints.c (process_alt_operands): Reload operand
27026 conflicting with earlier clobber only if no more other conflicting
27029 2013-02-22 Jakub Jelinek <jakub@redhat.com>
27032 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
27033 if not linking a shared library.
27035 2013-02-22 Seth LaForge <sethml@google.com>
27037 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
27039 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
27041 * config/arm/arm.md (split for extendsidi): Update condition.
27042 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
27043 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
27044 (qhs_zextenddi_cstr): Likewise.
27046 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27048 PR middle-end/56420
27049 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
27050 avoid signed wrapping.
27051 (expand_mult): Handle properly multiplication by
27052 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
27053 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
27054 in the compiler if coeff is HOST_WIDE_INT_MIN.
27055 (expand_divmod): Don't make ext_op1 static, change it's type to
27056 uhwi. Avoid undefined behavior in -INTVAL (op1).
27058 PR rtl-optimization/50339
27059 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
27061 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
27062 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
27063 into splitting_ashiftrt field.
27064 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
27066 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
27069 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
27071 PR middle-end/56108
27072 * trans-mem.c (execute_tm_mark): Do not expand transactions that
27073 are sure to go irrevocable.
27075 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
27077 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
27078 scalars are valid operands.
27080 2013-02-21 Martin Jambor <mjambor@suse.cz>
27082 PR tree-optimization/56310
27083 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
27084 only matching indices and non-negative final offsets.
27085 (intersect_aggregates_with_edge): Pass src_idx to
27086 agg_replacements_to_vector. Pass src_idx insstead of index to
27087 intersect_with_agg_replacements.
27089 2013-02-21 Martin Jambor <mjambor@suse.cz>
27091 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
27092 instead of hard-wired defaults.
27094 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
27096 * doc/invoke.texi (MIPS Options): Update documentation of the
27097 floating-point multiply-accumulate instruction restrictions.
27099 2013-02-21 Kostya Serebryany <kcc@google.com>
27101 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
27102 asan_shadow_offset on x86_64 linux.
27104 2013-02-21 Richard Biener <rguenther@suse.de>
27106 PR tree-optimization/56415
27108 2013-02-11 Richard Biener <rguenther@suse.de>
27110 PR tree-optimization/56273
27111 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27114 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27117 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
27120 PR inline-asm/56405
27121 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
27122 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
27124 2013-02-20 Jan Hubicka <jh@suse.cz>
27126 PR tree-optimization/56265
27127 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
27128 when target is referenced for first time.
27130 2013-02-20 Richard Biener <rguenther@suse.de>
27132 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
27133 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
27134 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
27135 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
27136 not return anything.
27137 (rename_ssa_copies): Do not remove unused locals.
27138 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
27139 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
27140 * passes.c (execute_function_todo): Do not schedule unused locals
27141 removal if cleanup_tree_cfg did something.
27142 * tree-ssa-live.c (remove_unused_locals): Dump statistics
27143 about the number of removed locals.
27145 2013-02-20 Richard Biener <rguenther@suse.de>
27147 PR tree-optimization/56398
27148 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
27150 2013-02-20 Martin Jambor <mjambor@suse.cz>
27152 PR tree-optimization/55334
27153 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
27154 restricted pointers to arrays.
27156 2013-02-20 Richard Biener <rguenther@suse.de>
27157 Jakub Jelinek <jakub@redhat.com>
27159 PR tree-optimization/56396
27160 * tree-ssa-ccp.c (n_const_val): New static variable.
27161 (get_value): Return NULL for SSA names we don't have a lattice
27163 (ccp_initialize): Initialize n_const_val.
27164 * tree-ssa-copy.c (n_copy_of): New static variable.
27165 (init_copy_prop): Initialize n_copy_of.
27166 (get_value): Return NULL_TREE for SSA names we don't have a
27169 2013-02-20 Martin Jambor <mjambor@suse.cz>
27171 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
27173 2013-02-20 Richard Biener <rguenther@suse.de>
27175 * genpreds.c (write_lookup_constraint): Do not compare first
27176 letter of the constraint again.
27178 2013-02-20 Richard Biener <rguenther@suse.de>
27180 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
27182 (get_use_iv_cost): Terminate hashtable walk when coming across
27185 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
27187 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
27188 reassociation for avx2 targets.
27190 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27192 * config/microblaze/microblaze.c: microblaze_has_clz = 0
27193 Add version check for v8.10.a to enable microblaze_has_clz
27194 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
27195 version and TARGET_PATTERN_COMPARE check
27196 * config/microblaze/microblaze.md: New clzsi2 instruction
27198 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27200 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
27201 function before branching.
27203 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27205 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
27207 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
27209 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27211 PR middle-end/55889
27212 * sel-sched.c: Include ira.h.
27213 (implicit_clobber_conflict_p): New function.
27214 (moveup_expr): Use it.
27215 * Makefile.in (sel-sched.o): Depend on ira.h.
27217 2013-02-19 Richard Biener <rguenther@suse.de>
27219 PR tree-optimization/56384
27220 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
27221 (vn_hash_type): Split out from ...
27222 (vn_hash_constant_with_type): ... here.
27223 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
27224 (vn_phi_eq): Compare types from vn_phi_s structure.
27225 (vn_phi_lookup): Populate vn_phi_s type.
27226 (vn_phi_insert): Likewise.
27228 2013-02-19 Jakub Jelinek <jakub@redhat.com>
27230 PR tree-optimization/56350
27231 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
27232 if haven't found reduction or nested cycle operand, rather than
27233 asserting we must find it.
27235 PR tree-optimization/56381
27236 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
27239 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
27240 Jakub Jelinek <jakub@redhat.com>
27243 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
27244 (swap_optab_enable): Same.
27245 (init_all_optabs): Use argument instead of global.
27246 * tree.h (struct tree_optimization_option): New field target_optabs.
27247 * expr.h (init_all_optabs): Add argument to prototype.
27248 (TREE_OPTIMIZATION_OPTABS): New.
27249 (save_optabs_if_changed): Protoize.
27250 * optabs.h: Declare this_fn_optabs.
27251 * optabs.c (save_optabs_if_changed): New.
27252 Declare this_fn_optabs.
27253 (init_optabs): Add argument to init_all_optabs() call.
27254 * function.c (invoke_set_current_function_hook): Handle per
27256 * function.h (struct function): New field optabs.
27257 * config/mips/mips.c (mips_set_mips16_mode): Handle when
27258 optimization_current_node has changed.
27259 * target-globals.h (save_target_globals_default_opts): Protoize.
27260 * target-globals.c (save_target_globals_default_opts): New.
27262 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27265 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
27266 registers %fr12 and %fr12R as call used.
27269 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
27270 and HImode, require all displacements to be an integer multiple of
27272 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
27273 only allow QImode and HImode when reload is in progress and strict is
27274 true. Likewise for symbolic addresses. Use base14_operand to check
27275 displacements in REG+BASE addresses.
27277 2013-02-18 Richard Biener <rguenther@suse.de>
27279 PR tree-optimization/56366
27280 * tree-vect-loop.c (get_initial_def_for_induction): Properly
27281 handle sign-conversion of outer-loop initial induction value.
27283 2013-02-18 Richard Biener <rguenther@suse.de>
27285 PR middle-end/56349
27286 * cfghooks.c (merge_blocks): If we merge a latch into another
27287 block adjust references to it.
27288 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
27289 (verify_loop_structure): Verify that a recorded latch is in fact
27292 2013-02-18 Richard Biener <rguenther@suse.de>
27294 PR tree-optimization/56321
27295 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
27296 order SSA name release and virtual operand unlinking.
27298 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27300 * config/microblaze/microblaze.md (save_stack_block): Define.
27301 (restore_stack_block): Likewise.
27303 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27305 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
27306 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
27307 * config/microblaze/microblaze.c (microblaze_option_override):
27308 Bail out early for PIC modes when target does not support PIC.
27310 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27312 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
27313 Replace with a microblaze version.
27314 (microblaze_trampoline_init): Adapt for microblaze.
27315 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
27318 2013-02-16 Jakub Jelinek <jakub@redhat.com>
27319 Dodji Seketeli <dodji@redhat.com>
27322 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
27323 (instrument_mem_region_access): Do not forget to always put
27324 instrumentation of the of 'base' and 'base + len' in a "if (len !=
27325 0) statement, even for cases where either 'base' or 'base + len'
27326 are not instrumented -- because they have been previously
27327 instrumented. Simplify the logic by putting all the statements
27328 instrument 'base + len' inside a sequence, and then insert that
27329 sequence right before the current insertion point. Then, to
27330 instrument 'base + len', just get an iterator on that statement.
27331 And do not forget to update the pointer to iterator the function
27332 received as argument.
27334 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
27336 PR rtl-optimization/56348
27337 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
27339 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
27341 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
27342 (clean_graph_dump_file): Pass base to start_graph_dump.
27344 2013-02-14 Richard Henderson <rth@redhat.com>
27347 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
27349 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
27351 * collect2-aix.h: Define F_LOADONLY.
27353 2013-02-14 Richard Biener <rguenther@suse.de>
27356 * varasm.c (output_constant_def_1): Get the decl representing
27357 the constant as argument.
27358 (output_constant_def): Wrap output_constant_def_1.
27359 (make_decl_rtl): Use output_constant_def_1 with the decl
27360 representing the constant.
27361 (build_constant_desc): Optionally re-use a decl already
27362 representing the constant.
27363 (tree_output_constant_def): Adjust.
27365 2013-02-14 Dodji Seketeli <dodji@redhat.com>
27368 * asan.c (instrument_builtin_call): Really put the length of the
27369 second source argument into src1_len.
27371 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27373 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
27374 argument. If it is false, don't create edge from then_bb to
27376 (insert_if_then_before_iter): Pass true to it.
27377 (build_check_stmt): Pass false to it.
27378 (transform_statements): Flush hash table only on extended basic
27379 block boundaries, rather than at the beginning of every bb.
27380 Don't flush hash table on nonfreeing_call_p calls.
27381 * tree-flow.h (nonfreeing_call_p): New prototype.
27382 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
27384 2013-02-13 David S. Miller <davem@davemloft.net>
27386 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
27388 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
27391 * ira.c (max_regno_before_ira): Move from ...
27393 (fix_reg_equiv_init): Use max_regno_before_ira instead of
27396 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27398 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
27400 2013-02-13 Richard Biener <rguenther@suse.de>
27403 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
27404 globals in MEM_REFs.
