1 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
3 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
4 set writeonly flag only for vars actually written to.
6 2014-05-20 Dehao Chen <dehao@google.com>
8 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
9 and callee count to get clone count.
10 * tree-inline.c (expand_call_inline): Use callee count instead of bb
13 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
15 PR rtl-optimization/61243
16 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
18 2014-05-20 Xinliang David Li <davidxl@google.com>
20 * cgraphunit.c (walk_polymorphic_call_targets): Add
21 dbgcnt and fopt-info support.
22 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
23 * ipa-devirt.c (ipa_devirt): Ditto.
24 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
26 * ipa.c (walk_polymorphic_call_targets): Ditto.
27 * gimple-fold.c (fold_gimple_assign): Ditto.
28 (gimple_fold_call): Ditto.
29 * dbgcnt.def: New counter.
31 2014-05-20 DJ Delorie <dj@redhat.com>
33 * config/msp430/msp430.md (split): Don't allow subregs when
34 splitting SImode adds.
35 (andneghi): Fix subtraction logic.
36 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
38 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
40 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
42 * except.c (switch_to_exception_section, resolve_unique_section,
43 get_named_text_section, default_function_rodata_section,
44 align_variable, get_block_for_decl, default_section_type_flags):
45 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
46 * symtab.c (symtab_add_to_same_comdat_group,
47 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
48 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
50 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
51 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
52 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
53 (c6x_function_in_section_p): Likewise.
54 * config/darwin.c (machopic_select_section): Likewise.
55 * config/arm/arm.c (arm_function_in_section_p): Likewise.
56 * config/mips/mips.c (mips_function_rodata_section): Likewise.
57 * config/mep/mep.c (mep_select_section): LIkewise.
58 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
60 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
62 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
63 EH region of calls to pure functions that can throw an exception.
64 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
65 (copy_reference_ops_from_call): Also copy the EH region of the call if
66 it can throw an exception.
68 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
70 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
71 nested VEC_SELECTs that are inverses of each other.
73 2014-05-20 Richard Biener <rguenther@suse.de>
75 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when
77 (extract_and_process_scc_for_name): not here.
78 (cond_dom_walker::before_dom_children): Only process
79 stmts that end the BB in interesting ways.
80 (run_scc_vn): Mark param uses as visited.
82 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
84 * config/arm/arm.md (arith_shiftsi): Do not predicate for
87 2014-05-20 Nick Clifton <nickc@redhat.com>
89 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
90 (msp430_gimplify_va_arg_expr): New function.
91 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
93 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
94 operand 0 in order to prevent confusion about the number of
97 2014-05-20 Richard Biener <rguenther@suse.de>
99 PR tree-optimization/61221
100 * tree-ssa-pre.c (el_to_update): Remove.
101 (eliminate_dom_walker::before_dom_children): Handle released
102 VDEFs by value-numbering them to the associated VUSE. Update
103 stmt immediately for substituted call address.
104 (eliminate): Remove delayed stmt updating code.
105 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
106 possibly late re-numbered vuses.
107 (vn_reference_lookup_2): Adjust.
108 (vn_reference_lookup_pieces): Likewise.
109 (vn_reference_lookup): Likewise.
111 2014-05-20 Richard Biener <rguenther@suse.de>
113 * config.gcc: Remove need_64bit_hwint.
114 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
115 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
117 * config.in: Regenerate.
118 * configure: Likewise.
120 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
122 * doc/extend.texi: Create Label Attributes section,
123 move all label attributes into it and reference it.
125 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
127 * arm.c (thumb1_reorg): When scanning backwards skip anything
128 that's not a proper insn.
130 2014-05-19 Richard Biener <rguenther@suse.de>
132 PR tree-optimization/61221
133 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
134 Do nothing for unreachable blocks.
135 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
136 Improve unreachability detection.
138 2014-05-19 Richard Biener <rguenther@suse.de>
140 PR tree-optimization/61209
141 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
143 2014-05-19 Nick Clifton <nickc@redhat.com>
145 * except.c (init_eh): Fix computation of builtin setjmp buffer
146 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
148 2014-05-19 Richard Biener <rguenther@suse.de>
150 PR tree-optimization/61184
151 * tree-vrp.c (is_negative_overflow_infinity): Use
152 TREE_OVERFLOW_P and do that check first.
153 (is_positive_overflow_infinity): Likewise.
154 (is_overflow_infinity): Likewise.
155 (vrp_operand_equal_p): Properly treat operands with
156 differing overflow as not equal.
158 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
160 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
161 shift simplification where it was intended.
163 2014-05-19 Christian Bruel <christian.bruel@st.com>
166 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
168 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
171 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
174 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
176 * reg-notes.def (CROSSING_JUMP): Likewise.
177 * rtl.h (rtx_def): Update comment for jump flag.
178 (CROSSING_JUMP_P): Define.
179 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
180 of a REG_CROSSING_JUMP note.
181 * cfghooks.c (tidy_fallthru_edges): Likewise.
182 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
183 * emit-rtl.c (try_split): Likewise.
184 * haifa-sched.c (sched_create_recovery_edges): Likewise.
185 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
186 * jump.c (redirect_jump_2): Likewise.
187 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
188 (relax_delay_slots): Likewise.
189 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
191 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
192 * config/sh/sh.md (jump_compact): Likewise.
193 * bb-reorder.c (rotate_loop): Likewise.
194 (pass_duplicate_computed_gotos::execute): Likewise.
195 (add_reg_crossing_jump_notes): Rename to...
196 (update_crossing_jump_flags): ...this.
197 (pass_partition_blocks::execute): Update accordingly.
199 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
201 * tree.h: Remove extraneous template <>.
203 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
205 * ipa.c (symtab_remove_unreachable_nodes): Remove
206 symbol from comdat group if its body was eliminated.
207 (comdat_can_be_unshared_p_1): Static symbols can always
209 * symtab.c (symtab_remove_from_same_comdat_group): Break out
211 (symtab_unregister_node): ... this one.
212 (verify_symtab_base): More strict checking of comdats.
213 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
215 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
217 * tree-pass.h (make_pass_ipa_comdats): New pass.
218 * timevar.def (TV_IPA_COMDATS): New timevar.
219 * passes.def (pass_ipa_comdats): Add.
220 * Makefile.in (OBJS): Add ipa-comdats.o
221 * ipa-comdats.c: New file.
223 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
225 * ipa.c (update_visibility_by_resolution_info): New function.
226 (function_and_variable_visibility): Use it.
228 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
230 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
232 (FOR_EACH_DEFINED_SYMBOL): New macro.
233 (varpool_first_static_initializer, varpool_next_static_initializer,
234 varpool_first_defined_variable, varpool_next_defined_variable): Fix comments.
235 (symtab_in_same_comdat_p): Correctly deal with inline functions.
237 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
239 * ggc-page.c (ggc_handle_finalizers): Add comment.
241 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
243 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
244 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
245 (ggc_internal_cleared_alloc): Likewise.
246 * ggc-page.c (finalizer): New class.
247 (vec_finalizer): Likewise.
248 (globals::finalizers): New member.
249 (globals::vec_finalizers): Likewise.
250 (ggc_internal_alloc): Record the finalizer if any for the block being
252 (ggc_handle_finalizers): New function.
253 (ggc_collect): Call ggc_handle_finalizers.
254 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
256 (ggc_internal_cleared_alloc): Likewise.
257 (finalize): New function.
258 (need_finalization_p): Likewise.
259 (ggc_alloc): Install the type's destructor as the finalizer if it
261 (ggc_cleared_alloc): Likewise.
262 (ggc_vec_alloc): Likewise.
263 (ggc_cleared_vec_alloc): Likewise.
265 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
267 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
269 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
271 * alias.c (record_alias_subset): Adjust.
272 * bitmap.c (bitmap_element_allocate): Likewise.
273 (bitmap_gc_alloc_stat): Likewise.
274 * cfg.c (init_flow): Likewise.
275 (alloc_block): Likewise.
276 (unchecked_make_edge): Likewise.
277 * cfgloop.c (alloc_loop): Likewise.
278 (flow_loops_find): Likewise.
279 (rescan_loop_exit): Likewise.
280 * cfgrtl.c (init_rtl_bb_info): Likewise.
281 * cgraph.c (insert_new_cgraph_node_version): Likewise.
282 (cgraph_allocate_node): Likewise.
283 (cgraph_create_edge_1): Likewise.
284 (cgraph_allocate_init_indirect_info): Likewise.
285 * cgraphclones.c (cgraph_clone_edge): Likewise.
286 * cgraphunit.c (add_asm_node): Likewise.
287 (init_lowered_empty_function): Likewise.
288 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
289 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
290 (alpha_use_linkage): Likewise.
291 * config/arc/arc.c (arc_init_machine_status): Likewise.
292 * config/arm/arm.c (arm_init_machine_status): Likewise.
293 * config/avr/avr.c (avr_init_machine_status): Likewise.
294 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
295 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
296 * config/cris/cris.c (cris_init_machine_status): Likewise.
297 * config/darwin.c (machopic_indirection_name): Likewise.
298 (darwin_build_constant_cfstring): Likewise.
299 (darwin_enter_string_into_cfstring_table): Likewise.
300 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
301 * config/frv/frv.c (frv_init_machine_status): Likewise.
302 * config/i386/i386.c (get_dllimport_decl): Likewise.
303 (ix86_init_machine_status): Likewise.
304 (assign_386_stack_local): Likewise.
305 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
306 (i386_pe_maybe_record_exported_symbol): Likewise.
307 (i386_pe_record_stub): Likewise.
308 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
309 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
310 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
311 (m32c_note_pragma_address): Likewise.
312 * config/mep/mep.c (mep_init_machine_status): Likewise.
313 (mep_note_pragma_flag): Likewise.
314 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
315 (mips16_local_alias): Likewise.
316 (mips_init_machine_status): Likewise.
317 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
318 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
319 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
320 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
321 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
322 * config/pa/pa.c (pa_init_machine_status): Likewise.
323 (pa_get_deferred_plabel): Likewise.
324 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
325 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
326 (rs6000_init_machine_status): Likewise.
327 (output_toc): Likewise.
328 * config/s390/s390.c (s390_init_machine_status): Likewise.
329 * config/score/score.c (score_output_external): Likewise.
330 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
331 * config/spu/spu.c (spu_init_machine_status): Likewise.
332 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
333 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
334 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
335 * coverage.c (coverage_end_function): Likewise.
336 * dbxout.c (dbxout_init): Likewise.
337 * doc/gty.texi: Don't mention variable_size attribute.
338 * dwarf2cfi.c (new_cfi): Adjust.
339 (new_cfi_row): Likewise.
340 (copy_cfi_row): Likewise.
341 (create_cie_data): Likewise.
342 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
343 (new_loc_descr): Likewise.
344 (find_AT_string_in_table): Likewise.
345 (add_addr_table_entry): Likewise.
347 (add_var_loc_to_decl): Likewise.
348 (clone_die): Likewise.
349 (clone_as_declaration): Likewise.
350 (break_out_comdat_types): Likewise.
351 (new_loc_list): Likewise.
352 (add_loc_descr_to_each): Likewise.
353 (add_location_or_const_value_attribute): Likewise.
354 (add_linkage_name): Likewise.
355 (lookup_filename): Likewise.
356 (dwarf2out_var_location): Likewise.
357 (new_line_info_table): Likewise.
358 (dwarf2out_init): Likewise.
359 (mem_loc_descriptor): Likewise.
360 (loc_descriptor): Likewise.
361 (add_const_value_attribute): Likewise.
362 (tree_add_const_value_attribute): Likewise.
363 (comp_dir_string): Likewise.
364 (dwarf2out_vms_debug_main_pointer): Likewise.
365 (string_cst_pool_decl): Likewise.
366 * emit-rtl.c (set_mem_attrs): Likewise.
367 (get_reg_attrs): Likewise.
368 (start_sequence): Likewise.
369 (init_emit): Likewise.
370 (init_emit_regs): Likewise.
371 * except.c (init_eh_for_function): Likewise.
372 (gen_eh_region): Likewise.
373 (gen_eh_region_catch): Likewise.
374 (gen_eh_landing_pad): Likewise.
375 (add_call_site): Likewise.
376 * function.c (add_frame_space): Likewise.
377 (insert_temp_slot_address): Likewise.
378 (assign_stack_temp_for_type): Likewise.
379 (get_hard_reg_initial_val): Likewise.
380 (allocate_struct_function): Likewise.
381 (prepare_function_start): Likewise.
382 (types_used_by_var_decl_insert): Likewise.
383 * gengtype.c (variable_size_p): Remove function.
384 (enum alloc_quantity): Remove enum.
385 (write_typed_alloc_def): Remove function.
386 (write_typed_struct_alloc_def): Likewise.
387 (write_typed_typedef_alloc_def): Likewise.
388 (write_typed_alloc_defns): Likewise.
390 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
391 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
392 * ggc.h (ggc_alloc): new function.
393 (ggc_cleared_alloc): Likewise.
394 (ggc_vec_alloc): Template on type of vector element, and remove
395 element size argument.
396 (ggc_cleared_vec_alloc): Likewise.
397 * gimple.c (gimple_build_omp_for): Adjust.
398 (gimple_copy): Likewise.
399 * ipa-cp.c (get_replacement_map): Likewise.
400 (find_aggregate_values_for_callers_subset): Likewise.
401 (known_aggs_to_agg_replacement_list): Likewise.
402 * ipa-devirt.c (get_odr_type): Likewise.
403 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
404 (read_agg_replacement_chain): Likewise.
405 * loop-iv.c (get_simple_loop_desc): Likewise.
406 * lto-cgraph.c (input_node_opt_summary): Likewise.
407 * lto-section-in.c (lto_new_in_decl_state): Likewise.
408 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
409 (input_eh_region): Likewise.
410 (input_eh_lp): Likewise.
411 (input_cfg): Likewise.
412 * optabs.c (set_optab_libfunc): Likewise.
413 (init_tree_optimization_optabs): Likewise.
414 (set_conv_libfunc): Likewise.
415 * passes.c (do_per_function_toporder): Likewise.
416 * rtl.h: Don't use variable_size gty attribute.
417 * sese.c (if_region_set_false_region): Adjust.
418 * stringpool.c (gt_pch_save_stringpool): Likewise.
419 * target-globals.c (save_target_globals): Likewise.
420 * toplev.c (general_init): Likewise.
421 * trans-mem.c (record_tm_replacement): Likewise.
422 (split_bb_make_tm_edge): Likewise.
423 * tree-cfg.c (move_sese_region_to_fn): Likewise.
424 * tree-data-ref.h (lambda_vector_new): Likewise.
425 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
426 * tree-iterator.c (tsi_link_before): Likewise.
427 (tsi_link_after): Likewise.
428 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
429 * tree-ssa-loop-niter.c (record_estimate): Likewise.
430 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
431 * tree-ssa-operands.h: Don't use variable_size gty attribute.
432 * tree-ssa.c (init_tree_ssa): Adjust.
433 * tree-ssanames.c (set_range_info): Likewise.
434 (get_ptr_info): Likewise.
435 (duplicate_ssa_name_ptr_info): Likewise.
436 (duplicate_ssa_name_range_info): Likewise.
437 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
438 (unpack_ts_fixed_cst_value_fields): Likewise.
439 * tree.c (build_fixed): Likewise.
440 (build_real): Likewise.
441 (build_string): Likewise.
442 (decl_priority_info): Likewise.
443 (decl_debug_expr_insert): Likewise.
444 (decl_value_expr_insert): Likewise.
445 (decl_debug_args_insert): Likewise.
446 (type_hash_add): Likewise.
447 (build_omp_clause): Likewise.
448 * ubsan.c (decl_for_type_insert): Likewise.
449 * varasm.c (get_unnamed_section): Likewise.
450 (get_noswitch_section): Likewise.
451 (get_section): Likewise.
452 (get_block_for_section): Likewise.
453 (create_block_symbol): Likewise.
454 (build_constant_desc): Likewise.
455 (create_constant_pool): Likewise.
456 (force_const_mem): Likewise.
457 (record_tm_clone_pair): Likewise.
458 * varpool.c (varpool_create_empty_node): Likewise.
460 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
462 * dwarf2out.c (tree_add_const_value_attribute): Call
463 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
464 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
465 instead of ggc_internal_<x>alloc_stat.
466 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
467 (ggc_realloc): Likewise.
468 * ggc-none.c (ggc_internal_alloc): Likewise.
469 (ggc_internal_cleared_alloc): Likewise.
470 * ggc-page.c: Likewise.
471 * ggc.h (ggc_internal_alloc_stat): Likewise.
472 (ggc_internal_alloc): Remove macro.
473 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
474 (ggc_internal_cleared_alloc): Remove macro.
475 (GGC_RESIZEVEC): Adjust.
476 (ggc_resizevar): Remove macro.
477 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
478 (ggc_internal_cleared_vec_alloc_stat): Likewise.
479 (ggc_internal_vec_cleared_alloc): Remove macro.
480 (ggc_alloc_atomic_stat): Drop _stat suffix.
481 (ggc_alloc_atomic): Remove macro.
482 (ggc_alloc_cleared_atomic): Remove macro.
483 (ggc_alloc_string_stat): Drop _stat suffix.
484 (ggc_alloc_string): Remove macro.
485 (ggc_alloc_rtx_def_stat): Adjust.
486 (ggc_alloc_tree_node_stat): Likewise.
487 (ggc_alloc_cleared_tree_node_stat): Likewise.
488 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
489 (ggc_alloc_cleared_simd_clone_stat): Likewise.
490 * gimple.c (gimple_build_omp_for): Likewise.
491 (gimple_copy): Likewise.
492 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
493 * toplev.c (realloc_for_line_map): Adjust.
494 * tree-data-ref.h (lambda_vector_new): Likewise.
495 * tree-phinodes.c (allocate_phi_node): Likewise.
496 * tree.c (grow_tree_vec_stat): Likewise.
497 * vec.h (va_gc::reserve): Adjust.
499 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
501 * config/microblaze/microblaze.c (break_handler): New Declaration.
502 (microblaze_break_function_p,microblaze_is_break_handler): New.
503 (compute_frame_size): Use microblaze_break_function_p.
504 Add the test of break_handler.
505 (microblaze_function_prologue) : Add the test of variable
506 break_handler. Check the fnname by BREAK_HANDLER_NAME.
507 (microblaze_function_epilogue) : Add the test of break_handler.
508 (microblaze_globalize_label) : Add the test of break_handler.
509 Check the name by BREAK_HANDLER_NAME.
511 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
513 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
514 microblaze_is_break_handler test.
515 (call_internal1,call_value_intern): Use microblaze_break_function_p.
518 * config/microblaze/microblaze-protos.h
519 (microblaze_break_function_p,microblaze_is_break_handler):
522 * doc/extend.texi (MicroBlaze break_handler Functions): Document
523 new MicroBlaze break_handler functions.
525 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
527 * doc/extend.texi (Size of an asm): Move node text according
528 to its @menu entry position.
530 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
532 PR tree-optimization/61140
533 PR tree-optimization/61150
534 PR tree-optimization/61197
535 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
537 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
539 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
541 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
543 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
544 __SIZEOF_INT128__ is defined.
546 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
548 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
549 (rs6000_delegitimize_address): Use it.
551 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
553 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
554 inplace argument. Store the new address in the original MEM when true.
555 * emit-rtl.c (change_address_1): Likewise.
556 (adjust_address_1, adjust_automodify_address_1, offset_address):
558 * rtl.h (plus_constant): Add an inplace argument.
559 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
560 when true. Avoid generating (plus X (const_int 0)).
561 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
562 in-place. Pass true to plus_constant.
563 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
565 2014-05-16 Dehao Chen <dehao@google.com>
567 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
569 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
572 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
574 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
576 2014-05-16 Dehao Chen <dehao@google.com>
578 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
579 optimize_function_for_size_p.
580 * regs.h (REG_FREQ_FROM_BB): Likewise.
582 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
585 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
586 negt_reg_operand cases.
587 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
589 * config/sh/predicates.md (cbranch_treg_value): Simplify.
591 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
593 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
596 2014-05-16 David Malcolm <dmalcolm@redhat.com>
599 2014-04-29 David Malcolm <dmalcolm@redhat.com>
601 * tree-cfg.c (dump_function_to_file): Dump the return type of
602 functions, in a line to itself before the function body, mimicking
603 the layout of a C function.
605 2014-05-16 Dehao Chen <dehao@google.com>
607 * cfghooks.c (make_forwarder_block): Use direct computation to
608 get fall-through edge's count and frequency.
610 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
612 * config/arc/arc.c (arc_init): Fix typo in error message.
613 * config/i386/i386.c (ix86_expand_builtin): Likewise.
614 (split_stack_prologue_scratch_regno): Likewise.
615 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
616 word from error message.
618 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
620 * ira-costs.c: Fix typo in comment.
622 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
624 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
626 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
628 * varpool.c (dump_varpool_node): Dump write-only flag.
629 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
631 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
632 write-only variables.
633 * ipa.c (process_references): New function.
634 (set_readonly_bit): New function.
635 (set_writeonly_bit): New function.
636 (clear_addressable_bit): New function.
637 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
638 fix handling of aliases.
639 * cgraph.h (struct varpool_node): Add writeonly flag.
641 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
643 PR rtl-optimization/60969
644 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
645 Calculate costs for this case.
647 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
649 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
650 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
652 2014-05-16 Richard Biener <rguenther@suse.de>
654 PR tree-optimization/61194
655 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
656 bool patterns ending in a COND_EXPR.
658 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
660 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
662 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
664 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
665 where we were unable to cost an RTX.
667 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
669 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
672 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
673 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
675 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
677 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
678 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
680 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
681 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
683 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
684 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
686 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
689 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
690 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
692 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
695 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
696 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
698 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
699 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
701 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
702 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
704 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
707 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
708 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
710 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
711 ZERO_EXTEND and SIGN_EXTEND better.
713 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
714 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
716 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
719 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
720 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
722 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
723 costs when costing loads and stores to memory.
725 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
726 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
728 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
731 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
733 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
735 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
736 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
738 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
740 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
742 (aarch64_rtx_mult_cost): New.
743 (aarch64_rtx_costs): Use it, refactor as appropriate.
745 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
746 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
748 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
749 emit instructions, return number of instructions which would
751 (aarch64_add_constant): Update call to aarch64_build_constant.
752 (aarch64_output_mi_thunk): Likewise.
753 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
756 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
758 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
759 (TARGET_RTX_COSTS): Call it.
761 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
763 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
764 (cortexa57_vector_cost): Likewise.
765 (cortexa57_tunings): Use them.
767 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
769 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
770 (cpu_addrcost_table): Use it.
771 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
772 (aarch64_address_cost): Rewrite using aarch64_classify_address,
775 2014-05-16 Richard Biener <rguenther@suse.de>
777 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
778 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
779 (visit_phi): Ignore edges marked as not executable.
780 (class cond_dom_walker): New.
781 (cond_dom_walker::before_dom_children): Value-number
782 control statements and mark successor edges as not
783 executable if possible.
784 (run_scc_vn): First walk all control statements in
785 dominator order, marking edges as not executable.
786 * tree-inline.c (copy_edges_for_bb): Be not confused
787 about random edge flags.
789 2014-05-16 Richard Biener <rguenther@suse.de>
791 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
793 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
796 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
797 (__TM_simple_begin): Use it.
798 (__TM_begin): Likewise.
800 2014-05-15 Martin Jambor <mjambor@suse.cz>
803 * ipa-prop.c (update_indirect_edges_after_inlining): Check
804 type_preserved flag when the indirect edge is polymorphic.
806 2014-05-15 Martin Jambor <mjambor@suse.cz>
808 PR tree-optimization/61090
809 * tree-sra.c (sra_modify_expr): Pass the current gsi to
812 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
814 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
815 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
817 2014-05-15 Jakub Jelinek <jakub@redhat.com>
819 PR tree-optimization/61158
820 * fold-const.c (fold_binary_loc): If X is zero-extended and
821 shiftc >= prec, make sure zerobits is all ones instead of
822 invoking undefined behavior.
824 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
826 * regcprop.h: New file.
827 * regcprop.c (skip_debug_insn_p): New decl.
828 (replace_oldest_value_reg): Check skip_debug_insn_p.
829 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
830 * shrink-wrap.c: Include regcprop.h.
831 (prepare_shrink_wrap): Call
832 copyprop_hardreg_forward_bb_without_debug_insn.
834 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
836 * shrink-wrap.h: Update comment.
837 * shrink-wrap.c: Update comment.
838 (next_block_for_reg): Rename to live_edge_for_reg.
839 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
840 (move_insn_for_shrink_wrap): Split live_edge.
841 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
843 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
845 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
847 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
848 * config/sparc/sparc.md (fptype_ut699): New attribute.
849 (in_branch_delay): Return false if -mfix-ut699 is specified and
850 fptype_ut699 is set to single.
851 (truncdfsf2): Add fptype_ut699 attribute.
852 (fix_truncdfsi2): Likewise.
853 (floatsisf2): Change fptype attribute.
854 (fix_truncsfsi2): Likewise.
855 (negtf2_notv9): Delete.
856 (negtf2_v9): Likewise.
857 (negtf2_hq): New instruction.
858 (negtf2): New instruction and splitter.
859 (negdf2_notv9): Rewrite.
860 (abstf2_notv9): Delete.
861 (abstf2_hq_v9): Likewise.
862 (abstf2_v9): Likewise.
863 (abstf2_hq): New instruction.
864 (abstf2): New instruction and splitter.
865 (absdf2_notv9): Rewrite.
867 2014-05-14 Cary Coutant <ccoutant@google.com>
870 * opts.c (common_handle_option): Don't special-case "-g".
871 (set_debug_level): Default to at least level 2 with "-g".
873 2014-05-14 DJ Delorie <dj@redhat.com>
875 * config/msp430/msp430.c (msp430_builtin): Add
876 MSP430_BUILTIN_DELAY_CYCLES.
877 (msp430_init_builtins): Register void __delay_cycles(long long).
878 (msp430_builtin_decl): Add it.
879 (cg_magic_constant): New.
880 (msp430_expand_delay_cycles): New.
881 (msp430_expand_builtin): Call it.
882 (msp430_print_operand_raw): Change integer printing from "int" to
884 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
885 (delay_cycles_start): New.
886 (delay_cycles_end): New.
887 (delay_cycles_32): New.
888 (delay_cycles_32x): New.
889 (delay_cycles_16): New.
890 (delay_cycles_16x): New.
891 (delay_cycles_2): New.
892 (delay_cycles_1): New.
893 * doc/extend.texi: Document __delay_cycles().
895 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
897 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
898 length attribute computation.
900 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
903 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
905 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
908 * config/sparc/sparc.md: Fix types of low and high in DI constant
909 splitter. Use gen_int_mode in some other splitters.
911 2014-05-14 Martin Jambor <mjambor@suse.cz>
914 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
916 2014-05-14 James Norris <jnorris@codesourcery.com>
918 * omp-low.c (expand_parallel_call): Remove shadow variable.
919 (expand_omp_taskreg): Likewise.
921 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
923 * common/config/i386/i386-common.c
924 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
925 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
926 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
927 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
928 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
929 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
930 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
931 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
932 xsavecintrin.h, xsavesintrin.h.
934 * config/i386/clflushoptintrin.h: New.
935 * config/i386/xsavecintrin.h: Ditto.
936 * config/i386/xsavesintrin.h: Ditto.
937 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
940 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
941 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
943 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
944 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
945 OPTION_MASK_ISA_XSAVES.
946 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
948 (PTA_CLFLUSHOPT) Define.
951 (ix86_option_override_internal): Handle new options.
952 (ix86_valid_target_attribute_inner_p): Ditto.
953 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
954 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
955 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
956 (bdesc_special_args): Add __builtin_ia32_xsaves,
957 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
958 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
959 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
960 (ix86_expand_builtin): Handle new builtins.
961 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
962 (TARGET_CLFLUSHOPT_P): Ditto.
963 (TARGET_XSAVEC): Ditto.
964 (TARGET_XSAVEC_P): Ditto.
965 (TARGET_XSAVES): Ditto.
966 (TARGET_XSAVES_P): Ditto.
967 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
968 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
969 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
971 (ANY_XRSTOR64): Ditto.
973 (xrstor): Change into <xrstor>.
974 (xrstor_rex64): Change into <xrstor>_rex64.
975 (xrstor64): Change into <xrstor>64
977 * config/i386/i386.opt (mclflushopt): New.
980 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
982 * doc/invoke.texi: Document new options.
984 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
986 PR rtl-optimization/60866
987 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
988 Default it to -1. Pass it down to init_simplejump_data.
989 (init_simplejump_data): New parameter old_seqno. Pass it down
990 to get_seqno_for_a_jump.
991 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
992 initializing new jump seqno as a last resort. Add comment.
993 (sel_redirect_edge_and_branch): Save old seqno of the conditional
994 jump and pass it down to sel_init_new_insn.
995 (sel_redirect_edge_and_branch_force): Likewise.
997 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
999 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
1000 shifted values to avoid build warning.
1002 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1004 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
1005 * cfgrtl.c (rtl_merge_blocks): Fix comment.
1006 (cfg_layout_merge_blocks): Likewise.
1007 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
1009 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
1011 PR rtl-optimization/60901
1012 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
1013 bb predecessor belongs to the same scheduling region. Adjust comment.
1015 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
1017 * doc/sourcebuild.texi: (dfp_hw): Document.
1018 (p8vector_hw): Likewise.
1019 (powerpc_eabi_ok): Likewise.
1020 (powerpc_elfv2): Likewise.
1021 (powerpc_htm_ok): Likewise.
1022 (ppc_recip_hw): Likewise.
1025 2014-05-13 Cary Coutant <ccoutant@google.com>
1027 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
1029 2014-05-13 David Malcolm <dmalcolm@redhat.com>
1031 * gengtype-parse.c (require3): Eliminate in favor of...
1033 (require_template_declaration): Update to support optional single *
1036 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
1037 (create_user_defined_type): Handle a single level of explicit
1038 pointerness within template arguments.
1039 (struct write_types_data): Add field "kind".
1040 (filter_type_name): Handle "*" character.
1041 (write_user_func_for_structure_ptr): Require a write_types_data
1042 rather than just a prefix string, so that we can look up the kind
1043 of the wtd and use it as an index into wrote_user_func_for_ptr,
1044 ensuring that such functions are written at most once. Support
1045 subclasses by invoking the marking function of the ultimate base class.
1046 (write_user_func_for_structure_body): Require a write_types_data
1047 rather than just a prefix string, so that we can pass this to
1048 write_user_func_for_structure_ptr.
1049 (write_func_for_structure): Likewise.
1050 (ggc_wtd): Add initializer of new "kind" field.
1051 (pch_wtd): Likewise.
1053 * gengtype.h (enum write_types_kinds): New.
1054 (struct type): Add field wrote_user_func_for_ptr to the "s"
1057 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
1059 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
1060 instead of const_binop.
1061 (fold_binary_loc): Likewise.
1063 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
1065 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
1066 calculation to match get_ref_base_and_extent.
1068 2014-05-13 Catherine Moore <clm@codesourcery.com>
1069 Sandra Loosemore <sandra@codesourcery.com>
1071 * configure.ac: Fix assembly for explicit JALR relocation check.
1072 * configure: Regenerate.
1074 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1076 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
1077 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
1078 Remove associated type declarations and initialisations.
1079 (arm_expand_neon_builtin): Likewise.
1080 (neon_emit_pair_result_insn): Delete.
1081 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
1082 * config/arm/neon.md (neon_vtrn<mode>): Delete.
1083 (neon_vzip<mode>): Likewise.
1084 (neon_vuzp<mode>): Likewise.
1086 2014-05-13 Richard Biener <rguenther@suse.de>
1089 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
1090 it needs revisiting whether the call still may be tail-called.
1092 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1094 * rtl.def (SYMBOL_REF): Remove middle "0" field.
1095 * rtl.h (block_symbol): Reduce number of fields to 2.
1096 (rtx_def): Add u2.symbol_ref_flags.
1097 (SYMBOL_REF_FLAGS): Use it.
1098 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
1099 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
1100 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
1101 Lower index of SYMBOL_REF_DATA.
1102 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
1103 Print SYMBOL_REF_FLAGS at the same time.
1104 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
1106 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1108 * rtl.def (VAR_LOCATION): Remove "i" field.
1109 * rtl.h (rtx_def): Add u2.var_location_status.
1110 (PAT_VAR_LOCATION_STATUS): Use it.
1111 (gen_rtx_VAR_LOCATION): Declare.
1112 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
1113 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
1114 * var-tracking.c (emit_note_insn_var_location): Remove casts.
1116 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1118 * rtl.def (scratch): Fix outdated comment and remove "0" field.
1119 * gengtype.c (adjust_field_rtx_def): Update accordingly.
1121 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1123 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
1124 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
1125 * rtl.h (rtx_def): Add insn_uid to u2 field.
1126 (RTX_FLAG_CHECK8): Delete in favor of...
1127 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
1128 (INSN_DELETED_P): Update accordingly.
1129 (INSN_UID): Use u2.insn_uid.
1130 (INSN_CHAIN_CODE_P): Define.
1131 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
1132 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
1133 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
1134 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
1135 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
1136 indices accordingly.
1137 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
1138 Update indices for insn-chain rtxes.
1139 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
1140 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
1141 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
1142 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
1143 * combine.c (try_combine): Likewise.
1144 * ira.c (setup_prohibited_mode_move_regs): Likewise.
1146 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1148 * rtl.def (REG): Remove middle field.
1149 * rtl.h (rtx_def): Add orignal_regno to u2.
1150 (ORIGINAL_REGNO): Use it instead of field 1.
1151 (REG_ATTRS): Lower field index accordingly.
1152 * gengtype.c (adjust_field_rtx_def): Remove handling of
1153 ORIGINAL_REGNO. Move REG_ATTRS index down.
1154 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
1155 code that prints the REGNO.
1157 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1159 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
1162 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
1164 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
1166 2014-05-13 Bin Cheng <bin.cheng@arm.com>
1168 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
1169 (alloc_iv): Lower base expressions containing ADDR_EXPR.
1171 2014-05-13 Ian Bolton <ian.bolton@arm.com>
1173 * config/aarch64/aarch64-protos.h
1174 (aarch64_hard_regno_caller_save_mode): New prototype.
1175 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
1177 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
1179 2014-05-13 Christian Bruel <christian.bruel@st.com>
1181 * target.def (mode_switching): New hook vector.
1182 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
1183 (mode_exit, modepriority_to_mode): Likewise.
1184 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
1185 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1186 * target.h: Include tm.h and hard-reg-set.h.
1187 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
1188 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
1189 * doc/tm.texi Regenerate.
1190 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
1191 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1192 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
1193 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
1194 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
1195 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1196 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
1197 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
1198 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
1199 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
1200 (ix86_emit_mode_set): Hookify.
1201 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
1203 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1204 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
1205 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
1206 (epiphany_mode_priority_to_mode): Remove declaration.
1207 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
1208 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
1209 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
1211 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
1212 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
1213 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
1215 2014-05-13 Jakub Jelinek <jakub@redhat.com>
1218 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
1219 is const0_rtx, return immediately. Don't test count == 0 when
1222 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1224 * Makefile.in: add shrink-wrap.o.
1225 * config/i386/i386.c: include "shrink-wrap.h"
1226 * function.c: Likewise.
1227 (requires_stack_frame_p, next_block_for_reg,
1228 move_insn_for_shrink_wrap, prepare_shrink_wrap,
1229 dup_block_and_redirect): Move to shrink-wrap.c
1230 (thread_prologue_and_epilogue_insns): Extract three code segments
1231 as functions in shrink-wrap.c
1232 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
1234 * shrink-wrap.c: New file.
1235 * shrink-wrap.h: New file.
1237 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
1239 * doc/extend.texi: Reflect current numbers of pragmas. Remove
1240 reference to Solaris.
1242 2014-05-12 Mike Stump <mikestump@comcast.net>
1245 * genattrtab.c (filename): Add.
1246 (convert_set_attr_alternative): Improve error message.
1247 (check_defs): Restore read_md_filename for error messages.
1248 (gen_insn): Save filename.
1250 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
1252 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
1253 -fno-local-ivars and -fivar-visibility.
1254 * c-family/c.opt: Make -Wshadow also implicitly enable
1257 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
1259 * doc/tm.texi: Remove reference to deleted macro.
1260 * doc/tm.texi.in: Likewise.
1262 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
1265 * config/avr/avr.c (avr_out_store_psi): Use correct constant
1268 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
1270 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
1271 * config/arm/aout.h (License): Same.
1272 * config/arm/bpabi.h (License): Same.
1273 * config/arm/elf.h (License): Same.
1274 * config/arm/linux-elf.h (License): Same.
1275 * config/arm/linux-gas.h (License): Same.
1276 * config/arm/netbsd-elf.h (License): Same.
1277 * config/arm/uclinux-eabi.h (License): Same.
1278 * config/arm/uclinux-elf.h (License): Same.
1279 * config/arm/vxworks.h (License): Same.
1281 2014-05-11 Jakub Jelinek <jakub@redhat.com>
1283 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
1284 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
1285 number of operands to 3.
1286 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
1287 * tree-nested.c (convert_nonlocal_omp_clauses,
1288 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
1289 * gimplify.c (gimplify_scan_omp_clauses): Handle
1290 OMP_CLAUSE_LINEAR_STMT.
1291 * omp-low.c (lower_rec_input_clauses): Fix typo.
1292 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
1293 cast between Fortran boolean_type_node and C _Bool if
1296 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
1298 PR tree-optimization/61136
1299 * wide-int.h (multiple_of_p): Define a version that doesn't return
1301 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
1302 integer_zerop/const_binop pair.
1303 (multiple_of_p): Likewise, converting both operands to widest_int
1306 2014-05-09 Teresa Johnson <tejohnson@google.com>
1308 * cgraphunit.c (analyze_functions): Use correct dump file.
1310 2014-05-09 Florian Weimer <fweimer@redhat.com>
1312 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
1314 (stack_protect_return_slot_p): New function.
1315 (expand_used_vars): Call stack_protect_decl_p and
1316 stack_protect_return_slot_p for -fstack-protector-strong.
1318 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
1321 * fold-const.c (fold_binary_loc): Changed width of mask.
1323 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
1325 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
1326 unsigned int initializers for regno_in, regno_out.
1328 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
1331 * config/avr/avr.md (cc): Add new attribute set_vzn.
1332 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
1333 Set cc insn attribute to set_vzn instead of set_zn for alternatives
1334 with INC, DEC or NEG.
1335 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
1336 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
1337 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
1339 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1342 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1344 * wide-int.cc (UTItype): Define.
1345 (UDWtype): Define for appropriate W_TYPE_SIZE.
1347 2014-05-09 Richard Biener <rguenther@suse.de>
1349 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
1350 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
1351 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
1352 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
1353 ssa_propagate): Adjust.
1355 2014-05-08 Jeff Law <law@redhat.com>
1357 PR tree-optimization/61009
1358 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
1359 tri-state rather than a boolean. When a block is too big to
1360 thread through, inform caller via negative return value.
1361 (thread_across_edge): If a block was too big for normal threading,
1362 then it's too big for a joiner too, so remove temporary equivalences
1363 and return immediately.
1365 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
1366 Matthias Klose <doko@ubuntu.com>
1369 * optc-gen.awk: Fix option handling for -Wunused-parameter.
1371 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
1374 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
1376 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
1379 * config/alpha/alpha.c: Include gimple-iterator.h.
1380 (alpha_gimple_fold_builtin): New function. Move
1381 ALPHA_BUILTIN_UMULH folding from ...
1382 (alpha_fold_builtin): ... here.
1383 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
1385 2014-05-08 Wei Mi <wmi@google.com>
1388 * config/i386/i386.c (ix86_compute_frame_layout): Update
1389 preferred_stack_boundary for call, expanded from tls descriptor.
1390 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
1391 to depend on SP register.
1392 (*tls_local_dynamic_base_32_gnu): Ditto.
1393 (*tls_local_dynamic_32_once): Ditto.
1394 (tls_global_dynamic_64_<mode>): Set
1395 ix86_tls_descriptor_calls_expanded_in_cfun.
1396 (tls_local_dynamic_base_64_<mode>): Ditto.
1397 (tls_global_dynamic_32): Set
1398 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
1399 to depend on SP register.
1400 (tls_local_dynamic_base_32): Ditto.
1402 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1404 * config/arm/arm_neon.h: Update comment.
1405 * config/arm/neon-docgen.ml: Delete.
1406 * config/arm/neon-gen.ml: Delete.
1407 * doc/arm-neon-intrinsics.texi: Update comment.
1409 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1411 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
1413 (vand, vorr, veor, vorn, vbic): Remove.
1414 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
1416 (neon_vsub_unspec): Likewise.
1417 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
1419 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1421 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
1422 (vadd_s16): Likewise.
1423 (vadd_s32): Likewise.
1424 (vadd_f32): Likewise.
1425 (vadd_u8): Likewise.
1426 (vadd_u16): Likewise.
1427 (vadd_u32): Likewise.
1428 (vadd_s64): Likewise.
1429 (vadd_u64): Likewise.
1430 (vaddq_s8): Likewise.
1431 (vaddq_s16): Likewise.
1432 (vaddq_s32): Likewise.
1433 (vaddq_s64): Likewise.
1434 (vaddq_f32): Likewise.
1435 (vaddq_u8): Likewise.
1436 (vaddq_u16): Likewise.
1437 (vaddq_u32): Likewise.
1438 (vaddq_u64): Likewise.
1439 (vmul_s8): Likewise.
1440 (vmul_s16): Likewise.
1441 (vmul_s32): Likewise.
1442 (vmul_f32): Likewise.
1443 (vmul_u8): Likewise.
1444 (vmul_u16): Likewise.
1445 (vmul_u32): Likewise.
1446 (vmul_p8): Likewise.
1447 (vmulq_s8): Likewise.
1448 (vmulq_s16): Likewise.
1449 (vmulq_s32): Likewise.
1450 (vmulq_f32): Likewise.
1451 (vmulq_u8): Likewise.
1452 (vmulq_u16): Likewise.
1453 (vmulq_u32): Likewise.
1454 (vsub_s8): Likewise.
1455 (vsub_s16): Likewise.
1456 (vsub_s32): Likewise.
1457 (vsub_f32): Likewise.
1458 (vsub_u8): Likewise.
1459 (vsub_u16): Likewise.
1460 (vsub_u32): Likewise.
1461 (vsub_s64): Likewise.
1462 (vsub_u64): Likewise.
1463 (vsubq_s8): Likewise.
1464 (vsubq_s16): Likewise.
1465 (vsubq_s32): Likewise.
1466 (vsubq_s64): Likewise.
1467 (vsubq_f32): Likewise.
1468 (vsubq_u8): Likewise.
1469 (vsubq_u16): Likewise.
1470 (vsubq_u32): Likewise.
1471 (vsubq_u64): Likewise.
1472 (vand_s8): Likewise.
1473 (vand_s16): Likewise.
1474 (vand_s32): Likewise.
1475 (vand_u8): Likewise.
1476 (vand_u16): Likewise.
1477 (vand_u32): Likewise.
1478 (vand_s64): Likewise.
1479 (vand_u64): Likewise.
1480 (vandq_s8): Likewise.
1481 (vandq_s16): Likewise.
1482 (vandq_s32): Likewise.
1483 (vandq_s64): Likewise.
1484 (vandq_u8): Likewise.
1485 (vandq_u16): Likewise.
1486 (vandq_u32): Likewise.
1487 (vandq_u64): Likewise.
1488 (vorr_s8): Likewise.
1489 (vorr_s16): Likewise.
1490 (vorr_s32): Likewise.
1491 (vorr_u8): Likewise.
1492 (vorr_u16): Likewise.
1493 (vorr_u32): Likewise.
1494 (vorr_s64): Likewise.
1495 (vorr_u64): Likewise.
1496 (vorrq_s8): Likewise.
1497 (vorrq_s16): Likewise.
