1 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
3 * ipa.c (symtab_remove_unreachable_nodes): Remove
4 symbol from comdat group if its body was eliminated.
5 (comdat_can_be_unshared_p_1): Static symbols can always
7 * symtab.c (symtab_remove_from_same_comdat_group): Break out
9 (symtab_unregister_node): ... this one.
10 (verify_symtab_base): More strict checking of comdats.
11 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
13 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
15 * tree-pass.h (make_pass_ipa_comdats): New pass.
16 * timevar.def (TV_IPA_COMDATS): New timevar.
17 * passes.def (pass_ipa_comdats): Add.
18 * Makefile.in (OBJS): Add ipa-comdats.o
19 * ipa-comdats.c: New file.
21 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
23 * ipa.c (update_visibility_by_resolution_info): New function.
24 (function_and_variable_visibility): Use it.
26 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
28 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
30 (FOR_EACH_DEFINED_SYMBOL): New macro.
31 (varpool_first_static_initializer, varpool_next_static_initializer,
32 varpool_first_defined_variable, varpool_next_defined_variable): Fix comments.
33 (symtab_in_same_comdat_p): Correctly deal with inline functions.
35 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
37 * ggc-page.c (ggc_handle_finalizers): Add comment.
39 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
41 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
42 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
43 (ggc_internal_cleared_alloc): Likewise.
44 * ggc-page.c (finalizer): New class.
45 (vec_finalizer): Likewise.
46 (globals::finalizers): New member.
47 (globals::vec_finalizers): Likewise.
48 (ggc_internal_alloc): Record the finalizer if any for the block being
50 (ggc_handle_finalizers): New function.
51 (ggc_collect): Call ggc_handle_finalizers.
52 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
54 (ggc_internal_cleared_alloc): Likewise.
55 (finalize): New function.
56 (need_finalization_p): Likewise.
57 (ggc_alloc): Install the type's destructor as the finalizer if it
59 (ggc_cleared_alloc): Likewise.
60 (ggc_vec_alloc): Likewise.
61 (ggc_cleared_vec_alloc): Likewise.
63 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
65 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
67 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
69 * alias.c (record_alias_subset): Adjust.
70 * bitmap.c (bitmap_element_allocate): Likewise.
71 (bitmap_gc_alloc_stat): Likewise.
72 * cfg.c (init_flow): Likewise.
73 (alloc_block): Likewise.
74 (unchecked_make_edge): Likewise.
75 * cfgloop.c (alloc_loop): Likewise.
76 (flow_loops_find): Likewise.
77 (rescan_loop_exit): Likewise.
78 * cfgrtl.c (init_rtl_bb_info): Likewise.
79 * cgraph.c (insert_new_cgraph_node_version): Likewise.
80 (cgraph_allocate_node): Likewise.
81 (cgraph_create_edge_1): Likewise.
82 (cgraph_allocate_init_indirect_info): Likewise.
83 * cgraphclones.c (cgraph_clone_edge): Likewise.
84 * cgraphunit.c (add_asm_node): Likewise.
85 (init_lowered_empty_function): Likewise.
86 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
87 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
88 (alpha_use_linkage): Likewise.
89 * config/arc/arc.c (arc_init_machine_status): Likewise.
90 * config/arm/arm.c (arm_init_machine_status): Likewise.
91 * config/avr/avr.c (avr_init_machine_status): Likewise.
92 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
93 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
94 * config/cris/cris.c (cris_init_machine_status): Likewise.
95 * config/darwin.c (machopic_indirection_name): Likewise.
96 (darwin_build_constant_cfstring): Likewise.
97 (darwin_enter_string_into_cfstring_table): Likewise.
98 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
99 * config/frv/frv.c (frv_init_machine_status): Likewise.
100 * config/i386/i386.c (get_dllimport_decl): Likewise.
101 (ix86_init_machine_status): Likewise.
102 (assign_386_stack_local): Likewise.
103 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
104 (i386_pe_maybe_record_exported_symbol): Likewise.
105 (i386_pe_record_stub): Likewise.
106 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
107 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
108 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
109 (m32c_note_pragma_address): Likewise.
110 * config/mep/mep.c (mep_init_machine_status): Likewise.
111 (mep_note_pragma_flag): Likewise.
112 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
113 (mips16_local_alias): Likewise.
114 (mips_init_machine_status): Likewise.
115 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
116 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
117 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
118 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
119 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
120 * config/pa/pa.c (pa_init_machine_status): Likewise.
121 (pa_get_deferred_plabel): Likewise.
122 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
123 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
124 (rs6000_init_machine_status): Likewise.
125 (output_toc): Likewise.
126 * config/s390/s390.c (s390_init_machine_status): Likewise.
127 * config/score/score.c (score_output_external): Likewise.
128 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
129 * config/spu/spu.c (spu_init_machine_status): Likewise.
130 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
131 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
132 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
133 * coverage.c (coverage_end_function): Likewise.
134 * dbxout.c (dbxout_init): Likewise.
135 * doc/gty.texi: Don't mention variable_size attribute.
136 * dwarf2cfi.c (new_cfi): Adjust.
137 (new_cfi_row): Likewise.
138 (copy_cfi_row): Likewise.
139 (create_cie_data): Likewise.
140 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
141 (new_loc_descr): Likewise.
142 (find_AT_string_in_table): Likewise.
143 (add_addr_table_entry): Likewise.
145 (add_var_loc_to_decl): Likewise.
146 (clone_die): Likewise.
147 (clone_as_declaration): Likewise.
148 (break_out_comdat_types): Likewise.
149 (new_loc_list): Likewise.
150 (add_loc_descr_to_each): Likewise.
151 (add_location_or_const_value_attribute): Likewise.
152 (add_linkage_name): Likewise.
153 (lookup_filename): Likewise.
154 (dwarf2out_var_location): Likewise.
155 (new_line_info_table): Likewise.
156 (dwarf2out_init): Likewise.
157 (mem_loc_descriptor): Likewise.
158 (loc_descriptor): Likewise.
159 (add_const_value_attribute): Likewise.
160 (tree_add_const_value_attribute): Likewise.
161 (comp_dir_string): Likewise.
162 (dwarf2out_vms_debug_main_pointer): Likewise.
163 (string_cst_pool_decl): Likewise.
164 * emit-rtl.c (set_mem_attrs): Likewise.
165 (get_reg_attrs): Likewise.
166 (start_sequence): Likewise.
167 (init_emit): Likewise.
168 (init_emit_regs): Likewise.
169 * except.c (init_eh_for_function): Likewise.
170 (gen_eh_region): Likewise.
171 (gen_eh_region_catch): Likewise.
172 (gen_eh_landing_pad): Likewise.
173 (add_call_site): Likewise.
174 * function.c (add_frame_space): Likewise.
175 (insert_temp_slot_address): Likewise.
176 (assign_stack_temp_for_type): Likewise.
177 (get_hard_reg_initial_val): Likewise.
178 (allocate_struct_function): Likewise.
179 (prepare_function_start): Likewise.
180 (types_used_by_var_decl_insert): Likewise.
181 * gengtype.c (variable_size_p): Remove function.
182 (enum alloc_quantity): Remove enum.
183 (write_typed_alloc_def): Remove function.
184 (write_typed_struct_alloc_def): Likewise.
185 (write_typed_typedef_alloc_def): Likewise.
186 (write_typed_alloc_defns): Likewise.
188 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
189 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
190 * ggc.h (ggc_alloc): new function.
191 (ggc_cleared_alloc): Likewise.
192 (ggc_vec_alloc): Template on type of vector element, and remove
193 element size argument.
194 (ggc_cleared_vec_alloc): Likewise.
195 * gimple.c (gimple_build_omp_for): Adjust.
196 (gimple_copy): Likewise.
197 * ipa-cp.c (get_replacement_map): Likewise.
198 (find_aggregate_values_for_callers_subset): Likewise.
199 (known_aggs_to_agg_replacement_list): Likewise.
200 * ipa-devirt.c (get_odr_type): Likewise.
201 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
202 (read_agg_replacement_chain): Likewise.
203 * loop-iv.c (get_simple_loop_desc): Likewise.
204 * lto-cgraph.c (input_node_opt_summary): Likewise.
205 * lto-section-in.c (lto_new_in_decl_state): Likewise.
206 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
207 (input_eh_region): Likewise.
208 (input_eh_lp): Likewise.
209 (input_cfg): Likewise.
210 * optabs.c (set_optab_libfunc): Likewise.
211 (init_tree_optimization_optabs): Likewise.
212 (set_conv_libfunc): Likewise.
213 * passes.c (do_per_function_toporder): Likewise.
214 * rtl.h: Don't use variable_size gty attribute.
215 * sese.c (if_region_set_false_region): Adjust.
216 * stringpool.c (gt_pch_save_stringpool): Likewise.
217 * target-globals.c (save_target_globals): Likewise.
218 * toplev.c (general_init): Likewise.
219 * trans-mem.c (record_tm_replacement): Likewise.
220 (split_bb_make_tm_edge): Likewise.
221 * tree-cfg.c (move_sese_region_to_fn): Likewise.
222 * tree-data-ref.h (lambda_vector_new): Likewise.
223 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
224 * tree-iterator.c (tsi_link_before): Likewise.
225 (tsi_link_after): Likewise.
226 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
227 * tree-ssa-loop-niter.c (record_estimate): Likewise.
228 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
229 * tree-ssa-operands.h: Don't use variable_size gty attribute.
230 * tree-ssa.c (init_tree_ssa): Adjust.
231 * tree-ssanames.c (set_range_info): Likewise.
232 (get_ptr_info): Likewise.
233 (duplicate_ssa_name_ptr_info): Likewise.
234 (duplicate_ssa_name_range_info): Likewise.
235 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
236 (unpack_ts_fixed_cst_value_fields): Likewise.
237 * tree.c (build_fixed): Likewise.
238 (build_real): Likewise.
239 (build_string): Likewise.
240 (decl_priority_info): Likewise.
241 (decl_debug_expr_insert): Likewise.
242 (decl_value_expr_insert): Likewise.
243 (decl_debug_args_insert): Likewise.
244 (type_hash_add): Likewise.
245 (build_omp_clause): Likewise.
246 * ubsan.c (decl_for_type_insert): Likewise.
247 * varasm.c (get_unnamed_section): Likewise.
248 (get_noswitch_section): Likewise.
249 (get_section): Likewise.
250 (get_block_for_section): Likewise.
251 (create_block_symbol): Likewise.
252 (build_constant_desc): Likewise.
253 (create_constant_pool): Likewise.
254 (force_const_mem): Likewise.
255 (record_tm_clone_pair): Likewise.
256 * varpool.c (varpool_create_empty_node): Likewise.
258 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
260 * dwarf2out.c (tree_add_const_value_attribute): Call
261 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
262 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
263 instead of ggc_internal_<x>alloc_stat.
264 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
265 (ggc_realloc): Likewise.
266 * ggc-none.c (ggc_internal_alloc): Likewise.
267 (ggc_internal_cleared_alloc): Likewise.
268 * ggc-page.c: Likewise.
269 * ggc.h (ggc_internal_alloc_stat): Likewise.
270 (ggc_internal_alloc): Remove macro.
271 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
272 (ggc_internal_cleared_alloc): Remove macro.
273 (GGC_RESIZEVEC): Adjust.
274 (ggc_resizevar): Remove macro.
275 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
276 (ggc_internal_cleared_vec_alloc_stat): Likewise.
277 (ggc_internal_vec_cleared_alloc): Remove macro.
278 (ggc_alloc_atomic_stat): Drop _stat suffix.
279 (ggc_alloc_atomic): Remove macro.
280 (ggc_alloc_cleared_atomic): Remove macro.
281 (ggc_alloc_string_stat): Drop _stat suffix.
282 (ggc_alloc_string): Remove macro.
283 (ggc_alloc_rtx_def_stat): Adjust.
284 (ggc_alloc_tree_node_stat): Likewise.
285 (ggc_alloc_cleared_tree_node_stat): Likewise.
286 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
287 (ggc_alloc_cleared_simd_clone_stat): Likewise.
288 * gimple.c (gimple_build_omp_for): Likewise.
289 (gimple_copy): Likewise.
290 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
291 * toplev.c (realloc_for_line_map): Adjust.
292 * tree-data-ref.h (lambda_vector_new): Likewise.
293 * tree-phinodes.c (allocate_phi_node): Likewise.
294 * tree.c (grow_tree_vec_stat): Likewise.
295 * vec.h (va_gc::reserve): Adjust.
297 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
299 * config/microblaze/microblaze.c (break_handler): New Declaration.
300 (microblaze_break_function_p,microblaze_is_break_handler): New.
301 (compute_frame_size): Use microblaze_break_function_p.
302 Add the test of break_handler.
303 (microblaze_function_prologue) : Add the test of variable
304 break_handler. Check the fnname by BREAK_HANDLER_NAME.
305 (microblaze_function_epilogue) : Add the test of break_handler.
306 (microblaze_globalize_label) : Add the test of break_handler.
307 Check the name by BREAK_HANDLER_NAME.
309 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
311 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
312 microblaze_is_break_handler test.
313 (call_internal1,call_value_intern): Use microblaze_break_function_p.
316 * config/microblaze/microblaze-protos.h
317 (microblaze_break_function_p,microblaze_is_break_handler):
320 * doc/extend.texi (MicroBlaze break_handler Functions): Document
321 new MicroBlaze break_handler functions.
323 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
325 * doc/extend.texi (Size of an asm): Move node text according
326 to its @menu entry position.
328 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
330 PR tree-optimization/61140
331 PR tree-optimization/61150
332 PR tree-optimization/61197
333 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
335 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
337 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
339 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
341 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
342 __SIZEOF_INT128__ is defined.
344 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
346 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
347 (rs6000_delegitimize_address): Use it.
349 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
351 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
352 inplace argument. Store the new address in the original MEM when true.
353 * emit-rtl.c (change_address_1): Likewise.
354 (adjust_address_1, adjust_automodify_address_1, offset_address):
356 * rtl.h (plus_constant): Add an inplace argument.
357 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
358 when true. Avoid generating (plus X (const_int 0)).
359 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
360 in-place. Pass true to plus_constant.
361 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
363 2014-05-16 Dehao Chen <dehao@google.com>
365 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
367 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
370 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
372 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
374 2014-05-16 Dehao Chen <dehao@google.com>
376 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
377 optimize_function_for_size_p.
378 * regs.h (REG_FREQ_FROM_BB): Likewise.
380 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
383 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
384 negt_reg_operand cases.
385 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
387 * config/sh/predicates.md (cbranch_treg_value): Simplify.
389 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
391 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
394 2014-05-16 David Malcolm <dmalcolm@redhat.com>
397 2014-04-29 David Malcolm <dmalcolm@redhat.com>
399 * tree-cfg.c (dump_function_to_file): Dump the return type of
400 functions, in a line to itself before the function body, mimicking
401 the layout of a C function.
403 2014-05-16 Dehao Chen <dehao@google.com>
405 * cfghooks.c (make_forwarder_block): Use direct computation to
406 get fall-through edge's count and frequency.
408 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
410 * config/arc/arc.c (arc_init): Fix typo in error message.
411 * config/i386/i386.c (ix86_expand_builtin): Likewise.
412 (split_stack_prologue_scratch_regno): Likewise.
413 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
414 word from error message.
416 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
418 * ira-costs.c: Fix typo in comment.
420 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
422 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
424 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
426 * varpool.c (dump_varpool_node): Dump write-only flag.
427 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
429 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
430 write-only variables.
431 * ipa.c (process_references): New function.
432 (set_readonly_bit): New function.
433 (set_writeonly_bit): New function.
434 (clear_addressable_bit): New function.
435 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
436 fix handling of aliases.
437 * cgraph.h (struct varpool_node): Add writeonly flag.
439 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
441 PR rtl-optimization/60969
442 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
443 Calculate costs for this case.
445 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
447 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
448 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
450 2014-05-16 Richard Biener <rguenther@suse.de>
452 PR tree-optimization/61194
453 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
454 bool patterns ending in a COND_EXPR.
456 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
458 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
460 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
462 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
463 where we were unable to cost an RTX.
465 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
467 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
470 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
471 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
473 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
475 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
476 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
478 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
479 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
481 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
482 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
484 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
487 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
488 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
490 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
493 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
494 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
496 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
497 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
499 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
500 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
502 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
505 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
506 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
508 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
509 ZERO_EXTEND and SIGN_EXTEND better.
511 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
512 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
514 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
517 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
518 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
520 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
521 costs when costing loads and stores to memory.
523 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
524 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
526 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
529 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
531 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
533 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
534 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
536 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
538 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
540 (aarch64_rtx_mult_cost): New.
541 (aarch64_rtx_costs): Use it, refactor as appropriate.
543 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
544 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
546 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
547 emit instructions, return number of instructions which would
549 (aarch64_add_constant): Update call to aarch64_build_constant.
550 (aarch64_output_mi_thunk): Likewise.
551 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
554 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
556 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
557 (TARGET_RTX_COSTS): Call it.
559 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
561 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
562 (cortexa57_vector_cost): Likewise.
563 (cortexa57_tunings): Use them.
565 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
567 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
568 (cpu_addrcost_table): Use it.
569 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
570 (aarch64_address_cost): Rewrite using aarch64_classify_address,
573 2014-05-16 Richard Biener <rguenther@suse.de>
575 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
576 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
577 (visit_phi): Ignore edges marked as not executable.
578 (class cond_dom_walker): New.
579 (cond_dom_walker::before_dom_children): Value-number
580 control statements and mark successor edges as not
581 executable if possible.
582 (run_scc_vn): First walk all control statements in
583 dominator order, marking edges as not executable.
584 * tree-inline.c (copy_edges_for_bb): Be not confused
585 about random edge flags.
587 2014-05-16 Richard Biener <rguenther@suse.de>
589 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
591 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
594 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
595 (__TM_simple_begin): Use it.
596 (__TM_begin): Likewise.
598 2014-05-15 Martin Jambor <mjambor@suse.cz>
601 * ipa-prop.c (update_indirect_edges_after_inlining): Check
602 type_preserved flag when the indirect edge is polymorphic.
604 2014-05-15 Martin Jambor <mjambor@suse.cz>
606 PR tree-optimization/61090
607 * tree-sra.c (sra_modify_expr): Pass the current gsi to
610 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
612 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
613 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
615 2014-05-15 Jakub Jelinek <jakub@redhat.com>
617 PR tree-optimization/61158
618 * fold-const.c (fold_binary_loc): If X is zero-extended and
619 shiftc >= prec, make sure zerobits is all ones instead of
620 invoking undefined behavior.
622 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
624 * regcprop.h: New file.
625 * regcprop.c (skip_debug_insn_p): New decl.
626 (replace_oldest_value_reg): Check skip_debug_insn_p.
627 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
628 * shrink-wrap.c: Include regcprop.h.
629 (prepare_shrink_wrap): Call
630 copyprop_hardreg_forward_bb_without_debug_insn.
632 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
634 * shrink-wrap.h: Update comment.
635 * shrink-wrap.c: Update comment.
636 (next_block_for_reg): Rename to live_edge_for_reg.
637 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
638 (move_insn_for_shrink_wrap): Split live_edge.
639 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
641 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
643 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
645 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
646 * config/sparc/sparc.md (fptype_ut699): New attribute.
647 (in_branch_delay): Return false if -mfix-ut699 is specified and
648 fptype_ut699 is set to single.
649 (truncdfsf2): Add fptype_ut699 attribute.
650 (fix_truncdfsi2): Likewise.
651 (floatsisf2): Change fptype attribute.
652 (fix_truncsfsi2): Likewise.
653 (negtf2_notv9): Delete.
654 (negtf2_v9): Likewise.
655 (negtf2_hq): New instruction.
656 (negtf2): New instruction and splitter.
657 (negdf2_notv9): Rewrite.
658 (abstf2_notv9): Delete.
659 (abstf2_hq_v9): Likewise.
660 (abstf2_v9): Likewise.
661 (abstf2_hq): New instruction.
662 (abstf2): New instruction and splitter.
663 (absdf2_notv9): Rewrite.
665 2014-05-14 Cary Coutant <ccoutant@google.com>
668 * opts.c (common_handle_option): Don't special-case "-g".
669 (set_debug_level): Default to at least level 2 with "-g".
671 2014-05-14 DJ Delorie <dj@redhat.com>
673 * config/msp430/msp430.c (msp430_builtin): Add
674 MSP430_BUILTIN_DELAY_CYCLES.
675 (msp430_init_builtins): Register void __delay_cycles(long long).
676 (msp430_builtin_decl): Add it.
677 (cg_magic_constant): New.
678 (msp430_expand_delay_cycles): New.
679 (msp430_expand_builtin): Call it.
680 (msp430_print_operand_raw): Change integer printing from "int" to
682 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
683 (delay_cycles_start): New.
684 (delay_cycles_end): New.
685 (delay_cycles_32): New.
686 (delay_cycles_32x): New.
687 (delay_cycles_16): New.
688 (delay_cycles_16x): New.
689 (delay_cycles_2): New.
690 (delay_cycles_1): New.
691 * doc/extend.texi: Document __delay_cycles().
693 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
695 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
696 length attribute computation.
698 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
701 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
703 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
706 * config/sparc/sparc.md: Fix types of low and high in DI constant
707 splitter. Use gen_int_mode in some other splitters.
709 2014-05-14 Martin Jambor <mjambor@suse.cz>
712 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
714 2014-05-14 James Norris <jnorris@codesourcery.com>
716 * omp-low.c (expand_parallel_call): Remove shadow variable.
717 (expand_omp_taskreg): Likewise.
719 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
721 * common/config/i386/i386-common.c
722 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
723 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
724 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
725 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
726 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
727 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
728 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
729 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
730 xsavecintrin.h, xsavesintrin.h.
732 * config/i386/clflushoptintrin.h: New.
733 * config/i386/xsavecintrin.h: Ditto.
734 * config/i386/xsavesintrin.h: Ditto.
735 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
738 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
739 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
741 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
742 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
743 OPTION_MASK_ISA_XSAVES.
744 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
746 (PTA_CLFLUSHOPT) Define.
749 (ix86_option_override_internal): Handle new options.
750 (ix86_valid_target_attribute_inner_p): Ditto.
751 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
752 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
753 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
754 (bdesc_special_args): Add __builtin_ia32_xsaves,
755 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
756 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
757 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
758 (ix86_expand_builtin): Handle new builtins.
759 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
760 (TARGET_CLFLUSHOPT_P): Ditto.
761 (TARGET_XSAVEC): Ditto.
762 (TARGET_XSAVEC_P): Ditto.
763 (TARGET_XSAVES): Ditto.
764 (TARGET_XSAVES_P): Ditto.
765 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
766 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
767 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
769 (ANY_XRSTOR64): Ditto.
771 (xrstor): Change into <xrstor>.
772 (xrstor_rex64): Change into <xrstor>_rex64.
773 (xrstor64): Change into <xrstor>64
775 * config/i386/i386.opt (mclflushopt): New.
778 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
780 * doc/invoke.texi: Document new options.
782 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
784 PR rtl-optimization/60866
785 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
786 Default it to -1. Pass it down to init_simplejump_data.
787 (init_simplejump_data): New parameter old_seqno. Pass it down
788 to get_seqno_for_a_jump.
789 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
790 initializing new jump seqno as a last resort. Add comment.
791 (sel_redirect_edge_and_branch): Save old seqno of the conditional
792 jump and pass it down to sel_init_new_insn.
793 (sel_redirect_edge_and_branch_force): Likewise.
795 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
797 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
798 shifted values to avoid build warning.
800 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
802 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
803 * cfgrtl.c (rtl_merge_blocks): Fix comment.
804 (cfg_layout_merge_blocks): Likewise.
805 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
807 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
809 PR rtl-optimization/60901
810 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
811 bb predecessor belongs to the same scheduling region. Adjust comment.
813 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
815 * doc/sourcebuild.texi: (dfp_hw): Document.
816 (p8vector_hw): Likewise.
817 (powerpc_eabi_ok): Likewise.
818 (powerpc_elfv2): Likewise.
819 (powerpc_htm_ok): Likewise.
820 (ppc_recip_hw): Likewise.
823 2014-05-13 Cary Coutant <ccoutant@google.com>
825 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
827 2014-05-13 David Malcolm <dmalcolm@redhat.com>
829 * gengtype-parse.c (require3): Eliminate in favor of...
831 (require_template_declaration): Update to support optional single *
834 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
835 (create_user_defined_type): Handle a single level of explicit
836 pointerness within template arguments.
837 (struct write_types_data): Add field "kind".
838 (filter_type_name): Handle "*" character.
839 (write_user_func_for_structure_ptr): Require a write_types_data
840 rather than just a prefix string, so that we can look up the kind
841 of the wtd and use it as an index into wrote_user_func_for_ptr,
842 ensuring that such functions are written at most once. Support
843 subclasses by invoking the marking function of the ultimate base class.
844 (write_user_func_for_structure_body): Require a write_types_data
845 rather than just a prefix string, so that we can pass this to
846 write_user_func_for_structure_ptr.
847 (write_func_for_structure): Likewise.
848 (ggc_wtd): Add initializer of new "kind" field.
851 * gengtype.h (enum write_types_kinds): New.
852 (struct type): Add field wrote_user_func_for_ptr to the "s"
855 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
857 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
858 instead of const_binop.
859 (fold_binary_loc): Likewise.
861 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
863 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
864 calculation to match get_ref_base_and_extent.
866 2014-05-13 Catherine Moore <clm@codesourcery.com>
867 Sandra Loosemore <sandra@codesourcery.com>
869 * configure.ac: Fix assembly for explicit JALR relocation check.
870 * configure: Regenerate.
872 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
874 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
875 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
876 Remove associated type declarations and initialisations.
877 (arm_expand_neon_builtin): Likewise.
878 (neon_emit_pair_result_insn): Delete.
879 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
880 * config/arm/neon.md (neon_vtrn<mode>): Delete.
881 (neon_vzip<mode>): Likewise.
882 (neon_vuzp<mode>): Likewise.
884 2014-05-13 Richard Biener <rguenther@suse.de>
887 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
888 it needs revisiting whether the call still may be tail-called.
890 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
892 * rtl.def (SYMBOL_REF): Remove middle "0" field.
893 * rtl.h (block_symbol): Reduce number of fields to 2.
894 (rtx_def): Add u2.symbol_ref_flags.
895 (SYMBOL_REF_FLAGS): Use it.
896 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
897 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
898 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
899 Lower index of SYMBOL_REF_DATA.
900 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
901 Print SYMBOL_REF_FLAGS at the same time.
902 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
904 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
906 * rtl.def (VAR_LOCATION): Remove "i" field.
907 * rtl.h (rtx_def): Add u2.var_location_status.
908 (PAT_VAR_LOCATION_STATUS): Use it.
909 (gen_rtx_VAR_LOCATION): Declare.
910 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
911 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
912 * var-tracking.c (emit_note_insn_var_location): Remove casts.
914 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
916 * rtl.def (scratch): Fix outdated comment and remove "0" field.
917 * gengtype.c (adjust_field_rtx_def): Update accordingly.
919 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
921 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
922 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
923 * rtl.h (rtx_def): Add insn_uid to u2 field.
924 (RTX_FLAG_CHECK8): Delete in favor of...
925 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
926 (INSN_DELETED_P): Update accordingly.
927 (INSN_UID): Use u2.insn_uid.
928 (INSN_CHAIN_CODE_P): Define.
929 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
930 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
931 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
932 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
933 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
935 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
936 Update indices for insn-chain rtxes.
937 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
938 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
939 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
940 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
941 * combine.c (try_combine): Likewise.
942 * ira.c (setup_prohibited_mode_move_regs): Likewise.
944 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
946 * rtl.def (REG): Remove middle field.
947 * rtl.h (rtx_def): Add orignal_regno to u2.
948 (ORIGINAL_REGNO): Use it instead of field 1.
949 (REG_ATTRS): Lower field index accordingly.
950 * gengtype.c (adjust_field_rtx_def): Remove handling of
951 ORIGINAL_REGNO. Move REG_ATTRS index down.
952 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
953 code that prints the REGNO.
955 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
957 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
960 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
962 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
964 2014-05-13 Bin Cheng <bin.cheng@arm.com>
966 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
967 (alloc_iv): Lower base expressions containing ADDR_EXPR.
969 2014-05-13 Ian Bolton <ian.bolton@arm.com>
971 * config/aarch64/aarch64-protos.h
972 (aarch64_hard_regno_caller_save_mode): New prototype.
973 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
975 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
977 2014-05-13 Christian Bruel <christian.bruel@st.com>
979 * target.def (mode_switching): New hook vector.
980 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
981 (mode_exit, modepriority_to_mode): Likewise.
982 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
983 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
984 * target.h: Include tm.h and hard-reg-set.h.
985 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
986 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
987 * doc/tm.texi Regenerate.
988 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
989 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
990 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
991 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
992 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
993 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
994 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
995 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
996 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
997 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
998 (ix86_emit_mode_set): Hookify.
999 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
1001 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
1002 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
1003 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
1004 (epiphany_mode_priority_to_mode): Remove declaration.
1005 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
1006 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
1007 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
1009 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
1010 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
1011 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
1013 2014-05-13 Jakub Jelinek <jakub@redhat.com>
1016 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
1017 is const0_rtx, return immediately. Don't test count == 0 when
1020 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1022 * Makefile.in: add shrink-wrap.o.
1023 * config/i386/i386.c: include "shrink-wrap.h"
1024 * function.c: Likewise.
1025 (requires_stack_frame_p, next_block_for_reg,
1026 move_insn_for_shrink_wrap, prepare_shrink_wrap,
1027 dup_block_and_redirect): Move to shrink-wrap.c
1028 (thread_prologue_and_epilogue_insns): Extract three code segments
1029 as functions in shrink-wrap.c
1030 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
1032 * shrink-wrap.c: New file.
1033 * shrink-wrap.h: New file.
1035 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
1037 * doc/extend.texi: Reflect current numbers of pragmas. Remove
1038 reference to Solaris.
1040 2014-05-12 Mike Stump <mikestump@comcast.net>
1043 * genattrtab.c (filename): Add.
1044 (convert_set_attr_alternative): Improve error message.
1045 (check_defs): Restore read_md_filename for error messages.
1046 (gen_insn): Save filename.
1048 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
1050 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
1051 -fno-local-ivars and -fivar-visibility.
1052 * c-family/c.opt: Make -Wshadow also implicitly enable
1055 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
1057 * doc/tm.texi: Remove reference to deleted macro.
1058 * doc/tm.texi.in: Likewise.
1060 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
1063 * config/avr/avr.c (avr_out_store_psi): Use correct constant
1066 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
1068 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
1069 * config/arm/aout.h (License): Same.
1070 * config/arm/bpabi.h (License): Same.
1071 * config/arm/elf.h (License): Same.
1072 * config/arm/linux-elf.h (License): Same.
1073 * config/arm/linux-gas.h (License): Same.
1074 * config/arm/netbsd-elf.h (License): Same.
1075 * config/arm/uclinux-eabi.h (License): Same.
1076 * config/arm/uclinux-elf.h (License): Same.
1077 * config/arm/vxworks.h (License): Same.
1079 2014-05-11 Jakub Jelinek <jakub@redhat.com>
1081 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
1082 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
1083 number of operands to 3.
1084 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
1085 * tree-nested.c (convert_nonlocal_omp_clauses,
1086 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
1087 * gimplify.c (gimplify_scan_omp_clauses): Handle
1088 OMP_CLAUSE_LINEAR_STMT.
1089 * omp-low.c (lower_rec_input_clauses): Fix typo.
1090 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
1091 cast between Fortran boolean_type_node and C _Bool if
1094 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
1096 PR tree-optimization/61136
1097 * wide-int.h (multiple_of_p): Define a version that doesn't return
1099 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
1100 integer_zerop/const_binop pair.
1101 (multiple_of_p): Likewise, converting both operands to widest_int
1104 2014-05-09 Teresa Johnson <tejohnson@google.com>
1106 * cgraphunit.c (analyze_functions): Use correct dump file.
1108 2014-05-09 Florian Weimer <fweimer@redhat.com>
1110 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
1112 (stack_protect_return_slot_p): New function.
1113 (expand_used_vars): Call stack_protect_decl_p and
1114 stack_protect_return_slot_p for -fstack-protector-strong.
1116 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
1119 * fold-const.c (fold_binary_loc): Changed width of mask.
1121 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
1123 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
1124 unsigned int initializers for regno_in, regno_out.
1126 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
1129 * config/avr/avr.md (cc): Add new attribute set_vzn.
1130 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
1131 Set cc insn attribute to set_vzn instead of set_zn for alternatives
1132 with INC, DEC or NEG.
1133 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
1134 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
1135 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
1137 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1140 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1142 * wide-int.cc (UTItype): Define.
1143 (UDWtype): Define for appropriate W_TYPE_SIZE.
1145 2014-05-09 Richard Biener <rguenther@suse.de>
1147 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
1148 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
1149 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
1150 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
1151 ssa_propagate): Adjust.
1153 2014-05-08 Jeff Law <law@redhat.com>
1155 PR tree-optimization/61009
1156 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
1157 tri-state rather than a boolean. When a block is too big to
1158 thread through, inform caller via negative return value.
1159 (thread_across_edge): If a block was too big for normal threading,
1160 then it's too big for a joiner too, so remove temporary equivalences
1161 and return immediately.
1163 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
1164 Matthias Klose <doko@ubuntu.com>
1167 * optc-gen.awk: Fix option handling for -Wunused-parameter.
1169 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
1172 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
1174 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
1177 * config/alpha/alpha.c: Include gimple-iterator.h.
1178 (alpha_gimple_fold_builtin): New function. Move
1179 ALPHA_BUILTIN_UMULH folding from ...
1180 (alpha_fold_builtin): ... here.
1181 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
1183 2014-05-08 Wei Mi <wmi@google.com>
1186 * config/i386/i386.c (ix86_compute_frame_layout): Update
1187 preferred_stack_boundary for call, expanded from tls descriptor.
1188 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
1189 to depend on SP register.
1190 (*tls_local_dynamic_base_32_gnu): Ditto.
1191 (*tls_local_dynamic_32_once): Ditto.
1192 (tls_global_dynamic_64_<mode>): Set
1193 ix86_tls_descriptor_calls_expanded_in_cfun.
1194 (tls_local_dynamic_base_64_<mode>): Ditto.
1195 (tls_global_dynamic_32): Set
1196 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
1197 to depend on SP register.
1198 (tls_local_dynamic_base_32): Ditto.
1200 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1202 * config/arm/arm_neon.h: Update comment.
1203 * config/arm/neon-docgen.ml: Delete.
1204 * config/arm/neon-gen.ml: Delete.
1205 * doc/arm-neon-intrinsics.texi: Update comment.
1207 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1209 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
1211 (vand, vorr, veor, vorn, vbic): Remove.
1212 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
1214 (neon_vsub_unspec): Likewise.
1215 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
1217 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1219 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
1220 (vadd_s16): Likewise.
1221 (vadd_s32): Likewise.
1222 (vadd_f32): Likewise.
1223 (vadd_u8): Likewise.
1224 (vadd_u16): Likewise.
1225 (vadd_u32): Likewise.
1226 (vadd_s64): Likewise.
1227 (vadd_u64): Likewise.
1228 (vaddq_s8): Likewise.
1229 (vaddq_s16): Likewise.
1230 (vaddq_s32): Likewise.
1231 (vaddq_s64): Likewise.
1232 (vaddq_f32): Likewise.
1233 (vaddq_u8): Likewise.
1234 (vaddq_u16): Likewise.
1235 (vaddq_u32): Likewise.
1236 (vaddq_u64): Likewise.
1237 (vmul_s8): Likewise.
1238 (vmul_s16): Likewise.
1239 (vmul_s32): Likewise.
1240 (vmul_f32): Likewise.
1241 (vmul_u8): Likewise.
1242 (vmul_u16): Likewise.
1243 (vmul_u32): Likewise.
1244 (vmul_p8): Likewise.
1245 (vmulq_s8): Likewise.
1246 (vmulq_s16): Likewise.
1247 (vmulq_s32): Likewise.
1248 (vmulq_f32): Likewise.
1249 (vmulq_u8): Likewise.
1250 (vmulq_u16): Likewise.
1251 (vmulq_u32): Likewise.
1252 (vsub_s8): Likewise.
1253 (vsub_s16): Likewise.
1254 (vsub_s32): Likewise.
1255 (vsub_f32): Likewise.
1256 (vsub_u8): Likewise.
1257 (vsub_u16): Likewise.
1258 (vsub_u32): Likewise.
1259 (vsub_s64): Likewise.
1260 (vsub_u64): Likewise.
1261 (vsubq_s8): Likewise.
1262 (vsubq_s16): Likewise.
1263 (vsubq_s32): Likewise.
1264 (vsubq_s64): Likewise.
1265 (vsubq_f32): Likewise.
1266 (vsubq_u8): Likewise.
1267 (vsubq_u16): Likewise.
1268 (vsubq_u32): Likewise.
1269 (vsubq_u64): Likewise.
1270 (vand_s8): Likewise.
1271 (vand_s16): Likewise.
1272 (vand_s32): Likewise.
1273 (vand_u8): Likewise.
1274 (vand_u16): Likewise.
1275 (vand_u32): Likewise.
1276 (vand_s64): Likewise.
1277 (vand_u64): Likewise.
1278 (vandq_s8): Likewise.
1279 (vandq_s16): Likewise.
1280 (vandq_s32): Likewise.
1281 (vandq_s64): Likewise.
1282 (vandq_u8): Likewise.
1283 (vandq_u16): Likewise.
1284 (vandq_u32): Likewise.
1285 (vandq_u64): Likewise.
1286 (vorr_s8): Likewise.
1287 (vorr_s16): Likewise.
1288 (vorr_s32): Likewise.
1289 (vorr_u8): Likewise.
1290 (vorr_u16): Likewise.
1291 (vorr_u32): Likewise.
1292 (vorr_s64): Likewise.
1293 (vorr_u64): Likewise.
1294 (vorrq_s8): Likewise.
1295 (vorrq_s16): Likewise.
1296 (vorrq_s32): Likewise.
1297 (vorrq_s64): Likewise.
1298 (vorrq_u8): Likewise.
1299 (vorrq_u16): Likewise.
1300 (vorrq_u32): Likewise.
1301 (vorrq_u64): Likewise.
1302 (veor_s8): Likewise.
1303 (veor_s16): Likewise.
1304 (veor_s32): Likewise.
1305 (veor_u8): Likewise.
1306 (veor_u16): Likewise.
1307 (veor_u32): Likewise.
1308 (veor_s64): Likewise.
1309 (veor_u64): Likewise.
1310 (veorq_s8): Likewise.
1311 (veorq_s16): Likewise.
1312 (veorq_s32): Likewise.
1313 (veorq_s64): Likewise.
1314 (veorq_u8): Likewise.
1315 (veorq_u16): Likewise.
1316 (veorq_u32): Likewise.
1317 (veorq_u64): Likewise.
1318 (vbic_s8): Likewise.
1319 (vbic_s16): Likewise.
1320 (vbic_s32): Likewise.
1321 (vbic_u8): Likewise.
1322 (vbic_u16): Likewise.
1323 (vbic_u32): Likewise.
1324 (vbic_s64): Likewise.
1325 (vbic_u64): Likewise.
1326 (vbicq_s8): Likewise.
1327 (vbicq_s16): Likewise.
1328 (vbicq_s32): Likewise.
1329 (vbicq_s64): Likewise.
1330 (vbicq_u8): Likewise.
1331 (vbicq_u16): Likewise.
1332 (vbicq_u32): Likewise.
1333 (vbicq_u64): Likewise.
1334 (vorn_s8): Likewise.
1335 (vorn_s16): Likewise.
1336 (vorn_s32): Likewise.
1337 (vorn_u8): Likewise.
1338 (vorn_u16): Likewise.
1339 (vorn_u32): Likewise.
1340 (vorn_s64): Likewise.
1341 (vorn_u64): Likewise.
1342 (vornq_s8): Likewise.
1343 (vornq_s16): Likewise.
1344 (vornq_s32): Likewise.
1345 (vornq_s64): Likewise.
1346 (vornq_u8): Likewise.
1347 (vornq_u16): Likewise.
1348 (vornq_u32): Likewise.
1349 (vornq_u64): Likewise.
1351 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1353 * wide-int.cc (UTItype): Define.
1354 (UDWtype): Define for appropriate W_TYPE_SIZE.
1356 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
1358 PR tree-optimization/59100
1359 * tree-ssa-phiopt.c: Include tree-inline.h.
1360 (neutral_element_p, absorbing_element_p): New functions.
1361 (value_replacement): Handle conditional binary operations with a
1362 neutral or absorbing element.
1364 2014-05-08 Richard Biener <rguenther@suse.de>
1366 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
1367 folding the expression.
1368 (valueize_expr): Remove.
1369 (visit_reference_op_load): Do not valueize the result of
1371 (simplify_binary_expression): Likewise.
1372 (simplify_unary_expression): Likewise.
1374 2014-05-08 Richard Biener <rguenther@suse.de>
1376 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
1377 looking at TYPE_ARG_TYPES.
1379 2014-05-08 Richard Biener <rguenther@suse.de>
1381 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
1382 pointer propagation special-case.
1384 2014-05-08 Bin Cheng <bin.cheng@arm.com>
1386 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
1387 core part of address expressions.
1389 2014-05-08 Alan Modra <amodra@gmail.com>
1392 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
1393 loads and stores when -mno-strict-align at any alignment.