27406 2013-02-13 Richard Biener <rguenther@suse.de>
27408 * loop-init.c (loop_optimizer_init): Clear loop state when
27409 re-initializing preserved loops.
27410 * loop-unswitch.c (unswitch_single_loop): Return whether
27411 we unswitched the loop. Do not verify loop state here.
27412 (unswitch_loops): When we unswitched a loop discover new loops.
27414 2013-02-13 Kostya Serebryany <kcc@google.com>
27416 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
27418 * sanitizer.def: Rename __asan_init to __asan_init_v1.
27420 2013-02-12 Dodji Seketeli <dodji@redhat.com>
27422 Avoid instrumenting duplicated memory access in the same basic block
27423 * Makefile.in (asan.o): Add new dependency on hash-table.h
27424 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
27425 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
27426 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
27427 (free_mem_ref_resources, has_mem_ref_been_instrumented)
27428 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
27429 (get_mem_ref_of_assignment): New functions.
27430 (get_mem_refs_of_builtin_call): Extract from
27431 instrument_builtin_call and tweak a little bit to make it fit with
27433 (instrument_builtin_call): Use the new
27434 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
27435 of is_gimple_builtin_call.
27436 (instrument_derefs, instrument_mem_region_access): Insert the
27437 instrumented memory reference into the hash table.
27438 (maybe_instrument_assignment): Renamed instrument_assignment into
27439 this, and change it to advance the iterator when instrumentation
27440 actually happened and return true in that case. This makes it
27441 homogeneous with maybe_instrument_assignment, and thus give a
27442 chance to callers to be more 'regular'.
27443 (transform_statements): Clear the memory reference hash table
27444 whenever we enter a new BB, when we cross a function call, or when
27445 we are done transforming statements. Use
27446 maybe_instrument_assignment instead of instrumentation. No more
27447 need to special case maybe_instrument_assignment and advance the
27448 iterator after calling it; it's now handled just like
27449 maybe_instrument_call. Update comment.
27451 2013-02-13 Richard Biener <rguenther@suse.de>
27453 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
27454 Fix loop discovery code.
27456 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
27458 PR inline-asm/56148
27459 * lra-constraints.c (process_alt_operands): Match early clobber
27460 operand with itself. Check conflicts with earlyclobber only if
27461 the operand is not reloaded. Prefer to reload conflicting operand
27462 if earlyclobber and matching operands are the same.
27464 2013-02-12 Richard Biener <rguenther@suse.de>
27467 * lto-streamer-out.c (write_symbol): Do not output symbols
27468 for hard register variables.
27470 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
27473 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
27474 (umulsidi3_insn, mulsidi3_insn): New insns.
27476 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
27478 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
27479 (struct tune_params): Add vec_costs field.
27480 * config/arm/arm.c (arm_builtin_vectorization_cost)
27481 (arm_add_stmt_cost): New functions.
27482 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
27483 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
27484 (arm_default_vec_cost): New struct of type cpu_vec_costs.
27485 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
27486 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
27487 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
27488 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
27490 2013-02-12 Richard Biener <rguenther@suse.de>
27493 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
27494 decls again if possible.
27496 2013-02-12 Richard Biener <rguenther@suse.de>
27498 PR middle-end/56288
27499 * tree-ssa.c (verify_ssa_name): Fix check, move
27500 SSA_NAME_IN_FREE_LIST check up.
27502 2013-02-12 Jakub Jelinek <jakub@redhat.com>
27503 Steven Bosscher <steven@gcc.gnu.org>
27505 PR rtl-optimization/56151
27506 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
27507 equal to op0 or op1, and last_insn pattern is CODE operation
27508 with MEM dest and one of the operands matches that MEM.
27510 2013-02-11 Sriraman Tallam <tmsriram@google.com>
27512 * doc/extend.texi: Document Function Multiversioning and "default"
27513 parameter string to target attribute.
27514 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
27515 target attribute parameter is "default".
27516 (ix86_compare_version_priority): Remove checks for target attribute.
27517 (ix86_mangle_function_version_assembler_name): Change error to sorry.
27518 Remove check for target attribute equal to NULL. Add assert.
27519 (ix86_generate_version_dispatcher_body): Change error to sorry.
27521 2013-02-11 Iain Sandoe <iain@codesourcery.com>
27522 Jack Howarth <howarth@bromo.med.uc.edu>
27523 Patrick Marlier <patrick.marlier@gmail.com>
27526 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
27527 define ENDFILE_SPEC as TM_DESTRUCTOR.
27528 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
27530 2013-02-11 Alexander Potapenko <glider@google.com>
27531 Jack Howarth <howarth@bromo.med.uc.edu>
27532 Jakub Jelinek <jakub@redhat.com>
27535 * config/darwin.c (cdtor_record): Rename ctor_record.
27536 (sort_cdtor_records): Rename sort_ctor_records.
27537 (finalize_dtors): New routine to sort destructors by
27538 priority before use in assemble_integer.
27539 (machopic_asm_out_destructor): Use finalize_dtors if needed.
27541 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
27543 PR rtl-optimization/56275
27544 * simplify-rtx.c (avoid_constant_pool_reference): Check that
27545 offset is non-negative and less than cmode size before
27546 calling simplify_subreg.
27548 2013-02-11 Richard Biener <rguenther@suse.de>
27550 PR tree-optimization/56264
27551 * cfgloop.h (fix_loop_structure): Adjust prototype.
27552 * loop-init.c (fix_loop_structure): Return the number of
27553 newly discovered loops.
27554 * tree-cfgcleanup.c (repair_loop_structures): When new loops
27555 are discovered, do a full loop-closed SSA rewrite.
27557 2013-02-11 Richard Biener <rguenther@suse.de>
27559 PR tree-optimization/56273
27560 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27562 (check_array_ref): Fix missing newline in dumps.
27563 (search_for_addr_array): Likewise.
27565 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
27567 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
27569 2013-02-09 Jakub Jelinek <jakub@redhat.com>
27572 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
27574 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
27576 PR rtl-optimization/56246
27577 * lra-constraints.c (simplify_operand_subreg): Try to reuse
27579 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
27580 constraints are satisfied.
27582 2013-02-08 Jeff Law <law@redhat.com>
27585 * emit-rtl.c (reg_is_parm_p): New function.
27586 * regs.h (reg_is_parm_p): New prototype.
27587 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
27588 callee-clobbered registers.
27590 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
27593 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
27594 If there is no implicit builtin declaration, just return NULL.
27596 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27598 * config/i386/sse.md (FMAMODEM): New mode iterator.
27599 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
27600 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
27602 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27604 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
27605 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
27606 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
27608 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27610 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
27611 (microblaze*-*-elf): Likewise.
27612 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
27614 * config/microblaze/microblaze-c.c: Add builtin defines for
27615 _LITTLE_ENDIAN and _BIG_ENDIAN.
27616 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
27617 add to TARGET_DEFAULT flags.
27618 Expand ASM_SPEC and LINK_SPEC.
27619 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
27620 * config/microblaze/microblaze.md: Update extendsidi2 and
27621 movdi_internal instructions to use low-order / high-order reg
27623 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
27624 options and inversemask / mask of LITTLE_ENDIAN.
27625 * config/microblaze/t-microblaze: Expand multilib options to
27626 include mlittle-endian (le) and update exceptions patterns.
27628 2013-02-08 Jakub Jelinek <jakub@redhat.com>
27630 PR rtl-optimization/56195
27631 * lra-constraints.c (get_reload_reg): Don't reuse regs
27632 if they have smaller mode than requested, if they have
27633 wider mode than requested, try to return a SUBREG.
27635 PR tree-optimization/56250
27636 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
27637 if type is unsigned and code isn't MULT_EXPR.
27639 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27641 PR tree-optimization/56064
27642 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
27643 bits according to mode.
27644 * fixed-value.h (fixed_from_double_int)
27645 (const_fixed_from_double_int): Adjust comments.
27647 2013-02-08 Richard Biener <rguenther@suse.de>
27650 * lto-streamer.h (struct data_in): Remove current_file, current_line
27651 and current_col members.
27652 * lto-streamer-out.c (lto_output_location): Stream changed bits
27653 en-block for efficiency.
27654 * lto-streamer-in.c (clear_line_info): Remove.
27655 (lto_input_location): Cache current file, line and column
27656 globally via local statics. Read changed bits en-block.
27657 (input_function): Do not call clear_line_info.
27658 (lto_read_body): Likewise.
27659 (lto_input_toplevel_asms): Likewise.
27661 2013-02-08 Michael Matz <matz@suse.de>
27663 PR tree-optimization/52448
27664 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
27665 (nt_call_phase): New static.
27666 (add_or_mark_expr): Only mark accesses with newer phase than any
27668 (nonfreeing_call_p): New.
27669 (nt_init_block): Update nt_call_phase, mark blocks as visited.
27670 (nt_fini_block): Keep blocks marked as visited.
27671 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
27673 2013-02-08 Richard Biener <rguenther@suse.de>
27675 * ira.c (ira): Free broken dominator information.
27677 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27679 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
27681 2013-02-08 Marek Polacek <polacek@redhat.com>
27683 * cfgloop.c (verify_loop_structure): Add more checking of headers.
27685 2013-02-08 Richard Biener <rguenther@suse.de>
27687 PR middle-end/56181
27688 * cfgloop.h (flow_loops_find): Adjust.
27689 (bb_loop_header_p): Declare.
27690 * cfgloop.c (bb_loop_header_p): New function split out from ...
27691 (flow_loops_find): ... here. Adjust function signature,
27692 support incremental loop structure update.
27693 (verify_loop_structure): Cleanup. Verify a loop is a loop.
27694 * cfgloopmanip.c (fix_loop_structure): Move ...
27695 * loop-init.c (fix_loop_structure): ... here.
27696 (apply_loop_flags): Split out from ...
27697 (loop_optimizer_init): ... here.
27698 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
27699 in incremental mode, only remove dead loops here.
27701 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27704 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
27705 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
27706 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
27707 (*round<mode>3.libgcc): New insns for fixed-modes.
27708 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
27709 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
27710 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
27711 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
27712 implementations. Define to __builtin_avr_absFX,
27713 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
27714 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
27715 __builtin_avr_countlsFX, respectively.
27716 * config/avr/avr-c.c (target.h): Include it.
27717 (enum avr_builtin_id): New enum.
27718 (avr_resolve_overloaded_builtin): New static function.