1498 (vorrq_s32): Likewise.
1499 (vorrq_s64): Likewise.
1500 (vorrq_u8): Likewise.
1501 (vorrq_u16): Likewise.
1502 (vorrq_u32): Likewise.
1503 (vorrq_u64): Likewise.
1504 (veor_s8): Likewise.
1505 (veor_s16): Likewise.
1506 (veor_s32): Likewise.
1507 (veor_u8): Likewise.
1508 (veor_u16): Likewise.
1509 (veor_u32): Likewise.
1510 (veor_s64): Likewise.
1511 (veor_u64): Likewise.
1512 (veorq_s8): Likewise.
1513 (veorq_s16): Likewise.
1514 (veorq_s32): Likewise.
1515 (veorq_s64): Likewise.
1516 (veorq_u8): Likewise.
1517 (veorq_u16): Likewise.
1518 (veorq_u32): Likewise.
1519 (veorq_u64): Likewise.
1520 (vbic_s8): Likewise.
1521 (vbic_s16): Likewise.
1522 (vbic_s32): Likewise.
1523 (vbic_u8): Likewise.
1524 (vbic_u16): Likewise.
1525 (vbic_u32): Likewise.
1526 (vbic_s64): Likewise.
1527 (vbic_u64): Likewise.
1528 (vbicq_s8): Likewise.
1529 (vbicq_s16): Likewise.
1530 (vbicq_s32): Likewise.
1531 (vbicq_s64): Likewise.
1532 (vbicq_u8): Likewise.
1533 (vbicq_u16): Likewise.
1534 (vbicq_u32): Likewise.
1535 (vbicq_u64): Likewise.
1536 (vorn_s8): Likewise.
1537 (vorn_s16): Likewise.
1538 (vorn_s32): Likewise.
1539 (vorn_u8): Likewise.
1540 (vorn_u16): Likewise.
1541 (vorn_u32): Likewise.
1542 (vorn_s64): Likewise.
1543 (vorn_u64): Likewise.
1544 (vornq_s8): Likewise.
1545 (vornq_s16): Likewise.
1546 (vornq_s32): Likewise.
1547 (vornq_s64): Likewise.
1548 (vornq_u8): Likewise.
1549 (vornq_u16): Likewise.
1550 (vornq_u32): Likewise.
1551 (vornq_u64): Likewise.
1553 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1555 * wide-int.cc (UTItype): Define.
1556 (UDWtype): Define for appropriate W_TYPE_SIZE.
1558 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
1560 PR tree-optimization/59100
1561 * tree-ssa-phiopt.c: Include tree-inline.h.
1562 (neutral_element_p, absorbing_element_p): New functions.
1563 (value_replacement): Handle conditional binary operations with a
1564 neutral or absorbing element.
1566 2014-05-08 Richard Biener <rguenther@suse.de>
1568 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
1569 folding the expression.
1570 (valueize_expr): Remove.
1571 (visit_reference_op_load): Do not valueize the result of
1573 (simplify_binary_expression): Likewise.
1574 (simplify_unary_expression): Likewise.
1576 2014-05-08 Richard Biener <rguenther@suse.de>
1578 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
1579 looking at TYPE_ARG_TYPES.
1581 2014-05-08 Richard Biener <rguenther@suse.de>
1583 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
1584 pointer propagation special-case.
1586 2014-05-08 Bin Cheng <bin.cheng@arm.com>
1588 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
1589 core part of address expressions.
1591 2014-05-08 Alan Modra <amodra@gmail.com>
1594 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
1595 loads and stores when -mno-strict-align at any alignment.
1596 (expand_block_clear): Similarly. Also correct calculation of
1599 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
1602 * tree-complex.c (expand_complex_move): Keep line info when expanding
1604 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
1605 of complex expression. Use new argument to display correct location
1606 for values coming from phi statement.
1607 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
1608 (warn_uninitialized_phi): Pass location of phi argument to
1610 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
1611 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
1613 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
1615 * config/rs6000/predicates.md (indexed_address_mem): New.
1616 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
1617 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
1618 fpstore_ux, fpstore_u.
1619 (sign_extend, indexed, update): New.
1620 (cell_micro): Adjust.
1621 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
1622 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
1623 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
1624 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
1625 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
1626 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
1627 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
1628 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
1629 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
1630 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
1631 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
1632 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
1633 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
1634 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
1635 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
1637 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
1638 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
1639 *vsx_extract_<mode>_store): Adjust.
1640 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
1641 is_cracked_insn, insn_must_be_first_in_group,
1642 insn_must_be_last_in_group): Adjust.
1644 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
1646 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
1647 ppc440-fpstore): Adjust.
1648 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
1649 ppc476-fpstore): Adjust.
1650 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
1651 ppc601-fpstore): Adjust.
1652 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
1654 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
1656 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
1657 ppc7450-fpstore): Adjust.
1658 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
1659 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
1660 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
1662 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
1663 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
1664 cell-fpstore, cell-fpstore-update): Adjust.
1665 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
1666 ppce300c3_store, ppce300c3_fpstore): Adjust.
1667 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
1668 e500mc_fpstore): Adjust.
1669 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
1670 e500mc64_store, e500mc64_fpstore): Adjust.
1671 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
1672 e5500_fpstore): Adjust.
1673 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
1674 e6500_fpstore): Adjust.
1675 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
1677 * config/rs6000/power4.md (power4-load, power4-load-ext,
1678 power4-load-ext-update, power4-load-ext-update-indexed,
1679 power4-load-update-indexed, power4-load-update, power4-fpload,
1680 power4-fpload-update, power4-store, power4-store-update,
1681 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
1683 * config/rs6000/power5.md (power5-load, power5-load-ext,
1684 power5-load-ext-update, power5-load-ext-update-indexed,
1685 power5-load-update-indexed, power5-load-update, power5-fpload,
1686 power5-fpload-update, power5-store, power5-store-update,
1687 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
1689 * config/rs6000/power6.md (power6-load, power6-load-ext,
1690 power6-load-update, power6-load-update-indexed,
1691 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
1692 power6-fpload-update, power6-store, power6-store-update,
1693 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
1695 * config/rs6000/power7.md (power7-load, power7-load-ext,
1696 power7-load-update, power7-load-update-indexed,
1697 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
1698 power7-fpload-update, power7-store, power7-store-update,
1699 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
1701 * config/rs6000/power8.md (power8-load, power8-load-update,
1702 power8-load-ext, power8-load-ext-update, power8-fpload,
1703 power8-fpload-update, power8-store, power8-store-update-indexed,
1704 power8-fpstore, power8-fpstore-update): Adjust.
1705 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
1707 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
1708 titan_lsu_store, titan_lsu_fpstore): Adjust.
1709 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
1711 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
1714 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
1715 unrolled byte insns. Emit address increments after move insns.
1717 2014-05-07 David Malcolm <dmalcolm@redhat.com>
1719 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
1720 const_gimple, rather than a gimple.
1721 (gimple_call_builtin_p): Likewise, for the three variants.
1723 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
1724 (gimple_call_builtin_p): Likewise, for the three variants.
1726 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1728 PR tree-optimization/61095
1729 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
1731 2014-05-07 Richard Biener <rguenther@suse.de>
1733 PR tree-optimization/61034
1734 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
1735 (maybe_skip_until): Use translate to take into account
1736 lattices when trying to do disambiguations.
1737 (get_continuation_for_phi_1): Likewise.
1738 (get_continuation_for_phi): Adjust for added translate arguments.
1739 (walk_non_aliased_vuses): Likewise.
1740 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
1741 (walk_non_aliased_vuses): Likewise.
1742 (call_may_clobber_ref_p_1): Declare.
1743 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
1744 calls. Stop early if we are only supposed to disambiguate.
1745 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
1747 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
1749 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
1750 Emit an error when the function has arguments.
1752 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
1754 * cfgloop.h (unswitch_loops): Remove.
1755 * doc/passes.texi: Remove references to loop-unswitch.c
1756 * timevar.def (TV_LOOP_UNSWITCH): Remove.
1758 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
1760 * tree-vect-data-refs.c (vect_grouped_load_supported): New
1761 check for loads group of length 3.
1762 (vect_permute_load_chain): New permutations for loads group of
1764 * tree-vect-stmts.c (vect_model_load_cost): Change cost
1765 of vec_perm_shuffle for the new permutations.
1767 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
1769 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
1770 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
1771 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
1772 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
1773 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
1774 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
1775 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
1776 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
1778 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
1780 * loop-unswitch.c: Delete.
1782 2014-05-07 Richard Biener <rguenther@suse.de>
1784 * config.gcc: Always set need_64bit_hwint to yes.
1786 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
1788 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
1789 of using optimize_size.
1791 2014-05-06 Mike Stump <mikestump@comcast.net>
1793 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
1795 2014-05-06 Joseph Myers <joseph@codesourcery.com>
1797 * config/i386/sse.md (*mov<mode>_internal)
1798 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
1799 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
1800 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
1801 (*<code><mode>3, *andnot<mode>3<mask_name>)
1802 (<mask_codefor><code><mode>3<mask_name>): Only consider
1803 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
1805 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
1808 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1810 * lra-constraints.c (valid_address_p): Move earlier in file.
1811 Add a constraint argument to the address_info version.
1812 (satisfies_memory_constraint_p): New function.
1813 (satisfies_address_constraint_p): Likewise.
1814 (process_alt_operands, curr_insn_transform): Use them.
1815 (process_address): Pass the constraint to valid_address_p when
1816 checking address operands.
1818 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
1820 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
1821 to their respective blocks. Fix inadvertent use of "node".
1823 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
1825 * emit-rtl.c (init_derived_machine_modes): New functionm, split
1827 (init_emit_once): ...here.
1828 * rtl.h (init_derived_machine_modes): Declare.
1829 * toplev.c (do_compile): Call it even if no_backend.
1831 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
1832 Mike Stump <mikestump@comcast.net>
1833 Richard Sandiford <rdsandiford@googlemail.com>
1834 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1836 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
1837 (rtx_equal_for_memref_p): Update comment.
1838 (adjust_offset_for_component_ref): Use wide-int interfaces.
1839 * builtins.c (get_object_alignment_2): Likewise.
1840 (c_readstr): Likewise.
1841 (target_char_cast): Add comment.
1842 (determine_block_size): Use wide-int interfaces.
1843 (expand_builtin_signbit): Likewise.
1844 (fold_builtin_int_roundingfn): Likewise.
1845 (fold_builtin_bitop): Likewise.
1846 (fold_builtin_bswap): Likewise.
1847 (fold_builtin_logarithm): Use signop.
1848 (fold_builtin_pow): Likewise.
1849 (fold_builtin_memory_op): Use wide-int interfaces.
1850 (fold_builtin_object_size): Likewise.
1851 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
1852 nb_iterations_estimate.
1853 (record_niter_bound): Use wide-int interfaces.
1854 (get_estimated_loop_iterations_int): Likewise.
1855 (get_estimated_loop_iterations): Likewise.
1856 (get_max_loop_iterations): Likewise.
1857 * cfgloop.h: Include wide-int.h.
1858 (struct nb_iter_bound): Change bound to widest_int.
1859 (struct loop): Change nb_iterations_upper_bound and
1860 nb_iterations_estimate to widest_int.
1861 (record_niter_bound): Switch to use widest_int.
1862 (get_estimated_loop_iterations): Likewise.
1863 (get_max_loop_iterations): Likewise.
1864 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
1865 update for wide-int.
1866 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
1867 * combine.c (try_combine): Likewise.
1868 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
1869 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
1871 (aarch64_float_const_representable_p): Likewise.
1872 * config/arc/arc.c: Include wide-int.h.
1873 (arc_can_use_doloop_p): Use wide-int interfaces.
1874 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
1875 (vfp3_const_double_index): Likewise.
1876 * config/avr/avr.c (avr_out_round): Likewise.
1877 (avr_fold_builtin): Likewise.
1878 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
1879 (bfin_can_use_doloop_p): Likewise.
1880 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
1881 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
1882 * config/i386/i386.c: Include wide-int.h.
1883 (ix86_data_alignment): Use wide-int interfaces.
1884 (ix86_local_alignment): Likewise.
1885 (ix86_emit_swsqrtsf): Update real_from_integer.
1886 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
1887 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
1888 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
1889 (zero_constant): Likewise.
1890 (input_operand): Likewise.
1891 (splat_input_operand): Likewise.
1892 (non_logical_cint_operand): Change const_double to const_wide_int.
1893 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
1894 (easy_altivec_constant): Remove comment.
1895 (paired_expand_vector_init): Use CONSTANT_P.
1896 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
1897 (rs6000_emit_move): Update checks.
1898 (rs6000_aggregate_candidate): Use wide-int interfaces.
1899 (rs6000_expand_ternop_builtin): Likewise.
1900 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
1901 (rs6000_assemble_integer): Likewise.
1902 (rs6000_hash_constant): Likewise.
1903 (output_toc): Likewise.
1904 (rs6000_rtx_costs): Likewise.
1905 (rs6000_emit_swrsqrt); Update call to real_from_integer.
1906 * config/rs6000/rs6000-c.c: Include wide-int.h.
1907 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
1908 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
1909 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
1910 Handle CONST_WIDE_INT.
1911 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
1912 Use tree_fits_uhwi_p.
1913 * config/sparc/sparc.c: Include wide-int.h.
1914 (sparc_fold_builtin): Use wide-int interfaces.
1915 * config/vax/vax.c: Include wide-int.h.
1916 (vax_float_literal): Use real_from_integer.
1917 * coretypes.h (struct hwivec_def): New.
1919 (const_hwivec): New.
1920 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
1921 (equiv_constant): Handle CONST_WIDE_INT.
1922 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
1923 (cselib_hash_rtx): Handle CONST_WIDE_INT.
1924 * dbxout.c (stabstr_U): Use wide-int interfaces.
1925 (dbxout_type): Update to use cst_fits_shwi_p.
1926 * defaults.h (LOG2_BITS_PER_UNIT): Define.
1927 (TARGET_SUPPORTS_WIDE_INT): Add default.
1928 * dfp.c: Include wide-int.h.
1929 (decimal_real_to_integer2): Use wide-int interfaces and rename to
1930 decimal_real_to_integer.
1931 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
1932 decimal_real_to_integer.
1933 * doc/generic.texi (Constant expressions): Update for wide_int.
1934 * doc/rtl.texi (const_double): Likewise.
1935 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
1936 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
1937 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
1938 (REAL_VALUE_FROM_INT): Remove.
1939 (TARGET_SUPPORTS_WIDE_INT): New.
1940 * doc/tm.texi: Regenerate.
1941 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
1942 * double-int.h: Include wide-int.h.
1943 (struct wi::int_traits): New.
1944 * dwarf2out.c (get_full_len): New.
1945 (dw_val_equal_p): Add case dw_val_class_wide_int.
1946 (size_of_loc_descr): Likewise.
1947 (output_loc_operands): Likewise.
1948 (insert_double): Remove.
1949 (insert_wide_int): New.
1951 (print_die): Add case dw_val_class_wide_int.
1952 (attr_checksum): Likewise.
1953 (attr_checksum_ordered): Likewise.
1954 (same_dw_val_p): Likewise.
1955 (size_of_die): Likewise.
1956 (value_format): Likewise.
1957 (output_die): Likewise.
1958 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
1960 (clz_loc_descriptor): Use wide-int interfaces.
1961 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
1962 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
1963 (round_up_to_align): Use wide-int interfaces.
1964 (field_byte_offset): Likewise.
1965 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
1966 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
1967 CONST_DOUBLE handling. Use wide-int interfaces.
1968 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
1969 (gen_enumeration_type_die): Use add_AT_wide.
1970 (hash_loc_operands): Add case dw_val_class_wide_int.
1971 (compare_loc_operands): Likewise.
1972 * dwarf2out.h: Include wide-int.h.
1973 (wide_int_ptr): New.
1974 (enum dw_val_class): Add dw_val_class_wide_int.
1975 (struct dw_val_struct): Add val_wide.
1976 * emit-rtl.c (const_wide_int_htab): New.
1977 (const_wide_int_htab_hash): New.
1978 (const_wide_int_htab_eq): New.
1979 (lookup_const_wide_int): New.
1980 (const_double_htab_hash): Use wide-int interfaces.
1981 (const_double_htab_eq): Likewise.
1982 (rtx_to_double_int): Conditionally compile for wide-int.
1983 (immed_double_int_const): Rename to immed_wide_int_const and
1984 update for wide-int.
1985 (immed_double_const): Conditionally compile for wide-int.
1986 (init_emit_once): Use wide-int interfaces.
1987 * explow.c (plus_constant): Likewise.
1988 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
1989 (lshift_value): Use wide-int interfaces.
1990 (expand_mult): Likewise.
1991 (choose_multiplier): Likewise.
1992 (expand_smod_pow2): Likewise.
1993 (make_tree): Likewise.
1994 * expr.c (convert_modes): Consolidate handling of constants.
1995 Use wide-int interfaces.
1996 (emit_group_load_1): Add note.
1997 (store_expr): Update comment.
1998 (get_inner_reference): Use wide-int interfaces.
1999 (expand_constructor): Update comment.
2000 (expand_expr_real_2): Use wide-int interfaces.
2001 (expand_expr_real_1): Likewise.
2002 (reduce_to_bit_field_precision): Likewise.
2003 (const_vector_from_tree): Likewise.
2004 * final.c: Include wide-int-print.h.
2005 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
2006 * fixed-value.c: Include wide-int.h.
2007 (fixed_from_string): Use wide-int interfaces.
2008 (fixed_to_decimal): Likewise.
2009 (fixed_convert_from_real): Likewise.
2010 (real_convert_from_fixed): Likewise.
2011 * fold-const.h (mem_ref_offset): Return an offset_int.
2012 (div_if_zero_remainder): Remove code parameter.
2013 * fold-const.c (div_if_zero_remainder): Remove code parameter.
2014 Use wide-int interfaces.
2015 (may_negate_without_overflow_p): Use wide-int interfaces.
2016 (negate_expr_p): Likewise.
2017 (fold_negate_expr): Likewise.
2018 (int_const_binop_1): Likewise.
2019 (const_binop): Likewise.
2020 (fold_convert_const_int_from_int): Likewise.
2021 (fold_convert_const_int_from_real): Likewise.
2022 (fold_convert_const_int_from_fixed): Likewise.
2023 (fold_convert_const_fixed_from_int): Likewise.
2024 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
2025 (sign_bit_p): Use wide-int interfaces.
2026 (make_range_step): Likewise.
2027 (build_range_check): Likewise. Pass an integer of the correct type
2028 instead of using integer_one_node.
2029 (range_predecessor): Pass an integer of the correct type instead
2030 of using integer_one_node.
2031 (range_successor): Likewise.
2032 (merge_ranges): Likewise.
2033 (unextend): Use wide-int interfaces.
2034 (extract_muldiv_1): Likewise.
2035 (fold_div_compare): Likewise.
2036 (fold_single_bit_test): Likewise.
2037 (fold_sign_changed_comparison): Likewise.
2038 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
2039 (fold_plusminus_mult_expr): Use wide-int interfaces.
2040 (native_encode_int): Likewise.
2041 (native_interpret_int): Likewise.
2042 (fold_unary_loc): Likewise.
2043 (pointer_may_wrap_p): Likewise.
2044 (size_low_cst): Likewise.
2045 (mask_with_tz): Likewise.
2046 (fold_binary_loc): Likewise.
2047 (fold_ternary_loc): Likewise.
2048 (multiple_of_p): Likewise.
2049 (tree_call_nonnegative_warnv_p): Update calls to
2050 tree_int_cst_min_precision and real_from_integer.
2051 (fold_negate_const): Use wide-int interfaces.
2052 (fold_abs_const): Likewise.
2053 (fold_relational_const): Use tree_int_cst_lt.
2054 (round_up_loc): Use wide-int interfaces.
2055 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
2056 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
2057 * gengtype.c: Remove include of double-int.h.
2058 (do_typedef): Use wide-int interfaces.
2059 (open_base_files): Add wide-int.h.
2060 (main): Add offset_int and widest_int typedefs.
2061 * gengtype-lex.l: Handle "^".
2062 (CXX_KEYWORD): Add "static".
2063 * gengtype-parse.c (require3): New.
2064 (require_template_declaration): Handle constant template arguments
2065 and nested templates.
2066 * gengtype-state.c: Don't include "double-int.h".
2067 * genpreds.c (write_one_predicate_function): Update comment.
2068 (write_tm_constrs_h): Add check for hval and lval use in
2070 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
2071 (add_to_sequence): Likewise.
2072 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
2073 and const_double_operand.
2074 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
2076 * gimple-fold.c (get_base_constructor): Likewise.
2077 (fold_array_ctor_reference): Likewise.
2078 (fold_nonarray_ctor_reference): Likewise.
2079 (fold_const_aggregate_ref_1): Likewise.
2080 (gimple_val_nonnegative_real_p): Likewise.
2081 (gimple_fold_indirect_ref): Likewise.
2082 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
2083 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
2084 (struct slsr_cand_d): Change index to be widest_int.
2085 (struct incr_info_d): Change incr to be widest_int.
2086 (alloc_cand_and_find_basis): Use wide-int interfaces.
2087 (slsr_process_phi): Likewise.
2088 (backtrace_base_for_ref): Likewise. Return a widest_int.
2089 (restructure_reference): Take a widest_int instead of a double_int.
2090 (slsr_process_ref): Use wide-int interfaces.
2091 (create_mul_ssa_cand): Likewise.
2092 (create_mul_imm_cand): Likewise.
2093 (create_add_ssa_cand): Likewise.
2094 (create_add_imm_cand): Take a widest_int instead of a double_int.
2095 (slsr_process_add): Use wide-int interfaces.
2096 (slsr_process_cast): Likewise.
2097 (slsr_process_copy): Likewise.
2098 (dump_candidate): Likewise.
2099 (dump_incr_vec): Likewise.
2100 (replace_ref): Likewise.
2101 (cand_increment): Likewise. Return a widest_int.
2102 (cand_abs_increment): Likewise.
2103 (replace_mult_candidate): Take a widest_int instead of a double_int.
2104 (replace_unconditional_candidate): Use wide-int interfaces.
2105 (incr_vec_index): Take a widest_int instead of a double_int.
2106 (create_add_on_incoming_edge): Likewise.
2107 (create_phi_basis): Use wide-int interfaces.
2108 (replace_conditional_candidate): Likewise.
2109 (record_increment): Take a widest_int instead of a double_int.
2110 (record_phi_increments): Use wide-int interfaces.
2111 (phi_incr_cost): Take a widest_int instead of a double_int.
2112 (lowest_cost_path): Likewise.
2113 (total_savings): Likewise.
2114 (analyze_increments): Use wide-int interfaces.
2115 (ncd_with_phi): Take a widest_int instead of a double_int.
2116 (ncd_of_cand_and_phis): Likewise.
2117 (nearest_common_dominator_for_cands): Likewise.
2118 (insert_initializers): Use wide-int interfaces.
2119 (all_phi_incrs_profitable): Likewise.
2120 (replace_one_candidate): Likewise.
2121 (replace_profitable_candidates): Likewise.
2122 * godump.c: Include wide-int-print.h.
2123 (go_output_typedef): Use wide-int interfaces.
2124 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
2125 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
2126 (build_loop_iteration_domains): Likewise.
2127 * hooks.h: Include wide-int.h rather than double-int.h.
2128 (hook_bool_dint_dint_uint_bool_true): Delete.
2129 (hook_bool_wint_wint_uint_bool_true): Declare.
2130 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
2131 (hook_bool_wint_wint_uint_bool_true): New.
2132 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
2134 (ubsan_expand_si_overflow_mul_check): Likewise.
2135 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
2136 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
2137 (get_ancestor_addr_info): Likewise.
2138 (ipa_modify_call_arguments): Likewise.
2139 * loop-doloop.c (doloop_modify): Likewise.
2140 (doloop_optimize): Likewise.
2141 * loop-iv.c (iv_number_of_iterations): Likewise.
2142 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
2143 (unroll_loop_constant_iterations): Likewise.
2144 (decide_unroll_runtime_iterations): Likewise.
2145 (unroll_loop_runtime_iterations): Likewise.
2146 (decide_peel_simple): Likewise.
2147 (decide_unroll_stupid): Likewise.
2148 * lto-streamer-in.c (streamer_read_wi): Add.
2149 (input_cfg): Use wide-int interfaces.
2150 (lto_input_tree_1): Likewise.
2151 * lto-streamer-out.c (streamer_write_wi): Add.
2152 (hash_tree): Use wide-int interfaces.
2153 (output_cfg): Likewise.
2154 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
2155 (GTFILES): Add wide-int.h and signop.h.
2156 (TAGS): Look for .cc files too.
2157 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
2158 * optabs.c (expand_subword_shift): Likewise.
2159 (expand_doubleword_shift): Likewise.
2160 (expand_absneg_bit): Likewise.
2161 (expand_copysign_absneg): Likewise.
2162 (expand_copysign_bit): Likewise.
2163 * postreload.c (reload_cse_simplify_set): Likewise.
2164 * predict.c (predict_iv_comparison): Likewise.
2165 * pretty-print.h: Include wide-int-print.h.
2167 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
2168 * print-tree.c: Include wide-int-print.h.
2169 (print_node_brief): Use wide-int interfaces.
2170 (print_node): Likewise.
2171 * read-rtl.c (validate_const_wide_int): New.
2172 (read_rtx_code): Add CONST_WIDE_INT case.
2173 * real.c: Include wide-int.h.
2174 (real_to_integer2): Delete.
2175 (real_to_integer): New function, returning a wide_int.
2176 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
2177 (ten_to_ptwo): Update call to real_from_integer.
2178 (real_digit): Likewise.
2179 * real.h: Include signop.h, wide-int.h and insn-modes.h.
2180 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
2181 (REAL_VALUE_TO_INT): Delete.
2182 (real_to_integer): Declare a wide-int form.
2183 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
2184 * recog.c (const_int_operand): Improve comment.
2185 (const_scalar_int_operand): New.
2186 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
2187 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
2188 (split_double): Likewise.
2189 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
2190 (rtx_size): Likewise.
2191 (rtx_alloc_stat_v): New.
2192 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
2193 (cwi_output_hex): New.
2194 (iterative_hash_rtx): Handle CONST_WIDE_INT.
2195 (cwi_check_failed_bounds): New.
2196 * rtl.def (CONST_WIDE_INT): New.
2197 * rtl.h: Include <utility> and wide-int.h.
2198 (struct hwivec_def): New.
2199 (CWI_GET_NUM_ELEM): New.
2200 (CWI_PUT_NUM_ELEM): New.
2201 (struct rtx_def): Add num_elem and hwiv.
2202 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
2203 (CASE_CONST_UNIQUE): Likewise.
2204 (CASE_CONST_ANY): Likewise.
2205 (CONST_SCALAR_INT_P): Likewise.
2206 (CONST_WIDE_INT_P): New.
2208 (HWIVEC_CHECK): New.
2209 (cwi_check_failed_bounds): New.
2211 (HWIVEC_CHECK): New.
2212 (CONST_WIDE_INT_VEC) New.
2213 (CONST_WIDE_INT_NUNITS) New.
2214 (CONST_WIDE_INT_ELT) New.
2215 (rtx_mode_t): New type.
2216 (wi::int_traits <rtx_mode_t>): New.
2218 (wi::min_value): New.
2219 (wi::max_value): New.
2221 (const_wide_int_alloc): New.
2222 (immed_wide_int_const): New.
2223 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
2224 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
2225 * signop.h: New file.
2226 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
2227 (simplify_const_unary_operation): Use wide-int interfaces.
2228 (simplify_binary_operation_1): Likewise.
2229 (simplify_const_binary_operation): Likewise.
2230 (simplify_const_relational_operation): Likewise.
2231 (simplify_immed_subreg): Likewise.
2232 * stmt.c (expand_case): Likewise.
2233 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
2234 signop rather than a bool.
2235 * stor-layout.c (layout_type): Use wide-int interfaces.
2236 (initialize_sizetypes): Update calls to
2237 set_min_and_max_values_for_integral_type.
2238 (set_min_and_max_values_for_integral_type): Take a signop rather
2239 than a bool. Use wide-int interfaces.
2240 (fixup_signed_type): Update accordingly. Remove
2241 HOST_BITS_PER_DOUBLE_INT limit.
2242 (fixup_unsigned_type): Likewise.
2243 * system.h (STATIC_CONSTANT_P): New.
2244 (STATIC_ASSERT): New.
2245 * target.def (can_use_doloop_p): Take widest_ints rather than
2247 * target.h: Include wide-int.h rather than double-int.h.
2248 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
2250 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
2251 rather than INT_CST_LT_UNSIGNED.
2252 (can_use_doloop_if_innermost): Take widest_ints rather than
2254 * tree-affine.c: Include wide-int-print.h.
2255 (double_int_ext_for_comb): Delete.
2256 (wide_int_ext_for_comb): New.
2257 (aff_combination_zero): Use wide-int interfaces.
2258 (aff_combination_const): Take a widest_int instead of a double_int.
2259 (aff_combination_elt): Use wide-int interfaces.
2260 (aff_combination_scale): Take a widest_int instead of a double_int.
2261 (aff_combination_add_elt): Likewise.
2262 (aff_combination_add_cst): Likewise.
2263 (aff_combination_add): Use wide-int interfaces.
2264 (aff_combination_convert): Likewise.
2265 (tree_to_aff_combination): Likewise.
2266 (add_elt_to_tree): Take a widest_int instead of a double_int.
2267 (aff_combination_to_tree): Use wide-int interfaces.
2268 (aff_combination_remove_elt): Likewise.
2269 (aff_combination_add_product): Take a widest_int instead of
2271 (aff_combination_mult): Use wide-int interfaces.
2272 (aff_combination_expand): Likewise.
2273 (double_int_constant_multiple_p): Delete.
2274 (wide_int_constant_multiple_p): New.
2275 (aff_combination_constant_multiple_p): Take a widest_int pointer
2276 instead of a double_int pointer.
2277 (print_aff): Use wide-int interfaces.
2278 (get_inner_reference_aff): Take a widest_int pointer
2279 instead of a double_int pointer.
2280 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
2281 * tree-affine.h: Include wide-int.h.
2282 (struct aff_comb_elt): Change type of coef to widest_int.
2283 (struct affine_tree_combination): Change type of offset to widest_int.
2284 (double_int_ext_for_comb): Delete.
2285 (wide_int_ext_for_comb): New.
2286 (aff_combination_const): Use widest_int instead of double_int.
2287 (aff_combination_scale): Likewise.
2288 (aff_combination_add_elt): Likewise.
2289 (aff_combination_constant_multiple_p): Likewise.
2290 (get_inner_reference_aff): Likewise.
2291 (aff_comb_cannot_overlap_p): Likewise.
2292 (aff_combination_zero_p): Use wide-int interfaces.
2293 * tree.c: Include tree.h.
2294 (init_ttree): Use make_int_cst.
2295 (tree_code_size): Removed code for INTEGER_CST case.
2296 (tree_size): Add INTEGER_CST case.
2297 (make_node_stat): Update comment.
2298 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
2299 (build_int_cst_type): Use wide-int interfaces.
2300 (double_int_to_tree): Likewise.
2301 (double_int_fits_to_tree_p): Delete.
2302 (force_fit_type_double): Delete.
2303 (force_fit_type): New.
2304 (int_cst_hash_hash): Use wide-int interfaces.
2305 (int_cst_hash_eq): Likewise.
2306 (build_int_cst_wide): Delete.
2307 (wide_int_to_tree): New.
2308 (cache_integer_cst): Use wide-int interfaces.
2309 (build_low_bits_mask): Likewise.
2310 (cst_and_fits_in_hwi): Likewise.
2311 (real_value_from_int_cst): Likewise.
2312 (make_int_cst_stat): New.
2313 (integer_zerop): Use wide_int interfaces.
2314 (integer_onep): Likewise.
2315 (integer_all_onesp): Likewise.
2316 (integer_pow2p): Likewise.
2317 (integer_nonzerop): Likewise.
2318 (tree_log2): Likewise.
2319 (tree_floor_log2): Likewise.
2320 (tree_ctz): Likewise.
2321 (int_size_in_bytes): Likewise.
2322 (mem_ref_offset): Return an offset_int rather than a double_int.
2323 (build_type_attribute_qual_variant): Use wide_int interfaces.
2324 (type_hash_eq): Likewise
2325 (tree_int_cst_equal): Likewise.
2326 (tree_int_cst_lt): Delete.
2327 (tree_int_cst_compare): Likewise.
2328 (tree_fits_shwi_p): Use wide_int interfaces.
2329 (tree_fits_uhwi_p): Likewise.
2330 (tree_int_cst_sign_bit): Likewise.
2331 (tree_int_cst_sgn): Likewise.
2332 (tree_int_cst_min_precision): Take a signop rather than a bool.
2333 (simple_cst_equal): Use wide_int interfaces.
2334 (compare_tree_int): Likewise.
2335 (iterative_hash_expr): Likewise.
2336 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
2338 (get_type_static_bounds): Use wide_int interfaces.
2339 (tree_int_cst_elt_check_failed): New.
2340 (build_common_tree_nodes): Reordered to set prec before filling in
2342 (int_cst_value): Check cst_and_fits_in_hwi.
2343 (widest_int_cst_value): Use wide_int interfaces.
2344 (upper_bound_in_type): Likewise.
2345 (lower_bound_in_type): Likewise.
2346 (num_ending_zeros): Likewise.
2347 (drop_tree_overflow): Likewise.
2348 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
2349 (gen_conditions_for_pow_cst_base): Likewise.
2350 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
2351 (group_case_labels_stmt): Use wide-int interfaces.
2352 (verify_gimple_assign_binary): Likewise.
2353 (print_loop): Likewise.
2354 * tree-chrec.c (tree_fold_binomial): Likewise.
2355 * tree-core.h (struct tree_base): Add int_length.
2356 (struct tree_int_cst): Change rep of value.
2357 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
2358 (dr_may_alias_p): Likewise.
2359 (max_stmt_executions_tree): Likewise.
2360 * tree.def (INTEGER_CST): Update comment.
2361 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
2362 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
2363 * tree-dump.c: Include wide-int.h and wide-int-print.h.
2364 (dequeue_and_dump): Use wide-int interfaces.
2365 * tree.h: Include wide-int.h.
2366 (NULL_TREE): Moved to earlier loc in file.
2367 (TREE_INT_CST_ELT_CHECK): New.
2368 (tree_int_cst_elt_check_failed): New.
2370 (TREE_INT_CST): Delete.
2371 (TREE_INT_CST_LOW): Use wide-int interfaces.
2372 (TREE_INT_CST_HIGH): Delete.
2373 (TREE_INT_CST_NUNITS): New.
2374 (TREE_INT_CST_EXT_NUNITS): Likewise.
2375 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
2376 (TREE_INT_CST_ELT): Likewise.
2377 (INT_CST_LT): Delete.
2378 (tree_int_cst_elt_check): New (two forms).
2379 (type_code_size): Update comment.
2380 (make_int_cst_stat, make_int_cst): New.
2381 (tree_to_double_int): Delete.
2382 (double_int_fits_to_tree_p): Delete.
2383 (force_fit_type_double): Delete.
2384 (build_int_cstu): Replace with out-of-line function.
2385 (build_int_cst_wide): Delete.
2386 (tree_int_cst_lt): Define inline.
2387 (tree_int_cst_le): New.
2388 (tree_int_cst_compare): Define inline.
2389 (tree_int_cst_min_precision): Take a signop rather than a bool.
2390 (wi::int_traits <const_tree>): New.
2391 (wi::int_traits <tree>): New.
2392 (wi::extended_tree): New.
2393 (wi::int_traits <wi::extended_tree>): New.
2394 (wi::to_widest): New.
2395 (wi::to_offset): New.
2396 (wi::fits_to_tree_p): New.
2397 (wi::min_value): New.
2398 (wi::max_value): New.
2399 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
2400 (copy_tree_body_r): Likewise.
2401 * tree-object-size.c (compute_object_offset): Likewise.
2402 (addr_object_size): Likewise.
2403 * tree-predcom.c: Include wide-int-print.h.
2404 (struct dref_d): Change type of offset to widest_int.
2405 (dump_dref): Call wide-int printer.
2406 (aff_combination_dr_offset): Use wide-int interfaces.
2407 (determine_offset): Take a widest_int pointer rather than a
2409 (split_data_refs_to_components): Use wide-int interfaces.
2410 (suitable_component_p): Likewise.
2411 (order_drefs): Likewise.
2412 (add_ref_to_chain): Likewise.
2413 (valid_initializer_p): Likewise.
2414 (determine_roots_comp): Likewise.
2415 * tree-pretty-print.c: Include wide-int-print.h.
2416 (dump_generic_node): Use wide-int interfaces.
2417 * tree-sra.c (sra_ipa_modify_expr): Likewise.
2418 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
2419 (move_fixed_address_to_symbol): Likewise.
2420 (move_hint_to_base): Likewise.
2421 (move_pointer_to_base): Likewise.
2422 (move_variant_to_index): Likewise.
2423 (most_expensive_mult_to_index): Likewise.
2424 (addr_to_parts): Likewise.
2425 (copy_ref_info): Likewise.
2426 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
2427 (indirect_refs_may_alias_p): Likewise.
2428 (stmt_kills_ref_p_1): Likewise.
2429 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
2430 * tree-ssa-ccp.c: Update comment at top of file. Include
2432 (struct prop_value_d): Change type of mask to widest_int.
2433 (extend_mask): New function.
2434 (dump_lattice_value): Use wide-int interfaces.
2435 (get_default_value): Likewise.
2436 (set_constant_value): Likewise.
2437 (set_value_varying): Likewise.
2438 (valid_lattice_transition): Likewise.
2439 (set_lattice_value): Likewise.
2440 (value_to_double_int): Delete.
2441 (value_to_wide_int): New.
2442 (get_value_from_alignment): Use wide-int interfaces.
2443 (get_value_for_expr): Likewise.
2444 (do_dbg_cnt): Likewise.
2445 (ccp_finalize): Likewise.
2446 (ccp_lattice_meet): Likewise.
2447 (bit_value_unop_1): Use widest_ints rather than double_ints.
2448 (bit_value_binop_1): Likewise.
2449 (bit_value_unop): Use wide-int interfaces.
2450 (bit_value_binop): Likewise.
2451 (bit_value_assume_aligned): Likewise.
2452 (evaluate_stmt): Likewise.
2453 (ccp_fold_stmt): Likewise.
2454 (visit_cond_stmt): Likewise.
2455 (ccp_visit_stmt): Likewise.
2456 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
2457 (constant_pointer_difference): Likewise.
2458 (associate_pointerplus): Likewise.
2459 (combine_conversions): Likewise.
2460 * tree-ssa-loop.h: Include wide-int.h.
2461 (struct tree_niter_desc): Change type of max to widest_int.
2462 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
2463 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
2464 (remove_redundant_iv_tests): Likewise.
2465 (canonicalize_loop_induction_variables): Likewise.
2466 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
2467 (constant_multiple_of): Take a widest_int pointer instead of
2468 a double_int pointer.
2469 (get_computation_aff): Use wide-int interfaces.
2470 (ptr_difference_cost): Likewise.
2471 (difference_cost): Likewise.
2472 (get_loop_invariant_expr_id): Likewise.
2473 (get_computation_cost_at): Likewise.
2474 (iv_elimination_compare_lt): Likewise.
2475 (may_eliminate_iv): Likewise.
2476 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
2477 instead of double_int.
2478 (max_loop_iterations): Likewise.
2479 (max_stmt_executions): Likewise.
2480 (estimated_stmt_executions): Likewise.
2481 * tree-ssa-loop-niter.c: Include wide-int-print.h.
2482 (split_to_var_and_offset): Use wide-int interfaces.
2483 (determine_value_range): Likewise.
2484 (bound_difference_of_offsetted_base): Likewise.
2485 (bounds_add): Take a widest_int instead of a double_int.
2486 (number_of_iterations_ne_max): Use wide-int interfaces.
2487 (number_of_iterations_ne): Likewise.
2488 (number_of_iterations_lt_to_ne): Likewise.
2489 (assert_loop_rolls_lt): Likewise.
2490 (number_of_iterations_lt): Likewise.
2491 (number_of_iterations_le): Likewise.
2492 (number_of_iterations_cond): Likewise.
2493 (number_of_iterations_exit): Likewise.
2494 (finite_loop_p): Likewise.
2495 (derive_constant_upper_bound_assign): Likewise.
2496 (derive_constant_upper_bound): Return a widest_int.
2497 (derive_constant_upper_bound_ops): Likewise.
2498 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
2499 (record_estimate): Take a widest_int rather than a double_int.
2500 (record_nonwrapping_iv): Use wide-int interfaces.
2501 (double_int_cmp): Delete.
2502 (wide_int_cmp): New.
2503 (bound_index): Take a widest_int rather than a double_int.
2504 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
2505 (maybe_lower_iteration_bound): Likewise.
2506 (estimate_numbers_of_iterations_loop): Likewise.
2507 (estimated_loop_iterations): Take a widest_int pointer than than
2508 a double_int pointer.
2509 (estimated_loop_iterations_int): Use wide-int interfaces.
2510 (max_loop_iterations): Take a widest_int pointer than than
2511 a double_int pointer.
2512 (max_loop_iterations_int): Use wide-int interfaces.
2513 (max_stmt_executions): Take a widest_int pointer than than
2514 a double_int pointer.
2515 (estimated_stmt_executions): Likewise.
2516 (n_of_executions_at_most): Use wide-int interfaces.
2517 (scev_probably_wraps_p): Likewise.
2518 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
2520 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
2522 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
2523 double_ints. Adjust for trailing_wide_ints <3> representation.
2524 (set_nonzero_bits): Likewise.
2525 (get_range_info): Return wide_ints rather than double_ints.
2526 Adjust for trailing_wide_ints <3> representation.
2527 (get_nonzero_bits): Likewise.
2528 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
2530 * tree-ssanames.h (struct range_info_def): Replace min, max and
2531 nonzero_bits with a trailing_wide_ints <3>.
2532 (set_range_info): Use wide_int_refs rather than double_ints.
2533 (set_nonzero_bits): Likewise.
2534 (get_range_info): Return wide_ints rather than double_ints.
2535 (get_nonzero_bits): Likewise.
2536 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
2537 * tree-ssa-pre.c (phi_translate_1): Likewise.
2538 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
2539 (acceptable_pow_call): Likewise.
2540 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
2542 (vn_reference_fold_indirect): Likewise.
2543 (vn_reference_maybe_forwprop_address): Likewise.
2544 (valueize_refs_1): Likewise.
2545 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
2546 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
2547 tree_int_cst_lt and tree_int_cst_le.
2548 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
2550 (streamer_alloc_tree): Likewise.
2551 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
2552 (streamer_write_tree_header): Likewise.
2553 (streamer_write_integer_cst): Likewise.
2554 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
2555 (build_constructors): Likewise.
2556 (array_value_type): Likewise.
2557 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
2558 (vect_check_gather): Likewise.
2559 * tree-vect-generic.c (build_replicated_const): Likewise.
2560 (expand_vector_divmod): Likewise.
2561 * tree-vect-loop.c (vect_transform_loop): Likewise.
2562 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
2563 (vect_do_peeling_for_alignment): Likewise.
2564 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
2565 * tree-vrp.c: Include wide-int.h.
2566 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
2567 (extract_range_from_assert): Use wide-int interfaces.
2568 (vrp_int_const_binop): Likewise.
2569 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
2570 double_int pointers.
2571 (ranges_from_anti_range): Use wide-int interfaces.
2572 (quad_int_cmp): Delete.
2573 (quad_int_pair_sort): Likewise.
2574 (extract_range_from_binary_expr_1): Use wide-int interfaces.
2575 (extract_range_from_unary_expr_1): Likewise.
2576 (adjust_range_with_scev): Likewise.
2577 (masked_increment): Take and return wide_ints rather than double_ints.
2578 (register_edge_assert_for_2): Use wide-int interfaces.