1394 (expand_block_clear): Similarly. Also correct calculation of
1397 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
1400 * tree-complex.c (expand_complex_move): Keep line info when expanding
1402 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
1403 of complex expression. Use new argument to display correct location
1404 for values coming from phi statement.
1405 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
1406 (warn_uninitialized_phi): Pass location of phi argument to
1408 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
1409 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
1411 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
1413 * config/rs6000/predicates.md (indexed_address_mem): New.
1414 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
1415 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
1416 fpstore_ux, fpstore_u.
1417 (sign_extend, indexed, update): New.
1418 (cell_micro): Adjust.
1419 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
1420 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
1421 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
1422 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
1423 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
1424 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
1425 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
1426 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
1427 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
1428 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
1429 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
1430 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
1431 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
1432 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
1433 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
1435 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
1436 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
1437 *vsx_extract_<mode>_store): Adjust.
1438 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
1439 is_cracked_insn, insn_must_be_first_in_group,
1440 insn_must_be_last_in_group): Adjust.
1442 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
1444 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
1445 ppc440-fpstore): Adjust.
1446 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
1447 ppc476-fpstore): Adjust.
1448 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
1449 ppc601-fpstore): Adjust.
1450 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
1452 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
1454 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
1455 ppc7450-fpstore): Adjust.
1456 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
1457 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
1458 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
1460 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
1461 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
1462 cell-fpstore, cell-fpstore-update): Adjust.
1463 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
1464 ppce300c3_store, ppce300c3_fpstore): Adjust.
1465 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
1466 e500mc_fpstore): Adjust.
1467 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
1468 e500mc64_store, e500mc64_fpstore): Adjust.
1469 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
1470 e5500_fpstore): Adjust.
1471 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
1472 e6500_fpstore): Adjust.
1473 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
1475 * config/rs6000/power4.md (power4-load, power4-load-ext,
1476 power4-load-ext-update, power4-load-ext-update-indexed,
1477 power4-load-update-indexed, power4-load-update, power4-fpload,
1478 power4-fpload-update, power4-store, power4-store-update,
1479 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
1481 * config/rs6000/power5.md (power5-load, power5-load-ext,
1482 power5-load-ext-update, power5-load-ext-update-indexed,
1483 power5-load-update-indexed, power5-load-update, power5-fpload,
1484 power5-fpload-update, power5-store, power5-store-update,
1485 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
1487 * config/rs6000/power6.md (power6-load, power6-load-ext,
1488 power6-load-update, power6-load-update-indexed,
1489 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
1490 power6-fpload-update, power6-store, power6-store-update,
1491 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
1493 * config/rs6000/power7.md (power7-load, power7-load-ext,
1494 power7-load-update, power7-load-update-indexed,
1495 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
1496 power7-fpload-update, power7-store, power7-store-update,
1497 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
1499 * config/rs6000/power8.md (power8-load, power8-load-update,
1500 power8-load-ext, power8-load-ext-update, power8-fpload,
1501 power8-fpload-update, power8-store, power8-store-update-indexed,
1502 power8-fpstore, power8-fpstore-update): Adjust.
1503 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
1505 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
1506 titan_lsu_store, titan_lsu_fpstore): Adjust.
1507 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
1509 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
1512 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
1513 unrolled byte insns. Emit address increments after move insns.
1515 2014-05-07 David Malcolm <dmalcolm@redhat.com>
1517 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
1518 const_gimple, rather than a gimple.
1519 (gimple_call_builtin_p): Likewise, for the three variants.
1521 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
1522 (gimple_call_builtin_p): Likewise, for the three variants.
1524 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1526 PR tree-optimization/61095
1527 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
1529 2014-05-07 Richard Biener <rguenther@suse.de>
1531 PR tree-optimization/61034
1532 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
1533 (maybe_skip_until): Use translate to take into account
1534 lattices when trying to do disambiguations.
1535 (get_continuation_for_phi_1): Likewise.
1536 (get_continuation_for_phi): Adjust for added translate arguments.
1537 (walk_non_aliased_vuses): Likewise.
1538 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
1539 (walk_non_aliased_vuses): Likewise.
1540 (call_may_clobber_ref_p_1): Declare.
1541 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
1542 calls. Stop early if we are only supposed to disambiguate.
1543 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
1545 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
1547 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
1548 Emit an error when the function has arguments.
1550 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
1552 * cfgloop.h (unswitch_loops): Remove.
1553 * doc/passes.texi: Remove references to loop-unswitch.c
1554 * timevar.def (TV_LOOP_UNSWITCH): Remove.
1556 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
1558 * tree-vect-data-refs.c (vect_grouped_load_supported): New
1559 check for loads group of length 3.
1560 (vect_permute_load_chain): New permutations for loads group of
1562 * tree-vect-stmts.c (vect_model_load_cost): Change cost
1563 of vec_perm_shuffle for the new permutations.
1565 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
1567 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
1568 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
1569 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
1570 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
1571 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
1572 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
1573 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
1574 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
1576 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
1578 * loop-unswitch.c: Delete.
1580 2014-05-07 Richard Biener <rguenther@suse.de>
1582 * config.gcc: Always set need_64bit_hwint to yes.
1584 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
1586 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
1587 of using optimize_size.
1589 2014-05-06 Mike Stump <mikestump@comcast.net>
1591 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
1593 2014-05-06 Joseph Myers <joseph@codesourcery.com>
1595 * config/i386/sse.md (*mov<mode>_internal)
1596 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
1597 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
1598 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
1599 (*<code><mode>3, *andnot<mode>3<mask_name>)
1600 (<mask_codefor><code><mode>3<mask_name>): Only consider
1601 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
1603 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
1606 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
1608 * lra-constraints.c (valid_address_p): Move earlier in file.
1609 Add a constraint argument to the address_info version.
1610 (satisfies_memory_constraint_p): New function.
1611 (satisfies_address_constraint_p): Likewise.
1612 (process_alt_operands, curr_insn_transform): Use them.
1613 (process_address): Pass the constraint to valid_address_p when
1614 checking address operands.
1616 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
1618 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
1619 to their respective blocks. Fix inadvertent use of "node".
1621 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
1623 * emit-rtl.c (init_derived_machine_modes): New functionm, split
1625 (init_emit_once): ...here.
1626 * rtl.h (init_derived_machine_modes): Declare.
1627 * toplev.c (do_compile): Call it even if no_backend.
1629 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
1630 Mike Stump <mikestump@comcast.net>
1631 Richard Sandiford <rdsandiford@googlemail.com>
1632 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1634 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
1635 (rtx_equal_for_memref_p): Update comment.
1636 (adjust_offset_for_component_ref): Use wide-int interfaces.
1637 * builtins.c (get_object_alignment_2): Likewise.
1638 (c_readstr): Likewise.
1639 (target_char_cast): Add comment.
1640 (determine_block_size): Use wide-int interfaces.
1641 (expand_builtin_signbit): Likewise.
1642 (fold_builtin_int_roundingfn): Likewise.
1643 (fold_builtin_bitop): Likewise.
1644 (fold_builtin_bswap): Likewise.
1645 (fold_builtin_logarithm): Use signop.
1646 (fold_builtin_pow): Likewise.
1647 (fold_builtin_memory_op): Use wide-int interfaces.
1648 (fold_builtin_object_size): Likewise.
1649 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
1650 nb_iterations_estimate.
1651 (record_niter_bound): Use wide-int interfaces.
1652 (get_estimated_loop_iterations_int): Likewise.
1653 (get_estimated_loop_iterations): Likewise.
1654 (get_max_loop_iterations): Likewise.
1655 * cfgloop.h: Include wide-int.h.
1656 (struct nb_iter_bound): Change bound to widest_int.
1657 (struct loop): Change nb_iterations_upper_bound and
1658 nb_iterations_estimate to widest_int.
1659 (record_niter_bound): Switch to use widest_int.
1660 (get_estimated_loop_iterations): Likewise.
1661 (get_max_loop_iterations): Likewise.
1662 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
1663 update for wide-int.
1664 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
1665 * combine.c (try_combine): Likewise.
1666 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
1667 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
1669 (aarch64_float_const_representable_p): Likewise.
1670 * config/arc/arc.c: Include wide-int.h.
1671 (arc_can_use_doloop_p): Use wide-int interfaces.
1672 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
1673 (vfp3_const_double_index): Likewise.
1674 * config/avr/avr.c (avr_out_round): Likewise.
1675 (avr_fold_builtin): Likewise.
1676 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
1677 (bfin_can_use_doloop_p): Likewise.
1678 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
1679 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
1680 * config/i386/i386.c: Include wide-int.h.
1681 (ix86_data_alignment): Use wide-int interfaces.
1682 (ix86_local_alignment): Likewise.
1683 (ix86_emit_swsqrtsf): Update real_from_integer.
1684 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
1685 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
1686 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
1687 (zero_constant): Likewise.
1688 (input_operand): Likewise.
1689 (splat_input_operand): Likewise.
1690 (non_logical_cint_operand): Change const_double to const_wide_int.
1691 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
1692 (easy_altivec_constant): Remove comment.
1693 (paired_expand_vector_init): Use CONSTANT_P.
1694 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
1695 (rs6000_emit_move): Update checks.
1696 (rs6000_aggregate_candidate): Use wide-int interfaces.
1697 (rs6000_expand_ternop_builtin): Likewise.
1698 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
1699 (rs6000_assemble_integer): Likewise.
1700 (rs6000_hash_constant): Likewise.
1701 (output_toc): Likewise.
1702 (rs6000_rtx_costs): Likewise.
1703 (rs6000_emit_swrsqrt); Update call to real_from_integer.
1704 * config/rs6000/rs6000-c.c: Include wide-int.h.
1705 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
1706 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
1707 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
1708 Handle CONST_WIDE_INT.
1709 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
1710 Use tree_fits_uhwi_p.
1711 * config/sparc/sparc.c: Include wide-int.h.
1712 (sparc_fold_builtin): Use wide-int interfaces.
1713 * config/vax/vax.c: Include wide-int.h.
1714 (vax_float_literal): Use real_from_integer.
1715 * coretypes.h (struct hwivec_def): New.
1717 (const_hwivec): New.
1718 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
1719 (equiv_constant): Handle CONST_WIDE_INT.
1720 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
1721 (cselib_hash_rtx): Handle CONST_WIDE_INT.
1722 * dbxout.c (stabstr_U): Use wide-int interfaces.
1723 (dbxout_type): Update to use cst_fits_shwi_p.
1724 * defaults.h (LOG2_BITS_PER_UNIT): Define.
1725 (TARGET_SUPPORTS_WIDE_INT): Add default.
1726 * dfp.c: Include wide-int.h.
1727 (decimal_real_to_integer2): Use wide-int interfaces and rename to
1728 decimal_real_to_integer.
1729 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
1730 decimal_real_to_integer.
1731 * doc/generic.texi (Constant expressions): Update for wide_int.
1732 * doc/rtl.texi (const_double): Likewise.
1733 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
1734 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
1735 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
1736 (REAL_VALUE_FROM_INT): Remove.
1737 (TARGET_SUPPORTS_WIDE_INT): New.
1738 * doc/tm.texi: Regenerate.
1739 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
1740 * double-int.h: Include wide-int.h.
1741 (struct wi::int_traits): New.
1742 * dwarf2out.c (get_full_len): New.
1743 (dw_val_equal_p): Add case dw_val_class_wide_int.
1744 (size_of_loc_descr): Likewise.
1745 (output_loc_operands): Likewise.
1746 (insert_double): Remove.
1747 (insert_wide_int): New.
1749 (print_die): Add case dw_val_class_wide_int.
1750 (attr_checksum): Likewise.
1751 (attr_checksum_ordered): Likewise.
1752 (same_dw_val_p): Likewise.
1753 (size_of_die): Likewise.
1754 (value_format): Likewise.
1755 (output_die): Likewise.
1756 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
1758 (clz_loc_descriptor): Use wide-int interfaces.
1759 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
1760 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
1761 (round_up_to_align): Use wide-int interfaces.
1762 (field_byte_offset): Likewise.
1763 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
1764 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
1765 CONST_DOUBLE handling. Use wide-int interfaces.
1766 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
1767 (gen_enumeration_type_die): Use add_AT_wide.
1768 (hash_loc_operands): Add case dw_val_class_wide_int.
1769 (compare_loc_operands): Likewise.
1770 * dwarf2out.h: Include wide-int.h.
1771 (wide_int_ptr): New.
1772 (enum dw_val_class): Add dw_val_class_wide_int.
1773 (struct dw_val_struct): Add val_wide.
1774 * emit-rtl.c (const_wide_int_htab): New.
1775 (const_wide_int_htab_hash): New.
1776 (const_wide_int_htab_eq): New.
1777 (lookup_const_wide_int): New.
1778 (const_double_htab_hash): Use wide-int interfaces.
1779 (const_double_htab_eq): Likewise.
1780 (rtx_to_double_int): Conditionally compile for wide-int.
1781 (immed_double_int_const): Rename to immed_wide_int_const and
1782 update for wide-int.
1783 (immed_double_const): Conditionally compile for wide-int.
1784 (init_emit_once): Use wide-int interfaces.
1785 * explow.c (plus_constant): Likewise.
1786 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
1787 (lshift_value): Use wide-int interfaces.
1788 (expand_mult): Likewise.
1789 (choose_multiplier): Likewise.
1790 (expand_smod_pow2): Likewise.
1791 (make_tree): Likewise.
1792 * expr.c (convert_modes): Consolidate handling of constants.
1793 Use wide-int interfaces.
1794 (emit_group_load_1): Add note.
1795 (store_expr): Update comment.
1796 (get_inner_reference): Use wide-int interfaces.
1797 (expand_constructor): Update comment.
1798 (expand_expr_real_2): Use wide-int interfaces.
1799 (expand_expr_real_1): Likewise.
1800 (reduce_to_bit_field_precision): Likewise.
1801 (const_vector_from_tree): Likewise.
1802 * final.c: Include wide-int-print.h.
1803 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
1804 * fixed-value.c: Include wide-int.h.
1805 (fixed_from_string): Use wide-int interfaces.
1806 (fixed_to_decimal): Likewise.
1807 (fixed_convert_from_real): Likewise.
1808 (real_convert_from_fixed): Likewise.
1809 * fold-const.h (mem_ref_offset): Return an offset_int.
1810 (div_if_zero_remainder): Remove code parameter.
1811 * fold-const.c (div_if_zero_remainder): Remove code parameter.
1812 Use wide-int interfaces.
1813 (may_negate_without_overflow_p): Use wide-int interfaces.
1814 (negate_expr_p): Likewise.
1815 (fold_negate_expr): Likewise.
1816 (int_const_binop_1): Likewise.
1817 (const_binop): Likewise.
1818 (fold_convert_const_int_from_int): Likewise.
1819 (fold_convert_const_int_from_real): Likewise.
1820 (fold_convert_const_int_from_fixed): Likewise.
1821 (fold_convert_const_fixed_from_int): Likewise.
1822 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
1823 (sign_bit_p): Use wide-int interfaces.
1824 (make_range_step): Likewise.
1825 (build_range_check): Likewise. Pass an integer of the correct type
1826 instead of using integer_one_node.
1827 (range_predecessor): Pass an integer of the correct type instead
1828 of using integer_one_node.
1829 (range_successor): Likewise.
1830 (merge_ranges): Likewise.
1831 (unextend): Use wide-int interfaces.
1832 (extract_muldiv_1): Likewise.
1833 (fold_div_compare): Likewise.
1834 (fold_single_bit_test): Likewise.
1835 (fold_sign_changed_comparison): Likewise.
1836 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
1837 (fold_plusminus_mult_expr): Use wide-int interfaces.
1838 (native_encode_int): Likewise.
1839 (native_interpret_int): Likewise.
1840 (fold_unary_loc): Likewise.
1841 (pointer_may_wrap_p): Likewise.
1842 (size_low_cst): Likewise.
1843 (mask_with_tz): Likewise.
1844 (fold_binary_loc): Likewise.
1845 (fold_ternary_loc): Likewise.
1846 (multiple_of_p): Likewise.
1847 (tree_call_nonnegative_warnv_p): Update calls to
1848 tree_int_cst_min_precision and real_from_integer.
1849 (fold_negate_const): Use wide-int interfaces.
1850 (fold_abs_const): Likewise.
1851 (fold_relational_const): Use tree_int_cst_lt.
1852 (round_up_loc): Use wide-int interfaces.
1853 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
1854 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
1855 * gengtype.c: Remove include of double-int.h.
1856 (do_typedef): Use wide-int interfaces.
1857 (open_base_files): Add wide-int.h.
1858 (main): Add offset_int and widest_int typedefs.
1859 * gengtype-lex.l: Handle "^".
1860 (CXX_KEYWORD): Add "static".
1861 * gengtype-parse.c (require3): New.
1862 (require_template_declaration): Handle constant template arguments
1863 and nested templates.
1864 * gengtype-state.c: Don't include "double-int.h".
1865 * genpreds.c (write_one_predicate_function): Update comment.
1866 (write_tm_constrs_h): Add check for hval and lval use in
1868 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
1869 (add_to_sequence): Likewise.
1870 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
1871 and const_double_operand.
1872 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
1874 * gimple-fold.c (get_base_constructor): Likewise.
1875 (fold_array_ctor_reference): Likewise.
1876 (fold_nonarray_ctor_reference): Likewise.
1877 (fold_const_aggregate_ref_1): Likewise.
1878 (gimple_val_nonnegative_real_p): Likewise.
1879 (gimple_fold_indirect_ref): Likewise.
1880 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
1881 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
1882 (struct slsr_cand_d): Change index to be widest_int.
1883 (struct incr_info_d): Change incr to be widest_int.
1884 (alloc_cand_and_find_basis): Use wide-int interfaces.
1885 (slsr_process_phi): Likewise.
1886 (backtrace_base_for_ref): Likewise. Return a widest_int.
1887 (restructure_reference): Take a widest_int instead of a double_int.
1888 (slsr_process_ref): Use wide-int interfaces.
1889 (create_mul_ssa_cand): Likewise.
1890 (create_mul_imm_cand): Likewise.
1891 (create_add_ssa_cand): Likewise.
1892 (create_add_imm_cand): Take a widest_int instead of a double_int.
1893 (slsr_process_add): Use wide-int interfaces.
1894 (slsr_process_cast): Likewise.
1895 (slsr_process_copy): Likewise.
1896 (dump_candidate): Likewise.
1897 (dump_incr_vec): Likewise.
1898 (replace_ref): Likewise.
1899 (cand_increment): Likewise. Return a widest_int.
1900 (cand_abs_increment): Likewise.
1901 (replace_mult_candidate): Take a widest_int instead of a double_int.
1902 (replace_unconditional_candidate): Use wide-int interfaces.
1903 (incr_vec_index): Take a widest_int instead of a double_int.
1904 (create_add_on_incoming_edge): Likewise.
1905 (create_phi_basis): Use wide-int interfaces.
1906 (replace_conditional_candidate): Likewise.
1907 (record_increment): Take a widest_int instead of a double_int.
1908 (record_phi_increments): Use wide-int interfaces.
1909 (phi_incr_cost): Take a widest_int instead of a double_int.
1910 (lowest_cost_path): Likewise.
1911 (total_savings): Likewise.
1912 (analyze_increments): Use wide-int interfaces.
1913 (ncd_with_phi): Take a widest_int instead of a double_int.
1914 (ncd_of_cand_and_phis): Likewise.
1915 (nearest_common_dominator_for_cands): Likewise.
1916 (insert_initializers): Use wide-int interfaces.
1917 (all_phi_incrs_profitable): Likewise.
1918 (replace_one_candidate): Likewise.
1919 (replace_profitable_candidates): Likewise.
1920 * godump.c: Include wide-int-print.h.
1921 (go_output_typedef): Use wide-int interfaces.
1922 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
1923 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
1924 (build_loop_iteration_domains): Likewise.
1925 * hooks.h: Include wide-int.h rather than double-int.h.
1926 (hook_bool_dint_dint_uint_bool_true): Delete.
1927 (hook_bool_wint_wint_uint_bool_true): Declare.
1928 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
1929 (hook_bool_wint_wint_uint_bool_true): New.
1930 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
1932 (ubsan_expand_si_overflow_mul_check): Likewise.
1933 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
1934 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
1935 (get_ancestor_addr_info): Likewise.
1936 (ipa_modify_call_arguments): Likewise.
1937 * loop-doloop.c (doloop_modify): Likewise.
1938 (doloop_optimize): Likewise.
1939 * loop-iv.c (iv_number_of_iterations): Likewise.
1940 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
1941 (unroll_loop_constant_iterations): Likewise.
1942 (decide_unroll_runtime_iterations): Likewise.
1943 (unroll_loop_runtime_iterations): Likewise.
1944 (decide_peel_simple): Likewise.
1945 (decide_unroll_stupid): Likewise.
1946 * lto-streamer-in.c (streamer_read_wi): Add.
1947 (input_cfg): Use wide-int interfaces.
1948 (lto_input_tree_1): Likewise.
1949 * lto-streamer-out.c (streamer_write_wi): Add.
1950 (hash_tree): Use wide-int interfaces.
1951 (output_cfg): Likewise.
1952 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
1953 (GTFILES): Add wide-int.h and signop.h.
1954 (TAGS): Look for .cc files too.
1955 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
1956 * optabs.c (expand_subword_shift): Likewise.
1957 (expand_doubleword_shift): Likewise.
1958 (expand_absneg_bit): Likewise.
1959 (expand_copysign_absneg): Likewise.
1960 (expand_copysign_bit): Likewise.
1961 * postreload.c (reload_cse_simplify_set): Likewise.
1962 * predict.c (predict_iv_comparison): Likewise.
1963 * pretty-print.h: Include wide-int-print.h.
1965 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
1966 * print-tree.c: Include wide-int-print.h.
1967 (print_node_brief): Use wide-int interfaces.
1968 (print_node): Likewise.
1969 * read-rtl.c (validate_const_wide_int): New.
1970 (read_rtx_code): Add CONST_WIDE_INT case.
1971 * real.c: Include wide-int.h.
1972 (real_to_integer2): Delete.
1973 (real_to_integer): New function, returning a wide_int.
1974 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1975 (ten_to_ptwo): Update call to real_from_integer.
1976 (real_digit): Likewise.
1977 * real.h: Include signop.h, wide-int.h and insn-modes.h.
1978 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
1979 (REAL_VALUE_TO_INT): Delete.
1980 (real_to_integer): Declare a wide-int form.
1981 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
1982 * recog.c (const_int_operand): Improve comment.
1983 (const_scalar_int_operand): New.
1984 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
1985 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
1986 (split_double): Likewise.
1987 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
1988 (rtx_size): Likewise.
1989 (rtx_alloc_stat_v): New.
1990 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
1991 (cwi_output_hex): New.
1992 (iterative_hash_rtx): Handle CONST_WIDE_INT.
1993 (cwi_check_failed_bounds): New.
1994 * rtl.def (CONST_WIDE_INT): New.
1995 * rtl.h: Include <utility> and wide-int.h.
1996 (struct hwivec_def): New.
1997 (CWI_GET_NUM_ELEM): New.
1998 (CWI_PUT_NUM_ELEM): New.
1999 (struct rtx_def): Add num_elem and hwiv.
2000 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
2001 (CASE_CONST_UNIQUE): Likewise.
2002 (CASE_CONST_ANY): Likewise.
2003 (CONST_SCALAR_INT_P): Likewise.
2004 (CONST_WIDE_INT_P): New.
2006 (HWIVEC_CHECK): New.
2007 (cwi_check_failed_bounds): New.
2009 (HWIVEC_CHECK): New.
2010 (CONST_WIDE_INT_VEC) New.
2011 (CONST_WIDE_INT_NUNITS) New.
2012 (CONST_WIDE_INT_ELT) New.
2013 (rtx_mode_t): New type.
2014 (wi::int_traits <rtx_mode_t>): New.
2016 (wi::min_value): New.
2017 (wi::max_value): New.
2019 (const_wide_int_alloc): New.
2020 (immed_wide_int_const): New.
2021 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
2022 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
2023 * signop.h: New file.
2024 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
2025 (simplify_const_unary_operation): Use wide-int interfaces.
2026 (simplify_binary_operation_1): Likewise.
2027 (simplify_const_binary_operation): Likewise.
2028 (simplify_const_relational_operation): Likewise.
2029 (simplify_immed_subreg): Likewise.
2030 * stmt.c (expand_case): Likewise.
2031 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
2032 signop rather than a bool.
2033 * stor-layout.c (layout_type): Use wide-int interfaces.
2034 (initialize_sizetypes): Update calls to
2035 set_min_and_max_values_for_integral_type.
2036 (set_min_and_max_values_for_integral_type): Take a signop rather
2037 than a bool. Use wide-int interfaces.
2038 (fixup_signed_type): Update accordingly. Remove
2039 HOST_BITS_PER_DOUBLE_INT limit.
2040 (fixup_unsigned_type): Likewise.
2041 * system.h (STATIC_CONSTANT_P): New.
2042 (STATIC_ASSERT): New.
2043 * target.def (can_use_doloop_p): Take widest_ints rather than
2045 * target.h: Include wide-int.h rather than double-int.h.
2046 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
2048 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
2049 rather than INT_CST_LT_UNSIGNED.
2050 (can_use_doloop_if_innermost): Take widest_ints rather than
2052 * tree-affine.c: Include wide-int-print.h.
2053 (double_int_ext_for_comb): Delete.
2054 (wide_int_ext_for_comb): New.
2055 (aff_combination_zero): Use wide-int interfaces.
2056 (aff_combination_const): Take a widest_int instead of a double_int.
2057 (aff_combination_elt): Use wide-int interfaces.
2058 (aff_combination_scale): Take a widest_int instead of a double_int.
2059 (aff_combination_add_elt): Likewise.
2060 (aff_combination_add_cst): Likewise.
2061 (aff_combination_add): Use wide-int interfaces.
2062 (aff_combination_convert): Likewise.
2063 (tree_to_aff_combination): Likewise.
2064 (add_elt_to_tree): Take a widest_int instead of a double_int.
2065 (aff_combination_to_tree): Use wide-int interfaces.
2066 (aff_combination_remove_elt): Likewise.
2067 (aff_combination_add_product): Take a widest_int instead of
2069 (aff_combination_mult): Use wide-int interfaces.
2070 (aff_combination_expand): Likewise.
2071 (double_int_constant_multiple_p): Delete.
2072 (wide_int_constant_multiple_p): New.
2073 (aff_combination_constant_multiple_p): Take a widest_int pointer
2074 instead of a double_int pointer.
2075 (print_aff): Use wide-int interfaces.
2076 (get_inner_reference_aff): Take a widest_int pointer
2077 instead of a double_int pointer.
2078 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
2079 * tree-affine.h: Include wide-int.h.
2080 (struct aff_comb_elt): Change type of coef to widest_int.
2081 (struct affine_tree_combination): Change type of offset to widest_int.
2082 (double_int_ext_for_comb): Delete.
2083 (wide_int_ext_for_comb): New.
2084 (aff_combination_const): Use widest_int instead of double_int.
2085 (aff_combination_scale): Likewise.
2086 (aff_combination_add_elt): Likewise.
2087 (aff_combination_constant_multiple_p): Likewise.
2088 (get_inner_reference_aff): Likewise.
2089 (aff_comb_cannot_overlap_p): Likewise.
2090 (aff_combination_zero_p): Use wide-int interfaces.
2091 * tree.c: Include tree.h.
2092 (init_ttree): Use make_int_cst.
2093 (tree_code_size): Removed code for INTEGER_CST case.
2094 (tree_size): Add INTEGER_CST case.
2095 (make_node_stat): Update comment.
2096 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
2097 (build_int_cst_type): Use wide-int interfaces.
2098 (double_int_to_tree): Likewise.
2099 (double_int_fits_to_tree_p): Delete.
2100 (force_fit_type_double): Delete.
2101 (force_fit_type): New.
2102 (int_cst_hash_hash): Use wide-int interfaces.
2103 (int_cst_hash_eq): Likewise.
2104 (build_int_cst_wide): Delete.
2105 (wide_int_to_tree): New.
2106 (cache_integer_cst): Use wide-int interfaces.
2107 (build_low_bits_mask): Likewise.
2108 (cst_and_fits_in_hwi): Likewise.
2109 (real_value_from_int_cst): Likewise.
2110 (make_int_cst_stat): New.
2111 (integer_zerop): Use wide_int interfaces.
2112 (integer_onep): Likewise.
2113 (integer_all_onesp): Likewise.
2114 (integer_pow2p): Likewise.
2115 (integer_nonzerop): Likewise.
2116 (tree_log2): Likewise.
2117 (tree_floor_log2): Likewise.
2118 (tree_ctz): Likewise.
2119 (int_size_in_bytes): Likewise.
2120 (mem_ref_offset): Return an offset_int rather than a double_int.
2121 (build_type_attribute_qual_variant): Use wide_int interfaces.
2122 (type_hash_eq): Likewise
2123 (tree_int_cst_equal): Likewise.
2124 (tree_int_cst_lt): Delete.
2125 (tree_int_cst_compare): Likewise.
2126 (tree_fits_shwi_p): Use wide_int interfaces.
2127 (tree_fits_uhwi_p): Likewise.
2128 (tree_int_cst_sign_bit): Likewise.
2129 (tree_int_cst_sgn): Likewise.
2130 (tree_int_cst_min_precision): Take a signop rather than a bool.
2131 (simple_cst_equal): Use wide_int interfaces.
2132 (compare_tree_int): Likewise.
2133 (iterative_hash_expr): Likewise.
2134 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
2136 (get_type_static_bounds): Use wide_int interfaces.
2137 (tree_int_cst_elt_check_failed): New.
2138 (build_common_tree_nodes): Reordered to set prec before filling in
2140 (int_cst_value): Check cst_and_fits_in_hwi.
2141 (widest_int_cst_value): Use wide_int interfaces.
2142 (upper_bound_in_type): Likewise.
2143 (lower_bound_in_type): Likewise.
2144 (num_ending_zeros): Likewise.
2145 (drop_tree_overflow): Likewise.
2146 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
2147 (gen_conditions_for_pow_cst_base): Likewise.
2148 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
2149 (group_case_labels_stmt): Use wide-int interfaces.
2150 (verify_gimple_assign_binary): Likewise.
2151 (print_loop): Likewise.
2152 * tree-chrec.c (tree_fold_binomial): Likewise.
2153 * tree-core.h (struct tree_base): Add int_length.
2154 (struct tree_int_cst): Change rep of value.
2155 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
2156 (dr_may_alias_p): Likewise.
2157 (max_stmt_executions_tree): Likewise.
2158 * tree.def (INTEGER_CST): Update comment.
2159 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
2160 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
2161 * tree-dump.c: Include wide-int.h and wide-int-print.h.
2162 (dequeue_and_dump): Use wide-int interfaces.
2163 * tree.h: Include wide-int.h.
2164 (NULL_TREE): Moved to earlier loc in file.
2165 (TREE_INT_CST_ELT_CHECK): New.
2166 (tree_int_cst_elt_check_failed): New.
2168 (TREE_INT_CST): Delete.
2169 (TREE_INT_CST_LOW): Use wide-int interfaces.
2170 (TREE_INT_CST_HIGH): Delete.
2171 (TREE_INT_CST_NUNITS): New.
2172 (TREE_INT_CST_EXT_NUNITS): Likewise.
2173 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
2174 (TREE_INT_CST_ELT): Likewise.
2175 (INT_CST_LT): Delete.
2176 (tree_int_cst_elt_check): New (two forms).
2177 (type_code_size): Update comment.
2178 (make_int_cst_stat, make_int_cst): New.
2179 (tree_to_double_int): Delete.
2180 (double_int_fits_to_tree_p): Delete.
2181 (force_fit_type_double): Delete.
2182 (build_int_cstu): Replace with out-of-line function.
2183 (build_int_cst_wide): Delete.
2184 (tree_int_cst_lt): Define inline.
2185 (tree_int_cst_le): New.
2186 (tree_int_cst_compare): Define inline.
2187 (tree_int_cst_min_precision): Take a signop rather than a bool.
2188 (wi::int_traits <const_tree>): New.
2189 (wi::int_traits <tree>): New.
2190 (wi::extended_tree): New.
2191 (wi::int_traits <wi::extended_tree>): New.
2192 (wi::to_widest): New.
2193 (wi::to_offset): New.
2194 (wi::fits_to_tree_p): New.
2195 (wi::min_value): New.
2196 (wi::max_value): New.
2197 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
2198 (copy_tree_body_r): Likewise.
2199 * tree-object-size.c (compute_object_offset): Likewise.
2200 (addr_object_size): Likewise.
2201 * tree-predcom.c: Include wide-int-print.h.
2202 (struct dref_d): Change type of offset to widest_int.
2203 (dump_dref): Call wide-int printer.
2204 (aff_combination_dr_offset): Use wide-int interfaces.
2205 (determine_offset): Take a widest_int pointer rather than a
2207 (split_data_refs_to_components): Use wide-int interfaces.
2208 (suitable_component_p): Likewise.
2209 (order_drefs): Likewise.
2210 (add_ref_to_chain): Likewise.
2211 (valid_initializer_p): Likewise.
2212 (determine_roots_comp): Likewise.
2213 * tree-pretty-print.c: Include wide-int-print.h.
2214 (dump_generic_node): Use wide-int interfaces.
2215 * tree-sra.c (sra_ipa_modify_expr): Likewise.
2216 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
2217 (move_fixed_address_to_symbol): Likewise.
2218 (move_hint_to_base): Likewise.
2219 (move_pointer_to_base): Likewise.
2220 (move_variant_to_index): Likewise.
2221 (most_expensive_mult_to_index): Likewise.
2222 (addr_to_parts): Likewise.
2223 (copy_ref_info): Likewise.
2224 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
2225 (indirect_refs_may_alias_p): Likewise.
2226 (stmt_kills_ref_p_1): Likewise.
2227 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
2228 * tree-ssa-ccp.c: Update comment at top of file. Include
2230 (struct prop_value_d): Change type of mask to widest_int.
2231 (extend_mask): New function.
2232 (dump_lattice_value): Use wide-int interfaces.
2233 (get_default_value): Likewise.
2234 (set_constant_value): Likewise.
2235 (set_value_varying): Likewise.
2236 (valid_lattice_transition): Likewise.
2237 (set_lattice_value): Likewise.
2238 (value_to_double_int): Delete.
2239 (value_to_wide_int): New.
2240 (get_value_from_alignment): Use wide-int interfaces.
2241 (get_value_for_expr): Likewise.
2242 (do_dbg_cnt): Likewise.
2243 (ccp_finalize): Likewise.
2244 (ccp_lattice_meet): Likewise.
2245 (bit_value_unop_1): Use widest_ints rather than double_ints.
2246 (bit_value_binop_1): Likewise.
2247 (bit_value_unop): Use wide-int interfaces.
2248 (bit_value_binop): Likewise.
2249 (bit_value_assume_aligned): Likewise.
2250 (evaluate_stmt): Likewise.
2251 (ccp_fold_stmt): Likewise.
2252 (visit_cond_stmt): Likewise.
2253 (ccp_visit_stmt): Likewise.
2254 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
2255 (constant_pointer_difference): Likewise.
2256 (associate_pointerplus): Likewise.
2257 (combine_conversions): Likewise.
2258 * tree-ssa-loop.h: Include wide-int.h.
2259 (struct tree_niter_desc): Change type of max to widest_int.
2260 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
2261 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
2262 (remove_redundant_iv_tests): Likewise.
2263 (canonicalize_loop_induction_variables): Likewise.
2264 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
2265 (constant_multiple_of): Take a widest_int pointer instead of
2266 a double_int pointer.
2267 (get_computation_aff): Use wide-int interfaces.
2268 (ptr_difference_cost): Likewise.
2269 (difference_cost): Likewise.
2270 (get_loop_invariant_expr_id): Likewise.
2271 (get_computation_cost_at): Likewise.
2272 (iv_elimination_compare_lt): Likewise.
2273 (may_eliminate_iv): Likewise.
2274 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
2275 instead of double_int.
2276 (max_loop_iterations): Likewise.
2277 (max_stmt_executions): Likewise.
2278 (estimated_stmt_executions): Likewise.
2279 * tree-ssa-loop-niter.c: Include wide-int-print.h.
2280 (split_to_var_and_offset): Use wide-int interfaces.
2281 (determine_value_range): Likewise.
2282 (bound_difference_of_offsetted_base): Likewise.
2283 (bounds_add): Take a widest_int instead of a double_int.
2284 (number_of_iterations_ne_max): Use wide-int interfaces.
2285 (number_of_iterations_ne): Likewise.
2286 (number_of_iterations_lt_to_ne): Likewise.
2287 (assert_loop_rolls_lt): Likewise.
2288 (number_of_iterations_lt): Likewise.
2289 (number_of_iterations_le): Likewise.
2290 (number_of_iterations_cond): Likewise.
2291 (number_of_iterations_exit): Likewise.
2292 (finite_loop_p): Likewise.
2293 (derive_constant_upper_bound_assign): Likewise.
2294 (derive_constant_upper_bound): Return a widest_int.
2295 (derive_constant_upper_bound_ops): Likewise.
2296 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
2297 (record_estimate): Take a widest_int rather than a double_int.
2298 (record_nonwrapping_iv): Use wide-int interfaces.
2299 (double_int_cmp): Delete.
2300 (wide_int_cmp): New.
2301 (bound_index): Take a widest_int rather than a double_int.
2302 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
2303 (maybe_lower_iteration_bound): Likewise.
2304 (estimate_numbers_of_iterations_loop): Likewise.
2305 (estimated_loop_iterations): Take a widest_int pointer than than
2306 a double_int pointer.
2307 (estimated_loop_iterations_int): Use wide-int interfaces.
2308 (max_loop_iterations): Take a widest_int pointer than than
2309 a double_int pointer.
2310 (max_loop_iterations_int): Use wide-int interfaces.
2311 (max_stmt_executions): Take a widest_int pointer than than
2312 a double_int pointer.
2313 (estimated_stmt_executions): Likewise.
2314 (n_of_executions_at_most): Use wide-int interfaces.
2315 (scev_probably_wraps_p): Likewise.
2316 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
2318 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
2320 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
2321 double_ints. Adjust for trailing_wide_ints <3> representation.
2322 (set_nonzero_bits): Likewise.
2323 (get_range_info): Return wide_ints rather than double_ints.
2324 Adjust for trailing_wide_ints <3> representation.
2325 (get_nonzero_bits): Likewise.
2326 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
2328 * tree-ssanames.h (struct range_info_def): Replace min, max and
2329 nonzero_bits with a trailing_wide_ints <3>.
2330 (set_range_info): Use wide_int_refs rather than double_ints.
2331 (set_nonzero_bits): Likewise.
2332 (get_range_info): Return wide_ints rather than double_ints.
2333 (get_nonzero_bits): Likewise.
2334 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
2335 * tree-ssa-pre.c (phi_translate_1): Likewise.
2336 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
2337 (acceptable_pow_call): Likewise.
2338 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
2340 (vn_reference_fold_indirect): Likewise.
2341 (vn_reference_maybe_forwprop_address): Likewise.
2342 (valueize_refs_1): Likewise.
2343 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
2344 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
2345 tree_int_cst_lt and tree_int_cst_le.
2346 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
2348 (streamer_alloc_tree): Likewise.
2349 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
2350 (streamer_write_tree_header): Likewise.
2351 (streamer_write_integer_cst): Likewise.
2352 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
2353 (build_constructors): Likewise.
2354 (array_value_type): Likewise.
2355 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
2356 (vect_check_gather): Likewise.
2357 * tree-vect-generic.c (build_replicated_const): Likewise.
2358 (expand_vector_divmod): Likewise.
2359 * tree-vect-loop.c (vect_transform_loop): Likewise.
2360 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
2361 (vect_do_peeling_for_alignment): Likewise.
2362 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
2363 * tree-vrp.c: Include wide-int.h.
2364 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
2365 (extract_range_from_assert): Use wide-int interfaces.
2366 (vrp_int_const_binop): Likewise.
2367 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
2368 double_int pointers.
2369 (ranges_from_anti_range): Use wide-int interfaces.
2370 (quad_int_cmp): Delete.
2371 (quad_int_pair_sort): Likewise.
2372 (extract_range_from_binary_expr_1): Use wide-int interfaces.
2373 (extract_range_from_unary_expr_1): Likewise.
2374 (adjust_range_with_scev): Likewise.
2375 (masked_increment): Take and return wide_ints rather than double_ints.
2376 (register_edge_assert_for_2): Use wide-int interfaces.
2377 (check_array_ref): Likewise.
2378 (search_for_addr_array): Likewise.
2379 (maybe_set_nonzero_bits): Likewise.
2380 (union_ranges): Pass an integer of the correct type instead of
2381 using integer_one_node.
2382 (intersect_ranges): Likewise.
2383 (simplify_truth_ops_using_ranges): Likewise.
2384 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
2385 (range_fits_type_p): Likewise.
2386 (simplify_cond_using_ranges): Likewise. Take a signop rather than
2388 (simplify_conversion_using_ranges): Use wide-int interfaces.
2389 (simplify_float_conversion_using_ranges): Likewise.
2390 (vrp_finalize): Likewise.
2391 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
2392 (gimple_stringops_transform): Likewise.
2393 * varasm.c (decode_addr_const): Likewise.
2394 (const_hash_1): Likewise.
2395 (const_rtx_hash_1): Likewise
2396 (output_constant): Likewise.
2397 (array_size_for_constructor): Likewise.
2398 (output_constructor_regular_field): Likewise.
2399 (output_constructor_bitfield): Likewise.