27719 (avr_register_target_pragmas): Use it to set
27720 targetm.resolve_overloaded_builtin.
27721 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
27722 tree nodes used by DEF_BUILTIN.
27723 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
27724 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
27725 <AVR_BUILTIN_xxBITS>: Same.
27727 2013-02-08 Richard Biener <rguenther@suse.de>
27729 * cfgloop.c (verify_loop_structure): Properly handle
27730 a loop exiting to another loop header.
27731 * ira-int.h (ira_loops): Remove.
27732 * ira.c (ira_loops): Remove.
27733 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
27734 (do_reload): Use loop_optimizer_finalize.
27735 * ira-build.c (create_loop_tree_nodes): Use get_loops and
27736 number_of_loops to access the loop tree.
27737 (more_one_region_p): Likewise.
27738 (finish_loop_tree_nodes): Likewise.
27739 (rebuild_regno_allocno_maps): Likewise.
27740 (mark_loops_for_removal): Likewise.
27741 (mark_all_loops_for_removal): Likewise.
27742 (remove_unnecessary_regions): Likewise.
27743 (ira_build): Likewise.
27744 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
27746 2013-02-08 Richard Biener <rguenther@suse.de>
27748 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
27749 * ipa-pure-const.c (analyze_function): Avoid calling
27750 mark_irreducible_loops twice.
27751 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
27753 2013-02-07 David S. Miller <davem@davemloft.net>
27755 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
27757 * var-tracking.c (vt_add_function_parameter): Test the presence of
27758 HAVE_window_save properly and do not remap argument registers when
27759 we have a leaf function.
27761 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
27764 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
27766 * config/i386/i386.c (ix86_print_operand): Ditto.
27768 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27770 * lra-constraints.c (process_alt_operands): Fix recently added comment.
27772 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27774 PR rtl-optimization/56225
27775 * lra-constraints.c (process_alt_operands): Check that reload hard
27776 reg can hold value for strict_low_part.
27778 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27781 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
27782 dwarf2out_end_function.
27783 (in_first_function_p, maybe_at_text_label_p,
27784 first_loclabel_num_not_at_text_label): New variables.
27785 (dwarf2out_var_location): In the first function find out
27786 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
27787 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
27790 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
27792 PR rtl-optimization/56178
27793 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
27794 SUBREG of a register. Tidy up related block of code.
27795 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
27796 note if the source is a register or a SUBREG of a register.
27798 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27801 * config/rs6000/rs6000.md (ptrm): New mode attr.
27802 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
27803 call_value_indirect_aix<pttrsize>,
27804 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
27807 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
27809 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
27810 if -bnortl. Convert to strcmp and strncmp.
27812 2013-02-07 Alan Modra <amodra@gmail.com>
27815 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
27816 addresses won't wrap when offsetting.
27817 (rs6000_secondary_reload): Provide secondary reloads needed for
27818 wrapping LO_SUM addresses.
27820 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
27822 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
27823 MACH, just __MACH__.
27825 2013-02-06 Richard Biener <rguenther@suse.de>
27827 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
27828 instead of calling fix_loop_structure.
27830 2013-02-06 Jakub Jelinek <jakub@redhat.com>
27832 PR middle-end/56217
27833 * omp-low.c (use_pointer_for_field): Return false if
27834 lower_send_shared_vars doesn't generate any copy-out code.
27836 2013-02-06 Tom de Vries <tom@codesourcery.com>
27838 PR rtl-optimization/56131
27839 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
27840 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
27841 of the label is NULL. Add comment.
27843 2013-02-05 Jakub Jelinek <jakub@redhat.com>
27845 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
27848 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
27849 (STATIC_LIBTSAN_LIBS): Likewise.
27850 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
27851 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
27852 is defined, don't add anything else beyond that.
27853 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
27854 (LINK_COMMAND_SPEC): Use them.
27856 PR tree-optimization/56205
27857 * tree-stdarg.c (check_all_va_list_escapes): Return true if
27858 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
27859 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
27861 2013-02-05 Richard Biener <rguenther@suse.de>
27863 PR tree-optimization/53342
27864 PR tree-optimization/53185
27865 * tree-vectorizer.h (vect_check_strided_load): Remove.
27866 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
27867 not disallow peeling for vectorized strided loads.
27868 (vect_check_strided_load): Make static and simplify.
27869 (vect_analyze_data_refs): Adjust.
27870 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
27871 correctly when vectorizing strided loads.
27873 2013-02-05 Richard Biener <rguenther@suse.de>
27875 * doc/install.texi: Refer to ISL, not PPL.
27877 2013-02-05 Jan Hubicka <jh@suse.cz>
27879 PR tree-optimization/55789
27880 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
27882 2013-02-05 Jan Hubicka <jh@suse.cz>
27884 PR tree-optimization/55789
27885 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
27886 the dead call anyway.
27888 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
27891 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
27893 2013-02-04 Alexander Potapenko <glider@google.com>
27894 Jack Howarth <howarth@bromo.med.uc.edu>
27895 Jakub Jelinek <jakub@redhat.com>
27898 * config/darwin.c (sort_ctor_records): Stabilized qsort
27899 on constructor priority by using original position.
27900 (finalize_ctors): New routine to sort constructors by
27901 priority before use in assemble_integer.
27902 (machopic_asm_out_constructor): Use finalize_ctors if needed.
27904 2013-02-04 Jakub Jelinek <jakub@redhat.com>
27907 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
27908 about visibility on artificial decls.
27909 * config/sol2.c (solaris_assemble_visibility): Likewise.
27911 2013-02-04 Kai Tietz <ktietz@redhat.com>
27914 * config/i386/i386.c (function_value_ms_64): Add additional valtype
27915 argument and improve checking of return-argument types for 16-byte
27917 (ix86_function_value_1): Add additional valtype argument on call
27918 of function_value_64.
27919 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
27920 handling infunction_value_64 function.
27922 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
27924 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
27926 2013-02-04 Richard Biener <rguenther@suse.de>
27928 PR tree-optimization/56188
27929 * tree-ssa-structalias.c (label_visit): Consider case with
27930 initially non-empty points-to set.
27931 (perform_var_substitution): Dump node mapping and clean up.
27933 2013-02-04 Richard Guenther <rguenther@suse.de>
27936 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
27937 node prevail as last resort.
27938 (lto_symtab_merge_decls): Remove guard on LTRANS here.
27939 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
27941 2013-02-04 Richard Biener <rguenther@suse.de>
27943 PR tree-optimization/56113
27944 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
27946 (equiv_class_lookup_or_add): ... this.
27947 (label_visit): Adjust and fix error in previous patch.
27948 (perform_var_substitution): Adjust.
27950 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
27952 * config/sh/divtab.c: Fix formatting and comments throughout the file.
27953 * config/sh/sh4-300.md: Likewise.
27954 * config/sh/sh4a.md: Likewise.
27955 * config/sh/constraints.md: Likewise.
27956 * config/sh/sh.md: Likewise.
27957 * config/sh/netbsd-elf.h: Likewise.
27958 * config/sh/predicates.md: Likewise.
27959 * config/sh/sh-protos.h: Likewise.
27960 * config/sh/ushmedia.h: Likewise.
27961 * config/sh/linux.h: Likewise.
27962 * config/sh/sh.c: Likewise.
27963 * config/sh/superh.h: Likewise.
27964 * config/sh/elf.h: Likewise.
27965 * config/sh/sh4.md: Likewise.
27966 * config/sh/sh.h: Likewise.
27968 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27970 * config/pa/constraints.md: Adjust unused letters. Change "T"
27971 constraint to match_test floating_point_store_memory_operand().
27972 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
27973 (base14_operand): New.
27974 (floating_point_store_memory_operand): New.
27975 (integer_store_memory_operand): Revise to use base14_operand and
27976 reg_plus_base_memory_operand.
27977 (move_dest_operand): Allow symbolic_memory_operands.
27978 (symbolic_memory_operand): Check for LO_SOM.
27979 (symbolic_operand): Change default case to break.
27980 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
27981 CONST_DOUBLE values to be reloaded by putting them into memory when
27982 the destination is a floating point register.
27983 (movdf): Remove code to handle CONST_DOUBLE.
27985 (reload_indf_r1): New.
27986 (reload_insf_r1): New.
27987 Consistently use "Q" and "T" constraints with integer and floating
27988 point move instructions, respectively.
27989 (movdi): Remove FAIL.
27990 Change predicate for source operand unamed DImode move from
27991 general_operand to move_src_operand.
27992 (umulsidi3): Change predicate for destination operand to
27994 Likewise for similar unamed patterns.
27995 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
27996 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
27997 (hppa_legitimize_address): Simplify mask calculation.
27998 (pa_emit_move_sequence): Revised handling of secondary reloads from
27999 REG+D addresses for floating point loads and stores. Directly handle
28000 loading CONST0_RTX (mode) to a floating point register.
28001 (pa_secondary_reload): Handle reloading DF and SFmode constant values
28002 to floating point registers. Don't restrict secondary reloads to
28003 floating point registers to integer modes. Revise some comments and
28005 (TARGET_LEGITIMATE_ADDRESS_P): Define.
28006 (pa_legitimate_address_p): New.
28007 (pa_legitimize_reload_address): New.
28008 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
28009 (STRICT_REG_OK_FOR_BASE_P): New.
28010 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
28011 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
28013 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
28014 Andrew Dixie <andrewd@gentrack.com>
28016 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
28019 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
28021 * expmed.c (extract_bit_field_1): Pass the full width of the
28022 structure to get_best_reg_extraction_insn.
28024 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28027 * configure.ac (use_cxa_atexit): Add AIX.
28028 * configure: Regenerate.
28030 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
28032 2013-02-01 Jakub Jelinek <jakub@redhat.com>
28035 * final.c (need_profile_function): New variable.
28036 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
28037 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
28038 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
28039 notes, targetm.asm_out.function_prologue doesn't emit anything,
28040 HAVE_prologue and profiler should be emitted before prologue,
28041 set need_profile_function instead of emitting it.
28042 (final_scan_insn): If need_profile_function, emit
28043 profile_function on the first NOTE_INSN_BASIC_BLOCK or
28044 NOTE_INSN_FUNCTION_BEG note.
28046 2013-02-01 Richard Henderson <rth@redhat.com>
28048 * config/rs6000/rs6000.md (smulditi3): New.
28051 * config/alpha/alpha.md (umulditi3): New.