2579 (check_array_ref): Likewise.
2580 (search_for_addr_array): Likewise.
2581 (maybe_set_nonzero_bits): Likewise.
2582 (union_ranges): Pass an integer of the correct type instead of
2583 using integer_one_node.
2584 (intersect_ranges): Likewise.
2585 (simplify_truth_ops_using_ranges): Likewise.
2586 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
2587 (range_fits_type_p): Likewise.
2588 (simplify_cond_using_ranges): Likewise. Take a signop rather than
2590 (simplify_conversion_using_ranges): Use wide-int interfaces.
2591 (simplify_float_conversion_using_ranges): Likewise.
2592 (vrp_finalize): Likewise.
2593 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
2594 (gimple_stringops_transform): Likewise.
2595 * varasm.c (decode_addr_const): Likewise.
2596 (const_hash_1): Likewise.
2597 (const_rtx_hash_1): Likewise
2598 (output_constant): Likewise.
2599 (array_size_for_constructor): Likewise.
2600 (output_constructor_regular_field): Likewise.
2601 (output_constructor_bitfield): Likewise.
2602 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
2603 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
2605 * gencheck.c: Define BITS_PER_UNIT.
2608 * wide-int-print.cc: New.
2609 * wide-int-print.h: New.
2611 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2613 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
2615 2014-05-06 Richard Biener <rguenther@suse.de>
2617 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
2618 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
2619 (TODO_verify_all): Adjust.
2620 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
2621 TODO_verify_stmts and TODO_verify_rtl_sharing.
2622 * bb-reorder.c: Likewise.
2623 * cfgexpand.c: Likewise.
2624 * cprop.c: Likewise.
2626 * function.c: Likewise.
2627 * fwprop.c: Likewise.
2629 * gimple-ssa-isolate-paths.c: Likewise.
2630 * gimple-ssa-strength-reduction.c: Likewise.
2631 * ipa-split.c: Likewise.
2632 * loop-init.c: Likewise.
2633 * loop-unroll.c: Likewise.
2634 * lower-subreg.c: Likewise.
2635 * modulo-sched.c: Likewise.
2636 * postreload-gcse.c: Likewise.
2637 * predict.c: Likewise.
2638 * recog.c: Likewise.
2639 * sched-rgn.c: Likewise.
2640 * store-motion.c: Likewise.
2641 * tracer.c: Likewise.
2642 * trans-mem.c: Likewise.
2643 * tree-call-cdce.c: Likewise.
2644 * tree-cfg.c: Likewise.
2645 * tree-cfgcleanup.c: Likewise.
2646 * tree-complex.c: Likewise.
2647 * tree-eh.c: Likewise.
2648 * tree-emutls.c: Likewise.
2649 * tree-if-conv.c: Likewise.
2650 * tree-into-ssa.c: Likewise.
2651 * tree-loop-distribution.c: Likewise.
2652 * tree-object-size.c: Likewise.
2653 * tree-parloops.c: Likewise.
2654 * tree-pass.h: Likewise.
2655 * tree-sra.c: Likewise.
2656 * tree-ssa-ccp.c: Likewise.
2657 * tree-ssa-copy.c: Likewise.
2658 * tree-ssa-copyrename.c: Likewise.
2659 * tree-ssa-dce.c: Likewise.
2660 * tree-ssa-dom.c: Likewise.
2661 * tree-ssa-dse.c: Likewise.
2662 * tree-ssa-forwprop.c: Likewise.
2663 * tree-ssa-ifcombine.c: Likewise.
2664 * tree-ssa-loop-ch.c: Likewise.
2665 * tree-ssa-loop-ivcanon.c: Likewise.
2666 * tree-ssa-loop.c: Likewise.
2667 * tree-ssa-math-opts.c: Likewise.
2668 * tree-ssa-phiopt.c: Likewise.
2669 * tree-ssa-phiprop.c: Likewise.
2670 * tree-ssa-pre.c: Likewise.
2671 * tree-ssa-reassoc.c: Likewise.
2672 * tree-ssa-sink.c: Likewise.
2673 * tree-ssa-strlen.c: Likewise.
2674 * tree-ssa-tail-merge.c: Likewise.
2675 * tree-ssa-uncprop.c: Likewise.
2676 * tree-switch-conversion.c: Likewise.
2677 * tree-tailcall.c: Likewise.
2678 * tree-vect-generic.c: Likewise.
2679 * tree-vectorizer.c: Likewise.
2680 * tree-vrp.c: Likewise.
2682 * var-tracking.c: Likewise.
2683 * bt-load.c: Likewise.
2684 * cfgcleanup.c: Likewise.
2685 * combine-stack-adj.c: Likewise.
2686 * combine.c: Likewise.
2687 * compare-elim.c: Likewise.
2688 * config/epiphany/resolve-sw-modes.c: Likewise.
2689 * config/i386/i386.c: Likewise.
2690 * config/mips/mips.c: Likewise.
2691 * config/s390/s390.c: Likewise.
2692 * config/sh/sh_treg_combine.cc: Likewise.
2693 * config/sparc/sparc.c: Likewise.
2696 * final.c: Likewise.
2697 * ifcvt.c: Likewise.
2698 * mode-switching.c: Likewise.
2699 * passes.c: Likewise.
2700 * postreload.c: Likewise.
2702 * reg-stack.c: Likewise.
2703 * regcprop.c: Likewise.
2704 * regrename.c: Likewise.
2707 2014-05-06 Richard Biener <rguenther@suse.de>
2710 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
2711 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
2713 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
2716 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
2718 2014-05-05 Radovan Obradovic <robradovic@mips.com>
2719 Tom de Vries <tom@codesourcery.com>
2721 * target.def (call_fusage_contains_non_callee_clobbers): New
2723 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
2725 (@node Miscellaneous Register Hooks): New node.
2726 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
2727 * doc/tm.texi: Regenerate.
2729 2014-05-05 Marek Polacek <polacek@redhat.com>
2732 * opts.c (common_handle_option): Call error_at instead of warning_at.
2734 2014-05-05 Richard Biener <rguenther@suse.de>
2736 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
2737 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
2738 under the TODO_verify_il umbrella.
2740 2014-05-05 Richard Biener <rguenther@suse.de>
2742 * passes.c (execute_function_todo): Move TODO_verify_flow under
2743 the TODO_verify_ul umbrella.
2745 2014-05-05 Richard Biener <rguenther@suse.de>
2748 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
2749 X & CST away from a CST that is the mask of a mode.
2751 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2753 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
2754 int argument to enum machine_mode.
2755 (picochip_class_max_nregs): Ditto.
2756 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
2757 (picochip_class_max_nregs): Ditto.
2759 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2761 * target.def: Add new target hook.
2762 * doc/tm.texi: Regenerate.
2763 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
2764 * targhooks.c (default_keep_leaf_when_profiled): New function.
2766 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
2767 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
2769 2014-05-05 Bin Cheng <bin.cheng@arm.com>
2771 PR tree-optimization/60363
2772 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
2773 (copy_phi_args): New parameters. Call get_value_locus_in_path.
2774 (update_destination_phis): New parameter.
2775 (create_edge_and_update_destination_phis): Ditto.
2776 (ssa_fix_duplicate_block_edges): Pass new arguments.
2777 (thread_single_edge): Ditto.
2779 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
2781 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
2782 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
2783 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
2784 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
2785 Use RS6000_BTM_HARD_FLOAT.
2786 (BU_MISC_2): Likewise.
2787 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
2788 RS6000_BTM_HARD_FLOAT.
2789 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
2791 (rs6000_invalid_builtin): Add hard floating builtin support.
2792 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
2793 hard float builtins.
2794 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
2796 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2798 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
2799 Add missing function* argument.
2801 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2803 * lra-constraints.c (valid_address_p): Move earlier in file.
2804 Add a constraint argument to the address_info version.
2805 (satisfies_memory_constraint_p): New function.
2806 (satisfies_address_constraint_p): Likewise.
2807 (process_alt_operands, curr_insn_transform): Use them.
2808 (process_address): Pass the constraint to valid_address_p when
2809 checking address operands.
2811 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2813 * config/mips/mips.c (mips_isa_rev): New variable.
2814 (mips_set_architecture): Set it.
2815 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
2817 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
2818 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
2819 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
2820 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
2821 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
2822 conditions in terms of mips_isa_rev.
2823 (mips_isa_rev): Declare.
2825 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2827 * config/sh/sh-mem.cc: Use tabs instead of spaces.
2828 (prob_unlikely, prob_likely): Make variables const.
2830 2014-05-03 Denis Chertykov <chertykov@gmail.com>
2832 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
2834 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2836 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
2838 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2840 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
2841 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
2842 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
2844 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
2845 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
2847 Replace usage of ROUND_REG with sh_round_reg.
2848 Use CEIL instead of ROUND_ADVANCE.
2850 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2853 * config/sh/sh.c: Include stdlib headers before everything else.
2855 2014-05-02 Jakub Jelinek <jakub@redhat.com>
2857 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
2858 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
2859 (gimplify_adjust_omp_clauses): Simd region is never
2860 directly nested in combined parallel. Instead, for linear
2861 with copyin/copyout, if in combined for simd loop, make decl
2862 firstprivate/lastprivate on OMP_FOR.
2863 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
2864 expand_omp_for_static_chunk): When setting endvar, also set
2865 fd->loop.v to the same value.
2867 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2869 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
2871 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
2873 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
2876 2014-05-02 Marek Polacek <polacek@redhat.com>
2878 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
2880 2014-05-02 Kito Cheng <kito@0xlab.org>
2882 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
2883 to a C expression marco.
2884 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
2885 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
2886 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
2887 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
2888 HONOR_REG_ALLOC_ORDER.
2889 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
2891 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2893 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
2895 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2897 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
2899 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
2901 * tree-if-conv.c (is_cond_scalar_reduction): New function.
2902 (convert_scalar_cond_reduction): Likewise.
2903 (predicate_scalar_phi): Add recognition and transformation
2904 of simple conditioanl reduction to be vectorizable.
2906 2014-05-01 Marek Polacek <polacek@redhat.com>
2909 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
2911 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
2913 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
2914 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
2915 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
2916 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
2917 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
2918 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
2919 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
2920 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
2922 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
2924 * config/arc/arc.opt (mlra): Move comment above option name
2925 to avoid mis-parsing as language options.
2927 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2929 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
2930 * config/sol2.h: ... here.
2931 * config/sol2-10.h: Remove.
2933 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
2934 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
2935 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
2936 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
2937 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
2938 * config/sol2.h: ... here.
2939 (SECTION_NAME_FORMAT): Don't redefine.
2940 (STARTFILE_ARCH32_SPEC): Rename to ...
2941 (STARTFILE_ARCH_SPEC): ... this.
2942 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
2943 * config/sparc/sol2.h: ... here.
2944 (SECTION_NAME_FORMAT): Don't undef.
2945 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
2946 (SUBTARGET_EXTRA_SPECS): Remove.
2947 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
2949 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
2950 (MD_STARTFILE_PREFIX): Remove.
2951 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
2952 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
2953 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
2954 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
2955 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
2956 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
2957 * config/i386/sol2.h: ... here.
2958 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
2959 * config/i386/sol2-bi.h: Remove.
2960 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
2961 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
2963 * config/i386/t-sol2-64: Rename to ...
2964 * config/i386/t-sol2: ... this.
2965 * config/sparc/t-sol2-64: Rename to ...
2966 * config/sparc/t-sol2: ... this.
2968 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
2969 sol2_tm_file_head, sol2_tm_file_tail.
2970 Include ${cpu_type}/sol2.h before sol2.h.
2972 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
2973 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
2974 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
2975 Reflect i386/t-sol2-64 renaming.
2976 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
2977 Reflect sparc/t-sol2-64 renaming.
2979 2014-04-30 Richard Biener <rguenther@suse.de>
2981 * passes.c (execute_function_todo): Move TODO_verify_stmts
2982 and TODO_verify_ssa under the TODO_verify_il umbrella.
2983 * tree-ssa.h (verify_ssa): Adjust prototype.
2984 * tree-ssa.c (verify_ssa): Add parameter to tell whether
2985 we should verify SSA operands.
2986 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
2987 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
2988 whether we should verify whether not throwing stmts have EH info.
2989 * graphite-scop-detection.c (create_sese_edges): Adjust.
2990 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
2991 * tree-eh.c (lower_try_finally_switch): Do not add the
2992 default case label twice.
2994 2014-04-30 Marek Polacek <polacek@redhat.com>
2996 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
2997 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
2998 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
2999 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
3001 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
3003 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
3004 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
3005 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
3006 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
3007 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
3008 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
3009 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
3010 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
3012 2014-04-29 David Malcolm <dmalcolm@redhat.com>
3014 * tree-cfg.c (dump_function_to_file): Dump the return type of
3015 functions, in a line to itself before the function body, mimicking
3016 the layout of a C function.
3018 2014-04-29 Jakub Jelinek <jakub@redhat.com>
3020 PR tree-optimization/60971
3021 * tree-tailcall.c (process_assignment): Reject conversions which
3024 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
3026 * calls.c (initialize_argument_information): Always treat
3027 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
3028 (expand_call): Likewise.
3029 (emit_library_call_calue_1): Likewise.
3030 * expr.c (PUSH_ARGS_REVERSED): Do not define.
3031 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
3034 2014-04-29 Nick Clifton <nickc@redhat.com>
3036 * config/msp430/msp430.md (umulsidi): Fix typo.
3037 (mulhisi3): Enable even inside interrupt handlers.
3038 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
3039 bigger return address pushed in large mode.
3041 2014-04-29 Nick Clifton <nickc@redhat.com>
3043 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
3044 (arc_init_reg_tables): Use a machine_mode enum to iterate over
3046 * config/m32r/m32r.c (init_reg_tables): Likewise.
3047 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
3048 enum to hold the modes.
3050 2014-04-29 Richard Biener <rguenther@suse.de>
3052 * dominance.c (free_dominance_info): Add overload with
3054 (dom_info_state): Likewise.
3055 (dom_info_available_p): Likewise.
3056 * basic-block.h (free_dominance_info, dom_info_state,
3057 dom_info_available_p): Declare overloads.
3058 * passes.c (execute_function_todo): Verify that verifiers
3059 don't change dominator info state. Drop dominator info
3060 for IPA pass invocations.
3061 * cgraph.c (release_function_body): Restore asserts that
3062 dominator information is released.
3064 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
3066 * doc/invoke.texi: Fix typo.
3067 * tree-vrp.c: Fix typos.
3068 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
3070 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3072 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
3074 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
3076 * config/aarch64/aarch64-builtins.c
3077 (aarch64_types_storestruct_lane_qualifiers): New.
3078 (TYPES_STORESTRUCT_LANE): Likewise.
3079 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
3080 (st3_lane): Likewise.
3081 (st4_lane): Likewise.
3082 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
3083 (vec_store_lanesci_lane<mode>): Likewise.
3084 (vec_store_lanesxi_lane<mode>): Likewise.
3085 (aarch64_st2_lane<VQ:mode>): Likewise.
3086 (aarch64_st3_lane<VQ:mode>): Likewise.
3087 (aarch64_st4_lane<VQ:mode>): Likewise.
3088 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
3089 * config/aarch64/arm_neon.h
3090 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
3091 use new macro arguments.
3092 (__ST3_LANE_FUNC): Likewise.
3093 (__ST4_LANE_FUNC): Likewise.
3094 * config/aarch64/iterators.md (V_TWO_ELEM): New.
3095 (V_THREE_ELEM): Likewise.
3096 (V_FOUR_ELEM): Likewise.
3098 2014-04-28 David Malcolm <dmalcolm@redhat.com>
3100 * doc/gimple.texi: Replace the description of the now-defunct
3101 union gimple_statement_d with a diagram showing the
3102 gimple_statement_base class hierarchy and its relationships to
3103 the GSS_ and GIMPLE_ enums.
3105 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
3107 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
3108 * config/aarch64/aarch64.c
3109 (aarch64_cannot_change_mode_class): Weaken conditions.
3110 (aarch64_modes_tieable_p): New.
3111 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
3113 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
3115 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
3116 (loadsync_<mode>): Change mode.
3117 (load_quadpti, store_quadpti): New.
3118 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
3119 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
3121 2014-04-28 Martin Jambor <mjambor@suse.cz>
3123 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
3124 same alias type as the original statement.
3125 (subreplacement_assignment_data): New type.
3126 (handle_unscalarized_data_in_subtree): New type of parameter,
3127 generate new memory accesses with same alias type as the original
3129 (load_assign_lhs_subreplacements): Likewise.
3130 (sra_modify_constructor_assign): Generate new memory accesses with
3131 same alias type as the original statement.
3133 2014-04-28 Richard Biener <rguenther@suse.de>
3135 * tree-pass.h (TODO_verify_il): Define.
3136 (TODO_verify_all): Complete properly.
3137 * passes.c (execute_function_todo): Move existing loop-closed
3138 SSA verification under TODO_verify_il.
3139 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
3140 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
3141 Fix tree sharing issue.
3143 2014-04-28 Richard Biener <rguenther@suse.de>
3146 * builtins.def (DEF_C11_BUILTIN): Add.
3147 (BUILT_IN_ALIGNED_ALLOC): Likewise.
3148 * coretypes.h (enum function_class): Add function_c11_misc.
3149 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
3150 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
3151 (call_may_clobber_ref_p_1): Likewise.
3152 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
3153 (mark_all_reaching_defs_necessary_1): Likewise.
3154 (propagate_necessity): Likewise.
3155 (eliminate_unnecessary_stmts): Likewise.
3156 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
3158 2014-04-28 Richard Biener <rguenther@suse.de>
3160 * tree-vrp.c (vrp_var_may_overflow): Remove.
3161 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
3162 with overflow immediately bump to one before that value and
3163 let iteration figure out overflow status.
3165 2014-04-28 Richard Biener <rguenther@suse.de>
3167 * configure.ac: Do valgrind header checks unconditionally.
3168 Add --enable-valgrind-annotations.
3169 * system.h: Guard valgrind header inclusion with
3170 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
3171 * alloc-pool.c (pool_alloc, pool_free): Use
3172 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
3173 to guard possibly dead code.
3174 * config.in: Regenerated.
3175 * configure: Likewise.
3177 2014-04-28 Jeff Law <law@redhat.com>
3179 PR tree-optimization/60902
3180 * tree-ssa-threadedge.c
3181 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
3182 over real defs when invalidating outputs from statements that do not
3183 produce useful outputs for threading.
3185 2014-04-28 Richard Biener <rguenther@suse.de>
3187 PR tree-optimization/60979
3188 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
3189 SCOPs that end in a block with a successor with abnormal
3192 2014-04-28 Richard Biener <rguenther@suse.de>
3194 * tree-pass.h (execute_pass_list): Adjust prototype.
3195 * passes.c (pass_manager::execute_early_local_passes): Adjust.
3196 (do_per_function): Change callback signature, push all actual
3197 work to the callbals.
3198 (do_per_function_toporder): Likewise.
3199 (execute_function_dump): Adjust.
3200 (execute_function_todo): Likewise.
3201 (clear_last_verified): Likewise.
3202 (verify_curr_properties): Likewise.
3203 (update_properties_after_pass): Likewise.
3204 (execute_pass_list_1): Split out from ...
3205 (execute_pass_list): ... here. Adjust.
3206 (execute_ipa_pass_list): Likewise.
3207 * cgraphunit.c (cgraph_add_new_function): Adjust.
3208 (analyze_function): Likewise.
3209 (expand_function): Likewise.
3210 * cgraph.c (release_function_body): Free dominance info
3211 here instead of asserting it was magically freed elsewhere.
3213 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
3215 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
3216 * configure: Regenerate.
3217 * config/sparc/sparc.opt (muser-mode): New option.
3218 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
3220 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
3221 * doc/invoke.texi (SPARC options): Document -muser-mode.
3223 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
3225 * cselib.c (find_slot_memmode): Delete.
3226 (cselib_hasher): Change compare_type to a struct.
3227 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
3229 (preserve_constants_and_equivs): Adjust for new compare_type.
3230 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
3231 (wrap_constant): Delete.
3232 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
3234 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
3236 * doc/install.texi (Building with profile feedback): Remove
3239 2014-04-26 Tom de Vries <tom@codesourcery.com>
3241 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
3244 2014-04-25 Cary Coutant <ccoutant@google.com>
3247 * dwarf2out.c (should_move_die_to_comdat): A type definition
3248 can contain a subprogram definition, but don't move it to a
3250 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
3251 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
3253 (clone_tree_hash): Rename to...
3254 (clone_tree_partial): ...this; change callers. Copy
3255 DW_TAG_subprogram DIEs as declarations.
3256 (copy_decls_walk): Don't copy children of a declaration into a
3259 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
3262 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
3265 2014-04-25 Jiong Wang <jiong.wang@arm.com>
3267 * config/arm/predicates.md (call_insn_operand): Add long_call check.
3268 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
3270 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
3273 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3275 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
3277 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3279 PR tree-optimization/60930
3280 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
3281 creating a multiply candidate by folding two constant
3282 multiplicands when the result overflows.
3284 2014-04-25 Jakub Jelinek <jakub@redhat.com>
3286 PR tree-optimization/60960
3287 * tree-vect-generic.c (expand_vector_operation): Only call
3288 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
3290 2014-04-25 Tom de Vries <tom@codesourcery.com>
3292 * expr.c (clobber_reg_mode): New function.
3293 * expr.h (clobber_reg): New function.
3295 2014-04-25 Tom de Vries <tom@codesourcery.com>
3297 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
3300 2014-04-25 Radovan Obradovic <robradovic@mips.com>
3301 Tom de Vries <tom@codesourcery.com>
3303 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
3305 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
3306 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
3307 new argument to find_all_hard_reg_sets call.
3309 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3311 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
3312 Use HOST_WIDE_INT_C for mask literal.
3313 (aarch_rev16_shleft_mask_imm_p): Likewise.
3315 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
3318 * config/sparc/sparc.md (ashlsi3_extend): Delete.
3320 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
3322 PR preprocessor/56540
3323 * config/i386/i386-c.c (ix86_target_macros): Define
3324 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
3326 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3328 * configure.ac (tga_func): Remove.
3329 (LIB_TLS_SPEC): Remove.
3330 * configure: Regenerate.
3331 * config.in: Regenerate.
3332 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
3334 2014-04-25 Richard Biener <rguenther@suse.de>
3337 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
3338 call stmt use/clobber sets during stmt walk instead of
3339 walking the possibly incomplete set of caller edges.
3341 2014-04-25 Richard Biener <rguenther@suse.de>
3344 * passes.c (apply_ipa_transforms): Inline into only caller ...
3345 (execute_one_pass): ... here. Properly bring in function
3346 bodies for nodes we want to apply IPA transforms to.
3348 2014-04-24 Cong Hou <congh@google.com>
3350 PR tree-optimization/60896
3351 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
3352 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
3353 (vect_mark_pattern_stmts): Set the def type of all statements in
3354 PATTERN_DEF_SEQ as vect_internal_def.
3356 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
3358 * doc/extend.texi (PowerPC Built-in Functions): Document new
3359 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
3360 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
3362 * config/rs6000/predicates.md (const_0_to_3_operand): New
3363 predicate to match 0..3 integer constants.
3365 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
3366 to support adding miscellaneous builtin functions.
3367 (BU_DFP_MISC_2): Likewise.
3368 (BU_P7_MISC_1): Likewise.
3369 (BU_P7_MISC_2): Likewise.
3370 (BU_P8V_MISC_3): Likewise.
3371 (BU_MISC_1): Likewise.
3372 (BU_MISC_2): Likewise.
3373 (DIVWE): Add extended divide builtin functions.
3376 (DIVWEUO): Likewise.
3380 (DIVDEUO): Likewise.
3381 (DXEX): Add decimal floating-point builtin functions.
3384 (DDEDPDQ): Likewise.
3386 (DENBCDQ): Likewise.
3393 (CDTBCD): Add new BCD builtin functions.
3397 (BCDADD_LT): Likewise.
3398 (BCDADD_EQ): Likewise.
3399 (BCDADD_GT): Likewise.
3400 (BCDADD_OV): Likewise.
3402 (BCDSUB_LT): Likewise.
3403 (BCDSUB_EQ): Likewise.
3404 (BCDSUB_GT): Likewise.
3405 (BCDSUB_OV): Likewise.
3406 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
3407 (UNPACK_TD): Likewise.
3408 (PACK_TF): Likewise.
3409 (UNPACK_TF): Likewise.
3410 (UNPACK_TF_0): Likewise.
3411 (UNPACK_TF_1): Likewise.
3412 (PACK_V1TI): Likewise.
3413 (UNPACK_V1TI): Likewise.
3415 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
3416 support for decimal floating point builtin functions.
3417 (rs6000_expand_ternop_builtin): Add checks for the new builtin
3418 functions that take constant arguments.
3419 (rs6000_invalid_builtin): Add decimal floating point builtin support.
3420 (rs6000_init_builtins): Setup long double, _Decimal64, and
3421 _Decimal128 types for new builtin functions.
3422 (builtin_function_type): Set the unsigned flags appropriately for
3423 the new builtin functions.
3424 (rs6000_opt_masks): Add support for decimal floating point builtin
3427 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
3428 floating point builtin functions.
3429 (RS6000_BTM_COMMON): Likewise.
3430 (RS6000_BTI_long_double): Likewise.
3431 (RS6000_BTI_dfloat64): Likewise.
3432 (RS6000_BTI_dfloat128): Likewise.
3433 (long_double_type_internal_node): Likewise.
3434 (dfloat64_type_internal_node): Likewise.
3435 (dfloat128_type_internal_node): Likewise.
3437 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
3438 2.07 bcd arithmetic instructions.
3439 (UNSPEC_BCDSUB): Likewise.
3440 (UNSPEC_BCD_OVERFLOW): Likewise.
3441 (UNSPEC_BCD_ADD_SUB): Likewise.
3442 (bcd_add_sub): Likewise.
3443 (BCD_TEST): Likewise.
3444 (bcd<bcd_add_sub>): Likewise.
3445 (bcd<bcd_add_sub>_test): Likewise.
3446 (bcd<bcd_add_sub>_test2): Likewise.
3447 (bcd<bcd_add_sub>_<code>): Likewise.
3448 (peephole2 for combined bcd ops): Likewise.
3450 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
3451 decimal floating point builtin functions.
3452 (UNSPEC_DENBCD): Likewise.
3453 (UNSPEC_DXEX): Likewise.
3454 (UNSPEC_DIEX): Likewise.
3455 (UNSPEC_DSCLI): Likewise.
3456 (UNSPEC_DSCRI): Likewise.
3457 (D64_D128): Likewise.
3458 (dfp_suffix): Likewise.
3459 (dfp_ddedpd_<mode>): Likewise.
3460 (dfp_denbcd_<mode>): Likewise.
3461 (dfp_dxex_<mode>): Likewise.
3462 (dfp_diex_<mode>): Likewise.
3463 (dfp_dscli_<mode>): Likewise.
3464 (dfp_dscri_<mode>): Likewise.
3466 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
3468 (UNSPEC_CDTBCD): Likewise.
3469 (UNSPEC_CBCDTD): Likewise.
3470 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
3471 (UNSPEC_DIVEO): Likewise.
3472 (UNSPEC_DIVEU): Likewise.
3473 (UNSPEC_DIVEUO): Likewise.
3474 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
3475 pack/unpack 128-bit types.
3476 (UNSPEC_PACK_128BIT): Likewise.
3477 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
3478 (udiv<mode>3): Use idiv_ldiv mode attribute.
3479 (div<mode>3): Likewise.
3480 (addg6s): Add new BCD builtin functions.
3483 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
3484 (div_extend): Likewise.
3485 (div<div_extend>_<mode>"): Likewise.
3486 (FP128_64): Add support for new builtin functions to pack/unpack
3488 (unpack<mode>): Likewise.
3489 (unpacktf_0): Likewise.
3490 (unpacktf_1): Likewise.
3491 (unpack<mode>_dm): Likewise.
3492 (unpack<mode>_nodm): Likewise.
3493 (pack<mode>): Likewise.
3494 (unpackv1ti): Likewise.
3495 (packv1ti): Likewise.
3497 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
3499 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
3502 2014-04-24 Jakub Jelinek <jakub@redhat.com>
3504 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
3505 * gimplify.c (omp_is_private): Change last argument's type to int.
3506 Only diagnose lastprivate if the simd argument is 1, only diagnose
3507 linear if the simd argument is 2.
3508 (gimplify_omp_for): Adjust omp_is_private callers. When adding
3509 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
3510 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
3511 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
3512 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
3513 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
3514 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
3515 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
3516 * tree-nested.c (convert_nonlocal_omp_clauses,
3517 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
3519 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
3522 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
3525 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
3527 * flag-types.h (enum ivar_visibility): Add.
3529 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
3531 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
3532 function * argument.
3534 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
3536 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
3538 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3539 Tom de Vries <tom@codesourcery.com>
3541 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
3542 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
3544 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
3545 * emit-rtl.c (try_split): Same.
3547 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3548 Tom de Vries <tom@codesourcery.com>
3550 * common.opt (fuse-caller-save): New option.
3552 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
3554 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
3555 elements for big-endian.
3557 2014-04-24 Richard Biener <rguenther@suse.de>
3559 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
3560 during TER and instead use the sepops interface for expanding
3561 non-GIMPLE_SINGLE_RHS.
3563 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3565 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
3566 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
3568 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3570 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
3571 assembler 64-bit option.
3572 * configure: Regenerate.
3574 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3576 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
3577 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
3578 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
3579 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
3580 (TARGET_CRYPTO): Take TARGET_SIMD into account.
3582 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3584 * config/aarch64/aarch64-builtins.c
3585 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
3586 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
3587 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
3588 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
3590 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
3591 (Vrevsuff): New mode attribute.
3593 2014-04-24 Terry Guo <terry.guo@arm.com>
3595 * config/arm/arm.h (machine_function): Define variable
3596 after_arm_reorg here.
3597 * config/arm/arm.c (after_arm_reorg): Remove the definition.
3598 (arm_split_constant): Update the way to access variable
3601 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
3603 2014-04-23 Tom de Vries <tom@codesourcery.com>
3605 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
3607 2014-04-23 David Malcolm <dmalcolm@redhat.com>
3609 * is-a.h: Update comments to reflect the following changes to the
3610 "pointerness" of the API, making the template parameter match the
3611 return type, allowing use of is-a.h with typedefs of pointers.
3612 (is_a_helper::cast): Return a T rather then a pointer to a T, so
3613 that the return type matches the parameter to the is_a_helper.
3615 (dyn_cast): Likewise.
3617 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
3618 pointer from the is-a.h API.
3620 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
3621 (is_a_helper <cgraph_node *>::test): ...this, matching change to
3623 (is_a_helper <varpool_node>::test): Likewise, convert to...
3624 (is_a_helper <varpool_node *>::test): ...this.
3626 (varpool_first_variable): Update for removal of implicit pointer
3627 from the is-a.h API.
3628 (varpool_next_variable): Likewise.
3629 (varpool_first_static_initializer): Likewise.
3630 (varpool_next_static_initializer): Likewise.
3631 (varpool_first_defined_variable): Likewise.
3632 (varpool_next_defined_variable): Likewise.
3633 (cgraph_first_defined_function): Likewise.
3634 (cgraph_next_defined_function): Likewise.
3635 (cgraph_first_function): Likewise.
3636 (cgraph_next_function): Likewise.
3637 (cgraph_first_function_with_gimple_body): Likewise.
3638 (cgraph_next_function_with_gimple_body): Likewise.
3639 (cgraph_alias_target): Likewise.
3640 (varpool_alias_target): Likewise.
3641 (cgraph_function_or_thunk_node): Likewise.
3642 (varpool_variable_node): Likewise.
3643 (symtab_real_symbol_p): Likewise.
3644 * cgraphunit.c (referred_to_p): Likewise.
3645 (analyze_functions): Likewise.
3646 (handle_alias_pairs): Likewise.
3647 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
3648 * gimple-ssa.h (gimple_vuse_op): Likewise.
3649 (gimple_vdef_op): Likewise.
3650 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
3651 * gimple.c (gimple_build_asm_1): Likewise.
3652 (gimple_build_try): Likewise.
3653 (gimple_build_resx): Likewise.
3654 (gimple_build_eh_dispatch): Likewise.
3655 (gimple_build_omp_for): Likewise.
3656 (gimple_omp_for_set_clauses): Likewise.
3658 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
3659 (is_a_helper <gimple_statement_asm *>::test): ...this.
3660 (is_a_helper <gimple_statement_bind>::test): Convert to...
3661 (is_a_helper <gimple_statement_bind *>::test): ...this.
3662 (is_a_helper <gimple_statement_call>::test): Convert to...
3663 (is_a_helper <gimple_statement_call *>::test): ...this.
3664 (is_a_helper <gimple_statement_catch>::test): Convert to...
3665 (is_a_helper <gimple_statement_catch *>::test): ...this.
3666 (is_a_helper <gimple_statement_resx>::test): Convert to...
3667 (is_a_helper <gimple_statement_resx *>::test): ...this.
3668 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
3669 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
3670 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
3671 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
3672 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
3673 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
3674 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
3675 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
3676 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
3677 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
3678 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
3679 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
3680 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
3681 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
3682 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
3683 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
3684 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
3685 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
3686 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
3687 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
3688 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
3689 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
3690 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
3691 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
3692 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
3693 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
3694 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
3695 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
3696 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
3697 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
3698 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
3699 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
3700 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
3701 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
3702 (is_a_helper <gimple_statement_phi>::test): Convert to...
3703 (is_a_helper <gimple_statement_phi *>::test): ...this.
3704 (is_a_helper <gimple_statement_transaction>::test): Convert to...
3705 (is_a_helper <gimple_statement_transaction *>::test): ...this.
3706 (is_a_helper <gimple_statement_try>::test): Convert to...
3707 (is_a_helper <gimple_statement_try *>::test): ...this.
3708 (is_a_helper <gimple_statement_wce>::test): Convert to...
3709 (is_a_helper <gimple_statement_wce *>::test): ...this.
3710 (is_a_helper <const gimple_statement_asm>::test): Convert to...
3711 (is_a_helper <const gimple_statement_asm *>::test): ...this.
3712 (is_a_helper <const gimple_statement_bind>::test): Convert to...
3713 (is_a_helper <const gimple_statement_bind *>::test): ...this.
3714 (is_a_helper <const gimple_statement_call>::test): Convert to...
3715 (is_a_helper <const gimple_statement_call *>::test): ...this.
3716 (is_a_helper <const gimple_statement_catch>::test): Convert to...
3717 (is_a_helper <const gimple_statement_catch *>::test): ...this.
3718 (is_a_helper <const gimple_statement_resx>::test): Convert to...
3719 (is_a_helper <const gimple_statement_resx *>::test): ...this.
3720 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
3721 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
3722 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
3723 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
3724 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
3726 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
3728 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
3730 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
3732 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
3733 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
3734 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
3736 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
3737 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
3739 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
3740 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
3741 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
3742 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
3743 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
3744 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
3746 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
3747 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
3748 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
3749 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
3751 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
3752 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
3753 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
3754 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
3755 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
3756 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
3757 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
3758 (is_a_helper <const gimple_statement_phi>::test): Convert to...
3759 (is_a_helper <const gimple_statement_phi *>::test): ...this.
3760 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
3761 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
3762 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
3763 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
3764 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
3765 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
3766 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
3768 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
3770 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
3771 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
3773 (gimple_use_ops): Update for removal of implicit pointer from the
3775 (gimple_set_use_ops): Likewise.
3776 (gimple_vuse): Likewise.
3777 (gimple_vdef): Likewise.
3778 (gimple_vuse_ptr): Likewise.
3779 (gimple_vdef_ptr): Likewise.
3780 (gimple_set_vuse): Likewise.
3781 (gimple_set_vdef): Likewise.
3782 (gimple_omp_return_set_lhs): Likewise.
3783 (gimple_omp_return_lhs): Likewise.
3784 (gimple_omp_return_lhs_ptr): Likewise.
3785 (gimple_call_fntype): Likewise.
3786 (gimple_call_set_fntype): Likewise.
3787 (gimple_call_set_internal_fn): Likewise.
3788 (gimple_call_use_set): Likewise.
3789 (gimple_call_clobber_set): Likewise.
3790 (gimple_bind_vars): Likewise.
3791 (gimple_bind_set_vars): Likewise.
3792 (gimple_bind_body_ptr): Likewise.
3793 (gimple_bind_set_body): Likewise.
3794 (gimple_bind_add_stmt): Likewise.
3795 (gimple_bind_block): Likewise.
3796 (gimple_bind_set_block): Likewise.
3797 (gimple_asm_ninputs): Likewise.
3798 (gimple_asm_noutputs): Likewise.
3799 (gimple_asm_nclobbers): Likewise.
3800 (gimple_asm_nlabels): Likewise.
3801 (gimple_asm_input_op): Likewise.
3802 (gimple_asm_input_op_ptr): Likewise.
3803 (gimple_asm_output_op): Likewise.
3804 (gimple_asm_output_op_ptr): Likewise.
3805 (gimple_asm_set_output_op): Likewise.
3806 (gimple_asm_clobber_op): Likewise.
3807 (gimple_asm_set_clobber_op): Likewise.
3808 (gimple_asm_label_op): Likewise.
3809 (gimple_asm_set_label_op): Likewise.
3810 (gimple_asm_string): Likewise.
3811 (gimple_catch_types): Likewise.
3812 (gimple_catch_types_ptr): Likewise.
3813 (gimple_catch_handler_ptr): Likewise.
3814 (gimple_catch_set_types): Likewise.
3815 (gimple_catch_set_handler): Likewise.
3816 (gimple_eh_filter_types): Likewise.
3817 (gimple_eh_filter_types_ptr): Likewise.
3818 (gimple_eh_filter_failure_ptr): Likewise.
3819 (gimple_eh_filter_set_types): Likewise.
3820 (gimple_eh_filter_set_failure): Likewise.
3821 (gimple_eh_must_not_throw_fndecl): Likewise.
3822 (gimple_eh_must_not_throw_set_fndecl): Likewise.
3823 (gimple_eh_else_n_body_ptr): Likewise.
3824 (gimple_eh_else_e_body_ptr): Likewise.
3825 (gimple_eh_else_set_n_body): Likewise.
3826 (gimple_eh_else_set_e_body): Likewise.
3827 (gimple_try_eval_ptr): Likewise.
3828 (gimple_try_cleanup_ptr): Likewise.
3829 (gimple_try_set_eval): Likewise.
3830 (gimple_try_set_cleanup): Likewise.
3831 (gimple_wce_cleanup_ptr): Likewise.
3832 (gimple_wce_set_cleanup): Likewise.
3833 (gimple_phi_capacity): Likewise.
3834 (gimple_phi_num_args): Likewise.
3835 (gimple_phi_result): Likewise.
3836 (gimple_phi_result_ptr): Likewise.
3837 (gimple_phi_set_result): Likewise.
3838 (gimple_phi_arg): Likewise.
3839 (gimple_phi_set_arg): Likewise.
3840 (gimple_resx_region): Likewise.
3841 (gimple_resx_set_region): Likewise.
3842 (gimple_eh_dispatch_region): Likewise.
3843 (gimple_eh_dispatch_set_region): Likewise.
3844 (gimple_omp_critical_name): Likewise.
3845 (gimple_omp_critical_name_ptr): Likewise.
3846 (gimple_omp_critical_set_name): Likewise.
3847 (gimple_omp_for_clauses): Likewise.
3848 (gimple_omp_for_clauses_ptr): Likewise.
3849 (gimple_omp_for_set_clauses): Likewise.
3850 (gimple_omp_for_collapse): Likewise.
3851 (gimple_omp_for_index): Likewise.
3852 (gimple_omp_for_index_ptr): Likewise.
3853 (gimple_omp_for_set_index): Likewise.
3854 (gimple_omp_for_initial): Likewise.
3855 (gimple_omp_for_initial_ptr): Likewise.
3856 (gimple_omp_for_set_initial): Likewise.
3857 (gimple_omp_for_final): Likewise.
3858 (gimple_omp_for_final_ptr): Likewise.
3859 (gimple_omp_for_set_final): Likewise.
3860 (gimple_omp_for_incr): Likewise.
3861 (gimple_omp_for_incr_ptr): Likewise.
3862 (gimple_omp_for_set_incr): Likewise.
3863 (gimple_omp_for_pre_body_ptr): Likewise.
3864 (gimple_omp_for_set_pre_body): Likewise.
3865 (gimple_omp_parallel_clauses): Likewise.
3866 (gimple_omp_parallel_clauses_ptr): Likewise.
3867 (gimple_omp_parallel_set_clauses): Likewise.
3868 (gimple_omp_parallel_child_fn): Likewise.
3869 (gimple_omp_parallel_child_fn_ptr): Likewise.
3870 (gimple_omp_parallel_set_child_fn): Likewise.
3871 (gimple_omp_parallel_data_arg): Likewise.
3872 (gimple_omp_parallel_data_arg_ptr): Likewise.
3873 (gimple_omp_parallel_set_data_arg): Likewise.
3874 (gimple_omp_task_clauses): Likewise.
3875 (gimple_omp_task_clauses_ptr): Likewise.
3876 (gimple_omp_task_set_clauses): Likewise.
3877 (gimple_omp_task_child_fn): Likewise.
3878 (gimple_omp_task_child_fn_ptr): Likewise.
3879 (gimple_omp_task_set_child_fn): Likewise.
3880 (gimple_omp_task_data_arg): Likewise.
3881 (gimple_omp_task_data_arg_ptr): Likewise.
3882 (gimple_omp_task_set_data_arg): Likewise.
3883 (gimple_omp_taskreg_clauses): Likewise.
3884 (gimple_omp_taskreg_clauses_ptr): Likewise.
3885 (gimple_omp_taskreg_set_clauses): Likewise.
3886 (gimple_omp_taskreg_child_fn): Likewise.
3887 (gimple_omp_taskreg_child_fn_ptr): Likewise.
3888 (gimple_omp_taskreg_set_child_fn): Likewise.
3889 (gimple_omp_taskreg_data_arg): Likewise.
3890 (gimple_omp_taskreg_data_arg_ptr): Likewise.
3891 (gimple_omp_taskreg_set_data_arg): Likewise.
3892 (gimple_omp_task_copy_fn): Likewise.
3893 (gimple_omp_task_copy_fn_ptr): Likewise.
3894 (gimple_omp_task_set_copy_fn): Likewise.
3895 (gimple_omp_task_arg_size): Likewise.
3896 (gimple_omp_task_arg_size_ptr): Likewise.
3897 (gimple_omp_task_set_arg_size): Likewise.
3898 (gimple_omp_task_arg_align): Likewise.
3899 (gimple_omp_task_arg_align_ptr): Likewise.
3900 (gimple_omp_task_set_arg_align): Likewise.
3901 (gimple_omp_single_clauses): Likewise.
3902 (gimple_omp_single_clauses_ptr): Likewise.
3903 (gimple_omp_single_set_clauses): Likewise.
3904 (gimple_omp_target_clauses): Likewise.
3905 (gimple_omp_target_clauses_ptr): Likewise.
3906 (gimple_omp_target_set_clauses): Likewise.
3907 (gimple_omp_target_child_fn): Likewise.
3908 (gimple_omp_target_child_fn_ptr): Likewise.
3909 (gimple_omp_target_set_child_fn): Likewise.
3910 (gimple_omp_target_data_arg): Likewise.
3911 (gimple_omp_target_data_arg_ptr): Likewise.
3912 (gimple_omp_target_set_data_arg): Likewise.
3913 (gimple_omp_teams_clauses): Likewise.
3914 (gimple_omp_teams_clauses_ptr): Likewise.
3915 (gimple_omp_teams_set_clauses): Likewise.
3916 (gimple_omp_sections_clauses): Likewise.
3917 (gimple_omp_sections_clauses_ptr): Likewise.
3918 (gimple_omp_sections_set_clauses): Likewise.
3919 (gimple_omp_sections_control): Likewise.