2400 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
2401 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
2403 * gencheck.c: Define BITS_PER_UNIT.
2406 * wide-int-print.cc: New.
2407 * wide-int-print.h: New.
2409 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2411 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
2413 2014-05-06 Richard Biener <rguenther@suse.de>
2415 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
2416 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
2417 (TODO_verify_all): Adjust.
2418 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
2419 TODO_verify_stmts and TODO_verify_rtl_sharing.
2420 * bb-reorder.c: Likewise.
2421 * cfgexpand.c: Likewise.
2422 * cprop.c: Likewise.
2424 * function.c: Likewise.
2425 * fwprop.c: Likewise.
2427 * gimple-ssa-isolate-paths.c: Likewise.
2428 * gimple-ssa-strength-reduction.c: Likewise.
2429 * ipa-split.c: Likewise.
2430 * loop-init.c: Likewise.
2431 * loop-unroll.c: Likewise.
2432 * lower-subreg.c: Likewise.
2433 * modulo-sched.c: Likewise.
2434 * postreload-gcse.c: Likewise.
2435 * predict.c: Likewise.
2436 * recog.c: Likewise.
2437 * sched-rgn.c: Likewise.
2438 * store-motion.c: Likewise.
2439 * tracer.c: Likewise.
2440 * trans-mem.c: Likewise.
2441 * tree-call-cdce.c: Likewise.
2442 * tree-cfg.c: Likewise.
2443 * tree-cfgcleanup.c: Likewise.
2444 * tree-complex.c: Likewise.
2445 * tree-eh.c: Likewise.
2446 * tree-emutls.c: Likewise.
2447 * tree-if-conv.c: Likewise.
2448 * tree-into-ssa.c: Likewise.
2449 * tree-loop-distribution.c: Likewise.
2450 * tree-object-size.c: Likewise.
2451 * tree-parloops.c: Likewise.
2452 * tree-pass.h: Likewise.
2453 * tree-sra.c: Likewise.
2454 * tree-ssa-ccp.c: Likewise.
2455 * tree-ssa-copy.c: Likewise.
2456 * tree-ssa-copyrename.c: Likewise.
2457 * tree-ssa-dce.c: Likewise.
2458 * tree-ssa-dom.c: Likewise.
2459 * tree-ssa-dse.c: Likewise.
2460 * tree-ssa-forwprop.c: Likewise.
2461 * tree-ssa-ifcombine.c: Likewise.
2462 * tree-ssa-loop-ch.c: Likewise.
2463 * tree-ssa-loop-ivcanon.c: Likewise.
2464 * tree-ssa-loop.c: Likewise.
2465 * tree-ssa-math-opts.c: Likewise.
2466 * tree-ssa-phiopt.c: Likewise.
2467 * tree-ssa-phiprop.c: Likewise.
2468 * tree-ssa-pre.c: Likewise.
2469 * tree-ssa-reassoc.c: Likewise.
2470 * tree-ssa-sink.c: Likewise.
2471 * tree-ssa-strlen.c: Likewise.
2472 * tree-ssa-tail-merge.c: Likewise.
2473 * tree-ssa-uncprop.c: Likewise.
2474 * tree-switch-conversion.c: Likewise.
2475 * tree-tailcall.c: Likewise.
2476 * tree-vect-generic.c: Likewise.
2477 * tree-vectorizer.c: Likewise.
2478 * tree-vrp.c: Likewise.
2480 * var-tracking.c: Likewise.
2481 * bt-load.c: Likewise.
2482 * cfgcleanup.c: Likewise.
2483 * combine-stack-adj.c: Likewise.
2484 * combine.c: Likewise.
2485 * compare-elim.c: Likewise.
2486 * config/epiphany/resolve-sw-modes.c: Likewise.
2487 * config/i386/i386.c: Likewise.
2488 * config/mips/mips.c: Likewise.
2489 * config/s390/s390.c: Likewise.
2490 * config/sh/sh_treg_combine.cc: Likewise.
2491 * config/sparc/sparc.c: Likewise.
2494 * final.c: Likewise.
2495 * ifcvt.c: Likewise.
2496 * mode-switching.c: Likewise.
2497 * passes.c: Likewise.
2498 * postreload.c: Likewise.
2500 * reg-stack.c: Likewise.
2501 * regcprop.c: Likewise.
2502 * regrename.c: Likewise.
2505 2014-05-06 Richard Biener <rguenther@suse.de>
2508 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
2509 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
2511 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
2514 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
2516 2014-05-05 Radovan Obradovic <robradovic@mips.com>
2517 Tom de Vries <tom@codesourcery.com>
2519 * target.def (call_fusage_contains_non_callee_clobbers): New
2521 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
2523 (@node Miscellaneous Register Hooks): New node.
2524 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
2525 * doc/tm.texi: Regenerate.
2527 2014-05-05 Marek Polacek <polacek@redhat.com>
2530 * opts.c (common_handle_option): Call error_at instead of warning_at.
2532 2014-05-05 Richard Biener <rguenther@suse.de>
2534 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
2535 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
2536 under the TODO_verify_il umbrella.
2538 2014-05-05 Richard Biener <rguenther@suse.de>
2540 * passes.c (execute_function_todo): Move TODO_verify_flow under
2541 the TODO_verify_ul umbrella.
2543 2014-05-05 Richard Biener <rguenther@suse.de>
2546 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
2547 X & CST away from a CST that is the mask of a mode.
2549 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2551 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
2552 int argument to enum machine_mode.
2553 (picochip_class_max_nregs): Ditto.
2554 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
2555 (picochip_class_max_nregs): Ditto.
2557 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2559 * target.def: Add new target hook.
2560 * doc/tm.texi: Regenerate.
2561 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
2562 * targhooks.c (default_keep_leaf_when_profiled): New function.
2564 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
2565 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
2567 2014-05-05 Bin Cheng <bin.cheng@arm.com>
2569 PR tree-optimization/60363
2570 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
2571 (copy_phi_args): New parameters. Call get_value_locus_in_path.
2572 (update_destination_phis): New parameter.
2573 (create_edge_and_update_destination_phis): Ditto.
2574 (ssa_fix_duplicate_block_edges): Pass new arguments.
2575 (thread_single_edge): Ditto.
2577 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
2579 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
2580 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
2581 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
2582 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
2583 Use RS6000_BTM_HARD_FLOAT.
2584 (BU_MISC_2): Likewise.
2585 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
2586 RS6000_BTM_HARD_FLOAT.
2587 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
2589 (rs6000_invalid_builtin): Add hard floating builtin support.
2590 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
2591 hard float builtins.
2592 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
2594 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2596 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
2597 Add missing function* argument.
2599 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2601 * lra-constraints.c (valid_address_p): Move earlier in file.
2602 Add a constraint argument to the address_info version.
2603 (satisfies_memory_constraint_p): New function.
2604 (satisfies_address_constraint_p): Likewise.
2605 (process_alt_operands, curr_insn_transform): Use them.
2606 (process_address): Pass the constraint to valid_address_p when
2607 checking address operands.
2609 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2611 * config/mips/mips.c (mips_isa_rev): New variable.
2612 (mips_set_architecture): Set it.
2613 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
2615 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
2616 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
2617 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
2618 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
2619 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
2620 conditions in terms of mips_isa_rev.
2621 (mips_isa_rev): Declare.
2623 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2625 * config/sh/sh-mem.cc: Use tabs instead of spaces.
2626 (prob_unlikely, prob_likely): Make variables const.
2628 2014-05-03 Denis Chertykov <chertykov@gmail.com>
2630 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
2632 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2634 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
2636 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2638 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
2639 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
2640 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
2642 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
2643 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
2645 Replace usage of ROUND_REG with sh_round_reg.
2646 Use CEIL instead of ROUND_ADVANCE.
2648 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2651 * config/sh/sh.c: Include stdlib headers before everything else.
2653 2014-05-02 Jakub Jelinek <jakub@redhat.com>
2655 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
2656 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
2657 (gimplify_adjust_omp_clauses): Simd region is never
2658 directly nested in combined parallel. Instead, for linear
2659 with copyin/copyout, if in combined for simd loop, make decl
2660 firstprivate/lastprivate on OMP_FOR.
2661 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
2662 expand_omp_for_static_chunk): When setting endvar, also set
2663 fd->loop.v to the same value.
2665 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2667 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
2669 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
2671 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
2674 2014-05-02 Marek Polacek <polacek@redhat.com>
2676 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
2678 2014-05-02 Kito Cheng <kito@0xlab.org>
2680 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
2681 to a C expression marco.
2682 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
2683 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
2684 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
2685 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
2686 HONOR_REG_ALLOC_ORDER.
2687 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
2689 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2691 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
2693 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2695 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
2697 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
2699 * tree-if-conv.c (is_cond_scalar_reduction): New function.
2700 (convert_scalar_cond_reduction): Likewise.
2701 (predicate_scalar_phi): Add recognition and transformation
2702 of simple conditioanl reduction to be vectorizable.
2704 2014-05-01 Marek Polacek <polacek@redhat.com>
2707 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
2709 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
2711 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
2712 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
2713 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
2714 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
2715 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
2716 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
2717 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
2718 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
2720 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
2722 * config/arc/arc.opt (mlra): Move comment above option name
2723 to avoid mis-parsing as language options.
2725 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2727 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
2728 * config/sol2.h: ... here.
2729 * config/sol2-10.h: Remove.
2731 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
2732 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
2733 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
2734 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
2735 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
2736 * config/sol2.h: ... here.
2737 (SECTION_NAME_FORMAT): Don't redefine.
2738 (STARTFILE_ARCH32_SPEC): Rename to ...
2739 (STARTFILE_ARCH_SPEC): ... this.
2740 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
2741 * config/sparc/sol2.h: ... here.
2742 (SECTION_NAME_FORMAT): Don't undef.
2743 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
2744 (SUBTARGET_EXTRA_SPECS): Remove.
2745 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
2747 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
2748 (MD_STARTFILE_PREFIX): Remove.
2749 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
2750 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
2751 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
2752 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
2753 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
2754 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
2755 * config/i386/sol2.h: ... here.
2756 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
2757 * config/i386/sol2-bi.h: Remove.
2758 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
2759 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
2761 * config/i386/t-sol2-64: Rename to ...
2762 * config/i386/t-sol2: ... this.
2763 * config/sparc/t-sol2-64: Rename to ...
2764 * config/sparc/t-sol2: ... this.
2766 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
2767 sol2_tm_file_head, sol2_tm_file_tail.
2768 Include ${cpu_type}/sol2.h before sol2.h.
2770 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
2771 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
2772 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
2773 Reflect i386/t-sol2-64 renaming.
2774 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
2775 Reflect sparc/t-sol2-64 renaming.
2777 2014-04-30 Richard Biener <rguenther@suse.de>
2779 * passes.c (execute_function_todo): Move TODO_verify_stmts
2780 and TODO_verify_ssa under the TODO_verify_il umbrella.
2781 * tree-ssa.h (verify_ssa): Adjust prototype.
2782 * tree-ssa.c (verify_ssa): Add parameter to tell whether
2783 we should verify SSA operands.
2784 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
2785 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
2786 whether we should verify whether not throwing stmts have EH info.
2787 * graphite-scop-detection.c (create_sese_edges): Adjust.
2788 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
2789 * tree-eh.c (lower_try_finally_switch): Do not add the
2790 default case label twice.
2792 2014-04-30 Marek Polacek <polacek@redhat.com>
2794 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
2795 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
2796 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
2797 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
2799 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
2801 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
2802 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
2803 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
2804 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
2805 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
2806 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
2807 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
2808 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
2810 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2812 * tree-cfg.c (dump_function_to_file): Dump the return type of
2813 functions, in a line to itself before the function body, mimicking
2814 the layout of a C function.
2816 2014-04-29 Jakub Jelinek <jakub@redhat.com>
2818 PR tree-optimization/60971
2819 * tree-tailcall.c (process_assignment): Reject conversions which
2822 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
2824 * calls.c (initialize_argument_information): Always treat
2825 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
2826 (expand_call): Likewise.
2827 (emit_library_call_calue_1): Likewise.
2828 * expr.c (PUSH_ARGS_REVERSED): Do not define.
2829 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
2832 2014-04-29 Nick Clifton <nickc@redhat.com>
2834 * config/msp430/msp430.md (umulsidi): Fix typo.
2835 (mulhisi3): Enable even inside interrupt handlers.
2836 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
2837 bigger return address pushed in large mode.
2839 2014-04-29 Nick Clifton <nickc@redhat.com>
2841 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
2842 (arc_init_reg_tables): Use a machine_mode enum to iterate over
2844 * config/m32r/m32r.c (init_reg_tables): Likewise.
2845 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
2846 enum to hold the modes.
2848 2014-04-29 Richard Biener <rguenther@suse.de>
2850 * dominance.c (free_dominance_info): Add overload with
2852 (dom_info_state): Likewise.
2853 (dom_info_available_p): Likewise.
2854 * basic-block.h (free_dominance_info, dom_info_state,
2855 dom_info_available_p): Declare overloads.
2856 * passes.c (execute_function_todo): Verify that verifiers
2857 don't change dominator info state. Drop dominator info
2858 for IPA pass invocations.
2859 * cgraph.c (release_function_body): Restore asserts that
2860 dominator information is released.
2862 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
2864 * doc/invoke.texi: Fix typo.
2865 * tree-vrp.c: Fix typos.
2866 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
2868 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2870 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
2872 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2874 * config/aarch64/aarch64-builtins.c
2875 (aarch64_types_storestruct_lane_qualifiers): New.
2876 (TYPES_STORESTRUCT_LANE): Likewise.
2877 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
2878 (st3_lane): Likewise.
2879 (st4_lane): Likewise.
2880 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
2881 (vec_store_lanesci_lane<mode>): Likewise.
2882 (vec_store_lanesxi_lane<mode>): Likewise.
2883 (aarch64_st2_lane<VQ:mode>): Likewise.
2884 (aarch64_st3_lane<VQ:mode>): Likewise.
2885 (aarch64_st4_lane<VQ:mode>): Likewise.
2886 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
2887 * config/aarch64/arm_neon.h
2888 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
2889 use new macro arguments.
2890 (__ST3_LANE_FUNC): Likewise.
2891 (__ST4_LANE_FUNC): Likewise.
2892 * config/aarch64/iterators.md (V_TWO_ELEM): New.
2893 (V_THREE_ELEM): Likewise.
2894 (V_FOUR_ELEM): Likewise.
2896 2014-04-28 David Malcolm <dmalcolm@redhat.com>
2898 * doc/gimple.texi: Replace the description of the now-defunct
2899 union gimple_statement_d with a diagram showing the
2900 gimple_statement_base class hierarchy and its relationships to
2901 the GSS_ and GIMPLE_ enums.
2903 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
2905 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
2906 * config/aarch64/aarch64.c
2907 (aarch64_cannot_change_mode_class): Weaken conditions.
2908 (aarch64_modes_tieable_p): New.
2909 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
2911 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
2913 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2914 (loadsync_<mode>): Change mode.
2915 (load_quadpti, store_quadpti): New.
2916 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2917 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2919 2014-04-28 Martin Jambor <mjambor@suse.cz>
2921 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
2922 same alias type as the original statement.
2923 (subreplacement_assignment_data): New type.
2924 (handle_unscalarized_data_in_subtree): New type of parameter,
2925 generate new memory accesses with same alias type as the original
2927 (load_assign_lhs_subreplacements): Likewise.
2928 (sra_modify_constructor_assign): Generate new memory accesses with
2929 same alias type as the original statement.
2931 2014-04-28 Richard Biener <rguenther@suse.de>
2933 * tree-pass.h (TODO_verify_il): Define.
2934 (TODO_verify_all): Complete properly.
2935 * passes.c (execute_function_todo): Move existing loop-closed
2936 SSA verification under TODO_verify_il.
2937 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
2938 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
2939 Fix tree sharing issue.
2941 2014-04-28 Richard Biener <rguenther@suse.de>
2944 * builtins.def (DEF_C11_BUILTIN): Add.
2945 (BUILT_IN_ALIGNED_ALLOC): Likewise.
2946 * coretypes.h (enum function_class): Add function_c11_misc.
2947 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
2948 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
2949 (call_may_clobber_ref_p_1): Likewise.
2950 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
2951 (mark_all_reaching_defs_necessary_1): Likewise.
2952 (propagate_necessity): Likewise.
2953 (eliminate_unnecessary_stmts): Likewise.
2954 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
2956 2014-04-28 Richard Biener <rguenther@suse.de>
2958 * tree-vrp.c (vrp_var_may_overflow): Remove.
2959 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
2960 with overflow immediately bump to one before that value and
2961 let iteration figure out overflow status.
2963 2014-04-28 Richard Biener <rguenther@suse.de>
2965 * configure.ac: Do valgrind header checks unconditionally.
2966 Add --enable-valgrind-annotations.
2967 * system.h: Guard valgrind header inclusion with
2968 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
2969 * alloc-pool.c (pool_alloc, pool_free): Use
2970 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
2971 to guard possibly dead code.
2972 * config.in: Regenerated.
2973 * configure: Likewise.
2975 2014-04-28 Jeff Law <law@redhat.com>
2977 PR tree-optimization/60902
2978 * tree-ssa-threadedge.c
2979 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
2980 over real defs when invalidating outputs from statements that do not
2981 produce useful outputs for threading.
2983 2014-04-28 Richard Biener <rguenther@suse.de>
2985 PR tree-optimization/60979
2986 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
2987 SCOPs that end in a block with a successor with abnormal
2990 2014-04-28 Richard Biener <rguenther@suse.de>
2992 * tree-pass.h (execute_pass_list): Adjust prototype.
2993 * passes.c (pass_manager::execute_early_local_passes): Adjust.
2994 (do_per_function): Change callback signature, push all actual
2995 work to the callbals.
2996 (do_per_function_toporder): Likewise.
2997 (execute_function_dump): Adjust.
2998 (execute_function_todo): Likewise.
2999 (clear_last_verified): Likewise.
3000 (verify_curr_properties): Likewise.
3001 (update_properties_after_pass): Likewise.
3002 (execute_pass_list_1): Split out from ...
3003 (execute_pass_list): ... here. Adjust.
3004 (execute_ipa_pass_list): Likewise.
3005 * cgraphunit.c (cgraph_add_new_function): Adjust.
3006 (analyze_function): Likewise.
3007 (expand_function): Likewise.
3008 * cgraph.c (release_function_body): Free dominance info
3009 here instead of asserting it was magically freed elsewhere.
3011 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
3013 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
3014 * configure: Regenerate.
3015 * config/sparc/sparc.opt (muser-mode): New option.
3016 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
3018 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
3019 * doc/invoke.texi (SPARC options): Document -muser-mode.
3021 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
3023 * cselib.c (find_slot_memmode): Delete.
3024 (cselib_hasher): Change compare_type to a struct.
3025 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
3027 (preserve_constants_and_equivs): Adjust for new compare_type.
3028 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
3029 (wrap_constant): Delete.
3030 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
3032 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
3034 * doc/install.texi (Building with profile feedback): Remove
3037 2014-04-26 Tom de Vries <tom@codesourcery.com>
3039 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
3042 2014-04-25 Cary Coutant <ccoutant@google.com>
3045 * dwarf2out.c (should_move_die_to_comdat): A type definition
3046 can contain a subprogram definition, but don't move it to a
3048 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
3049 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
3051 (clone_tree_hash): Rename to...
3052 (clone_tree_partial): ...this; change callers. Copy
3053 DW_TAG_subprogram DIEs as declarations.
3054 (copy_decls_walk): Don't copy children of a declaration into a
3057 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
3060 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
3063 2014-04-25 Jiong Wang <jiong.wang@arm.com>
3065 * config/arm/predicates.md (call_insn_operand): Add long_call check.
3066 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
3068 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
3071 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3073 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
3075 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3077 PR tree-optimization/60930
3078 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
3079 creating a multiply candidate by folding two constant
3080 multiplicands when the result overflows.
3082 2014-04-25 Jakub Jelinek <jakub@redhat.com>
3084 PR tree-optimization/60960
3085 * tree-vect-generic.c (expand_vector_operation): Only call
3086 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
3088 2014-04-25 Tom de Vries <tom@codesourcery.com>
3090 * expr.c (clobber_reg_mode): New function.
3091 * expr.h (clobber_reg): New function.
3093 2014-04-25 Tom de Vries <tom@codesourcery.com>
3095 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
3098 2014-04-25 Radovan Obradovic <robradovic@mips.com>
3099 Tom de Vries <tom@codesourcery.com>
3101 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
3103 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
3104 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
3105 new argument to find_all_hard_reg_sets call.
3107 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3109 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
3110 Use HOST_WIDE_INT_C for mask literal.
3111 (aarch_rev16_shleft_mask_imm_p): Likewise.
3113 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
3116 * config/sparc/sparc.md (ashlsi3_extend): Delete.
3118 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
3120 PR preprocessor/56540
3121 * config/i386/i386-c.c (ix86_target_macros): Define
3122 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
3124 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3126 * configure.ac (tga_func): Remove.
3127 (LIB_TLS_SPEC): Remove.
3128 * configure: Regenerate.
3129 * config.in: Regenerate.
3130 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
3132 2014-04-25 Richard Biener <rguenther@suse.de>
3135 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
3136 call stmt use/clobber sets during stmt walk instead of
3137 walking the possibly incomplete set of caller edges.
3139 2014-04-25 Richard Biener <rguenther@suse.de>
3142 * passes.c (apply_ipa_transforms): Inline into only caller ...
3143 (execute_one_pass): ... here. Properly bring in function
3144 bodies for nodes we want to apply IPA transforms to.
3146 2014-04-24 Cong Hou <congh@google.com>
3148 PR tree-optimization/60896
3149 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
3150 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
3151 (vect_mark_pattern_stmts): Set the def type of all statements in
3152 PATTERN_DEF_SEQ as vect_internal_def.
3154 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
3156 * doc/extend.texi (PowerPC Built-in Functions): Document new
3157 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
3158 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
3160 * config/rs6000/predicates.md (const_0_to_3_operand): New
3161 predicate to match 0..3 integer constants.
3163 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
3164 to support adding miscellaneous builtin functions.
3165 (BU_DFP_MISC_2): Likewise.
3166 (BU_P7_MISC_1): Likewise.
3167 (BU_P7_MISC_2): Likewise.
3168 (BU_P8V_MISC_3): Likewise.
3169 (BU_MISC_1): Likewise.
3170 (BU_MISC_2): Likewise.
3171 (DIVWE): Add extended divide builtin functions.
3174 (DIVWEUO): Likewise.
3178 (DIVDEUO): Likewise.
3179 (DXEX): Add decimal floating-point builtin functions.
3182 (DDEDPDQ): Likewise.
3184 (DENBCDQ): Likewise.
3191 (CDTBCD): Add new BCD builtin functions.
3195 (BCDADD_LT): Likewise.
3196 (BCDADD_EQ): Likewise.
3197 (BCDADD_GT): Likewise.
3198 (BCDADD_OV): Likewise.
3200 (BCDSUB_LT): Likewise.
3201 (BCDSUB_EQ): Likewise.
3202 (BCDSUB_GT): Likewise.
3203 (BCDSUB_OV): Likewise.
3204 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
3205 (UNPACK_TD): Likewise.
3206 (PACK_TF): Likewise.
3207 (UNPACK_TF): Likewise.
3208 (UNPACK_TF_0): Likewise.
3209 (UNPACK_TF_1): Likewise.
3210 (PACK_V1TI): Likewise.
3211 (UNPACK_V1TI): Likewise.
3213 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
3214 support for decimal floating point builtin functions.
3215 (rs6000_expand_ternop_builtin): Add checks for the new builtin
3216 functions that take constant arguments.
3217 (rs6000_invalid_builtin): Add decimal floating point builtin support.
3218 (rs6000_init_builtins): Setup long double, _Decimal64, and
3219 _Decimal128 types for new builtin functions.
3220 (builtin_function_type): Set the unsigned flags appropriately for
3221 the new builtin functions.
3222 (rs6000_opt_masks): Add support for decimal floating point builtin
3225 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
3226 floating point builtin functions.
3227 (RS6000_BTM_COMMON): Likewise.
3228 (RS6000_BTI_long_double): Likewise.
3229 (RS6000_BTI_dfloat64): Likewise.
3230 (RS6000_BTI_dfloat128): Likewise.
3231 (long_double_type_internal_node): Likewise.
3232 (dfloat64_type_internal_node): Likewise.
3233 (dfloat128_type_internal_node): Likewise.
3235 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
3236 2.07 bcd arithmetic instructions.
3237 (UNSPEC_BCDSUB): Likewise.
3238 (UNSPEC_BCD_OVERFLOW): Likewise.
3239 (UNSPEC_BCD_ADD_SUB): Likewise.
3240 (bcd_add_sub): Likewise.
3241 (BCD_TEST): Likewise.
3242 (bcd<bcd_add_sub>): Likewise.
3243 (bcd<bcd_add_sub>_test): Likewise.
3244 (bcd<bcd_add_sub>_test2): Likewise.
3245 (bcd<bcd_add_sub>_<code>): Likewise.
3246 (peephole2 for combined bcd ops): Likewise.
3248 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
3249 decimal floating point builtin functions.
3250 (UNSPEC_DENBCD): Likewise.
3251 (UNSPEC_DXEX): Likewise.
3252 (UNSPEC_DIEX): Likewise.
3253 (UNSPEC_DSCLI): Likewise.
3254 (UNSPEC_DSCRI): Likewise.
3255 (D64_D128): Likewise.
3256 (dfp_suffix): Likewise.
3257 (dfp_ddedpd_<mode>): Likewise.
3258 (dfp_denbcd_<mode>): Likewise.
3259 (dfp_dxex_<mode>): Likewise.
3260 (dfp_diex_<mode>): Likewise.
3261 (dfp_dscli_<mode>): Likewise.
3262 (dfp_dscri_<mode>): Likewise.
3264 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
3266 (UNSPEC_CDTBCD): Likewise.
3267 (UNSPEC_CBCDTD): Likewise.
3268 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
3269 (UNSPEC_DIVEO): Likewise.
3270 (UNSPEC_DIVEU): Likewise.
3271 (UNSPEC_DIVEUO): Likewise.
3272 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
3273 pack/unpack 128-bit types.
3274 (UNSPEC_PACK_128BIT): Likewise.
3275 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
3276 (udiv<mode>3): Use idiv_ldiv mode attribute.
3277 (div<mode>3): Likewise.
3278 (addg6s): Add new BCD builtin functions.
3281 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
3282 (div_extend): Likewise.
3283 (div<div_extend>_<mode>"): Likewise.
3284 (FP128_64): Add support for new builtin functions to pack/unpack
3286 (unpack<mode>): Likewise.
3287 (unpacktf_0): Likewise.
3288 (unpacktf_1): Likewise.
3289 (unpack<mode>_dm): Likewise.
3290 (unpack<mode>_nodm): Likewise.
3291 (pack<mode>): Likewise.
3292 (unpackv1ti): Likewise.
3293 (packv1ti): Likewise.
3295 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
3297 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
3300 2014-04-24 Jakub Jelinek <jakub@redhat.com>
3302 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
3303 * gimplify.c (omp_is_private): Change last argument's type to int.
3304 Only diagnose lastprivate if the simd argument is 1, only diagnose
3305 linear if the simd argument is 2.
3306 (gimplify_omp_for): Adjust omp_is_private callers. When adding
3307 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
3308 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
3309 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
3310 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
3311 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
3312 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
3313 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
3314 * tree-nested.c (convert_nonlocal_omp_clauses,
3315 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
3317 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
3320 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
3323 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
3325 * flag-types.h (enum ivar_visibility): Add.
3327 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
3329 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
3330 function * argument.
3332 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
3334 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
3336 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3337 Tom de Vries <tom@codesourcery.com>
3339 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
3340 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
3342 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
3343 * emit-rtl.c (try_split): Same.
3345 2014-04-24 Radovan Obradovic <robradovic@mips.com>
3346 Tom de Vries <tom@codesourcery.com>
3348 * common.opt (fuse-caller-save): New option.
3350 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
3352 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
3353 elements for big-endian.
3355 2014-04-24 Richard Biener <rguenther@suse.de>
3357 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
3358 during TER and instead use the sepops interface for expanding
3359 non-GIMPLE_SINGLE_RHS.
3361 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3363 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
3364 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
3366 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3368 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
3369 assembler 64-bit option.
3370 * configure: Regenerate.
3372 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3374 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
3375 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
3376 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
3377 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
3378 (TARGET_CRYPTO): Take TARGET_SIMD into account.
3380 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3382 * config/aarch64/aarch64-builtins.c
3383 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
3384 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
3385 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
3386 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
3388 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
3389 (Vrevsuff): New mode attribute.
3391 2014-04-24 Terry Guo <terry.guo@arm.com>
3393 * config/arm/arm.h (machine_function): Define variable
3394 after_arm_reorg here.
3395 * config/arm/arm.c (after_arm_reorg): Remove the definition.
3396 (arm_split_constant): Update the way to access variable
3399 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
3401 2014-04-23 Tom de Vries <tom@codesourcery.com>
3403 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
3405 2014-04-23 David Malcolm <dmalcolm@redhat.com>
3407 * is-a.h: Update comments to reflect the following changes to the
3408 "pointerness" of the API, making the template parameter match the
3409 return type, allowing use of is-a.h with typedefs of pointers.
3410 (is_a_helper::cast): Return a T rather then a pointer to a T, so
3411 that the return type matches the parameter to the is_a_helper.
3413 (dyn_cast): Likewise.
3415 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
3416 pointer from the is-a.h API.
3418 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
3419 (is_a_helper <cgraph_node *>::test): ...this, matching change to
3421 (is_a_helper <varpool_node>::test): Likewise, convert to...
3422 (is_a_helper <varpool_node *>::test): ...this.
3424 (varpool_first_variable): Update for removal of implicit pointer
3425 from the is-a.h API.
3426 (varpool_next_variable): Likewise.
3427 (varpool_first_static_initializer): Likewise.
3428 (varpool_next_static_initializer): Likewise.
3429 (varpool_first_defined_variable): Likewise.
3430 (varpool_next_defined_variable): Likewise.
3431 (cgraph_first_defined_function): Likewise.
3432 (cgraph_next_defined_function): Likewise.
3433 (cgraph_first_function): Likewise.
3434 (cgraph_next_function): Likewise.
3435 (cgraph_first_function_with_gimple_body): Likewise.
3436 (cgraph_next_function_with_gimple_body): Likewise.
3437 (cgraph_alias_target): Likewise.
3438 (varpool_alias_target): Likewise.
3439 (cgraph_function_or_thunk_node): Likewise.
3440 (varpool_variable_node): Likewise.
3441 (symtab_real_symbol_p): Likewise.
3442 * cgraphunit.c (referred_to_p): Likewise.
3443 (analyze_functions): Likewise.
3444 (handle_alias_pairs): Likewise.
3445 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
3446 * gimple-ssa.h (gimple_vuse_op): Likewise.
3447 (gimple_vdef_op): Likewise.
3448 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
3449 * gimple.c (gimple_build_asm_1): Likewise.
3450 (gimple_build_try): Likewise.
3451 (gimple_build_resx): Likewise.
3452 (gimple_build_eh_dispatch): Likewise.
3453 (gimple_build_omp_for): Likewise.
3454 (gimple_omp_for_set_clauses): Likewise.
3456 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
3457 (is_a_helper <gimple_statement_asm *>::test): ...this.
3458 (is_a_helper <gimple_statement_bind>::test): Convert to...
3459 (is_a_helper <gimple_statement_bind *>::test): ...this.
3460 (is_a_helper <gimple_statement_call>::test): Convert to...
3461 (is_a_helper <gimple_statement_call *>::test): ...this.
3462 (is_a_helper <gimple_statement_catch>::test): Convert to...
3463 (is_a_helper <gimple_statement_catch *>::test): ...this.
3464 (is_a_helper <gimple_statement_resx>::test): Convert to...
3465 (is_a_helper <gimple_statement_resx *>::test): ...this.
3466 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
3467 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
3468 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
3469 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
3470 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
3471 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
3472 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
3473 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
3474 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
3475 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
3476 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
3477 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
3478 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
3479 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
3480 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
3481 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
3482 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
3483 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
3484 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
3485 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
3486 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
3487 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
3488 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
3489 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
3490 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
3491 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
3492 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
3493 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
3494 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
3495 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
3496 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
3497 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
3498 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
3499 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
3500 (is_a_helper <gimple_statement_phi>::test): Convert to...
3501 (is_a_helper <gimple_statement_phi *>::test): ...this.
3502 (is_a_helper <gimple_statement_transaction>::test): Convert to...
3503 (is_a_helper <gimple_statement_transaction *>::test): ...this.
3504 (is_a_helper <gimple_statement_try>::test): Convert to...
3505 (is_a_helper <gimple_statement_try *>::test): ...this.
3506 (is_a_helper <gimple_statement_wce>::test): Convert to...
3507 (is_a_helper <gimple_statement_wce *>::test): ...this.
3508 (is_a_helper <const gimple_statement_asm>::test): Convert to...
3509 (is_a_helper <const gimple_statement_asm *>::test): ...this.
3510 (is_a_helper <const gimple_statement_bind>::test): Convert to...
3511 (is_a_helper <const gimple_statement_bind *>::test): ...this.
3512 (is_a_helper <const gimple_statement_call>::test): Convert to...
3513 (is_a_helper <const gimple_statement_call *>::test): ...this.
3514 (is_a_helper <const gimple_statement_catch>::test): Convert to...
3515 (is_a_helper <const gimple_statement_catch *>::test): ...this.
3516 (is_a_helper <const gimple_statement_resx>::test): Convert to...
3517 (is_a_helper <const gimple_statement_resx *>::test): ...this.
3518 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
3519 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
3520 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
3521 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
3522 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
3524 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
3526 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
3528 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
3530 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
3531 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
3532 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
3534 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
3535 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
3537 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
3538 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
3539 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
3540 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
3541 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
3542 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
3544 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
3545 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
3546 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
3547 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
3549 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
3550 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
3551 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
3552 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
3553 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
3554 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
3555 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
3556 (is_a_helper <const gimple_statement_phi>::test): Convert to...
3557 (is_a_helper <const gimple_statement_phi *>::test): ...this.
3558 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
3559 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
3560 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
3561 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
3562 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
3563 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
3564 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
3566 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
3568 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
3569 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
3571 (gimple_use_ops): Update for removal of implicit pointer from the
3573 (gimple_set_use_ops): Likewise.
3574 (gimple_vuse): Likewise.
3575 (gimple_vdef): Likewise.
3576 (gimple_vuse_ptr): Likewise.
3577 (gimple_vdef_ptr): Likewise.
3578 (gimple_set_vuse): Likewise.
3579 (gimple_set_vdef): Likewise.
3580 (gimple_omp_return_set_lhs): Likewise.
3581 (gimple_omp_return_lhs): Likewise.
3582 (gimple_omp_return_lhs_ptr): Likewise.
3583 (gimple_call_fntype): Likewise.
3584 (gimple_call_set_fntype): Likewise.
3585 (gimple_call_set_internal_fn): Likewise.
3586 (gimple_call_use_set): Likewise.
3587 (gimple_call_clobber_set): Likewise.
3588 (gimple_bind_vars): Likewise.
3589 (gimple_bind_set_vars): Likewise.
3590 (gimple_bind_body_ptr): Likewise.
3591 (gimple_bind_set_body): Likewise.
3592 (gimple_bind_add_stmt): Likewise.
3593 (gimple_bind_block): Likewise.
3594 (gimple_bind_set_block): Likewise.
3595 (gimple_asm_ninputs): Likewise.
3596 (gimple_asm_noutputs): Likewise.
3597 (gimple_asm_nclobbers): Likewise.
3598 (gimple_asm_nlabels): Likewise.
3599 (gimple_asm_input_op): Likewise.
3600 (gimple_asm_input_op_ptr): Likewise.
3601 (gimple_asm_output_op): Likewise.
3602 (gimple_asm_output_op_ptr): Likewise.
3603 (gimple_asm_set_output_op): Likewise.
3604 (gimple_asm_clobber_op): Likewise.
3605 (gimple_asm_set_clobber_op): Likewise.
3606 (gimple_asm_label_op): Likewise.
3607 (gimple_asm_set_label_op): Likewise.
3608 (gimple_asm_string): Likewise.
3609 (gimple_catch_types): Likewise.
3610 (gimple_catch_types_ptr): Likewise.
3611 (gimple_catch_handler_ptr): Likewise.
3612 (gimple_catch_set_types): Likewise.
3613 (gimple_catch_set_handler): Likewise.
3614 (gimple_eh_filter_types): Likewise.
3615 (gimple_eh_filter_types_ptr): Likewise.
3616 (gimple_eh_filter_failure_ptr): Likewise.
3617 (gimple_eh_filter_set_types): Likewise.
3618 (gimple_eh_filter_set_failure): Likewise.
3619 (gimple_eh_must_not_throw_fndecl): Likewise.
3620 (gimple_eh_must_not_throw_set_fndecl): Likewise.
3621 (gimple_eh_else_n_body_ptr): Likewise.
3622 (gimple_eh_else_e_body_ptr): Likewise.
3623 (gimple_eh_else_set_n_body): Likewise.
3624 (gimple_eh_else_set_e_body): Likewise.
3625 (gimple_try_eval_ptr): Likewise.
3626 (gimple_try_cleanup_ptr): Likewise.
3627 (gimple_try_set_eval): Likewise.
3628 (gimple_try_set_cleanup): Likewise.
3629 (gimple_wce_cleanup_ptr): Likewise.
3630 (gimple_wce_set_cleanup): Likewise.
3631 (gimple_phi_capacity): Likewise.
3632 (gimple_phi_num_args): Likewise.
3633 (gimple_phi_result): Likewise.
3634 (gimple_phi_result_ptr): Likewise.
3635 (gimple_phi_set_result): Likewise.
3636 (gimple_phi_arg): Likewise.
3637 (gimple_phi_set_arg): Likewise.
3638 (gimple_resx_region): Likewise.
3639 (gimple_resx_set_region): Likewise.
3640 (gimple_eh_dispatch_region): Likewise.
3641 (gimple_eh_dispatch_set_region): Likewise.
3642 (gimple_omp_critical_name): Likewise.
3643 (gimple_omp_critical_name_ptr): Likewise.
3644 (gimple_omp_critical_set_name): Likewise.
3645 (gimple_omp_for_clauses): Likewise.
3646 (gimple_omp_for_clauses_ptr): Likewise.
3647 (gimple_omp_for_set_clauses): Likewise.
3648 (gimple_omp_for_collapse): Likewise.
3649 (gimple_omp_for_index): Likewise.
3650 (gimple_omp_for_index_ptr): Likewise.
3651 (gimple_omp_for_set_index): Likewise.
3652 (gimple_omp_for_initial): Likewise.
3653 (gimple_omp_for_initial_ptr): Likewise.
3654 (gimple_omp_for_set_initial): Likewise.
3655 (gimple_omp_for_final): Likewise.
3656 (gimple_omp_for_final_ptr): Likewise.
3657 (gimple_omp_for_set_final): Likewise.
3658 (gimple_omp_for_incr): Likewise.
3659 (gimple_omp_for_incr_ptr): Likewise.
3660 (gimple_omp_for_set_incr): Likewise.
3661 (gimple_omp_for_pre_body_ptr): Likewise.
3662 (gimple_omp_for_set_pre_body): Likewise.
3663 (gimple_omp_parallel_clauses): Likewise.
3664 (gimple_omp_parallel_clauses_ptr): Likewise.
3665 (gimple_omp_parallel_set_clauses): Likewise.
3666 (gimple_omp_parallel_child_fn): Likewise.
3667 (gimple_omp_parallel_child_fn_ptr): Likewise.
3668 (gimple_omp_parallel_set_child_fn): Likewise.
3669 (gimple_omp_parallel_data_arg): Likewise.
3670 (gimple_omp_parallel_data_arg_ptr): Likewise.
3671 (gimple_omp_parallel_set_data_arg): Likewise.
3672 (gimple_omp_task_clauses): Likewise.
3673 (gimple_omp_task_clauses_ptr): Likewise.
3674 (gimple_omp_task_set_clauses): Likewise.
3675 (gimple_omp_task_child_fn): Likewise.
3676 (gimple_omp_task_child_fn_ptr): Likewise.
3677 (gimple_omp_task_set_child_fn): Likewise.
3678 (gimple_omp_task_data_arg): Likewise.
3679 (gimple_omp_task_data_arg_ptr): Likewise.
3680 (gimple_omp_task_set_data_arg): Likewise.
3681 (gimple_omp_taskreg_clauses): Likewise.
3682 (gimple_omp_taskreg_clauses_ptr): Likewise.
3683 (gimple_omp_taskreg_set_clauses): Likewise.
3684 (gimple_omp_taskreg_child_fn): Likewise.
3685 (gimple_omp_taskreg_child_fn_ptr): Likewise.
3686 (gimple_omp_taskreg_set_child_fn): Likewise.
3687 (gimple_omp_taskreg_data_arg): Likewise.
3688 (gimple_omp_taskreg_data_arg_ptr): Likewise.
3689 (gimple_omp_taskreg_set_data_arg): Likewise.
3690 (gimple_omp_task_copy_fn): Likewise.
3691 (gimple_omp_task_copy_fn_ptr): Likewise.
3692 (gimple_omp_task_set_copy_fn): Likewise.
3693 (gimple_omp_task_arg_size): Likewise.
3694 (gimple_omp_task_arg_size_ptr): Likewise.
3695 (gimple_omp_task_set_arg_size): Likewise.