28053 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28055 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
28056 (ASM_OUTPUT_ALIGNED_LOCAL): New.
28058 2013-02-01 Richard Biener <rguenther@suse.de>
28060 PR tree-optimization/56113
28061 * tree-ssa-structalias.c (label_visit): Reduce work for
28062 single-predecessor nodes.
28064 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
28066 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
28067 range isn't testing for zero.
28069 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
28071 PR middle-end/56113
28072 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
28074 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
28075 Nick Clifton <nickc@redhat.com>
28077 * config/v850/constraints.md (Q): Define as a memory constraint.
28078 * config/v850/predicates.md (label_ref_operand): New predicate.
28079 (e3v5_shift_operand): New predicate.
28080 (ior_operator): New predicate.
28081 * config/v850/t-v850: Add e3v5 multilib.
28082 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
28083 (v850_gen_movdi): Prototype.
28084 * config/v850/v850.c: Add support for e3v5 architecture.
28085 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
28087 (construct_save_jarl): Add e3v5 long JARL support.
28088 (v850_adjust_insn_length): New function. Adjust length of call
28089 insns when using e3v5 instructions.
28090 (v850_gen_movdi): New function: Generate instructions to move a
28092 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
28093 (CPP_SPEC): Define __v850e3v5__ as appropriate.
28094 (TARGET_USE_FPU): Enable for e3v5.
28095 (CONST_OK_FOR_W): New macro.
28096 (ADJUST_INSN_LENGTH): Define.
28097 * config/v850/v850.md (UNSPEC_LOOP): Define.
28098 (attr cpu): Add v850e3v5.
28099 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
28100 (movdi): New pattern.
28101 (movdi_internal): New pattern.
28102 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
28103 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
28104 (cstoresf4): Likewise.
28105 (cstoredf4): Likewise.
28106 (insv): New pattern.
28107 (rotlso3_a): New pattern.
28108 (rotlsi3_b): New pattern
28109 (rotlsi3_v850e3v5): New pattern.
28110 (doloop_begin): New pattern.
28111 (fix_loop_counter): New pattern.
28112 (doloop_end): New pattern.
28113 (branch_normal): Add e3v5 long branch support.
28114 (branch_invert): Likewise.
28115 (branch_z_normal): Likewise.
28116 (branch_z_invert): Likewise.
28117 (branch_nz_normal): Likewise.
28118 (branch_nz_invert): Likewise.
28119 (call_internal_short): Add e3v5 register-indirect JARL support.
28120 (call_internal_long): Likewise.
28121 (call_value_internal_short): Likewise.
28122 (call_value_internal_long): Likewise.
28123 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
28124 (mloop): New option.
28125 * config.gcc: Add support for configuring v840e3v5 target.
28126 * doc/invoke.texi: Document new v850 specific command line options.
28128 2013-01-31 Paul Koning <ni1d@arrl.net>
28132 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
28133 children if parent is a class.
28134 (prune_unused_types_prune): Don't add DW_AT_declaration.
28136 2013-01-31 Richard Biener <rguenther@suse.de>
28138 PR tree-optimization/56157
28139 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
28140 match up operand with SLP child.
28142 2013-01-31 Jason Merrill <jason@redhat.com>
28145 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
28146 parameters the first time.
28147 (gen_scheduled_generic_parms_dies): Check completeness here.
28149 2013-01-31 Richard Biener <rguenther@suse.de>
28151 PR middle-end/53073
28152 * common.opt (faggressive-loop-optimizations): New flag,
28153 enabled by default.
28154 * doc/invoke.texi (faggressive-loop-optimizations): Document.
28155 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
28156 infer_loop_bounds_from_undefined by it.
28158 2013-01-31 Richard Biener <rguenther@suse.de>
28160 PR tree-optimization/56150
28161 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
28162 visit virtual operands.
28163 (find_uses_to_rename_bb): Likewise.
28165 2013-01-31 Richard Biener <rguenther@suse.de>
28167 PR tree-optimization/56150
28168 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
28169 mixed store non-store stmts.
28171 2013-01-30 Jakub Jelinek <jakub@redhat.com>
28174 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
28175 LIBASAN_EARLY_SPEC is defined.
28176 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
28177 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
28179 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
28182 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
28183 invalid args instead of ICEing on it.
28184 (ix86_valid_target_attribute_tree): Return error_mark_node if
28185 ix86_valid_target_attribute_inner_p failed.
28186 (ix86_valid_target_attribute_p): Return false only if
28187 ix86_valid_target_attribute_tree returned error_mark_node. Allow
28188 target("default") attribute.
28189 (sorted_attr_string): Change argument from const char * to tree,
28190 merge in all target attribute arguments rather than just one.
28191 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
28192 instead of free. Avoid using strcat.
28193 (ix86_mangle_function_version_assembler_name): Mangle
28194 target("default") as if no target attribute is present. Adjust
28195 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
28196 instead of xmalloc and XDELETEVEC instead of free.
28197 (ix86_function_versions): Don't return true if one of the decls
28198 doesn't have target attribute. If they don't and one of the decls
28199 is DECL_FUNCTION_VERSIONED, report an error. Adjust
28200 sorted_attr_string caller. Use XDELETEVEC instead of free.
28201 (ix86_supports_function_versions): Remove.
28202 (make_name): Fix up formatting.
28203 (make_dispatcher_decl): Remove resolver_name and its initialization.
28204 Avoid leaking memory.
28205 (is_function_default_version): Return true if there is
28206 target("default") attribute rather than no target attribute at all.
28207 (make_resolver_func): Avoid leaking memory.
28208 (ix86_generate_version_dispatcher_body): Likewise.
28209 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
28210 * target.def (supports_function_versions): Remove.
28211 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
28212 * doc/tm.texi: Regenerated.
28214 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
28216 PR rtl-optimization/56144
28217 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
28218 for values with side effects.
28220 2013-01-30 Richard Biener <rguenther@suse.de>
28222 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
28223 (sparseset_pop): Likewise.
28224 * cfganal.c (compute_idf): Likewise. Increase work-stack size
28225 to be able to use quick_push in the worker loop.
28227 2013-01-30 Marek Polacek <polacek@redhat.com>
28229 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
28231 2013-01-30 Richard Biener <rguenther@suse.de>
28234 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
28236 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
28238 PR tree-optimization/56064
28239 * fixed-value.c (fixed_from_double_int): New function.
28240 * fixed-value.h (fixed_from_double_int): New prototype.
28241 (const_fixed_from_double_int): New static inline function.
28242 * fold-const.c (native_interpret_fixed): New static function.
28243 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
28244 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
28245 (native_encode_fixed): New static function.
28246 (native_encode_expr) <FIXED_CST>: Use it.
28247 (native_interpret_int): Move double_int worker code to...
28248 * double-int.c (double_int::from_buffer): ...this new static method.
28249 * double-int.h (double_int::from_buffer): Prototype it.
28251 2013-01-30 Richard Biener <rguenther@suse.de>
28253 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
28254 New pointer-map and obstack.
28255 (init_alias_vars): Allocate pointer-map and obstack.
28256 (delete_points_to_sets): Free them.
28257 (find_what_var_points_to): Cache result.
28258 (find_what_p_points_to): Adjust for changed interface of
28259 find_what_var_points_to.
28260 (compute_points_to_sets): Likewise.
28261 (ipa_pta_execute): Likewise.
28263 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
28265 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
28266 * configure: Regenerate.
28267 * config.in: Regenerate.
28268 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
28269 #nobits/#progbits if supported.
28271 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
28274 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
28275 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
28276 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
28278 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28280 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
28281 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
28283 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28285 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
28286 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
28288 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28290 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
28292 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
28293 * config/arm/cortex-a7.md: New bypasses using
28294 arm_mac_accumulator_is_result.
28296 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28298 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
28299 (cortex_a7_neon_mla): Likewise.
28300 (cortex_a7_fpfmad): New reservation.
28301 (cortex_a7_fpmacs): Use ffmas and update required units.
28302 (cortex_a7_fpmuld): Update required units and latency.
28303 (cortex_a7_fpmacd): Likewise.
28304 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
28305 (cortex_a7_neon). Likewise.
28306 (bypass) Update participating units.
28308 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28310 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
28311 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
28313 * config/arm/vfp11.md (vfp_farith): Use ffmas.
28314 (vfp_fmul): Use ffmad.
28315 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
28316 (cortex_r4_fmacd): Use ffmad.
28317 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
28318 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
28319 (cortex_a9_fmacd): Use ffmad.
28320 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
28321 (cortex_a8_vfp_macd): Use ffmad.
28322 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
28323 (cortex_a5_fpmacd): Use ffmad.
28324 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
28325 (cortex_a15_vfp_macd): Use ffmad.
28326 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
28328 2013-01-29 Jason Merrill <jason@redhat.com>
28331 * varasm.c (default_assemble_visibility): Don't warn about
28332 visibility on artificial decls.
28334 2013-01-29 Richard Biener <rguenther@suse.de>
28336 PR tree-optimization/56113
28337 * tree-ssa-structalias.c (equiv_class_lookup): Also return
28339 (label_visit): Free duplicate bitmaps and record the leader instead.
28340 (perform_var_substitution): Adjust.
28342 2013-01-29 Richard Biener <rguenther@suse.de>
28344 PR tree-optimization/55270
28345 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
28346 the CFG, schedule loops for fixup.
28348 2013-01-29 Nick Clifton <nickc@redhat.com>
28350 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
28353 2013-01-28 Leif Ekblad <leif@rdos.net>
28355 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
28356 * config/i386/i386.h (TARGET_RDOS): New macro.
28357 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
28358 * config/i386/i386.c (ix86_option_override_internal): For 64bit
28359 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
28360 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
28361 DEFAULT_LARGE_SECTION_THRESHOLD.
28362 * config/i386/i386.md (R14_REG, R15_REG): New constants.
28363 * config/i386/rdos.h: New file.
28364 * config/i386/rdos64.h: New file.
28366 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
28369 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
28372 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28374 PR rtl-optimization/56117
28375 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
28376 call cselib_lookup_from_insn on the MEM before calling
28377 add_insn_mem_dependence.
28379 2013-01-28 Richard Biener <rguenther@suse.de>
28381 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
28382 to a stmt that didn't have one.
28383 (copy_phis_for_bb): Likewise for PHI arguments.
28384 (copy_debug_stmt): Likewise for debug stmts.