3920 (gimple_omp_sections_control_ptr): Likewise.
3921 (gimple_omp_sections_set_control): Likewise.
3922 (gimple_omp_for_set_cond): Likewise.
3923 (gimple_omp_for_cond): Likewise.
3924 (gimple_omp_atomic_store_set_val): Likewise.
3925 (gimple_omp_atomic_store_val): Likewise.
3926 (gimple_omp_atomic_store_val_ptr): Likewise.
3927 (gimple_omp_atomic_load_set_lhs): Likewise.
3928 (gimple_omp_atomic_load_lhs): Likewise.
3929 (gimple_omp_atomic_load_lhs_ptr): Likewise.
3930 (gimple_omp_atomic_load_set_rhs): Likewise.
3931 (gimple_omp_atomic_load_rhs): Likewise.
3932 (gimple_omp_atomic_load_rhs_ptr): Likewise.
3933 (gimple_omp_continue_control_def): Likewise.
3934 (gimple_omp_continue_control_def_ptr): Likewise.
3935 (gimple_omp_continue_set_control_def): Likewise.
3936 (gimple_omp_continue_control_use): Likewise.
3937 (gimple_omp_continue_control_use_ptr): Likewise.
3938 (gimple_omp_continue_set_control_use): Likewise.
3939 (gimple_transaction_body_ptr): Likewise.
3940 (gimple_transaction_label): Likewise.
3941 (gimple_transaction_label_ptr): Likewise.
3942 (gimple_transaction_set_body): Likewise.
3943 (gimple_transaction_set_label): Likewise.
3945 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
3946 * ipa-inline-analysis.c (inline_write_summary): Likewise.
3947 * ipa-ref.c (ipa_record_reference): Likewise.
3948 * ipa-reference.c (analyze_function): Likewise.
3949 (ipa_reference_write_optimization_summary): Likewise.
3950 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
3951 (address_taken_from_non_vtable_p): Likewise.
3952 (comdat_can_be_unshared_p_1): Likewise.
3953 * lto-cgraph.c (lto_output_ref): Likewise.
3954 (add_references): Likewise.
3955 (compute_ltrans_boundary): Likewise.
3956 (output_symtab): Likewise.
3957 (input_ref): Likewise.
3958 (input_cgraph_1): Likewise.
3959 (output_cgraph_opt_summary): Likewise.
3960 * lto-streamer-out.c (lto_output): Likewise.
3961 (output_symbol_p): Likewise.
3962 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
3963 (lsei_start_function_in_partition): Likewise.
3964 (lsei_next_variable_in_partition): Likewise.
3965 (lsei_start_variable_in_partition): Likewise.
3966 * symtab.c (insert_to_assembler_name_hash): Likewise.
3967 (unlink_from_assembler_name_hash): Likewise.
3968 (symtab_unregister_node): Likewise.
3969 (symtab_remove_node): Likewise.
3970 (dump_symtab_node): Likewise.
3971 (verify_symtab_base): Likewise.
3972 (verify_symtab_node): Likewise.
3973 (symtab_make_decl_local): Likewise.
3974 (symtab_alias_ultimate_target): Likewise.
3975 (symtab_resolve_alias): Likewise.
3976 (symtab_get_symbol_partitioning_class): Likewise.
3977 * tree-phinodes.c (allocate_phi_node): Likewise.
3978 (reserve_phi_args_for_new_edge): Likewise.
3979 (remove_phi_args): Likewise.
3980 * varpool.c (varpool_node_for_asm): Likewise.
3981 (varpool_remove_unreferenced_decls): Likewise.
3983 2014-04-23 Jeff Law <law@redhat.com>
3985 PR tree-optimization/60902
3986 * tree-ssa-threadedge.c
3987 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
3988 invalidate outputs from statements that do not produce useful
3989 outputs for threading.
3991 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3993 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
3994 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
3995 machine descriptions for Stack Smashing Protector.
3997 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
3999 * aarch64.md (<optab>_rol<mode>3): New pattern.
4000 (<optab>_rolsi3_uxtw): Likewise.
4001 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
4003 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
4005 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
4006 (arm_cortex_a12_tune): Likewise.
4008 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4010 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
4012 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4014 * config/arm/arm.md (arm_rev16si2): New pattern.
4015 (arm_rev16si2_alt): Likewise.
4016 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
4018 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4020 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
4021 (rev16<mode>2_alt): Likewise.
4022 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
4023 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
4024 (aarch_rev16_shleft_mask_imm_p): Likewise.
4025 (aarch_rev16_p_1): Likewise.
4026 (aarch_rev16_p): Likewise.
4027 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
4028 (aarch_rev16_shright_mask_imm_p): Likewise.
4029 (aarch_rev16_shleft_mask_imm_p): Likewise.
4031 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4033 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
4034 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
4036 (cortex_a53_extra_costs): Likewise.
4037 (cortex_a57_extra_costs): Likewise.
4038 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
4039 (cortexa7_extra_costs): Likewise.
4040 (cortexa8_extra_costs): Likewise.
4041 (cortexa12_extra_costs): Likewise.
4042 (cortexa15_extra_costs): Likewise.
4043 (v7m_extra_costs): Likewise.
4044 (arm_new_rtx_costs): Handle BSWAP.
4046 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4048 * config/arm/arm.c (cortexa8_extra_costs): New table.
4049 (arm_cortex_a8_tune): New tuning struct.
4050 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
4052 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4054 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
4056 2014-04-23 Richard Biener <rguenther@suse.de>
4058 * Makefile.in (OBJS): Remove loop-unswitch.o.
4059 * tree-pass.h (make_pass_rtl_unswitch): Remove.
4060 * passes.def (pass_rtl_unswitch): Likewise.
4061 * loop-init.c (gate_rtl_unswitch): Likewise.
4062 (rtl_unswitch): Likewise.
4063 (pass_data_rtl_unswitch): Likewise.
4064 (pass_rtl_unswitch): Likewise.
4065 (make_pass_rtl_unswitch): Likewise.
4066 * rtl.h (reversed_condition): Likewise.
4067 (compare_and_jump_seq): Likewise.
4068 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
4070 * loop-unroll.c (compare_and_jump_seq): Likewise.
4072 2014-04-23 Richard Biener <rguenther@suse.de>
4074 PR tree-optimization/60903
4075 * tree-ssa-loop-im.c (analyze_memory_references): Remove
4076 commented code block.
4077 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
4078 loop flags to newly created BBs and edges.
4080 2014-04-23 Nick Clifton <nickc@redhat.com>
4082 * config/msp430/msp430.c (msp430_handle_option): Move function
4084 (msp430_option_override): Simplify mcu and mcpu option handling.
4085 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
4086 support for -mhwmult command line option.
4087 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
4088 -mhwmult command line option.
4089 (msp430_hwmult_enabled): Delete.
4090 (msp43o_output_labelref): Add support for -mhwmult command line option.
4091 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
4092 (umulsidi3): Likewise.
4093 * config/msp430/msp430.opt (mmcu): Add Report attribute.
4094 (mcpu, mlarge, msmall): Likewise.
4095 (mhwmult): New option.
4096 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
4098 (msp430_is_f5_mcu): Remove prototype.
4099 (msp430_use_f5_series_hwmult): Add prototype.
4100 * config/msp430/msp430-opts.h: New file.
4101 * common/config/msp430: New directory.
4102 * common/config/msp430/msp430-common.c: New file.
4103 * config.gcc (msp430): Remove target_has_targetm_common.
4104 * doc/invoke.texi: Document -mhwmult command line option.
4106 2014-04-23 Nick Clifton <nickc@redhat.com>
4108 * config/i386/cygwin.h (ENDFILE_SPEC): Include
4109 default-manifest.o if it can be found in the search path.
4110 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
4112 2014-04-23 Terry Guo <terry.guo@arm.com>
4114 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
4116 2014-04-23 Richard Biener <rguenther@suse.de>
4119 * tree-inline.c (declare_return_variable): Use mark_addressable.
4121 2014-04-23 Richard Biener <rguenther@suse.de>
4124 * loop-init.c (loop_optimizer_init): Make sure to apply
4125 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
4127 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4130 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
4132 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
4133 if flag_sanitize_undefined_trap_on_error.
4134 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
4135 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
4136 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
4137 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
4138 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
4139 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
4140 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
4141 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
4142 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
4143 * ubsan.c (ubsan_instrument_unreachable): Return
4144 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
4145 (ubsan_expand_null_ifn): Emit __builtin_trap ()
4146 if flag_sanitize_undefined_trap_on_error and
4147 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
4148 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
4149 instrument_bool_enum_load): Emit __builtin_trap () if
4150 flag_sanitize_undefined_trap_on_error and
4151 __builtin_handle_*_abort () if !flag_sanitize_recover.
4152 * doc/invoke.texi (-fsanitize-recover,
4153 -fsanitize-undefined-trap-on-error): Document.
4155 2014-04-22 Christian Bruel <christian.bruel@st.com>
4157 * config/sh/sh.md (mov<mode>): Replace movQIHI.
4158 Force immediates to SImode.
4160 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
4162 * config/nios2/nios2.md (UNSPEC_ROUND): New.
4164 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
4165 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
4166 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
4167 (nios2_fpu_insn): Add entry for round.
4168 (N2FPU_NO_ERRNO_P): Define.
4169 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
4171 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
4173 2014-04-22 Richard Henderson <rth@redhat.com>
4175 * config/aarch64/aarch64 (addti3, subti3): New expanders.
4176 (add<GPI>3_compare0): Remove leading * from name.
4177 (add<GPI>3_carryin): Likewise.
4178 (sub<GPI>3_compare0): Likewise.
4179 (sub<GPI>3_carryin): Likewise.
4180 (<su_optab>mulditi3): New expander.
4181 (multi3): New expander.
4182 (madd<GPI>): Remove leading * from name.
4184 2014-04-22 Martin Jambor <mjambor@suse.cz>
4186 * cgraphclones.c (cgraph_function_versioning): Copy
4187 ipa_transforms_to_apply instead of asserting it is empty.
4189 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
4192 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
4193 on count_exp to get mode.
4195 2014-04-22 Andrew Pinski <apinski@cavium.com>
4197 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
4198 Handle TLS for ILP32.
4199 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
4200 (tlsie_small_<mode>): this and handle PTR.
4201 (tlsie_small_sidi): New pattern.
4202 (tlsle_small): Change to an expand to handle ILP32.
4203 (tlsle_small_<mode>): New pattern.
4204 (tlsdesc_small): Rename to ...
4205 (tlsdesc_small_<mode>): this and handle PTR.
4207 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4209 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
4211 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4213 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
4214 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
4215 (aarch64_types_signed_poly_qualifiers): Likewise.
4216 (aarch64_types_unsigned_signed_qualifiers): Likewise.
4217 (aarch64_types_poly_signed_qualifiers): Likewise.
4218 (TYPES_REINTERP_SS): Type macro added.
4219 (TYPES_REINTERP_SU): Likewise.
4220 (TYPES_REINTERP_SP): Likewise.
4221 (TYPES_REINTERP_US): Likewise.
4222 (TYPES_REINTERP_PS): Likewise.
4223 (aarch64_fold_builtin): New expression folding added.
4224 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
4225 Declarations removed.
4226 (REINTERP_SS): Declarations added.
4227 (REINTERP_US): Likewise.
4228 (REINTERP_PS): Likewise.
4229 (REINTERP_SU): Likewise.
4230 (REINTERP_SP): Likewise.
4231 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
4232 (vreinterpretq_p8_f64): Likewise.
4233 (vreinterpret_p16_f64): Likewise.
4234 (vreinterpretq_p16_f64): Likewise.
4235 (vreinterpret_f32_f64): Likewise.
4236 (vreinterpretq_f32_f64): Likewise.
4237 (vreinterpret_f64_f32): Likewise.
4238 (vreinterpret_f64_p8): Likewise.
4239 (vreinterpret_f64_p16): Likewise.
4240 (vreinterpret_f64_s8): Likewise.
4241 (vreinterpret_f64_s16): Likewise.
4242 (vreinterpret_f64_s32): Likewise.
4243 (vreinterpret_f64_s64): Likewise.
4244 (vreinterpret_f64_u8): Likewise.
4245 (vreinterpret_f64_u16): Likewise.
4246 (vreinterpret_f64_u32): Likewise.
4247 (vreinterpret_f64_u64): Likewise.
4248 (vreinterpretq_f64_f32): Likewise.
4249 (vreinterpretq_f64_p8): Likewise.
4250 (vreinterpretq_f64_p16): Likewise.
4251 (vreinterpretq_f64_s8): Likewise.
4252 (vreinterpretq_f64_s16): Likewise.
4253 (vreinterpretq_f64_s32): Likewise.
4254 (vreinterpretq_f64_s64): Likewise.
4255 (vreinterpretq_f64_u8): Likewise.
4256 (vreinterpretq_f64_u16): Likewise.
4257 (vreinterpretq_f64_u32): Likewise.
4258 (vreinterpretq_f64_u64): Likewise.
4259 (vreinterpret_s64_f64): Likewise.
4260 (vreinterpretq_s64_f64): Likewise.
4261 (vreinterpret_u64_f64): Likewise.
4262 (vreinterpretq_u64_f64): Likewise.
4263 (vreinterpret_s8_f64): Likewise.
4264 (vreinterpretq_s8_f64): Likewise.
4265 (vreinterpret_s16_f64): Likewise.
4266 (vreinterpretq_s16_f64): Likewise.
4267 (vreinterpret_s32_f64): Likewise.
4268 (vreinterpretq_s32_f64): Likewise.
4269 (vreinterpret_u8_f64): Likewise.
4270 (vreinterpretq_u8_f64): Likewise.
4271 (vreinterpret_u16_f64): Likewise.
4272 (vreinterpretq_u16_f64): Likewise.
4273 (vreinterpret_u32_f64): Likewise.
4274 (vreinterpretq_u32_f64): Likewise.
4276 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4278 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
4279 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
4280 (vreinterpret_p8_s8): Likewise.
4281 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
4282 (vreinterpret_p8_s16): Likewise.
4283 (vreinterpret_p8_s32): Likewise.
4284 (vreinterpret_p8_s64): Likewise.
4285 (vreinterpret_p8_f32): Likewise.
4286 (vreinterpret_p8_u8): Likewise.
4287 (vreinterpret_p8_u16): Likewise.
4288 (vreinterpret_p8_u32): Likewise.
4289 (vreinterpret_p8_u64): Likewise.
4290 (vreinterpret_p8_p16): Likewise.
4291 (vreinterpretq_p8_s8): Likewise.
4292 (vreinterpretq_p8_s16): Likewise.
4293 (vreinterpretq_p8_s32): Likewise.
4294 (vreinterpretq_p8_s64): Likewise.
4295 (vreinterpretq_p8_f32): Likewise.
4296 (vreinterpretq_p8_u8): Likewise.
4297 (vreinterpretq_p8_u16): Likewise.
4298 (vreinterpretq_p8_u32): Likewise.
4299 (vreinterpretq_p8_u64): Likewise.
4300 (vreinterpretq_p8_p16): Likewise.
4301 (vreinterpret_p16_s8): Likewise.
4302 (vreinterpret_p16_s16): Likewise.
4303 (vreinterpret_p16_s32): Likewise.
4304 (vreinterpret_p16_s64): Likewise.
4305 (vreinterpret_p16_f32): Likewise.
4306 (vreinterpret_p16_u8): Likewise.
4307 (vreinterpret_p16_u16): Likewise.
4308 (vreinterpret_p16_u32): Likewise.
4309 (vreinterpret_p16_u64): Likewise.
4310 (vreinterpret_p16_p8): Likewise.
4311 (vreinterpretq_p16_s8): Likewise.
4312 (vreinterpretq_p16_s16): Likewise.
4313 (vreinterpretq_p16_s32): Likewise.
4314 (vreinterpretq_p16_s64): Likewise.
4315 (vreinterpretq_p16_f32): Likewise.
4316 (vreinterpretq_p16_u8): Likewise.
4317 (vreinterpretq_p16_u16): Likewise.
4318 (vreinterpretq_p16_u32): Likewise.
4319 (vreinterpretq_p16_u64): Likewise.
4320 (vreinterpretq_p16_p8): Likewise.
4321 (vreinterpret_f32_s8): Likewise.
4322 (vreinterpret_f32_s16): Likewise.
4323 (vreinterpret_f32_s32): Likewise.
4324 (vreinterpret_f32_s64): Likewise.
4325 (vreinterpret_f32_u8): Likewise.
4326 (vreinterpret_f32_u16): Likewise.
4327 (vreinterpret_f32_u32): Likewise.
4328 (vreinterpret_f32_u64): Likewise.
4329 (vreinterpret_f32_p8): Likewise.
4330 (vreinterpret_f32_p16): Likewise.
4331 (vreinterpretq_f32_s8): Likewise.
4332 (vreinterpretq_f32_s16): Likewise.
4333 (vreinterpretq_f32_s32): Likewise.
4334 (vreinterpretq_f32_s64): Likewise.
4335 (vreinterpretq_f32_u8): Likewise.
4336 (vreinterpretq_f32_u16): Likewise.
4337 (vreinterpretq_f32_u32): Likewise.
4338 (vreinterpretq_f32_u64): Likewise.
4339 (vreinterpretq_f32_p8): Likewise.
4340 (vreinterpretq_f32_p16): Likewise.
4341 (vreinterpret_s64_s8): Likewise.
4342 (vreinterpret_s64_s16): Likewise.
4343 (vreinterpret_s64_s32): Likewise.
4344 (vreinterpret_s64_f32): Likewise.
4345 (vreinterpret_s64_u8): Likewise.
4346 (vreinterpret_s64_u16): Likewise.
4347 (vreinterpret_s64_u32): Likewise.
4348 (vreinterpret_s64_u64): Likewise.
4349 (vreinterpret_s64_p8): Likewise.
4350 (vreinterpret_s64_p16): Likewise.
4351 (vreinterpretq_s64_s8): Likewise.
4352 (vreinterpretq_s64_s16): Likewise.
4353 (vreinterpretq_s64_s32): Likewise.
4354 (vreinterpretq_s64_f32): Likewise.
4355 (vreinterpretq_s64_u8): Likewise.
4356 (vreinterpretq_s64_u16): Likewise.
4357 (vreinterpretq_s64_u32): Likewise.
4358 (vreinterpretq_s64_u64): Likewise.
4359 (vreinterpretq_s64_p8): Likewise.
4360 (vreinterpretq_s64_p16): Likewise.
4361 (vreinterpret_u64_s8): Likewise.
4362 (vreinterpret_u64_s16): Likewise.
4363 (vreinterpret_u64_s32): Likewise.
4364 (vreinterpret_u64_s64): Likewise.
4365 (vreinterpret_u64_f32): Likewise.
4366 (vreinterpret_u64_u8): Likewise.
4367 (vreinterpret_u64_u16): Likewise.
4368 (vreinterpret_u64_u32): Likewise.
4369 (vreinterpret_u64_p8): Likewise.
4370 (vreinterpret_u64_p16): Likewise.
4371 (vreinterpretq_u64_s8): Likewise.
4372 (vreinterpretq_u64_s16): Likewise.
4373 (vreinterpretq_u64_s32): Likewise.
4374 (vreinterpretq_u64_s64): Likewise.
4375 (vreinterpretq_u64_f32): Likewise.
4376 (vreinterpretq_u64_u8): Likewise.
4377 (vreinterpretq_u64_u16): Likewise.
4378 (vreinterpretq_u64_u32): Likewise.
4379 (vreinterpretq_u64_p8): Likewise.
4380 (vreinterpretq_u64_p16): Likewise.
4381 (vreinterpret_s8_s16): Likewise.
4382 (vreinterpret_s8_s32): Likewise.
4383 (vreinterpret_s8_s64): Likewise.
4384 (vreinterpret_s8_f32): Likewise.
4385 (vreinterpret_s8_u8): Likewise.
4386 (vreinterpret_s8_u16): Likewise.
4387 (vreinterpret_s8_u32): Likewise.
4388 (vreinterpret_s8_u64): Likewise.
4389 (vreinterpret_s8_p8): Likewise.
4390 (vreinterpret_s8_p16): Likewise.
4391 (vreinterpretq_s8_s16): Likewise.
4392 (vreinterpretq_s8_s32): Likewise.
4393 (vreinterpretq_s8_s64): Likewise.
4394 (vreinterpretq_s8_f32): Likewise.
4395 (vreinterpretq_s8_u8): Likewise.
4396 (vreinterpretq_s8_u16): Likewise.
4397 (vreinterpretq_s8_u32): Likewise.
4398 (vreinterpretq_s8_u64): Likewise.
4399 (vreinterpretq_s8_p8): Likewise.
4400 (vreinterpretq_s8_p16): Likewise.
4401 (vreinterpret_s16_s8): Likewise.
4402 (vreinterpret_s16_s32): Likewise.
4403 (vreinterpret_s16_s64): Likewise.
4404 (vreinterpret_s16_f32): Likewise.
4405 (vreinterpret_s16_u8): Likewise.
4406 (vreinterpret_s16_u16): Likewise.
4407 (vreinterpret_s16_u32): Likewise.
4408 (vreinterpret_s16_u64): Likewise.
4409 (vreinterpret_s16_p8): Likewise.
4410 (vreinterpret_s16_p16): Likewise.
4411 (vreinterpretq_s16_s8): Likewise.
4412 (vreinterpretq_s16_s32): Likewise.
4413 (vreinterpretq_s16_s64): Likewise.
4414 (vreinterpretq_s16_f32): Likewise.
4415 (vreinterpretq_s16_u8): Likewise.
4416 (vreinterpretq_s16_u16): Likewise.
4417 (vreinterpretq_s16_u32): Likewise.
4418 (vreinterpretq_s16_u64): Likewise.
4419 (vreinterpretq_s16_p8): Likewise.
4420 (vreinterpretq_s16_p16): Likewise.
4421 (vreinterpret_s32_s8): Likewise.
4422 (vreinterpret_s32_s16): Likewise.
4423 (vreinterpret_s32_s64): Likewise.
4424 (vreinterpret_s32_f32): Likewise.
4425 (vreinterpret_s32_u8): Likewise.
4426 (vreinterpret_s32_u16): Likewise.
4427 (vreinterpret_s32_u32): Likewise.
4428 (vreinterpret_s32_u64): Likewise.
4429 (vreinterpret_s32_p8): Likewise.
4430 (vreinterpret_s32_p16): Likewise.
4431 (vreinterpretq_s32_s8): Likewise.
4432 (vreinterpretq_s32_s16): Likewise.
4433 (vreinterpretq_s32_s64): Likewise.
4434 (vreinterpretq_s32_f32): Likewise.
4435 (vreinterpretq_s32_u8): Likewise.
4436 (vreinterpretq_s32_u16): Likewise.
4437 (vreinterpretq_s32_u32): Likewise.
4438 (vreinterpretq_s32_u64): Likewise.
4439 (vreinterpretq_s32_p8): Likewise.
4440 (vreinterpretq_s32_p16): Likewise.
4441 (vreinterpret_u8_s8): Likewise.
4442 (vreinterpret_u8_s16): Likewise.
4443 (vreinterpret_u8_s32): Likewise.
4444 (vreinterpret_u8_s64): Likewise.
4445 (vreinterpret_u8_f32): Likewise.
4446 (vreinterpret_u8_u16): Likewise.
4447 (vreinterpret_u8_u32): Likewise.
4448 (vreinterpret_u8_u64): Likewise.
4449 (vreinterpret_u8_p8): Likewise.
4450 (vreinterpret_u8_p16): Likewise.
4451 (vreinterpretq_u8_s8): Likewise.
4452 (vreinterpretq_u8_s16): Likewise.
4453 (vreinterpretq_u8_s32): Likewise.
4454 (vreinterpretq_u8_s64): Likewise.
4455 (vreinterpretq_u8_f32): Likewise.
4456 (vreinterpretq_u8_u16): Likewise.
4457 (vreinterpretq_u8_u32): Likewise.
4458 (vreinterpretq_u8_u64): Likewise.
4459 (vreinterpretq_u8_p8): Likewise.
4460 (vreinterpretq_u8_p16): Likewise.
4461 (vreinterpret_u16_s8): Likewise.
4462 (vreinterpret_u16_s16): Likewise.
4463 (vreinterpret_u16_s32): Likewise.
4464 (vreinterpret_u16_s64): Likewise.
4465 (vreinterpret_u16_f32): Likewise.
4466 (vreinterpret_u16_u8): Likewise.
4467 (vreinterpret_u16_u32): Likewise.
4468 (vreinterpret_u16_u64): Likewise.
4469 (vreinterpret_u16_p8): Likewise.
4470 (vreinterpret_u16_p16): Likewise.
4471 (vreinterpretq_u16_s8): Likewise.
4472 (vreinterpretq_u16_s16): Likewise.
4473 (vreinterpretq_u16_s32): Likewise.
4474 (vreinterpretq_u16_s64): Likewise.
4475 (vreinterpretq_u16_f32): Likewise.
4476 (vreinterpretq_u16_u8): Likewise.
4477 (vreinterpretq_u16_u32): Likewise.
4478 (vreinterpretq_u16_u64): Likewise.
4479 (vreinterpretq_u16_p8): Likewise.
4480 (vreinterpretq_u16_p16): Likewise.
4481 (vreinterpret_u32_s8): Likewise.
4482 (vreinterpret_u32_s16): Likewise.
4483 (vreinterpret_u32_s32): Likewise.
4484 (vreinterpret_u32_s64): Likewise.
4485 (vreinterpret_u32_f32): Likewise.
4486 (vreinterpret_u32_u8): Likewise.
4487 (vreinterpret_u32_u16): Likewise.
4488 (vreinterpret_u32_u64): Likewise.
4489 (vreinterpret_u32_p8): Likewise.
4490 (vreinterpret_u32_p16): Likewise.
4491 (vreinterpretq_u32_s8): Likewise.
4492 (vreinterpretq_u32_s16): Likewise.
4493 (vreinterpretq_u32_s32): Likewise.
4494 (vreinterpretq_u32_s64): Likewise.
4495 (vreinterpretq_u32_f32): Likewise.
4496 (vreinterpretq_u32_u8): Likewise.
4497 (vreinterpretq_u32_u16): Likewise.
4498 (vreinterpretq_u32_u64): Likewise.
4499 (vreinterpretq_u32_p8): Likewise.
4500 (vreinterpretq_u32_p16): Likewise.
4502 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4504 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
4506 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
4508 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
4509 (vqnegd_s64): Likewise.
4510 (vqabs_s64): Likewise.
4511 (vqabsd_s64): Likewise.
4513 2014-04-22 Richard Henderson <rth@redhat.com>
4515 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
4516 computation to the top of the loop.
4518 2014-04-22 Renlin <renlin.li@arm.com>
4519 Jiong Wang <jiong.wang@arm.com>
4521 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
4522 * config/aarch64/aarch64.c (aarch64_layout_frame)
4523 (aarch64_initial_elimination_offset): Likewise.
4525 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
4527 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
4530 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
4532 * machmode.h (bitwise_mode_for_mode): Declare.
4533 * stor-layout.h (bitwise_type_for_mode): Likewise.
4534 * stor-layout.c (bitwise_mode_for_mode): New function.
4535 (bitwise_type_for_mode): Likewise.
4536 * builtins.c (fold_builtin_memory_op): Use it instead of
4537 int_mode_for_mode and build_nonstandard_integer_type.
4539 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4541 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
4542 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
4543 (*-*-solaris2*): Simplify.
4544 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
4545 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
4546 *-*-solaris2.9* handling.
4548 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
4550 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
4551 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
4553 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
4554 * configure: Regenerate.
4556 * config/i386/sol2-9.h: Remove.
4558 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
4559 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
4560 Remove Solaris 9 references.
4562 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
4564 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
4565 (floatuns<GPI:mode><GPF:mode>2): Remove.
4566 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
4567 and floatuns conversions.
4568 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
4569 and floatuns conversions.
4570 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
4571 (w1,w2): New mode attributes for inequal width conversions.
4573 2014-04-22 Renlin Li <Renlin.Li@arm.com>
4575 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
4576 the output asm format.
4578 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
4580 * config/aarch64/aarch64-simd.md
4581 (aarch64_cm<optab>di): Always split.
4582 (*aarch64_cm<optab>di): New.
4583 (aarch64_cmtstdi): Always split.
4584 (*aarch64_cmtstdi): New.
4586 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4588 PR tree-optimization/60823
4589 * omp-low.c (ipa_simd_modify_function_body): Go through
4590 all SSA_NAMEs and for those refering to vector arguments
4591 which are going to be replaced adjust SSA_NAME_VAR and,
4592 if it is a default definition, change it into a non-default
4593 definition assigned at the beginning of function from new_decl.
4594 (ipa_simd_modify_stmt_ops): Rewritten.
4595 * tree-dfa.c (set_ssa_default_def): When removing default def,
4596 check for NULL loc instead of NULL *loc.
4598 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4600 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
4601 restrictions on core registers for DImode values in Thumb2.
4603 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4605 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
4606 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
4608 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4610 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
4611 (*iordi_notzesidi_di): Likewise.
4612 (*iordi_notsesidi_di): Likewise.
4614 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4616 * config/arm/arm-protos.h (tune_params): New struct members.
4617 * config/arm/arm.c: Initialise tune_params per processor.
4618 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
4619 for speed, based on new tune_params.
4621 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4623 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
4624 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
4625 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
4626 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
4627 * config/aarch64/arm_neon.h (vrnd_f64): Added.
4628 (vrnda_f64): Likewise.
4629 (vrndi_f64): Likewise.
4630 (vrndm_f64): Likewise.
4631 (vrndn_f64): Likewise.
4632 (vrndp_f64): Likewise.
4633 (vrndx_f64): Likewise.
4635 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4637 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
4638 GET_MODE_SIZE argument is enum machine_mode.
4640 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4643 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
4644 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
4646 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
4649 * asan.c (asan_emit_stack_protection): Force the base to align to
4650 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
4651 appropriate bits if STRICT_ALIGNMENT.
4652 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
4654 (expand_used_vars): Leave a space in the stack frame for alignment
4655 if STRICT_ALIGNMENT.
4657 2014-04-21 David Malcolm <dmalcolm@redhat.com>
4659 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
4661 (gimple_store_p): Likewise.
4662 (gimple_assign_load_p): Likewise.
4663 (gimple_assign_cast_p): Likewise.
4664 (gimple_clobber_p): Likewise.
4666 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
4667 rather than a gimple.
4668 (gimple_assign_cast_p): Likewise.
4670 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
4673 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
4674 If mode is DDmode and TARGET_E500_DOUBLE allow move.
4676 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
4677 more debug information for E500 if -mdebug=reg.
4679 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
4682 * config/i386/i386.c (ix86_expand_builtin)
4683 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
4684 register for target RTX.
4685 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
4687 2014-04-18 Cong Hou <congh@google.com>
4689 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
4690 the widen-mult pattern by handling two operands with different sizes,
4691 and operands whose size is smaller than half of the result type.
4693 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
4695 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
4696 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
4697 (do_estimate_edge_time): Compute it.
4698 * ipa-inline.c (want_inline_small_function_p): Bypass
4699 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
4701 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
4703 * ipa-inline.c (spec_rem): New static variable.
4704 (dump_overall_stats): New function.
4705 (dump_inline_stats): New function.
4707 2014-04-18 Richard Henderson <rth@redhat.com>
4709 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
4710 to GET_MODE_SIZE, not a reg_class_t.
4712 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4714 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
4715 (vsx_xxmrglw_<mode>): Likewise.
4717 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
4720 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
4721 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
4722 (rs6000_init_hard_regno_mode_ok): Likewise.
4724 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
4726 * ipa-inline.c (inline_small_functions): Account only non-cold
4728 * doc/invoke.texi (inline-unit-growth): Update documentation.
4730 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
4732 * config/rs6000/rs6000.md (addti3, subti3): New.
4734 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
4737 * config/i386/i386.c (ix86_expand_clear): Remove outdated
4738 comment. Check optimize_insn_for_size_p instead of
4739 optimize_insn_for_speed_p.
4741 2014-04-17 Martin Jambor <mjambor@suse.cz>
4743 * gimple-iterator.c (gsi_start_edge): New function.
4744 * gimple-iterator.h (gsi_start_edge): Declare.
4745 * tree-sra.c (single_non_eh_succ): New function.
4746 (disqualify_ops_if_throwing_stmt): Renamed to
4747 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
4748 having one non-EH successor BB.
4749 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
4750 generate loads into replacements.
4751 (sra_modify_assign): Likewise and and also use the simple path for
4753 (sra_modify_function_body): Commit statements on edges.
4755 2014-04-17 Richard Biener <rguenther@suse.de>
4758 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
4759 comparison results and add clarifying comment.
4761 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4763 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
4764 (blank_mode): Initialize it.
4765 (emit_mode_size_inline, emit_mode_nunits_inline,
4766 emit_mode_inner_inline): New functions.
4767 (emit_insn_modes_h): Call them and surround their output with
4768 #if GCC_VERSION >= 4001 ... #endif.
4769 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
4770 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
4771 mode_* arrays if the argument is __builtin_constant_p.
4772 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
4773 is enum machine_mode.
4775 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4777 * passes.c (opt_pass::execute): Adjust.
4778 (pass_manager::execute_pass_mode_switching): Likewise.
4779 (early_local_passes::execute): Likewise.
4780 (execute_one_pass): Pass cfun to the pass's execute method.
4781 * tree-pass.h (opt_pass::execute): Add function * argument.
4782 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
4783 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
4784 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
4785 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4786 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
4787 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
4788 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
4789 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
4790 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
4791 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
4792 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
4793 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
4794 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
4795 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
4796 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
4797 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
4798 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
4799 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
4800 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
4801 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4802 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4803 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4804 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4805 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4806 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4807 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4808 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
4809 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
4810 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
4813 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4815 * passes.c (opt_pass::gate): Take function * argument.
4816 (gate_all_early_local_passes): Merge into
4817 (early_local_passes::gate): this.
4818 (gate_all_early_optimizations): Merge into
4819 (all_early_optimizations::gate): this.
4820 (gate_all_optimizations): Mege into
4821 (all_optimizations::gate): this.
4822 (gate_all_optimizations_g): Merge into
4823 (all_optimizations_g::gate): this.
4824 (gate_rest_of_compilation): Mege into
4825 (rest_of_compilation::gate): this.
4826 (gate_postreload): Merge into
4827 (postreload::gate): this.
4828 (dump_one_pass): Pass cfun to the pass's gate method.
4829 (execute_ipa_summary_passes): Likewise.
4830 (execute_one_pass): Likewise.
4831 (ipa_write_summaries_2): Likewise.
4832 (ipa_write_optimization_summaries_1): Likewise.
4833 (ipa_read_summaries_1): Likewise.
4834 (ipa_read_optimization_summaries_1): Likewise.
4835 (execute_ipa_stmt_fixups): Likewise.
4836 * tree-pass.h (opt_pass::gate): Add function * argument.
4837 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
4838 combine-stack-adj.c, combine.c, compare-elim.c,
4839 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4840 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
4841 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
4842 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
4843 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
4844 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
4845 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4846 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
4847 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
4848 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
4849 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
4850 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
4851 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
4852 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4853 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4854 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4855 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4856 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4857 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4858 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4859 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4860 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4861 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
4862 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
4863 var-tracking.c, vtable-verify.c, web.c: Adjust.
4865 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4867 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
4868 * configure: Regenerate.
4870 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4872 * passes.c (dump_one_pass): don't check pass->has_gate.
4873 (execute_ipa_summary_passes): Likewise.
4874 (execute_one_pass): Likewise.
4875 (ipa_write_summaries_2): Likewise.
4876 (ipa_write_optimization_summaries_1): Likewise.
4877 (ipa_read_optimization_summaries_1): Likewise.
4878 (execute_ipa_stmt_fixups): Likewise.
4879 * tree-pass.h (pass_data::has_gate): Remove.
4880 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
4881 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
4882 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
4883 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4884 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
4885 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
4886 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
4887 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
4888 gimple-low.c, gimple-ssa-isolate-paths.c,
4889 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
4890 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
4891 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4892 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
4893 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
4894 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
4895 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
4896 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
4897 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
4898 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
4899 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4900 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4901 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4902 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4903 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4904 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4905 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4906 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4907 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4908 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
4909 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
4910 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
4913 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4915 * pass_manager.h (pass_manager::register_dump_files_1): Remove
4917 * passes.c (pass_manager::register_dump_files_1): Merge into
4918 (pass_manager::register_dump_files): this, and remove its handling of
4919 properties since the pass always has the properties anyway.
4920 (pass_manager::pass_manager): Adjust.
4922 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4924 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
4925 * passes.c (pass_manager::register_dump_files_1): Remove dead code
4926 dealing with properties.
4927 (pass_manager::register_dump_files): Adjust.
4929 2014-03-20 Mark Wielaard <mjw@redhat.com>
4931 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
4932 then represent the bound as normal constant value.
4934 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4937 Forward port from 4.8 branch
4938 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
4940 * config/i386/bmiintrin.h (_blsi_u32): New.
4944 (_blsmsk_u32): Ditto.
4945 (_blsmsk_u64): Ditto.
4946 (_tzcnt_u32): Ditto.
4947 (_tzcnt_u64): Ditto.
4949 2014-04-17 Kito Cheng <kito@0xlab.org>
4951 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
4953 2014-04-17 Richard Biener <rguenther@suse.de>
4956 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
4957 boolean results for comparisons.
4959 2014-04-17 Richard Biener <rguenther@suse.de>
4961 PR tree-optimization/60836
4962 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
4963 initial PHI args to be gimple values.
4965 2014-04-17 Richard Biener <rguenther@suse.de>
4967 PR tree-optimization/60841
4968 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
4969 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
4970 of stmts to SLP build.
4971 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
4972 (vect_analyze_slp): Likewise.
4973 (vect_analyze_slp_instance): Likewise.
4974 (vect_build_slp_tree): Limit overall SLP tree growth.
4975 * tree-vectorizer.h (vect_analyze_data_refs,
4976 vect_analyze_slp): Adjust prototypes.
4978 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4980 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
4983 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4985 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
4986 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
4987 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
4988 for TARGET_SLOW_PSHUFB
4990 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4992 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
4993 * config/i386/i386.c (intel_cost): Ditto.
4995 2014-04-17 Joey Ye <joey.ye@arm.com>
4997 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
4999 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5001 * opts.c (common_handle_option): Disable -fipa-reference coorectly
5004 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5006 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
5007 (type_all_derivations_known_p): New predicate.
5008 (type_all_ctors_visible_p): New predicate.
5009 (type_possibly_instantiated_p): New predicate.
5010 (get_odr_type): Compute all_derivations_known.
5011 (dump_odr_type): Dump the flag.
5012 (maybe_record_type): Cleanup.
5013 (record_target_from_binfo): Add bases_to_consider array;
5014 record bases for types w/o instances and skip CXX destructor.
5015 (possible_polymorphic_call_targets_1): Add bases_to_consider
5016 and consider_construction parameters; check if type may have instance.
5017 (get_polymorphic_call_info): Set maybe_in_construction to true
5018 when we know nothing.
5019 (record_targets_from_bases): Skip CXX destructors; they are
5020 never called for types in construction.
5021 (possible_polymorphic_call_targets): Do not record target when
5022 type may not have instance.
5024 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5027 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
5028 external aliases alive, too.
5030 2014-04-16 Andrew Pinski <apinski@cavium.com>
5032 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
5035 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
5037 * final.c (compute_alignments): Do not apply loop alignment to a block
5038 falling through to the exit.
5040 2014-04-16 Catherine Moore <clm@codesourcery.com>
5042 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
5043 Adjust constraints for microMIPS store patterns.
5045 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
5047 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
5049 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
5051 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
5052 (append_use): Run at -O0.
5053 (append_vdef): Likewise.
5054 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
5055 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
5057 2014-04-16 Jakub Jelinek <jakub@redhat.com>
5059 PR tree-optimization/60844
5060 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
5061 (propagate_op_to_single_use, remove_visited_stmt_chain,
5062 linearize_expr, repropagate_negates, reassociate_bb): Use it
5063 instead of gsi_remove.
5065 2014-04-16 Martin Jambor <mjambor@suse.cz>
5067 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
5068 ipa_transforms_to_apply.
5069 (cgraph_function_versioning): Assert that old_node has empty
5070 ipa_transforms_to_apply.
5071 * trans-mem.c (ipa_tm_create_version): Likewise.
5072 * tree-inline.c (tree_function_versioning): Do not duplicate
5073 ipa_transforms_to_apply.
5075 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5078 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
5080 Pass necessary as flags on 64-bit Solaris/x86.
5081 Use lowercase relocs for x86_64-*-*.
5082 * configure: Regenerate.
5084 2014-04-15 Jan Hubicka <jh@suse.cz>
5086 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
5087 (maybe_record_node, likely_target_p): Use it.
5089 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5092 Revert following patch
5094 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5097 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
5098 software floating point or no floating point registers, do not
5099 allow any type in the FPRs. Eliminate a test for SPE SIMD types
5100 in GPRs that occurs after we tested for GPRs that would never be
5103 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
5104 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
5105 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
5106 specifically allow DDmode, since that does not use the SPE SIMD
5109 2014-03-21 Mark Wielaard <mjw@redhat.com>
5111 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
5112 as unsigned or int depending on type and value used.
5114 2014-04-15 Richard Biener <rguenther@suse.de>
5116 PR rtl-optimization/56965
5117 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
5118 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
5120 * alias.c (true_dependence_1): Do not call
5121 nonoverlapping_component_refs_p.
5122 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
5123 nonoverlapping_component_refs_p.
5124 (indirect_refs_may_alias_p): Likewise.
5126 2014-04-15 Teresa Johnson <tejohnson@google.com>
5128 * cfg.c (dump_bb_info): Fix flags check.
5129 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
5131 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5133 PR rtl-optimization/60663
5134 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
5137 2014-04-15 Richard Biener <rguenther@suse.de>
5139 * lto-streamer.h (LTO_major_version): Bump to 4.
5141 2014-04-15 Richard Biener <rguenther@suse.de>
5143 * common.opt (lto_partition_model): New enum.
5144 (flto-partition=): Merge separate options with a single with argument,
5145 add -flto-partition=one support.
5146 * flag-types.h (enum lto_partition_model): Declare.
5147 * opts.c (finish_options): Remove duplicate -flto-partition=
5149 * lto-wrapper.c (run_gcc): Adjust.
5151 2014-04-15 Richard Biener <rguenther@suse.de>
5153 * alias.c (ncr_compar): New function.
5154 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
5156 2014-04-15 Richard Biener <rguenther@suse.de>
5158 * alias.c (record_component_aliases): Do not walk BINFOs.
5160 2014-04-15 Richard Biener <rguenther@suse.de>
5162 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5163 Add struct function argument and adjust.
5164 (find_func_aliases_for_call): Likewise.
5165 (find_func_aliases): Likewise.
5166 (find_func_clobbers): Likewise.
5167 (intra_create_variable_infos): Likewise.
5168 (compute_points_to_sets): Likewise.
5169 (ipa_pta_execute): Adjust. Do not push/pop cfun.
5171 2014-04-15 Richard Biener <rguenther@suse.de>
5173 * tree.c (iterative_hash_expr): Use enum tree_code_class
5174 to store TREE_CODE_CLASS.
5175 (tree_block): Likewise.
5176 (tree_set_block): Likewise.
5177 * tree.h (fold_build_pointer_plus_loc): Use
5178 convert_to_ptrofftype_loc.
5180 2014-04-15 Jakub Jelinek <jakub@redhat.com>
5183 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
5186 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
5188 * cfgloop.h (struct loop): Move force_vectorize down.
5189 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
5190 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
5191 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
5192 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
5193 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
5194 * tree-core.h (enum annot_expr_kind): Add new kind values.
5195 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
5196 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
5198 * tree.def (ANNOTATE_EXPR): Tweak comment.
5200 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5202 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
5205 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
5207 * tree.h (TYPE_IDENTIFIER): Declare.