3696 (gimple_omp_task_arg_align): Likewise.
3697 (gimple_omp_task_arg_align_ptr): Likewise.
3698 (gimple_omp_task_set_arg_align): Likewise.
3699 (gimple_omp_single_clauses): Likewise.
3700 (gimple_omp_single_clauses_ptr): Likewise.
3701 (gimple_omp_single_set_clauses): Likewise.
3702 (gimple_omp_target_clauses): Likewise.
3703 (gimple_omp_target_clauses_ptr): Likewise.
3704 (gimple_omp_target_set_clauses): Likewise.
3705 (gimple_omp_target_child_fn): Likewise.
3706 (gimple_omp_target_child_fn_ptr): Likewise.
3707 (gimple_omp_target_set_child_fn): Likewise.
3708 (gimple_omp_target_data_arg): Likewise.
3709 (gimple_omp_target_data_arg_ptr): Likewise.
3710 (gimple_omp_target_set_data_arg): Likewise.
3711 (gimple_omp_teams_clauses): Likewise.
3712 (gimple_omp_teams_clauses_ptr): Likewise.
3713 (gimple_omp_teams_set_clauses): Likewise.
3714 (gimple_omp_sections_clauses): Likewise.
3715 (gimple_omp_sections_clauses_ptr): Likewise.
3716 (gimple_omp_sections_set_clauses): Likewise.
3717 (gimple_omp_sections_control): Likewise.
3718 (gimple_omp_sections_control_ptr): Likewise.
3719 (gimple_omp_sections_set_control): Likewise.
3720 (gimple_omp_for_set_cond): Likewise.
3721 (gimple_omp_for_cond): Likewise.
3722 (gimple_omp_atomic_store_set_val): Likewise.
3723 (gimple_omp_atomic_store_val): Likewise.
3724 (gimple_omp_atomic_store_val_ptr): Likewise.
3725 (gimple_omp_atomic_load_set_lhs): Likewise.
3726 (gimple_omp_atomic_load_lhs): Likewise.
3727 (gimple_omp_atomic_load_lhs_ptr): Likewise.
3728 (gimple_omp_atomic_load_set_rhs): Likewise.
3729 (gimple_omp_atomic_load_rhs): Likewise.
3730 (gimple_omp_atomic_load_rhs_ptr): Likewise.
3731 (gimple_omp_continue_control_def): Likewise.
3732 (gimple_omp_continue_control_def_ptr): Likewise.
3733 (gimple_omp_continue_set_control_def): Likewise.
3734 (gimple_omp_continue_control_use): Likewise.
3735 (gimple_omp_continue_control_use_ptr): Likewise.
3736 (gimple_omp_continue_set_control_use): Likewise.
3737 (gimple_transaction_body_ptr): Likewise.
3738 (gimple_transaction_label): Likewise.
3739 (gimple_transaction_label_ptr): Likewise.
3740 (gimple_transaction_set_body): Likewise.
3741 (gimple_transaction_set_label): Likewise.
3743 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
3744 * ipa-inline-analysis.c (inline_write_summary): Likewise.
3745 * ipa-ref.c (ipa_record_reference): Likewise.
3746 * ipa-reference.c (analyze_function): Likewise.
3747 (ipa_reference_write_optimization_summary): Likewise.
3748 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
3749 (address_taken_from_non_vtable_p): Likewise.
3750 (comdat_can_be_unshared_p_1): Likewise.
3751 * lto-cgraph.c (lto_output_ref): Likewise.
3752 (add_references): Likewise.
3753 (compute_ltrans_boundary): Likewise.
3754 (output_symtab): Likewise.
3755 (input_ref): Likewise.
3756 (input_cgraph_1): Likewise.
3757 (output_cgraph_opt_summary): Likewise.
3758 * lto-streamer-out.c (lto_output): Likewise.
3759 (output_symbol_p): Likewise.
3760 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
3761 (lsei_start_function_in_partition): Likewise.
3762 (lsei_next_variable_in_partition): Likewise.
3763 (lsei_start_variable_in_partition): Likewise.
3764 * symtab.c (insert_to_assembler_name_hash): Likewise.
3765 (unlink_from_assembler_name_hash): Likewise.
3766 (symtab_unregister_node): Likewise.
3767 (symtab_remove_node): Likewise.
3768 (dump_symtab_node): Likewise.
3769 (verify_symtab_base): Likewise.
3770 (verify_symtab_node): Likewise.
3771 (symtab_make_decl_local): Likewise.
3772 (symtab_alias_ultimate_target): Likewise.
3773 (symtab_resolve_alias): Likewise.
3774 (symtab_get_symbol_partitioning_class): Likewise.
3775 * tree-phinodes.c (allocate_phi_node): Likewise.
3776 (reserve_phi_args_for_new_edge): Likewise.
3777 (remove_phi_args): Likewise.
3778 * varpool.c (varpool_node_for_asm): Likewise.
3779 (varpool_remove_unreferenced_decls): Likewise.
3781 2014-04-23 Jeff Law <law@redhat.com>
3783 PR tree-optimization/60902
3784 * tree-ssa-threadedge.c
3785 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
3786 invalidate outputs from statements that do not produce useful
3787 outputs for threading.
3789 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3791 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
3792 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
3793 machine descriptions for Stack Smashing Protector.
3795 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
3797 * aarch64.md (<optab>_rol<mode>3): New pattern.
3798 (<optab>_rolsi3_uxtw): Likewise.
3799 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
3801 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
3803 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
3804 (arm_cortex_a12_tune): Likewise.
3806 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3808 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
3810 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3812 * config/arm/arm.md (arm_rev16si2): New pattern.
3813 (arm_rev16si2_alt): Likewise.
3814 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
3816 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3818 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
3819 (rev16<mode>2_alt): Likewise.
3820 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
3821 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
3822 (aarch_rev16_shleft_mask_imm_p): Likewise.
3823 (aarch_rev16_p_1): Likewise.
3824 (aarch_rev16_p): Likewise.
3825 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
3826 (aarch_rev16_shright_mask_imm_p): Likewise.
3827 (aarch_rev16_shleft_mask_imm_p): Likewise.
3829 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3831 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
3832 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
3834 (cortex_a53_extra_costs): Likewise.
3835 (cortex_a57_extra_costs): Likewise.
3836 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
3837 (cortexa7_extra_costs): Likewise.
3838 (cortexa8_extra_costs): Likewise.
3839 (cortexa12_extra_costs): Likewise.
3840 (cortexa15_extra_costs): Likewise.
3841 (v7m_extra_costs): Likewise.
3842 (arm_new_rtx_costs): Handle BSWAP.
3844 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3846 * config/arm/arm.c (cortexa8_extra_costs): New table.
3847 (arm_cortex_a8_tune): New tuning struct.
3848 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
3850 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3852 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
3854 2014-04-23 Richard Biener <rguenther@suse.de>
3856 * Makefile.in (OBJS): Remove loop-unswitch.o.
3857 * tree-pass.h (make_pass_rtl_unswitch): Remove.
3858 * passes.def (pass_rtl_unswitch): Likewise.
3859 * loop-init.c (gate_rtl_unswitch): Likewise.
3860 (rtl_unswitch): Likewise.
3861 (pass_data_rtl_unswitch): Likewise.
3862 (pass_rtl_unswitch): Likewise.
3863 (make_pass_rtl_unswitch): Likewise.
3864 * rtl.h (reversed_condition): Likewise.
3865 (compare_and_jump_seq): Likewise.
3866 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
3868 * loop-unroll.c (compare_and_jump_seq): Likewise.
3870 2014-04-23 Richard Biener <rguenther@suse.de>
3872 PR tree-optimization/60903
3873 * tree-ssa-loop-im.c (analyze_memory_references): Remove
3874 commented code block.
3875 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
3876 loop flags to newly created BBs and edges.
3878 2014-04-23 Nick Clifton <nickc@redhat.com>
3880 * config/msp430/msp430.c (msp430_handle_option): Move function
3882 (msp430_option_override): Simplify mcu and mcpu option handling.
3883 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
3884 support for -mhwmult command line option.
3885 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
3886 -mhwmult command line option.
3887 (msp430_hwmult_enabled): Delete.
3888 (msp43o_output_labelref): Add support for -mhwmult command line option.
3889 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
3890 (umulsidi3): Likewise.
3891 * config/msp430/msp430.opt (mmcu): Add Report attribute.
3892 (mcpu, mlarge, msmall): Likewise.
3893 (mhwmult): New option.
3894 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
3896 (msp430_is_f5_mcu): Remove prototype.
3897 (msp430_use_f5_series_hwmult): Add prototype.
3898 * config/msp430/msp430-opts.h: New file.
3899 * common/config/msp430: New directory.
3900 * common/config/msp430/msp430-common.c: New file.
3901 * config.gcc (msp430): Remove target_has_targetm_common.
3902 * doc/invoke.texi: Document -mhwmult command line option.
3904 2014-04-23 Nick Clifton <nickc@redhat.com>
3906 * config/i386/cygwin.h (ENDFILE_SPEC): Include
3907 default-manifest.o if it can be found in the search path.
3908 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
3910 2014-04-23 Terry Guo <terry.guo@arm.com>
3912 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
3914 2014-04-23 Richard Biener <rguenther@suse.de>
3917 * tree-inline.c (declare_return_variable): Use mark_addressable.
3919 2014-04-23 Richard Biener <rguenther@suse.de>
3922 * loop-init.c (loop_optimizer_init): Make sure to apply
3923 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
3925 2014-04-22 Jakub Jelinek <jakub@redhat.com>
3928 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
3930 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
3931 if flag_sanitize_undefined_trap_on_error.
3932 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
3933 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
3934 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
3935 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
3936 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
3937 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
3938 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
3939 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
3940 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
3941 * ubsan.c (ubsan_instrument_unreachable): Return
3942 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
3943 (ubsan_expand_null_ifn): Emit __builtin_trap ()
3944 if flag_sanitize_undefined_trap_on_error and
3945 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
3946 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
3947 instrument_bool_enum_load): Emit __builtin_trap () if
3948 flag_sanitize_undefined_trap_on_error and
3949 __builtin_handle_*_abort () if !flag_sanitize_recover.
3950 * doc/invoke.texi (-fsanitize-recover,
3951 -fsanitize-undefined-trap-on-error): Document.
3953 2014-04-22 Christian Bruel <christian.bruel@st.com>
3955 * config/sh/sh.md (mov<mode>): Replace movQIHI.
3956 Force immediates to SImode.
3958 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
3960 * config/nios2/nios2.md (UNSPEC_ROUND): New.
3962 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
3963 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
3964 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
3965 (nios2_fpu_insn): Add entry for round.
3966 (N2FPU_NO_ERRNO_P): Define.
3967 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
3969 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
3971 2014-04-22 Richard Henderson <rth@redhat.com>
3973 * config/aarch64/aarch64 (addti3, subti3): New expanders.
3974 (add<GPI>3_compare0): Remove leading * from name.
3975 (add<GPI>3_carryin): Likewise.
3976 (sub<GPI>3_compare0): Likewise.
3977 (sub<GPI>3_carryin): Likewise.
3978 (<su_optab>mulditi3): New expander.
3979 (multi3): New expander.
3980 (madd<GPI>): Remove leading * from name.
3982 2014-04-22 Martin Jambor <mjambor@suse.cz>
3984 * cgraphclones.c (cgraph_function_versioning): Copy
3985 ipa_transforms_to_apply instead of asserting it is empty.
3987 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
3990 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
3991 on count_exp to get mode.
3993 2014-04-22 Andrew Pinski <apinski@cavium.com>
3995 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
3996 Handle TLS for ILP32.
3997 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
3998 (tlsie_small_<mode>): this and handle PTR.
3999 (tlsie_small_sidi): New pattern.
4000 (tlsle_small): Change to an expand to handle ILP32.
4001 (tlsle_small_<mode>): New pattern.
4002 (tlsdesc_small): Rename to ...
4003 (tlsdesc_small_<mode>): this and handle PTR.
4005 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4007 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
4009 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4011 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
4012 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
4013 (aarch64_types_signed_poly_qualifiers): Likewise.
4014 (aarch64_types_unsigned_signed_qualifiers): Likewise.
4015 (aarch64_types_poly_signed_qualifiers): Likewise.
4016 (TYPES_REINTERP_SS): Type macro added.
4017 (TYPES_REINTERP_SU): Likewise.
4018 (TYPES_REINTERP_SP): Likewise.
4019 (TYPES_REINTERP_US): Likewise.
4020 (TYPES_REINTERP_PS): Likewise.
4021 (aarch64_fold_builtin): New expression folding added.
4022 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
4023 Declarations removed.
4024 (REINTERP_SS): Declarations added.
4025 (REINTERP_US): Likewise.
4026 (REINTERP_PS): Likewise.
4027 (REINTERP_SU): Likewise.
4028 (REINTERP_SP): Likewise.
4029 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
4030 (vreinterpretq_p8_f64): Likewise.
4031 (vreinterpret_p16_f64): Likewise.
4032 (vreinterpretq_p16_f64): Likewise.
4033 (vreinterpret_f32_f64): Likewise.
4034 (vreinterpretq_f32_f64): Likewise.
4035 (vreinterpret_f64_f32): Likewise.
4036 (vreinterpret_f64_p8): Likewise.
4037 (vreinterpret_f64_p16): Likewise.
4038 (vreinterpret_f64_s8): Likewise.
4039 (vreinterpret_f64_s16): Likewise.
4040 (vreinterpret_f64_s32): Likewise.
4041 (vreinterpret_f64_s64): Likewise.
4042 (vreinterpret_f64_u8): Likewise.
4043 (vreinterpret_f64_u16): Likewise.
4044 (vreinterpret_f64_u32): Likewise.
4045 (vreinterpret_f64_u64): Likewise.
4046 (vreinterpretq_f64_f32): Likewise.
4047 (vreinterpretq_f64_p8): Likewise.
4048 (vreinterpretq_f64_p16): Likewise.
4049 (vreinterpretq_f64_s8): Likewise.
4050 (vreinterpretq_f64_s16): Likewise.
4051 (vreinterpretq_f64_s32): Likewise.
4052 (vreinterpretq_f64_s64): Likewise.
4053 (vreinterpretq_f64_u8): Likewise.
4054 (vreinterpretq_f64_u16): Likewise.
4055 (vreinterpretq_f64_u32): Likewise.
4056 (vreinterpretq_f64_u64): Likewise.
4057 (vreinterpret_s64_f64): Likewise.
4058 (vreinterpretq_s64_f64): Likewise.
4059 (vreinterpret_u64_f64): Likewise.
4060 (vreinterpretq_u64_f64): Likewise.
4061 (vreinterpret_s8_f64): Likewise.
4062 (vreinterpretq_s8_f64): Likewise.
4063 (vreinterpret_s16_f64): Likewise.
4064 (vreinterpretq_s16_f64): Likewise.
4065 (vreinterpret_s32_f64): Likewise.
4066 (vreinterpretq_s32_f64): Likewise.
4067 (vreinterpret_u8_f64): Likewise.
4068 (vreinterpretq_u8_f64): Likewise.
4069 (vreinterpret_u16_f64): Likewise.
4070 (vreinterpretq_u16_f64): Likewise.
4071 (vreinterpret_u32_f64): Likewise.
4072 (vreinterpretq_u32_f64): Likewise.
4074 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4076 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
4077 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
4078 (vreinterpret_p8_s8): Likewise.
4079 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
4080 (vreinterpret_p8_s16): Likewise.
4081 (vreinterpret_p8_s32): Likewise.
4082 (vreinterpret_p8_s64): Likewise.
4083 (vreinterpret_p8_f32): Likewise.
4084 (vreinterpret_p8_u8): Likewise.
4085 (vreinterpret_p8_u16): Likewise.
4086 (vreinterpret_p8_u32): Likewise.
4087 (vreinterpret_p8_u64): Likewise.
4088 (vreinterpret_p8_p16): Likewise.
4089 (vreinterpretq_p8_s8): Likewise.
4090 (vreinterpretq_p8_s16): Likewise.
4091 (vreinterpretq_p8_s32): Likewise.
4092 (vreinterpretq_p8_s64): Likewise.
4093 (vreinterpretq_p8_f32): Likewise.
4094 (vreinterpretq_p8_u8): Likewise.
4095 (vreinterpretq_p8_u16): Likewise.
4096 (vreinterpretq_p8_u32): Likewise.
4097 (vreinterpretq_p8_u64): Likewise.
4098 (vreinterpretq_p8_p16): Likewise.
4099 (vreinterpret_p16_s8): Likewise.
4100 (vreinterpret_p16_s16): Likewise.
4101 (vreinterpret_p16_s32): Likewise.
4102 (vreinterpret_p16_s64): Likewise.
4103 (vreinterpret_p16_f32): Likewise.
4104 (vreinterpret_p16_u8): Likewise.
4105 (vreinterpret_p16_u16): Likewise.
4106 (vreinterpret_p16_u32): Likewise.
4107 (vreinterpret_p16_u64): Likewise.
4108 (vreinterpret_p16_p8): Likewise.
4109 (vreinterpretq_p16_s8): Likewise.
4110 (vreinterpretq_p16_s16): Likewise.
4111 (vreinterpretq_p16_s32): Likewise.
4112 (vreinterpretq_p16_s64): Likewise.
4113 (vreinterpretq_p16_f32): Likewise.
4114 (vreinterpretq_p16_u8): Likewise.
4115 (vreinterpretq_p16_u16): Likewise.
4116 (vreinterpretq_p16_u32): Likewise.
4117 (vreinterpretq_p16_u64): Likewise.
4118 (vreinterpretq_p16_p8): Likewise.
4119 (vreinterpret_f32_s8): Likewise.
4120 (vreinterpret_f32_s16): Likewise.
4121 (vreinterpret_f32_s32): Likewise.
4122 (vreinterpret_f32_s64): Likewise.
4123 (vreinterpret_f32_u8): Likewise.
4124 (vreinterpret_f32_u16): Likewise.
4125 (vreinterpret_f32_u32): Likewise.
4126 (vreinterpret_f32_u64): Likewise.
4127 (vreinterpret_f32_p8): Likewise.
4128 (vreinterpret_f32_p16): Likewise.
4129 (vreinterpretq_f32_s8): Likewise.
4130 (vreinterpretq_f32_s16): Likewise.
4131 (vreinterpretq_f32_s32): Likewise.
4132 (vreinterpretq_f32_s64): Likewise.
4133 (vreinterpretq_f32_u8): Likewise.
4134 (vreinterpretq_f32_u16): Likewise.
4135 (vreinterpretq_f32_u32): Likewise.
4136 (vreinterpretq_f32_u64): Likewise.
4137 (vreinterpretq_f32_p8): Likewise.
4138 (vreinterpretq_f32_p16): Likewise.
4139 (vreinterpret_s64_s8): Likewise.
4140 (vreinterpret_s64_s16): Likewise.
4141 (vreinterpret_s64_s32): Likewise.
4142 (vreinterpret_s64_f32): Likewise.
4143 (vreinterpret_s64_u8): Likewise.
4144 (vreinterpret_s64_u16): Likewise.
4145 (vreinterpret_s64_u32): Likewise.
4146 (vreinterpret_s64_u64): Likewise.
4147 (vreinterpret_s64_p8): Likewise.
4148 (vreinterpret_s64_p16): Likewise.
4149 (vreinterpretq_s64_s8): Likewise.
4150 (vreinterpretq_s64_s16): Likewise.
4151 (vreinterpretq_s64_s32): Likewise.
4152 (vreinterpretq_s64_f32): Likewise.
4153 (vreinterpretq_s64_u8): Likewise.
4154 (vreinterpretq_s64_u16): Likewise.
4155 (vreinterpretq_s64_u32): Likewise.
4156 (vreinterpretq_s64_u64): Likewise.
4157 (vreinterpretq_s64_p8): Likewise.
4158 (vreinterpretq_s64_p16): Likewise.
4159 (vreinterpret_u64_s8): Likewise.
4160 (vreinterpret_u64_s16): Likewise.
4161 (vreinterpret_u64_s32): Likewise.
4162 (vreinterpret_u64_s64): Likewise.
4163 (vreinterpret_u64_f32): Likewise.
4164 (vreinterpret_u64_u8): Likewise.
4165 (vreinterpret_u64_u16): Likewise.
4166 (vreinterpret_u64_u32): Likewise.
4167 (vreinterpret_u64_p8): Likewise.
4168 (vreinterpret_u64_p16): Likewise.
4169 (vreinterpretq_u64_s8): Likewise.
4170 (vreinterpretq_u64_s16): Likewise.
4171 (vreinterpretq_u64_s32): Likewise.
4172 (vreinterpretq_u64_s64): Likewise.
4173 (vreinterpretq_u64_f32): Likewise.
4174 (vreinterpretq_u64_u8): Likewise.
4175 (vreinterpretq_u64_u16): Likewise.
4176 (vreinterpretq_u64_u32): Likewise.
4177 (vreinterpretq_u64_p8): Likewise.
4178 (vreinterpretq_u64_p16): Likewise.
4179 (vreinterpret_s8_s16): Likewise.
4180 (vreinterpret_s8_s32): Likewise.
4181 (vreinterpret_s8_s64): Likewise.
4182 (vreinterpret_s8_f32): Likewise.
4183 (vreinterpret_s8_u8): Likewise.
4184 (vreinterpret_s8_u16): Likewise.
4185 (vreinterpret_s8_u32): Likewise.
4186 (vreinterpret_s8_u64): Likewise.
4187 (vreinterpret_s8_p8): Likewise.
4188 (vreinterpret_s8_p16): Likewise.
4189 (vreinterpretq_s8_s16): Likewise.
4190 (vreinterpretq_s8_s32): Likewise.
4191 (vreinterpretq_s8_s64): Likewise.
4192 (vreinterpretq_s8_f32): Likewise.
4193 (vreinterpretq_s8_u8): Likewise.
4194 (vreinterpretq_s8_u16): Likewise.
4195 (vreinterpretq_s8_u32): Likewise.
4196 (vreinterpretq_s8_u64): Likewise.
4197 (vreinterpretq_s8_p8): Likewise.
4198 (vreinterpretq_s8_p16): Likewise.
4199 (vreinterpret_s16_s8): Likewise.
4200 (vreinterpret_s16_s32): Likewise.
4201 (vreinterpret_s16_s64): Likewise.
4202 (vreinterpret_s16_f32): Likewise.
4203 (vreinterpret_s16_u8): Likewise.
4204 (vreinterpret_s16_u16): Likewise.
4205 (vreinterpret_s16_u32): Likewise.
4206 (vreinterpret_s16_u64): Likewise.
4207 (vreinterpret_s16_p8): Likewise.
4208 (vreinterpret_s16_p16): Likewise.
4209 (vreinterpretq_s16_s8): Likewise.
4210 (vreinterpretq_s16_s32): Likewise.
4211 (vreinterpretq_s16_s64): Likewise.
4212 (vreinterpretq_s16_f32): Likewise.
4213 (vreinterpretq_s16_u8): Likewise.
4214 (vreinterpretq_s16_u16): Likewise.
4215 (vreinterpretq_s16_u32): Likewise.
4216 (vreinterpretq_s16_u64): Likewise.
4217 (vreinterpretq_s16_p8): Likewise.
4218 (vreinterpretq_s16_p16): Likewise.
4219 (vreinterpret_s32_s8): Likewise.
4220 (vreinterpret_s32_s16): Likewise.
4221 (vreinterpret_s32_s64): Likewise.
4222 (vreinterpret_s32_f32): Likewise.
4223 (vreinterpret_s32_u8): Likewise.
4224 (vreinterpret_s32_u16): Likewise.
4225 (vreinterpret_s32_u32): Likewise.
4226 (vreinterpret_s32_u64): Likewise.
4227 (vreinterpret_s32_p8): Likewise.
4228 (vreinterpret_s32_p16): Likewise.
4229 (vreinterpretq_s32_s8): Likewise.
4230 (vreinterpretq_s32_s16): Likewise.
4231 (vreinterpretq_s32_s64): Likewise.
4232 (vreinterpretq_s32_f32): Likewise.
4233 (vreinterpretq_s32_u8): Likewise.
4234 (vreinterpretq_s32_u16): Likewise.
4235 (vreinterpretq_s32_u32): Likewise.
4236 (vreinterpretq_s32_u64): Likewise.
4237 (vreinterpretq_s32_p8): Likewise.
4238 (vreinterpretq_s32_p16): Likewise.
4239 (vreinterpret_u8_s8): Likewise.
4240 (vreinterpret_u8_s16): Likewise.
4241 (vreinterpret_u8_s32): Likewise.
4242 (vreinterpret_u8_s64): Likewise.
4243 (vreinterpret_u8_f32): Likewise.
4244 (vreinterpret_u8_u16): Likewise.
4245 (vreinterpret_u8_u32): Likewise.
4246 (vreinterpret_u8_u64): Likewise.
4247 (vreinterpret_u8_p8): Likewise.
4248 (vreinterpret_u8_p16): Likewise.
4249 (vreinterpretq_u8_s8): Likewise.
4250 (vreinterpretq_u8_s16): Likewise.
4251 (vreinterpretq_u8_s32): Likewise.
4252 (vreinterpretq_u8_s64): Likewise.
4253 (vreinterpretq_u8_f32): Likewise.
4254 (vreinterpretq_u8_u16): Likewise.
4255 (vreinterpretq_u8_u32): Likewise.
4256 (vreinterpretq_u8_u64): Likewise.
4257 (vreinterpretq_u8_p8): Likewise.
4258 (vreinterpretq_u8_p16): Likewise.
4259 (vreinterpret_u16_s8): Likewise.
4260 (vreinterpret_u16_s16): Likewise.
4261 (vreinterpret_u16_s32): Likewise.
4262 (vreinterpret_u16_s64): Likewise.
4263 (vreinterpret_u16_f32): Likewise.
4264 (vreinterpret_u16_u8): Likewise.
4265 (vreinterpret_u16_u32): Likewise.
4266 (vreinterpret_u16_u64): Likewise.
4267 (vreinterpret_u16_p8): Likewise.
4268 (vreinterpret_u16_p16): Likewise.
4269 (vreinterpretq_u16_s8): Likewise.
4270 (vreinterpretq_u16_s16): Likewise.
4271 (vreinterpretq_u16_s32): Likewise.
4272 (vreinterpretq_u16_s64): Likewise.
4273 (vreinterpretq_u16_f32): Likewise.
4274 (vreinterpretq_u16_u8): Likewise.
4275 (vreinterpretq_u16_u32): Likewise.
4276 (vreinterpretq_u16_u64): Likewise.
4277 (vreinterpretq_u16_p8): Likewise.
4278 (vreinterpretq_u16_p16): Likewise.
4279 (vreinterpret_u32_s8): Likewise.
4280 (vreinterpret_u32_s16): Likewise.
4281 (vreinterpret_u32_s32): Likewise.
4282 (vreinterpret_u32_s64): Likewise.
4283 (vreinterpret_u32_f32): Likewise.
4284 (vreinterpret_u32_u8): Likewise.
4285 (vreinterpret_u32_u16): Likewise.
4286 (vreinterpret_u32_u64): Likewise.
4287 (vreinterpret_u32_p8): Likewise.
4288 (vreinterpret_u32_p16): Likewise.
4289 (vreinterpretq_u32_s8): Likewise.
4290 (vreinterpretq_u32_s16): Likewise.
4291 (vreinterpretq_u32_s32): Likewise.
4292 (vreinterpretq_u32_s64): Likewise.
4293 (vreinterpretq_u32_f32): Likewise.
4294 (vreinterpretq_u32_u8): Likewise.
4295 (vreinterpretq_u32_u16): Likewise.
4296 (vreinterpretq_u32_u64): Likewise.
4297 (vreinterpretq_u32_p8): Likewise.
4298 (vreinterpretq_u32_p16): Likewise.
4300 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4302 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
4304 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
4306 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
4307 (vqnegd_s64): Likewise.
4308 (vqabs_s64): Likewise.
4309 (vqabsd_s64): Likewise.
4311 2014-04-22 Richard Henderson <rth@redhat.com>
4313 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
4314 computation to the top of the loop.
4316 2014-04-22 Renlin <renlin.li@arm.com>
4317 Jiong Wang <jiong.wang@arm.com>
4319 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
4320 * config/aarch64/aarch64.c (aarch64_layout_frame)
4321 (aarch64_initial_elimination_offset): Likewise.
4323 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
4325 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
4328 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
4330 * machmode.h (bitwise_mode_for_mode): Declare.
4331 * stor-layout.h (bitwise_type_for_mode): Likewise.
4332 * stor-layout.c (bitwise_mode_for_mode): New function.
4333 (bitwise_type_for_mode): Likewise.
4334 * builtins.c (fold_builtin_memory_op): Use it instead of
4335 int_mode_for_mode and build_nonstandard_integer_type.
4337 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4339 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
4340 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
4341 (*-*-solaris2*): Simplify.
4342 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
4343 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
4344 *-*-solaris2.9* handling.
4346 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
4348 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
4349 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
4351 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
4352 * configure: Regenerate.
4354 * config/i386/sol2-9.h: Remove.
4356 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
4357 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
4358 Remove Solaris 9 references.
4360 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
4362 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
4363 (floatuns<GPI:mode><GPF:mode>2): Remove.
4364 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
4365 and floatuns conversions.
4366 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
4367 and floatuns conversions.
4368 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
4369 (w1,w2): New mode attributes for inequal width conversions.
4371 2014-04-22 Renlin Li <Renlin.Li@arm.com>
4373 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
4374 the output asm format.
4376 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
4378 * config/aarch64/aarch64-simd.md
4379 (aarch64_cm<optab>di): Always split.
4380 (*aarch64_cm<optab>di): New.
4381 (aarch64_cmtstdi): Always split.
4382 (*aarch64_cmtstdi): New.
4384 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4386 PR tree-optimization/60823
4387 * omp-low.c (ipa_simd_modify_function_body): Go through
4388 all SSA_NAMEs and for those refering to vector arguments
4389 which are going to be replaced adjust SSA_NAME_VAR and,
4390 if it is a default definition, change it into a non-default
4391 definition assigned at the beginning of function from new_decl.
4392 (ipa_simd_modify_stmt_ops): Rewritten.
4393 * tree-dfa.c (set_ssa_default_def): When removing default def,
4394 check for NULL loc instead of NULL *loc.
4396 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4398 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
4399 restrictions on core registers for DImode values in Thumb2.
4401 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4403 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
4404 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
4406 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4408 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
4409 (*iordi_notzesidi_di): Likewise.
4410 (*iordi_notsesidi_di): Likewise.
4412 2014-04-22 Ian Bolton <ian.bolton@arm.com>
4414 * config/arm/arm-protos.h (tune_params): New struct members.
4415 * config/arm/arm.c: Initialise tune_params per processor.
4416 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
4417 for speed, based on new tune_params.
4419 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
4421 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
4422 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
4423 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
4424 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
4425 * config/aarch64/arm_neon.h (vrnd_f64): Added.
4426 (vrnda_f64): Likewise.
4427 (vrndi_f64): Likewise.
4428 (vrndm_f64): Likewise.
4429 (vrndn_f64): Likewise.
4430 (vrndp_f64): Likewise.
4431 (vrndx_f64): Likewise.
4433 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4435 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
4436 GET_MODE_SIZE argument is enum machine_mode.
4438 2014-04-22 Jakub Jelinek <jakub@redhat.com>
4441 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
4442 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
4444 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
4447 * asan.c (asan_emit_stack_protection): Force the base to align to
4448 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
4449 appropriate bits if STRICT_ALIGNMENT.
4450 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
4452 (expand_used_vars): Leave a space in the stack frame for alignment
4453 if STRICT_ALIGNMENT.
4455 2014-04-21 David Malcolm <dmalcolm@redhat.com>
4457 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
4459 (gimple_store_p): Likewise.
4460 (gimple_assign_load_p): Likewise.
4461 (gimple_assign_cast_p): Likewise.
4462 (gimple_clobber_p): Likewise.
4464 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
4465 rather than a gimple.
4466 (gimple_assign_cast_p): Likewise.
4468 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
4471 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
4472 If mode is DDmode and TARGET_E500_DOUBLE allow move.
4474 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
4475 more debug information for E500 if -mdebug=reg.
4477 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
4480 * config/i386/i386.c (ix86_expand_builtin)
4481 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
4482 register for target RTX.
4483 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
4485 2014-04-18 Cong Hou <congh@google.com>
4487 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
4488 the widen-mult pattern by handling two operands with different sizes,
4489 and operands whose size is smaller than half of the result type.
4491 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
4493 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
4494 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
4495 (do_estimate_edge_time): Compute it.
4496 * ipa-inline.c (want_inline_small_function_p): Bypass
4497 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
4499 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
4501 * ipa-inline.c (spec_rem): New static variable.
4502 (dump_overall_stats): New function.
4503 (dump_inline_stats): New function.
4505 2014-04-18 Richard Henderson <rth@redhat.com>
4507 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
4508 to GET_MODE_SIZE, not a reg_class_t.
4510 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4512 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
4513 (vsx_xxmrglw_<mode>): Likewise.
4515 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
4518 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
4519 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
4520 (rs6000_init_hard_regno_mode_ok): Likewise.
4522 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
4524 * ipa-inline.c (inline_small_functions): Account only non-cold
4526 * doc/invoke.texi (inline-unit-growth): Update documentation.
4528 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
4530 * config/rs6000/rs6000.md (addti3, subti3): New.
4532 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
4535 * config/i386/i386.c (ix86_expand_clear): Remove outdated
4536 comment. Check optimize_insn_for_size_p instead of
4537 optimize_insn_for_speed_p.
4539 2014-04-17 Martin Jambor <mjambor@suse.cz>
4541 * gimple-iterator.c (gsi_start_edge): New function.
4542 * gimple-iterator.h (gsi_start_edge): Declare.
4543 * tree-sra.c (single_non_eh_succ): New function.
4544 (disqualify_ops_if_throwing_stmt): Renamed to
4545 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
4546 having one non-EH successor BB.
4547 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
4548 generate loads into replacements.
4549 (sra_modify_assign): Likewise and and also use the simple path for
4551 (sra_modify_function_body): Commit statements on edges.
4553 2014-04-17 Richard Biener <rguenther@suse.de>
4556 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
4557 comparison results and add clarifying comment.
4559 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4561 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
4562 (blank_mode): Initialize it.
4563 (emit_mode_size_inline, emit_mode_nunits_inline,
4564 emit_mode_inner_inline): New functions.
4565 (emit_insn_modes_h): Call them and surround their output with
4566 #if GCC_VERSION >= 4001 ... #endif.
4567 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
4568 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
4569 mode_* arrays if the argument is __builtin_constant_p.
4570 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
4571 is enum machine_mode.
4573 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4575 * passes.c (opt_pass::execute): Adjust.
4576 (pass_manager::execute_pass_mode_switching): Likewise.
4577 (early_local_passes::execute): Likewise.
4578 (execute_one_pass): Pass cfun to the pass's execute method.
4579 * tree-pass.h (opt_pass::execute): Add function * argument.
4580 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
4581 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
4582 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
4583 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4584 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
4585 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
4586 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
4587 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
4588 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
4589 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
4590 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
4591 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
4592 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
4593 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
4594 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
4595 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
4596 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
4597 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
4598 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
4599 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4600 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4601 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4602 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4603 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4604 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4605 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4606 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
4607 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
4608 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
4611 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4613 * passes.c (opt_pass::gate): Take function * argument.
4614 (gate_all_early_local_passes): Merge into
4615 (early_local_passes::gate): this.
4616 (gate_all_early_optimizations): Merge into
4617 (all_early_optimizations::gate): this.
4618 (gate_all_optimizations): Mege into
4619 (all_optimizations::gate): this.
4620 (gate_all_optimizations_g): Merge into
4621 (all_optimizations_g::gate): this.
4622 (gate_rest_of_compilation): Mege into
4623 (rest_of_compilation::gate): this.
4624 (gate_postreload): Merge into
4625 (postreload::gate): this.
4626 (dump_one_pass): Pass cfun to the pass's gate method.
4627 (execute_ipa_summary_passes): Likewise.
4628 (execute_one_pass): Likewise.
4629 (ipa_write_summaries_2): Likewise.
4630 (ipa_write_optimization_summaries_1): Likewise.
4631 (ipa_read_summaries_1): Likewise.
4632 (ipa_read_optimization_summaries_1): Likewise.
4633 (execute_ipa_stmt_fixups): Likewise.
4634 * tree-pass.h (opt_pass::gate): Add function * argument.
4635 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
4636 combine-stack-adj.c, combine.c, compare-elim.c,
4637 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4638 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
4639 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
4640 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
4641 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
4642 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
4643 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4644 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
4645 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
4646 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
4647 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
4648 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
4649 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
4650 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4651 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4652 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4653 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4654 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4655 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4656 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4657 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4658 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4659 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
4660 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
4661 var-tracking.c, vtable-verify.c, web.c: Adjust.
4663 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4665 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
4666 * configure: Regenerate.
4668 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4670 * passes.c (dump_one_pass): don't check pass->has_gate.
4671 (execute_ipa_summary_passes): Likewise.
4672 (execute_one_pass): Likewise.
4673 (ipa_write_summaries_2): Likewise.
4674 (ipa_write_optimization_summaries_1): Likewise.
4675 (ipa_read_optimization_summaries_1): Likewise.
4676 (execute_ipa_stmt_fixups): Likewise.
4677 * tree-pass.h (pass_data::has_gate): Remove.
4678 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
4679 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
4680 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
4681 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
4682 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
4683 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
4684 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
4685 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
4686 gimple-low.c, gimple-ssa-isolate-paths.c,
4687 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
4688 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
4689 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
4690 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
4691 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
4692 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
4693 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
4694 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
4695 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
4696 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
4697 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
4698 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4699 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4700 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
4701 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
4702 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
4703 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4704 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4705 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
4706 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
4707 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
4708 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
4711 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4713 * pass_manager.h (pass_manager::register_dump_files_1): Remove
4715 * passes.c (pass_manager::register_dump_files_1): Merge into
4716 (pass_manager::register_dump_files): this, and remove its handling of
4717 properties since the pass always has the properties anyway.
4718 (pass_manager::pass_manager): Adjust.
4720 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
4722 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
4723 * passes.c (pass_manager::register_dump_files_1): Remove dead code
4724 dealing with properties.
4725 (pass_manager::register_dump_files): Adjust.
4727 2014-03-20 Mark Wielaard <mjw@redhat.com>
4729 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
4730 then represent the bound as normal constant value.
4732 2014-04-17 Jakub Jelinek <jakub@redhat.com>
4735 Forward port from 4.8 branch
4736 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
4738 * config/i386/bmiintrin.h (_blsi_u32): New.
4742 (_blsmsk_u32): Ditto.
4743 (_blsmsk_u64): Ditto.
4744 (_tzcnt_u32): Ditto.
4745 (_tzcnt_u64): Ditto.
4747 2014-04-17 Kito Cheng <kito@0xlab.org>
4749 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
4751 2014-04-17 Richard Biener <rguenther@suse.de>
4754 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
4755 boolean results for comparisons.
4757 2014-04-17 Richard Biener <rguenther@suse.de>
4759 PR tree-optimization/60836
4760 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
4761 initial PHI args to be gimple values.
4763 2014-04-17 Richard Biener <rguenther@suse.de>
4765 PR tree-optimization/60841
4766 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
4767 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
4768 of stmts to SLP build.
4769 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
4770 (vect_analyze_slp): Likewise.
4771 (vect_analyze_slp_instance): Likewise.
4772 (vect_build_slp_tree): Limit overall SLP tree growth.
4773 * tree-vectorizer.h (vect_analyze_data_refs,
4774 vect_analyze_slp): Adjust prototypes.
4776 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4778 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
4781 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4783 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
4784 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
4785 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
4786 for TARGET_SLOW_PSHUFB
4788 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
4790 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
4791 * config/i386/i386.c (intel_cost): Ditto.
4793 2014-04-17 Joey Ye <joey.ye@arm.com>
4795 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
4797 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4799 * opts.c (common_handle_option): Disable -fipa-reference coorectly
4802 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4804 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
4805 (type_all_derivations_known_p): New predicate.
4806 (type_all_ctors_visible_p): New predicate.
4807 (type_possibly_instantiated_p): New predicate.
4808 (get_odr_type): Compute all_derivations_known.
4809 (dump_odr_type): Dump the flag.
4810 (maybe_record_type): Cleanup.
4811 (record_target_from_binfo): Add bases_to_consider array;
4812 record bases for types w/o instances and skip CXX destructor.
4813 (possible_polymorphic_call_targets_1): Add bases_to_consider
4814 and consider_construction parameters; check if type may have instance.
4815 (get_polymorphic_call_info): Set maybe_in_construction to true
4816 when we know nothing.
4817 (record_targets_from_bases): Skip CXX destructors; they are
4818 never called for types in construction.
4819 (possible_polymorphic_call_targets): Do not record target when
4820 type may not have instance.
4822 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
4825 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
4826 external aliases alive, too.
4828 2014-04-16 Andrew Pinski <apinski@cavium.com>
4830 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
4833 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4835 * final.c (compute_alignments): Do not apply loop alignment to a block
4836 falling through to the exit.
4838 2014-04-16 Catherine Moore <clm@codesourcery.com>
4840 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
4841 Adjust constraints for microMIPS store patterns.
4843 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
4845 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
4847 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
4849 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
4850 (append_use): Run at -O0.
4851 (append_vdef): Likewise.