28386 2013-01-28 Richard Biener <rguenther@suse.de>
28388 PR tree-optimization/56034
28389 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
28390 (partition_builtin_p): Adjust.
28391 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
28392 it is the last partition.
28393 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
28394 up the vertex for the definition.
28395 (classify_partition): Classify whether a partition is a
28396 PKIND_REDUCTION, thus has uses outside of the loop.
28397 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
28398 Merge all PKIND_REDUCTION partitions into the last partition.
28399 (tree_loop_distribution): Seed partitions from reductions as well.
28401 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28403 PR tree-optimization/56125
28404 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
28405 pow(x,c) into sqrt(x) * powi(x, n/2) or
28406 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
28407 optimizing for size.
28408 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
28409 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
28412 PR tree-optimization/56094
28413 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
28414 to UNKNOWN_LOCATION while gimplifying expr.
28416 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
28419 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
28420 operand 0 in movabs insn template for -masm=intel asm alternative.
28421 (*movabs<mode>_2): Ditto for operand 1.
28423 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
28426 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
28429 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28431 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
28434 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
28436 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
28437 Undef to avoid warning.
28439 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28441 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
28442 * configure: Regenerate.
28444 2013-01-25 Jakub Jelinek <jakub@redhat.com>
28446 PR tree-optimization/56098
28447 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
28448 for stmts with volatile ops.
28449 (cond_store_replacement): Don't optimize if assign has volatile ops.
28450 (cond_if_else_store_replacement_1): Don't optimize if either
28451 then_assign or else_assign have volatile ops.
28452 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
28455 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28457 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
28459 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28461 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
28462 missing ':' in asm example.
28464 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
28466 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
28467 entries into lane and laneq entries.
28468 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
28469 Remove AdvSIMD scalar modes.
28470 (aarch64_sq<r>dmulh_laneq<mode>): New.
28471 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
28473 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
28474 builtin implementations to relfect changes in RTL in aarch64-simd.md.
28475 * config/aarch64/iterators.md (VCOND): New.
28478 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28481 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
28482 Add NULL LIBNAME argument to existing definitions.
28483 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
28484 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
28485 * config/avr/avr.c (DEF_BUILTIN): Same.
28486 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
28487 (avr_expand_builtin): Expand to a vanilla call if a libgcc
28488 implementation is available (DECL_ASSEMBLER_NAME is set).
28489 (avr_fold_absfx): New static function.
28490 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
28491 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
28492 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
28493 AVR_BUILTIN_ABSLLK.
28494 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
28495 (abshk, absk, abslk, absllk): Provide as static inline functions.
28497 2013-01-25 Marek Polacek <polacek@redhat.com>
28499 PR tree-optimization/56035
28500 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
28502 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
28504 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
28505 (*movtf_internal_rex64): Add (!o,C) alternative
28506 (*movxf_internal_rex64): Ditto.
28507 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
28509 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
28511 * doc/invoke.texi: fix typo.
28512 * doc/objc.texi: fix typo.
28514 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
28516 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
28517 for the first two alternatives.
28519 2013-01-24 Diego Novillo <dnovillo@google.com>
28521 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
28522 (ggc-zone.o): Remove.
28523 * configure.ac: Remove option --with-gc.
28524 * configure: Re-generate.
28525 * doc/install.texi: Remove documentation for --with-gc.
28526 * gengtype.c (write_enum_defn): Remove. Update all users.
28527 (write_Types_process_field): Remove generation of gt_e_* argument.
28528 (output_type_enum): Remove. Update all users.
28529 (write_enum_defn): Remove. Update all users.
28530 (enum alloc_zone): Remove. Update all users.
28531 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
28532 * ggc-common.c (ggc_splay_alloc): Remove first argument.
28533 Update all callers.
28534 (struct ptr_data): Remove field TYPE. Update all users.
28535 (gt_pch_note_object): Remove argument TYPE. Update all users.
28536 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
28538 * ggc-none.c (ggc_alloc_typed_stat): Remove.
28539 (struct alloc_zone): Remove.
28540 (ggc_internal_alloc_zone_stat): Remove.
28541 (ggc_internal_cleared_alloc_zone_stat): Remove.
28542 * ggc-page.c (ggc_alloc_typed_stat): Remove.
28543 (ggc_pch_count_object): Remove last argument. Update all users.
28544 (ggc_pch_alloc_object): Remove last argument. Update all users.
28545 (struct alloc_zone): Remove.
28546 * ggc-zone.c: Remove.
28547 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
28548 (struct alloc_zone): Remove.
28549 (ggc_alloc_typed_stat): Remove.
28550 (ggc_alloc_typed): Remove.
28551 (ggc_splay_alloc): Remove first argument.
28552 (rtl_zone): Remove. Update all users.
28553 (tree_zone): Remove. Update all users.
28554 (tree_id_zone): Remove. Update all users.
28555 (ggc_internal_zone_alloc_stat): Remove. Update all users.
28556 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
28557 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
28558 * tree-ssanames.c: Remove references to zone allocator in comments.
28560 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
28562 * config/avr/avr.c (avr_out_fract): Make register numbers that
28563 might be outside of source operand signed.
28565 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
28567 * config/i386/constraints.md (Yf): New constraint.
28568 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
28569 of f constraint to conditionaly disable x87 register preferences.
28570 (*movdf_internal): Ditto.
28571 (*movsf_internal): Ditto.
28573 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
28575 PR inline-asm/55934
28576 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
28577 that have operands with impossible constraints.
28578 Add a FIXME for a speed-up opportunity.
28579 * lra-constraints.c (process_alt_operands): Verify that a class
28580 selected from constraints on asms is valid for the operand mode.
28581 (curr_insn_transform): Remove incorrect comment.
28583 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
28585 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
28586 TOC operand is a valid symbol ref in the constant pool.
28588 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28590 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
28592 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
28595 * config/avr/stdfix.h: New file.
28596 * t-avr (stdfix-gcc.h): New rule to build it.
28597 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
28599 2013-01-23 Kostya Serebryany <kcc@google.com>
28601 * config/darwin.h: remove dependency on
28602 CoreFoundation (asan on Mac OS).
28604 2013-01-23 Jakub Jelinek <jakub@redhat.com>
28607 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
28608 instead of cmpdi_operand for first comparison operand.
28609 Don't assert that comparison operands aren't both constants.
28611 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28613 * doc/install.texi (Downloading the Source): Update references to
28614 downloading separate components.
28616 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28618 * doc/extend.texi (__int128): Improve grammar.
28620 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
28623 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
28624 alternative to (o,r).
28625 (*movdi_internal_rex64): Remove (!o,n) alternative.
28626 (DImode immediate->memory splitter): Remove.
28627 (DImode immediate->memory peephole2): Remove.
28628 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
28629 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
28630 alternative to (!o,*r).
28631 (*movtf_internal_sse): New pattern.
28632 (*movxf_internal_rex64): New pattern.
28633 (*movxf_internal): Disable for TARGET_64BIT.
28634 (*movdf_internal_rex64): Remove (!o,F) alternative.
28636 2013-01-22 Jakub Jelinek <jakub@redhat.com>
28638 PR middle-end/56074
28639 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
28640 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
28641 * tree-vect-loop-manip.c (find_loop_location): Also ignore
28642 stmt locations where LOCATION_LOCUS of the stmt location is
28643 UNKNOWN_LOCATION or BUILTINS_LOCATION.
28646 * config/i386/i386.md (UNSPEC_STOS): New.
28647 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
28648 *strsetqi_1): Add UNSPEC_STOS.
28650 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
28653 * doc/invoke.texi: Remove left over -Wsynth example.
28655 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28657 PR tree-optimization/56051
28658 * fold-const.c (fold_binary_loc): Don't fold
28659 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
28660 a narrowing conversion, or widening conversion from signed
28663 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28665 PR rtl-optimization/56023
28666 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
28667 dependent on debug instruction.
28669 2013-01-21 Martin Jambor <mjambor@suse.cz>
28671 PR middle-end/56022
28672 * function.c (allocate_struct_function): Call
28673 invoke_set_current_function_hook earlier.
28675 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28677 * reload1.c (init_reload): Only initialize reload_obstack
28678 during the first call.
28680 2013-01-21 Marek Polacek <polacek@redhat.com>
28682 * cfgloop.c (verify_loop_structure): Fix up grammar.
28684 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
28686 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
28687 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
28689 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28692 * config/arm/marvell-pj4.md: Update copyright year.
28693 Fix up use of alu to alu_reg and simple_alu_imm.
28695 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28697 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
28699 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
28702 * lra-constraints.c (curr_insn_transform): Don't reuse original
28703 insn for secondary memory move when memory mode should be different.
28705 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28707 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
28708 atomic_storedi_1): New patterns.
28710 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
28712 btver2 pipeline descriptions.
28713 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
28715 * config/i386/i386.md (btver2_decode): New type attributes.
28716 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
28718 * config/i386/btver2.md: New file describing btver2 pipelines.
28720 2013-01-19 Andrew Pinski <apinski@cavium.com>
28722 PR tree-optimization/52631
28723 * tree-ssa-sccvn (visit_use): Before looking up the original
28724 statement, try looking up the simplified expression.
28726 2013-01-19 Anthony Green <green@moxielogic.com>
28728 * config/moxie/moxie.c (moxie_expand_prologue): Set
28729 current_function_static_stack_size.
28731 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28733 PR tree-optimization/56029
28734 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
28735 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
28737 2013-01-18 Sharad Singhai <singhai@google.com>
28739 PR tree-optimization/55995
28740 * dumpfile.c (dump_loc): Print location only if available.
28741 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
28743 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
28746 * lra-constraints.c (curr_insn_transform): Reuse original insn for
28747 secondary memory move.
28748 (inherit_reload_reg): Use rclass instead of cl for
28749 check_secondary_memory_needed_p.
28751 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28753 PR middle-end/56015
28754 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
28755 the case where writing real complex part of target modifies op1.
28757 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
28759 * config/aarch64/aarch64-simd.md
28760 (aarch64_vcond_internal<mode>): Handle unordered cases.
28761 * config/aarch64/iterators.md (v_cmp_result): New.
28763 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
28764 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28766 * config/arm/marvell-pj4.md: New file.
28767 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
28768 * config/arm/arm.md (generic_sched): Add marvell_pj4.
28769 (generic_vfp): Likewise.
28770 * config/arm/arm-cores.def: Add marvell-pj4.
28771 * config/arm/arm-tune.md: Regenerate.