5208 * tree.c (subrange_type_for_debug_p): Use it.
5209 * godump.c (go_format_type): Likewise.
5210 * dwarf2out.c (is_cxx_auto, modified_type_die,
5211 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
5212 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
5214 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5217 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
5219 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
5221 * config/i386/i386.c (examine_argument): Return bool. Return true if
5222 parameter should be passed in memory.
5223 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
5224 (construct_container): Update calls to examine_argument.
5225 (function_arg_advance_64): Ditto.
5226 (return_in_memory_32): Merge with ix86_return_in_memory.
5227 (return_in_memory_64): Ditto.
5228 (return_in_memory_ms_64): Ditto.
5230 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5232 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
5233 * coverage.c (coverage_compute_profile_id): Handle externally visible
5236 2014-04-14 Martin Jambor <mjambor@suse.cz>
5238 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
5239 DECL_DISREGARD_INLINE_LIMITS functions.
5241 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
5244 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
5246 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
5249 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
5250 optimize_insn_for_speed_p instead of
5251 optimize_function_for_speed_p.
5253 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
5255 * doc/invoke.texi (free): Document AArch64.
5257 2014-04-14 Richard Biener <rguenther@suse.de>
5259 PR tree-optimization/60042
5260 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
5261 (insert_into_preds_of_block): Do not prevent PHI insertion
5262 for REFERENCE exprs here ...
5263 (eliminate_dom_walker::before_dom_children): ... but prevent
5264 their use here under similar conditions when applied to the
5265 IL after PRE optimizations.
5267 2014-04-14 Richard Biener <rguenther@suse.de>
5269 * passes.def: Move early points-to after early SRA.
5271 2014-04-14 Richard Biener <rguenther@suse.de>
5273 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
5274 check for which sign-changes we allow when forwarding
5275 a converted value into a switch.
5277 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
5279 * stor-layout.c (place_field): Finalize non-constant offset for the
5282 2014-04-14 Richard Biener <rguenther@suse.de>
5284 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
5286 (expand_switch_using_bit_tests_p): Likewise.
5287 (process_switch): Compute and pass on speed_p based on the
5289 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
5290 optimize_bb_for_speed_p.
5292 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
5294 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
5295 * function.h (struct function): Rename has_force_vect_loops into
5296 has_force_vectorize_loops.
5297 * lto-streamer-in.c (input_cfg): Adjust for renaming.
5298 (input_struct_function_base): Likewise.
5299 * lto-streamer-out.c (output_cfg): Likewise.
5300 (output_struct_function_base): Likewise.
5301 * omp-low.c (expand_omp_simd): Likewise.
5302 * tree-cfg.c (move_sese_region_to_fn): Likewise.
5303 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
5304 (version_loop_for_if_conversion): Likewise.
5305 (tree_if_conversion): Likewise.
5306 (main_tree_if_conversion): Likewise.
5307 (gate_tree_if_conversion): Likewise.
5308 * tree-inline.c (copy_loops): Likewise.
5309 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
5310 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
5311 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
5312 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
5313 * tree-vectorizer.c (vectorize_loops): Likewise.
5314 * tree-vectorizer.h (unlimited_cost_model): Likewise.
5316 2014-04-14 Richard Biener <rguenther@suse.de>
5319 * lto-streamer-out.c (wrap_refs): New function.
5320 (lto_output): Wrap symbol references in global initializes in
5321 type-preserving MEM_REFs.
5323 2014-04-14 Christian Bruel <christian.bruel@st.com>
5325 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
5327 2014-04-14 Christian Bruel <christian.bruel@st.com>
5329 * config/sh/sh.md (setmemqi): New expand pattern.
5330 * config/sh/sh.h (CLEAR_RATIO): Define.
5331 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
5332 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
5334 2014-04-14 Richard Biener <rguenther@suse.de>
5337 * fold-const.c (negate_expr_p): Don't negate directional rounding
5339 (fold_negate_expr): Likewise.
5341 2014-04-14 Richard Biener <rguenther@suse.de>
5343 PR tree-optimization/59817
5344 PR tree-optimization/60453
5345 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
5346 recursion to catch all CHRECs in the scalar evolution and restrict
5347 the predicate for the remains appropriately.
5349 2014-04-12 Catherine Moore <clm@codesourcery.com>
5351 * config/mips/constraints.md: Add new register constraint "kb".
5352 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
5353 (*movhi_internal): Likewise.
5354 (*movqi_internal): Likewise.
5355 * config/mips/mips.h (M16_STORE_REGS): New register class.
5356 (REG_CLASS_NAMES): Add M16_STORE_REGS.
5357 (REG_CLASS_CONTENTS): Likewise.
5358 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
5360 2014-04-11 Tobias Burnus <burnus@net-b.de>
5363 * doc/invoke.texi (-Wformat-signedness): Document it.
5364 (Wformat=2): Mention that this enables -Wformat-signedness.
5366 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5368 * common/config/epiphany/epiphany-common.c
5369 (epiphany_option_optimization_table): Enable section anchors by
5370 default at -O1 or higher.
5371 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
5372 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
5373 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
5374 carries no extra cost.
5375 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
5376 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
5377 * config/epiphany/predicates.md (memclob_operand): New predicate.
5378 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
5379 Use memclob_operand predicate and X constraint for operand 3.
5381 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5383 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
5384 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
5387 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5389 PR rtl-optimization/60651
5390 * mode-switching.c (optimize_mode_switching): Make sure to emit
5391 sets of a lower numbered entity before sets of a higher numbered
5392 entity to a mode of the same or lower priority.
5393 When creating a seginfo for a basic block that starts with a code
5394 label, move the insertion point past the code label.
5395 (new_seginfo): Document and enforce requirement that
5396 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
5397 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
5398 * doc/tm.texi: Regenerate.
5400 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
5403 * config/arc/arc.c (arc_save_restore): Fix assert typo.
5405 2013-04-11 Jakub Jelinek <jakub@redhat.com>
5407 * BASE-VER: Set to 4.10.0.
5409 2014-04-11 Tobias Burnus <burnus@net-b.de>
5412 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
5413 * doc/gcc.texi (Service): Update description in the @menu
5414 * doc/invoke.texi (Option Summary): Remove misplaced and
5417 2014-04-11 Steve Ellcey <sellcey@mips.com>
5418 Jakub Jelinek <jakub@redhat.com>
5421 * expr.c (convert_move): Use emit_store_flag_force instead of
5422 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
5425 2014-04-11 Richard Biener <rguenther@suse.de>
5428 * varasm.c (assemble_alias): Avoid endless error reporting
5429 recursion by setting TREE_ASM_WRITTEN.
5431 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5433 * config/s390/s390.md: Add a splitter for NOT rtx.
5435 2014-04-11 Jakub Jelinek <jakub@redhat.com>
5437 PR rtl-optimization/60663
5438 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
5440 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
5441 Jakub Jelinek <jakub@redhat.com>
5444 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
5445 flag from decl_node to node.
5447 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5450 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
5451 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
5452 ameliorating the cases where it can be.
5454 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
5457 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
5459 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5460 (loadsync_<mode>): Change mode.
5461 (load_quadpti, store_quadpti): New.
5462 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5463 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5464 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
5466 2014-04-09 Cong Hou <congh@google.com>
5469 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
5472 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5474 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
5475 instead of vnor to exploit possible fusion opportunity in the
5477 (altivec_expand_vec_perm_const_le): Likewise.
5479 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
5481 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5482 (loadsync_<mode>): Change mode.
5483 (load_quadpti, store_quadpti): New.
5484 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5485 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5487 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
5490 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
5491 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
5492 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
5494 2014-04-08 Richard Biener <rguenther@suse.de>
5497 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
5498 a 64bit widest int print double-int similar to on HWI64 hosts.
5500 2014-04-08 Richard Biener <rguenther@suse.de>
5502 PR tree-optimization/60785
5503 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
5504 default defs properly.
5506 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
5508 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
5509 (Weffc++): Likewise.
5511 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
5513 * ipa-devirt.c (maybe_record_node): When node is not recorded,
5514 set completep to false rather than true.
5516 2014-04-07 Douglas B Rupp <rupp@adacore.com>
5519 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
5520 ARM_TARGET2_DWARF_FORMAT.
5522 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
5525 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
5526 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
5529 2014-04-07 Richard Biener <rguenther@suse.de>
5531 PR tree-optimization/60766
5532 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
5533 (may_eliminate_iv): Convert cand_value_at result to desired type.
5535 2014-04-07 Jason Merrill <jason@redhat.com>
5538 * common.opt (-fno-gnu-unique): Add.
5539 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
5541 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5543 * haifa-sched.c: Fix outdated function reference and minor
5544 grammar errors in introductory comment.
5546 2014-04-07 Richard Biener <rguenther@suse.de>
5549 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
5551 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
5553 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
5556 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
5557 size accounting for thunks.
5558 (pa_asm_output_mi_thunk): Use final_start_function() and
5559 final_end_function() to output function start and end directives.
5561 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
5563 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
5564 device specific ISA/ feature information. Remove short_sp and
5565 errata_skip ds. Add avr_device_specific_features enum to have device
5567 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
5568 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
5569 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
5570 updated device specific info.
5571 * config/avr/avr-mcus.def: Merge device specific details to
5572 dev_attribute field.
5573 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
5575 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
5576 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
5577 assembler if RMW isa supported by current device.
5578 * config/avr/genmultilib.awk: Update as device info structure changed.
5579 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
5581 2014-04-04 Cong Hou <congh@google.com>
5583 PR tree-optimization/60656
5584 * tree-vect-stmts.c (supportable_widening_operation):
5585 Fix a bug that elements in a vector with vect_used_by_reduction
5586 property are incorrectly reordered when the operation on it is not
5587 consistant with the one in reduction operation.
5589 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
5591 PR rtl-optimization/60155
5592 * gcse.c (record_set_data): New function.
5593 (single_set_gcse): New function.
5594 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
5595 (hoist_code): Likewise.
5596 (get_pressure_class_and_nregs): Likewise.
5598 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
5600 * explow.c (probe_stack_range): Emit a final optimization blockage.
5602 2014-04-04 Anthony Green <green@moxielogic.com>
5604 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
5607 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
5610 * lto-cgraph.c (input_overwrite_node): Check that partitioning
5611 flags are set only during streaming.
5612 * ipa.c (process_references, walk_polymorphic_call_targets,
5613 symtab_remove_unreachable_nodes): Drop bodies of always inline
5614 after early inlining.
5615 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
5617 2014-04-04 Jakub Jelinek <jakub@redhat.com>
5618 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5621 * dwarf2out.c (const_ok_for_output_1): Reject expressions
5624 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5627 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
5629 (cortex_a53_fdivd): Likewise.
5631 2014-04-04 Martin Jambor <mjambor@suse.cz>
5634 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
5636 * cgraph.c (clone_of_p): Also return true if thunks match.
5637 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
5638 cgraph_function_or_thunk_node and an obsolete comment.
5639 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
5641 (build_function_decl_skip_args): Likewise.
5642 (set_new_clone_decl_and_node_flags): New function.
5643 (duplicate_thunk_for_node): Likewise.
5644 (redirect_edge_duplicating_thunks): Likewise.
5645 (cgraph_clone_node): New parameter args_to_skip, pass it to
5646 redirect_edge_duplicating_thunks which is called instead of
5647 cgraph_redirect_edge_callee.
5648 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
5649 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
5651 2014-04-04 Jeff Law <law@redhat.com>
5654 * config/arm/predicates.md (const_int_I_operand): New predicate.
5655 (const_int_M_operand): Similarly.
5656 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
5658 (insv_t2, extv_reg, extzv_t2): Likewise.
5659 (load_multiple_with_writeback): Similarly for const_int_I_operand.
5660 (pop_multiple_with_writeback_and_return): Likewise.
5661 (vfp_pop_multiple_with_writeback): Likewise
5663 2014-04-04 Richard Biener <rguenther@suse.de>
5666 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
5667 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
5669 * gimplify.h (gimple_add_tmp_var_fn): Declare.
5670 * gimplify.c (gimple_add_tmp_var_fn): New function.
5671 * gimple-expr.h (create_tmp_reg_fn): Declare.
5672 * gimple-expr.c (create_tmp_reg_fn): New function.
5673 * gimple-low.c (record_vars_into): Don't change cfun.
5674 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
5675 code generation without cfun.
5677 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
5680 * Makefile.in (install-driver): Fix shell scripting.
5682 2014-04-03 Cong Hou <congh@google.com>
5684 PR tree-optimization/60505
5685 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
5686 threshold of number of iterations below which no vectorization
5688 * tree-vect-loop.c (new_loop_vec_info):
5689 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
5690 * tree-vect-loop.c (vect_analyze_loop_operations):
5691 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
5692 * tree-vect-loop.c (vect_transform_loop):
5693 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
5694 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
5695 of iterations of the loop and see if we should build the epilogue.
5697 2014-04-03 Richard Biener <rguenther@suse.de>
5699 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
5700 (streamer_tree_cache_create): Adjust.
5701 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
5702 to allow optional nodes array.
5703 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
5704 (streamer_tree_cache_append): Likewise.
5705 (streamer_tree_cache_create): Create nodes array optionally
5706 as specified by parameter.
5707 * lto-streamer-out.c (create_output_block): Avoid maintaining
5708 the node array in the writer cache.
5709 (DFS_write_tree): Remove assertion.
5710 (produce_asm_for_decls): Free the out decl state hash table early.
5711 * lto-streamer-in.c (lto_data_in_create): Adjust for
5712 streamer_tree_cache_create prototype change.
5714 2014-04-03 Richard Biener <rguenther@suse.de>
5716 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
5717 set TREE_CHAIN to NULL_TREE.
5719 2014-04-03 Richard Biener <rguenther@suse.de>
5721 PR tree-optimization/60740
5722 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
5723 over all GIMPLE_COND operands.
5725 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
5727 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
5728 (Weffc++): Remove Scott's numbering, merge lists and reference
5731 2014-04-03 Nick Clifton <nickc@redhat.com>
5733 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
5736 2014-04-03 Martin Jambor <mjambor@suse.cz>
5738 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
5739 mention gcc_unreachable before failing.
5740 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
5743 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
5746 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
5747 inconsistent code and instead mark the context inconsistent.
5748 (possible_polymorphic_call_targets): For inconsistent contexts
5749 return empty complete list.
5751 2014-04-02 Anthony Green <green@moxielogic.com>
5753 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
5754 (extendqisi2, extendhisi2): Define.
5755 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
5756 (WCHAR_TYPE): Change to unsigned int.
5758 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5760 PR tree-optimization/60733
5761 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
5762 insertion point for PHI candidates to be the end of the feeding
5763 block for the PHI argument.
5765 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
5767 PR rtl-optimization/60650
5768 * lra-constraints.c (process_alt_operands): Decrease reject for
5769 earlyclobber matching.
5771 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5773 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
5775 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5777 * config/spu/spu.c (pad_bb): Do not crash when the last
5778 insn is CODE_FOR_blockage.
5780 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5782 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
5783 lies outside the target mode.
5785 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5788 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
5789 software floating point or no floating point registers, do not
5790 allow any type in the FPRs. Eliminate a test for SPE SIMD types
5791 in GPRs that occurs after we tested for GPRs that would never be
5794 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
5795 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
5796 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
5797 specifically allow DDmode, since that does not use the SPE SIMD
5800 2014-04-02 Richard Biener <rguenther@suse.de>
5803 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
5804 MODE_INTs. Properly use negv_optab.
5805 (expand_abs): Likewise.
5807 2014-04-02 Richard Biener <rguenther@suse.de>
5810 * Makefile.in (install-driver): Guard extra installs with special
5813 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
5815 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5818 2014-04-01 Richard Henderson <rth@redhat.com>
5821 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
5822 alternative enabled before register allocation.
5824 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
5826 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
5827 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
5829 (nios2_large_got_address): Remove unneeded 'sym' parameter.
5830 (nios2_got_address): Update nios2_large_got_address call site.
5831 (nios2_delegitimize_address): New function.
5832 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
5833 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
5834 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
5836 2014-04-01 Martin Husemann <martin@duskware.de>
5838 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
5841 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
5843 PR rtl-optimization/60604
5844 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
5845 check from register_operand.
5846 (register_operand): Redefine in terms of general_operand.
5847 (nonmemory_operand): Use register_operand for the non-constant cases.
5849 2014-04-01 Richard Biener <rguenther@suse.de>
5851 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
5853 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
5855 * doc/invoke.texi (mapp-regs): Clarify.
5857 2014-03-31 Ulrich Drepper <drepper@gmail.com>
5859 * config/i386/avx512fintrin.h (__v32hi): Define type.
5860 (__v64qi): Likewise.
5861 (_mm512_set1_epi8): Define.
5862 (_mm512_set1_epi16): Define.
5863 (_mm512_set4_epi32): Define.
5864 (_mm512_set4_epi64): Define.
5865 (_mm512_set4_pd): Define.
5866 (_mm512_set4_ps): Define.
5867 (_mm512_setr4_epi64): Define.
5868 (_mm512_setr4_epi32): Define.
5869 (_mm512_setr4_pd): Define.
5870 (_mm512_setr4_ps): Define.
5871 (_mm512_setzero_epi32): Define.
5873 2014-03-31 Martin Jambor <mjambor@suse.cz>
5876 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
5877 callsite_arguments_match_p. Updated all callers. Also check types of
5878 corresponding formal parameters and actual arguments.
5879 (not_all_callers_have_enough_arguments_p) Renamed to
5880 some_callers_have_mismatched_arguments_p.
5882 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
5884 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
5886 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
5889 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
5892 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
5894 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
5895 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
5897 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
5898 Use FMAMODE_NOVF512 mode iterator.
5899 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
5900 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
5901 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
5902 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
5904 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
5905 Use VF_128_256 mode iterator.
5906 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
5909 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5911 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
5912 static chain if needed.
5914 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5917 * lra-constraints.c (index_part_to_reg): New.
5918 (process_address): Use it.
5920 2014-03-27 Jeff Law <law@redhat.com>
5921 Jakub Jelinek <jakub@redhat.com>
5924 * expr.c (do_tablejump): Use simplify_gen_binary rather than
5925 gen_rtx_{PLUS,MULT} to build up the address expression.
5927 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
5928 creating non-canonical RTL.
5930 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5933 * ipa-inline.c (want_inline_small_function_p): Short circuit large
5934 functions; reorganize to make cheap checks first.
5935 (inline_small_functions): Do not estimate growth when dumping;
5937 * ipa-inline.h (inline_summary): Add min_size.
5938 (growth_likely_positive): New function.
5939 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
5940 (set_cond_stmt_execution_predicate): Cleanup.
5941 (estimate_edge_size_and_time): Compute min_size.
5942 (estimate_calls_size_and_time): Likewise.
5943 (estimate_node_size_and_time): Likewise.
5944 (inline_update_overall_summary): Update min_size.
5945 (do_estimate_edge_time): Likewise.
5946 (do_estimate_edge_size): Update.
5947 (do_estimate_edge_hints): Update.
5948 (growth_likely_positive): New function.
5950 2014-03-28 Jakub Jelinek <jakub@redhat.com>
5953 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
5954 also if addr has VOIDmode.
5956 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5958 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
5959 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
5961 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
5962 instructions as well as AdvancedSIMD loads.
5964 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5966 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
5967 Use crypto_aese type.
5968 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
5969 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
5970 crypto_aese, crypto_aesmc. Move to types.md.
5971 * config/arm/types.md (crypto_aes): Split into crypto_aese,
5973 * config/arm/iterators.md (crypto_type): Likewise.
5975 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5977 * cgraph.c: Include expr.h and tree-dfa.h.
5978 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
5981 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5984 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
5985 regs from checking multi-reg pseudos.
5987 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5989 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
5991 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5993 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
5994 if it would clobber the stack pointer, even temporarily.
5996 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
5998 * mode-switching.c: Make small adjustments to the top comment.
6000 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6002 * config/rs6000/constraints.md (wD constraint): New constraint to
6003 match the constant integer to get the top DImode/DFmode out of a
6004 vector in a VSX register.
6006 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
6007 match the constant integer to get the top DImode/DFmode out of a
6008 vector in a VSX register.
6010 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
6013 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6016 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
6017 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
6019 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
6020 Optimize vec_extract of 64-bit values, where the value being
6021 extracted is in the top word, where we can use scalar
6022 instructions. Add direct move and store support. Combine the big
6023 endian/little endian vector select load support into a single insn.
6024 (vsx_extract_<mode>_internal1): Likewise.
6025 (vsx_extract_<mode>_internal2): Likewise.
6026 (vsx_extract_<mode>_load): Likewise.
6027 (vsx_extract_<mode>_store): Likewise.
6028 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
6029 combined into vsx_extract_<mode>_load.
6030 (vsx_extract_<mode>_one_le): Likewise.
6032 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
6033 define the top 64-bit vector element.
6035 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
6038 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6039 Document vec_vbpermq builtin.
6042 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
6043 enable use of xxsldwi and xxpermdi builtin functions.
6044 (vec_xxpermdi): Likewise.
6046 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6047 Document use of vec_xxsldwi and vec_xxpermdi builtins.
6049 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
6051 PR rtl-optimization/60650
6052 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
6054 (find_spills_for): New.
6055 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
6056 Spill all pseudos on the second iteration.
6058 2014-03-27 Marek Polacek <polacek@redhat.com>
6061 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
6064 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6066 * config/s390/s390.c (s390_can_use_return_insn): Check for
6067 call-saved FPRs on 31 bit.
6069 2014-03-27 Jakub Jelinek <jakub@redhat.com>
6072 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
6073 if they need regimplification, just drop them instead of
6074 calling gimple_regimplify_operands on them.
6076 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
6079 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
6080 (aarch64_frame_pointer_required): Adjust logic.
6081 (aarch64_can_eliminate): Adjust logic.
6082 (aarch64_override_options_after_change): Adjust logic.
6084 2014-03-27 Dehao Chen <dehao@google.com>
6086 * ipa-inline.c (early_inliner): Update node's inline info.
6088 2014-03-26 Dehao Chen <dehao@google.com>
6090 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
6091 compiler inserted conditional jumps for NAN float check.
6093 2014-03-26 Jakub Jelinek <jakub@redhat.com>
6095 * ubsan.h (ubsan_create_data): Change second argument's type
6096 to const location_t *.
6097 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
6099 (ubsan_create_data): Change second argument to const location_t *PLOC.
6100 Create Loc field whenever PLOC is non-NULL.
6101 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
6102 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
6106 * real.c (real_to_integer2): Change type of low to UHWI.
6108 2014-03-26 Tobias Burnus <burnus@net-b.de>
6110 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
6111 (CILK_SELF_SPECS): New define.
6112 (driver_self_specs): Use it.
6114 2014-03-26 Richard Biener <rguenther@suse.de>
6116 * tree-pretty-print.c (percent_K_format): Implement special
6117 case for LTO and its stripped down BLOCK tree.
6119 2014-03-26 Jakub Jelinek <jakub@redhat.com>
6122 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
6124 * tree-vrp.c (simplify_internal_call_using_ranges): If only
6125 one range is range_int_cst_p, but not both, at least optimize
6126 addition/subtraction of 0 and multiplication by 0 or 1.
6127 * gimple-fold.c (gimple_fold_call): Fold
6128 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
6129 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
6130 INTEGER_CSTs, try to fold at least x * 0 and y - y.
6132 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
6134 PR rtl-optimization/60452
6135 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
6136 <case REG>: Return 1 for invalid offsets from the frame pointer.
6138 2014-03-26 Marek Polacek <polacek@redhat.com>
6141 * doc/extend.texi (C Extensions): Mention variable-length arrays in
6144 2014-03-26 Marek Polacek <polacek@redhat.com>
6147 * doc/extend.texi (Designated Inits): Describe what happens to omitted
6150 2014-03-26 Marek Polacek <polacek@redhat.com>
6153 * ira-color.c (update_conflict_hard_regno_costs): Perform the
6154 multiplication in unsigned type.
6156 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
6158 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
6160 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
6162 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
6164 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
6167 * cif-code.def (UNREACHABLE) New code.
6168 * ipa-inline.c (inline_small_functions): Skip edges to
6169 __builtlin_unreachable.
6170 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
6171 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
6172 predicate to __bulitin_unreachable.
6173 (set_cond_stmt_execution_predicate): Fix issue when
6174 invert_tree_comparison returns ERROR_MARK.
6175 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
6176 propagate to inline clones.
6177 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
6179 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
6180 * cgraphclones.c (cgraph_clone_node): If call destination is already
6181 ureachable, do not redirect it back.
6182 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
6185 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
6187 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
6188 Do not modify inline clones.
6190 2014-03-25 Jakub Jelinek <jakub@redhat.com>
6192 * config/i386/i386.md (general_sext_operand): New mode attr.
6193 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
6194 don't generate (sign_extend (const_int)).
6195 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
6196 operands[2]. Use We constraint instead of <i> and
6197 <general_sext_operand> predicate instead of <general_operand>.
6198 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
6199 * config/i386/constraints.md (We): New constraint.
6200 * config/i386/predicates.md (x86_64_sext_operand,
6201 sext_operand): New predicates.
6203 2014-03-25 Martin Jambor <mjambor@suse.cz>
6206 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
6207 inconsistent devirtualizations to __builtin_unreachable.
6209 2014-03-25 Marek Polacek <polacek@redhat.com>
6212 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
6214 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
6216 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
6217 order of elements for big-endian.
6219 2014-03-25 Richard Biener <rguenther@suse.de>
6222 * gimplify-me.c (gimple_regimplify_operands): Update the
6225 2014-03-25 Martin Jambor <mjambor@suse.cz>
6228 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
6229 (lto_output_varpool_node): Likewise.
6230 (input_overwrite_node): Likewise.
6231 (input_varpool_node): Likewise.
6233 2014-03-25 Richard Biener <rguenther@suse.de>
6235 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
6236 (run_gcc): Likewise.
6238 2014-03-25 Jakub Jelinek <jakub@redhat.com>
6240 * combine.c (simplify_compare_const): Add MODE argument.
6241 Handle mode_width 0 as very large mode_width.
6242 (try_combine, simplify_comparison): Adjust callers.
6244 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
6245 type to avoid signed integer overflow.
6246 * explow.c (plus_constant): Likewise.
6248 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
6250 * doc/generic.texi: Correct typos.
6252 2014-03-24 Tobias Burnus <burnus@net-b.de>
6254 * doc/invoke.texi (-flto): Expand section about
6255 using static libraries with LTO.
6257 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6259 PR rtl-optimization/60501
6260 * optabs.def (addptr3_optab): New optab.
6261 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
6262 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
6263 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
6265 * lra.c (emit_add3_insn): Use the addptr pattern if available.
6267 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
6269 2014-03-24 Ulrich Drepper <drepper@gmail.com>
6271 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
6274 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
6275 (_mm256_undefined_ps): Define.
6276 (_mm256_undefined_pd): Define.
6277 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
6278 (_mm_undefined_pd): Define.
6279 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
6280 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
6281 (_mm512_undefined_ps): Define.
6282 (_mm512_undefined_pd): Define.
6283 Use _mm*_undefined_*.
6284 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
6286 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
6288 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
6289 (lshr_simd): DI mode added.
6290 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
6291 (aarch64_ushr_simddi): Likewise.
6292 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
6293 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
6294 (vshrd_n_u64): Likewise.
6296 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6298 * Makefile.in (s-macro_list): Depend on cc1.
6300 2014-03-23 Teresa Johnson <tejohnson@google.com>
6302 * ipa-utils.c (ipa_print_order): Use specified dump file.
6304 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
6306 PR rtl-optimization/60601
6307 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
6309 * gcc.c (eval_spec_function): Initialize save_growing_value.
6311 2014-03-22 Jakub Jelinek <jakub@redhat.com>
6314 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
6315 code == MINUS_EXPR, never swap op0 with op1.
6317 * toplev.c (init_local_tick): Avoid signed integer multiplication
6319 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
6320 shift by first operand's bitsize.
6322 2014-03-21 Jakub Jelinek <jakub@redhat.com>
6325 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
6327 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
6328 TARGET_ISA_64BIT_P(x).
6330 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6332 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
6333 pattern for vector nor instead of subtract from splat(-1).
6334 (altivec_expand_vec_perm_const_le): Likewise.
6336 2014-03-21 Richard Henderson <rth@twiddle.net>
6339 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
6340 related insns after epilogue_completed.
6342 2014-03-21 Martin Jambor <mjambor@suse.cz>
6345 * cgraph.h (symtab_node): New flag body_removed.
6346 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
6347 when removing bodies.
6348 * symtab.c (dump_symtab_base): Dump body_removed flag.
6349 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
6350 had their bodies removed.
6352 2014-03-21 Martin Jambor <mjambor@suse.cz>
6355 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
6358 2014-03-21 Richard Biener <rguenther@suse.de>
6360 PR tree-optimization/60577
6361 * tree-core.h (struct tree_base): Document nothrow_flag use
6362 in VAR_DECL_NONALIASED.
6363 * tree.h (VAR_DECL_NONALIASED): New.
6364 (may_be_aliased): Adjust.
6365 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
6367 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6369 * expr.c (expand_expr_real_1): Remove outdated comment.
6371 2014-03-20 Jakub Jelinek <jakub@redhat.com>
6374 * ira.c (adjust_cleared_regs): Call copy_rtx on
6375 *reg_equiv[REGNO (loc)].src_p before passing it to
6376 simplify_replace_fn_rtx.
6379 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
6380 into CONST, put pic register as first operand of PLUS. Use
6381 gen_const_mem for both 32-bit and 64-bit PIC got loads.
6383 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6385 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
6387 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6389 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
6390 around for store forwarding issue in the FPU on the UT699.
6391 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
6392 loads and operations if -mfix-ut699 is specified.
6393 (divtf3_hq): Tweak attribute.
6394 (sqrttf2_hq): Likewise.
6396 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6398 * calls.c (store_one_arg): Remove incorrect const qualification on the
6399 type of the temporary.
6400 * cfgexpand.c (expand_return): Likewise.
6401 * expr.c (expand_constructor): Likewise.
6402 (expand_expr_real_1): Likewise.
6404 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6406 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
6409 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
6412 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
6414 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
6416 * config/arm/aarch-common-protos.h
6417 (alu_cost_table): Fix spelling of "extend".
6418 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
6420 2014-03-19 Richard Biener <rguenther@suse.de>
6423 * tree-core.h (tree_type_common): Re-order pointer members
6424 to reduce recursion depth during GC walks.
6426 2014-03-19 Marek Polacek <polacek@redhat.com>
6429 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
6430 before accessing it.
6432 2014-03-19 Richard Biener <rguenther@suse.de>
6435 * lto-streamer-in.c (input_function): In WPA stage do not drop
6438 2014-03-19 Jakub Jelinek <jakub@redhat.com>
6440 PR tree-optimization/60559
6441 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
6442 with build_zero_cst assignment.
6444 2014-03-18 Kai Tietz <ktietz@redhat.com>
6446 PR rtl-optimization/56356
6447 * sdbout.c (sdbout_parms): Verify that parms'
6448 incoming argument is valid.
6449 (sdbout_reg_parms): Likewise.
6451 2014-03-18 Richard Henderson <rth@redhat.com>
6454 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
6455 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
6456 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
6458 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
6460 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
6461 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
6462 Italicize plugin event names in description. Explain that
6463 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
6464 Remind that no GCC functions should be called after PLUGIN_FINISH.
6465 Explain what pragmas with expansion are.
6467 2014-03-18 Martin Liska <mliska@suse.cz>
6469 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
6470 gimple call statement is update.
6471 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
6472 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
6474 2014-03-18 Jakub Jelinek <jakub@redhat.com>
6477 * ubsan.c (ubsan_instrument_unreachable): Call
6478 initialize_sanitizer_builtins.
6479 (ubsan_pass): Likewise.
6482 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
6483 varpool_finalize_decl instead of rest_of_decl_compilation.
6485 2014-03-18 Richard Biener <rguenther@suse.de>
6487 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
6488 by using bitmap_and_compl instead of bitmap_and_compl_into.
6489 (df_rd_transfer_function): Likewise.
6491 2014-03-18 Richard Biener <rguenther@suse.de>
6493 * doc/lto.texi (fresolution): Fix typo.
6495 2014-03-18 Richard Biener <rguenther@suse.de>
6497 * doc/invoke.texi (flto): Update for changes in 4.9.
6499 2014-03-18 Richard Biener <rguenther@suse.de>
6501 * doc/loop.texi: Remove section on the removed lambda framework.
6502 Update loop docs with recent changes in preserving loop structure.
6504 2014-03-18 Richard Biener <rguenther@suse.de>
6506 * doc/lto.texi (-fresolution): Document.
6508 2014-03-18 Richard Biener <rguenther@suse.de>
6510 * doc/contrib.texi: Adjust my name.
6512 2014-03-18 Jakub Jelinek <jakub@redhat.com>
6515 * internal-fn.c: Include diagnostic-core.h.
6516 (expand_BUILTIN_EXPECT): New function.
6517 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
6518 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
6519 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
6520 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
6522 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
6523 Revert 3 argument __builtin_expect code.
6524 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
6525 * gimple-fold.c (gimple_fold_call): Likewise.
6526 * tree.h (fold_builtin_expect): New prototype.
6527 * builtins.c (build_builtin_expect_predicate): Add predictor
6528 argument, if non-NULL, create 3 argument __builtin_expect.
6529 (fold_builtin_expect): No longer static. Add ARG2 argument,
6530 pass it through to build_builtin_expect_predicate.
6531 (fold_builtin_2): Adjust caller.
6532 (fold_builtin_3): Handle BUILT_IN_EXPECT.
6533 * internal-fn.def (BUILTIN_EXPECT): New.
6535 2014-03-18 Tobias Burnus <burnus@net-b.de>
6538 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
6539 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
6540 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
6542 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
6545 * predict.c (combine_predictions_for_bb): Fix up formatting.
6546 (expr_expected_value_1, expr_expected_value): Add predictor argument,
6547 fill what it points to if non-NULL.
6548 (tree_predict_by_opcode): Adjust caller, use the predictor.
6549 * predict.def (PRED_COMPARE_AND_SWAP): Add.
6551 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
6553 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
6554 proper constant for the store mode.
6556 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
6558 * symtab.c (change_decl_assembler_name): Fix transparent alias
6561 2014-03-16 Renlin Li <Renlin.Li@arm.com>
6563 * config/aarch64/aarch64.c: Correct the comments about the
6564 aarch64 stack layout.
6566 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
6568 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
6569 check for GF_OMP_FOR_KIND_FOR.
6571 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
6573 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
6574 ymm and zmm register names.
6576 2014-03-17 Jakub Jelinek <jakub@redhat.com>
6579 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
6580 note creation for the 2010-08-31 changes.
6582 2014-03-17 Marek Polacek <polacek@redhat.com>
6585 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
6586 as -fno-tree-loop-vectorize.
6587 (expand_omp_simd): Likewise.
6589 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
6591 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
6592 (eligible_for_call_delay): New prototype.
6593 * config/sparc/sparc.c (tls_call_delay): Rename into...
6594 (eligible_for_call_delay): ...this. Return false if the instruction
6595 cannot be put in the delay slot of a branch.
6596 (eligible_for_restore_insn): Simplify.
6597 (eligible_for_return_delay): Return false if the instruction cannot be
6598 put in the delay slot of a branch and simplify.
6599 (eligible_for_sibcall_delay): Return false if the instruction cannot be
6600 put in the delay slot of a branch.
6601 * config/sparc/sparc.md (fix_ut699): New attribute.
6602 (tls_call_delay): Delete.
6603 (in_call_delay): Reimplement.
6604 (eligible_for_sibcall_delay): Rename into...
6605 (in_sibcall_delay): ...this.
6606 (eligible_for_return_delay): Rename into...
6607 (in_return_delay): ...this.
6608 (in_branch_delay): Reimplement.
6609 (in_uncond_branch_delay): Delete.
6610 (in_annul_branch_delay): Delete.
6612 2014-03-14 Richard Henderson <rth@redhat.com>
6615 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
6616 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
6617 (*floathi<X87MODEF>2_i387_with_temp): Remove.
6618 (floathi splitters): Remove.
6619 (float<SWI48x>xf2): New pattern.
6620 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
6621 code that tried to handle DImode for 32-bit, but which was excluded
6622 by the pattern's condition. Drop allocation of stack temporary.
6623 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
6624 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
6625 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
6626 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
6627 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
6628 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
6629 (*float<SWI48><MODEF>2_sse_interunit): Remove.
6630 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
6631 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
6632 (*float<SWI48x><X87MODEF>2_i387): Remove.
6633 (all float _with_temp splitters): Remove.
6634 (*float<SWI48x><MODEF>2_i387): New pattern.
6635 (*float<SWI48><MODEF>2_sse): New pattern.
6636 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
6637 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
6639 2014-03-14 Jakub Jelinek <jakub@redhat.com>
6640 Marek Polacek <polacek@redhat.com>
6643 * common.opt (dump_base_name_prefixed): New Variable.
6644 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
6645 if x_dump_base_name_prefixed is already set, set it at the end.
6647 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
6649 PR rtl-optimization/60508
6650 * lra-constraints.c (get_reload_reg): Add new parameter
6652 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
6653 Pass the new parameter values.
6655 2014-03-14 Richard Biener <rguenther@suse.de>
6657 * common.opt: Revert unintented changes from r205065.
6660 2014-03-14 Richard Biener <rguenther@suse.de>
6663 * cfghooks.c (split_block): Properly adjust all loops the
6664 block was a latch of.
6666 2014-03-14 Martin Jambor <mjambor@suse.cz>
6669 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
6672 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
6675 * config/avr/avr.c (avr_set_current_function): Pass function name
6676 through default_strip_name_encoding before sanity checking instead
6677 of skipping the first char of the assembler name.
6679 2014-03-13 Richard Henderson <rth@redhat.com>
6682 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
6683 (ix86_force_to_memory, ix86_free_from_memory): Remove.
6684 * config/i386/i386-protos.h: Likewise.
6685 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
6686 in the expander instead of a splitter.
6687 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
6688 any possibility of requiring a memory.
6689 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
6690 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
6691 (fp branch splitters): Update for ix86_split_fp_branch.
6692 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
6693 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
6694 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
6695 (*fop_<MODEF>_2_i387): Remove f/r alternative.
6696 (*fop_<MODEF>_3_i387): Likewise.
6697 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
6698 (splitters for the fop_* register patterns): Remove.
6699 (fscalexf4_i387): Rename from *fscalexf4_i387.
6700 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
6702 2014-03-13 Jakub Jelinek <jakub@redhat.com>
6704 PR tree-optimization/59779
6705 * tree-dfa.c (get_ref_base_and_extent): Use double_int
6706 type for bitsize and maxsize instead of HOST_WIDE_INT.
6708 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
6710 PR rtl-optimization/57320
6711 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
6712 the CFG after thread_prologue_and_epilogue_insns.
6714 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
6716 PR rtl-optimization/57189
6717 * lra-constraints.c (process_alt_operands): Disfavor spilling
6720 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
6722 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
6724 2014-03-13 Jakub Jelinek <jakub@redhat.com>
6726 PR tree-optimization/59025
6728 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
6729 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
6731 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
6734 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
6735 calls of avr_out_plus_1.
6737 2014-03-13 Bin Cheng <bin.cheng@arm.com>
6739 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
6740 BB's single pred and update the father loop's latch info later.
6742 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
6744 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
6748 (VEC_base): Likewise.
6749 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
6750 registers, we need to swap double words in little endian mode.
6752 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
6753 to be a container mode for 128-bit integer operations added in ISA
6754 2.07. Unlike TImode and PTImode, the preferred register set is
6755 the Altivec/VMX registers for the 128-bit operations.
6757 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
6759 (rs6000_split_128bit_ok_p): Likewise.
6761 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
6762 macros for creating ISA 2.07 normal and overloaded builtin
6763 functions with 3 arguments.
6764 (BU_P8V_OVERLOAD_3): Likewise.
6765 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
6766 for use as overloaded functions.
6767 (VPERM_1TI_UNS): Likewise.
6768 (VSEL_1TI): Likewise.
6769 (VSEL_1TI_UNS): Likewise.
6770 (ST_INTERNAL_1ti): Likewise.
6771 (LD_INTERNAL_1ti): Likewise.
6772 (XXSEL_1TI): Likewise.
6773 (XXSEL_1TI_UNS): Likewise.
6774 (VPERM_1TI): Likewise.
6775 (VPERM_1TI_UNS): Likewise.
6776 (XXPERMDI_1TI): Likewise.
6777 (SET_1TI): Likewise.
6778 (LXVD2X_V1TI): Likewise.
6779 (STXVD2X_V1TI): Likewise.
6780 (VEC_INIT_V1TI): Likewise.
6781 (VEC_SET_V1TI): Likewise.
6782 (VEC_EXT_V1TI): Likewise.
6783 (EQV_V1TI): Likewise.
6784 (NAND_V1TI): Likewise.
6785 (ORC_V1TI): Likewise.
6786 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
6787 added in ISA 2.07. Add both normal 'altivec' builtins, and the
6789 (VADDUQM): Likewise.
6790 (VSUBCUQ): Likewise.
6791 (VADDEUQM): Likewise.
6792 (VADDECUQ): Likewise.
6793 (VSUBEUQM): Likewise.
6794 (VSUBECUQ): Likewise.
6796 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
6797 __int128_t and __uint128_t types.
6798 (__uint128_type): Likewise.
6799 (altivec_categorize_keyword): Add support for vector __int128_t,
6800 vector __uint128_t, vector __int128, and vector unsigned __int128
6801 as a container type for TImode operations that need to be done in
6802 VSX/Altivec registers.
6803 (rs6000_macro_to_expand): Likewise.
6804 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
6805 to support 128-bit integer instructions vaddcuq, vadduqm,
6806 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
6807 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
6809 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
6810 for V1TImode, and set up preferences to use VSX/Altivec registers.
6811 Setup VSX reload handlers.
6812 (rs6000_debug_reg_global): Likewise.
6813 (rs6000_init_hard_regno_mode_ok): Likewise.
6814 (rs6000_preferred_simd_mode): Likewise.
6815 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
6816 (easy_altivec_constant): Likewise.
6817 (output_vec_const_move): Likewise.
6818 (rs6000_expand_vector_set): Convert V1TImode set and extract to
6820 (rs6000_expand_vector_extract): Likewise.
6821 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
6823 (rs6000_const_vec): Add support for V1TImode.
6824 (rs6000_emit_le_vsx_load): Swap double words when loading or
6825 storing TImode/V1TImode.
6826 (rs6000_emit_le_vsx_store): Likewise.
6827 (rs6000_emit_le_vsx_move): Likewise.
6828 (rs6000_emit_move): Add support for V1TImode.
6829 (altivec_expand_ld_builtin): Likewise.
6830 (altivec_expand_st_builtin): Likewise.
6831 (altivec_expand_vec_init_builtin): Likewise.
6832 (altivec_expand_builtin): Likewise.
6833 (rs6000_init_builtins): Add support for V1TImode type. Add
6834 support for ISA 2.07 128-bit integer builtins. Define type names
6835 for the VSX/Altivec vector types.
6836 (altivec_init_builtins): Add support for overloaded vector
6837 functions with V1TImode type.
6838 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
6839 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
6841 (rs6000_split_128bit_ok_p): Likewise.
6842 (rs6000_handle_altivec_attribute): Create V1TImode from vector
6843 __int128_t and vector __uint128_t.
6845 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
6846 and mode attributes.
6853 (VS_scalar): Likewise.
6854 (VS_double): Likewise.
6855 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
6857 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
6858 we support the ISA 2.07 128-bit integer arithmetic instructions.
6859 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
6860 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
6861 and TImode types for use with the builtin functions.
6862 (V1TI_type_node): Likewise.
6863 (unsigned_V1TI_type_node): Likewise.
6864 (intTI_type_internal_node): Likewise.
6865 (uintTI_type_internal_node): Likewise.
6867 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
6868 128-bit builtin functions.