4852 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
4853 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
4855 2014-04-16 Jakub Jelinek <jakub@redhat.com>
4857 PR tree-optimization/60844
4858 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
4859 (propagate_op_to_single_use, remove_visited_stmt_chain,
4860 linearize_expr, repropagate_negates, reassociate_bb): Use it
4861 instead of gsi_remove.
4863 2014-04-16 Martin Jambor <mjambor@suse.cz>
4865 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
4866 ipa_transforms_to_apply.
4867 (cgraph_function_versioning): Assert that old_node has empty
4868 ipa_transforms_to_apply.
4869 * trans-mem.c (ipa_tm_create_version): Likewise.
4870 * tree-inline.c (tree_function_versioning): Do not duplicate
4871 ipa_transforms_to_apply.
4873 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4876 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
4878 Pass necessary as flags on 64-bit Solaris/x86.
4879 Use lowercase relocs for x86_64-*-*.
4880 * configure: Regenerate.
4882 2014-04-15 Jan Hubicka <jh@suse.cz>
4884 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
4885 (maybe_record_node, likely_target_p): Use it.
4887 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4890 Revert following patch
4892 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4895 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
4896 software floating point or no floating point registers, do not
4897 allow any type in the FPRs. Eliminate a test for SPE SIMD types
4898 in GPRs that occurs after we tested for GPRs that would never be
4901 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
4902 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
4903 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
4904 specifically allow DDmode, since that does not use the SPE SIMD
4907 2014-03-21 Mark Wielaard <mjw@redhat.com>
4909 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
4910 as unsigned or int depending on type and value used.
4912 2014-04-15 Richard Biener <rguenther@suse.de>
4914 PR rtl-optimization/56965
4915 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
4916 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
4918 * alias.c (true_dependence_1): Do not call
4919 nonoverlapping_component_refs_p.
4920 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
4921 nonoverlapping_component_refs_p.
4922 (indirect_refs_may_alias_p): Likewise.
4924 2014-04-15 Teresa Johnson <tejohnson@google.com>
4926 * cfg.c (dump_bb_info): Fix flags check.
4927 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
4929 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4931 PR rtl-optimization/60663
4932 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
4935 2014-04-15 Richard Biener <rguenther@suse.de>
4937 * lto-streamer.h (LTO_major_version): Bump to 4.
4939 2014-04-15 Richard Biener <rguenther@suse.de>
4941 * common.opt (lto_partition_model): New enum.
4942 (flto-partition=): Merge separate options with a single with argument,
4943 add -flto-partition=one support.
4944 * flag-types.h (enum lto_partition_model): Declare.
4945 * opts.c (finish_options): Remove duplicate -flto-partition=
4947 * lto-wrapper.c (run_gcc): Adjust.
4949 2014-04-15 Richard Biener <rguenther@suse.de>
4951 * alias.c (ncr_compar): New function.
4952 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
4954 2014-04-15 Richard Biener <rguenther@suse.de>
4956 * alias.c (record_component_aliases): Do not walk BINFOs.
4958 2014-04-15 Richard Biener <rguenther@suse.de>
4960 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
4961 Add struct function argument and adjust.
4962 (find_func_aliases_for_call): Likewise.
4963 (find_func_aliases): Likewise.
4964 (find_func_clobbers): Likewise.
4965 (intra_create_variable_infos): Likewise.
4966 (compute_points_to_sets): Likewise.
4967 (ipa_pta_execute): Adjust. Do not push/pop cfun.
4969 2014-04-15 Richard Biener <rguenther@suse.de>
4971 * tree.c (iterative_hash_expr): Use enum tree_code_class
4972 to store TREE_CODE_CLASS.
4973 (tree_block): Likewise.
4974 (tree_set_block): Likewise.
4975 * tree.h (fold_build_pointer_plus_loc): Use
4976 convert_to_ptrofftype_loc.
4978 2014-04-15 Jakub Jelinek <jakub@redhat.com>
4981 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
4984 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
4986 * cfgloop.h (struct loop): Move force_vectorize down.
4987 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
4988 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
4989 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
4990 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
4991 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
4992 * tree-core.h (enum annot_expr_kind): Add new kind values.
4993 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
4994 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
4996 * tree.def (ANNOTATE_EXPR): Tweak comment.
4998 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5000 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
5003 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
5005 * tree.h (TYPE_IDENTIFIER): Declare.
5006 * tree.c (subrange_type_for_debug_p): Use it.
5007 * godump.c (go_format_type): Likewise.
5008 * dwarf2out.c (is_cxx_auto, modified_type_die,
5009 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
5010 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
5012 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5015 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
5017 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
5019 * config/i386/i386.c (examine_argument): Return bool. Return true if
5020 parameter should be passed in memory.
5021 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
5022 (construct_container): Update calls to examine_argument.
5023 (function_arg_advance_64): Ditto.
5024 (return_in_memory_32): Merge with ix86_return_in_memory.
5025 (return_in_memory_64): Ditto.
5026 (return_in_memory_ms_64): Ditto.
5028 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
5030 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
5031 * coverage.c (coverage_compute_profile_id): Handle externally visible
5034 2014-04-14 Martin Jambor <mjambor@suse.cz>
5036 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
5037 DECL_DISREGARD_INLINE_LIMITS functions.
5039 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
5042 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
5044 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
5047 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
5048 optimize_insn_for_speed_p instead of
5049 optimize_function_for_speed_p.
5051 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
5053 * doc/invoke.texi (free): Document AArch64.
5055 2014-04-14 Richard Biener <rguenther@suse.de>
5057 PR tree-optimization/60042
5058 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
5059 (insert_into_preds_of_block): Do not prevent PHI insertion
5060 for REFERENCE exprs here ...
5061 (eliminate_dom_walker::before_dom_children): ... but prevent
5062 their use here under similar conditions when applied to the
5063 IL after PRE optimizations.
5065 2014-04-14 Richard Biener <rguenther@suse.de>
5067 * passes.def: Move early points-to after early SRA.
5069 2014-04-14 Richard Biener <rguenther@suse.de>
5071 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
5072 check for which sign-changes we allow when forwarding
5073 a converted value into a switch.
5075 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
5077 * stor-layout.c (place_field): Finalize non-constant offset for the
5080 2014-04-14 Richard Biener <rguenther@suse.de>
5082 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
5084 (expand_switch_using_bit_tests_p): Likewise.
5085 (process_switch): Compute and pass on speed_p based on the
5087 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
5088 optimize_bb_for_speed_p.
5090 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
5092 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
5093 * function.h (struct function): Rename has_force_vect_loops into
5094 has_force_vectorize_loops.
5095 * lto-streamer-in.c (input_cfg): Adjust for renaming.
5096 (input_struct_function_base): Likewise.
5097 * lto-streamer-out.c (output_cfg): Likewise.
5098 (output_struct_function_base): Likewise.
5099 * omp-low.c (expand_omp_simd): Likewise.
5100 * tree-cfg.c (move_sese_region_to_fn): Likewise.
5101 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
5102 (version_loop_for_if_conversion): Likewise.
5103 (tree_if_conversion): Likewise.
5104 (main_tree_if_conversion): Likewise.
5105 (gate_tree_if_conversion): Likewise.
5106 * tree-inline.c (copy_loops): Likewise.
5107 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
5108 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
5109 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
5110 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
5111 * tree-vectorizer.c (vectorize_loops): Likewise.
5112 * tree-vectorizer.h (unlimited_cost_model): Likewise.
5114 2014-04-14 Richard Biener <rguenther@suse.de>
5117 * lto-streamer-out.c (wrap_refs): New function.
5118 (lto_output): Wrap symbol references in global initializes in
5119 type-preserving MEM_REFs.
5121 2014-04-14 Christian Bruel <christian.bruel@st.com>
5123 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
5125 2014-04-14 Christian Bruel <christian.bruel@st.com>
5127 * config/sh/sh.md (setmemqi): New expand pattern.
5128 * config/sh/sh.h (CLEAR_RATIO): Define.
5129 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
5130 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
5132 2014-04-14 Richard Biener <rguenther@suse.de>
5135 * fold-const.c (negate_expr_p): Don't negate directional rounding
5137 (fold_negate_expr): Likewise.
5139 2014-04-14 Richard Biener <rguenther@suse.de>
5141 PR tree-optimization/59817
5142 PR tree-optimization/60453
5143 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
5144 recursion to catch all CHRECs in the scalar evolution and restrict
5145 the predicate for the remains appropriately.
5147 2014-04-12 Catherine Moore <clm@codesourcery.com>
5149 * config/mips/constraints.md: Add new register constraint "kb".
5150 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
5151 (*movhi_internal): Likewise.
5152 (*movqi_internal): Likewise.
5153 * config/mips/mips.h (M16_STORE_REGS): New register class.
5154 (REG_CLASS_NAMES): Add M16_STORE_REGS.
5155 (REG_CLASS_CONTENTS): Likewise.
5156 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
5158 2014-04-11 Tobias Burnus <burnus@net-b.de>
5161 * doc/invoke.texi (-Wformat-signedness): Document it.
5162 (Wformat=2): Mention that this enables -Wformat-signedness.
5164 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5166 * common/config/epiphany/epiphany-common.c
5167 (epiphany_option_optimization_table): Enable section anchors by
5168 default at -O1 or higher.
5169 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
5170 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
5171 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
5172 carries no extra cost.
5173 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
5174 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
5175 * config/epiphany/predicates.md (memclob_operand): New predicate.
5176 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
5177 Use memclob_operand predicate and X constraint for operand 3.
5179 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5181 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
5182 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
5185 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
5187 PR rtl-optimization/60651
5188 * mode-switching.c (optimize_mode_switching): Make sure to emit
5189 sets of a lower numbered entity before sets of a higher numbered
5190 entity to a mode of the same or lower priority.
5191 When creating a seginfo for a basic block that starts with a code
5192 label, move the insertion point past the code label.
5193 (new_seginfo): Document and enforce requirement that
5194 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
5195 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
5196 * doc/tm.texi: Regenerate.
5198 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
5201 * config/arc/arc.c (arc_save_restore): Fix assert typo.
5203 2013-04-11 Jakub Jelinek <jakub@redhat.com>
5205 * BASE-VER: Set to 4.10.0.
5207 2014-04-11 Tobias Burnus <burnus@net-b.de>
5210 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
5211 * doc/gcc.texi (Service): Update description in the @menu
5212 * doc/invoke.texi (Option Summary): Remove misplaced and
5215 2014-04-11 Steve Ellcey <sellcey@mips.com>
5216 Jakub Jelinek <jakub@redhat.com>
5219 * expr.c (convert_move): Use emit_store_flag_force instead of
5220 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
5223 2014-04-11 Richard Biener <rguenther@suse.de>
5226 * varasm.c (assemble_alias): Avoid endless error reporting
5227 recursion by setting TREE_ASM_WRITTEN.
5229 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5231 * config/s390/s390.md: Add a splitter for NOT rtx.
5233 2014-04-11 Jakub Jelinek <jakub@redhat.com>
5235 PR rtl-optimization/60663
5236 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
5238 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
5239 Jakub Jelinek <jakub@redhat.com>
5242 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
5243 flag from decl_node to node.
5245 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5248 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
5249 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
5250 ameliorating the cases where it can be.
5252 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
5255 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
5257 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5258 (loadsync_<mode>): Change mode.
5259 (load_quadpti, store_quadpti): New.
5260 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5261 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5262 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
5264 2014-04-09 Cong Hou <congh@google.com>
5267 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
5270 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5272 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
5273 instead of vnor to exploit possible fusion opportunity in the
5275 (altivec_expand_vec_perm_const_le): Likewise.
5277 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
5279 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5280 (loadsync_<mode>): Change mode.
5281 (load_quadpti, store_quadpti): New.
5282 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5283 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5285 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
5288 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
5289 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
5290 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
5292 2014-04-08 Richard Biener <rguenther@suse.de>
5295 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
5296 a 64bit widest int print double-int similar to on HWI64 hosts.
5298 2014-04-08 Richard Biener <rguenther@suse.de>
5300 PR tree-optimization/60785
5301 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
5302 default defs properly.
5304 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
5306 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
5307 (Weffc++): Likewise.
5309 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
5311 * ipa-devirt.c (maybe_record_node): When node is not recorded,
5312 set completep to false rather than true.
5314 2014-04-07 Douglas B Rupp <rupp@adacore.com>
5317 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
5318 ARM_TARGET2_DWARF_FORMAT.
5320 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
5323 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
5324 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
5327 2014-04-07 Richard Biener <rguenther@suse.de>
5329 PR tree-optimization/60766
5330 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
5331 (may_eliminate_iv): Convert cand_value_at result to desired type.
5333 2014-04-07 Jason Merrill <jason@redhat.com>
5336 * common.opt (-fno-gnu-unique): Add.
5337 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
5339 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5341 * haifa-sched.c: Fix outdated function reference and minor
5342 grammar errors in introductory comment.
5344 2014-04-07 Richard Biener <rguenther@suse.de>
5347 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
5349 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
5351 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
5354 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
5355 size accounting for thunks.
5356 (pa_asm_output_mi_thunk): Use final_start_function() and
5357 final_end_function() to output function start and end directives.
5359 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
5361 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
5362 device specific ISA/ feature information. Remove short_sp and
5363 errata_skip ds. Add avr_device_specific_features enum to have device
5365 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
5366 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
5367 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
5368 updated device specific info.
5369 * config/avr/avr-mcus.def: Merge device specific details to
5370 dev_attribute field.
5371 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
5373 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
5374 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
5375 assembler if RMW isa supported by current device.
5376 * config/avr/genmultilib.awk: Update as device info structure changed.
5377 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
5379 2014-04-04 Cong Hou <congh@google.com>
5381 PR tree-optimization/60656
5382 * tree-vect-stmts.c (supportable_widening_operation):
5383 Fix a bug that elements in a vector with vect_used_by_reduction
5384 property are incorrectly reordered when the operation on it is not
5385 consistant with the one in reduction operation.
5387 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
5389 PR rtl-optimization/60155
5390 * gcse.c (record_set_data): New function.
5391 (single_set_gcse): New function.
5392 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
5393 (hoist_code): Likewise.
5394 (get_pressure_class_and_nregs): Likewise.
5396 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
5398 * explow.c (probe_stack_range): Emit a final optimization blockage.
5400 2014-04-04 Anthony Green <green@moxielogic.com>
5402 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
5405 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
5408 * lto-cgraph.c (input_overwrite_node): Check that partitioning
5409 flags are set only during streaming.
5410 * ipa.c (process_references, walk_polymorphic_call_targets,
5411 symtab_remove_unreachable_nodes): Drop bodies of always inline
5412 after early inlining.
5413 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
5415 2014-04-04 Jakub Jelinek <jakub@redhat.com>
5416 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5419 * dwarf2out.c (const_ok_for_output_1): Reject expressions
5422 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5425 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
5427 (cortex_a53_fdivd): Likewise.
5429 2014-04-04 Martin Jambor <mjambor@suse.cz>
5432 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
5434 * cgraph.c (clone_of_p): Also return true if thunks match.
5435 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
5436 cgraph_function_or_thunk_node and an obsolete comment.
5437 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
5439 (build_function_decl_skip_args): Likewise.
5440 (set_new_clone_decl_and_node_flags): New function.
5441 (duplicate_thunk_for_node): Likewise.
5442 (redirect_edge_duplicating_thunks): Likewise.
5443 (cgraph_clone_node): New parameter args_to_skip, pass it to
5444 redirect_edge_duplicating_thunks which is called instead of
5445 cgraph_redirect_edge_callee.
5446 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
5447 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
5449 2014-04-04 Jeff Law <law@redhat.com>
5452 * config/arm/predicates.md (const_int_I_operand): New predicate.
5453 (const_int_M_operand): Similarly.
5454 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
5456 (insv_t2, extv_reg, extzv_t2): Likewise.
5457 (load_multiple_with_writeback): Similarly for const_int_I_operand.
5458 (pop_multiple_with_writeback_and_return): Likewise.
5459 (vfp_pop_multiple_with_writeback): Likewise
5461 2014-04-04 Richard Biener <rguenther@suse.de>
5464 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
5465 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
5467 * gimplify.h (gimple_add_tmp_var_fn): Declare.
5468 * gimplify.c (gimple_add_tmp_var_fn): New function.
5469 * gimple-expr.h (create_tmp_reg_fn): Declare.
5470 * gimple-expr.c (create_tmp_reg_fn): New function.
5471 * gimple-low.c (record_vars_into): Don't change cfun.
5472 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
5473 code generation without cfun.
5475 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
5478 * Makefile.in (install-driver): Fix shell scripting.
5480 2014-04-03 Cong Hou <congh@google.com>
5482 PR tree-optimization/60505
5483 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
5484 threshold of number of iterations below which no vectorization
5486 * tree-vect-loop.c (new_loop_vec_info):
5487 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
5488 * tree-vect-loop.c (vect_analyze_loop_operations):
5489 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
5490 * tree-vect-loop.c (vect_transform_loop):
5491 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
5492 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
5493 of iterations of the loop and see if we should build the epilogue.
5495 2014-04-03 Richard Biener <rguenther@suse.de>
5497 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
5498 (streamer_tree_cache_create): Adjust.
5499 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
5500 to allow optional nodes array.
5501 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
5502 (streamer_tree_cache_append): Likewise.
5503 (streamer_tree_cache_create): Create nodes array optionally
5504 as specified by parameter.
5505 * lto-streamer-out.c (create_output_block): Avoid maintaining
5506 the node array in the writer cache.
5507 (DFS_write_tree): Remove assertion.
5508 (produce_asm_for_decls): Free the out decl state hash table early.
5509 * lto-streamer-in.c (lto_data_in_create): Adjust for
5510 streamer_tree_cache_create prototype change.
5512 2014-04-03 Richard Biener <rguenther@suse.de>
5514 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
5515 set TREE_CHAIN to NULL_TREE.
5517 2014-04-03 Richard Biener <rguenther@suse.de>
5519 PR tree-optimization/60740
5520 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
5521 over all GIMPLE_COND operands.
5523 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
5525 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
5526 (Weffc++): Remove Scott's numbering, merge lists and reference
5529 2014-04-03 Nick Clifton <nickc@redhat.com>
5531 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
5534 2014-04-03 Martin Jambor <mjambor@suse.cz>
5536 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
5537 mention gcc_unreachable before failing.
5538 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
5541 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
5544 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
5545 inconsistent code and instead mark the context inconsistent.
5546 (possible_polymorphic_call_targets): For inconsistent contexts
5547 return empty complete list.
5549 2014-04-02 Anthony Green <green@moxielogic.com>
5551 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
5552 (extendqisi2, extendhisi2): Define.
5553 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
5554 (WCHAR_TYPE): Change to unsigned int.
5556 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5558 PR tree-optimization/60733
5559 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
5560 insertion point for PHI candidates to be the end of the feeding
5561 block for the PHI argument.
5563 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
5565 PR rtl-optimization/60650
5566 * lra-constraints.c (process_alt_operands): Decrease reject for
5567 earlyclobber matching.
5569 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5571 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
5573 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5575 * config/spu/spu.c (pad_bb): Do not crash when the last
5576 insn is CODE_FOR_blockage.
5578 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5580 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
5581 lies outside the target mode.
5583 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5586 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
5587 software floating point or no floating point registers, do not
5588 allow any type in the FPRs. Eliminate a test for SPE SIMD types
5589 in GPRs that occurs after we tested for GPRs that would never be
5592 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
5593 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
5594 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
5595 specifically allow DDmode, since that does not use the SPE SIMD
5598 2014-04-02 Richard Biener <rguenther@suse.de>
5601 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
5602 MODE_INTs. Properly use negv_optab.
5603 (expand_abs): Likewise.
5605 2014-04-02 Richard Biener <rguenther@suse.de>
5608 * Makefile.in (install-driver): Guard extra installs with special
5611 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
5613 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5616 2014-04-01 Richard Henderson <rth@redhat.com>
5619 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
5620 alternative enabled before register allocation.
5622 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
5624 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
5625 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
5627 (nios2_large_got_address): Remove unneeded 'sym' parameter.
5628 (nios2_got_address): Update nios2_large_got_address call site.
5629 (nios2_delegitimize_address): New function.
5630 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
5631 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
5632 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
5634 2014-04-01 Martin Husemann <martin@duskware.de>
5636 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
5639 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
5641 PR rtl-optimization/60604
5642 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
5643 check from register_operand.
5644 (register_operand): Redefine in terms of general_operand.
5645 (nonmemory_operand): Use register_operand for the non-constant cases.
5647 2014-04-01 Richard Biener <rguenther@suse.de>
5649 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
5651 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
5653 * doc/invoke.texi (mapp-regs): Clarify.
5655 2014-03-31 Ulrich Drepper <drepper@gmail.com>
5657 * config/i386/avx512fintrin.h (__v32hi): Define type.
5658 (__v64qi): Likewise.
5659 (_mm512_set1_epi8): Define.
5660 (_mm512_set1_epi16): Define.
5661 (_mm512_set4_epi32): Define.
5662 (_mm512_set4_epi64): Define.
5663 (_mm512_set4_pd): Define.
5664 (_mm512_set4_ps): Define.
5665 (_mm512_setr4_epi64): Define.
5666 (_mm512_setr4_epi32): Define.
5667 (_mm512_setr4_pd): Define.
5668 (_mm512_setr4_ps): Define.
5669 (_mm512_setzero_epi32): Define.
5671 2014-03-31 Martin Jambor <mjambor@suse.cz>
5674 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
5675 callsite_arguments_match_p. Updated all callers. Also check types of
5676 corresponding formal parameters and actual arguments.
5677 (not_all_callers_have_enough_arguments_p) Renamed to
5678 some_callers_have_mismatched_arguments_p.
5680 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
5682 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
5684 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
5687 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
5690 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
5692 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
5693 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
5695 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
5696 Use FMAMODE_NOVF512 mode iterator.
5697 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
5698 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
5699 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
5700 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
5702 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
5703 Use VF_128_256 mode iterator.
5704 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
5707 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5709 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
5710 static chain if needed.
5712 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5715 * lra-constraints.c (index_part_to_reg): New.
5716 (process_address): Use it.
5718 2014-03-27 Jeff Law <law@redhat.com>
5719 Jakub Jelinek <jakub@redhat.com>
5722 * expr.c (do_tablejump): Use simplify_gen_binary rather than
5723 gen_rtx_{PLUS,MULT} to build up the address expression.
5725 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
5726 creating non-canonical RTL.
5728 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5731 * ipa-inline.c (want_inline_small_function_p): Short circuit large
5732 functions; reorganize to make cheap checks first.
5733 (inline_small_functions): Do not estimate growth when dumping;
5735 * ipa-inline.h (inline_summary): Add min_size.
5736 (growth_likely_positive): New function.
5737 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
5738 (set_cond_stmt_execution_predicate): Cleanup.
5739 (estimate_edge_size_and_time): Compute min_size.
5740 (estimate_calls_size_and_time): Likewise.
5741 (estimate_node_size_and_time): Likewise.
5742 (inline_update_overall_summary): Update min_size.
5743 (do_estimate_edge_time): Likewise.
5744 (do_estimate_edge_size): Update.
5745 (do_estimate_edge_hints): Update.
5746 (growth_likely_positive): New function.
5748 2014-03-28 Jakub Jelinek <jakub@redhat.com>
5751 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
5752 also if addr has VOIDmode.
5754 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5756 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
5757 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
5759 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
5760 instructions as well as AdvancedSIMD loads.
5762 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5764 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
5765 Use crypto_aese type.
5766 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
5767 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
5768 crypto_aese, crypto_aesmc. Move to types.md.
5769 * config/arm/types.md (crypto_aes): Split into crypto_aese,
5771 * config/arm/iterators.md (crypto_type): Likewise.
5773 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
5775 * cgraph.c: Include expr.h and tree-dfa.h.
5776 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
5779 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
5782 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
5783 regs from checking multi-reg pseudos.
5785 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5787 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
5789 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5791 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
5792 if it would clobber the stack pointer, even temporarily.
5794 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
5796 * mode-switching.c: Make small adjustments to the top comment.
5798 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
5800 * config/rs6000/constraints.md (wD constraint): New constraint to
5801 match the constant integer to get the top DImode/DFmode out of a
5802 vector in a VSX register.
5804 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
5805 match the constant integer to get the top DImode/DFmode out of a
5806 vector in a VSX register.
5808 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
5811 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
5814 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
5815 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
5817 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
5818 Optimize vec_extract of 64-bit values, where the value being
5819 extracted is in the top word, where we can use scalar
5820 instructions. Add direct move and store support. Combine the big
5821 endian/little endian vector select load support into a single insn.
5822 (vsx_extract_<mode>_internal1): Likewise.
5823 (vsx_extract_<mode>_internal2): Likewise.
5824 (vsx_extract_<mode>_load): Likewise.
5825 (vsx_extract_<mode>_store): Likewise.
5826 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
5827 combined into vsx_extract_<mode>_load.
5828 (vsx_extract_<mode>_one_le): Likewise.
5830 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
5831 define the top 64-bit vector element.
5833 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
5836 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5837 Document vec_vbpermq builtin.
5840 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
5841 enable use of xxsldwi and xxpermdi builtin functions.
5842 (vec_xxpermdi): Likewise.
5844 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
5845 Document use of vec_xxsldwi and vec_xxpermdi builtins.
5847 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
5849 PR rtl-optimization/60650
5850 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
5852 (find_spills_for): New.
5853 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
5854 Spill all pseudos on the second iteration.
5856 2014-03-27 Marek Polacek <polacek@redhat.com>
5859 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
5862 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5864 * config/s390/s390.c (s390_can_use_return_insn): Check for
5865 call-saved FPRs on 31 bit.
5867 2014-03-27 Jakub Jelinek <jakub@redhat.com>
5870 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
5871 if they need regimplification, just drop them instead of
5872 calling gimple_regimplify_operands on them.
5874 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
5877 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
5878 (aarch64_frame_pointer_required): Adjust logic.
5879 (aarch64_can_eliminate): Adjust logic.
5880 (aarch64_override_options_after_change): Adjust logic.
5882 2014-03-27 Dehao Chen <dehao@google.com>
5884 * ipa-inline.c (early_inliner): Update node's inline info.
5886 2014-03-26 Dehao Chen <dehao@google.com>
5888 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
5889 compiler inserted conditional jumps for NAN float check.
5891 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5893 * ubsan.h (ubsan_create_data): Change second argument's type
5894 to const location_t *.
5895 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
5897 (ubsan_create_data): Change second argument to const location_t *PLOC.
5898 Create Loc field whenever PLOC is non-NULL.
5899 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
5900 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
5904 * real.c (real_to_integer2): Change type of low to UHWI.
5906 2014-03-26 Tobias Burnus <burnus@net-b.de>
5908 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
5909 (CILK_SELF_SPECS): New define.
5910 (driver_self_specs): Use it.
5912 2014-03-26 Richard Biener <rguenther@suse.de>
5914 * tree-pretty-print.c (percent_K_format): Implement special
5915 case for LTO and its stripped down BLOCK tree.
5917 2014-03-26 Jakub Jelinek <jakub@redhat.com>
5920 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
5922 * tree-vrp.c (simplify_internal_call_using_ranges): If only
5923 one range is range_int_cst_p, but not both, at least optimize
5924 addition/subtraction of 0 and multiplication by 0 or 1.
5925 * gimple-fold.c (gimple_fold_call): Fold
5926 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
5927 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
5928 INTEGER_CSTs, try to fold at least x * 0 and y - y.
5930 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
5932 PR rtl-optimization/60452
5933 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
5934 <case REG>: Return 1 for invalid offsets from the frame pointer.
5936 2014-03-26 Marek Polacek <polacek@redhat.com>
5939 * doc/extend.texi (C Extensions): Mention variable-length arrays in
5942 2014-03-26 Marek Polacek <polacek@redhat.com>
5945 * doc/extend.texi (Designated Inits): Describe what happens to omitted
5948 2014-03-26 Marek Polacek <polacek@redhat.com>
5951 * ira-color.c (update_conflict_hard_regno_costs): Perform the
5952 multiplication in unsigned type.
5954 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5956 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
5958 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
5960 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
5962 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5965 * cif-code.def (UNREACHABLE) New code.
5966 * ipa-inline.c (inline_small_functions): Skip edges to
5967 __builtlin_unreachable.
5968 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
5969 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
5970 predicate to __bulitin_unreachable.
5971 (set_cond_stmt_execution_predicate): Fix issue when
5972 invert_tree_comparison returns ERROR_MARK.
5973 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
5974 propagate to inline clones.
5975 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
5977 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
5978 * cgraphclones.c (cgraph_clone_node): If call destination is already
5979 ureachable, do not redirect it back.
5980 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
5983 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
5985 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
5986 Do not modify inline clones.
5988 2014-03-25 Jakub Jelinek <jakub@redhat.com>
5990 * config/i386/i386.md (general_sext_operand): New mode attr.
5991 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
5992 don't generate (sign_extend (const_int)).
5993 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
5994 operands[2]. Use We constraint instead of <i> and
5995 <general_sext_operand> predicate instead of <general_operand>.
5996 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
5997 * config/i386/constraints.md (We): New constraint.
5998 * config/i386/predicates.md (x86_64_sext_operand,
5999 sext_operand): New predicates.
6001 2014-03-25 Martin Jambor <mjambor@suse.cz>
6004 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
6005 inconsistent devirtualizations to __builtin_unreachable.
6007 2014-03-25 Marek Polacek <polacek@redhat.com>
6010 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
6012 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
6014 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
6015 order of elements for big-endian.
6017 2014-03-25 Richard Biener <rguenther@suse.de>
6020 * gimplify-me.c (gimple_regimplify_operands): Update the
6023 2014-03-25 Martin Jambor <mjambor@suse.cz>
6026 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
6027 (lto_output_varpool_node): Likewise.
6028 (input_overwrite_node): Likewise.
6029 (input_varpool_node): Likewise.
6031 2014-03-25 Richard Biener <rguenther@suse.de>
6033 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
6034 (run_gcc): Likewise.
6036 2014-03-25 Jakub Jelinek <jakub@redhat.com>
6038 * combine.c (simplify_compare_const): Add MODE argument.
6039 Handle mode_width 0 as very large mode_width.
6040 (try_combine, simplify_comparison): Adjust callers.
6042 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
6043 type to avoid signed integer overflow.
6044 * explow.c (plus_constant): Likewise.
6046 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
6048 * doc/generic.texi: Correct typos.
6050 2014-03-24 Tobias Burnus <burnus@net-b.de>
6052 * doc/invoke.texi (-flto): Expand section about
6053 using static libraries with LTO.
6055 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6057 PR rtl-optimization/60501
6058 * optabs.def (addptr3_optab): New optab.
6059 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
6060 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
6061 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
6063 * lra.c (emit_add3_insn): Use the addptr pattern if available.
6065 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
6067 2014-03-24 Ulrich Drepper <drepper@gmail.com>
6069 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
6072 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
6073 (_mm256_undefined_ps): Define.
6074 (_mm256_undefined_pd): Define.
6075 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
6076 (_mm_undefined_pd): Define.
6077 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
6078 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
6079 (_mm512_undefined_ps): Define.
6080 (_mm512_undefined_pd): Define.
6081 Use _mm*_undefined_*.
6082 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
6084 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
6086 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
6087 (lshr_simd): DI mode added.
6088 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
6089 (aarch64_ushr_simddi): Likewise.
6090 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
6091 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
6092 (vshrd_n_u64): Likewise.
6094 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6096 * Makefile.in (s-macro_list): Depend on cc1.
6098 2014-03-23 Teresa Johnson <tejohnson@google.com>
6100 * ipa-utils.c (ipa_print_order): Use specified dump file.
6102 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
6104 PR rtl-optimization/60601
6105 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
6107 * gcc.c (eval_spec_function): Initialize save_growing_value.
6109 2014-03-22 Jakub Jelinek <jakub@redhat.com>
6112 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
6113 code == MINUS_EXPR, never swap op0 with op1.
6115 * toplev.c (init_local_tick): Avoid signed integer multiplication
6117 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
6118 shift by first operand's bitsize.
6120 2014-03-21 Jakub Jelinek <jakub@redhat.com>
6123 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
6125 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
6126 TARGET_ISA_64BIT_P(x).
6128 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6130 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
6131 pattern for vector nor instead of subtract from splat(-1).
6132 (altivec_expand_vec_perm_const_le): Likewise.
6134 2014-03-21 Richard Henderson <rth@twiddle.net>
6137 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
6138 related insns after epilogue_completed.
6140 2014-03-21 Martin Jambor <mjambor@suse.cz>
6143 * cgraph.h (symtab_node): New flag body_removed.
6144 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
6145 when removing bodies.
6146 * symtab.c (dump_symtab_base): Dump body_removed flag.
6147 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
6148 had their bodies removed.
6150 2014-03-21 Martin Jambor <mjambor@suse.cz>
6153 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
6156 2014-03-21 Richard Biener <rguenther@suse.de>
6158 PR tree-optimization/60577
6159 * tree-core.h (struct tree_base): Document nothrow_flag use
6160 in VAR_DECL_NONALIASED.
6161 * tree.h (VAR_DECL_NONALIASED): New.
6162 (may_be_aliased): Adjust.
6163 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
6165 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6167 * expr.c (expand_expr_real_1): Remove outdated comment.
6169 2014-03-20 Jakub Jelinek <jakub@redhat.com>
6172 * ira.c (adjust_cleared_regs): Call copy_rtx on
6173 *reg_equiv[REGNO (loc)].src_p before passing it to
6174 simplify_replace_fn_rtx.
6177 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
6178 into CONST, put pic register as first operand of PLUS. Use
6179 gen_const_mem for both 32-bit and 64-bit PIC got loads.
6181 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6183 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
6185 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6187 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
6188 around for store forwarding issue in the FPU on the UT699.
6189 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
6190 loads and operations if -mfix-ut699 is specified.
6191 (divtf3_hq): Tweak attribute.
6192 (sqrttf2_hq): Likewise.
6194 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
6196 * calls.c (store_one_arg): Remove incorrect const qualification on the
6197 type of the temporary.
6198 * cfgexpand.c (expand_return): Likewise.
6199 * expr.c (expand_constructor): Likewise.
6200 (expand_expr_real_1): Likewise.
6202 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6204 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
6207 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
6210 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
6212 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
6214 * config/arm/aarch-common-protos.h
6215 (alu_cost_table): Fix spelling of "extend".
6216 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
6218 2014-03-19 Richard Biener <rguenther@suse.de>
6221 * tree-core.h (tree_type_common): Re-order pointer members
6222 to reduce recursion depth during GC walks.
6224 2014-03-19 Marek Polacek <polacek@redhat.com>
6227 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
6228 before accessing it.
6230 2014-03-19 Richard Biener <rguenther@suse.de>
6233 * lto-streamer-in.c (input_function): In WPA stage do not drop
6236 2014-03-19 Jakub Jelinek <jakub@redhat.com>
6238 PR tree-optimization/60559
6239 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
6240 with build_zero_cst assignment.
6242 2014-03-18 Kai Tietz <ktietz@redhat.com>
6244 PR rtl-optimization/56356
6245 * sdbout.c (sdbout_parms): Verify that parms'
6246 incoming argument is valid.
6247 (sdbout_reg_parms): Likewise.
6249 2014-03-18 Richard Henderson <rth@redhat.com>
6252 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
6253 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
6254 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
6256 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
6258 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
6259 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
6260 Italicize plugin event names in description. Explain that
6261 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
6262 Remind that no GCC functions should be called after PLUGIN_FINISH.
6263 Explain what pragmas with expansion are.
6265 2014-03-18 Martin Liska <mliska@suse.cz>
6267 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
6268 gimple call statement is update.
6269 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
6270 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
6272 2014-03-18 Jakub Jelinek <jakub@redhat.com>
6275 * ubsan.c (ubsan_instrument_unreachable): Call
6276 initialize_sanitizer_builtins.
6277 (ubsan_pass): Likewise.
6280 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
6281 varpool_finalize_decl instead of rest_of_decl_compilation.
6283 2014-03-18 Richard Biener <rguenther@suse.de>
6285 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
6286 by using bitmap_and_compl instead of bitmap_and_compl_into.
6287 (df_rd_transfer_function): Likewise.
6289 2014-03-18 Richard Biener <rguenther@suse.de>
6291 * doc/lto.texi (fresolution): Fix typo.
6293 2014-03-18 Richard Biener <rguenther@suse.de>
6295 * doc/invoke.texi (flto): Update for changes in 4.9.
6297 2014-03-18 Richard Biener <rguenther@suse.de>
6299 * doc/loop.texi: Remove section on the removed lambda framework.
6300 Update loop docs with recent changes in preserving loop structure.
6302 2014-03-18 Richard Biener <rguenther@suse.de>
6304 * doc/lto.texi (-fresolution): Document.
6306 2014-03-18 Richard Biener <rguenther@suse.de>
6308 * doc/contrib.texi: Adjust my name.
6310 2014-03-18 Jakub Jelinek <jakub@redhat.com>
6313 * internal-fn.c: Include diagnostic-core.h.
6314 (expand_BUILTIN_EXPECT): New function.
6315 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
6316 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
6317 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
6318 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
6320 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
6321 Revert 3 argument __builtin_expect code.
6322 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
6323 * gimple-fold.c (gimple_fold_call): Likewise.
6324 * tree.h (fold_builtin_expect): New prototype.
6325 * builtins.c (build_builtin_expect_predicate): Add predictor
6326 argument, if non-NULL, create 3 argument __builtin_expect.
6327 (fold_builtin_expect): No longer static. Add ARG2 argument,
6328 pass it through to build_builtin_expect_predicate.
6329 (fold_builtin_2): Adjust caller.
6330 (fold_builtin_3): Handle BUILT_IN_EXPECT.
6331 * internal-fn.def (BUILTIN_EXPECT): New.
6333 2014-03-18 Tobias Burnus <burnus@net-b.de>
6336 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
6337 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
6338 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
6340 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
6343 * predict.c (combine_predictions_for_bb): Fix up formatting.
6344 (expr_expected_value_1, expr_expected_value): Add predictor argument,
6345 fill what it points to if non-NULL.
6346 (tree_predict_by_opcode): Adjust caller, use the predictor.
6347 * predict.def (PRED_COMPARE_AND_SWAP): Add.
6349 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
6351 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
6352 proper constant for the store mode.
6354 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
6356 * symtab.c (change_decl_assembler_name): Fix transparent alias
6359 2014-03-16 Renlin Li <Renlin.Li@arm.com>
6361 * config/aarch64/aarch64.c: Correct the comments about the
6362 aarch64 stack layout.
6364 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
6366 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
6367 check for GF_OMP_FOR_KIND_FOR.
6369 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
6371 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
6372 ymm and zmm register names.
6374 2014-03-17 Jakub Jelinek <jakub@redhat.com>
6377 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
6378 note creation for the 2010-08-31 changes.
6380 2014-03-17 Marek Polacek <polacek@redhat.com>
6383 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
6384 as -fno-tree-loop-vectorize.
6385 (expand_omp_simd): Likewise.
6387 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
6389 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
6390 (eligible_for_call_delay): New prototype.
6391 * config/sparc/sparc.c (tls_call_delay): Rename into...
6392 (eligible_for_call_delay): ...this. Return false if the instruction
6393 cannot be put in the delay slot of a branch.
6394 (eligible_for_restore_insn): Simplify.
6395 (eligible_for_return_delay): Return false if the instruction cannot be
6396 put in the delay slot of a branch and simplify.
6397 (eligible_for_sibcall_delay): Return false if the instruction cannot be
6398 put in the delay slot of a branch.
6399 * config/sparc/sparc.md (fix_ut699): New attribute.
6400 (tls_call_delay): Delete.
6401 (in_call_delay): Reimplement.
6402 (eligible_for_sibcall_delay): Rename into...
6403 (in_sibcall_delay): ...this.
6404 (eligible_for_return_delay): Rename into...
6405 (in_return_delay): ...this.
6406 (in_branch_delay): Reimplement.
6407 (in_uncond_branch_delay): Delete.
6408 (in_annul_branch_delay): Delete.
6410 2014-03-14 Richard Henderson <rth@redhat.com>
6413 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
6414 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
6415 (*floathi<X87MODEF>2_i387_with_temp): Remove.
6416 (floathi splitters): Remove.
6417 (float<SWI48x>xf2): New pattern.
6418 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
6419 code that tried to handle DImode for 32-bit, but which was excluded
6420 by the pattern's condition. Drop allocation of stack temporary.
6421 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
6422 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
6423 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
6424 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
6425 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
6426 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
6427 (*float<SWI48><MODEF>2_sse_interunit): Remove.
6428 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
6429 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
6430 (*float<SWI48x><X87MODEF>2_i387): Remove.
6431 (all float _with_temp splitters): Remove.
6432 (*float<SWI48x><MODEF>2_i387): New pattern.
6433 (*float<SWI48><MODEF>2_sse): New pattern.
6434 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
6435 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
6437 2014-03-14 Jakub Jelinek <jakub@redhat.com>
6438 Marek Polacek <polacek@redhat.com>
6441 * common.opt (dump_base_name_prefixed): New Variable.
6442 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
6443 if x_dump_base_name_prefixed is already set, set it at the end.
6445 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
6447 PR rtl-optimization/60508
6448 * lra-constraints.c (get_reload_reg): Add new parameter
6450 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
6451 Pass the new parameter values.
6453 2014-03-14 Richard Biener <rguenther@suse.de>
6455 * common.opt: Revert unintented changes from r205065.
6458 2014-03-14 Richard Biener <rguenther@suse.de>
6461 * cfghooks.c (split_block): Properly adjust all loops the
6462 block was a latch of.