28772 * config/arm/arm-tables.opt: Regenerate.
28773 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
28774 * doc/invoke.texi: Document marvell-pj4.
28776 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
28778 * config/aarch64/arm_neon.h: Map scalar types to standard types.
28780 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28785 * var-tracking.c (negative_power_of_two_p): New.
28786 (global_get_addr_cache, local_get_addr_cache): New.
28787 (get_addr_from_global_cache, get_addr_from_local_cache): New.
28788 (vt_canonicalize_addr): Rewrite using the above. Adjust the
28790 (vt_stack_offset_p): Remove.
28791 (vt_canon_true_dep): Always canonicalize loc's address.
28792 (clobber_overlapping_mems): Make sure we have a MEM.
28793 (local_get_addr_clear_given_value): New.
28794 (val_reset): Clear local cached entries.
28795 (compute_bb_dataflow): Create and release the local cache.
28796 Disable duplicate MEMs clobbering.
28797 (emit_notes_in_bb): Clobber MEMs likewise.
28798 (vt_emit_notes): Create and release the local cache.
28799 (vt_initialize, vt_finalize): Create and release the global
28800 cache, respectively.
28801 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
28803 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28805 PR libmudflap/53359
28806 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
28807 not found in the symtab.
28809 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28812 PR rtl-optimization/55547
28813 PR rtl-optimization/53827
28816 * alias.c (offset_overlap_p): New, factored out of...
28817 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
28818 the conservative special case for symbolic constants. Don't
28819 adjust zero sizes on alignment.
28821 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
28823 PR rtl-optimization/52573
28824 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
28825 REG_UNUSED for the same register.
28827 2013-01-17 Richard Biener <rguenther@suse.de>
28828 Marek Polacek <polacek@redhat.com>
28830 PR rtl-optimization/55833
28831 * loop-unswitch.c (unswitch_loops): Move loop verification...
28832 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
28833 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
28834 Set it to true when we're removing a loop from hierarchy tree in
28835 an irreducible region.
28836 (fix_bb_placements): Adjust caller.
28837 (fix_loop_placements): Likewise.
28839 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
28841 * config/avr/builtins.def (DEF_BUILTIN): Factor out
28842 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
28843 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
28844 Remove ID. Adjust comments.
28845 * config/avr/avr-c.c (avr_builtin_name): Remove.
28846 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
28847 * config/avr/avr.c (avr_tolower): New static function.
28848 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
28849 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
28850 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
28853 2013-01-17 Jan Hubicka <jh@suse.cz>
28855 PR tree-optimization/55273
28856 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
28858 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
28861 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
28862 store through atomic_store<mode>_1.
28863 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
28865 2013-01-17 Martin Jambor <mjambor@suse.cz>
28867 PR tree-optimizations/55264
28868 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
28869 for virtual methods.
28870 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
28871 virtual methods before inlining is over.
28872 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
28874 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
28877 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
28879 PR rtl-optimization/56005
28880 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
28881 pending reads for prefetch.
28883 2013-01-16 Ian Bolton <ian.bolton@arm.com>
28885 * config/aarch64/aarch64.md
28886 (*cstoresi_neg_uxtw): New pattern.
28887 (*cmovsi_insn_uxtw): New pattern.
28888 (*<optab>si3_uxtw): New pattern.
28889 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
28890 (*<optab>si3_insn_uxtw): New pattern.
28891 (*bswapsi2_uxtw): New pattern.
28893 2013-01-16 Richard Biener <rguenther@suse.de>
28895 * tree-inline.c (tree_function_versioning): Remove set but
28896 never used variable.
28898 2013-01-16 Richard Biener <rguenther@suse.de>
28900 PR tree-optimization/55964
28901 * tree-flow.h (rename_variables_in_loop): Remove.
28902 (rename_variables_in_bb): Likewise.
28903 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
28904 (copy_loop_before): Adjust and delete update-ssa status.
28905 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
28906 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
28907 (rename_variables_in_loop): Remove.
28908 (slpeel_update_phis_for_duplicate_loop): Likewise.
28909 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
28910 use available cfg machinery instead of duplicating it.
28911 Update PHI nodes and perform poor-mans SSA update here.
28912 (slpeel_tree_peel_loop_to_edge): Adjust.
28914 2013-01-16 Richard Biener <rguenther@suse.de>
28916 PR tree-optimization/54767
28917 PR tree-optimization/53465
28918 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
28919 (vrp_visit_phi_node): For PHI arguments coming via backedges
28920 drop all symbolical range information.
28921 (execute_vrp): Compute backedges.
28923 2013-01-16 Richard Biener <rguenther@suse.de>
28925 * doc/install.texi: Update CLooG and ISL requirements to
28928 2013-01-16 Christian Bruel <christian.bruel@st.com>
28931 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
28932 (broken_move): Handle UNSPECV_SP_SWITCH_B.
28933 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
28935 2013-01-16 DJ Delorie <dj@redhat.com>
28937 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
28938 (UNSPECV_SP_SWITCH_E): New.
28939 (sp_switch_1): Change to an unspec.
28940 (sp_switch_2): Change to an unspec. Don't use post-inc when we
28943 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
28945 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
28946 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
28947 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
28948 (expand_mem_thread_fence): Ditto.
28949 (expand_mem_signal_fence): Ditto.
28950 (expand_atomic_load): Ditto.
28951 (expand_atomic_store): Ditto.
28953 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
28955 PR rtl-optimization/55547
28956 PR rtl-optimization/53827
28959 * alias.c (memrefs_conflict_p): Set sizes to negative after
28962 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28965 * function.c (thread_prologue_and_epilogue_insns): Always
28966 add crtl->drap_reg to set_up_by_prologue.set, even if
28967 stack_realign_drap is false.
28969 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
28971 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
28972 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
28973 *call): Fix indention.
28975 2013-01-15 Tom de Vries <tom@codesourcery.com>
28978 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
28981 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
28983 PR rtl-optimization/55153
28984 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
28986 2013-01-15 Martin Jambor <mjambor@suse.cz>
28988 PR tree-optimization/55920
28989 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
28990 accesses as grp_to_be_debug_replaced.
28992 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28994 PR tree-optimization/55920
28995 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
28996 there is non-useless type conversion needed from debug rhs to lhs,
28997 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
28999 2013-01-15 Joseph Myers <joseph@codesourcery.com>
29000 Mikael Pettersson <mikpe@it.uu.se>
29003 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
29005 (ASM_OUTPUT_CASE_LABEL): Remove.
29006 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
29007 * final.c (shorten_branches): Update alignment of labels before
29008 jump tables if CASE_VECTOR_SHORTEN_MODE.
29010 2013-01-15 Richard Biener <rguenther@suse.de>
29013 * system.h: Do not include gmp.h for building host tools.
29015 2013-01-15 Richard Biener <rguenther@suse.de>
29017 PR middle-end/55882
29018 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
29019 account for bitpos when computing alignment.
29021 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
29023 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
29024 (ix86_target_macros_internal): Likewise.
29026 * config/i386/i386.c (m_CORE2I7): Removed.
29027 (m_CORE_HASWELL): New macro.
29028 (m_CORE_ALL): Likewise.
29029 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
29030 (initial_ix86_arch_features): Likewise.
29031 (processor_target_table): Initializations for Core avx2.
29032 (cpu_names): New names "core-avx2".
29033 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
29034 PROCESSOR_CORE_HASWELL.
29035 (ix86_issue_rate): New case.
29036 (ia32_multipass_dfa_lookahead): Likewise.
29037 (ix86_sched_init_global): Likewise.
29039 * config/i386/i386.h (TARGET_HASWELL): New macro.
29040 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
29041 (processor_type): New PROCESSOR_HASWELL.
29043 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29045 PR tree-optimization/55955
29046 * tree-vect-loop.c (vectorizable_reduction): Give up early on
29047 *SHIFT_EXPR and *ROTATE_EXPR codes.
29049 PR tree-optimization/48766
29050 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
29051 -ftrapv disable -fwrapv.
29053 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29056 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
29057 etc. to 1 and not to __flash.
29058 Use LL suffix for __INT24_MAX__ with -mint8.
29059 Use ULL suffix for __UINT24_MAX__ with -mint8.
29061 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29063 * config/avr/avr-arch.h
29064 (struct base_arch_s): Use typedef avr_arch_t instead.
29065 (struct arch_info_s): Use typedef avr_arch_info_t instead.
29066 (struct mcu_type_s): Use typedef avr_mcu_t instead.
29067 * config/avr/avr.c: Same.
29068 * config/avr/avr-devices.c: Same.
29069 * config/avr/driver-avr.c: Same.
29070 * config/avr/gen-avr-mmcu-texi.c: Same.
29071 * config/avr/avr-mcus.def: Adjust comment.
29073 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
29075 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
29076 * config/aarch64/iterators.md (VALLDI): New.
29078 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
29079 Andi Kleen <ak@linux.intel.com>
29082 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
29083 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
29086 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29088 * config/avr/avr-stdint.h: Remove trailing blanks.
29089 * config/avr/avr-log.h: Same.
29090 * config/avr/avr-arch.h: Same.
29091 * config/avr/avr-devices.c: Same.
29092 * config/avr/avr-dimode.md: Same.
29093 * config/avr/predicates.md: Same.
29094 * config/avr/avr-c.c: Same. And fix typo.
29096 * config/avr/avr-protos.h: Same. And:
29097 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
29098 (init_cumulative_args): Rename to avr_init_cumulative_args.
29099 (expand_prologue): Rename to avr_expand_prologue.
29100 (expand_epilogue): Rename to avr_expand_epilogue.
29101 (adjust_insn_length): Rename to avr_adjust_insn_length.
29102 (notice_update_cc): Rename to avr_notice_update_cc.
29103 (final_prescan_insn): Rename to avr_final_prescan_insn.
29104 * config/avr/avr.c: Same.
29105 * config/avr/avr.h: Same.
29106 * config/avr/avr.md: Remove trailing blanks.
29107 (prologue): Use avr_expand_prologue.
29108 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
29110 2013-01-14 Richard Biener <rguenther@suse.de>
29112 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
29113 verify_location, collect_subblocks): New functions.
29114 (verify_gimple_in_cfg): Verify that locations only reference
29115 BLOCKs in the functions BLOCK tree.
29117 2013-01-14 Richard Biener <rguenther@suse.de>
29119 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
29121 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
29123 (insert_out_of_ssa_copy_on_edge): Likewise.