6869 (UNSPEC_VADDEUQM): Likewise.
6870 (UNSPEC_VADDECUQ): Likewise.
6871 (UNSPEC_VSUBCUQ): Likewise.
6872 (UNSPEC_VSUBEUQM): Likewise.
6873 (UNSPEC_VSUBECUQ): Likewise.
6874 (VM): Add V1TImode to vector mode iterators.
6876 (VI_unit): Likewise.
6877 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
6878 (altivec_vaddcuq): Likewise.
6879 (altivec_vsubuqm): Likewise.
6880 (altivec_vsubcuq): Likewise.
6881 (altivec_vaddeuqm): Likewise.
6882 (altivec_vaddecuq): Likewise.
6883 (altivec_vsubeuqm): Likewise.
6884 (altivec_vsubecuq): Likewise.
6886 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
6888 (BOOL_128): Likewise.
6889 (BOOL_REGS_OUTPUT): Likewise.
6890 (BOOL_REGS_OP1): Likewise.
6891 (BOOL_REGS_OP2): Likewise.
6892 (BOOL_REGS_UNARY): Likewise.
6893 (BOOL_REGS_AND_CR0): Likewise.
6895 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
6896 128-bit integer builtin support.
6897 (vec_vadduqm): Likewise.
6898 (vec_vaddecuq): Likewise.
6899 (vec_vaddeuqm): Likewise.
6900 (vec_vsubecuq): Likewise.
6901 (vec_vsubeuqm): Likewise.
6902 (vec_vsubcuq): Likewise.
6903 (vec_vsubuqm): Likewise.
6905 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6906 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
6907 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
6908 128-bit integer add/subtract to ISA 2.07.
6910 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
6912 * config/arc/arc.c (arc_predicate_delay_insns):
6913 Fix third argument passed to conditionalize_nonjump.
6915 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
6917 * config/aarch64/aarch64-builtins.c
6918 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
6919 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
6920 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
6921 instead of __builtin_lfloor.
6922 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
6924 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6926 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
6927 (tree_ssa_ifcombine_bb_1): New function.
6928 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
6929 is an empty forwarder block to then_bb or vice versa and then_bb
6930 and else_bb are effectively swapped.
6932 2014-03-12 Christian Bruel <christian.bruel@st.com>
6935 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
6936 REG_CFA_DEF_CFA note.
6937 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
6938 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
6940 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
6942 PR tree-optimization/60454
6943 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
6945 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6947 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
6948 Do not define target_cpu_default2 to generic.
6949 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
6950 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
6951 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
6953 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6954 Marc Glisse <marc.glisse@inria.fr>
6956 PR tree-optimization/60502
6957 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
6958 instead of build_low_bits_mask.
6960 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6963 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
6964 if there are multiple uses, but op doesn't live on E edge.
6965 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
6966 clobber stmts before __builtin_unreachable.
6968 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
6970 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
6971 hard_frame_pointer_rtx.
6972 * cse.c (cse_insn): Remove volatile check.
6973 * cselib.c (cselib_process_insn): Likewise.
6974 * dse.c (scan_insn): Likewise.
6976 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6978 * config/arc/arc.c (conditionalize_nonjump): New function,
6980 (arc_ifcvt): ... this.
6981 (arc_predicate_delay_insns): Use it.
6983 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6985 * config/arc/predicates.md (extend_operand): During/after reload,
6986 allow const_int_operand.
6987 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
6988 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
6989 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
6991 (umulsi3_highpart_i): Likewise.
6993 2014-03-11 Richard Biener <rguenther@suse.de>
6995 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
6996 Add asserts to guard possible wrong-code bugs.
6998 2014-03-11 Richard Biener <rguenther@suse.de>
7000 PR tree-optimization/60429
7001 PR tree-optimization/60485
7002 * tree-ssa-structalias.c (set_union_with_increment): Properly
7003 take into account all fields that overlap the shifted vars.
7004 (do_sd_constraint): Likewise.
7005 (do_ds_constraint): Likewise.
7006 (get_constraint_for_ptr_offset): Likewise.
7008 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
7010 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
7011 (nios2_compute_frame_layout):
7012 Add calculation of cfun->machine->fp_save_offset.
7013 (nios2_expand_prologue): Correct setting of frame pointer register
7015 (nios2_expand_epilogue): Update recovery of stack pointer from
7016 frame pointer accordingly.
7017 (nios2_initial_elimination_offset): Update calculation of offset
7018 for eliminating to HARD_FRAME_POINTER_REGNUM.
7020 2014-03-10 Jakub Jelinek <jakub@redhat.com>
7023 * ipa.c (symtab_remove_unreachable_nodes): Don't call
7024 cgraph_get_create_node on VAR_DECLs.
7026 2014-03-10 Richard Biener <rguenther@suse.de>
7029 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
7031 2014-03-08 Douglas B Rupp <rupp@gnat.com>
7033 * config/vms/vms.opt (vms_float_format): New variable.
7035 2014-03-08 Tobias Burnus <burnus@net-b.de>
7037 * doc/invoke.texi (-fcilkplus): Update implementation status.
7039 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
7040 Richard Biener <rguenther@suse.de>
7042 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
7043 consistently accross all TUs.
7044 (run_gcc): Enable -fshort-double automatically at link at link-time
7045 and disallow override.
7047 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
7050 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
7051 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
7052 if they can't be used.
7054 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7056 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
7057 for Solaris 11/x86 ld.
7058 * configure: Regenerate.
7060 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7062 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
7063 (LIB_TLS_SPEC): Save as ld_tls_libs.
7064 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
7065 (HAVE_AS_IX86_TLSLDM): New test.
7066 * configure, config.in: Regenerate.
7067 * config/i386/i386.c (legitimize_tls_address): Fall back to
7068 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
7069 cannot support TLS_MODEL_LOCAL_DYNAMIC.
7070 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
7071 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
7073 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
7075 * common.opt (fira-loop-pressure): Mark as optimization.
7077 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
7079 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
7080 an OpenMP mappable type.
7082 2014-03-06 Matthias Klose <doko@ubuntu.com>
7084 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
7085 MULTILIB_OSDIRNAMES is not defined.
7087 2014-03-06 Jakub Jelinek <jakub@redhat.com>
7088 Meador Inge <meadori@codesourcery.com>
7091 * config/arm/arm.c (arm_tls_symbol_p): Remove.
7092 (arm_legitimize_address): Call legitimize_tls_address for any
7093 arm_tls_referenced_p expression, handle constant addend. Call it
7094 before testing for !TARGET_ARM.
7095 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
7097 2014-03-06 Richard Biener <rguenther@suse.de>
7103 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
7105 * tree-streamer.c (record_common_node): Assert we don't record
7106 nodes with type double.
7107 (preload_common_node): Skip type double, complex double and double
7108 pointer since it is now frontend dependent due to fshort-double option.
7110 2014-03-06 Richard Biener <rguenther@suse.de>
7112 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
7113 or -fno-lto is specified and the linker has full plugin support.
7114 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
7115 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
7116 * lto-wrapper.c (merge_and_complain): Merge compile-time
7117 optimization levels.
7118 (run_gcc): And pass it through to the link options.
7120 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
7124 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7126 * cselib.c (remove_useless_values): Skip to avoid quadratic
7127 behavior if the condition moved from...
7128 (cselib_process_insn): ... here holds.
7130 2014-03-05 Jakub Jelinek <jakub@redhat.com>
7133 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
7134 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
7137 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
7138 (TM_H): Add x86-tune.def.
7140 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7142 * config/aarch64/aarch64.c (generic_tunings):
7143 Use cortexa57_extra_costs.
7145 2014-03-05 Jakub Jelinek <jakub@redhat.com>
7148 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
7149 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
7150 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
7153 2014-03-04 Heiher <r@hev.cc>
7155 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
7156 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
7158 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
7160 * config/i386/predicates.md (const2356_operand): Change to ...
7161 (const2367_operand): ... this.
7162 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
7164 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7165 (*avx512pf_scatterpf<mode>sf): Ditto.
7166 (avx512pf_scatterpf<mode>df): Ditto.
7167 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7168 (*avx512pf_scatterpf<mode>df): Ditto.
7169 * config/i386/i386.c (ix86_expand_builtin): Update
7170 incorrect hint operand error message.
7172 2014-03-04 Richard Biener <rguenther@suse.de>
7174 * lto-section-in.c (lto_get_section_data): Fix const cast.
7176 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
7178 * tree-streamer.c (record_common_node): Assert we don't record
7179 nodes with type double.
7180 (preload_common_node): Skip type double, complex double and double
7181 pointer since it is now frontend dependent due to fshort-double option.
7183 2014-03-04 Richard Biener <rguenther@suse.de>
7186 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
7187 (lto_input_toplevel_asms): Likewise.
7188 * lto-section-in.c (lto_get_section_data): Instead do it here
7191 2014-03-04 Richard Biener <rguenther@suse.de>
7193 PR tree-optimization/60382
7194 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
7195 dead PHIs a reduction.
7197 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
7199 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
7201 (_mm_prefetch): Move out of GCC target("sse") pragma.
7202 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
7203 GCC target("prfchw") pragma.
7204 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
7206 * config/i386/i386.c (ix86_option_override_internal): Enable
7207 -mprfchw with -mprefetchwt1.
7209 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
7211 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
7214 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
7216 * opts.h (CL_PCH_IGNORE): Define.
7217 * targhooks.c (option_affects_pch_p):
7218 Return false for options that have CL_PCH_IGNORE set.
7219 * opt-functions.awk: Process PchIgnore.
7220 * doc/options.texi: Document PchIgnore.
7222 * config/arc/arc.opt (misize): Add PchIgnore property.
7224 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7226 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
7227 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
7228 constraint on constants to permit them being loaded into
7229 GENERAL_REGS or BASE_REGS.
7231 2014-03-03 Nick Clifton <nickc@redhat.com>
7233 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
7234 anti-cacnonical alternatives.
7235 (negandhi3_real): New pattern.
7236 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
7238 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7240 * config/avr/avr-mcus.def: Remove atxmega16x1.
7241 * config/avr/avr-tables.opt: Regenerate.
7242 * config/avr/t-multilib: Regenerate.
7243 * doc/avr-mmcu.texi: Regenerate.
7245 2014-03-03 Tobias Grosser <tobias@grosser.es>
7246 Mircea Namolaru <mircea.namolaru@inria.fr>
7248 PR tree-optimization/58028
7249 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
7252 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7254 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
7255 not handled by recognizers.
7257 2014-03-03 Jakub Jelinek <jakub@redhat.com>
7260 * function.c (expand_function_end): Don't emit
7261 clobber_return_register sequence if clobber_after is a BARRIER.
7262 * cfgexpand.c (construct_exit_block): Append instructions before
7263 return_label to prev_bb.
7265 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7267 * config/rs6000/constraints.md: Document reserved use of "wc".
7269 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
7272 * ipa.c (function_and_variable_visibility): When dissolving comdat
7273 group, also set all symbols to local.
7275 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
7280 2013-12-14 Jan Hubicka <jh@suse.cz>
7282 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
7284 2014-03-02 Jon Beniston <jon@beniston.com>
7290 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
7291 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
7292 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
7293 (simple_return, *simple_return): New patterns
7294 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
7295 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
7297 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
7299 * dwarf2out.c (gen_subprogram_die): Tidy.
7301 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
7304 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
7305 (*mov_t_msb_neg_negc): ... this new insn.
7307 2014-02-28 Jason Merrill <jason@redhat.com>
7310 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
7313 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
7316 * dwarf2out.c (decltype_auto_die): New static.
7317 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
7318 (gen_type_die_with_usage): Handle 'decltype(auto)'.
7319 (is_cxx_auto): Likewise.
7321 2014-02-28 Ian Bolton <ian.bolton@arm.com>
7323 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
7324 we are not using general regs only.
7326 2014-02-28 Richard Biener <rguenther@suse.de>
7329 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
7330 previous fix and only allow to remove trivial pre-headers
7331 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
7332 (remove_forwarder_block): Properly update the latch of a loop.
7334 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7337 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
7338 (cselib_preserved_hash_table): New.
7339 (preserve_constants_and_equivs): Move preserved vals to it.
7340 (cselib_find_slot): Look it up first.
7341 (cselib_init): Initialize it.
7342 (cselib_finish): Release it.
7343 (dump_cselib_table): Dump it.
7345 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7348 * cselib.c (remove_useless_values): Skip to avoid quadratic
7349 behavior if the condition moved from...
7350 (cselib_process_insn): ... here holds.
7352 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7355 * var-tracking.c (vt_initialize): Apply the same condition to
7356 preserve the CFA base value.
7358 2014-02-28 Joey Ye <joey.ye@arm.com>
7361 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
7362 if reload in progress or completed.
7364 2014-02-28 Tobias Burnus <burnus@net-b.de>
7367 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
7370 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
7372 * doc/tm.texi.in (Condition Code Status): Update documention for
7373 relative locations of cc0-setter and cc0-user.
7375 2014-02-27 Jeff Law <law@redhat.com>
7377 PR rtl-optimization/52714
7378 * combine.c (try_combine): When splitting an unrecognized PARALLEL
7379 into two independent simple sets, if I3 is a jump, ensure the
7380 pattern we place into I3 is a (set (pc) ...).
7382 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
7383 Jeff Law <law@redhat.com>
7385 PR rtl-optimization/49847
7386 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
7387 are in different blocks.
7388 * doc/tm.texi (Condition Code Status): Update documention for
7389 relative locations of cc0-setter and cc0-user.
7391 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
7394 * lra.c (lra_emit_add): Check SUBREG too.
7396 2014-02-27 Andreas Schwab <schwab@suse.de>
7398 * config/m68k/m68k.c (m68k_option_override): Disable
7399 -flive-range-shrinkage for classic m68k.
7400 (m68k_override_options_after_change): Likewise.
7402 2014-02-27 Marek Polacek <polacek@redhat.com>
7405 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
7406 -Wmaybe-uninitialized.
7408 2014-02-27 Alan Modra <amodra@gmail.com>
7411 * reload1.c (emit_input_reload_insns): When reload_override_in,
7412 set old to rl->in_reg when rl->in_reg is a subreg.
7414 2014-02-26 Richard Biener <rguenther@suse.de>
7417 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
7419 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
7421 * common/config/i386/predicates.md (const1256_operand): Remove.
7422 (const2356_operand): New.
7423 (const_1_to_2_operand): Remove.
7424 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7425 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7426 (*avx512pf_gatherpf<mode>sf): Ditto.
7427 (avx512pf_gatherpf<mode>df): Ditto.
7428 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7429 (*avx512pf_gatherpf<mode>df): Ditto.
7430 (avx512pf_scatterpf<mode>sf): Ditto.
7431 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7432 (*avx512pf_scatterpf<mode>sf): Ditto.
7433 (avx512pf_scatterpf<mode>df): Ditto.
7434 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7435 (*avx512pf_scatterpf<mode>df): Ditto.
7436 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
7438 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
7440 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
7441 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
7442 (_mm512_mask_testn_epi64_mask): Move to ...
7443 * config/i386/avx512cdintrin.h: Here.
7444 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
7445 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
7446 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
7447 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
7448 TARGET_AVX512F from TARGET_AVX512CD.
7450 2014-02-26 Richard Biener <rguenther@suse.de>
7453 * ipa.c (walk_polymorphic_call_targets): Properly guard
7454 call to inline_update_overall_summary.
7456 2014-02-26 Bin Cheng <bin.cheng@arm.com>
7459 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
7460 and latches only if requested. Fix latch if it is removed.
7461 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
7462 LOOPS_HAVE_PREHEADERS.
7464 2014-02-25 Andrew Pinski <apinski@cavium.com>
7466 * builtins.c (expand_builtin_thread_pointer): Create a new target
7467 when the target is NULL.
7469 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
7471 PR rtl-optimization/60317
7472 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
7473 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
7474 * lra-assigns.c: Include params.h.
7475 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
7476 other reload pseudos considerations.
7478 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7480 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
7481 to use canonical form for nor<mode>3.
7483 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7486 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
7489 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
7491 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
7492 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
7493 (ix86_handle_option): Handle OPT_mprefetchwt1.
7494 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
7495 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7497 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7498 OPTION_MASK_ISA_PREFETCHWT1.
7499 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
7500 (PTA_PREFETCHWT1): New.
7501 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
7502 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
7503 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
7504 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
7505 (*prefetch_avx512pf_<mode>_: Change into ...
7506 (*prefetch_prefetchwt1_<mode>: This.
7507 * config/i386/i386.opt (mprefetchwt1): New.
7508 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
7509 (_mm_prefetch): Handle intent to write.
7510 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
7512 2014-02-25 Richard Biener <rguenther@suse.de>
7515 * emit-rtl.c (mem_attrs_htab): Remove.
7516 (mem_attrs_htab_hash): Likewise.
7517 (mem_attrs_htab_eq): Likewise.
7518 (set_mem_attrs): Always allocate new mem-attrs when something changed.
7519 (init_emit_once): Do not allocate mem_attrs_htab.
7521 2014-02-25 Richard Biener <rguenther@suse.de>
7524 * lto-opts.c (lto_write_options): Output non-explicit conservative
7525 -fwrapv, -fno-trapv and -fno-strict-overflow.
7526 * lto-wrapper.c (merge_and_complain): Handle merging those options.
7527 (run_gcc): And pass them through.
7529 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7531 * sel-sched.c (calculate_new_fences): New parameter ptime.
7532 Calculate it as a maximum over all fence cycles.
7533 (sel_sched_region_2): Adjust the call to calculate_new_fences.
7534 Print the final schedule timing when sched_verbose.
7536 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7538 PR rtl-optimization/60292
7539 * sel-sched.c (fill_vec_av_set): Do not reset target availability
7540 bit fot the fence instruction.
7542 2014-02-24 Alangi Derick <alangiderick@gmail.com>
7544 * calls.h: Fix typo in comment.
7546 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
7548 * config/pa/pa.c (pa_output_move_double): Don't valididate when
7549 adjusting offsetable addresses.
7551 2014-02-24 Guozhi Wei <carrot@google.com>
7553 * sparseset.h (sparseset_pop): Fix the wrong index.
7555 2014-02-24 Walter Lee <walt@tilera.com>
7557 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
7558 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
7560 * common/config/tilegx/tilegx-common.c
7561 (TARGET_DEFAULT_TARGET_FLAGS): Define.
7562 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
7564 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
7565 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
7566 (tilegx_gimplify_va_arg_expr): Handle big endian.
7567 (tilegx_expand_unaligned_load): Ditto.
7568 (tilegx_expand_unaligned_store): Ditto.
7569 (TARGET_RETURN_IN_MSB): New.
7570 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
7571 (TARGET_ENDIAN_DEFAULT): New.
7572 (TARGET_BIG_ENDIAN): Handle big endian.
7573 (BYTES_BIG_ENDIAN): Ditto.
7574 (WORDS_BIG_ENDIAN): Ditto.
7575 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
7578 * config/tilegx/tilegx.md (extv): Handle big endian.
7580 (insn_st<n>): Ditto.
7581 (insn_st<n>_add<bitsuffix>): Ditto.
7582 (insn_stnt<n>): Ditto.
7583 (insn_stnt<n>_add<bitsuffix>):Ditto.
7584 (vec_interleave_highv8qi): Handle big endian.
7585 (vec_interleave_highv8qi_be): New.
7586 (vec_interleave_highv8qi_le): New.
7587 (insn_v1int_h): Handle big endian.
7588 (vec_interleave_lowv8qi): Handle big endian.
7589 (vec_interleave_lowv8qi_be): New.
7590 (vec_interleave_lowv8qi_le): New.
7591 (insn_v1int_l): Handle big endian.
7592 (vec_interleave_highv4hi): Handle big endian.
7593 (vec_interleave_highv4hi_be): New.
7594 (vec_interleave_highv4hi_le): New.
7595 (insn_v2int_h): Handle big endian.
7596 (vec_interleave_lowv4hi): Handle big endian.
7597 (vec_interleave_lowv4hi_be): New.
7598 (vec_interleave_lowv4hi_le): New.
7599 (insn_v2int_l): Handle big endian.
7600 (vec_interleave_highv2si): Handle big endian.
7601 (vec_interleave_highv2si_be): New.
7602 (vec_interleave_highv2si_le): New.
7603 (insn_v4int_h): Handle big endian.
7604 (vec_interleave_lowv2si): Handle big endian.
7605 (vec_interleave_lowv2si_be): New.
7606 (vec_interleave_lowv2si_le): New.
7607 (insn_v4int_l): Handle big endian.
7608 * config/tilegx/tilegx.opt (mbig-endian): New option.
7609 (mlittle-endian): New option.
7610 * doc/install.texi: Document tilegxbe-linux.
7611 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
7613 2014-02-24 Martin Jambor <mjambor@suse.cz>
7616 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
7617 there are no parameter descriptors.
7619 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
7621 PR rtl-optimization/60268
7622 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
7623 initialization to ...
7624 (sched_rgn_init): ... here.
7625 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
7627 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7629 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
7632 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
7634 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
7637 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7639 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
7640 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
7642 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7644 * config/microblaze/predicates.md: Add cmp_op predicate.
7645 * config/microblaze/microblaze.md: Add branch_compare instruction
7646 which uses cmp_op predicate and emits cmp insn before branch.
7647 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
7648 to microblaze_expand_conditional_branch and consolidate logic.
7649 (microblaze_expand_conditional_branch): emit branch_compare
7650 insn instead of handling cmp op separate from branch insn.
7652 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7654 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
7657 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7659 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
7660 define_insn with define_expand and new define_insn
7661 *altivec_lve<VI_char>x_internal.
7662 (altivec_stve<VI_char>x): Replace define_insn with define_expand
7663 and new define_insn *altivec_stve<VI_char>x_internal.
7664 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
7666 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
7668 (altivec_expand_stvex_be): New function.
7670 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
7672 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
7673 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
7674 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
7675 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
7677 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
7680 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
7681 instead of emit_move_insn.
7683 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7685 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
7687 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
7688 gen_altivec_vsumsws.
7690 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7692 * config/rs6000/altivec.md (altivec_lvxl): Rename as
7693 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
7694 (altivec_lvxl_<mode>): New define_expand incorporating
7695 -maltivec=be semantics where needed.
7696 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
7697 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
7698 semantics where needed.
7699 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
7700 (altivec_stvx_<mode>): New define_expand incorporating
7701 -maltivec=be semantics where needed.
7702 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
7703 VM2 iterator instead of V4SI.
7704 (altivec_stvxl_<mode>): New define_expand incorporating
7705 -maltivec=be semantics where needed.
7706 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
7707 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
7708 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
7709 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
7710 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
7711 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
7712 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
7713 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
7714 ALTIVEC_BUILTIN_STVXL.
7715 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
7716 (altivec_expand_stvx_be): Likewise.
7717 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
7718 (altivec_expand_lvx_be): Likewise.
7719 (altivec_expand_stvx_be): Likewise.
7720 (altivec_expand_builtin): Add cases for
7721 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
7722 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
7723 (altivec_init_builtins): Add definitions for
7724 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
7725 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
7727 2014-02-21 Catherine Moore <clm@codesourcery.com>
7729 * doc/invoke.texi (mvirt, mno-virt): Document.
7730 * config/mips/mips.opt (mvirt): New option.
7731 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
7733 2014-02-21 Richard Biener <rguenther@suse.de>
7735 PR tree-optimization/60276
7736 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
7737 (STMT_VINFO_MIN_NEG_DIST): New macro.
7738 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
7739 STMT_VINFO_MIN_NEG_DIST.
7740 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
7741 made for negative dependence distances still hold.
7743 2014-02-21 Richard Biener <rguenther@suse.de>
7746 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
7747 DECL_INITIAL for globals not in the current function context.
7749 2014-02-21 Jakub Jelinek <jakub@redhat.com>
7751 PR tree-optimization/56490
7752 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
7753 * tree-ssa-uninit.c: Include params.h.
7754 (compute_control_dep_chain): Add num_calls argument, return false
7755 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
7756 num_calls to recursive call.
7757 (find_predicates): Change dep_chain into normal array,
7758 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
7759 variable and adjust compute_control_dep_chain caller.
7760 (find_def_preds): Likewise.
7762 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
7764 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
7765 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
7767 2014-02-21 Nick Clifton <nickc@redhat.com>
7769 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
7770 (pushhi1): Likewise.
7771 (popqi1): Add mode to pre_dec.
7774 2014-02-21 Jakub Jelinek <jakub@redhat.com>
7776 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
7777 mode for mask of V8SFmode permutation.
7779 2014-02-20 Richard Henderson <rth@redhat.com>
7782 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
7783 a new pseudo for OLDVAL.
7785 2014-02-20 Jakub Jelinek <jakub@redhat.com>
7788 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
7789 gen_reg_rtx if d->testing_p.
7790 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
7791 if d->testing_p and we will certainly return true.
7792 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
7795 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
7797 * emit-rtl.c (gen_reg_rtx): Assert that
7798 crtl->emit.regno_pointer_align_length is non-zero.
7800 2014-02-20 Richard Henderson <rth@redhat.com>
7803 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
7804 on failure the store back into EXPECT.
7806 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
7807 Sandra Loosemore <sandra@codesourcery.com>
7809 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
7810 * config/nios2/nios2.c (nios2_function_profiler): Add
7811 -fPIC (flag_pic == 2) support.
7812 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
7813 (nios2_large_offset_p): New function.
7814 (nios2_unspec_reloc_p): Move up position, update to use
7815 nios2_large_offset_p.
7816 (nios2_unspec_address): Remove function.
7817 (nios2_unspec_offset): New function.
7818 (nios2_large_got_address): New function.
7819 (nios2_got_address): Add large offset support.
7820 (nios2_legitimize_tls_address): Update usage of removed and new
7822 (nios2_symbol_binds_local_p): New function.
7823 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
7824 (nios2_legitimize_address): Update to use nios2_large_offset_p.
7825 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
7826 (nios2_print_operand): Merge H/L processing, add hiadj/lo
7827 processing for (const (unspec ...)).
7828 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
7830 2014-02-20 Richard Biener <rguenther@suse.de>
7832 * tree-cfg.c (replace_uses_by): Mark altered BBs before
7833 doing the substitution.
7834 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
7836 2014-02-20 Martin Jambor <mjambor@suse.cz>
7839 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
7840 info when checking whether lattices are bottom.
7842 2014-02-20 Richard Biener <rguenther@suse.de>
7845 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
7848 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
7851 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
7852 parameter specifying the scaling.
7853 (inline_call): Update.
7854 (want_inline_recursively): Guard division by zero.
7855 (recursive_inlining): Update.
7856 * ipa-inline.h (clone_inlined_nodes): Update.
7858 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7861 * config/i386/i386.c (classify_argument): Pass structures of size
7862 64 bytes or less in register.
7864 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7865 Kirill Yukhin <kirill.yukhin@intel.com>
7867 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
7868 (_mm_rcp28_round_ss): Ditto.
7869 (_mm_rsqrt28_round_sd): Ditto.
7870 (_mm_rsqrt28_round_ss): Ditto.
7871 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
7872 (_mm_rcp14_round_ss): Ditto.
7873 (_mm_rsqrt14_round_sd): Ditto.
7874 (_mm_rsqrt14_round_ss): Ditto.
7875 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
7876 the first input operand, get rid of match_dup.
7877 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
7879 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7881 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
7882 operand as the first input operand, set type attribute.
7883 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7885 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
7886 operand as the first input operand, set type attribute.
7888 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7890 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
7893 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
7896 * config/i386/i386.c (construct_container): Remove TFmode check
7897 for X86_64_INTEGER_CLASS.
7899 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
7902 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
7903 only when -Wpsabi is enabled.
7905 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
7908 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
7909 passing arrays in registers are the same as for structs, so remove the
7910 special case for them.
7912 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
7914 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
7915 destination type, extract only the valid bits if the source type is not
7916 integral and has a different mode.
7918 2014-02-19 Richard Biener <rguenther@suse.de>
7921 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
7924 2014-02-19 Richard Biener <rguenther@suse.de>
7927 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
7928 (ipa_modify_call_arguments): Emit an argument load explicitely and
7929 preserve virtual SSA form there and for the replacement call.
7930 Do not update SSA form nor free dominance info.
7932 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7934 * ipa.c (function_and_variable_visibility): Also clear WEAK
7935 flag when disolving COMDAT_GROUP.
7937 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7939 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
7940 * ipa-prop.c (ipa_set_jf_known_type): Return early when
7942 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
7943 do more sanity checks.
7944 (detect_type_change): Return true when giving up early.
7945 (compute_complex_assign_jump_func): Fix type parameter of
7946 ipa_set_ancestor_jf.
7947 (compute_complex_ancestor_jump_func): Likewise.
7948 (update_jump_functions_after_inlining): Fix updating of
7950 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
7952 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7954 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
7955 inline clones when edge disappears.
7957 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7960 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
7961 Split 64-bit moves into 2 patterns. Do not allow the use of
7962 direct move for TDmode in little endian, since the decimal value
7963 has little endian bytes within a word, but the 64-bit pieces are
7964 ordered in a big endian fashion, and normal subreg's of TDmode are
7966 (mov<mode>_64bit_dm): Likewise.
7967 (movtd_64bit_nodm): Likewise.
7969 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7971 PR tree-optimization/60174
7972 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
7973 statement of an SSA_NAME that occurs in an abnormal PHI node.
7975 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7978 * final.c (SEEN_BB): Remove.
7979 (SEEN_NOTE, SEEN_EMITTED): Renumber.
7980 (final_scan_insn): Don't force_source_line on second
7981 NOTE_INSN_BASIC_BLOCK.
7983 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
7986 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
7987 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
7988 (type_natural_mode): Warn ABI change when %zmm register is not
7989 available for AVX512F vector value passing.
7991 2014-02-18 Kai Tietz <ktietz@redhat.com>
7994 * config/i386/i386.c (ix86_expand_prologue): Use value in
7995 rax register as displacement when restoring %r10 or %rax.
7996 Fix wrong offset when restoring both registers.
7998 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
8000 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
8001 assertion with conditional return.
8003 2014-02-18 Jakub Jelinek <jakub@redhat.com>
8004 Uros Bizjak <ubizjak@gmail.com>
8007 * config/i386/driver-i386.c (host_detect_local_cpu): If
8008 YMM state is not saved by the OS, also clear has_f16c. Move
8009 CPUID 0x80000001 handling before YMM state saving checking.
8011 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
8013 PR rtl-optimization/58960
8014 * haifa-sched.c (alloc_global_sched_pressure_data): New,
8015 factored out from ...
8016 (sched_init): ... here.
8017 (free_global_sched_pressure_data): New, factored out from ...
8018 (sched_finish): ... here.
8019 * sched-int.h (free_global_sched_pressure_data): Declare.
8020 * sched-rgn.c (nr_regions_initial): New static global.
8021 (haifa_find_rgns): Initialize it.
8022 (schedule_region): Disable sched-pressure for the newly
8025 2014-02-17 Richard Biener <rguenther@suse.de>
8027 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
8028 release SSA defs of pattern stmts.
8030 2014-02-17 Richard Biener <rguenther@suse.de>
8032 * tree-inline.c (expand_call_inline): Release the virtual
8033 operand defined by the call we are about to inline.
8035 2014-02-17 Richard Biener <rguenther@suse.de>
8037 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
8039 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
8040 Ilya Tocar <ilya.tocar@intel.com>
8042 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
8043 arguments order in builtin.
8044 (_mm512_permutexvar_epi64): Ditto.
8045 (_mm512_mask_permutexvar_epi64): Ditto
8046 (_mm512_maskz_permutexvar_epi32): Ditto
8047 (_mm512_permutexvar_epi32): Ditto
8048 (_mm512_mask_permutexvar_epi32): Ditto
8050 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8052 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
8053 (p8_vmrgow): Likewise.
8055 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8057 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
8060 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
8063 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
8064 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
8065 into 64-bit and 32-bit moves. On 64-bit moves, add support for
8066 using direct move instructions on ISA 2.07. Also adjust
8067 instruction length for 64-bit.
8068 (mov<mode>_64bit, TFmode/TDmode): Likewise.
8069 (mov<mode>_32bit, TFmode/TDmode): Likewise.
8071 2014-02-15 Alan Modra <amodra@gmail.com>
8075 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
8076 find_replacement on parts of insn rtl that might be reloaded.
8078 2014-02-15 Richard Biener <rguenther@suse.de>
8080 PR tree-optimization/60183
8081 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
8082 (tree_ssa_phiprop): Calculate and free post-dominators.
8084 2014-02-14 Jeff Law <law@redhat.com>
8086 PR rtl-optimization/60131
8087 * ree.c (get_extended_src_reg): New function.
8088 (combine_reaching_defs): Use it rather than assuming location of REG.
8089 (find_and_remove_re): Verify first operand of extension is
8090 a REG before adding the insns to the copy list.
8092 2014-02-14 Roland McGrath <mcgrathr@google.com>
8094 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
8095 * configure: Regenerated.
8096 * config.in: Regenerated.
8097 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
8098 instead of ASM_SHORT.
8100 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
8101 Richard Earnshaw <rearnsha@arm.com>
8103 PR rtl-optimization/59535
8104 * lra-constraints.c (process_alt_operands): Encourage alternative
8105 when unassigned pseudo class is superset of the alternative class.
8106 (inherit_reload_reg): Don't inherit when optimizing for code size.
8107 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
8108 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
8109 modes not less than 4 for Thumb1.
8111 2014-02-14 Kyle McMartin <kyle@redhat.com>
8114 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
8116 2014-02-14 Richard Biener <rguenther@suse.de>
8118 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
8119 (get_frame_arg): Drop the assert with langhook types_compatible_p.
8120 Do not strip INDIRECT_REFs.
8122 2014-02-14 Richard Biener <rguenther@suse.de>
8125 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
8126 DECL_FUNCTION_SPECIFIC_TARGET.
8127 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
8128 * tree-streamer-out.c (pack_ts_target_option): Remove.
8129 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
8130 (write_ts_function_decl_tree_pointers): Do not stream
8131 DECL_FUNCTION_SPECIFIC_TARGET.
8132 * tree-streamer-in.c (unpack_ts_target_option): Remove.
8133 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
8134 (lto_input_ts_function_decl_tree_pointers): Do not stream
8135 DECL_FUNCTION_SPECIFIC_TARGET.
8137 2014-02-14 Jakub Jelinek <jakub@redhat.com>
8139 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
8140 (get_initial_def_for_induction, vectorizable_induction): Ignore
8141 debug stmts when looking for exit_phi.
8142 (vectorizable_live_operation): Fix up condition.
8144 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8146 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
8147 nreverse() because it changes the content of original tree list.
8149 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8151 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
8152 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
8154 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
8156 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
8157 GNU coding standards.
8159 2014-02-13 Jakub Jelinek <jakub@redhat.com>
8162 * dwarf2out.c (gen_subprogram_die): Don't call
8163 add_calling_convention_attribute if subr_die is old_die.
8165 2014-02-13 Sharad Singhai <singhai@google.com>
8167 * doc/optinfo.texi: Fix order of nodes.
8169 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
8171 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
8172 operands[2], not operands[3].
8174 2014-02-13 Richard Biener <rguenther@suse.de>
8177 * doc/install.texi (ISL): Update recommended version to 0.12.2,
8178 mention the possibility of an in-tree build.
8179 (CLooG): Update recommended version to 0.18.1, mention the
8180 possibility of an in-tree build and clarify that the ISL
8181 bundled with CLooG does not work.
8183 2014-02-13 Jakub Jelinek <jakub@redhat.com>
8186 * expr.c (compress_float_constant): If x is a hard register,
8187 extend into a pseudo and then move to x.
8189 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
8191 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
8192 caused by bad second argument to warning_at() with -mhotpatch and
8193 nested functions (e.g. with gfortran).
8195 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
8197 * opts.c (option_name): Remove "enabled by default" rider.
8199 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
8201 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
8203 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
8204 Uros Bizjak <ubizjak@gmail.com>
8207 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
8208 * configure: Regenerated.
8210 2014-02-12 Richard Biener <rguenther@suse.de>
8212 * vec.c (vec_prefix::calculate_allocation): Move as
8213 inline variant to vec.h.
8214 (vec_prefix::calculate_allocation_1): New out-of-line version.
8215 * vec.h (vec_prefix::calculate_allocation_1): Declare.
8216 (vec_prefix::m_has_auto_buf): Rename to ...
8217 (vec_prefix::m_using_auto_storage): ... this.
8218 (vec_prefix::calculate_allocation): Inline the easy cases
8219 and dispatch to calculate_allocation_1 which doesn't need the
8221 (va_heap::reserve): Use gcc_checking_assert.
8222 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
8223 m_using_auto_storage.
8224 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
8226 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
8227 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
8228 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
8230 2014-02-12 Richard Biener <rguenther@suse.de>
8232 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
8233 when we found a dependence.
8235 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
8237 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
8239 (maybe_fold_stmt): ... into this new function.
8240 * omp-low.c (lower_omp): Update comment.
8242 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
8245 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
8248 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
8250 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
8251 identifiers in comments.
8252 (cortexa53_extra_costs): Likewise.
8253 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
8254 (cortexa7_extra_costs): Likewise.
8255 (cortexa12_extra_costs): Likewise.
8256 (cortexa15_extra_costs): Likewise.
8257 (v7m_extra_costs): Likewise.
8259 2014-02-12 Richard Biener <rguenther@suse.de>
8262 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
8263 of posix_memalign being successful.
8264 (lower_stmt): Restrict lowering of posix_memalign to when
8265 -ftree-bit-ccp is enabled.
8267 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
8269 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
8271 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
8273 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
8275 PR rtl-optimization/60116
8276 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
8277 other_insn once the combination has been validated.
8279 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
8282 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
8284 * ipa-devirt.c: Include demangle.h
8285 (odr_violation_reported): New static variable.
8286 (add_type_duplicate): Update odr_violations.
8287 (maybe_record_node): Add completep parameter; update it.
8288 (record_target_from_binfo): Add COMPLETEP parameter;
8289 update it as needed.
8290 (possible_polymorphic_call_targets_1): Likewise.
8291 (struct polymorphic_call_target_d): Add nonconstruction_targets;
8292 rename FINAL to COMPLETE.
8293 (record_targets_from_bases): Sanity check we found the binfo;
8294 fix COMPLETEP updating.
8295 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
8296 parameter, fix computing of COMPLETEP.
8297 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
8298 at LTO time do demangling.
8299 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
8300 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
8302 (gimple_get_virt_method_for_binfo): Likewise.
8303 * gimple-fold.h (gimple_get_virt_method_for_binfo,
8304 gimple_get_virt_method_for_vtable): Update prototypes.
8306 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
8309 * genautomata.c (add_presence_absence): Fix typo with
8310 {final_}presence_list.
8312 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
8315 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
8316 for VSX/Altivec vectors that land in GPR registers.
8318 2014-02-11 Richard Henderson <rth@redhat.com>
8319 Jakub Jelinek <jakub@redhat.com>
8322 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
8323 around drhs if type conversion to lacc->type is not useless.
8325 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8327 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
8329 (cortex-a57.cortex-a53): Likewise.
8330 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
8332 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8334 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
8337 2014-02-11 Renlin Li <Renlin.Li@arm.com>
8339 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
8340 add_options_for_arm_vfp3.
8342 2014-02-11 Jeff Law <law@redhat.com>
8345 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
8346 object with an undesirable mode.
8348 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8351 * config/i386/sol2-9.h: New file.
8352 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
8353 *-*-solaris2.9*): Use it.
8355 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
8357 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
8358 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
8360 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
8362 * config/microblaze/microblaze.c: Extend mcpu version format
8364 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
8366 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
8368 2014-02-10 Richard Henderson <rth@redhat.com>
8371 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
8372 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
8373 ms-abi vs -mno-accumulate-outgoing-args.
8374 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
8375 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
8378 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
8381 * cfgexpand.c (expand_asm_operands): Attach source location to
8382 ASM_INPUT rtx objects.
8383 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
8385 2014-02-10 Nick Clifton <nickc@redhat.com>
8387 * config/mn10300/mn10300.c (popcount): New function.
8388 (mn10300_expand_prologue): Include saved registers in stack usage
8391 2014-02-10 Jeff Law <law@redhat.com>
8394 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
8395 when changing the SET_DEST of a prior insn to avoid an input reload.
8397 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8399 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
8400 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
8401 -mcall-openbsd, or -mcall-linux.
8402 (CC1_ENDIAN_BIG_SPEC): Remove.
8403 (CC1_ENDIAN_LITTLE_SPEC): Remove.
8404 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
8405 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
8406 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
8407 and %cc1_endian_default.
8408 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
8410 2014-02-10 Richard Biener <rguenther@suse.de>
8412 PR tree-optimization/60115
8413 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
8414 MEM_REF handling. Properly verify that the accesses are not
8415 out of the objects bound.
8417 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8419 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
8422 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
8424 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
8425 proper constants and fix formatting.
8426 (possible_polymorphic_call_targets): Fix formatting.
8428 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
8429 Ilya Tocar <ilya.tocar@intel.com>
8431 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
8432 (_mm512_loadu_epi32): Renamed into...
8433 (_mm512_loadu_si512): This.
8434 (_mm512_storeu_epi32): Renamed into...
8435 (_mm512_storeu_si512): This.
8436 (_mm512_maskz_ceil_ps): Removed.
8437 (_mm512_maskz_ceil_pd): Ditto.
8438 (_mm512_maskz_floor_ps): Ditto.
8439 (_mm512_maskz_floor_pd): Ditto.
8440 (_mm512_floor_round_ps): Ditto.
8441 (_mm512_floor_round_pd): Ditto.
8442 (_mm512_ceil_round_ps): Ditto.
8443 (_mm512_ceil_round_pd): Ditto.
8444 (_mm512_mask_floor_round_ps): Ditto.
8445 (_mm512_mask_floor_round_pd): Ditto.
8446 (_mm512_mask_ceil_round_ps): Ditto.
8447 (_mm512_mask_ceil_round_pd): Ditto.
8448 (_mm512_maskz_floor_round_ps): Ditto.
8449 (_mm512_maskz_floor_round_pd): Ditto.
8450 (_mm512_maskz_ceil_round_ps): Ditto.
8451 (_mm512_maskz_ceil_round_pd): Ditto.
8452 (_mm512_expand_pd): Ditto.
8453 (_mm512_expand_ps): Ditto.
8454 * config/i386/i386.c (ix86_builtins): Remove
8455 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
8456 (bdesc_args): Ditto.
8457 * config/i386/predicates.md (const1256_operand): New.
8458 (const_1_to_2_operand): Ditto.
8459 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
8460 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8461 (*avx512pf_gatherpf<mode>sf): Ditto.
8462 (avx512pf_gatherpf<mode>df): Ditto.
8463 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8464 (*avx512pf_gatherpf<mode>df): Ditto.
8465 (avx512pf_scatterpf<mode>sf): Ditto.
8466 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8467 (*avx512pf_scatterpf<mode>sf): Ditto.
8468 (avx512pf_scatterpf<mode>df): Ditto.
8469 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8470 (*avx512pf_scatterpf<mode>df): Ditto.
8471 (avx512f_expand<mode>): Removed.
8472 (<shift_insn><mode>3<mask_name>): Change predicate type.
8474 2014-02-08 Jakub Jelinek <jakub@redhat.com>
8476 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
8477 not at the end of datarefs vector use ordered_remove to avoid
8478 reordering datarefs vector.
8481 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
8482 mark local addressable non-static vars as GOVD_PRIVATE
8483 instead of GOVD_LOCAL.
8484 * omp-low.c (lower_omp_for): Move gimple_bind_vars
8485 and BLOCK_VARS of gimple_bind_block to new_stmt rather
8489 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
8490 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
8491 assume_aligned or alloc_align attributes.
8492 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
8493 arguments. Handle also assume_aligned and alloc_align attributes.
8494 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
8495 calls to functions with assume_aligned or alloc_align attributes.
8496 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
8498 2014-02-08 Terry Guo <terry.guo@arm.com>
8500 * doc/invoke.texi: Document ARM -march=armv7e-m.