6464 2014-03-14 Martin Jambor <mjambor@suse.cz>
6467 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
6470 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
6473 * config/avr/avr.c (avr_set_current_function): Pass function name
6474 through default_strip_name_encoding before sanity checking instead
6475 of skipping the first char of the assembler name.
6477 2014-03-13 Richard Henderson <rth@redhat.com>
6480 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
6481 (ix86_force_to_memory, ix86_free_from_memory): Remove.
6482 * config/i386/i386-protos.h: Likewise.
6483 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
6484 in the expander instead of a splitter.
6485 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
6486 any possibility of requiring a memory.
6487 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
6488 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
6489 (fp branch splitters): Update for ix86_split_fp_branch.
6490 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
6491 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
6492 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
6493 (*fop_<MODEF>_2_i387): Remove f/r alternative.
6494 (*fop_<MODEF>_3_i387): Likewise.
6495 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
6496 (splitters for the fop_* register patterns): Remove.
6497 (fscalexf4_i387): Rename from *fscalexf4_i387.
6498 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
6500 2014-03-13 Jakub Jelinek <jakub@redhat.com>
6502 PR tree-optimization/59779
6503 * tree-dfa.c (get_ref_base_and_extent): Use double_int
6504 type for bitsize and maxsize instead of HOST_WIDE_INT.
6506 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
6508 PR rtl-optimization/57320
6509 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
6510 the CFG after thread_prologue_and_epilogue_insns.
6512 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
6514 PR rtl-optimization/57189
6515 * lra-constraints.c (process_alt_operands): Disfavor spilling
6518 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
6520 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
6522 2014-03-13 Jakub Jelinek <jakub@redhat.com>
6524 PR tree-optimization/59025
6526 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
6527 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
6529 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
6532 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
6533 calls of avr_out_plus_1.
6535 2014-03-13 Bin Cheng <bin.cheng@arm.com>
6537 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
6538 BB's single pred and update the father loop's latch info later.
6540 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
6542 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
6546 (VEC_base): Likewise.
6547 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
6548 registers, we need to swap double words in little endian mode.
6550 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
6551 to be a container mode for 128-bit integer operations added in ISA
6552 2.07. Unlike TImode and PTImode, the preferred register set is
6553 the Altivec/VMX registers for the 128-bit operations.
6555 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
6557 (rs6000_split_128bit_ok_p): Likewise.
6559 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
6560 macros for creating ISA 2.07 normal and overloaded builtin
6561 functions with 3 arguments.
6562 (BU_P8V_OVERLOAD_3): Likewise.
6563 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
6564 for use as overloaded functions.
6565 (VPERM_1TI_UNS): Likewise.
6566 (VSEL_1TI): Likewise.
6567 (VSEL_1TI_UNS): Likewise.
6568 (ST_INTERNAL_1ti): Likewise.
6569 (LD_INTERNAL_1ti): Likewise.
6570 (XXSEL_1TI): Likewise.
6571 (XXSEL_1TI_UNS): Likewise.
6572 (VPERM_1TI): Likewise.
6573 (VPERM_1TI_UNS): Likewise.
6574 (XXPERMDI_1TI): Likewise.
6575 (SET_1TI): Likewise.
6576 (LXVD2X_V1TI): Likewise.
6577 (STXVD2X_V1TI): Likewise.
6578 (VEC_INIT_V1TI): Likewise.
6579 (VEC_SET_V1TI): Likewise.
6580 (VEC_EXT_V1TI): Likewise.
6581 (EQV_V1TI): Likewise.
6582 (NAND_V1TI): Likewise.
6583 (ORC_V1TI): Likewise.
6584 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
6585 added in ISA 2.07. Add both normal 'altivec' builtins, and the
6587 (VADDUQM): Likewise.
6588 (VSUBCUQ): Likewise.
6589 (VADDEUQM): Likewise.
6590 (VADDECUQ): Likewise.
6591 (VSUBEUQM): Likewise.
6592 (VSUBECUQ): Likewise.
6594 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
6595 __int128_t and __uint128_t types.
6596 (__uint128_type): Likewise.
6597 (altivec_categorize_keyword): Add support for vector __int128_t,
6598 vector __uint128_t, vector __int128, and vector unsigned __int128
6599 as a container type for TImode operations that need to be done in
6600 VSX/Altivec registers.
6601 (rs6000_macro_to_expand): Likewise.
6602 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
6603 to support 128-bit integer instructions vaddcuq, vadduqm,
6604 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
6605 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
6607 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
6608 for V1TImode, and set up preferences to use VSX/Altivec registers.
6609 Setup VSX reload handlers.
6610 (rs6000_debug_reg_global): Likewise.
6611 (rs6000_init_hard_regno_mode_ok): Likewise.
6612 (rs6000_preferred_simd_mode): Likewise.
6613 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
6614 (easy_altivec_constant): Likewise.
6615 (output_vec_const_move): Likewise.
6616 (rs6000_expand_vector_set): Convert V1TImode set and extract to
6618 (rs6000_expand_vector_extract): Likewise.
6619 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
6621 (rs6000_const_vec): Add support for V1TImode.
6622 (rs6000_emit_le_vsx_load): Swap double words when loading or
6623 storing TImode/V1TImode.
6624 (rs6000_emit_le_vsx_store): Likewise.
6625 (rs6000_emit_le_vsx_move): Likewise.
6626 (rs6000_emit_move): Add support for V1TImode.
6627 (altivec_expand_ld_builtin): Likewise.
6628 (altivec_expand_st_builtin): Likewise.
6629 (altivec_expand_vec_init_builtin): Likewise.
6630 (altivec_expand_builtin): Likewise.
6631 (rs6000_init_builtins): Add support for V1TImode type. Add
6632 support for ISA 2.07 128-bit integer builtins. Define type names
6633 for the VSX/Altivec vector types.
6634 (altivec_init_builtins): Add support for overloaded vector
6635 functions with V1TImode type.
6636 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
6637 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
6639 (rs6000_split_128bit_ok_p): Likewise.
6640 (rs6000_handle_altivec_attribute): Create V1TImode from vector
6641 __int128_t and vector __uint128_t.
6643 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
6644 and mode attributes.
6651 (VS_scalar): Likewise.
6652 (VS_double): Likewise.
6653 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
6655 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
6656 we support the ISA 2.07 128-bit integer arithmetic instructions.
6657 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
6658 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
6659 and TImode types for use with the builtin functions.
6660 (V1TI_type_node): Likewise.
6661 (unsigned_V1TI_type_node): Likewise.
6662 (intTI_type_internal_node): Likewise.
6663 (uintTI_type_internal_node): Likewise.
6665 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
6666 128-bit builtin functions.
6667 (UNSPEC_VADDEUQM): Likewise.
6668 (UNSPEC_VADDECUQ): Likewise.
6669 (UNSPEC_VSUBCUQ): Likewise.
6670 (UNSPEC_VSUBEUQM): Likewise.
6671 (UNSPEC_VSUBECUQ): Likewise.
6672 (VM): Add V1TImode to vector mode iterators.
6674 (VI_unit): Likewise.
6675 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
6676 (altivec_vaddcuq): Likewise.
6677 (altivec_vsubuqm): Likewise.
6678 (altivec_vsubcuq): Likewise.
6679 (altivec_vaddeuqm): Likewise.
6680 (altivec_vaddecuq): Likewise.
6681 (altivec_vsubeuqm): Likewise.
6682 (altivec_vsubecuq): Likewise.
6684 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
6686 (BOOL_128): Likewise.
6687 (BOOL_REGS_OUTPUT): Likewise.
6688 (BOOL_REGS_OP1): Likewise.
6689 (BOOL_REGS_OP2): Likewise.
6690 (BOOL_REGS_UNARY): Likewise.
6691 (BOOL_REGS_AND_CR0): Likewise.
6693 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
6694 128-bit integer builtin support.
6695 (vec_vadduqm): Likewise.
6696 (vec_vaddecuq): Likewise.
6697 (vec_vaddeuqm): Likewise.
6698 (vec_vsubecuq): Likewise.
6699 (vec_vsubeuqm): Likewise.
6700 (vec_vsubcuq): Likewise.
6701 (vec_vsubuqm): Likewise.
6703 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6704 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
6705 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
6706 128-bit integer add/subtract to ISA 2.07.
6708 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
6710 * config/arc/arc.c (arc_predicate_delay_insns):
6711 Fix third argument passed to conditionalize_nonjump.
6713 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
6715 * config/aarch64/aarch64-builtins.c
6716 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
6717 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
6718 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
6719 instead of __builtin_lfloor.
6720 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
6722 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6724 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
6725 (tree_ssa_ifcombine_bb_1): New function.
6726 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
6727 is an empty forwarder block to then_bb or vice versa and then_bb
6728 and else_bb are effectively swapped.
6730 2014-03-12 Christian Bruel <christian.bruel@st.com>
6733 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
6734 REG_CFA_DEF_CFA note.
6735 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
6736 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
6738 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
6740 PR tree-optimization/60454
6741 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
6743 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6745 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
6746 Do not define target_cpu_default2 to generic.
6747 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
6748 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
6749 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
6751 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6752 Marc Glisse <marc.glisse@inria.fr>
6754 PR tree-optimization/60502
6755 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
6756 instead of build_low_bits_mask.
6758 2014-03-12 Jakub Jelinek <jakub@redhat.com>
6761 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
6762 if there are multiple uses, but op doesn't live on E edge.
6763 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
6764 clobber stmts before __builtin_unreachable.
6766 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
6768 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
6769 hard_frame_pointer_rtx.
6770 * cse.c (cse_insn): Remove volatile check.
6771 * cselib.c (cselib_process_insn): Likewise.
6772 * dse.c (scan_insn): Likewise.
6774 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6776 * config/arc/arc.c (conditionalize_nonjump): New function,
6778 (arc_ifcvt): ... this.
6779 (arc_predicate_delay_insns): Use it.
6781 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
6783 * config/arc/predicates.md (extend_operand): During/after reload,
6784 allow const_int_operand.
6785 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
6786 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
6787 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
6789 (umulsi3_highpart_i): Likewise.
6791 2014-03-11 Richard Biener <rguenther@suse.de>
6793 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
6794 Add asserts to guard possible wrong-code bugs.
6796 2014-03-11 Richard Biener <rguenther@suse.de>
6798 PR tree-optimization/60429
6799 PR tree-optimization/60485
6800 * tree-ssa-structalias.c (set_union_with_increment): Properly
6801 take into account all fields that overlap the shifted vars.
6802 (do_sd_constraint): Likewise.
6803 (do_ds_constraint): Likewise.
6804 (get_constraint_for_ptr_offset): Likewise.
6806 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
6808 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
6809 (nios2_compute_frame_layout):
6810 Add calculation of cfun->machine->fp_save_offset.
6811 (nios2_expand_prologue): Correct setting of frame pointer register
6813 (nios2_expand_epilogue): Update recovery of stack pointer from
6814 frame pointer accordingly.
6815 (nios2_initial_elimination_offset): Update calculation of offset
6816 for eliminating to HARD_FRAME_POINTER_REGNUM.
6818 2014-03-10 Jakub Jelinek <jakub@redhat.com>
6821 * ipa.c (symtab_remove_unreachable_nodes): Don't call
6822 cgraph_get_create_node on VAR_DECLs.
6824 2014-03-10 Richard Biener <rguenther@suse.de>
6827 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
6829 2014-03-08 Douglas B Rupp <rupp@gnat.com>
6831 * config/vms/vms.opt (vms_float_format): New variable.
6833 2014-03-08 Tobias Burnus <burnus@net-b.de>
6835 * doc/invoke.texi (-fcilkplus): Update implementation status.
6837 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
6838 Richard Biener <rguenther@suse.de>
6840 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
6841 consistently accross all TUs.
6842 (run_gcc): Enable -fshort-double automatically at link at link-time
6843 and disallow override.
6845 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
6848 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
6849 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
6850 if they can't be used.
6852 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6854 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
6855 for Solaris 11/x86 ld.
6856 * configure: Regenerate.
6858 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6860 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
6861 (LIB_TLS_SPEC): Save as ld_tls_libs.
6862 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
6863 (HAVE_AS_IX86_TLSLDM): New test.
6864 * configure, config.in: Regenerate.
6865 * config/i386/i386.c (legitimize_tls_address): Fall back to
6866 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
6867 cannot support TLS_MODEL_LOCAL_DYNAMIC.
6868 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
6869 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
6871 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
6873 * common.opt (fira-loop-pressure): Mark as optimization.
6875 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
6877 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
6878 an OpenMP mappable type.
6880 2014-03-06 Matthias Klose <doko@ubuntu.com>
6882 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
6883 MULTILIB_OSDIRNAMES is not defined.
6885 2014-03-06 Jakub Jelinek <jakub@redhat.com>
6886 Meador Inge <meadori@codesourcery.com>
6889 * config/arm/arm.c (arm_tls_symbol_p): Remove.
6890 (arm_legitimize_address): Call legitimize_tls_address for any
6891 arm_tls_referenced_p expression, handle constant addend. Call it
6892 before testing for !TARGET_ARM.
6893 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
6895 2014-03-06 Richard Biener <rguenther@suse.de>
6901 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6903 * tree-streamer.c (record_common_node): Assert we don't record
6904 nodes with type double.
6905 (preload_common_node): Skip type double, complex double and double
6906 pointer since it is now frontend dependent due to fshort-double option.
6908 2014-03-06 Richard Biener <rguenther@suse.de>
6910 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
6911 or -fno-lto is specified and the linker has full plugin support.
6912 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
6913 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
6914 * lto-wrapper.c (merge_and_complain): Merge compile-time
6915 optimization levels.
6916 (run_gcc): And pass it through to the link options.
6918 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
6922 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
6924 * cselib.c (remove_useless_values): Skip to avoid quadratic
6925 behavior if the condition moved from...
6926 (cselib_process_insn): ... here holds.
6928 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6931 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
6932 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
6935 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
6936 (TM_H): Add x86-tune.def.
6938 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6940 * config/aarch64/aarch64.c (generic_tunings):
6941 Use cortexa57_extra_costs.
6943 2014-03-05 Jakub Jelinek <jakub@redhat.com>
6946 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
6947 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
6948 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
6951 2014-03-04 Heiher <r@hev.cc>
6953 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
6954 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
6956 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
6958 * config/i386/predicates.md (const2356_operand): Change to ...
6959 (const2367_operand): ... this.
6960 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
6962 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6963 (*avx512pf_scatterpf<mode>sf): Ditto.
6964 (avx512pf_scatterpf<mode>df): Ditto.
6965 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6966 (*avx512pf_scatterpf<mode>df): Ditto.
6967 * config/i386/i386.c (ix86_expand_builtin): Update
6968 incorrect hint operand error message.
6970 2014-03-04 Richard Biener <rguenther@suse.de>
6972 * lto-section-in.c (lto_get_section_data): Fix const cast.
6974 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
6976 * tree-streamer.c (record_common_node): Assert we don't record
6977 nodes with type double.
6978 (preload_common_node): Skip type double, complex double and double
6979 pointer since it is now frontend dependent due to fshort-double option.
6981 2014-03-04 Richard Biener <rguenther@suse.de>
6984 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
6985 (lto_input_toplevel_asms): Likewise.
6986 * lto-section-in.c (lto_get_section_data): Instead do it here
6989 2014-03-04 Richard Biener <rguenther@suse.de>
6991 PR tree-optimization/60382
6992 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
6993 dead PHIs a reduction.
6995 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
6997 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
6999 (_mm_prefetch): Move out of GCC target("sse") pragma.
7000 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
7001 GCC target("prfchw") pragma.
7002 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
7004 * config/i386/i386.c (ix86_option_override_internal): Enable
7005 -mprfchw with -mprefetchwt1.
7007 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
7009 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
7012 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
7014 * opts.h (CL_PCH_IGNORE): Define.
7015 * targhooks.c (option_affects_pch_p):
7016 Return false for options that have CL_PCH_IGNORE set.
7017 * opt-functions.awk: Process PchIgnore.
7018 * doc/options.texi: Document PchIgnore.
7020 * config/arc/arc.opt (misize): Add PchIgnore property.
7022 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7024 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
7025 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
7026 constraint on constants to permit them being loaded into
7027 GENERAL_REGS or BASE_REGS.
7029 2014-03-03 Nick Clifton <nickc@redhat.com>
7031 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
7032 anti-cacnonical alternatives.
7033 (negandhi3_real): New pattern.
7034 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
7036 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
7038 * config/avr/avr-mcus.def: Remove atxmega16x1.
7039 * config/avr/avr-tables.opt: Regenerate.
7040 * config/avr/t-multilib: Regenerate.
7041 * doc/avr-mmcu.texi: Regenerate.
7043 2014-03-03 Tobias Grosser <tobias@grosser.es>
7044 Mircea Namolaru <mircea.namolaru@inria.fr>
7046 PR tree-optimization/58028
7047 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
7050 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7052 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
7053 not handled by recognizers.
7055 2014-03-03 Jakub Jelinek <jakub@redhat.com>
7058 * function.c (expand_function_end): Don't emit
7059 clobber_return_register sequence if clobber_after is a BARRIER.
7060 * cfgexpand.c (construct_exit_block): Append instructions before
7061 return_label to prev_bb.
7063 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7065 * config/rs6000/constraints.md: Document reserved use of "wc".
7067 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
7070 * ipa.c (function_and_variable_visibility): When dissolving comdat
7071 group, also set all symbols to local.
7073 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
7078 2013-12-14 Jan Hubicka <jh@suse.cz>
7080 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
7082 2014-03-02 Jon Beniston <jon@beniston.com>
7088 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
7089 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
7090 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
7091 (simple_return, *simple_return): New patterns
7092 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
7093 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
7095 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
7097 * dwarf2out.c (gen_subprogram_die): Tidy.
7099 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
7102 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
7103 (*mov_t_msb_neg_negc): ... this new insn.
7105 2014-02-28 Jason Merrill <jason@redhat.com>
7108 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
7111 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
7114 * dwarf2out.c (decltype_auto_die): New static.
7115 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
7116 (gen_type_die_with_usage): Handle 'decltype(auto)'.
7117 (is_cxx_auto): Likewise.
7119 2014-02-28 Ian Bolton <ian.bolton@arm.com>
7121 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
7122 we are not using general regs only.
7124 2014-02-28 Richard Biener <rguenther@suse.de>
7127 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
7128 previous fix and only allow to remove trivial pre-headers
7129 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
7130 (remove_forwarder_block): Properly update the latch of a loop.
7132 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7135 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
7136 (cselib_preserved_hash_table): New.
7137 (preserve_constants_and_equivs): Move preserved vals to it.
7138 (cselib_find_slot): Look it up first.
7139 (cselib_init): Initialize it.
7140 (cselib_finish): Release it.
7141 (dump_cselib_table): Dump it.
7143 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7146 * cselib.c (remove_useless_values): Skip to avoid quadratic
7147 behavior if the condition moved from...
7148 (cselib_process_insn): ... here holds.
7150 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
7153 * var-tracking.c (vt_initialize): Apply the same condition to
7154 preserve the CFA base value.
7156 2014-02-28 Joey Ye <joey.ye@arm.com>
7159 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
7160 if reload in progress or completed.
7162 2014-02-28 Tobias Burnus <burnus@net-b.de>
7165 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
7168 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
7170 * doc/tm.texi.in (Condition Code Status): Update documention for
7171 relative locations of cc0-setter and cc0-user.
7173 2014-02-27 Jeff Law <law@redhat.com>
7175 PR rtl-optimization/52714
7176 * combine.c (try_combine): When splitting an unrecognized PARALLEL
7177 into two independent simple sets, if I3 is a jump, ensure the
7178 pattern we place into I3 is a (set (pc) ...).
7180 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
7181 Jeff Law <law@redhat.com>
7183 PR rtl-optimization/49847
7184 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
7185 are in different blocks.
7186 * doc/tm.texi (Condition Code Status): Update documention for
7187 relative locations of cc0-setter and cc0-user.
7189 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
7192 * lra.c (lra_emit_add): Check SUBREG too.
7194 2014-02-27 Andreas Schwab <schwab@suse.de>
7196 * config/m68k/m68k.c (m68k_option_override): Disable
7197 -flive-range-shrinkage for classic m68k.
7198 (m68k_override_options_after_change): Likewise.
7200 2014-02-27 Marek Polacek <polacek@redhat.com>
7203 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
7204 -Wmaybe-uninitialized.
7206 2014-02-27 Alan Modra <amodra@gmail.com>
7209 * reload1.c (emit_input_reload_insns): When reload_override_in,
7210 set old to rl->in_reg when rl->in_reg is a subreg.
7212 2014-02-26 Richard Biener <rguenther@suse.de>
7215 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
7217 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
7219 * common/config/i386/predicates.md (const1256_operand): Remove.
7220 (const2356_operand): New.
7221 (const_1_to_2_operand): Remove.
7222 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
7223 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
7224 (*avx512pf_gatherpf<mode>sf): Ditto.
7225 (avx512pf_gatherpf<mode>df): Ditto.
7226 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7227 (*avx512pf_gatherpf<mode>df): Ditto.
7228 (avx512pf_scatterpf<mode>sf): Ditto.
7229 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
7230 (*avx512pf_scatterpf<mode>sf): Ditto.
7231 (avx512pf_scatterpf<mode>df): Ditto.
7232 (*avx512pf_scatterpf<mode>df_mask): Ditto.
7233 (*avx512pf_scatterpf<mode>df): Ditto.
7234 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
7236 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
7238 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
7239 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
7240 (_mm512_mask_testn_epi64_mask): Move to ...
7241 * config/i386/avx512cdintrin.h: Here.
7242 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
7243 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
7244 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
7245 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
7246 TARGET_AVX512F from TARGET_AVX512CD.
7248 2014-02-26 Richard Biener <rguenther@suse.de>
7251 * ipa.c (walk_polymorphic_call_targets): Properly guard
7252 call to inline_update_overall_summary.
7254 2014-02-26 Bin Cheng <bin.cheng@arm.com>
7257 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
7258 and latches only if requested. Fix latch if it is removed.
7259 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
7260 LOOPS_HAVE_PREHEADERS.
7262 2014-02-25 Andrew Pinski <apinski@cavium.com>
7264 * builtins.c (expand_builtin_thread_pointer): Create a new target
7265 when the target is NULL.
7267 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
7269 PR rtl-optimization/60317
7270 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
7271 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
7272 * lra-assigns.c: Include params.h.
7273 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
7274 other reload pseudos considerations.
7276 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7278 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
7279 to use canonical form for nor<mode>3.
7281 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7284 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
7287 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
7289 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
7290 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
7291 (ix86_handle_option): Handle OPT_mprefetchwt1.
7292 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
7293 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7295 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7296 OPTION_MASK_ISA_PREFETCHWT1.
7297 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
7298 (PTA_PREFETCHWT1): New.
7299 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
7300 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
7301 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
7302 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
7303 (*prefetch_avx512pf_<mode>_: Change into ...
7304 (*prefetch_prefetchwt1_<mode>: This.
7305 * config/i386/i386.opt (mprefetchwt1): New.
7306 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
7307 (_mm_prefetch): Handle intent to write.
7308 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
7310 2014-02-25 Richard Biener <rguenther@suse.de>
7313 * emit-rtl.c (mem_attrs_htab): Remove.
7314 (mem_attrs_htab_hash): Likewise.
7315 (mem_attrs_htab_eq): Likewise.
7316 (set_mem_attrs): Always allocate new mem-attrs when something changed.
7317 (init_emit_once): Do not allocate mem_attrs_htab.
7319 2014-02-25 Richard Biener <rguenther@suse.de>
7322 * lto-opts.c (lto_write_options): Output non-explicit conservative
7323 -fwrapv, -fno-trapv and -fno-strict-overflow.
7324 * lto-wrapper.c (merge_and_complain): Handle merging those options.
7325 (run_gcc): And pass them through.
7327 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7329 * sel-sched.c (calculate_new_fences): New parameter ptime.
7330 Calculate it as a maximum over all fence cycles.
7331 (sel_sched_region_2): Adjust the call to calculate_new_fences.
7332 Print the final schedule timing when sched_verbose.
7334 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
7336 PR rtl-optimization/60292
7337 * sel-sched.c (fill_vec_av_set): Do not reset target availability
7338 bit fot the fence instruction.
7340 2014-02-24 Alangi Derick <alangiderick@gmail.com>
7342 * calls.h: Fix typo in comment.
7344 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
7346 * config/pa/pa.c (pa_output_move_double): Don't valididate when
7347 adjusting offsetable addresses.
7349 2014-02-24 Guozhi Wei <carrot@google.com>
7351 * sparseset.h (sparseset_pop): Fix the wrong index.
7353 2014-02-24 Walter Lee <walt@tilera.com>
7355 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
7356 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
7358 * common/config/tilegx/tilegx-common.c
7359 (TARGET_DEFAULT_TARGET_FLAGS): Define.
7360 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
7362 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
7363 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
7364 (tilegx_gimplify_va_arg_expr): Handle big endian.
7365 (tilegx_expand_unaligned_load): Ditto.
7366 (tilegx_expand_unaligned_store): Ditto.
7367 (TARGET_RETURN_IN_MSB): New.
7368 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
7369 (TARGET_ENDIAN_DEFAULT): New.
7370 (TARGET_BIG_ENDIAN): Handle big endian.
7371 (BYTES_BIG_ENDIAN): Ditto.
7372 (WORDS_BIG_ENDIAN): Ditto.
7373 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
7376 * config/tilegx/tilegx.md (extv): Handle big endian.
7378 (insn_st<n>): Ditto.
7379 (insn_st<n>_add<bitsuffix>): Ditto.
7380 (insn_stnt<n>): Ditto.
7381 (insn_stnt<n>_add<bitsuffix>):Ditto.
7382 (vec_interleave_highv8qi): Handle big endian.
7383 (vec_interleave_highv8qi_be): New.
7384 (vec_interleave_highv8qi_le): New.
7385 (insn_v1int_h): Handle big endian.
7386 (vec_interleave_lowv8qi): Handle big endian.
7387 (vec_interleave_lowv8qi_be): New.
7388 (vec_interleave_lowv8qi_le): New.
7389 (insn_v1int_l): Handle big endian.
7390 (vec_interleave_highv4hi): Handle big endian.
7391 (vec_interleave_highv4hi_be): New.
7392 (vec_interleave_highv4hi_le): New.
7393 (insn_v2int_h): Handle big endian.
7394 (vec_interleave_lowv4hi): Handle big endian.
7395 (vec_interleave_lowv4hi_be): New.
7396 (vec_interleave_lowv4hi_le): New.
7397 (insn_v2int_l): Handle big endian.
7398 (vec_interleave_highv2si): Handle big endian.
7399 (vec_interleave_highv2si_be): New.
7400 (vec_interleave_highv2si_le): New.
7401 (insn_v4int_h): Handle big endian.
7402 (vec_interleave_lowv2si): Handle big endian.
7403 (vec_interleave_lowv2si_be): New.
7404 (vec_interleave_lowv2si_le): New.
7405 (insn_v4int_l): Handle big endian.
7406 * config/tilegx/tilegx.opt (mbig-endian): New option.
7407 (mlittle-endian): New option.
7408 * doc/install.texi: Document tilegxbe-linux.
7409 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
7411 2014-02-24 Martin Jambor <mjambor@suse.cz>
7414 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
7415 there are no parameter descriptors.
7417 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
7419 PR rtl-optimization/60268
7420 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
7421 initialization to ...
7422 (sched_rgn_init): ... here.
7423 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
7425 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7427 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
7430 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
7432 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
7435 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7437 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
7438 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
7440 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
7442 * config/microblaze/predicates.md: Add cmp_op predicate.
7443 * config/microblaze/microblaze.md: Add branch_compare instruction
7444 which uses cmp_op predicate and emits cmp insn before branch.
7445 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
7446 to microblaze_expand_conditional_branch and consolidate logic.
7447 (microblaze_expand_conditional_branch): emit branch_compare
7448 insn instead of handling cmp op separate from branch insn.
7450 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7452 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
7455 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7457 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
7458 define_insn with define_expand and new define_insn
7459 *altivec_lve<VI_char>x_internal.
7460 (altivec_stve<VI_char>x): Replace define_insn with define_expand
7461 and new define_insn *altivec_stve<VI_char>x_internal.
7462 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
7464 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
7466 (altivec_expand_stvex_be): New function.
7468 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
7470 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
7471 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
7472 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
7473 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
7475 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
7478 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
7479 instead of emit_move_insn.
7481 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7483 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
7485 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
7486 gen_altivec_vsumsws.
7488 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7490 * config/rs6000/altivec.md (altivec_lvxl): Rename as
7491 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
7492 (altivec_lvxl_<mode>): New define_expand incorporating
7493 -maltivec=be semantics where needed.
7494 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
7495 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
7496 semantics where needed.
7497 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
7498 (altivec_stvx_<mode>): New define_expand incorporating
7499 -maltivec=be semantics where needed.
7500 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
7501 VM2 iterator instead of V4SI.
7502 (altivec_stvxl_<mode>): New define_expand incorporating
7503 -maltivec=be semantics where needed.
7504 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
7505 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
7506 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
7507 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
7508 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
7509 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
7510 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
7511 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
7512 ALTIVEC_BUILTIN_STVXL.
7513 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
7514 (altivec_expand_stvx_be): Likewise.
7515 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
7516 (altivec_expand_lvx_be): Likewise.
7517 (altivec_expand_stvx_be): Likewise.
7518 (altivec_expand_builtin): Add cases for
7519 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
7520 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
7521 (altivec_init_builtins): Add definitions for
7522 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
7523 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
7525 2014-02-21 Catherine Moore <clm@codesourcery.com>
7527 * doc/invoke.texi (mvirt, mno-virt): Document.
7528 * config/mips/mips.opt (mvirt): New option.
7529 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
7531 2014-02-21 Richard Biener <rguenther@suse.de>
7533 PR tree-optimization/60276
7534 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
7535 (STMT_VINFO_MIN_NEG_DIST): New macro.
7536 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
7537 STMT_VINFO_MIN_NEG_DIST.
7538 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
7539 made for negative dependence distances still hold.
7541 2014-02-21 Richard Biener <rguenther@suse.de>
7544 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
7545 DECL_INITIAL for globals not in the current function context.
7547 2014-02-21 Jakub Jelinek <jakub@redhat.com>
7549 PR tree-optimization/56490
7550 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
7551 * tree-ssa-uninit.c: Include params.h.
7552 (compute_control_dep_chain): Add num_calls argument, return false
7553 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
7554 num_calls to recursive call.
7555 (find_predicates): Change dep_chain into normal array,
7556 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
7557 variable and adjust compute_control_dep_chain caller.
7558 (find_def_preds): Likewise.
7560 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
7562 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
7563 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
7565 2014-02-21 Nick Clifton <nickc@redhat.com>
7567 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
7568 (pushhi1): Likewise.
7569 (popqi1): Add mode to pre_dec.
7572 2014-02-21 Jakub Jelinek <jakub@redhat.com>
7574 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
7575 mode for mask of V8SFmode permutation.
7577 2014-02-20 Richard Henderson <rth@redhat.com>
7580 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
7581 a new pseudo for OLDVAL.
7583 2014-02-20 Jakub Jelinek <jakub@redhat.com>
7586 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
7587 gen_reg_rtx if d->testing_p.
7588 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
7589 if d->testing_p and we will certainly return true.
7590 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
7593 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
7595 * emit-rtl.c (gen_reg_rtx): Assert that
7596 crtl->emit.regno_pointer_align_length is non-zero.
7598 2014-02-20 Richard Henderson <rth@redhat.com>
7601 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
7602 on failure the store back into EXPECT.
7604 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
7605 Sandra Loosemore <sandra@codesourcery.com>
7607 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
7608 * config/nios2/nios2.c (nios2_function_profiler): Add
7609 -fPIC (flag_pic == 2) support.
7610 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
7611 (nios2_large_offset_p): New function.
7612 (nios2_unspec_reloc_p): Move up position, update to use
7613 nios2_large_offset_p.
7614 (nios2_unspec_address): Remove function.
7615 (nios2_unspec_offset): New function.
7616 (nios2_large_got_address): New function.
7617 (nios2_got_address): Add large offset support.
7618 (nios2_legitimize_tls_address): Update usage of removed and new
7620 (nios2_symbol_binds_local_p): New function.
7621 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
7622 (nios2_legitimize_address): Update to use nios2_large_offset_p.
7623 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
7624 (nios2_print_operand): Merge H/L processing, add hiadj/lo
7625 processing for (const (unspec ...)).
7626 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
7628 2014-02-20 Richard Biener <rguenther@suse.de>
7630 * tree-cfg.c (replace_uses_by): Mark altered BBs before
7631 doing the substitution.
7632 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
7634 2014-02-20 Martin Jambor <mjambor@suse.cz>
7637 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
7638 info when checking whether lattices are bottom.
7640 2014-02-20 Richard Biener <rguenther@suse.de>
7643 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
7646 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
7649 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
7650 parameter specifying the scaling.
7651 (inline_call): Update.
7652 (want_inline_recursively): Guard division by zero.
7653 (recursive_inlining): Update.
7654 * ipa-inline.h (clone_inlined_nodes): Update.
7656 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7659 * config/i386/i386.c (classify_argument): Pass structures of size
7660 64 bytes or less in register.
7662 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
7663 Kirill Yukhin <kirill.yukhin@intel.com>
7665 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
7666 (_mm_rcp28_round_ss): Ditto.
7667 (_mm_rsqrt28_round_sd): Ditto.
7668 (_mm_rsqrt28_round_ss): Ditto.
7669 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
7670 (_mm_rcp14_round_ss): Ditto.
7671 (_mm_rsqrt14_round_sd): Ditto.
7672 (_mm_rsqrt14_round_ss): Ditto.
7673 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
7674 the first input operand, get rid of match_dup.
7675 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
7677 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7679 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
7680 operand as the first input operand, set type attribute.
7681 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7683 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
7684 operand as the first input operand, set type attribute.
7686 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7688 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
7691 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
7694 * config/i386/i386.c (construct_container): Remove TFmode check
7695 for X86_64_INTEGER_CLASS.
7697 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
7700 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
7701 only when -Wpsabi is enabled.
7703 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
7706 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
7707 passing arrays in registers are the same as for structs, so remove the
7708 special case for them.
7710 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
7712 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
7713 destination type, extract only the valid bits if the source type is not
7714 integral and has a different mode.
7716 2014-02-19 Richard Biener <rguenther@suse.de>
7719 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
7722 2014-02-19 Richard Biener <rguenther@suse.de>
7725 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
7726 (ipa_modify_call_arguments): Emit an argument load explicitely and
7727 preserve virtual SSA form there and for the replacement call.
7728 Do not update SSA form nor free dominance info.
7730 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7732 * ipa.c (function_and_variable_visibility): Also clear WEAK
7733 flag when disolving COMDAT_GROUP.
7735 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7737 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
7738 * ipa-prop.c (ipa_set_jf_known_type): Return early when
7740 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
7741 do more sanity checks.
7742 (detect_type_change): Return true when giving up early.
7743 (compute_complex_assign_jump_func): Fix type parameter of
7744 ipa_set_ancestor_jf.
7745 (compute_complex_ancestor_jump_func): Likewise.
7746 (update_jump_functions_after_inlining): Fix updating of
7748 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
7750 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
7752 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
7753 inline clones when edge disappears.
7755 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7758 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
7759 Split 64-bit moves into 2 patterns. Do not allow the use of
7760 direct move for TDmode in little endian, since the decimal value
7761 has little endian bytes within a word, but the 64-bit pieces are
7762 ordered in a big endian fashion, and normal subreg's of TDmode are
7764 (mov<mode>_64bit_dm): Likewise.
7765 (movtd_64bit_nodm): Likewise.
7767 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7769 PR tree-optimization/60174
7770 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
7771 statement of an SSA_NAME that occurs in an abnormal PHI node.
7773 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7776 * final.c (SEEN_BB): Remove.
7777 (SEEN_NOTE, SEEN_EMITTED): Renumber.
7778 (final_scan_insn): Don't force_source_line on second
7779 NOTE_INSN_BASIC_BLOCK.
7781 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
7784 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
7785 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
7786 (type_natural_mode): Warn ABI change when %zmm register is not
7787 available for AVX512F vector value passing.
7789 2014-02-18 Kai Tietz <ktietz@redhat.com>
7792 * config/i386/i386.c (ix86_expand_prologue): Use value in
7793 rax register as displacement when restoring %r10 or %rax.
7794 Fix wrong offset when restoring both registers.
7796 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
7798 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
7799 assertion with conditional return.
7801 2014-02-18 Jakub Jelinek <jakub@redhat.com>
7802 Uros Bizjak <ubizjak@gmail.com>
7805 * config/i386/driver-i386.c (host_detect_local_cpu): If
7806 YMM state is not saved by the OS, also clear has_f16c. Move
7807 CPUID 0x80000001 handling before YMM state saving checking.
7809 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
7811 PR rtl-optimization/58960
7812 * haifa-sched.c (alloc_global_sched_pressure_data): New,
7813 factored out from ...
7814 (sched_init): ... here.
7815 (free_global_sched_pressure_data): New, factored out from ...
7816 (sched_finish): ... here.
7817 * sched-int.h (free_global_sched_pressure_data): Declare.
7818 * sched-rgn.c (nr_regions_initial): New static global.
7819 (haifa_find_rgns): Initialize it.
7820 (schedule_region): Disable sched-pressure for the newly
7823 2014-02-17 Richard Biener <rguenther@suse.de>
7825 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
7826 release SSA defs of pattern stmts.
7828 2014-02-17 Richard Biener <rguenther@suse.de>
7830 * tree-inline.c (expand_call_inline): Release the virtual
7831 operand defined by the call we are about to inline.
7833 2014-02-17 Richard Biener <rguenther@suse.de>
7835 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
7837 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
7838 Ilya Tocar <ilya.tocar@intel.com>
7840 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
7841 arguments order in builtin.
7842 (_mm512_permutexvar_epi64): Ditto.
7843 (_mm512_mask_permutexvar_epi64): Ditto
7844 (_mm512_maskz_permutexvar_epi32): Ditto
7845 (_mm512_permutexvar_epi32): Ditto
7846 (_mm512_mask_permutexvar_epi32): Ditto
7848 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7850 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
7851 (p8_vmrgow): Likewise.
7853 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7855 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
7858 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
7861 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
7862 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
7863 into 64-bit and 32-bit moves. On 64-bit moves, add support for
7864 using direct move instructions on ISA 2.07. Also adjust
7865 instruction length for 64-bit.
7866 (mov<mode>_64bit, TFmode/TDmode): Likewise.
7867 (mov<mode>_32bit, TFmode/TDmode): Likewise.
7869 2014-02-15 Alan Modra <amodra@gmail.com>
7873 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
7874 find_replacement on parts of insn rtl that might be reloaded.
7876 2014-02-15 Richard Biener <rguenther@suse.de>
7878 PR tree-optimization/60183
7879 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
7880 (tree_ssa_phiprop): Calculate and free post-dominators.
7882 2014-02-14 Jeff Law <law@redhat.com>
7884 PR rtl-optimization/60131
7885 * ree.c (get_extended_src_reg): New function.
7886 (combine_reaching_defs): Use it rather than assuming location of REG.
7887 (find_and_remove_re): Verify first operand of extension is
7888 a REG before adding the insns to the copy list.
7890 2014-02-14 Roland McGrath <mcgrathr@google.com>
7892 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
7893 * configure: Regenerated.
7894 * config.in: Regenerated.
7895 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
7896 instead of ASM_SHORT.
7898 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
7899 Richard Earnshaw <rearnsha@arm.com>
7901 PR rtl-optimization/59535
7902 * lra-constraints.c (process_alt_operands): Encourage alternative
7903 when unassigned pseudo class is superset of the alternative class.
7904 (inherit_reload_reg): Don't inherit when optimizing for code size.
7905 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
7906 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
7907 modes not less than 4 for Thumb1.
7909 2014-02-14 Kyle McMartin <kyle@redhat.com>
7912 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
7914 2014-02-14 Richard Biener <rguenther@suse.de>
7916 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
7917 (get_frame_arg): Drop the assert with langhook types_compatible_p.
7918 Do not strip INDIRECT_REFs.
7920 2014-02-14 Richard Biener <rguenther@suse.de>
7923 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
7924 DECL_FUNCTION_SPECIFIC_TARGET.
7925 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
7926 * tree-streamer-out.c (pack_ts_target_option): Remove.
7927 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
7928 (write_ts_function_decl_tree_pointers): Do not stream
7929 DECL_FUNCTION_SPECIFIC_TARGET.
7930 * tree-streamer-in.c (unpack_ts_target_option): Remove.
7931 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
7932 (lto_input_ts_function_decl_tree_pointers): Do not stream
7933 DECL_FUNCTION_SPECIFIC_TARGET.
7935 2014-02-14 Jakub Jelinek <jakub@redhat.com>
7937 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
7938 (get_initial_def_for_induction, vectorizable_induction): Ignore
7939 debug stmts when looking for exit_phi.
7940 (vectorizable_live_operation): Fix up condition.
7942 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7944 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
7945 nreverse() because it changes the content of original tree list.
7947 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7949 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
7950 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
7952 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
7954 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
7955 GNU coding standards.
7957 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7960 * dwarf2out.c (gen_subprogram_die): Don't call
7961 add_calling_convention_attribute if subr_die is old_die.
7963 2014-02-13 Sharad Singhai <singhai@google.com>
7965 * doc/optinfo.texi: Fix order of nodes.
7967 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
7969 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
7970 operands[2], not operands[3].