29124 (rewrite_close_phi_out_of_ssa): Likewise.
29125 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
29127 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
29128 propagated constants.
29129 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
29132 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29134 * config/avr/avr-modes.def: Add GPL copyright notice.
29136 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
29138 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
29139 MEMMODEL_MASK to determine memory model.
29140 (atomic_store<mode>): Ditto from operands[2].
29141 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
29143 2013-01-13 Jakub Jelinek <jakub@redhat.com>
29146 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
29147 (fold_gimple_assign): Don't call unshare_expr here.
29148 (fold_ctor_reference): Call unshare_expr.
29150 2013-01-13 Terry Guo <terry.guo@arm.com>
29152 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
29153 * doc/fragments.texi: Document MULTILIB_REUSE.
29154 * gcc.c (multilib_reuse): New internal spec.
29155 (set_multilib_dir): Also search multilib from multilib_reuse.
29156 * genmultilib (tmpmultilib3): Refactor code.
29157 (tmpmultilib4): Ditto.
29158 (multilib_reuse): New multilib argument.
29160 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
29162 * Makefile.in: Update copyright.
29164 2013-01-12 Tom de Vries <tom@codesourcery.com>
29166 PR middle-end/55890
29167 * calls.c (expand_call): Check if arg_nr is valid.
29169 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
29171 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
29172 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
29173 documentation. Add missing '__' in front of
29174 __builtin_ia32_packssdw256.
29176 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29179 * config/s390/s390.c (s390_preferred_reload_class): Do not return
29180 NO_REGS for larl operands.
29181 (s390_reload_larl_operand): Use s390_load_address instead of
29184 2013-01-11 Richard Biener <rguenther@suse.de>
29186 * tree-cfg.c (verify_node_sharing_1): Split out from ...
29187 (verify_node_sharing): ... here.
29188 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
29190 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
29192 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
29193 Substitute TREECHECKING.
29194 * configure: Regenerate.
29195 * Makefile.in (TREECHECKING): New.
29197 2013-01-11 Richard Guenther <rguenther@suse.de>
29199 PR tree-optimization/44061
29200 * tree-vrp.c (extract_range_basic): Compute zero as
29201 value-range for __builtin_constant_p of function parameters.
29203 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
29205 Update copyright years.
29207 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
29209 PR rtl-optimization/55672
29210 * lra-eliminations.c (mark_not_eliminable): Permit addition with
29211 const to be eliminable.
29213 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
29215 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
29216 * configure: Regenerate.
29218 2013-01-10 Richard Biener <rguenther@suse.de>
29220 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
29222 2013-01-10 Richard Biener <rguenther@suse.de>
29225 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
29226 locations for virtual PHI arguments.
29227 (rewrite_update_phi_arguments): Likewise.
29229 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
29231 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
29234 2013-01-10 Jakub Jelinek <jakub@redhat.com>
29236 PR tree-optimization/55921
29237 * tree-complex.c (expand_complex_asm): New function.
29238 (expand_complex_operations_1): Call it for GIMPLE_ASM.
29240 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29243 * config/s390/s390.c (s390_symref_operand_p)
29244 (s390_loadrelative_operand_p): Merge the two functions.
29245 (s390_check_qrst_address, print_operand_address): Add parameters
29246 to s390_loadrelative_operand_p invokation.
29247 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
29248 (s390_reload_larl_operand, s390_secondary_reload): Use
29249 s390_loadrelative_operand_p instead of s390_symref_operand_p.
29250 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
29252 2013-01-09 Mike Stump <mikestump@comcast.net>
29254 * dse.c (record_store): Remove unnecessary assert.
29256 2013-01-09 Jan Hubicka <jh@suse.cz>
29258 PR tree-optimization/55569
29259 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
29260 * cfgloop.h (scale_loop_profile): Likewise.
29262 2013-01-09 Jan Hubicka <jh@suse.cz>
29265 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
29267 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
29269 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
29271 PR middle-end/55114
29272 * expr.h (maybe_emit_group_store): Declare.
29273 * expr.c (maybe_emit_group_store): New function.
29274 * builtins.c (expand_builtin_int_roundingfn): Call it.
29275 (expand_builtin_int_roundingfn_2): Likewise.
29277 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
29279 PR rtl-optimization/55829
29280 * lra-constraints.c (match_reload): Add code for absent output.
29281 (curr_insn_transform): Add code for reloads of matched inputs
29284 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
29286 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
29287 attribute of movddup insn to DF.
29288 (*vec_interleave_lowv2df): Ditto.
29289 (vec_dupv2df): Ditto.
29291 2013-01-09 Jan Hubicka <jh@suse.cz>
29293 PR tree-optimiation/55875
29294 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
29295 EVERY_ITERATION parameter.
29296 (number_of_iterations_exit): Check if exit is executed every iteration.
29297 (idx_infer_loop_bounds): Similarly here.
29298 (n_of_executions_at_most): Simplify
29299 to only test for cases where statement is dominated by the
29300 particular bound; handle correctly the "postdominance" test.
29301 (scev_probably_wraps_p): Use max loop iterations info
29302 as a global bound first.
29304 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
29305 Nick Clifton <nickc@redhat.com>
29307 * config/v850/v850.md (cbranchsf4): New pattern.
29308 (cstoresf4): New pattern.
29309 (cbranchdf4): New pattern.
29310 (cstoredf4): New pattern.
29311 (movsicc): Disallow floating point comparisons.
29312 (cmpsf_le_insn): Fix order of operators.
29313 (cmpsf_lt_insn): Likewise.
29314 (cmpsf_eq_insn): Likewise.
29315 (cmpdf_le_insn): Likewise.
29316 (cmpdf_lt_insn): Likewise.
29317 (cmpdf_eq_insn): Likewise.
29318 (cmpsf_ge_insn): Use LE comparison.
29319 (cmpdf_ge_insn): Likewise.
29320 (cmpsf_gt_insn): Use LT comparison.
29321 (cmpdf_gt_insn): Likewise.
29322 (cmpsf_ne_insn): Delete pattern.
29323 (cmpdf_ne_insn): Delete pattern.
29324 * config/v850/v850.c (v850_gen_float_compare): Use
29325 gen_cmpdf_eq_insn for NE comparison.
29326 (v850_float_z_comparison_operator)
29327 (v850_float_nz_comparison_operator): Move from here ...
29328 * config/v850/predicates.md: ... to here. Move GT and GE
29329 comparisons into v850_float_z_comparison_operator.
29330 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
29332 (v850_float_nz_comparison_operator): Likewise.
29334 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29336 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
29337 with calls to gen_insvsi/gen_insvdi.
29339 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29341 * config/i386/i386.c (initial_ix86_tune_features): Set up
29342 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
29344 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
29345 Jakub Jelinek <jakub@redhat.com>
29347 PR tree-optimization/48189
29348 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
29349 If nitercst is 0, don't predict the exit edge.
29351 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29353 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
29354 in asm_fprintf with reg_names.
29355 (aarch64_print_operand_address): Likewise.
29356 (aarch64_return_addr): Likewise.
29357 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
29359 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29361 * config/pa/pa.h (VAL_U6_BITS_P): Define.
29362 (INT_U6_BITS): Likewise.
29363 * config/pa/predicates.md (uint6_operand): New predicate.
29364 (shift5_operand, shift6_operand): Likewise.
29365 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
29367 (lshrdi3): Use shift6_operand.
29368 (shrpsi4, shrpdi4): New insn patterns.
29369 (extzv): Delete expander.
29370 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
29371 predicates in unamed zero extract patterns. Tighten common constraint.
29372 (extv): Delete expander.
29373 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
29374 predicates in unamed sign extract patterns. Tighten common constraint.
29375 (insv): Delete expander.
29376 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
29377 predicates in unamed insert patterns. Tighten common constraint.
29378 Change uint32_operand predicate to uint6_operand predicate in unamed
29379 DImode pattern to insert constant values of type 1...1xxxx.
29381 2013-01-04 Jan Hubicka <jh@suse.cz>
29383 PR tree-optimization/55823
29384 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
29387 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29388 Uros Bizjak <ubizjak@gmail.com>
29390 PR rtl-optimization/55845
29391 * df-problems.c (can_move_insns_across): Stop scanning at
29392 volatile_insn_p source instruction or give up if
29393 across_from .. across_to range contains any volatile_insn_p
29396 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29398 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
29399 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
29401 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
29402 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
29404 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29407 * asan.c (asan_clear_shadow): New function.
29408 (asan_emit_stack_protection): Use it.
29410 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29412 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
29413 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
29414 with tab instead of space.
29416 2013-01-08 Nick Clifton <nickc@redhat.com>
29418 * config/rl78/rl78.c (rl78_expand_prologue): Always select
29419 register bank 0 at the start of an interrupt handler.
29420 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
29423 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29425 * config/aarch64/aarch64-simd.md
29426 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
29427 (aarch64_simd_bsl): Likewise.
29428 (aarch64_vcond_internal<mode>): Likewise.
29429 (vcond<mode><mode>): Likewise.
29430 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
29431 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
29433 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29435 * config/aarch64/aarch64-builtins.c
29436 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
29438 2013-01-08 Martin Jambor <mjambor@suse.cz>
29441 * tree-sra.c (analyze_access_subtree): Return true also after
29442 potentially creating a debug-only replacement.
29444 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29446 PR middle-end/55890
29447 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
29449 PR tree-optimization/54120
29450 * tree-vrp.c (range_fits_type_p): Don't allow
29451 src_precision < precision from signed vr to unsigned_p
29452 if vr->min or vr->max is negative.
29453 (simplify_float_conversion_using_ranges): Test can_float_p
29454 against CODE_FOR_nothing.
29456 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29457 Richard Biener <rguenther@suse.de>
29459 PR middle-end/55851
29460 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
29461 types instead of just INTEGER_TYPE types.
29463 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
29465 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
29468 2013-01-07 Steve Ellcey <sellcey@mips.com>
29471 * config/mips/mips.opt: Change mad to mmad to match documentation.
29473 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29476 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
29477 .progmemx.data now.
29479 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29482 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
29483 (avr_addrspace_t): Add .section_name field.
29484 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
29486 (avr_addrspace): Same. Initialize .section_name. Remove last
29487 NULL entry. Put __memx into .progmemx.data.
29488 (progmem_section_prefix): Remove.
29489 (avr_asm_init_sections): No need to initialize progmem_section.