8502 2014-02-08 Jakub Jelinek <jakub@redhat.com>
8504 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
8505 flag on __cilkrts_rethrow builtin.
8508 * ipa-cp.c (determine_versionability): Fail at -O0
8509 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
8510 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
8513 2014-02-04 Jakub Jelinek <jakub@redhat.com>
8516 * tree-inline.c (copy_forbidden): Fail for
8517 __attribute__((optimize (0))) functions.
8519 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
8521 * varpool.c: Include pointer-set.h.
8522 (varpool_remove_unreferenced_decls): Variables in other partitions
8523 will not be output; be however careful to not lose information
8526 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
8528 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
8529 lookup in the vtable constructor.
8531 2014-02-07 Jeff Law <law@redhat.com>
8534 * config/m68k/m68k.md (ashldi_extsi): Turn into a
8535 define_insn_and_split.
8537 * ipa-inline.c (inline_small_functions): Fix typos.
8539 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8541 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
8542 (s390_can_use_return_insn): Declare.
8543 * config/s390/s390.h (EPILOGUE_USES): Define.
8544 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
8546 (s390_chunkify_start): Handle return JUMP_LABELs.
8547 (s390_early_mach): Emit a main_pool instruction on the entry edge.
8548 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
8549 (s390_can_use_return_insn): New functions.
8550 (s390_fix_long_loop_prediction): Handle conditional returns.
8551 (TARGET_SET_UP_BY_PROLOGUE): Define.
8552 * config/s390/s390.md (ANY_RETURN): New code iterator.
8553 (*creturn, *csimple_return, return, simple_return): New patterns.
8555 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8557 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
8558 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
8559 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
8560 REG_CFA_RESTORE list when deciding not to restore a register.
8562 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8564 * config/s390/s390.c: Include tree-pass.h and context.h.
8565 (s390_early_mach): New function, split out from...
8566 (s390_emit_prologue): ...here.
8567 (pass_data_s390_early_mach): New pass structure.
8568 (pass_s390_early_mach): New class.
8569 (s390_option_override): Create and register early_mach pass.
8570 Move to end of file.
8572 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8574 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
8575 to match for the exit block.
8577 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8579 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
8580 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
8581 Reject misaligned operands.
8583 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8585 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
8587 2014-02-07 Richard Biener <rguenther@suse.de>
8590 * gimple-low.c (lower_builtin_posix_memalign): New function.
8591 (lower_stmt): Call it to lower posix_memalign in a way
8592 to make alignment info accessible.
8594 2014-02-07 Jakub Jelinek <jakub@redhat.com>
8597 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
8598 __builtin_setjmp_receiver.
8600 2014-02-07 Richard Biener <rguenther@suse.de>
8603 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
8604 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
8605 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
8606 Handle BUILT_IN_POSIX_MEMALIGN.
8607 (find_func_clobbers): Likewise.
8608 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
8609 (call_may_clobber_ref_p_1): Likewise.
8611 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8614 * ipa-devirt.c (record_target_from_binfo): Remove overactive
8617 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8620 * lto-cgraph.c (lto_output_node): Use
8621 symtab_get_symbol_partitioning_class.
8622 (lto_output_varpool_node): likewise.
8623 (symtab_get_symbol_partitioning_class): Move here from
8625 * cgraph.h (symbol_partitioning_class): Likewise.
8626 (symtab_get_symbol_partitioning_class): Declare.
8628 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8630 * ggc.h (ggc_internal_cleared_alloc): New macro.
8631 * vec.h (vec_safe_copy): Handle memory stats.
8632 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
8633 * target-globals.c (save_target_globals): Likewise.
8635 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8638 * expr.c (emit_move_resolve_push): Export; be bit more selective
8639 on when to clear alias set.
8640 * expr.h (emit_move_resolve_push): Declare.
8641 * function.h (struct function): Add tail_call_marked.
8642 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
8643 * config/i386/i386-protos.h (ix86_expand_push): Remove.
8644 * config/i386/i386.md (TImode move expander): De not call
8646 (FP push expanders): Preserve memory attributes.
8647 * config/i386/sse.md (push<mode>1): Remove.
8648 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
8649 (ix86_expand_push): Remove.
8650 * config/i386/mmx.md (push<mode>1): Remove.
8652 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8654 PR rtl-optimization/60030
8655 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
8656 lopart with paradoxical subreg before shifting it up by hprec.
8658 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8660 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
8661 Remove extra newline at end of file.
8662 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
8663 (arm_issue_rate): Handle cortexa57.
8664 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
8665 (cortex-a57.cortex-a53): Likewise.
8667 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8670 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
8671 don't record in REG_FRAME_RELATED_EXPR registers not set in that
8673 (arm_expand_prologue): Adjust all callers.
8674 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
8675 info, registers also at the lowest numbered registers side. Use
8676 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
8680 * var-tracking.c (adjust_mems): Before adding a SET to
8681 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
8683 2014-02-06 Alan Modra <amodra@gmail.com>
8686 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
8687 change SDmode to DDmode when lra_in_progress.
8689 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8692 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
8693 free_data_ref on the dr first, and before goto again also set dr
8694 to the next dr. For simd_lane_access, free old datarefs[i] before
8695 overwriting it. For get_vectype_for_scalar_type failure, don't
8696 free_data_ref if simd_lane_access.
8698 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
8701 * tree.h (opts_for_fn): New inline function.
8702 (opt_for_fn): Define.
8703 * config/i386/i386.c (ix86_function_regparm): Use
8704 opt_for_fn (decl, optimize) instead of optimize.
8706 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
8708 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
8709 for SYMBOL_REF in large memory model.
8711 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8713 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
8715 (cortex-a57): Likewise.
8716 (cortex-a57.cortex-a53): Likewise.
8718 2014-02-06 Yury Gribov <y.gribov@samsung.com>
8719 Kugan Vivekanandarajah <kuganv@linaro.org>
8721 * config/arm/arm.c (arm_vector_alignment_reachable): Check
8723 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
8725 2014-02-06 Richard Biener <rguenther@suse.de>
8727 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
8728 set_loop_copy and initialize_original_copy_tables.
8730 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
8732 * config/aarch64/aarch64-simd.md
8733 (aarch64_ashr_simddi): Change QI to SI.
8735 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
8736 Jakub Jelinek <jakub@redhat.com>
8739 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
8742 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8744 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
8745 CODE_FOR_altivec_vpku[hw]um to
8746 CODE_FOR_altivec_vpku[hw]um_direct.
8747 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
8748 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
8749 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
8750 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
8752 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8754 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
8755 generation for -maltivec=be.
8756 (altivec_vsumsws): Simplify redundant test.
8758 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8760 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
8761 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
8762 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
8763 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
8764 gen_altivec_vpkuwum.
8765 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
8767 (altivec_vpks<VI_char>ss): Likewise.
8768 (altivec_vpks<VI_char>us): Likewise.
8769 (altivec_vpku<VI_char>us): Likewise.
8770 (altivec_vpku<VI_char>um): Likewise.
8771 (altivec_vpku<VI_char>um_direct): New (copy of
8772 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
8774 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
8775 target is little endian and -maltivec=be is not specified.
8776 (*altivec_vupkhs<VU_char>_direct): New (copy of
8777 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
8778 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
8779 target is little endian and -maltivec=be is not specified.
8780 (*altivec_vupkls<VU_char>_direct): New (copy of
8781 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
8782 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
8783 little endian and -maltivec=be is not specified.
8784 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
8785 little endian and -maltivec=be is not specified.
8787 2014-02-05 Richard Henderson <rth@redhat.com>
8790 * combine-stack-adj.c: Revert r206943.
8791 * sched-int.h (struct deps_desc): Add last_args_size.
8792 * sched-deps.c (init_deps): Initialize it.
8793 (sched_analyze_insn): Add OUTPUT dependencies between insns that
8794 contain REG_ARGS_SIZE notes.
8796 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
8798 * lto-cgraph.c (asm_nodes_output): Make global.
8799 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
8800 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
8801 (driver_handle_option): Handle OPT_fwpa.
8803 2014-02-05 Jakub Jelinek <jakub@redhat.com>
8806 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
8807 a comment typo and formatting issue. If odr_hash hasn't been
8808 created, return vNULL and set *completep to false.
8811 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
8812 bb with no successors.
8814 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
8817 * doc/invoke.texi (-march): Clarify documentation for ARM.
8821 2014-02-05 Richard Biener <rguenther@suse.de>
8823 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
8824 when not vectorizing because of too many alias checks.
8825 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
8826 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
8828 2014-02-05 Nick Clifton <nickc@redhat.com>
8830 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
8831 accept extended registers in any mode when compiling for the MN10300.
8833 2014-02-05 Yury Gribov <y.gribov@samsung.com>
8835 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
8836 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
8837 sanitization attributes.
8838 (can_inline_edge_p): Likewise.
8839 (sanitize_attrs_match_for_inline_p): New function.
8841 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8843 * ipa-prop.c (detect_type_change): Shor circuit testing of
8844 type changes on THIS pointer.
8846 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
8849 * config/pa/pa.c (legitimize_tls_address): Return original address
8850 if not passed a SYMBOL_REF rtx.
8851 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
8853 (pa_emit_move_sequence): Simplify TLS source operands.
8854 (pa_legitimate_constant_p): Reject all TLS constants.
8855 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
8856 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
8858 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8860 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
8861 groups when we know they are controlled by LTO.
8862 * varasm.c (default_binds_local_p_1): If object is in other partition,
8863 it will be resolved locally.
8865 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8867 * config/host-linux.c (linux_gt_pch_use_address): Don't
8868 use SSIZE_MAX because it is not always defined.
8870 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
8873 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
8874 threshold for pseudo splitting.
8875 (update_ebb_live_info): Process call argument hard registers and
8876 hard registers from insn definition too.
8877 (max_small_class_regs_num): New constant.
8878 (inherit_in_ebb): Update live hard regs through EBBs. Update
8879 reloads_num only for small register classes. Don't split for
8882 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
8885 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
8888 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8890 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
8893 2014-02-04 Marek Polacek <polacek@redhat.com>
8895 * gdbinit.in (pel): Define.
8897 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8899 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
8902 2014-02-04 Richard Biener <rguenther@suse.de>
8905 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
8906 in function context local.
8907 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
8908 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
8909 similar to LTO_imported_decl_ref.
8911 2014-02-04 Jakub Jelinek <jakub@redhat.com>
8913 PR tree-optimization/60002
8914 * cgraphclones.c (build_function_decl_skip_args): Clear
8917 PR tree-optimization/60023
8918 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
8919 false to gsi_replace.
8920 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
8921 has been in some EH region and vec_stmt could throw, add
8922 vec_stmt into the same EH region.
8923 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
8924 has no lhs, ignore it.
8925 * internal-fn.c (expand_MASK_LOAD): Likewise.
8928 * tree-inline.c (copy_forbidden): Fail for
8929 __attribute__((optimize (0))) functions.
8932 * omp-low.c (simd_clone_struct_copy): If from->inbranch
8933 is set, copy one less argument.
8934 (expand_simd_clones): Don't subtract clone_info->inbranch
8935 from simd_clone_struct_alloc argument.
8937 PR rtl-optimization/57915
8938 * recog.c (simplify_while_replacing): If all unary/binary/relational
8939 operation arguments are constant, attempt to simplify those.
8942 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
8943 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
8945 2014-02-04 Richard Biener <rguenther@suse.de>
8947 PR tree-optimization/60012
8948 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
8949 TBAA disambiguation to all DDRs.
8951 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8954 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
8955 (LINK_SPEC): Use it for -shared, -shared-libgcc.
8957 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8960 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
8962 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8964 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
8965 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
8967 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8970 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
8971 to figure out targets of polymorphic calls with known decl.
8972 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8973 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
8974 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
8975 (get_polymorphic_call_info): ... here.
8976 (get_polymorphic_call_info_from_invariant): New function.
8978 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8980 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
8981 lookup via vtable pointer; check for type consistency
8982 and turn inconsitent facts into UNREACHABLE.
8983 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8984 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
8985 type inconsistent querries; return UNREACHABLE instead.
8987 2014-02-03 Richard Henderson <rth@twiddle.net>
8990 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
8991 already processed this node.
8992 (normalize_one_pred_1): Pass along mark_set.
8993 (normalize_one_pred): Create and destroy a pointer_set_t.
8994 (normalize_one_pred_chain): Likewise.
8996 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
8998 PR gcov-profile/58602
8999 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
9001 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9004 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
9005 -fno-devirtualize; try to devirtualize by the knowledge of
9006 virtual table pointer given by aggregate propagation.
9007 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9008 (ipa_print_node_jump_functions): Dump also offset that
9009 is relevant for polymorphic calls.
9010 (determine_known_aggregate_parts): Add arg_type parameter; use it
9011 instead of determining the type from pointer type.
9012 (ipa_compute_jump_functions_for_edge): Update call of
9013 determine_known_aggregate_parts.
9014 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
9015 (gimple_get_virt_method_for_binfo): ... here; simplify using
9016 vtable_pointer_value_to_vtable.
9017 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
9018 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
9019 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
9020 (vtable_pointer_value_to_vtable): Break out from ...; handle also
9022 (vtable_pointer_value_to_binfo): ... here.
9023 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
9025 2014-02-03 Teresa Johnson <tejohnson@google.com>
9027 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
9028 redef of outer loop index variable.
9030 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
9034 * doc/extend.texi (Function Attributes): Typo.
9036 2014-02-03 Cong Hou <congh@google.com>
9038 PR tree-optimization/60000
9039 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
9040 if the vectorized statement is a store. A store statement can only
9041 appear at the end of pattern statements.
9043 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
9045 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
9046 (ix86_option_override_internal): Default long double to 64-bit for
9047 32-bit Bionic and to 128-bit for 64-bit Bionic.
9049 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
9050 TARGET_LONG_DOUBLE_128 is true.
9051 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
9053 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
9054 (mlong-double-64): Negate -mlong-double-128.
9055 (mlong-double-128): New option.
9057 * config/i386/i386-c.c (ix86_target_macros): Define
9058 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
9060 * doc/invoke.texi: Document -mlong-double-128.
9062 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
9064 PR rtl-optimization/60024
9065 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
9067 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
9069 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
9071 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
9073 PR rtl-optimization/57662
9074 * sel-sched.c (code_motion_path_driver): Do not mark already not
9075 existing blocks in the visiting bitmap.
9077 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
9079 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
9080 on the insn being emitted.
9082 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
9083 Will Deacon <will.deacon@arm.com>
9085 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
9087 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9089 * config/arm/arm-tables.opt: Regenerate.
9091 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9093 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
9094 for vector types other than V16QImode.
9095 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
9096 define_expand, and call altivec_expand_vec_perm_le when producing
9097 code with little endian element order.
9098 (*altivec_vperm_<mode>_internal): New insn having previous
9099 behavior of altivec_vperm_<mode>.
9100 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
9101 altivec_expand_vec_perm_le when producing code with little endian
9103 (*altivec_vperm_<mode>_uns_internal): New insn having previous
9104 behavior of altivec_vperm_<mode>_uns.
9106 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9108 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
9109 (altivec_vsumsws): Add handling for -maltivec=be with a little
9111 (altivec_vsumsws_direct): New.
9112 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
9113 gen_altivec_vsumsws.
9115 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
9117 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
9118 vtable_pointer_value_to_binfo): New functions.
9119 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
9120 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
9122 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
9124 * config/nios2/nios2.md (load_got_register): Initialize GOT
9125 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
9126 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
9128 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
9130 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
9131 preserverd by passthrough, do not propagate the type.
9133 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9135 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
9136 (mips_atomic_assign_expand_fenv): New function.
9137 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
9139 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9141 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
9142 (__builtin_mips_set_fcsr): Likewise.
9143 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
9144 MIPS_USI_FTYPE_VOID.
9145 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
9146 (mips16_expand_set_fcsr): Likewise.
9147 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
9148 (mips16_set_fcsr_stub): Likewise.
9149 (mips16_get_fcsr_one_only_stub): New class.
9150 (mips16_set_fcsr_one_only_stub): Likewise.
9151 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
9152 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
9153 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
9154 (hard_float): New availability predicate.
9155 (mips_builtins): Add get_fcsr and set_fcsr.
9156 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
9157 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
9158 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
9159 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
9160 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
9163 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
9165 * config/mips/mips.c (mips_one_only_stub): New class.
9166 (mips_need_mips16_rdhwr_p): Replace with...
9167 (mips16_rdhwr_stub): ...this new variable.
9168 (mips16_stub_call_address): New function.
9169 (mips16_rdhwr_one_only_stub): New class.
9170 (mips_expand_thread_pointer): Use mips16_stub_call_address.
9171 (mips_output_mips16_rdhwr): Delete.
9172 (mips_finish_stub): New function.
9173 (mips_code_end): Use it to handle rdhwr stubs.
9175 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
9178 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
9179 when calculating size of integer atomic types.
9181 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
9183 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
9185 2014-02-01 Jakub Jelinek <jakub@redhat.com>
9187 PR tree-optimization/60003
9188 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
9189 * profile.c (branch_prob): Use gimple_call_builtin_p
9190 to check for BUILT_IN_SETJMP_RECEIVER.
9191 * tree-inline.c (copy_bb): Call notice_special_calls.
9193 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
9196 * lra-constraints.c (process_alt_operands): Update reload_sum only
9199 2014-01-31 Richard Henderson <rth@redhat.com>
9202 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
9203 until after else_eh is processed.
9205 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
9207 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
9208 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
9209 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
9210 in smmintrin.h, remove them.
9211 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
9212 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
9213 * config/i386/i386.md (ROUND_SAE): Fix value.
9214 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
9215 (const48_operand): New.
9216 * config/i386/subst.md (round), (round_expand): Use
9217 const_4_or_8_to_11_operand.
9218 (round_saeonly), (round_saeonly_expand): Use const48_operand.
9220 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
9222 * config/i386/constraints.md (Yk): Swap meaning with k.
9223 * config/i386/i386.md (movhi_internal): Change Yk to k.
9224 (movqi_internal): Ditto.
9225 (*k<logic><mode>): Ditto.
9228 (kandn<mode>): Ditto.
9229 (*<code>hi_1): Ditto.
9230 (*<code>qi_1): Ditto.
9231 (kxnor<mode>): Ditto.
9232 (kortestzhi): Ditto.
9233 (kortestchi): Ditto.
9235 (*one_cmplhi2_1): Ditto.
9236 (*one_cmplqi2_1): Ditto.
9237 * config/i386/sse.md (): Change k to Yk.
9238 (avx512f_load<mode>_mask): Ditto.
9239 (avx512f_blendm<mode>): Ditto.
9240 (avx512f_store<mode>_mask): Ditto.
9241 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
9242 (avx512f_storedqu<mode>_mask): Ditto.
9243 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
9245 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
9246 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
9247 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
9248 (avx512f_maskcmp<mode>3): Ditto.
9249 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
9250 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
9251 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
9252 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
9253 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
9254 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
9255 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
9256 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
9257 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
9258 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
9259 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
9260 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
9261 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
9262 (vec_extract_lo_<mode>_maskm): Ditto.
9263 (vec_extract_hi_<mode>_maskm): Ditto.
9264 (avx512f_vternlog<mode>_mask): Ditto.
9265 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
9266 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
9267 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
9268 (avx512f_<code>v8div16qi2_mask): Ditto.
9269 (avx512f_<code>v8div16qi2_mask_store): Ditto.
9270 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
9271 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
9272 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
9273 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
9274 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9275 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9276 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9277 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9278 (avx512cd_maskb_vec_dupv8di): Ditto.
9279 (avx512cd_maskw_vec_dupv16si): Ditto.
9280 (avx512f_vpermi2var<mode>3_maskz): Ditto.
9281 (avx512f_vpermi2var<mode>3_mask): Ditto.
9282 (avx512f_vpermi2var<mode>3_mask): Ditto.
9283 (avx512f_vpermt2var<mode>3_maskz): Ditto.
9284 (*avx512f_gathersi<mode>): Ditto.
9285 (*avx512f_gathersi<mode>_2): Ditto.
9286 (*avx512f_gatherdi<mode>): Ditto.
9287 (*avx512f_gatherdi<mode>_2): Ditto.
9288 (*avx512f_scattersi<mode>): Ditto.
9289 (*avx512f_scatterdi<mode>): Ditto.
9290 (avx512f_compress<mode>_mask): Ditto.
9291 (avx512f_compressstore<mode>_mask): Ditto.
9292 (avx512f_expand<mode>_mask): Ditto.
9293 * config/i386/subst.md (mask): Change k to Yk.
9294 (mask_scalar_merge): Ditto.
9297 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
9299 * doc/extend.texi (Vector Extensions): Document ?: in C++.
9301 2014-01-31 Richard Biener <rguenther@suse.de>
9304 * builtins.c (fold_builtin_memory_op): Make sure to not
9305 use a floating-point mode or a boolean or enumeral type for
9308 2014-01-30 DJ Delorie <dj@redhat.com>
9310 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
9311 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
9312 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
9313 whenever main() has an epilogue.
9315 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9317 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
9318 unused variable "field".
9319 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
9320 (vsx_mergeh_<mode>): Likewise.
9321 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
9322 (altivec_vmrghh): Likewise.
9323 (altivec_vmrghw): Likewise.
9324 (altivec_vmrglb): Likewise.
9325 (altivec_vmrglh): Likewise.
9326 (altivec_vmrglw): Likewise.
9327 (altivec_vspltb): Add missing uses.
9328 (altivec_vsplth): Likewise.
9329 (altivec_vspltw): Likewise.
9330 (altivec_vspltsf): Likewise.
9332 2014-01-30 Jakub Jelinek <jakub@redhat.com>
9335 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
9336 frame related instructions.
9338 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
9340 PR rtl-optimization/59959
9341 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
9342 any reload of register whose subreg is invalid.
9344 2014-01-30 Jakub Jelinek <jakub@redhat.com>
9346 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
9347 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
9348 Add missing return type - void.
9350 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9352 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
9353 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
9354 remove element index adjustment for endian (now handled in vsx.md
9356 (altivec_expand_vec_perm_const): Use
9357 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
9358 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
9359 (vsx_xxspltw_<mode>): Adjust element index for little endian.
9360 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
9361 define_expand and a new define_insn *altivec_vspltb_internal;
9362 adjust for -maltivec=be on a little endian target.
9363 (altivec_vspltb_direct): New.
9364 (altivec_vsplth): Divide into a define_expand and a new
9365 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
9366 little endian target.
9367 (altivec_vsplth_direct): New.
9368 (altivec_vspltw): Divide into a define_expand and a new
9369 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
9370 little endian target.
9371 (altivec_vspltw_direct): New.
9372 (altivec_vspltsf): Divide into a define_expand and a new
9373 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
9374 a little endian target.
9376 2014-01-30 Richard Biener <rguenther@suse.de>
9378 PR tree-optimization/59993
9379 * tree-ssa-forwprop.c (associate_pointerplus): Check we
9380 can propagate form the earlier stmt and avoid the transform
9381 when the intermediate result is needed.
9383 2014-01-30 Alangi Derick <alangiderick@gmail.com>
9385 * README.Portability: Fix typo.
9387 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
9389 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
9390 comparison_operator with ordered_comparison_operator.
9392 2014-01-30 Nick Clifton <nickc@redhat.com>
9394 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
9395 Rename to mn10300_store_multiple_regs.
9396 * config/mn10300/mn10300.c: Likewise.
9397 * config/mn10300/mn10300.md (store_movm): Fix typo: call
9398 store_multiple_regs.
9399 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
9400 Call mn10300_store_multiple_regs.
9402 2014-01-30 Nick Clifton <nickc@redhat.com>
9403 DJ Delorie <dj@redhat.com>
9405 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
9406 %fp 2 to keep registers after it properly word-aligned.
9407 (rl78_alloc_physical_registers_umul): Handle the case where both
9408 input operands are the same.
9410 2014-01-30 Richard Biener <rguenther@suse.de>
9412 PR tree-optimization/59903
9413 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
9416 2014-01-30 Jason Merrill <jason@redhat.com>
9419 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
9422 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
9424 2014-01-30 Richard Biener <rguenther@suse.de>
9426 PR tree-optimization/59951
9427 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
9429 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
9432 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
9433 SFmode to DFmode case.
9435 2014-01-29 DJ Delorie <dj@redhat.com>
9437 * config/msp430/msp430.opt (-minrt): New.
9438 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
9440 (ENDFILE_SPEC): Likewise.
9442 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
9444 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
9445 (estimate_function_body_sizes): Use it.
9447 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
9450 * dwarf2out.c (is_cxx_auto): New.
9451 (is_base_type): Use it.
9452 (gen_type_die_with_usage): Likewise.
9454 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9456 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
9457 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
9458 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
9459 -maltivec=be with LE targets.
9460 (vsx_mergeh_<mode>): Likewise.
9461 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
9462 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
9463 (altivec_vmrghb): Replace with define_expand and new
9464 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
9465 (altivec_vmrghb_direct): New define_insn.
9466 (altivec_vmrghh): Replace with define_expand and new
9467 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
9468 (altivec_vmrghh_direct): New define_insn.
9469 (altivec_vmrghw): Replace with define_expand and new
9470 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
9471 (altivec_vmrghw_direct): New define_insn.
9472 (*altivec_vmrghsf): Adjust for endianness.
9473 (altivec_vmrglb): Replace with define_expand and new
9474 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
9475 (altivec_vmrglb_direct): New define_insn.
9476 (altivec_vmrglh): Replace with define_expand and new
9477 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
9478 (altivec_vmrglh_direct): New define_insn.
9479 (altivec_vmrglw): Replace with define_expand and new
9480 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
9481 (altivec_vmrglw_direct): New define_insn.
9482 (*altivec_vmrglsf): Adjust for endianness.
9483 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
9484 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
9485 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
9486 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
9487 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
9488 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
9489 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
9490 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
9492 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
9494 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
9495 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
9498 2014-01-29 Richard Biener <rguenther@suse.de>
9500 PR tree-optimization/58742
9501 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
9502 associate_pointerplus_align.
9503 (associate_pointerplus_diff): New function.
9504 (associate_pointerplus): Likewise. Call associate_pointerplus_align
9505 and associate_pointerplus_diff.
9507 2014-01-29 Richard Biener <rguenther@suse.de>
9509 * lto-streamer.h (LTO_major_version): Bump to 3.
9510 (LTO_minor_version): Reset to 0.
9512 2014-01-29 Renlin Li <Renlin.Li@arm.com>
9514 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
9515 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
9516 (arm_file_start): Generate correct asm header for armv7ve.
9517 * config/arm/bpabi.h: Add multilib support for armv7ve.
9518 * config/arm/driver-arm.c: Change the architectures of cortex-a7
9519 and cortex-a15 to armv7ve.
9520 * config/arm/t-aprofile: Add multilib support for armv7ve.
9521 * doc/invoke.texi: Document -march=armv7ve.
9523 2014-01-29 Richard Biener <rguenther@suse.de>
9525 PR tree-optimization/58742
9526 * tree-ssa-forwprop.c (associate_plusminus): Return true
9527 if we changed sth, defer EH cleanup to ...
9528 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
9529 (simplify_mult): New function.
9531 2014-01-29 Jakub Jelinek <jakub@redhat.com>
9534 PR tree-optimization/59920
9535 * tree.c (build_common_builtin_nodes): Remove
9536 __builtin_setjmp_dispatcher initialization.
9537 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
9538 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
9539 instead of gsi_after_labels + manually skipping debug stmts.
9540 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
9541 ignore bbs with IFN_ABNORMAL_DISPATCHER.
9542 * tree-inline.c (copy_edges_for_bb): Remove
9543 can_make_abnormal_goto argument, instead add abnormal_goto_dest
9544 argument. Ignore computed_goto_p stmts. Don't call
9545 make_abnormal_goto_edges. If a call might need abnormal edges
9546 for non-local gotos, see if it already has an edge to
9547 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
9548 with true argument, don't do anything then, otherwise add
9549 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
9550 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
9552 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
9553 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
9554 (lower_stmt): Don't set data->calls_builtin_setjmp.
9555 (lower_builtin_setjmp): Adjust comment.
9556 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
9557 * tree-cfg.c (found_computed_goto): Remove.
9558 (factor_computed_gotos): Remove.
9559 (make_goto_expr_edges): Return bool, true for computed gotos.
9560 Don't call make_abnormal_goto_edges.
9561 (build_gimple_cfg): Don't set found_computed_goto, don't call
9562 factor_computed_gotos.
9563 (computed_goto_p): No longer static.
9564 (make_blocks): Don't set found_computed_goto.
9565 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
9566 (make_edges): If make_goto_expr_edges returns true, push bb
9567 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
9568 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
9569 vector. Record mapping between bbs and OpenMP regions if there
9570 are any, adjust make_gimple_omp_edges caller. Call
9571 handle_abnormal_edges.
9572 (make_abnormal_goto_edges): Remove.
9573 * tree-cfg.h (make_abnormal_goto_edges): Remove.
9574 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
9575 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
9576 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
9577 * internal-fn.def (ABNORMAL_DISPATCHER): New.
9578 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
9579 filling *region also set *region_idx to (*region)->entry->index.
9582 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
9583 For REGs set ORIGINAL_REGNO.
9585 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
9587 * doc/md.texi: Mention that a target shouldn't implement
9588 vec_widen_(s|u)mul_even/odd pair if it is less efficient
9591 2014-01-29 Jakub Jelinek <jakub@redhat.com>
9593 PR tree-optimization/59594
9594 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
9595 a copy of the datarefs vector rather than the vector itself.
9597 2014-01-28 Jason Merrill <jason@redhat.com>
9600 * dwarf2out.c (auto_die): New static.
9601 (gen_type_die_with_usage): Handle C++1y 'auto'.
9602 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
9605 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
9608 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
9609 (SPEC_X32): Likewise.
9610 (SPEC_64): Likewise.
9611 * config/i386/i386.c (ix86_option_override_internal): Turn off
9612 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
9614 (x86_file_start): Output .code16gcc for TARGET_16BIT.
9615 * config/i386/i386.h (TARGET_16BIT): New macro.
9616 (TARGET_16BIT_P): Likewise.
9617 * config/i386/i386.opt: Add m16.
9618 * doc/invoke.texi: Document -m16.
9620 2014-01-28 Jakub Jelinek <jakub@redhat.com>
9622 PR preprocessor/59935
9623 * input.c (location_get_source_line): Bail out on when line number
9624 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
9626 2014-01-28 Richard Biener <rguenther@suse.de>
9628 PR tree-optimization/58742
9629 * tree-ssa-forwprop.c (associate_plusminus): Handle
9630 pointer subtraction of the form (T)(P + A) - (T)P.
9632 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9634 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
9637 2014-01-28 Richard Biener <rguenther@suse.de>
9640 2014-01-28 Richard Biener <rguenther@suse.de>
9642 PR rtl-optimization/45364
9643 PR rtl-optimization/59890
9644 * var-tracking.c (local_get_addr_clear_given_value): Handle
9645 already cleared slot.
9646 (val_reset): Handle not allocated local_get_addr_cache.
9647 (vt_find_locations): Use post-order on the inverted CFG.
9649 2014-01-28 Richard Biener <rguenther@suse.de>
9651 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
9653 2014-01-28 Richard Biener <rguenther@suse.de>
9655 PR rtl-optimization/45364
9656 PR rtl-optimization/59890
9657 * var-tracking.c (local_get_addr_clear_given_value): Handle
9658 already cleared slot.
9659 (val_reset): Handle not allocated local_get_addr_cache.
9660 (vt_find_locations): Use post-order on the inverted CFG.
9662 2014-01-28 Alan Modra <amodra@gmail.com>
9664 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
9665 * configure.ac <recursive call for build != host>: Define
9666 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
9667 and LD_FOR_BUILD too.
9668 * configure: Regenerate.
9670 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
9672 * config/i386/i386.c (get_builtin_code_for_version): Separate
9673 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
9674 Broadwell from Haswell.
9676 2014-01-27 Steve Ellcey <sellcey@mips.com>
9678 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
9679 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
9680 * config/mips/mips.c (mips_option_override): Change setting
9682 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
9683 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
9684 Change from Mask to Var.
9686 2014-01-27 Jeff Law <law@redhat.com>
9688 * ipa-inline.c (inline_small_functions): Fix typo.
9690 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
9692 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
9693 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
9694 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
9695 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
9696 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
9697 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
9698 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
9699 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
9700 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
9701 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
9702 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
9703 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
9704 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
9705 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
9706 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
9707 (_mm512_storeu_epi64): Ditto.
9708 (_mm512_cmpge_epi32_mask): Ditto.
9709 (_mm512_cmpge_epu32_mask): Ditto.
9710 (_mm512_cmpge_epi64_mask): Ditto.
9711 (_mm512_cmpge_epu64_mask): Ditto.
9712 (_mm512_cmple_epi32_mask): Ditto.
9713 (_mm512_cmple_epu32_mask): Ditto.
9714 (_mm512_cmple_epi64_mask): Ditto.
9715 (_mm512_cmple_epu64_mask): Ditto.
9716 (_mm512_cmplt_epi32_mask): Ditto.
9717 (_mm512_cmplt_epu32_mask): Ditto.
9718 (_mm512_cmplt_epi64_mask): Ditto.
9719 (_mm512_cmplt_epu64_mask): Ditto.
9720 (_mm512_cmpneq_epi32_mask): Ditto.
9721 (_mm512_cmpneq_epu32_mask): Ditto.
9722 (_mm512_cmpneq_epi64_mask): Ditto.
9723 (_mm512_cmpneq_epu64_mask): Ditto.
9724 (_mm512_expand_pd): Ditto.
9725 (_mm512_expand_ps): Ditto.
9726 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
9727 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
9728 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
9729 * config/i386/i386.c (ix86_builtins): Add
9730 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
9731 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
9732 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
9733 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
9734 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
9735 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
9736 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
9737 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
9738 IX86_BUILTIN_PMOVUSQW512_MEM.
9739 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
9740 __builtin_ia32_pmovsqd512mem_mask,
9741 __builtin_ia32_pmovqd512mem_mask,
9742 __builtin_ia32_pmovusqw512mem_mask,
9743 __builtin_ia32_pmovsqw512mem_mask,
9744 __builtin_ia32_pmovqw512mem_mask,
9745 __builtin_ia32_pmovusdw512mem_mask,
9746 __builtin_ia32_pmovsdw512mem_mask,
9747 __builtin_ia32_pmovdw512mem_mask,
9748 __builtin_ia32_pmovqb512mem_mask,
9749 __builtin_ia32_pmovusqb512mem_mask,
9750 __builtin_ia32_pmovsqb512mem_mask,
9751 __builtin_ia32_pmovusdb512mem_mask,
9752 __builtin_ia32_pmovsdb512mem_mask,
9753 __builtin_ia32_pmovdb512mem_mask.
9754 (bdesc_args): Add __builtin_ia32_expanddf512,
9755 __builtin_ia32_expandsf512.
9756 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
9757 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
9758 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
9759 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
9760 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
9761 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
9762 (avx512f_<code>v8div16qi2_mask_store): This.
9763 (avx512f_expand<mode>): New.
9765 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
9767 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
9769 (_mm512_mask_prefetch_i64gather_pd): Ditto.
9770 (_mm512_prefetch_i32scatter_pd): Ditto.
9771 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
9772 (_mm512_prefetch_i64scatter_pd): Ditto.
9773 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
9774 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
9775 (_mm512_mask_prefetch_i64gather_ps): Ditto.
9776 (_mm512_prefetch_i32scatter_ps): Ditto.
9777 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
9778 (_mm512_prefetch_i64scatter_ps): Ditto.
9779 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
9780 * config/i386/i386-builtin-types.def: Define
9781 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
9782 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
9783 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
9784 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
9785 IX86_BUILTIN_SCATTERPFQPD.
9786 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
9787 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
9788 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
9789 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
9790 __builtin_ia32_scatterpfqps.
9791 (ix86_expand_builtin): Expand new built-ins.
9792 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
9793 fix memory access data type.
9794 (*avx512pf_gatherpf<mode>_mask): Ditto.
9795 (*avx512pf_gatherpf<mode>): Ditto.
9796 (avx512pf_scatterpf<mode>): Ditto.
9797 (*avx512pf_scatterpf<mode>_mask): Ditto.
9798 (*avx512pf_scatterpf<mode>): Ditto.
9799 (GATHER_SCATTER_SF_MEM_MODE): New.
9800 (avx512pf_gatherpf<mode>df): Ditto.
9801 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9802 (*avx512pf_scatterpf<mode>df): Ditto.
9804 2014-01-27 Jakub Jelinek <jakub@redhat.com>
9807 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
9808 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
9811 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
9813 * common/config/arm/arm-common.c
9814 (arm_rewrite_mcpu): Handle multiple names.
9816 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
9818 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
9820 * gimple-builder.h (create_gimple_tmp): Delete.
9822 2014-01-27 Christian Bruel <christian.bruel@st.com>
9824 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
9827 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
9829 * config/pa/pa.md (call): Generate indirect long calls to non-local
9830 functions when outputing 32-bit code.
9831 (call_value): Likewise except for special call to buggy powf function.
9833 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
9834 portable runtime and PIC indirect calls.
9835 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
9836 and PIC call sequences. Use ldo instead of blr to set return register
9837 in PIC call sequence.
9839 2014-01-25 Walter Lee <walt@tilera.com>
9841 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
9842 avoid clobbering a live register.
9844 2014-01-25 Walter Lee <walt@tilera.com>
9846 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
9847 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
9848 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
9849 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
9851 2014-01-25 Walter Lee <walt@tilera.com>
9853 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
9854 arguments on even registers.
9855 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
9857 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
9858 (BIGGEST_ALIGNMENT): Ditto.
9859 (BIGGEST_FIELD_ALIGNMENT): Ditto.
9861 2014-01-25 Walter Lee <walt@tilera.com>
9863 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
9864 insns before bundling.
9865 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
9867 2014-01-25 Walter Lee <walt@tilera.com>
9869 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
9870 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
9871 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
9873 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9875 * config/mips/constraints.md (kl): Delete.
9876 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
9877 define expands, using...
9878 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
9879 instructions for MIPS16.
9880 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
9881 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
9883 2014-01-25 Walter Lee <walt@tilera.com>
9885 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
9889 2014-01-25 Walter Lee <walt@tilera.com>
9891 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
9892 (TARGET_EXPAND_TO_RTL_HOOK): Define.
9894 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9896 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
9899 2014-01-25 Jakub Jelinek <jakub@redhat.com>
9901 * print-rtl.c (in_call_function_usage): New var.
9902 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
9903 EXPR_LIST mode as mode and not as reg note name.
9906 * cfgloopmanip.c (copy_loop_info): If
9907 loop->warned_aggressive_loop_optimizations, make sure
9908 the flag is set in target loop too.
9910 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
9912 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
9914 * builtins.def: Likewise.
9915 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
9916 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
9917 * ira.c (ira_setup_eliminable_regset): Likewise.
9918 * omp-low.c (gate_expand_omp): Likewise.
9919 (execute_lower_omp): Likewise.
9920 (diagnose_sb_0): Likewise.
9921 (gate_diagnose_omp_blocks): Likewise.
9922 (simd_clone_clauses_extract): Likewise.
9925 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9927 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
9928 correction for little endian...
9929 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
9932 2014-01-24 Jeff Law <law@redhat.com>
9934 PR tree-optimization/59919
9935 * tree-vrp.c (find_assert_locations_1): Do not register asserts
9936 for non-returning calls.
9938 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
9940 * common/config/aarch64/aarch64-common.c
9941 (aarch64_rewrite_mcpu): Handle multiple names.
9942 * config/aarch64/aarch64.h
9943 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
9945 2014-01-24 Dodji Seketeli <dodji@redhat.com>
9947 * input.c (add_file_to_cache_tab): Handle the case where fopen
9950 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
9953 * config/i386/i386.md (pushsf splitter): Get stack adjustment
9954 from push operand if code of push isn't PRE_DEC.
9956 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
9959 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
9960 -mquad-memory-atomic. Update -mquad-memory documentation to say
9961 it is only used for non-atomic loads/stores.
9963 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
9964 -mquad-memory or -mquad-memory-atomic switches.
9966 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
9967 -mquad-memory-atomic to ISA 2.07 support.
9969 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
9970 to separate support of normal quad word memory operations (ldq, stq)
9971 from the atomic quad word memory operations.
9973 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
9974 support to separate non-atomic quad word operations from atomic
9975 quad word operations. Disable non-atomic quad word operations in
9976 little endian mode so that we don't have to swap words after the
9977 load and before the store.
9978 (quad_load_store_p): Add comment about atomic quad word support.
9979 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
9980 options printed with -mdebug=reg.
9982 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
9983 -mquad-memory-atomic as the test for whether we have quad word
9984 atomic instructions.
9985 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
9986 or -mp8-vector are used, allow byte/half-word atomic operations.
9988 * config/rs6000/sync.md (load_lockedti): Insure that the address
9989 is a proper indexed or indirect address for the lqarx instruction.
9990 On little endian systems, swap the hi/lo registers after the lqarx
9992 (load_lockedpti): Use indexed_or_indirect_operand predicate to
9993 insure the address is valid for the lqarx instruction.
9994 (store_conditionalti): Insure that the address is a proper indexed
9995 or indirect address for the stqcrx. instruction. On little endian
9996 systems, swap the hi/lo registers before doing the stqcrx.
9998 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
9999 insure the address is valid for the stqcrx. instruction.
10001 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
10002 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
10003 type of quad memory support is available.
10005 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
10007 PR regression/59915
10008 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
10009 there is a danger of looping.
10011 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
10013 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
10014 force flag_ira_loop_pressure if set via command line.
10016 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
10018 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
10019 (ashr_simd): New builtin handling DI mode.
10020 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
10021 (aarch64_sshr_simddi): New match pattern.
10022 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
10023 (vshrd_n_s64): Likewise.
10024 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
10026 2014-01-23 Nick Clifton <nickc@redhat.com>
10028 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
10029 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
10030 favour of mcu specific scripts.
10031 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
10034 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
10035 Alex Velenko <Alex.Velenko@arm.com>
10037 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
10038 (vaddv_s16): Likewise.
10039 (vaddv_s32): Likewise.
10040 (vaddv_u8): Likewise.
10041 (vaddv_u16): Likewise.
10042 (vaddv_u32): Likewise.
10043 (vaddvq_s8): Likewise.
10044 (vaddvq_s16): Likewise.
10045 (vaddvq_s32): Likewise.
10046 (vaddvq_s64): Likewise.
10047 (vaddvq_u8): Likewise.
10048 (vaddvq_u16): Likewise.
10049 (vaddvq_u32): Likewise.
10050 (vaddvq_u64): Likewise.
10051 (vaddv_f32): Likewise.
10052 (vaddvq_f32): Likewise.
10053 (vaddvq_f64): Likewise.
10054 (vmaxv_f32): Likewise.
10055 (vmaxv_s8): Likewise.
10056 (vmaxv_s16): Likewise.
10057 (vmaxv_s32): Likewise.
10058 (vmaxv_u8): Likewise.
10059 (vmaxv_u16): Likewise.
10060 (vmaxv_u32): Likewise.
10061 (vmaxvq_f32): Likewise.
10062 (vmaxvq_f64): Likewise.
10063 (vmaxvq_s8): Likewise.
10064 (vmaxvq_s16): Likewise.
10065 (vmaxvq_s32): Likewise.
10066 (vmaxvq_u8): Likewise.
10067 (vmaxvq_u16): Likewise.
10068 (vmaxvq_u32): Likewise.
10069 (vmaxnmv_f32): Likewise.
10070 (vmaxnmvq_f32): Likewise.
10071 (vmaxnmvq_f64): Likewise.
10072 (vminv_f32): Likewise.
10073 (vminv_s8): Likewise.
10074 (vminv_s16): Likewise.
10075 (vminv_s32): Likewise.
10076 (vminv_u8): Likewise.
10077 (vminv_u16): Likewise.