7972 2014-02-13 Richard Biener <rguenther@suse.de>
7975 * doc/install.texi (ISL): Update recommended version to 0.12.2,
7976 mention the possibility of an in-tree build.
7977 (CLooG): Update recommended version to 0.18.1, mention the
7978 possibility of an in-tree build and clarify that the ISL
7979 bundled with CLooG does not work.
7981 2014-02-13 Jakub Jelinek <jakub@redhat.com>
7984 * expr.c (compress_float_constant): If x is a hard register,
7985 extend into a pseudo and then move to x.
7987 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
7989 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
7990 caused by bad second argument to warning_at() with -mhotpatch and
7991 nested functions (e.g. with gfortran).
7993 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
7995 * opts.c (option_name): Remove "enabled by default" rider.
7997 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
7999 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
8001 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
8002 Uros Bizjak <ubizjak@gmail.com>
8005 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
8006 * configure: Regenerated.
8008 2014-02-12 Richard Biener <rguenther@suse.de>
8010 * vec.c (vec_prefix::calculate_allocation): Move as
8011 inline variant to vec.h.
8012 (vec_prefix::calculate_allocation_1): New out-of-line version.
8013 * vec.h (vec_prefix::calculate_allocation_1): Declare.
8014 (vec_prefix::m_has_auto_buf): Rename to ...
8015 (vec_prefix::m_using_auto_storage): ... this.
8016 (vec_prefix::calculate_allocation): Inline the easy cases
8017 and dispatch to calculate_allocation_1 which doesn't need the
8019 (va_heap::reserve): Use gcc_checking_assert.
8020 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
8021 m_using_auto_storage.
8022 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
8024 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
8025 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
8026 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
8028 2014-02-12 Richard Biener <rguenther@suse.de>
8030 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
8031 when we found a dependence.
8033 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
8035 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
8037 (maybe_fold_stmt): ... into this new function.
8038 * omp-low.c (lower_omp): Update comment.
8040 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
8043 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
8046 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
8048 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
8049 identifiers in comments.
8050 (cortexa53_extra_costs): Likewise.
8051 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
8052 (cortexa7_extra_costs): Likewise.
8053 (cortexa12_extra_costs): Likewise.
8054 (cortexa15_extra_costs): Likewise.
8055 (v7m_extra_costs): Likewise.
8057 2014-02-12 Richard Biener <rguenther@suse.de>
8060 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
8061 of posix_memalign being successful.
8062 (lower_stmt): Restrict lowering of posix_memalign to when
8063 -ftree-bit-ccp is enabled.
8065 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
8067 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
8069 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
8071 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
8073 PR rtl-optimization/60116
8074 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
8075 other_insn once the combination has been validated.
8077 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
8080 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
8082 * ipa-devirt.c: Include demangle.h
8083 (odr_violation_reported): New static variable.
8084 (add_type_duplicate): Update odr_violations.
8085 (maybe_record_node): Add completep parameter; update it.
8086 (record_target_from_binfo): Add COMPLETEP parameter;
8087 update it as needed.
8088 (possible_polymorphic_call_targets_1): Likewise.
8089 (struct polymorphic_call_target_d): Add nonconstruction_targets;
8090 rename FINAL to COMPLETE.
8091 (record_targets_from_bases): Sanity check we found the binfo;
8092 fix COMPLETEP updating.
8093 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
8094 parameter, fix computing of COMPLETEP.
8095 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
8096 at LTO time do demangling.
8097 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
8098 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
8100 (gimple_get_virt_method_for_binfo): Likewise.
8101 * gimple-fold.h (gimple_get_virt_method_for_binfo,
8102 gimple_get_virt_method_for_vtable): Update prototypes.
8104 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
8107 * genautomata.c (add_presence_absence): Fix typo with
8108 {final_}presence_list.
8110 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
8113 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
8114 for VSX/Altivec vectors that land in GPR registers.
8116 2014-02-11 Richard Henderson <rth@redhat.com>
8117 Jakub Jelinek <jakub@redhat.com>
8120 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
8121 around drhs if type conversion to lacc->type is not useless.
8123 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8125 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
8127 (cortex-a57.cortex-a53): Likewise.
8128 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
8130 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8132 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
8135 2014-02-11 Renlin Li <Renlin.Li@arm.com>
8137 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
8138 add_options_for_arm_vfp3.
8140 2014-02-11 Jeff Law <law@redhat.com>
8143 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
8144 object with an undesirable mode.
8146 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8149 * config/i386/sol2-9.h: New file.
8150 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
8151 *-*-solaris2.9*): Use it.
8153 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
8155 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
8156 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
8158 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
8160 * config/microblaze/microblaze.c: Extend mcpu version format
8162 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
8164 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
8166 2014-02-10 Richard Henderson <rth@redhat.com>
8169 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
8170 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
8171 ms-abi vs -mno-accumulate-outgoing-args.
8172 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
8173 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
8176 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
8179 * cfgexpand.c (expand_asm_operands): Attach source location to
8180 ASM_INPUT rtx objects.
8181 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
8183 2014-02-10 Nick Clifton <nickc@redhat.com>
8185 * config/mn10300/mn10300.c (popcount): New function.
8186 (mn10300_expand_prologue): Include saved registers in stack usage
8189 2014-02-10 Jeff Law <law@redhat.com>
8192 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
8193 when changing the SET_DEST of a prior insn to avoid an input reload.
8195 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8197 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
8198 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
8199 -mcall-openbsd, or -mcall-linux.
8200 (CC1_ENDIAN_BIG_SPEC): Remove.
8201 (CC1_ENDIAN_LITTLE_SPEC): Remove.
8202 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
8203 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
8204 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
8205 and %cc1_endian_default.
8206 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
8208 2014-02-10 Richard Biener <rguenther@suse.de>
8210 PR tree-optimization/60115
8211 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
8212 MEM_REF handling. Properly verify that the accesses are not
8213 out of the objects bound.
8215 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8217 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
8220 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
8222 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
8223 proper constants and fix formatting.
8224 (possible_polymorphic_call_targets): Fix formatting.
8226 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
8227 Ilya Tocar <ilya.tocar@intel.com>
8229 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
8230 (_mm512_loadu_epi32): Renamed into...
8231 (_mm512_loadu_si512): This.
8232 (_mm512_storeu_epi32): Renamed into...
8233 (_mm512_storeu_si512): This.
8234 (_mm512_maskz_ceil_ps): Removed.
8235 (_mm512_maskz_ceil_pd): Ditto.
8236 (_mm512_maskz_floor_ps): Ditto.
8237 (_mm512_maskz_floor_pd): Ditto.
8238 (_mm512_floor_round_ps): Ditto.
8239 (_mm512_floor_round_pd): Ditto.
8240 (_mm512_ceil_round_ps): Ditto.
8241 (_mm512_ceil_round_pd): Ditto.
8242 (_mm512_mask_floor_round_ps): Ditto.
8243 (_mm512_mask_floor_round_pd): Ditto.
8244 (_mm512_mask_ceil_round_ps): Ditto.
8245 (_mm512_mask_ceil_round_pd): Ditto.
8246 (_mm512_maskz_floor_round_ps): Ditto.
8247 (_mm512_maskz_floor_round_pd): Ditto.
8248 (_mm512_maskz_ceil_round_ps): Ditto.
8249 (_mm512_maskz_ceil_round_pd): Ditto.
8250 (_mm512_expand_pd): Ditto.
8251 (_mm512_expand_ps): Ditto.
8252 * config/i386/i386.c (ix86_builtins): Remove
8253 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
8254 (bdesc_args): Ditto.
8255 * config/i386/predicates.md (const1256_operand): New.
8256 (const_1_to_2_operand): Ditto.
8257 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
8258 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8259 (*avx512pf_gatherpf<mode>sf): Ditto.
8260 (avx512pf_gatherpf<mode>df): Ditto.
8261 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8262 (*avx512pf_gatherpf<mode>df): Ditto.
8263 (avx512pf_scatterpf<mode>sf): Ditto.
8264 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8265 (*avx512pf_scatterpf<mode>sf): Ditto.
8266 (avx512pf_scatterpf<mode>df): Ditto.
8267 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8268 (*avx512pf_scatterpf<mode>df): Ditto.
8269 (avx512f_expand<mode>): Removed.
8270 (<shift_insn><mode>3<mask_name>): Change predicate type.
8272 2014-02-08 Jakub Jelinek <jakub@redhat.com>
8274 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
8275 not at the end of datarefs vector use ordered_remove to avoid
8276 reordering datarefs vector.
8279 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
8280 mark local addressable non-static vars as GOVD_PRIVATE
8281 instead of GOVD_LOCAL.
8282 * omp-low.c (lower_omp_for): Move gimple_bind_vars
8283 and BLOCK_VARS of gimple_bind_block to new_stmt rather
8287 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
8288 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
8289 assume_aligned or alloc_align attributes.
8290 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
8291 arguments. Handle also assume_aligned and alloc_align attributes.
8292 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
8293 calls to functions with assume_aligned or alloc_align attributes.
8294 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
8296 2014-02-08 Terry Guo <terry.guo@arm.com>
8298 * doc/invoke.texi: Document ARM -march=armv7e-m.
8300 2014-02-08 Jakub Jelinek <jakub@redhat.com>
8302 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
8303 flag on __cilkrts_rethrow builtin.
8306 * ipa-cp.c (determine_versionability): Fail at -O0
8307 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
8308 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
8311 2014-02-04 Jakub Jelinek <jakub@redhat.com>
8314 * tree-inline.c (copy_forbidden): Fail for
8315 __attribute__((optimize (0))) functions.
8317 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
8319 * varpool.c: Include pointer-set.h.
8320 (varpool_remove_unreferenced_decls): Variables in other partitions
8321 will not be output; be however careful to not lose information
8324 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
8326 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
8327 lookup in the vtable constructor.
8329 2014-02-07 Jeff Law <law@redhat.com>
8332 * config/m68k/m68k.md (ashldi_extsi): Turn into a
8333 define_insn_and_split.
8335 * ipa-inline.c (inline_small_functions): Fix typos.
8337 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8339 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
8340 (s390_can_use_return_insn): Declare.
8341 * config/s390/s390.h (EPILOGUE_USES): Define.
8342 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
8344 (s390_chunkify_start): Handle return JUMP_LABELs.
8345 (s390_early_mach): Emit a main_pool instruction on the entry edge.
8346 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
8347 (s390_can_use_return_insn): New functions.
8348 (s390_fix_long_loop_prediction): Handle conditional returns.
8349 (TARGET_SET_UP_BY_PROLOGUE): Define.
8350 * config/s390/s390.md (ANY_RETURN): New code iterator.
8351 (*creturn, *csimple_return, return, simple_return): New patterns.
8353 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8355 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
8356 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
8357 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
8358 REG_CFA_RESTORE list when deciding not to restore a register.
8360 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8362 * config/s390/s390.c: Include tree-pass.h and context.h.
8363 (s390_early_mach): New function, split out from...
8364 (s390_emit_prologue): ...here.
8365 (pass_data_s390_early_mach): New pass structure.
8366 (pass_s390_early_mach): New class.
8367 (s390_option_override): Create and register early_mach pass.
8368 Move to end of file.
8370 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8372 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
8373 to match for the exit block.
8375 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8377 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
8378 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
8379 Reject misaligned operands.
8381 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8383 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
8385 2014-02-07 Richard Biener <rguenther@suse.de>
8388 * gimple-low.c (lower_builtin_posix_memalign): New function.
8389 (lower_stmt): Call it to lower posix_memalign in a way
8390 to make alignment info accessible.
8392 2014-02-07 Jakub Jelinek <jakub@redhat.com>
8395 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
8396 __builtin_setjmp_receiver.
8398 2014-02-07 Richard Biener <rguenther@suse.de>
8401 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
8402 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
8403 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
8404 Handle BUILT_IN_POSIX_MEMALIGN.
8405 (find_func_clobbers): Likewise.
8406 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
8407 (call_may_clobber_ref_p_1): Likewise.
8409 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8412 * ipa-devirt.c (record_target_from_binfo): Remove overactive
8415 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8418 * lto-cgraph.c (lto_output_node): Use
8419 symtab_get_symbol_partitioning_class.
8420 (lto_output_varpool_node): likewise.
8421 (symtab_get_symbol_partitioning_class): Move here from
8423 * cgraph.h (symbol_partitioning_class): Likewise.
8424 (symtab_get_symbol_partitioning_class): Declare.
8426 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8428 * ggc.h (ggc_internal_cleared_alloc): New macro.
8429 * vec.h (vec_safe_copy): Handle memory stats.
8430 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
8431 * target-globals.c (save_target_globals): Likewise.
8433 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
8436 * expr.c (emit_move_resolve_push): Export; be bit more selective
8437 on when to clear alias set.
8438 * expr.h (emit_move_resolve_push): Declare.
8439 * function.h (struct function): Add tail_call_marked.
8440 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
8441 * config/i386/i386-protos.h (ix86_expand_push): Remove.
8442 * config/i386/i386.md (TImode move expander): De not call
8444 (FP push expanders): Preserve memory attributes.
8445 * config/i386/sse.md (push<mode>1): Remove.
8446 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
8447 (ix86_expand_push): Remove.
8448 * config/i386/mmx.md (push<mode>1): Remove.
8450 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8452 PR rtl-optimization/60030
8453 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
8454 lopart with paradoxical subreg before shifting it up by hprec.
8456 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8458 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
8459 Remove extra newline at end of file.
8460 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
8461 (arm_issue_rate): Handle cortexa57.
8462 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
8463 (cortex-a57.cortex-a53): Likewise.
8465 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8468 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
8469 don't record in REG_FRAME_RELATED_EXPR registers not set in that
8471 (arm_expand_prologue): Adjust all callers.
8472 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
8473 info, registers also at the lowest numbered registers side. Use
8474 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
8478 * var-tracking.c (adjust_mems): Before adding a SET to
8479 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
8481 2014-02-06 Alan Modra <amodra@gmail.com>
8484 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
8485 change SDmode to DDmode when lra_in_progress.
8487 2014-02-06 Jakub Jelinek <jakub@redhat.com>
8490 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
8491 free_data_ref on the dr first, and before goto again also set dr
8492 to the next dr. For simd_lane_access, free old datarefs[i] before
8493 overwriting it. For get_vectype_for_scalar_type failure, don't
8494 free_data_ref if simd_lane_access.
8496 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
8499 * tree.h (opts_for_fn): New inline function.
8500 (opt_for_fn): Define.
8501 * config/i386/i386.c (ix86_function_regparm): Use
8502 opt_for_fn (decl, optimize) instead of optimize.
8504 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
8506 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
8507 for SYMBOL_REF in large memory model.
8509 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8511 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
8513 (cortex-a57): Likewise.
8514 (cortex-a57.cortex-a53): Likewise.
8516 2014-02-06 Yury Gribov <y.gribov@samsung.com>
8517 Kugan Vivekanandarajah <kuganv@linaro.org>
8519 * config/arm/arm.c (arm_vector_alignment_reachable): Check
8521 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
8523 2014-02-06 Richard Biener <rguenther@suse.de>
8525 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
8526 set_loop_copy and initialize_original_copy_tables.
8528 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
8530 * config/aarch64/aarch64-simd.md
8531 (aarch64_ashr_simddi): Change QI to SI.
8533 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
8534 Jakub Jelinek <jakub@redhat.com>
8537 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
8540 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8542 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
8543 CODE_FOR_altivec_vpku[hw]um to
8544 CODE_FOR_altivec_vpku[hw]um_direct.
8545 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
8546 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
8547 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
8548 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
8550 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8552 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
8553 generation for -maltivec=be.
8554 (altivec_vsumsws): Simplify redundant test.
8556 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8558 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
8559 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
8560 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
8561 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
8562 gen_altivec_vpkuwum.
8563 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
8565 (altivec_vpks<VI_char>ss): Likewise.
8566 (altivec_vpks<VI_char>us): Likewise.
8567 (altivec_vpku<VI_char>us): Likewise.
8568 (altivec_vpku<VI_char>um): Likewise.
8569 (altivec_vpku<VI_char>um_direct): New (copy of
8570 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
8572 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
8573 target is little endian and -maltivec=be is not specified.
8574 (*altivec_vupkhs<VU_char>_direct): New (copy of
8575 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
8576 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
8577 target is little endian and -maltivec=be is not specified.
8578 (*altivec_vupkls<VU_char>_direct): New (copy of
8579 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
8580 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
8581 little endian and -maltivec=be is not specified.
8582 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
8583 little endian and -maltivec=be is not specified.
8585 2014-02-05 Richard Henderson <rth@redhat.com>
8588 * combine-stack-adj.c: Revert r206943.
8589 * sched-int.h (struct deps_desc): Add last_args_size.
8590 * sched-deps.c (init_deps): Initialize it.
8591 (sched_analyze_insn): Add OUTPUT dependencies between insns that
8592 contain REG_ARGS_SIZE notes.
8594 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
8596 * lto-cgraph.c (asm_nodes_output): Make global.
8597 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
8598 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
8599 (driver_handle_option): Handle OPT_fwpa.
8601 2014-02-05 Jakub Jelinek <jakub@redhat.com>
8604 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
8605 a comment typo and formatting issue. If odr_hash hasn't been
8606 created, return vNULL and set *completep to false.
8609 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
8610 bb with no successors.
8612 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
8615 * doc/invoke.texi (-march): Clarify documentation for ARM.
8619 2014-02-05 Richard Biener <rguenther@suse.de>
8621 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
8622 when not vectorizing because of too many alias checks.
8623 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
8624 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
8626 2014-02-05 Nick Clifton <nickc@redhat.com>
8628 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
8629 accept extended registers in any mode when compiling for the MN10300.
8631 2014-02-05 Yury Gribov <y.gribov@samsung.com>
8633 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
8634 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
8635 sanitization attributes.
8636 (can_inline_edge_p): Likewise.
8637 (sanitize_attrs_match_for_inline_p): New function.
8639 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8641 * ipa-prop.c (detect_type_change): Shor circuit testing of
8642 type changes on THIS pointer.
8644 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
8647 * config/pa/pa.c (legitimize_tls_address): Return original address
8648 if not passed a SYMBOL_REF rtx.
8649 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
8651 (pa_emit_move_sequence): Simplify TLS source operands.
8652 (pa_legitimate_constant_p): Reject all TLS constants.
8653 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
8654 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
8656 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8658 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
8659 groups when we know they are controlled by LTO.
8660 * varasm.c (default_binds_local_p_1): If object is in other partition,
8661 it will be resolved locally.
8663 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8665 * config/host-linux.c (linux_gt_pch_use_address): Don't
8666 use SSIZE_MAX because it is not always defined.
8668 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
8671 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
8672 threshold for pseudo splitting.
8673 (update_ebb_live_info): Process call argument hard registers and
8674 hard registers from insn definition too.
8675 (max_small_class_regs_num): New constant.
8676 (inherit_in_ebb): Update live hard regs through EBBs. Update
8677 reloads_num only for small register classes. Don't split for
8680 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
8683 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
8686 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
8688 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
8691 2014-02-04 Marek Polacek <polacek@redhat.com>
8693 * gdbinit.in (pel): Define.
8695 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
8697 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
8700 2014-02-04 Richard Biener <rguenther@suse.de>
8703 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
8704 in function context local.
8705 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
8706 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
8707 similar to LTO_imported_decl_ref.
8709 2014-02-04 Jakub Jelinek <jakub@redhat.com>
8711 PR tree-optimization/60002
8712 * cgraphclones.c (build_function_decl_skip_args): Clear
8715 PR tree-optimization/60023
8716 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
8717 false to gsi_replace.
8718 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
8719 has been in some EH region and vec_stmt could throw, add
8720 vec_stmt into the same EH region.
8721 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
8722 has no lhs, ignore it.
8723 * internal-fn.c (expand_MASK_LOAD): Likewise.
8726 * tree-inline.c (copy_forbidden): Fail for
8727 __attribute__((optimize (0))) functions.
8730 * omp-low.c (simd_clone_struct_copy): If from->inbranch
8731 is set, copy one less argument.
8732 (expand_simd_clones): Don't subtract clone_info->inbranch
8733 from simd_clone_struct_alloc argument.
8735 PR rtl-optimization/57915
8736 * recog.c (simplify_while_replacing): If all unary/binary/relational
8737 operation arguments are constant, attempt to simplify those.
8740 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
8741 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
8743 2014-02-04 Richard Biener <rguenther@suse.de>
8745 PR tree-optimization/60012
8746 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
8747 TBAA disambiguation to all DDRs.
8749 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8752 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
8753 (LINK_SPEC): Use it for -shared, -shared-libgcc.
8755 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8758 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
8760 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8762 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
8763 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
8765 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8768 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
8769 to figure out targets of polymorphic calls with known decl.
8770 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8771 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
8772 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
8773 (get_polymorphic_call_info): ... here.
8774 (get_polymorphic_call_info_from_invariant): New function.
8776 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8778 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
8779 lookup via vtable pointer; check for type consistency
8780 and turn inconsitent facts into UNREACHABLE.
8781 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8782 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
8783 type inconsistent querries; return UNREACHABLE instead.
8785 2014-02-03 Richard Henderson <rth@twiddle.net>
8788 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
8789 already processed this node.
8790 (normalize_one_pred_1): Pass along mark_set.
8791 (normalize_one_pred): Create and destroy a pointer_set_t.
8792 (normalize_one_pred_chain): Likewise.
8794 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
8796 PR gcov-profile/58602
8797 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
8799 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
8802 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
8803 -fno-devirtualize; try to devirtualize by the knowledge of
8804 virtual table pointer given by aggregate propagation.
8805 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
8806 (ipa_print_node_jump_functions): Dump also offset that
8807 is relevant for polymorphic calls.
8808 (determine_known_aggregate_parts): Add arg_type parameter; use it
8809 instead of determining the type from pointer type.
8810 (ipa_compute_jump_functions_for_edge): Update call of
8811 determine_known_aggregate_parts.
8812 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
8813 (gimple_get_virt_method_for_binfo): ... here; simplify using
8814 vtable_pointer_value_to_vtable.
8815 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
8816 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
8817 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
8818 (vtable_pointer_value_to_vtable): Break out from ...; handle also
8820 (vtable_pointer_value_to_binfo): ... here.
8821 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
8823 2014-02-03 Teresa Johnson <tejohnson@google.com>
8825 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
8826 redef of outer loop index variable.
8828 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
8832 * doc/extend.texi (Function Attributes): Typo.
8834 2014-02-03 Cong Hou <congh@google.com>
8836 PR tree-optimization/60000
8837 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
8838 if the vectorized statement is a store. A store statement can only
8839 appear at the end of pattern statements.
8841 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8843 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
8844 (ix86_option_override_internal): Default long double to 64-bit for
8845 32-bit Bionic and to 128-bit for 64-bit Bionic.
8847 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
8848 TARGET_LONG_DOUBLE_128 is true.
8849 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
8851 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
8852 (mlong-double-64): Negate -mlong-double-128.
8853 (mlong-double-128): New option.
8855 * config/i386/i386-c.c (ix86_target_macros): Define
8856 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
8858 * doc/invoke.texi: Document -mlong-double-128.
8860 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
8862 PR rtl-optimization/60024
8863 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
8865 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
8867 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
8869 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8871 PR rtl-optimization/57662
8872 * sel-sched.c (code_motion_path_driver): Do not mark already not
8873 existing blocks in the visiting bitmap.
8875 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
8877 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
8878 on the insn being emitted.
8880 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
8881 Will Deacon <will.deacon@arm.com>
8883 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
8885 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8887 * config/arm/arm-tables.opt: Regenerate.
8889 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8891 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
8892 for vector types other than V16QImode.
8893 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
8894 define_expand, and call altivec_expand_vec_perm_le when producing
8895 code with little endian element order.
8896 (*altivec_vperm_<mode>_internal): New insn having previous
8897 behavior of altivec_vperm_<mode>.
8898 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
8899 altivec_expand_vec_perm_le when producing code with little endian
8901 (*altivec_vperm_<mode>_uns_internal): New insn having previous
8902 behavior of altivec_vperm_<mode>_uns.
8904 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8906 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
8907 (altivec_vsumsws): Add handling for -maltivec=be with a little
8909 (altivec_vsumsws_direct): New.
8910 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
8911 gen_altivec_vsumsws.
8913 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8915 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
8916 vtable_pointer_value_to_binfo): New functions.
8917 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
8918 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
8920 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
8922 * config/nios2/nios2.md (load_got_register): Initialize GOT
8923 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
8924 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
8926 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
8928 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
8929 preserverd by passthrough, do not propagate the type.
8931 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8933 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
8934 (mips_atomic_assign_expand_fenv): New function.
8935 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
8937 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8939 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
8940 (__builtin_mips_set_fcsr): Likewise.
8941 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
8942 MIPS_USI_FTYPE_VOID.
8943 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
8944 (mips16_expand_set_fcsr): Likewise.
8945 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
8946 (mips16_set_fcsr_stub): Likewise.
8947 (mips16_get_fcsr_one_only_stub): New class.
8948 (mips16_set_fcsr_one_only_stub): Likewise.
8949 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
8950 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
8951 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
8952 (hard_float): New availability predicate.
8953 (mips_builtins): Add get_fcsr and set_fcsr.
8954 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
8955 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
8956 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
8957 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
8958 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
8961 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
8963 * config/mips/mips.c (mips_one_only_stub): New class.
8964 (mips_need_mips16_rdhwr_p): Replace with...
8965 (mips16_rdhwr_stub): ...this new variable.
8966 (mips16_stub_call_address): New function.
8967 (mips16_rdhwr_one_only_stub): New class.
8968 (mips_expand_thread_pointer): Use mips16_stub_call_address.
8969 (mips_output_mips16_rdhwr): Delete.
8970 (mips_finish_stub): New function.
8971 (mips_code_end): Use it to handle rdhwr stubs.
8973 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
8976 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
8977 when calculating size of integer atomic types.
8979 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
8981 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
8983 2014-02-01 Jakub Jelinek <jakub@redhat.com>
8985 PR tree-optimization/60003
8986 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
8987 * profile.c (branch_prob): Use gimple_call_builtin_p
8988 to check for BUILT_IN_SETJMP_RECEIVER.
8989 * tree-inline.c (copy_bb): Call notice_special_calls.
8991 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
8994 * lra-constraints.c (process_alt_operands): Update reload_sum only
8997 2014-01-31 Richard Henderson <rth@redhat.com>
9000 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
9001 until after else_eh is processed.
9003 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
9005 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
9006 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
9007 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
9008 in smmintrin.h, remove them.
9009 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
9010 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
9011 * config/i386/i386.md (ROUND_SAE): Fix value.
9012 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
9013 (const48_operand): New.
9014 * config/i386/subst.md (round), (round_expand): Use
9015 const_4_or_8_to_11_operand.
9016 (round_saeonly), (round_saeonly_expand): Use const48_operand.
9018 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
9020 * config/i386/constraints.md (Yk): Swap meaning with k.
9021 * config/i386/i386.md (movhi_internal): Change Yk to k.
9022 (movqi_internal): Ditto.
9023 (*k<logic><mode>): Ditto.
9026 (kandn<mode>): Ditto.
9027 (*<code>hi_1): Ditto.
9028 (*<code>qi_1): Ditto.
9029 (kxnor<mode>): Ditto.
9030 (kortestzhi): Ditto.
9031 (kortestchi): Ditto.
9033 (*one_cmplhi2_1): Ditto.
9034 (*one_cmplqi2_1): Ditto.
9035 * config/i386/sse.md (): Change k to Yk.
9036 (avx512f_load<mode>_mask): Ditto.
9037 (avx512f_blendm<mode>): Ditto.
9038 (avx512f_store<mode>_mask): Ditto.
9039 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
9040 (avx512f_storedqu<mode>_mask): Ditto.
9041 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
9043 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
9044 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
9045 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
9046 (avx512f_maskcmp<mode>3): Ditto.
9047 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
9048 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
9049 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
9050 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
9051 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
9052 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
9053 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
9054 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
9055 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
9056 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
9057 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
9058 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
9059 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
9060 (vec_extract_lo_<mode>_maskm): Ditto.
9061 (vec_extract_hi_<mode>_maskm): Ditto.
9062 (avx512f_vternlog<mode>_mask): Ditto.
9063 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
9064 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
9065 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
9066 (avx512f_<code>v8div16qi2_mask): Ditto.
9067 (avx512f_<code>v8div16qi2_mask_store): Ditto.
9068 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
9069 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
9070 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
9071 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
9072 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9073 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9074 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9075 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9076 (avx512cd_maskb_vec_dupv8di): Ditto.
9077 (avx512cd_maskw_vec_dupv16si): Ditto.
9078 (avx512f_vpermi2var<mode>3_maskz): Ditto.
9079 (avx512f_vpermi2var<mode>3_mask): Ditto.
9080 (avx512f_vpermi2var<mode>3_mask): Ditto.
9081 (avx512f_vpermt2var<mode>3_maskz): Ditto.
9082 (*avx512f_gathersi<mode>): Ditto.
9083 (*avx512f_gathersi<mode>_2): Ditto.
9084 (*avx512f_gatherdi<mode>): Ditto.
9085 (*avx512f_gatherdi<mode>_2): Ditto.
9086 (*avx512f_scattersi<mode>): Ditto.
9087 (*avx512f_scatterdi<mode>): Ditto.
9088 (avx512f_compress<mode>_mask): Ditto.
9089 (avx512f_compressstore<mode>_mask): Ditto.
9090 (avx512f_expand<mode>_mask): Ditto.
9091 * config/i386/subst.md (mask): Change k to Yk.
9092 (mask_scalar_merge): Ditto.
9095 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
9097 * doc/extend.texi (Vector Extensions): Document ?: in C++.
9099 2014-01-31 Richard Biener <rguenther@suse.de>
9102 * builtins.c (fold_builtin_memory_op): Make sure to not
9103 use a floating-point mode or a boolean or enumeral type for
9106 2014-01-30 DJ Delorie <dj@redhat.com>
9108 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
9109 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
9110 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
9111 whenever main() has an epilogue.
9113 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9115 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
9116 unused variable "field".
9117 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
9118 (vsx_mergeh_<mode>): Likewise.
9119 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
9120 (altivec_vmrghh): Likewise.
9121 (altivec_vmrghw): Likewise.
9122 (altivec_vmrglb): Likewise.
9123 (altivec_vmrglh): Likewise.
9124 (altivec_vmrglw): Likewise.
9125 (altivec_vspltb): Add missing uses.
9126 (altivec_vsplth): Likewise.
9127 (altivec_vspltw): Likewise.
9128 (altivec_vspltsf): Likewise.
9130 2014-01-30 Jakub Jelinek <jakub@redhat.com>
9133 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
9134 frame related instructions.
9136 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
9138 PR rtl-optimization/59959
9139 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
9140 any reload of register whose subreg is invalid.
9142 2014-01-30 Jakub Jelinek <jakub@redhat.com>
9144 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
9145 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
9146 Add missing return type - void.
9148 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9150 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
9151 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
9152 remove element index adjustment for endian (now handled in vsx.md
9154 (altivec_expand_vec_perm_const): Use
9155 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
9156 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
9157 (vsx_xxspltw_<mode>): Adjust element index for little endian.
9158 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
9159 define_expand and a new define_insn *altivec_vspltb_internal;
9160 adjust for -maltivec=be on a little endian target.
9161 (altivec_vspltb_direct): New.
9162 (altivec_vsplth): Divide into a define_expand and a new
9163 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
9164 little endian target.
9165 (altivec_vsplth_direct): New.
9166 (altivec_vspltw): Divide into a define_expand and a new
9167 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
9168 little endian target.
9169 (altivec_vspltw_direct): New.
9170 (altivec_vspltsf): Divide into a define_expand and a new
9171 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
9172 a little endian target.
9174 2014-01-30 Richard Biener <rguenther@suse.de>
9176 PR tree-optimization/59993
9177 * tree-ssa-forwprop.c (associate_pointerplus): Check we
9178 can propagate form the earlier stmt and avoid the transform
9179 when the intermediate result is needed.
9181 2014-01-30 Alangi Derick <alangiderick@gmail.com>
9183 * README.Portability: Fix typo.
9185 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
9187 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
9188 comparison_operator with ordered_comparison_operator.
9190 2014-01-30 Nick Clifton <nickc@redhat.com>
9192 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
9193 Rename to mn10300_store_multiple_regs.
9194 * config/mn10300/mn10300.c: Likewise.
9195 * config/mn10300/mn10300.md (store_movm): Fix typo: call
9196 store_multiple_regs.
9197 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
9198 Call mn10300_store_multiple_regs.
9200 2014-01-30 Nick Clifton <nickc@redhat.com>
9201 DJ Delorie <dj@redhat.com>
9203 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
9204 %fp 2 to keep registers after it properly word-aligned.
9205 (rl78_alloc_physical_registers_umul): Handle the case where both
9206 input operands are the same.
9208 2014-01-30 Richard Biener <rguenther@suse.de>
9210 PR tree-optimization/59903
9211 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
9214 2014-01-30 Jason Merrill <jason@redhat.com>
9217 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
9220 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
9222 2014-01-30 Richard Biener <rguenther@suse.de>
9224 PR tree-optimization/59951
9225 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
9227 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
9230 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
9231 SFmode to DFmode case.
9233 2014-01-29 DJ Delorie <dj@redhat.com>
9235 * config/msp430/msp430.opt (-minrt): New.
9236 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
9238 (ENDFILE_SPEC): Likewise.
9240 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
9242 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
9243 (estimate_function_body_sizes): Use it.
9245 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
9248 * dwarf2out.c (is_cxx_auto): New.
9249 (is_base_type): Use it.
9250 (gen_type_die_with_usage): Likewise.
9252 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9254 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
9255 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
9256 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
9257 -maltivec=be with LE targets.
9258 (vsx_mergeh_<mode>): Likewise.
9259 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
9260 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
9261 (altivec_vmrghb): Replace with define_expand and new
9262 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
9263 (altivec_vmrghb_direct): New define_insn.
9264 (altivec_vmrghh): Replace with define_expand and new
9265 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
9266 (altivec_vmrghh_direct): New define_insn.
9267 (altivec_vmrghw): Replace with define_expand and new
9268 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
9269 (altivec_vmrghw_direct): New define_insn.
9270 (*altivec_vmrghsf): Adjust for endianness.
9271 (altivec_vmrglb): Replace with define_expand and new
9272 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
9273 (altivec_vmrglb_direct): New define_insn.
9274 (altivec_vmrglh): Replace with define_expand and new
9275 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
9276 (altivec_vmrglh_direct): New define_insn.
9277 (altivec_vmrglw): Replace with define_expand and new
9278 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
9279 (altivec_vmrglw_direct): New define_insn.
9280 (*altivec_vmrglsf): Adjust for endianness.
9281 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
9282 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
9283 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
9284 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
9285 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
9286 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
9287 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
9288 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
9290 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
9292 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
9293 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
9296 2014-01-29 Richard Biener <rguenther@suse.de>
9298 PR tree-optimization/58742
9299 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
9300 associate_pointerplus_align.
9301 (associate_pointerplus_diff): New function.
9302 (associate_pointerplus): Likewise. Call associate_pointerplus_align
9303 and associate_pointerplus_diff.
9305 2014-01-29 Richard Biener <rguenther@suse.de>
9307 * lto-streamer.h (LTO_major_version): Bump to 3.
9308 (LTO_minor_version): Reset to 0.
9310 2014-01-29 Renlin Li <Renlin.Li@arm.com>
9312 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
9313 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
9314 (arm_file_start): Generate correct asm header for armv7ve.
9315 * config/arm/bpabi.h: Add multilib support for armv7ve.
9316 * config/arm/driver-arm.c: Change the architectures of cortex-a7
9317 and cortex-a15 to armv7ve.
9318 * config/arm/t-aprofile: Add multilib support for armv7ve.
9319 * doc/invoke.texi: Document -march=armv7ve.
9321 2014-01-29 Richard Biener <rguenther@suse.de>
9323 PR tree-optimization/58742
9324 * tree-ssa-forwprop.c (associate_plusminus): Return true
9325 if we changed sth, defer EH cleanup to ...
9326 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
9327 (simplify_mult): New function.
9329 2014-01-29 Jakub Jelinek <jakub@redhat.com>
9332 PR tree-optimization/59920
9333 * tree.c (build_common_builtin_nodes): Remove
9334 __builtin_setjmp_dispatcher initialization.
9335 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
9336 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
9337 instead of gsi_after_labels + manually skipping debug stmts.
9338 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
9339 ignore bbs with IFN_ABNORMAL_DISPATCHER.
9340 * tree-inline.c (copy_edges_for_bb): Remove
9341 can_make_abnormal_goto argument, instead add abnormal_goto_dest
9342 argument. Ignore computed_goto_p stmts. Don't call
9343 make_abnormal_goto_edges. If a call might need abnormal edges
9344 for non-local gotos, see if it already has an edge to
9345 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
9346 with true argument, don't do anything then, otherwise add
9347 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
9348 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
9350 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
9351 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
9352 (lower_stmt): Don't set data->calls_builtin_setjmp.
9353 (lower_builtin_setjmp): Adjust comment.
9354 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
9355 * tree-cfg.c (found_computed_goto): Remove.
9356 (factor_computed_gotos): Remove.
9357 (make_goto_expr_edges): Return bool, true for computed gotos.
9358 Don't call make_abnormal_goto_edges.
9359 (build_gimple_cfg): Don't set found_computed_goto, don't call
9360 factor_computed_gotos.
9361 (computed_goto_p): No longer static.
9362 (make_blocks): Don't set found_computed_goto.
9363 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
9364 (make_edges): If make_goto_expr_edges returns true, push bb
9365 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
9366 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
9367 vector. Record mapping between bbs and OpenMP regions if there
9368 are any, adjust make_gimple_omp_edges caller. Call
9369 handle_abnormal_edges.
9370 (make_abnormal_goto_edges): Remove.
9371 * tree-cfg.h (make_abnormal_goto_edges): Remove.
9372 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
9373 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
9374 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
9375 * internal-fn.def (ABNORMAL_DISPATCHER): New.
9376 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
9377 filling *region also set *region_idx to (*region)->entry->index.
9380 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
9381 For REGs set ORIGINAL_REGNO.
9383 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
9385 * doc/md.texi: Mention that a target shouldn't implement
9386 vec_widen_(s|u)mul_even/odd pair if it is less efficient
9389 2014-01-29 Jakub Jelinek <jakub@redhat.com>
9391 PR tree-optimization/59594
9392 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
9393 a copy of the datarefs vector rather than the vector itself.
9395 2014-01-28 Jason Merrill <jason@redhat.com>
9398 * dwarf2out.c (auto_die): New static.
9399 (gen_type_die_with_usage): Handle C++1y 'auto'.
9400 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
9403 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
9406 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
9407 (SPEC_X32): Likewise.
9408 (SPEC_64): Likewise.
9409 * config/i386/i386.c (ix86_option_override_internal): Turn off
9410 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
9412 (x86_file_start): Output .code16gcc for TARGET_16BIT.
9413 * config/i386/i386.h (TARGET_16BIT): New macro.
9414 (TARGET_16BIT_P): Likewise.
9415 * config/i386/i386.opt: Add m16.
9416 * doc/invoke.texi: Document -m16.
9418 2014-01-28 Jakub Jelinek <jakub@redhat.com>
9420 PR preprocessor/59935
9421 * input.c (location_get_source_line): Bail out on when line number
9422 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
9424 2014-01-28 Richard Biener <rguenther@suse.de>
9426 PR tree-optimization/58742
9427 * tree-ssa-forwprop.c (associate_plusminus): Handle
9428 pointer subtraction of the form (T)(P + A) - (T)P.
9430 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9432 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
9435 2014-01-28 Richard Biener <rguenther@suse.de>
9438 2014-01-28 Richard Biener <rguenther@suse.de>
9440 PR rtl-optimization/45364
9441 PR rtl-optimization/59890
9442 * var-tracking.c (local_get_addr_clear_given_value): Handle
9443 already cleared slot.
9444 (val_reset): Handle not allocated local_get_addr_cache.
9445 (vt_find_locations): Use post-order on the inverted CFG.
9447 2014-01-28 Richard Biener <rguenther@suse.de>
9449 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
9451 2014-01-28 Richard Biener <rguenther@suse.de>
9453 PR rtl-optimization/45364
9454 PR rtl-optimization/59890
9455 * var-tracking.c (local_get_addr_clear_given_value): Handle
9456 already cleared slot.
9457 (val_reset): Handle not allocated local_get_addr_cache.
9458 (vt_find_locations): Use post-order on the inverted CFG.
9460 2014-01-28 Alan Modra <amodra@gmail.com>
9462 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
9463 * configure.ac <recursive call for build != host>: Define
9464 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
9465 and LD_FOR_BUILD too.
9466 * configure: Regenerate.
9468 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
9470 * config/i386/i386.c (get_builtin_code_for_version): Separate
9471 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
9472 Broadwell from Haswell.
9474 2014-01-27 Steve Ellcey <sellcey@mips.com>
9476 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
9477 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
9478 * config/mips/mips.c (mips_option_override): Change setting
9480 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
9481 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
9482 Change from Mask to Var.
9484 2014-01-27 Jeff Law <law@redhat.com>
9486 * ipa-inline.c (inline_small_functions): Fix typo.
9488 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
9490 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
9491 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
9492 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
9493 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
9494 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
9495 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
9496 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
9497 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
9498 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
9499 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
9500 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
9501 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
9502 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
9503 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
9504 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
9505 (_mm512_storeu_epi64): Ditto.
9506 (_mm512_cmpge_epi32_mask): Ditto.
9507 (_mm512_cmpge_epu32_mask): Ditto.