29490 (avr_asm_named_section): Use avr_addrspace[].section_name to get
29491 section name prefix.
29492 (avr_asm_select_section): Ditto. And use get_unnamed_section to
29493 retrieve the progmem section.
29494 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
29495 boundary to run over avr_addrspace[].
29496 (avr_register_target_pragmas): Ditto.
29498 2013-01-06 Jakub Jelinek <jakub@redhat.com>
29500 * varasm.c (output_constant_def_contents): For asan_protect_global
29501 protected strings, adjust DECL_ALIGN if needed, before testing for
29503 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
29504 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
29506 (output_object_block): For asan protected decls, emit asan padding
29507 after their contents.
29508 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
29509 (asan_finish_file): Test it here instead.
29511 2013-01-07 Nick Clifton <nickc@redhat.com>
29512 Matthias Klose <doko@debian.org>
29513 Doug Kwan <dougkwan@google.com>
29514 H.J. Lu <hongjiu.lu@intel.com>
29517 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
29519 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
29521 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
29523 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
29525 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
29527 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29530 * doc/install.texi (Cross-Compiler-Specific Options): Document
29533 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
29535 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
29536 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
29537 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
29538 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
29539 vqmovun_high_s64): Fix source operand number and update copyright.
29541 2013-01-07 Richard Biener <rguenther@suse.de>
29543 PR middle-end/55890
29544 * gimple.h (gimple_call_builtin_p): New overload.
29545 * gimple.c (validate_call): New function.
29546 (gimple_call_builtin_p): Likewise.
29547 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
29548 Use gimple_call_builtin_p.
29549 (find_func_clobbers): Likewise.
29550 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
29551 (strlen_optimize_stmt): Likewise.
29553 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29555 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
29556 (vld1q_dup_*): Likewise.
29557 (vld1_*): Likewise.
29558 (vld1q_*): Likewise.
29559 (vld1_lane_*): Likewise.
29560 (vld1q_lane_*): Likewise.
29562 2013-01-07 Richard Biener <rguenther@suse.de>
29564 * lto-streamer.h (LTO_minor_version): Bump to 2.
29566 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29568 * config/aarch64/aarch64-protos.h
29569 (aarch64_const_double_zero_rtx_p): Rename to...
29570 (aarch64_float_const_zero_rtx_p): ...this.
29571 (aarch64_float_const_representable_p): New.
29572 (aarch64_output_simd_mov_immediate): Likewise.
29573 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
29574 move immediate case.
29575 * config/aarch64/aarch64.c
29576 (aarch64_const_double_zero_rtx_p): Rename to...
29577 (aarch64_float_const_zero_rtx_p): ...this.
29578 (aarch64_print_operand): Allow printing of new constants.
29579 (aarch64_valid_floating_const): New.
29580 (aarch64_legitimate_constant_p): Check for valid floating-point
29582 (aarch64_simd_valid_immediate): Likewise.
29583 (aarch64_vect_float_const_representable_p): New.
29584 (aarch64_float_const_representable_p): Likewise.
29585 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
29586 (aarch64_output_simd_mov_immediate): New.
29587 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
29588 (*movdf_aarch64): Likewise.
29589 * config/aarch64/constraints.md (Ufc): New.
29590 (Y): call aarch64_float_const_zero_rtx.
29591 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
29593 2013-01-07 Richard Biener <rguenther@suse.de>
29595 PR tree-optimization/55888
29596 PR tree-optimization/55862
29597 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
29598 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
29599 not if it is contained therein.
29601 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29603 * config/avr/t-avr: Typo.
29605 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29608 * config/avr/t-avr: Don't automatically rebuild
29609 $(srcdir)/config/avr/t-multilib
29610 $(srcdir)/config/avr/avr-tables.opt
29611 $(srcdir)/doc/avr-mmcu.texi
29612 (avr-mcus): New phony target to build them on request.
29613 (s-avr-mlib, s-avr-mmcu-texi): Remove.
29614 * avr/avr-mcus.def: Adjust comments.
29616 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
29618 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
29620 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29622 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
29624 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29626 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
29628 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
29630 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
29631 to generate profiling.
29632 * config/rs6000/aix64.h (LIB_SPEC): Same.
29634 2013-01-04 Andrew Pinski <apinski@cavium.com>
29636 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
29638 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
29640 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
29642 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
29644 (ix86_expand_move): Ditto.
29645 (ix86_zero_extend_to_Pmode): Ditto.
29646 (ix86_expand_call): Ditto.
29647 (ix86_expand_special_args_builtin): Ditto.
29648 (ix86_expand_builtin): Ditto.
29650 2013-01-04 Richard Biener <rguenther@suse.de>
29652 PR tree-optimization/55862
29653 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
29654 translating them through PHI nodes.
29656 2013-01-04 Martin Jambor <mjambor@suse.cz>
29658 PR tree-optimization/55755
29659 * tree-sra.c (sra_modify_assign): Do not check that an access has no
29660 children when trying to avoid producing a VIEW_CONVERT_EXPR.
29662 2013-01-04 Marek Polacek <polacek@redhat.com>
29664 PR middle-end/55859
29665 * opts.c (default_options_optimization): Clarify error message.
29667 2013-01-04 Richard Biener <rguenther@suse.de>
29669 PR middle-end/55863
29670 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
29673 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29676 * config/pa/pa.md (movsi): Revert previous change.
29677 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
29680 2013-01-03 Richard Henderson <rth@redhat.com>
29682 * config/i386/i386.c (ix86_expand_move): Always assign to op1
29683 after eliminating TLS symbols.
29685 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
29688 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
29689 * graphite-poly.c (debug_gmp_value): Likewise.
29691 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
29694 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
29695 selected code model, define __code_mode_small__, __code_model_medium__,
29696 __code_model_large__, __code_model_32__ or __code_model_kernel__.
29697 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
29698 xchg temporary register with %k. Declare temporary register as
29700 [__x86_64__]: For medium and large code models, preserve %rbx register.
29702 2013-01-03 Richard Biener <rguenther@suse.de>
29704 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
29705 (dump_subscript): Adjust.
29706 (finalize_ddr_dependent): Do not dump redundant info.
29707 (analyze_siv_subscript): Adjust.
29708 (subscript_dependence_tester): Likewise.
29709 (compute_affine_dependence): Likewise.
29711 2013-01-03 Richard Biener <rguenther@suse.de>
29714 2013-01-03 Richard Biener <rguenther@suse.de>
29716 PR tree-optimization/55857
29717 * tree-vect-stmts.c (vectorizable_load): Do not setup
29718 re-alignment for invariant loads.
29720 2013-01-02 Richard Biener <rguenther@suse.de>
29722 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29723 invariant load do not generate a vector load from the scalar location.
29725 2013-01-03 Richard Biener <rguenther@suse.de>
29727 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
29728 for not vectorizing.
29729 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
29730 not build INDIRECT_REFs, call get_name once only.
29731 (vect_create_data_ref_ptr): Likewise. Dump base object kind
29732 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
29734 2013-01-03 Richard Biener <rguenther@suse.de>
29736 PR tree-optimization/55857
29737 * tree-vect-stmts.c (vectorizable_load): Do not setup
29738 re-alignment for invariant loads.
29740 2013-01-03 Richard Biener <rguenther@suse.de>
29743 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
29744 prefer a built-in decl.
29746 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29748 * gcc.c (process_command): Update copyright notice dates.
29749 * gcov.c (print_version): Likewise.
29750 * gcov-dump.c (print_version): Likewise.
29752 PR rtl-optimization/55838
29753 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
29754 iv0.step, iv1.step and step.
29756 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29757 Marc Glisse <marc.glisse@inria.fr>
29759 PR tree-optimization/55832
29760 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
29761 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
29762 integer_{one,zero}_node.
29764 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29767 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
29768 * var-tracking.c (reverse_op): Don't add reverse ops to
29769 VALUEs that have already
29770 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
29772 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
29774 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
29776 2013-01-02 Teresa Johnson <tejohnson@google.com>
29778 * dumpfile.c (dump_loc): Print filename with location.
29779 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
29780 new location_t parameter to emit complete unroll message with
29781 new dump framework.
29782 (canonicalize_loop_induction_variables): Compute loops location
29783 and pass to try_unroll_loop_completely.
29784 * loop-unroll.c (report_unroll_peel): New function.
29785 (peel_loops_completely): Use new dump format with location
29786 for main dumpfile message, and invoke report_unroll_peel on success.
29787 (decide_unrolling_and_peeling): Ditto.
29788 (decide_peel_once_rolling): Remove old dumpfile message subsumed
29789 by report_unroll_peel.
29790 (decide_peel_completely): Ditto.
29791 (decide_unroll_constant_iterations): Ditto.
29792 (decide_unroll_runtime_iterations): Ditto.
29793 (decide_peel_simple): Ditto.
29794 (decide_unroll_stupid): Ditto.
29795 * cfgloop.c (get_loop_location): New function.
29796 * cfgloop.h (get_loop_location): Declare.
29798 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29800 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
29803 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29805 PR middle-end/55198
29806 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
29807 BLKmode objects when EXPAND_MEMORY is specified.
29809 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29811 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
29813 (fold_builtin_cpu): Do not share cpu model decls across statements.
29815 2013-01-02 Jason Merrill <jason@redhat.com>
29818 * tree.c (build_array_type_1): Revert earlier change.
29820 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
29822 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
29824 * config/aarch64/aarch64-tune.md: Re-generate.
29826 2013-01-02 Richard Biener <rguenther@suse.de>
29828 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29829 invariant load do not generate a vector load from the scalar location.
29831 2013-01-02 Richard Biener <rguenther@suse.de>
29834 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
29835 * configure: Regenerate.
29837 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29839 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
29840 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
29841 (expand_builtin_int_roundingfn_2): Keep the original target around
29842 for the fallback case.
29844 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29846 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
29847 to be clear for sign changes.
29849 2013-01-01 Jan Hubicka <jh@suse.cz>
29851 * ipa-inline-analysis.c: Fix formatting.
29853 2013-01-01 Jakub Jelinek <jakub@redhat.com>
29855 PR tree-optimization/55831
29856 * tree-vect-loop.c (get_initial_def_for_induction): Use
29857 gsi_after_labels instead of gsi_start_bb.
29859 Copyright (C) 2013 Free Software Foundation, Inc.
29861 Copying and distribution of this file, with or without modification,
29862 are permitted in any medium without royalty provided the copyright
29863 notice and this notice are preserved.