10078 (vminv_u32): Likewise.
10079 (vminvq_f32): Likewise.
10080 (vminvq_f64): Likewise.
10081 (vminvq_s8): Likewise.
10082 (vminvq_s16): Likewise.
10083 (vminvq_s32): Likewise.
10084 (vminvq_u8): Likewise.
10085 (vminvq_u16): Likewise.
10086 (vminvq_u32): Likewise.
10087 (vminnmv_f32): Likewise.
10088 (vminnmvq_f32): Likewise.
10089 (vminnmvq_f64): Likewise.
10091 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
10093 * config/aarch64/aarch64-simd.md
10094 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
10095 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
10096 (*aarch64_mul3_elt<mode>): Likewise.
10097 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
10098 (*aarch64_mul3_elt_to_64v2df): Likewise.
10099 (*aarch64_mla_elt<mode>): Likewise.
10100 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
10101 (*aarch64_mls_elt<mode>): Likewise.
10102 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
10103 (*aarch64_fma4_elt<mode>): Likewise.
10104 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
10105 (*aarch64_fma4_elt_to_64v2df): Likewise.
10106 (*aarch64_fnma4_elt<mode>): Likewise.
10107 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
10108 (*aarch64_fnma4_elt_to_64v2df): Likewise.
10109 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
10110 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
10111 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
10112 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
10113 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
10114 (aarch64_sqdmull_lane<mode>_internal): Likewise.
10115 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
10117 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
10119 * config/aarch64/aarch64-simd.md
10120 (aarch64_be_checked_get_lane<mode>): New define_expand.
10121 * config/aarch64/aarch64-simd-builtins.def
10122 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
10123 New builtin definition.
10124 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
10125 Use new safe be builtin.
10127 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
10129 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
10131 (aarch64_be_st1<mode>): Likewise.
10132 (aarch_ld1<VALL:mode>): Define_expand modified.
10133 (aarch_st1<VALL:mode>): Likewise.
10134 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
10135 (UNSPEC_ST1): Likewise.
10137 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
10139 * config/microblaze/microblaze.md: Add trap insn and attribute
10141 2014-01-23 Dodji Seketeli <dodji@redhat.com>
10143 PR preprocessor/58580
10144 * input.h (location_get_source_line): Take an additional line_size
10146 (void diagnostics_file_cache_fini): Declare new function.
10147 * input.c (struct fcache): New type.
10148 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
10149 New static constants.
10150 (diagnostic_file_cache_init, total_lines_num)
10151 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
10152 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
10153 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
10154 (get_next_line, read_next_line, goto_next_line, read_line_num):
10155 New static function definitions.
10156 (diagnostic_file_cache_fini): New function.
10157 (location_get_source_line): Take an additional output line_len
10158 parameter. Re-write using lookup_or_add_file_to_cache_tab and
10160 * diagnostic.c (diagnostic_finish): Call
10161 diagnostic_file_cache_fini.
10162 (adjust_line): Take an additional input parameter for the length
10163 of the line, rather than calculating it with strlen.
10164 (diagnostic_show_locus): Adjust the use of
10165 location_get_source_line and adjust_line with respect to their new
10166 signature. While displaying a line now, do not stop at the first
10167 null byte. Rather, display the zero byte as a space and keep
10168 going until we reach the size of the line.
10169 * Makefile.in: Add vec.o to OBJS-libcommon
10171 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
10172 Ilya Tocar <ilya.tocar@intel.com>
10174 * config/i386/avx512fintrin.h (_mm512_kmov): New.
10175 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
10176 (__builtin_ia32_kmov16): Ditto.
10177 * config/i386/i386.md (UNSPEC_KMOV): New.
10180 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
10182 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
10183 (_mm512_storeu_si512): Ditto.
10185 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
10188 * rtl.h (get_referenced_operands): Declare.
10189 * recog.c (get_referenced_operands): New function.
10190 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
10191 operands have been referenced when recording LO_SUM references.
10193 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
10195 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
10197 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
10199 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
10200 Enable for generic and recent AMD targets.
10202 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
10204 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
10205 ARG_SIZE note when adjustment was eliminated.
10207 2014-01-22 Jeff Law <law@redhat.com>
10209 PR tree-optimization/59597
10210 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
10211 in file. Accept new argument REGISTERING and use it to modify
10212 dump output appropriately.
10213 (register_jump_thread): Corresponding changes.
10214 (mark_threaded_blocks): Reinstate code to cancel unprofitable
10215 thread paths involving joiner blocks. Add code to dump cancelled
10216 jump threading paths.
10218 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
10220 PR rtl-optimization/59477
10221 * lra-constraints.c (inherit_in_ebb): Process call for living hard
10222 regs. Update reloads_num and potential_reload_hard_regs for all insns.
10224 2014-01-22 Tom Tromey <tromey@redhat.com>
10226 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
10228 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
10230 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
10232 PR rtl-optimization/59896
10233 * lra-constraints.c (process_alt_operands): Check unused note for
10234 matched operands of insn with no output reloads.
10236 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
10238 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
10239 (mips_move_from_gpr_cost): Likewise.
10241 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
10243 PR rtl-optimization/59858
10244 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
10245 ira_class_hard_regs_num.
10246 (process_alt_operands): Increase reject for dying matched operand.
10248 2014-01-21 Jakub Jelinek <jakub@redhat.com>
10251 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
10252 smaller than size, perform several stores or loads and stores
10253 at dst + count - size to store or copy all of size bytes, rather
10254 than just last modesize bytes.
10256 2014-01-20 DJ Delorie <dj@redhat.com>
10258 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
10259 that CLOBBERs are REGs before propogating their values.
10261 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
10263 PR middle-end/59789
10264 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
10265 (cgraph_inline_failed_type): New function.
10266 * cgraph.h (DEFCIFCODE): Add type.
10267 (cgraph_inline_failed_type_t): New enum.
10268 (cgraph_inline_failed_type): New prototype.
10269 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
10270 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
10271 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
10272 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
10273 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
10274 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
10275 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
10276 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
10277 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
10278 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
10279 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
10280 OPTIMIZATION_MISMATCH.
10281 * tree-inline.c (expand_call_inline): Emit errors during
10282 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
10284 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
10287 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
10288 mode attribute in insn output.
10290 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
10292 * output.h (output_constant): Delete.
10293 * varasm.c (output_constant): Make private.
10295 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
10297 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
10299 2014-01-20 Jakub Jelinek <jakub@redhat.com>
10301 PR middle-end/59860
10302 * tree.h (fold_builtin_strcat): New prototype.
10303 * builtins.c (fold_builtin_strcat): No longer static. Add len
10304 argument, if non-NULL, don't call c_strlen. Optimize
10305 directly into __builtin_memcpy instead of __builtin_strcpy.
10306 (fold_builtin_2): Adjust fold_builtin_strcat caller.
10307 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
10309 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
10311 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
10312 for SImode_address_operand operands, having only a REG argument.
10314 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
10316 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
10317 loader name using mbig-endian.
10318 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
10320 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
10322 * doc/invoke.texi (-march): Clarify documentation for AArch64.
10323 (-mtune): Likewise.
10326 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
10328 * config/aarch64/aarch64-protos.h
10329 (aarch64_cannot_change_mode_class_ptr): Declare.
10330 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
10331 aarch64_cannot_change_mode_class_ptr): New.
10332 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
10333 backend hook aarch64_cannot_change_mode_class.
10335 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
10337 * common/config/aarch64/aarch64-common.c
10338 (aarch64_handle_option): Don't handle any option order logic here.
10339 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
10340 selected_cpu, warn on architecture version mismatch.
10341 (aarch64_override_options): Fix parsing order for option strings.
10343 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10344 Iain Sandoe <iain@codesourcery.com>
10347 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
10348 warning. Amend comment to reflect current functionality.
10350 2014-01-20 Richard Biener <rguenther@suse.de>
10352 PR middle-end/59860
10353 * builtins.c (fold_builtin_strcat): Remove case better handled
10354 by tree-ssa-strlen.c.
10356 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
10358 * config/aarch64/aarch64.opt
10359 (mcpu, march, mtune): Make case-insensitive.
10361 2014-01-20 Jakub Jelinek <jakub@redhat.com>
10364 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
10365 if operands[1] is a REG or ZERO_EXTEND of a REG.
10367 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
10369 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
10371 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
10373 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
10374 long non-pic millicode calls.
10376 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10378 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
10380 2014-01-19 Kito Cheng <kito@0xlab.org>
10382 * builtins.c (expand_movstr): Check movstr expand done or fail.
10384 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10385 H.J. Lu <hongjiu.lu@intel.com>
10388 * config/i386/i386.md (*lea<mode>): Zero-extend return register
10389 to DImode for zero-extended addresses.
10391 2014-01-19 Jakub Jelinek <jakub@redhat.com>
10393 PR rtl-optimization/57763
10394 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
10395 on the new indirect jump_insn and increment LABEL_NUSES (label).
10397 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
10401 * config.gcc (x86_archs): New variable.
10402 (x86_64_archs): Likewise.
10403 (x86_cpus): Likewise.
10404 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
10405 --with-arch/--with-cpu= options.
10406 Support --with-arch=/--with-cpu={nehalem,westmere,
10407 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
10409 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10411 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
10412 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
10414 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10416 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
10418 2014-01-18 Jakub Jelinek <jakub@redhat.com>
10421 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
10422 clear cpp_get_options (parse_in)->warn_unused_macros for
10423 ix86_target_macros_internal with cpp_define.
10425 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
10427 * jump.c (delete_related_insns): Keep (use (insn))s.
10428 * reorg.c (redundant_insn): Check for barriers too.
10430 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10432 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
10434 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
10436 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
10437 call to $$dyncall when TARGET_LONG_CALLS is true.
10439 2014-01-17 Jeff Law <law@redhat.com>
10441 * ree.c (combine_set_extension): Temporarily disable test for
10442 changing number of hard registers.
10444 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
10446 PR middle-end/58125
10447 * ipa-inline-analysis.c (inline_free_summary):
10448 Do not free summary of aliases.
10450 2014-01-17 Jakub Jelinek <jakub@redhat.com>
10452 PR middle-end/59706
10453 * gimplify.c (gimplify_expr): Use create_tmp_var
10454 instead of create_tmp_var_raw. If cond doesn't have
10455 integral type, don't add the IFN_ANNOTATE builtin at all.
10457 2014-01-17 Martin Jambor <mjambor@suse.cz>
10460 * ipa-cp.c (prev_edge_clone): New variable.
10461 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
10462 Also resize prev_edge_clone vector.
10463 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
10464 (ipcp_edge_removal_hook): New function.
10465 (ipcp_driver): Register ipcp_edge_removal_hook.
10467 2014-01-17 Andrew Pinski <apinski@cavium.com>
10468 Steve Ellcey <sellcey@mips.com>
10471 * config/mips/mips.c (mips_print_operand): Check operand mode instead
10474 2014-01-17 Jeff Law <law@redhat.com>
10476 PR middle-end/57904
10477 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
10478 so that pass_ccp runs first.
10480 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10482 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
10483 (ix86_adjust_cost): Use !TARGET_XXX.
10484 (do_reorder_for_imul): Likewise.
10485 (swap_top_of_ready_list): Likewise.
10486 (ix86_sched_reorder): Likewise.
10488 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10490 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
10491 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
10492 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
10493 (intel_memset): New. Duplicate slm_memset.
10494 (intel_cost): New. Duplicate slm_cost.
10495 (m_INTEL): New macro.
10496 (processor_target_table): Add "intel".
10497 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
10498 with PROCESSOR_INTEL for "intel".
10499 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
10500 PROCESSOR_SILVERMONT.
10501 (ix86_issue_rate): Likewise.
10502 (ix86_adjust_cost): Likewise.
10503 (ia32_multipass_dfa_lookahead): Likewise.
10504 (swap_top_of_ready_list): Likewise.
10505 (ix86_sched_reorder): Likewise.
10506 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
10507 instead of TARGET_OPT_AGU.
10508 * config/i386/i386.h (TARGET_INTEL): New.
10509 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
10510 (processor_type): Add PROCESSOR_INTEL.
10511 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
10512 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
10514 2014-01-17 Marek Polacek <polacek@redhat.com>
10517 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
10520 2014-01-17 Richard Biener <rguenther@suse.de>
10522 PR tree-optimization/46590
10523 * opts.c (default_options_table): Add entries for
10524 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
10525 all enabled at -O1 but not for -Og.
10526 * common.opt (fbranch-count-reg): Remove Init(1).
10527 (fmove-loop-invariants): Likewise.
10528 (ftree-pta): Likewise.
10530 2014-01-17 Jakub Jelinek <jakub@redhat.com>
10532 * config/i386/i386.c (ix86_data_alignment): For compatibility with
10533 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
10534 decls to at least the GCC 4.8 used alignments.
10537 * tree-nested.c (convert_nonlocal_reference_stmt,
10538 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
10539 of GIMPLE_BIND stmts, adjust associated decls.
10541 2014-01-17 Richard Biener <rguenther@suse.de>
10543 PR tree-optimization/46590
10544 * vec.h (vec<>::bseach): New member function implementing
10545 binary search according to C89 bsearch.
10546 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
10547 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
10548 bitmap pointer again. Make accesses_in_loop a flat array.
10549 (mem_ref_obstack): New global.
10550 (outermost_indep_loop): Adjust for mem_ref->stored changes.
10551 (mark_ref_stored): Likewise.
10552 (ref_indep_loop_p_2): Likewise.
10553 (set_ref_stored_in_loop): New helper function.
10554 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
10555 (memref_free): Adjust.
10556 (record_mem_ref_loc): Simplify.
10557 (gather_mem_refs_stmt): Adjust.
10558 (sort_locs_in_loop_postorder_cmp): New function.
10559 (analyze_memory_references): Sort accesses_in_loop after
10560 loop postorder number.
10561 (find_ref_loc_in_loop_cmp): New function.
10562 (for_all_locs_in_loop): Find relevant cluster of locs in
10563 accesses_in_loop and iterate without recursion.
10564 (execute_sm): Avoid uninit warning.
10565 (struct ref_always_accessed): Simplify.
10566 (ref_always_accessed::operator ()): Likewise.
10567 (ref_always_accessed_p): Likewise.
10568 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
10569 loop postorder numbers here.
10570 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
10573 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
10576 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
10577 on decls for which assemble_alias has been called.
10579 2014-01-17 Nick Clifton <nickc@redhat.com>
10581 * config/msp430/msp430.opt: (mcpu): New option.
10582 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
10583 (msp430_option_override): Parse target_cpu. If the MCU name
10584 matches a generic string, clear target_mcu.
10585 (msp430_attr): Allow numeric interrupt values up to 63.
10586 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
10587 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
10589 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
10591 * config/msp430/msp430.md (popm): Use %J rather than %I.
10592 (addsi3): Use msp430_nonimmediate_operand for operand 2.
10593 (addhi_cy_i): Use immediate_operand for operand 2.
10594 * doc/invoke.texi: Document -mcpu option.
10596 2014-01-17 Richard Biener <rguenther@suse.de>
10598 PR rtl-optimization/38518
10599 * df.h (df_analyze_loop): Declare.
10600 * df-core.c: Include cfgloop.h.
10601 (df_analyze_1): Split out main part of df_analyze.
10602 (df_analyze): Adjust.
10603 (loop_inverted_post_order_compute): New function.
10604 (loop_post_order_compute): Likewise.
10605 (df_analyze_loop): New function avoiding whole-function
10606 postorder computes.
10607 * loop-invariant.c (find_defs): Use df_analyze_loop.
10608 (find_invariants): Adjust.
10609 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
10611 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
10613 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
10614 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
10616 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
10618 * ipa-ref.c (ipa_remove_stmt_references): Fix references
10619 traversal when removing references.
10621 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
10624 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
10626 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
10628 PR middle-end/56791
10629 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
10630 pushing a reload for an autoinc when we had previously reloaded an
10631 inner part of the address.
10633 2014-01-16 Jakub Jelinek <jakub@redhat.com>
10635 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
10637 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
10638 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
10639 when not giving up or versioning for alias only because of
10641 (vect_analyze_data_ref_dependences): Set to true.
10642 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
10644 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
10645 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
10648 PR middle-end/58344
10649 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
10652 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
10653 operand 0 predicate for gathers, use a new pseudo as subtarget.
10655 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
10657 PR middle-end/59609
10658 * lra-constraints.c (process_alt_operands): Add printing debug info.
10659 Check absence of input/output reloads for matched operands too.
10661 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
10663 PR rtl-optimization/59835
10664 * ira.c (ira_init_register_move_cost): Increase cost for
10667 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
10669 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
10671 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
10674 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
10675 non-register objects. Use gen_(high/low)part more consistently.
10678 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
10681 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
10682 endian support, remove tests for WORDS_BIG_ENDIAN.
10683 (p8_mfvsrd_3_<mode>): Likewise.
10684 (reload_gpr_from_vsx<mode>): Likewise.
10685 (reload_gpr_from_vsxsf): Likewise.
10686 (p8_mfvsrd_4_disf): Likewise.
10688 2014-01-16 Richard Biener <rguenther@suse.de>
10690 PR rtl-optimization/46590
10691 * lcm.c (compute_antinout_edge): Use postorder iteration.
10692 (compute_laterin): Use inverted postorder iteration.
10694 2014-01-16 Nick Clifton <nickc@redhat.com>
10696 PR middle-end/28865
10697 * varasm.c (output_constant): Return the number of bytes actually
10699 (output_constructor_array_range): Update the field size with the
10700 number of bytes emitted by output_constant.
10701 (output_constructor_regular_field): Likewise. Also do not
10702 complain if the total number of bytes emitted is now greater
10703 than the expected fieldpos.
10704 * output.h (output_constant): Update prototype and descriptive comment.
10706 2014-01-16 Marek Polacek <polacek@redhat.com>
10708 PR middle-end/59827
10709 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
10710 it is error_mark_node.
10712 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
10714 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
10715 VALID_AVX256_REG_OR_OI_MODE.
10717 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
10719 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
10720 current procedure should be profiled.
10722 2014-01-15 Andrew Pinski <apinski@cavium.com>
10724 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
10725 of moving from/to the STACK_REG register class.
10727 2014-01-15 Richard Henderson <rth@redhat.com>
10730 * reginfo.c (global_regs_decl): Globalize.
10731 * rtl.h (global_regs_decl): Declare.
10732 * ira.c (do_reload): Diagnose frame_pointer_needed and it
10733 reserved via global_regs.
10735 2014-01-15 Teresa Johnson <tejohnson@google.com>
10737 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
10739 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10741 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
10742 and vmulosh rather than call gen_vec_widen_smult_*.
10743 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
10744 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
10745 (vec_widen_smult_even_v16qi): Likewise.
10746 (vec_widen_umult_even_v8hi): Likewise.
10747 (vec_widen_smult_even_v8hi): Likewise.
10748 (vec_widen_umult_odd_v16qi): Likewise.
10749 (vec_widen_smult_odd_v16qi): Likewise.
10750 (vec_widen_umult_odd_v8hi): Likewise.
10751 (vec_widen_smult_odd_v8hi): Likewise.
10752 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
10753 vmuloub rather than call gen_vec_widen_umult_*.
10754 (vec_widen_umult_lo_v16qi): Likewise.
10755 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
10756 vmulosb rather than call gen_vec_widen_smult_*.
10757 (vec_widen_smult_lo_v16qi): Likewise.
10758 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
10759 rather than call gen_vec_widen_umult_*.
10760 (vec_widen_umult_lo_v8hi): Likewise.
10761 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
10762 rather than call gen_vec_widen_smult_*.
10763 (vec_widen_smult_lo_v8hi): Likewise.
10765 2014-01-15 Jeff Law <law@redhat.com>
10767 PR tree-optimization/59747
10768 * ree.c (find_and_remove_re): Properly handle case where a second
10769 eliminated extension requires widening a copy created for elimination
10770 of a prior extension.
10771 (combine_set_extension): Ensure that the number of hard regs needed
10772 for a destination register does not change when we widen it.
10774 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
10776 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
10777 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
10778 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
10779 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
10780 (avr-*-rtems*): Likewise.
10781 (bfin*-rtems*): Likewise.
10782 (moxie-*-rtems*): Likewise.
10783 (h8300-*-rtems*): Likewise.
10784 (i[34567]86-*-rtems*): Likewise.
10785 (lm32-*-rtems*): Likewise.
10786 (m32r-*-rtems*): Likewise.
10787 (m68k-*-rtems*): Likewise.
10788 (microblaze*-*-rtems*): Likewise.
10789 (mips*-*-rtems*): Likewise.
10790 (powerpc-*-rtems*): Likewise.
10791 (sh-*-rtems*): Likewise.
10792 (sparc-*-rtems*): Likewise.
10793 (sparc64-*-rtems*): Likewise.
10794 (v850-*-rtems*): Likewise.
10795 (m32c-*-rtems*): Likewise.
10797 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
10799 PR rtl-optimization/59511
10800 * ira.c (ira_init_register_move_cost): Use memory costs for some
10801 cases of register move cost calculations.
10802 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
10803 instead of BB frequency.
10804 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
10805 * lra-assigns.c (find_hard_regno_for): Ditto.
10807 2014-01-15 Richard Biener <rguenther@suse.de>
10809 PR tree-optimization/59822
10810 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
10811 (vectorizable_load): Use it to hoist defs of uses of invariant
10812 loads out of the loop.
10814 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
10815 Kugan Vivekanandarajah <kuganv@linaro.org>
10818 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
10821 2014-01-15 Richard Biener <rguenther@suse.de>
10823 PR rtl-optimization/59802
10824 * lcm.c (compute_available): Use inverted postorder to seed
10825 the initial worklist.
10827 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10830 * config/s390/s390.c (s390_preferred_reload_class): Don't return
10831 ADDR_REGS for invalid symrefs in non-PIC code.
10833 2014-01-15 Jakub Jelinek <jakub@redhat.com>
10836 * builtins.c (determine_block_size): Initialize *probable_max_size
10837 even if len_rtx is CONST_INT.
10839 2014-01-14 Andrew Pinski <apinski@cavium.com>
10841 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
10842 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
10843 (cortexa53_tunings): Likewise.
10844 (aarch64_sched_issue_rate): New function.
10845 (TARGET_SCHED_ISSUE_RATE): Define.
10847 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10849 * ira-costs.c (find_costs_and_classes): Add missed
10850 ira_init_register_move_cost_if_necessary.
10852 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10855 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
10857 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
10860 * config/i386/i386.c (type_natural_mode): Add a bool parameter
10861 to indicate if type is used for function return value. Warn ABI
10862 change if the vector mode isn't available for function return value.
10863 (ix86_function_arg_advance): Pass false to type_natural_mode.
10864 (ix86_function_arg): Likewise.
10865 (ix86_gimplify_va_arg): Likewise.
10866 (function_arg_32): Don't warn ABI change.
10867 (ix86_function_value): Pass true to type_natural_mode.
10868 (ix86_return_in_memory): Likewise.
10869 (ix86_struct_value_rtx): Removed.
10870 (TARGET_STRUCT_VALUE_RTX): Likewise.
10872 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10874 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
10875 converting a conditional jump into a conditional return.
10877 2014-01-14 Richard Biener <rguenther@suse.de>
10879 PR tree-optimization/58921
10880 PR tree-optimization/59006
10881 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
10882 hoisting invariant stmts.
10883 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
10884 invariant loads on the preheader edge if possible.
10886 2014-01-14 Joey Ye <joey.ye@arm.com>
10888 * doc/plugin.texi (Building GCC plugins): Update to C++.
10890 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
10892 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
10893 (_mm_rcp28_round_ss): Ditto.
10894 (_mm_rsqrt28_round_sd): Ditto.
10895 (_mm_rsqrt28_round_ss): Ditto.
10896 (_mm_rcp28_sd): Ditto.
10897 (_mm_rcp28_ss): Ditto.
10898 (_mm_rsqrt28_sd): Ditto.
10899 (_mm_rsqrt28_ss): Ditto.
10900 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
10901 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
10902 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
10903 (IX86_BUILTIN_RCP28SD): Ditto.
10904 (IX86_BUILTIN_RCP28SS): Ditto.
10905 (IX86_BUILTIN_RSQRT28SD): Ditto.
10906 (IX86_BUILTIN_RSQRT28SS): Ditto.
10907 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
10908 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
10909 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
10910 (ix86_expand_special_args_builtin): Expand new FTYPE.
10911 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
10912 (srcp14<mode>): Make insn unary.
10913 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
10914 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
10915 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
10916 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
10917 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
10918 Fix rounding: make it SAE only.
10919 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
10921 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
10923 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
10924 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
10925 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
10926 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
10927 (round_saeonly_mask_scalar_operand4): Ditto.
10928 (round_saeonly_mask_scalar_op3): Ditto.
10929 (round_saeonly_mask_scalar_op4): Ditto.
10931 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10933 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
10934 Implement -maltivec=be for vec_insert and vec_extract.
10936 2014-01-10 DJ Delorie <dj@redhat.com>
10938 * config/msp430/msp430.md (call_internal): Don't allow memory
10939 references with SP as the base register.
10940 (call_value_internal): Likewise.
10941 * config/msp430/constraints.md (Yc): New. For memory references
10942 that don't use SP as a base register.
10944 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
10945 "an integer without a # prefix"
10946 * config/msp430/msp430.md (epilogue_helper): Use it.
10948 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10951 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
10952 AVX512F gather builtins.
10953 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
10954 on gather decls with INTEGER_TYPE masktype.
10955 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
10956 directly into the builtin rather than hoisting it before loop.
10958 PR tree-optimization/59387
10959 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
10960 (scev_const_prop): If folded_casts and type has undefined overflow,
10961 use force_gimple_operand instead of force_gimple_operand_gsi and
10962 for each added stmt if it is assign with
10963 arith_code_with_undefined_signed_overflow, call
10964 rewrite_to_defined_overflow.
10965 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
10966 gimple-fold.h instead.
10967 (arith_code_with_undefined_signed_overflow,
10968 rewrite_to_defined_overflow): Moved to ...
10969 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
10970 rewrite_to_defined_overflow): ... here. No longer static.
10971 Include gimplify-me.h.
10972 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
10973 rewrite_to_defined_overflow): New prototypes.
10975 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10977 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
10979 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
10981 * builtins.c (get_object_alignment_2): Minor tweak.
10982 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
10984 2014-01-13 Christian Bruel <christian.bruel@st.com>
10986 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
10987 optimized non constant lengths.
10989 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10992 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
10993 load as __atomic_load_N if possible.
10995 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10997 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
10999 (rs6000_expand_builtin): Adjust call.
11001 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
11004 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
11005 * config/rs6000/rs6000.c: Include target-globals.h.
11006 (rs6000_set_current_function): Instead of doing target_reinit
11007 unconditionally, use save_target_globals_default_opts and
11008 restore_target_globals.
11010 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
11012 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
11013 (rs6000_expand_builtin): Handle mffs and mtfsf.
11014 (rs6000_init_builtins): Define mffs and mtfsf.
11015 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
11016 (rs6000_mffs): New pattern.
11017 (rs6000_mtfsf): New pattern.
11019 2014-01-11 Bin Cheng <bin.cheng@arm.com>
11021 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
11022 Start narrowing with START. Apply candidate-use pair
11023 and check overall cost in narrowing.
11024 (iv_ca_prune): Pass new argument.
11026 2014-01-10 Jeff Law <law@redhat.com>
11028 PR middle-end/59743
11029 * ree.c (combine_reaching_defs): Ensure the defining statement
11030 occurs before the extension when optimizing extensions with
11031 different source and destination hard registers.
11033 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
11036 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
11037 vtables into the type inheritance graph.
11039 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11041 PR rtl-optimization/59754
11042 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
11043 modes in the REGNO != REGNO case.
11045 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11047 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
11049 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11051 PR tree-optimization/59745
11052 * tree-predcom.c (tree_predictive_commoning_loop): Call
11053 free_affine_expand_cache if giving up because components is NULL.
11055 * target-globals.c (save_target_globals): Allocate < 4KB structs using
11056 GC in payload of target_globals struct instead of allocating them on
11057 the heap and the larger structs separately using GC.
11058 * target-globals.h (struct target_globals): Make regs, hard_regs,
11059 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
11060 of GTY((skip)) and change type to void *.
11061 (reset_target_globals): Cast loads from those fields to corresponding
11064 2014-01-10 Steve Ellcey <sellcey@mips.com>
11067 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
11068 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
11069 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
11071 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
11074 * aarch64-modes.def (CC_Zmode): New flags mode.
11075 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
11076 represents an equality.
11077 (aarch64_get_condition_code): Handle CC_Zmode.
11078 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
11080 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11082 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
11083 extraction in good case.
11085 2014-01-10 Richard Biener <rguenther@suse.de>
11087 PR tree-optimization/59374
11088 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
11089 checking after SLP discovery. Mark stmts not participating
11090 in any SLP instance properly.
11092 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11094 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
11095 when handling a SET rtx.
11097 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11099 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
11100 (cortex-a57): Likewise.
11101 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
11103 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11105 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
11106 non-iwmmxt builtins.
11108 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
11112 * ipa-devirt.c record_target_from_binfo): Take as argument
11113 stack of binfos and lookup matching one for virtual inheritance.
11114 (possible_polymorphic_call_targets_1): Update.
11116 2014-01-10 Huacai Chen <chenhc@lemote.com>
11118 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
11119 kernel strings for Loongson-2E/2F/3A.
11121 2014-01-10 Jakub Jelinek <jakub@redhat.com>
11123 PR middle-end/59670
11124 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
11125 is_gimple_call before calling gimple_call_internal_p.
11127 2014-01-09 Steve Ellcey <sellcey@mips.com>
11129 * Makefile.in (TREE_FLOW_H): Remove.
11130 (TREE_SSA_H): Add file names from tree-flow.h.
11131 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
11132 * tree.h: Remove tree-flow.h reference.
11133 * hash-table.h: Remove tree-flow.h reference.
11134 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
11135 reference with tree-ssa-loop.h.
11137 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11139 * doc/invoke.texi: Add -maltivec={be,le} options, and document
11140 default element-order behavior for -maltivec.
11141 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
11142 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
11143 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
11144 when targeting big endian, at least for now.
11145 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
11147 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11149 PR middle-end/47735
11150 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
11151 var satisfies use_register_for_decl, just take into account type
11152 alignment, rather than decl alignment.
11154 PR tree-optimization/59622
11155 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
11156 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
11157 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
11158 Don't devirtualize for inplace at all. For targets.length () == 1,
11159 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
11161 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
11163 * config/i386/i386.md (cpu): Remove the unused btver1.
11165 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
11167 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
11169 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11172 * tree-core.h (struct target_globals): New forward declaration.
11173 (struct tree_target_option): Add globals field.
11174 * tree.h (TREE_TARGET_GLOBALS): Define.
11175 (prepare_target_option_nodes_for_pch): New prototype.
11176 * target-globals.h (struct target_globals): Define even if
11177 !SWITCHABLE_TARGET.
11178 * tree.c (prepare_target_option_node_for_pch,
11179 prepare_target_option_nodes_for_pch): New functions.
11180 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
11181 * config/i386/i386.c: Include target-globals.h.
11182 (ix86_set_current_function): Instead of doing target_reinit
11183 unconditionally, use save_target_globals_default_opts and
11184 restore_target_globals.
11186 2014-01-09 Richard Biener <rguenther@suse.de>
11188 PR tree-optimization/59715
11189 * tree-cfg.h (split_critical_edges): Declare.
11190 * tree-cfg.c (split_critical_edges): Export.
11191 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
11193 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
11195 * cfgexpand.c (expand_stack_vars): Optionally disable
11196 asan stack protection.
11197 (expand_used_vars): Likewise.
11198 (partition_stack_vars): Likewise.
11199 * asan.c (asan_emit_stack_protection): Optionally disable
11200 after return stack usage.
11201 (instrument_derefs): Optionally disable memory access instrumentation.
11202 (instrument_builtin_call): Likewise.
11203 (instrument_strlen_call): Likewise.
11204 (asan_protect_global): Optionally disable global variables protection.
11205 * doc/invoke.texi: Added doc for new options.
11206 * params.def: Added new options.
11207 * params.h: Likewise.
11209 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11211 PR rtl-optimization/59724
11212 * ifcvt.c (cond_exec_process_if_block): Don't call
11213 flow_find_head_matching_sequence with 0 longest_match.
11214 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
11215 non-active insns if !stop_after.
11216 (try_head_merge_bb): Revert 2014-01-07 changes.
11218 2014-01-08 Jeff Law <law@redhat.com>
11220 * ree.c (get_sub_rtx): New function, extracted from...
11221 (merge_def_and_ext): Here.
11222 (combine_reaching_defs): Use get_sub_rtx.
11224 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
11226 * cgraph.h (varpool_variable_node): Do not choke on null node.
11228 2014-01-08 Catherine Moore <clm@codesourcery.com>
11230 * config/mips/mips.md (simple_return): Attempt to use JRC
11232 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
11234 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
11236 PR rtl-optimization/59137
11237 * reorg.c (steal_delay_list_from_target): Call update_block for
11239 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
11241 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11243 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
11244 two duplicate entries.
11246 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
11249 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
11251 * config/mips/mips.c (mips_truncated_op_cost): New function.
11252 (mips_rtx_costs): Adjust test for BADDU.
11253 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
11255 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
11257 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
11258 (*baddu_si): ...this new pattern.
11260 2014-01-08 Jakub Jelinek <jakub@redhat.com>
11263 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
11265 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
11267 PR middle-end/57748
11268 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
11270 (expand_expr, expand_normal): Adjust.
11271 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
11272 inner_reference_p. Use inner_reference_p to expand inner references.
11273 (store_expr): Adjust.
11274 * cfgexpand.c (expand_call_stmt): Adjust.
11276 2014-01-08 Rong Xu <xur@google.com>
11278 * gcov-io.c (gcov_var): Move from gcov-io.h.
11279 (gcov_position): Ditto.
11280 (gcov_is_error): Ditto.
11281 (gcov_rewrite): Ditto.
11282 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
11283 only part to libgcc/libgcov.h.
11285 2014-01-08 Marek Polacek <polacek@redhat.com>
11287 PR middle-end/59669
11288 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
11290 2014-01-08 Marek Polacek <polacek@redhat.com>
11293 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
11295 2014-01-08 Jakub Jelinek <jakub@redhat.com>
11297 PR rtl-optimization/59649
11298 * stor-layout.c (get_mode_bounds): For BImode return
11299 0 and STORE_FLAG_VALUE.
11301 2014-01-08 Richard Biener <rguenther@suse.de>
11303 PR middle-end/59630
11304 * gimple.h (is_gimple_builtin_call): Remove.
11305 (gimple_builtin_call_types_compatible_p): New.
11306 (gimple_call_builtin_p): New overload.
11307 * gimple.c (is_gimple_builtin_call): Remove.
11308 (validate_call): Rename to ...
11309 (gimple_builtin_call_types_compatible_p): ... this and export. Also
11310 check return types.
11311 (validate_type): New static function.
11312 (gimple_call_builtin_p): New overload and adjust.
11313 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
11314 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
11315 (gimple_fold_stmt_to_constant_1): Likewise.
11316 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
11318 2014-01-08 Richard Biener <rguenther@suse.de>
11320 PR middle-end/59471
11321 * gimplify.c (gimplify_expr): Gimplify register-register type
11322 VIEW_CONVERT_EXPRs to separate stmts.
11324 2014-01-07 Jeff Law <law@redhat.com>
11326 PR middle-end/53623
11327 * ree.c (combine_set_extension): Handle case where source
11328 and destination registers in an extension insn are different.
11329 (combine_reaching_defs): Allow source and destination registers
11330 in extension to be different under limited circumstances.
11331 (add_removable_extension): Remove restriction that the
11332 source and destination registers in the extension are the same.
11333 (find_and_remove_re): Emit a copy from the extension's
11334 destination to its source after the defining insn if
11335 the source and destination registers are different.
11337 PR middle-end/59285
11338 * ifcvt.c (merge_if_block): If we are merging a block with more than
11339 one successor with a block with no successors, remove any BARRIER
11340 after the second block.
11342 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
11344 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
11346 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
11349 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
11350 for 14-bit register offsets when INT14_OK_STRICT is false.
11352 2014-01-07 Roland Stigge <stigge@antcom.de>
11353 Michael Meissner <meissner@linux.vnet.ibm.com>
11356 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
11357 Only check TFmode for SPE constants. Don't check TImode or TDmode.
11359 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
11361 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
11364 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
11366 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
11367 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
11368 rtx is const0_rtx or not.
11370 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
11373 * target-globals.c (save_target_globals): Remove this_fn_optab
11375 * toplev.c: Include optabs.h.
11376 (target_reinit): Temporarily restore the global options if another
11377 set of options are in force.
11379 2014-01-07 Jakub Jelinek <jakub@redhat.com>
11381 PR rtl-optimization/58668
11382 * cfgcleanup.c (flow_find_cross_jump): Don't count
11383 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
11384 to determine what is counted.
11385 (flow_find_head_matching_sequence): Use active_insn_p to determine
11387 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
11389 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
11390 determine what is counted.
11392 PR tree-optimization/59643
11393 * tree-predcom.c (split_data_refs_to_components): If one dr is
11394 read and one write, determine_offset fails and the write isn't
11395 in the bad component, just put the read into the bad component.
11397 2014-01-07 Mike Stump <mikestump@comcast.net>
11398 Jakub Jelinek <jakub@redhat.com>
11401 * tree-core.h (struct tree_optimization_option): Change optabs
11402 type from unsigned char * to void *.
11403 * optabs.c (init_tree_optimization_optabs): Adjust
11404 TREE_OPTIMIZATION_OPTABS initialization.
11406 2014-01-06 Jakub Jelinek <jakub@redhat.com>
11409 * config/i386/i386.h (struct machine_function): Add
11410 no_drap_save_restore field.
11411 * config/i386/i386.c (ix86_save_reg): Use
11412 !cfun->machine->no_drap_save_restore instead of
11413 crtl->stack_realign_needed.
11414 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
11415 this function clears frame_pointer_needed. Set
11416 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
11417 and DRAP reg is needed.
11419 2014-01-06 Marek Polacek <polacek@redhat.com>
11422 * doc/implement-c.texi: Mention that other integer types are
11423 permitted as bit-field types in strictly conforming mode.
11425 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
11427 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
11428 is newly allocated.
11430 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
11432 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
11434 2014-01-06 Martin Jambor <mjambor@suse.cz>
11437 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
11439 * ipa-prop.c (ipa_print_node_params): Fix indentation.
11441 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
11445 * var-tracking.c (add_stores): Preserve the value of the source even if
11446 we don't record the store.
11448 2014-01-06 Terry Guo <terry.guo@arm.com>
11450 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
11452 2014-01-05 Iain Sandoe <iain@codesourcery.com>
11455 * config/darwin.c (darwin_function_section): Adjust return values to
11456 correspond to optimisation changes made in r206070.
11458 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
11460 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
11461 from prefetch_block tune setting.
11462 (nocona_cost): Correct size of prefetch block to 64.
11464 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
11466 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
11467 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
11468 used to save the static chain register in the computation of the offset
11469 from which the FP registers need to be restored.
11471 2014-01-04 Jakub Jelinek <jakub@redhat.com>
11473 PR tree-optimization/59519
11474 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
11475 ICE if get_current_def (current_new_name) is already non-NULL, as long
11476 as it is a phi result of some other phi in *new_exit_bb that has
11479 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
11480 or vmovdqu* for misaligned_operand.
11481 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
11482 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
11483 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
11484 aligned_mem for AVX512F masked aligned load and store builtins and for
11485 non-temporal moves.
11487 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
11489 PR tree-optimization/59651
11490 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
11491 Address range for negative step should be added by TYPE_SIZE_UNIT.
11493 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
11495 * config/m68k/m68k.c (handle_move_double): Handle pushes with
11496 overlapping registers also for registers other than the stack pointer.
11498 2014-01-03 Marek Polacek <polacek@redhat.com>
11501 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
11504 2014-01-03 Jakub Jelinek <jakub@redhat.com>
11507 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
11510 * config/i386/i386.md (MODE_SIZE): New mode attribute.
11511 (push splitter): Use <P:MODE_SIZE> instead of
11512 GET_MODE_SIZE (<P:MODE>mode).
11513 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
11514 (mov -1, reg peephole2): Likewise.
11515 * config/i386/sse.md (*mov<mode>_internal,
11516 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
11517 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
11518 *<code><mode>3, *andnot<mode>3<mask_name>,
11519 <mask_codefor><code><mode>3<mask_name>): Likewise.
11520 * config/i386/subst.md (mask_mode512bit_condition,
11521 sd_mask_mode512bit_condition): Likewise.
11523 2014-01-02 Xinliang David Li <davidxl@google.com>
11525 PR tree-optimization/59303
11526 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
11527 (dump_predicates): Better output format.
11528 (pred_equal_p): New function.
11529 (is_neq_relop_p): Ditto.
11530 (is_neq_zero_form_p): Ditto.
11531 (pred_expr_equal_p): Ditto.
11532 (pred_neg_p): Ditto.
11533 (simplify_pred): Ditto.
11534 (simplify_preds_2): Ditto.
11535 (simplify_preds_3): Ditto.
11536 (simplify_preds_4): Ditto.
11537 (simplify_preds): Ditto.
11538 (push_pred): Ditto.
11539 (push_to_worklist): Ditto.
11540 (get_pred_info_from_cmp): Ditto.
11541 (is_degenerated_phi): Ditto.
11542 (normalize_one_pred_1): Ditto.
11543 (normalize_one_pred): Ditto.
11544 (normalize_one_pred_chain): Ditto.
11545 (normalize_preds): Ditto.
11546 (normalize_cond_1): Remove function.
11547 (normalize_cond): Ditto.
11548 (is_gcond_subset_of): Ditto.
11549 (is_subset_of_any): Ditto.
11550 (is_or_set_subset_of): Ditto.
11551 (is_and_set_subset_of): Ditto.
11552 (is_norm_cond_subset_of): Ditto.
11553 (pred_chain_length_cmp): Ditto.
11554 (convert_control_dep_chain_into_preds): Type change.
11555 (find_predicates): Ditto.
11556 (find_def_preds): Ditto.
11557 (destroy_predicates_vecs): Ditto.
11558 (find_matching_predicates_in_rest_chains): Ditto.
11559 (use_pred_not_overlap_with_undef_path_pred): Ditto.
11560 (is_pred_expr_subset): Ditto.
11561 (is_pred_chain_subset_of): Ditto.
11562 (is_included_in): Ditto.
11563 (is_superset_of): Ditto.
11565 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
11567 Update copyright years.
11569 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
11571 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
11572 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
11573 config/arc/arc.md, config/arc/arc.opt,
11574 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
11575 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
11576 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
11577 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
11578 config/linux-protos.h, config/linux.c, config/winnt-c.c,
11579 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
11580 vtable-verify.c, vtable-verify.h: Use the standard form for the
11583 2014-01-02 Tobias Burnus <burnus@net-b.de>
11585 * gcc.c (process_command): Update copyright notice dates.
11586 * gcov-dump.c: Ditto.
11588 * doc/cpp.texi: Bump @copying's copyright year.
11589 * doc/cppinternals.texi: Ditto.
11590 * doc/gcc.texi: Ditto.
11591 * doc/gccint.texi: Ditto.
11592 * doc/gcov.texi: Ditto.
11593 * doc/install.texi: Ditto.
11594 * doc/invoke.texi: Ditto.
11596 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11598 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
11600 2014-01-01 Jakub Jelinek <jakub@redhat.com>
11602 * config/i386/sse.md (*mov<mode>_internal): Guard
11603 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
11605 PR rtl-optimization/59647
11606 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
11607 new_rtx into UNSIGNED_FLOAT rtxes.
11609 Copyright (C) 2014 Free Software Foundation, Inc.
11611 Copying and distribution of this file, with or without modification,
11612 are permitted in any medium without royalty provided the copyright
11613 notice and this notice are preserved.