9508 (_mm512_cmpge_epi64_mask): Ditto.
9509 (_mm512_cmpge_epu64_mask): Ditto.
9510 (_mm512_cmple_epi32_mask): Ditto.
9511 (_mm512_cmple_epu32_mask): Ditto.
9512 (_mm512_cmple_epi64_mask): Ditto.
9513 (_mm512_cmple_epu64_mask): Ditto.
9514 (_mm512_cmplt_epi32_mask): Ditto.
9515 (_mm512_cmplt_epu32_mask): Ditto.
9516 (_mm512_cmplt_epi64_mask): Ditto.
9517 (_mm512_cmplt_epu64_mask): Ditto.
9518 (_mm512_cmpneq_epi32_mask): Ditto.
9519 (_mm512_cmpneq_epu32_mask): Ditto.
9520 (_mm512_cmpneq_epi64_mask): Ditto.
9521 (_mm512_cmpneq_epu64_mask): Ditto.
9522 (_mm512_expand_pd): Ditto.
9523 (_mm512_expand_ps): Ditto.
9524 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
9525 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
9526 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
9527 * config/i386/i386.c (ix86_builtins): Add
9528 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
9529 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
9530 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
9531 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
9532 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
9533 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
9534 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
9535 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
9536 IX86_BUILTIN_PMOVUSQW512_MEM.
9537 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
9538 __builtin_ia32_pmovsqd512mem_mask,
9539 __builtin_ia32_pmovqd512mem_mask,
9540 __builtin_ia32_pmovusqw512mem_mask,
9541 __builtin_ia32_pmovsqw512mem_mask,
9542 __builtin_ia32_pmovqw512mem_mask,
9543 __builtin_ia32_pmovusdw512mem_mask,
9544 __builtin_ia32_pmovsdw512mem_mask,
9545 __builtin_ia32_pmovdw512mem_mask,
9546 __builtin_ia32_pmovqb512mem_mask,
9547 __builtin_ia32_pmovusqb512mem_mask,
9548 __builtin_ia32_pmovsqb512mem_mask,
9549 __builtin_ia32_pmovusdb512mem_mask,
9550 __builtin_ia32_pmovsdb512mem_mask,
9551 __builtin_ia32_pmovdb512mem_mask.
9552 (bdesc_args): Add __builtin_ia32_expanddf512,
9553 __builtin_ia32_expandsf512.
9554 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
9555 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
9556 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
9557 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
9558 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
9559 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
9560 (avx512f_<code>v8div16qi2_mask_store): This.
9561 (avx512f_expand<mode>): New.
9563 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
9565 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
9567 (_mm512_mask_prefetch_i64gather_pd): Ditto.
9568 (_mm512_prefetch_i32scatter_pd): Ditto.
9569 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
9570 (_mm512_prefetch_i64scatter_pd): Ditto.
9571 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
9572 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
9573 (_mm512_mask_prefetch_i64gather_ps): Ditto.
9574 (_mm512_prefetch_i32scatter_ps): Ditto.
9575 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
9576 (_mm512_prefetch_i64scatter_ps): Ditto.
9577 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
9578 * config/i386/i386-builtin-types.def: Define
9579 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
9580 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
9581 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
9582 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
9583 IX86_BUILTIN_SCATTERPFQPD.
9584 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
9585 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
9586 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
9587 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
9588 __builtin_ia32_scatterpfqps.
9589 (ix86_expand_builtin): Expand new built-ins.
9590 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
9591 fix memory access data type.
9592 (*avx512pf_gatherpf<mode>_mask): Ditto.
9593 (*avx512pf_gatherpf<mode>): Ditto.
9594 (avx512pf_scatterpf<mode>): Ditto.
9595 (*avx512pf_scatterpf<mode>_mask): Ditto.
9596 (*avx512pf_scatterpf<mode>): Ditto.
9597 (GATHER_SCATTER_SF_MEM_MODE): New.
9598 (avx512pf_gatherpf<mode>df): Ditto.
9599 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9600 (*avx512pf_scatterpf<mode>df): Ditto.
9602 2014-01-27 Jakub Jelinek <jakub@redhat.com>
9605 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
9606 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
9609 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
9611 * common/config/arm/arm-common.c
9612 (arm_rewrite_mcpu): Handle multiple names.
9614 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
9616 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
9618 * gimple-builder.h (create_gimple_tmp): Delete.
9620 2014-01-27 Christian Bruel <christian.bruel@st.com>
9622 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
9625 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
9627 * config/pa/pa.md (call): Generate indirect long calls to non-local
9628 functions when outputing 32-bit code.
9629 (call_value): Likewise except for special call to buggy powf function.
9631 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
9632 portable runtime and PIC indirect calls.
9633 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
9634 and PIC call sequences. Use ldo instead of blr to set return register
9635 in PIC call sequence.
9637 2014-01-25 Walter Lee <walt@tilera.com>
9639 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
9640 avoid clobbering a live register.
9642 2014-01-25 Walter Lee <walt@tilera.com>
9644 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
9645 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
9646 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
9647 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
9649 2014-01-25 Walter Lee <walt@tilera.com>
9651 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
9652 arguments on even registers.
9653 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
9655 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
9656 (BIGGEST_ALIGNMENT): Ditto.
9657 (BIGGEST_FIELD_ALIGNMENT): Ditto.
9659 2014-01-25 Walter Lee <walt@tilera.com>
9661 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
9662 insns before bundling.
9663 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
9665 2014-01-25 Walter Lee <walt@tilera.com>
9667 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
9668 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
9669 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
9671 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9673 * config/mips/constraints.md (kl): Delete.
9674 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
9675 define expands, using...
9676 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
9677 instructions for MIPS16.
9678 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
9679 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
9681 2014-01-25 Walter Lee <walt@tilera.com>
9683 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
9687 2014-01-25 Walter Lee <walt@tilera.com>
9689 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
9690 (TARGET_EXPAND_TO_RTL_HOOK): Define.
9692 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
9694 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
9697 2014-01-25 Jakub Jelinek <jakub@redhat.com>
9699 * print-rtl.c (in_call_function_usage): New var.
9700 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
9701 EXPR_LIST mode as mode and not as reg note name.
9704 * cfgloopmanip.c (copy_loop_info): If
9705 loop->warned_aggressive_loop_optimizations, make sure
9706 the flag is set in target loop too.
9708 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
9710 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
9712 * builtins.def: Likewise.
9713 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
9714 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
9715 * ira.c (ira_setup_eliminable_regset): Likewise.
9716 * omp-low.c (gate_expand_omp): Likewise.
9717 (execute_lower_omp): Likewise.
9718 (diagnose_sb_0): Likewise.
9719 (gate_diagnose_omp_blocks): Likewise.
9720 (simd_clone_clauses_extract): Likewise.
9723 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9725 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
9726 correction for little endian...
9727 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
9730 2014-01-24 Jeff Law <law@redhat.com>
9732 PR tree-optimization/59919
9733 * tree-vrp.c (find_assert_locations_1): Do not register asserts
9734 for non-returning calls.
9736 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
9738 * common/config/aarch64/aarch64-common.c
9739 (aarch64_rewrite_mcpu): Handle multiple names.
9740 * config/aarch64/aarch64.h
9741 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
9743 2014-01-24 Dodji Seketeli <dodji@redhat.com>
9745 * input.c (add_file_to_cache_tab): Handle the case where fopen
9748 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
9751 * config/i386/i386.md (pushsf splitter): Get stack adjustment
9752 from push operand if code of push isn't PRE_DEC.
9754 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
9757 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
9758 -mquad-memory-atomic. Update -mquad-memory documentation to say
9759 it is only used for non-atomic loads/stores.
9761 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
9762 -mquad-memory or -mquad-memory-atomic switches.
9764 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
9765 -mquad-memory-atomic to ISA 2.07 support.
9767 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
9768 to separate support of normal quad word memory operations (ldq, stq)
9769 from the atomic quad word memory operations.
9771 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
9772 support to separate non-atomic quad word operations from atomic
9773 quad word operations. Disable non-atomic quad word operations in
9774 little endian mode so that we don't have to swap words after the
9775 load and before the store.
9776 (quad_load_store_p): Add comment about atomic quad word support.
9777 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
9778 options printed with -mdebug=reg.
9780 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
9781 -mquad-memory-atomic as the test for whether we have quad word
9782 atomic instructions.
9783 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
9784 or -mp8-vector are used, allow byte/half-word atomic operations.
9786 * config/rs6000/sync.md (load_lockedti): Insure that the address
9787 is a proper indexed or indirect address for the lqarx instruction.
9788 On little endian systems, swap the hi/lo registers after the lqarx
9790 (load_lockedpti): Use indexed_or_indirect_operand predicate to
9791 insure the address is valid for the lqarx instruction.
9792 (store_conditionalti): Insure that the address is a proper indexed
9793 or indirect address for the stqcrx. instruction. On little endian
9794 systems, swap the hi/lo registers before doing the stqcrx.
9796 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
9797 insure the address is valid for the stqcrx. instruction.
9799 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
9800 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
9801 type of quad memory support is available.
9803 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
9806 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
9807 there is a danger of looping.
9809 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
9811 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
9812 force flag_ira_loop_pressure if set via command line.
9814 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9816 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
9817 (ashr_simd): New builtin handling DI mode.
9818 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
9819 (aarch64_sshr_simddi): New match pattern.
9820 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
9821 (vshrd_n_s64): Likewise.
9822 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
9824 2014-01-23 Nick Clifton <nickc@redhat.com>
9826 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
9827 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
9828 favour of mcu specific scripts.
9829 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
9832 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9833 Alex Velenko <Alex.Velenko@arm.com>
9835 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
9836 (vaddv_s16): Likewise.
9837 (vaddv_s32): Likewise.
9838 (vaddv_u8): Likewise.
9839 (vaddv_u16): Likewise.
9840 (vaddv_u32): Likewise.
9841 (vaddvq_s8): Likewise.
9842 (vaddvq_s16): Likewise.
9843 (vaddvq_s32): Likewise.
9844 (vaddvq_s64): Likewise.
9845 (vaddvq_u8): Likewise.
9846 (vaddvq_u16): Likewise.
9847 (vaddvq_u32): Likewise.
9848 (vaddvq_u64): Likewise.
9849 (vaddv_f32): Likewise.
9850 (vaddvq_f32): Likewise.
9851 (vaddvq_f64): Likewise.
9852 (vmaxv_f32): Likewise.
9853 (vmaxv_s8): Likewise.
9854 (vmaxv_s16): Likewise.
9855 (vmaxv_s32): Likewise.
9856 (vmaxv_u8): Likewise.
9857 (vmaxv_u16): Likewise.
9858 (vmaxv_u32): Likewise.
9859 (vmaxvq_f32): Likewise.
9860 (vmaxvq_f64): Likewise.
9861 (vmaxvq_s8): Likewise.
9862 (vmaxvq_s16): Likewise.
9863 (vmaxvq_s32): Likewise.
9864 (vmaxvq_u8): Likewise.
9865 (vmaxvq_u16): Likewise.
9866 (vmaxvq_u32): Likewise.
9867 (vmaxnmv_f32): Likewise.
9868 (vmaxnmvq_f32): Likewise.
9869 (vmaxnmvq_f64): Likewise.
9870 (vminv_f32): Likewise.
9871 (vminv_s8): Likewise.
9872 (vminv_s16): Likewise.
9873 (vminv_s32): Likewise.
9874 (vminv_u8): Likewise.
9875 (vminv_u16): Likewise.
9876 (vminv_u32): Likewise.
9877 (vminvq_f32): Likewise.
9878 (vminvq_f64): Likewise.
9879 (vminvq_s8): Likewise.
9880 (vminvq_s16): Likewise.
9881 (vminvq_s32): Likewise.
9882 (vminvq_u8): Likewise.
9883 (vminvq_u16): Likewise.
9884 (vminvq_u32): Likewise.
9885 (vminnmv_f32): Likewise.
9886 (vminnmvq_f32): Likewise.
9887 (vminnmvq_f64): Likewise.
9889 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
9891 * config/aarch64/aarch64-simd.md
9892 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
9893 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
9894 (*aarch64_mul3_elt<mode>): Likewise.
9895 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
9896 (*aarch64_mul3_elt_to_64v2df): Likewise.
9897 (*aarch64_mla_elt<mode>): Likewise.
9898 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
9899 (*aarch64_mls_elt<mode>): Likewise.
9900 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
9901 (*aarch64_fma4_elt<mode>): Likewise.
9902 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
9903 (*aarch64_fma4_elt_to_64v2df): Likewise.
9904 (*aarch64_fnma4_elt<mode>): Likewise.
9905 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
9906 (*aarch64_fnma4_elt_to_64v2df): Likewise.
9907 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
9908 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
9909 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9910 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
9911 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
9912 (aarch64_sqdmull_lane<mode>_internal): Likewise.
9913 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
9915 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
9917 * config/aarch64/aarch64-simd.md
9918 (aarch64_be_checked_get_lane<mode>): New define_expand.
9919 * config/aarch64/aarch64-simd-builtins.def
9920 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
9921 New builtin definition.
9922 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
9923 Use new safe be builtin.
9925 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
9927 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
9929 (aarch64_be_st1<mode>): Likewise.
9930 (aarch_ld1<VALL:mode>): Define_expand modified.
9931 (aarch_st1<VALL:mode>): Likewise.
9932 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
9933 (UNSPEC_ST1): Likewise.
9935 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
9937 * config/microblaze/microblaze.md: Add trap insn and attribute
9939 2014-01-23 Dodji Seketeli <dodji@redhat.com>
9941 PR preprocessor/58580
9942 * input.h (location_get_source_line): Take an additional line_size
9944 (void diagnostics_file_cache_fini): Declare new function.
9945 * input.c (struct fcache): New type.
9946 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
9947 New static constants.
9948 (diagnostic_file_cache_init, total_lines_num)
9949 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
9950 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
9951 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
9952 (get_next_line, read_next_line, goto_next_line, read_line_num):
9953 New static function definitions.
9954 (diagnostic_file_cache_fini): New function.
9955 (location_get_source_line): Take an additional output line_len
9956 parameter. Re-write using lookup_or_add_file_to_cache_tab and
9958 * diagnostic.c (diagnostic_finish): Call
9959 diagnostic_file_cache_fini.
9960 (adjust_line): Take an additional input parameter for the length
9961 of the line, rather than calculating it with strlen.
9962 (diagnostic_show_locus): Adjust the use of
9963 location_get_source_line and adjust_line with respect to their new
9964 signature. While displaying a line now, do not stop at the first
9965 null byte. Rather, display the zero byte as a space and keep
9966 going until we reach the size of the line.
9967 * Makefile.in: Add vec.o to OBJS-libcommon
9969 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9970 Ilya Tocar <ilya.tocar@intel.com>
9972 * config/i386/avx512fintrin.h (_mm512_kmov): New.
9973 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
9974 (__builtin_ia32_kmov16): Ditto.
9975 * config/i386/i386.md (UNSPEC_KMOV): New.
9978 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
9980 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
9981 (_mm512_storeu_si512): Ditto.
9983 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
9986 * rtl.h (get_referenced_operands): Declare.
9987 * recog.c (get_referenced_operands): New function.
9988 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
9989 operands have been referenced when recording LO_SUM references.
9991 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
9993 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
9995 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
9997 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
9998 Enable for generic and recent AMD targets.
10000 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
10002 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
10003 ARG_SIZE note when adjustment was eliminated.
10005 2014-01-22 Jeff Law <law@redhat.com>
10007 PR tree-optimization/59597
10008 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
10009 in file. Accept new argument REGISTERING and use it to modify
10010 dump output appropriately.
10011 (register_jump_thread): Corresponding changes.
10012 (mark_threaded_blocks): Reinstate code to cancel unprofitable
10013 thread paths involving joiner blocks. Add code to dump cancelled
10014 jump threading paths.
10016 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
10018 PR rtl-optimization/59477
10019 * lra-constraints.c (inherit_in_ebb): Process call for living hard
10020 regs. Update reloads_num and potential_reload_hard_regs for all insns.
10022 2014-01-22 Tom Tromey <tromey@redhat.com>
10024 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
10026 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
10028 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
10030 PR rtl-optimization/59896
10031 * lra-constraints.c (process_alt_operands): Check unused note for
10032 matched operands of insn with no output reloads.
10034 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
10036 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
10037 (mips_move_from_gpr_cost): Likewise.
10039 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
10041 PR rtl-optimization/59858
10042 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
10043 ira_class_hard_regs_num.
10044 (process_alt_operands): Increase reject for dying matched operand.
10046 2014-01-21 Jakub Jelinek <jakub@redhat.com>
10049 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
10050 smaller than size, perform several stores or loads and stores
10051 at dst + count - size to store or copy all of size bytes, rather
10052 than just last modesize bytes.
10054 2014-01-20 DJ Delorie <dj@redhat.com>
10056 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
10057 that CLOBBERs are REGs before propogating their values.
10059 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
10061 PR middle-end/59789
10062 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
10063 (cgraph_inline_failed_type): New function.
10064 * cgraph.h (DEFCIFCODE): Add type.
10065 (cgraph_inline_failed_type_t): New enum.
10066 (cgraph_inline_failed_type): New prototype.
10067 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
10068 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
10069 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
10070 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
10071 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
10072 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
10073 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
10074 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
10075 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
10076 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
10077 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
10078 OPTIMIZATION_MISMATCH.
10079 * tree-inline.c (expand_call_inline): Emit errors during
10080 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
10082 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
10085 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
10086 mode attribute in insn output.
10088 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
10090 * output.h (output_constant): Delete.
10091 * varasm.c (output_constant): Make private.
10093 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
10095 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
10097 2014-01-20 Jakub Jelinek <jakub@redhat.com>
10099 PR middle-end/59860
10100 * tree.h (fold_builtin_strcat): New prototype.
10101 * builtins.c (fold_builtin_strcat): No longer static. Add len
10102 argument, if non-NULL, don't call c_strlen. Optimize
10103 directly into __builtin_memcpy instead of __builtin_strcpy.
10104 (fold_builtin_2): Adjust fold_builtin_strcat caller.
10105 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
10107 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
10109 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
10110 for SImode_address_operand operands, having only a REG argument.
10112 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
10114 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
10115 loader name using mbig-endian.
10116 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
10118 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
10120 * doc/invoke.texi (-march): Clarify documentation for AArch64.
10121 (-mtune): Likewise.
10124 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
10126 * config/aarch64/aarch64-protos.h
10127 (aarch64_cannot_change_mode_class_ptr): Declare.
10128 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
10129 aarch64_cannot_change_mode_class_ptr): New.
10130 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
10131 backend hook aarch64_cannot_change_mode_class.
10133 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
10135 * common/config/aarch64/aarch64-common.c
10136 (aarch64_handle_option): Don't handle any option order logic here.
10137 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
10138 selected_cpu, warn on architecture version mismatch.
10139 (aarch64_override_options): Fix parsing order for option strings.
10141 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10142 Iain Sandoe <iain@codesourcery.com>
10145 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
10146 warning. Amend comment to reflect current functionality.
10148 2014-01-20 Richard Biener <rguenther@suse.de>
10150 PR middle-end/59860
10151 * builtins.c (fold_builtin_strcat): Remove case better handled
10152 by tree-ssa-strlen.c.
10154 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
10156 * config/aarch64/aarch64.opt
10157 (mcpu, march, mtune): Make case-insensitive.
10159 2014-01-20 Jakub Jelinek <jakub@redhat.com>
10162 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
10163 if operands[1] is a REG or ZERO_EXTEND of a REG.
10165 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
10167 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
10169 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
10171 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
10172 long non-pic millicode calls.
10174 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10176 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
10178 2014-01-19 Kito Cheng <kito@0xlab.org>
10180 * builtins.c (expand_movstr): Check movstr expand done or fail.
10182 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10183 H.J. Lu <hongjiu.lu@intel.com>
10186 * config/i386/i386.md (*lea<mode>): Zero-extend return register
10187 to DImode for zero-extended addresses.
10189 2014-01-19 Jakub Jelinek <jakub@redhat.com>
10191 PR rtl-optimization/57763
10192 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
10193 on the new indirect jump_insn and increment LABEL_NUSES (label).
10195 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
10199 * config.gcc (x86_archs): New variable.
10200 (x86_64_archs): Likewise.
10201 (x86_cpus): Likewise.
10202 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
10203 --with-arch/--with-cpu= options.
10204 Support --with-arch=/--with-cpu={nehalem,westmere,
10205 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
10207 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10209 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
10210 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
10212 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
10214 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
10216 2014-01-18 Jakub Jelinek <jakub@redhat.com>
10219 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
10220 clear cpp_get_options (parse_in)->warn_unused_macros for
10221 ix86_target_macros_internal with cpp_define.
10223 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
10225 * jump.c (delete_related_insns): Keep (use (insn))s.
10226 * reorg.c (redundant_insn): Check for barriers too.
10228 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10230 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
10232 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
10234 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
10235 call to $$dyncall when TARGET_LONG_CALLS is true.
10237 2014-01-17 Jeff Law <law@redhat.com>
10239 * ree.c (combine_set_extension): Temporarily disable test for
10240 changing number of hard registers.
10242 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
10244 PR middle-end/58125
10245 * ipa-inline-analysis.c (inline_free_summary):
10246 Do not free summary of aliases.
10248 2014-01-17 Jakub Jelinek <jakub@redhat.com>
10250 PR middle-end/59706
10251 * gimplify.c (gimplify_expr): Use create_tmp_var
10252 instead of create_tmp_var_raw. If cond doesn't have
10253 integral type, don't add the IFN_ANNOTATE builtin at all.
10255 2014-01-17 Martin Jambor <mjambor@suse.cz>
10258 * ipa-cp.c (prev_edge_clone): New variable.
10259 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
10260 Also resize prev_edge_clone vector.
10261 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
10262 (ipcp_edge_removal_hook): New function.
10263 (ipcp_driver): Register ipcp_edge_removal_hook.
10265 2014-01-17 Andrew Pinski <apinski@cavium.com>
10266 Steve Ellcey <sellcey@mips.com>
10269 * config/mips/mips.c (mips_print_operand): Check operand mode instead
10272 2014-01-17 Jeff Law <law@redhat.com>
10274 PR middle-end/57904
10275 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
10276 so that pass_ccp runs first.
10278 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10280 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
10281 (ix86_adjust_cost): Use !TARGET_XXX.
10282 (do_reorder_for_imul): Likewise.
10283 (swap_top_of_ready_list): Likewise.
10284 (ix86_sched_reorder): Likewise.
10286 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
10288 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
10289 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
10290 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
10291 (intel_memset): New. Duplicate slm_memset.
10292 (intel_cost): New. Duplicate slm_cost.
10293 (m_INTEL): New macro.
10294 (processor_target_table): Add "intel".
10295 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
10296 with PROCESSOR_INTEL for "intel".
10297 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
10298 PROCESSOR_SILVERMONT.
10299 (ix86_issue_rate): Likewise.
10300 (ix86_adjust_cost): Likewise.
10301 (ia32_multipass_dfa_lookahead): Likewise.
10302 (swap_top_of_ready_list): Likewise.
10303 (ix86_sched_reorder): Likewise.
10304 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
10305 instead of TARGET_OPT_AGU.
10306 * config/i386/i386.h (TARGET_INTEL): New.
10307 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
10308 (processor_type): Add PROCESSOR_INTEL.
10309 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
10310 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
10312 2014-01-17 Marek Polacek <polacek@redhat.com>
10315 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
10318 2014-01-17 Richard Biener <rguenther@suse.de>
10320 PR tree-optimization/46590
10321 * opts.c (default_options_table): Add entries for
10322 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
10323 all enabled at -O1 but not for -Og.
10324 * common.opt (fbranch-count-reg): Remove Init(1).
10325 (fmove-loop-invariants): Likewise.
10326 (ftree-pta): Likewise.
10328 2014-01-17 Jakub Jelinek <jakub@redhat.com>
10330 * config/i386/i386.c (ix86_data_alignment): For compatibility with
10331 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
10332 decls to at least the GCC 4.8 used alignments.
10335 * tree-nested.c (convert_nonlocal_reference_stmt,
10336 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
10337 of GIMPLE_BIND stmts, adjust associated decls.
10339 2014-01-17 Richard Biener <rguenther@suse.de>
10341 PR tree-optimization/46590
10342 * vec.h (vec<>::bseach): New member function implementing
10343 binary search according to C89 bsearch.
10344 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
10345 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
10346 bitmap pointer again. Make accesses_in_loop a flat array.
10347 (mem_ref_obstack): New global.
10348 (outermost_indep_loop): Adjust for mem_ref->stored changes.
10349 (mark_ref_stored): Likewise.
10350 (ref_indep_loop_p_2): Likewise.
10351 (set_ref_stored_in_loop): New helper function.
10352 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
10353 (memref_free): Adjust.
10354 (record_mem_ref_loc): Simplify.
10355 (gather_mem_refs_stmt): Adjust.
10356 (sort_locs_in_loop_postorder_cmp): New function.
10357 (analyze_memory_references): Sort accesses_in_loop after
10358 loop postorder number.
10359 (find_ref_loc_in_loop_cmp): New function.
10360 (for_all_locs_in_loop): Find relevant cluster of locs in
10361 accesses_in_loop and iterate without recursion.
10362 (execute_sm): Avoid uninit warning.
10363 (struct ref_always_accessed): Simplify.
10364 (ref_always_accessed::operator ()): Likewise.
10365 (ref_always_accessed_p): Likewise.
10366 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
10367 loop postorder numbers here.
10368 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
10371 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
10374 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
10375 on decls for which assemble_alias has been called.
10377 2014-01-17 Nick Clifton <nickc@redhat.com>
10379 * config/msp430/msp430.opt: (mcpu): New option.
10380 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
10381 (msp430_option_override): Parse target_cpu. If the MCU name
10382 matches a generic string, clear target_mcu.
10383 (msp430_attr): Allow numeric interrupt values up to 63.
10384 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
10385 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
10387 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
10389 * config/msp430/msp430.md (popm): Use %J rather than %I.
10390 (addsi3): Use msp430_nonimmediate_operand for operand 2.
10391 (addhi_cy_i): Use immediate_operand for operand 2.
10392 * doc/invoke.texi: Document -mcpu option.
10394 2014-01-17 Richard Biener <rguenther@suse.de>
10396 PR rtl-optimization/38518
10397 * df.h (df_analyze_loop): Declare.
10398 * df-core.c: Include cfgloop.h.
10399 (df_analyze_1): Split out main part of df_analyze.
10400 (df_analyze): Adjust.
10401 (loop_inverted_post_order_compute): New function.
10402 (loop_post_order_compute): Likewise.
10403 (df_analyze_loop): New function avoiding whole-function
10404 postorder computes.
10405 * loop-invariant.c (find_defs): Use df_analyze_loop.
10406 (find_invariants): Adjust.
10407 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
10409 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
10411 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
10412 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
10414 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
10416 * ipa-ref.c (ipa_remove_stmt_references): Fix references
10417 traversal when removing references.
10419 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
10422 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
10424 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
10426 PR middle-end/56791
10427 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
10428 pushing a reload for an autoinc when we had previously reloaded an
10429 inner part of the address.
10431 2014-01-16 Jakub Jelinek <jakub@redhat.com>
10433 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
10435 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
10436 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
10437 when not giving up or versioning for alias only because of
10439 (vect_analyze_data_ref_dependences): Set to true.
10440 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
10442 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
10443 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
10446 PR middle-end/58344
10447 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
10450 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
10451 operand 0 predicate for gathers, use a new pseudo as subtarget.
10453 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
10455 PR middle-end/59609
10456 * lra-constraints.c (process_alt_operands): Add printing debug info.
10457 Check absence of input/output reloads for matched operands too.
10459 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
10461 PR rtl-optimization/59835
10462 * ira.c (ira_init_register_move_cost): Increase cost for
10465 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
10467 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
10469 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
10472 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
10473 non-register objects. Use gen_(high/low)part more consistently.
10476 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
10479 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
10480 endian support, remove tests for WORDS_BIG_ENDIAN.
10481 (p8_mfvsrd_3_<mode>): Likewise.
10482 (reload_gpr_from_vsx<mode>): Likewise.
10483 (reload_gpr_from_vsxsf): Likewise.
10484 (p8_mfvsrd_4_disf): Likewise.
10486 2014-01-16 Richard Biener <rguenther@suse.de>
10488 PR rtl-optimization/46590
10489 * lcm.c (compute_antinout_edge): Use postorder iteration.
10490 (compute_laterin): Use inverted postorder iteration.
10492 2014-01-16 Nick Clifton <nickc@redhat.com>
10494 PR middle-end/28865
10495 * varasm.c (output_constant): Return the number of bytes actually
10497 (output_constructor_array_range): Update the field size with the
10498 number of bytes emitted by output_constant.
10499 (output_constructor_regular_field): Likewise. Also do not
10500 complain if the total number of bytes emitted is now greater
10501 than the expected fieldpos.
10502 * output.h (output_constant): Update prototype and descriptive comment.
10504 2014-01-16 Marek Polacek <polacek@redhat.com>
10506 PR middle-end/59827
10507 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
10508 it is error_mark_node.
10510 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
10512 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
10513 VALID_AVX256_REG_OR_OI_MODE.
10515 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
10517 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
10518 current procedure should be profiled.
10520 2014-01-15 Andrew Pinski <apinski@cavium.com>
10522 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
10523 of moving from/to the STACK_REG register class.
10525 2014-01-15 Richard Henderson <rth@redhat.com>
10528 * reginfo.c (global_regs_decl): Globalize.
10529 * rtl.h (global_regs_decl): Declare.
10530 * ira.c (do_reload): Diagnose frame_pointer_needed and it
10531 reserved via global_regs.
10533 2014-01-15 Teresa Johnson <tejohnson@google.com>
10535 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
10537 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10539 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
10540 and vmulosh rather than call gen_vec_widen_smult_*.
10541 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
10542 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
10543 (vec_widen_smult_even_v16qi): Likewise.
10544 (vec_widen_umult_even_v8hi): Likewise.
10545 (vec_widen_smult_even_v8hi): Likewise.
10546 (vec_widen_umult_odd_v16qi): Likewise.
10547 (vec_widen_smult_odd_v16qi): Likewise.
10548 (vec_widen_umult_odd_v8hi): Likewise.
10549 (vec_widen_smult_odd_v8hi): Likewise.
10550 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
10551 vmuloub rather than call gen_vec_widen_umult_*.
10552 (vec_widen_umult_lo_v16qi): Likewise.
10553 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
10554 vmulosb rather than call gen_vec_widen_smult_*.
10555 (vec_widen_smult_lo_v16qi): Likewise.
10556 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
10557 rather than call gen_vec_widen_umult_*.
10558 (vec_widen_umult_lo_v8hi): Likewise.
10559 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
10560 rather than call gen_vec_widen_smult_*.
10561 (vec_widen_smult_lo_v8hi): Likewise.
10563 2014-01-15 Jeff Law <law@redhat.com>
10565 PR tree-optimization/59747
10566 * ree.c (find_and_remove_re): Properly handle case where a second
10567 eliminated extension requires widening a copy created for elimination
10568 of a prior extension.
10569 (combine_set_extension): Ensure that the number of hard regs needed
10570 for a destination register does not change when we widen it.
10572 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
10574 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
10575 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
10576 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
10577 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
10578 (avr-*-rtems*): Likewise.
10579 (bfin*-rtems*): Likewise.
10580 (moxie-*-rtems*): Likewise.
10581 (h8300-*-rtems*): Likewise.
10582 (i[34567]86-*-rtems*): Likewise.
10583 (lm32-*-rtems*): Likewise.
10584 (m32r-*-rtems*): Likewise.
10585 (m68k-*-rtems*): Likewise.
10586 (microblaze*-*-rtems*): Likewise.
10587 (mips*-*-rtems*): Likewise.
10588 (powerpc-*-rtems*): Likewise.
10589 (sh-*-rtems*): Likewise.
10590 (sparc-*-rtems*): Likewise.
10591 (sparc64-*-rtems*): Likewise.
10592 (v850-*-rtems*): Likewise.
10593 (m32c-*-rtems*): Likewise.
10595 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
10597 PR rtl-optimization/59511
10598 * ira.c (ira_init_register_move_cost): Use memory costs for some
10599 cases of register move cost calculations.
10600 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
10601 instead of BB frequency.
10602 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
10603 * lra-assigns.c (find_hard_regno_for): Ditto.
10605 2014-01-15 Richard Biener <rguenther@suse.de>
10607 PR tree-optimization/59822
10608 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
10609 (vectorizable_load): Use it to hoist defs of uses of invariant
10610 loads out of the loop.
10612 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
10613 Kugan Vivekanandarajah <kuganv@linaro.org>
10616 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
10619 2014-01-15 Richard Biener <rguenther@suse.de>
10621 PR rtl-optimization/59802
10622 * lcm.c (compute_available): Use inverted postorder to seed
10623 the initial worklist.
10625 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10628 * config/s390/s390.c (s390_preferred_reload_class): Don't return
10629 ADDR_REGS for invalid symrefs in non-PIC code.
10631 2014-01-15 Jakub Jelinek <jakub@redhat.com>
10634 * builtins.c (determine_block_size): Initialize *probable_max_size
10635 even if len_rtx is CONST_INT.
10637 2014-01-14 Andrew Pinski <apinski@cavium.com>
10639 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
10640 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
10641 (cortexa53_tunings): Likewise.
10642 (aarch64_sched_issue_rate): New function.
10643 (TARGET_SCHED_ISSUE_RATE): Define.
10645 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10647 * ira-costs.c (find_costs_and_classes): Add missed
10648 ira_init_register_move_cost_if_necessary.
10650 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
10653 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
10655 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
10658 * config/i386/i386.c (type_natural_mode): Add a bool parameter
10659 to indicate if type is used for function return value. Warn ABI
10660 change if the vector mode isn't available for function return value.
10661 (ix86_function_arg_advance): Pass false to type_natural_mode.
10662 (ix86_function_arg): Likewise.
10663 (ix86_gimplify_va_arg): Likewise.
10664 (function_arg_32): Don't warn ABI change.
10665 (ix86_function_value): Pass true to type_natural_mode.
10666 (ix86_return_in_memory): Likewise.
10667 (ix86_struct_value_rtx): Removed.
10668 (TARGET_STRUCT_VALUE_RTX): Likewise.
10670 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10672 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
10673 converting a conditional jump into a conditional return.
10675 2014-01-14 Richard Biener <rguenther@suse.de>
10677 PR tree-optimization/58921
10678 PR tree-optimization/59006
10679 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
10680 hoisting invariant stmts.
10681 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
10682 invariant loads on the preheader edge if possible.
10684 2014-01-14 Joey Ye <joey.ye@arm.com>
10686 * doc/plugin.texi (Building GCC plugins): Update to C++.
10688 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
10690 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
10691 (_mm_rcp28_round_ss): Ditto.
10692 (_mm_rsqrt28_round_sd): Ditto.
10693 (_mm_rsqrt28_round_ss): Ditto.
10694 (_mm_rcp28_sd): Ditto.
10695 (_mm_rcp28_ss): Ditto.
10696 (_mm_rsqrt28_sd): Ditto.
10697 (_mm_rsqrt28_ss): Ditto.
10698 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
10699 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
10700 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
10701 (IX86_BUILTIN_RCP28SD): Ditto.
10702 (IX86_BUILTIN_RCP28SS): Ditto.
10703 (IX86_BUILTIN_RSQRT28SD): Ditto.
10704 (IX86_BUILTIN_RSQRT28SS): Ditto.
10705 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
10706 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
10707 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
10708 (ix86_expand_special_args_builtin): Expand new FTYPE.
10709 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
10710 (srcp14<mode>): Make insn unary.
10711 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
10712 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
10713 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
10714 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
10715 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
10716 Fix rounding: make it SAE only.
10717 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
10719 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
10721 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
10722 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
10723 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
10724 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
10725 (round_saeonly_mask_scalar_operand4): Ditto.
10726 (round_saeonly_mask_scalar_op3): Ditto.
10727 (round_saeonly_mask_scalar_op4): Ditto.
10729 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10731 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
10732 Implement -maltivec=be for vec_insert and vec_extract.
10734 2014-01-10 DJ Delorie <dj@redhat.com>
10736 * config/msp430/msp430.md (call_internal): Don't allow memory
10737 references with SP as the base register.
10738 (call_value_internal): Likewise.
10739 * config/msp430/constraints.md (Yc): New. For memory references
10740 that don't use SP as a base register.
10742 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
10743 "an integer without a # prefix"
10744 * config/msp430/msp430.md (epilogue_helper): Use it.
10746 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10749 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
10750 AVX512F gather builtins.
10751 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
10752 on gather decls with INTEGER_TYPE masktype.
10753 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
10754 directly into the builtin rather than hoisting it before loop.
10756 PR tree-optimization/59387
10757 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
10758 (scev_const_prop): If folded_casts and type has undefined overflow,
10759 use force_gimple_operand instead of force_gimple_operand_gsi and
10760 for each added stmt if it is assign with
10761 arith_code_with_undefined_signed_overflow, call
10762 rewrite_to_defined_overflow.
10763 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
10764 gimple-fold.h instead.
10765 (arith_code_with_undefined_signed_overflow,
10766 rewrite_to_defined_overflow): Moved to ...
10767 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
10768 rewrite_to_defined_overflow): ... here. No longer static.
10769 Include gimplify-me.h.
10770 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
10771 rewrite_to_defined_overflow): New prototypes.
10773 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10775 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
10777 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
10779 * builtins.c (get_object_alignment_2): Minor tweak.
10780 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
10782 2014-01-13 Christian Bruel <christian.bruel@st.com>
10784 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
10785 optimized non constant lengths.
10787 2014-01-13 Jakub Jelinek <jakub@redhat.com>
10790 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
10791 load as __atomic_load_N if possible.
10793 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10795 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
10797 (rs6000_expand_builtin): Adjust call.
10799 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
10802 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
10803 * config/rs6000/rs6000.c: Include target-globals.h.
10804 (rs6000_set_current_function): Instead of doing target_reinit
10805 unconditionally, use save_target_globals_default_opts and
10806 restore_target_globals.
10808 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
10810 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
10811 (rs6000_expand_builtin): Handle mffs and mtfsf.
10812 (rs6000_init_builtins): Define mffs and mtfsf.
10813 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
10814 (rs6000_mffs): New pattern.
10815 (rs6000_mtfsf): New pattern.
10817 2014-01-11 Bin Cheng <bin.cheng@arm.com>
10819 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
10820 Start narrowing with START. Apply candidate-use pair
10821 and check overall cost in narrowing.
10822 (iv_ca_prune): Pass new argument.
10824 2014-01-10 Jeff Law <law@redhat.com>
10826 PR middle-end/59743
10827 * ree.c (combine_reaching_defs): Ensure the defining statement
10828 occurs before the extension when optimizing extensions with
10829 different source and destination hard registers.
10831 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10834 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
10835 vtables into the type inheritance graph.
10837 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10839 PR rtl-optimization/59754
10840 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
10841 modes in the REGNO != REGNO case.
10843 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10845 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
10847 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10849 PR tree-optimization/59745
10850 * tree-predcom.c (tree_predictive_commoning_loop): Call
10851 free_affine_expand_cache if giving up because components is NULL.
10853 * target-globals.c (save_target_globals): Allocate < 4KB structs using
10854 GC in payload of target_globals struct instead of allocating them on
10855 the heap and the larger structs separately using GC.
10856 * target-globals.h (struct target_globals): Make regs, hard_regs,
10857 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
10858 of GTY((skip)) and change type to void *.
10859 (reset_target_globals): Cast loads from those fields to corresponding
10862 2014-01-10 Steve Ellcey <sellcey@mips.com>
10865 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
10866 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
10867 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
10869 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
10872 * aarch64-modes.def (CC_Zmode): New flags mode.
10873 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
10874 represents an equality.
10875 (aarch64_get_condition_code): Handle CC_Zmode.
10876 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
10878 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10880 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
10881 extraction in good case.
10883 2014-01-10 Richard Biener <rguenther@suse.de>
10885 PR tree-optimization/59374
10886 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
10887 checking after SLP discovery. Mark stmts not participating
10888 in any SLP instance properly.
10890 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10892 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
10893 when handling a SET rtx.
10895 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10897 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
10898 (cortex-a57): Likewise.
10899 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
10901 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10903 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
10904 non-iwmmxt builtins.
10906 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
10910 * ipa-devirt.c record_target_from_binfo): Take as argument
10911 stack of binfos and lookup matching one for virtual inheritance.
10912 (possible_polymorphic_call_targets_1): Update.
10914 2014-01-10 Huacai Chen <chenhc@lemote.com>
10916 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
10917 kernel strings for Loongson-2E/2F/3A.
10919 2014-01-10 Jakub Jelinek <jakub@redhat.com>
10921 PR middle-end/59670
10922 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
10923 is_gimple_call before calling gimple_call_internal_p.
10925 2014-01-09 Steve Ellcey <sellcey@mips.com>
10927 * Makefile.in (TREE_FLOW_H): Remove.
10928 (TREE_SSA_H): Add file names from tree-flow.h.
10929 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
10930 * tree.h: Remove tree-flow.h reference.
10931 * hash-table.h: Remove tree-flow.h reference.
10932 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
10933 reference with tree-ssa-loop.h.
10935 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10937 * doc/invoke.texi: Add -maltivec={be,le} options, and document
10938 default element-order behavior for -maltivec.
10939 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
10940 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
10941 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
10942 when targeting big endian, at least for now.
10943 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
10945 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10947 PR middle-end/47735
10948 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
10949 var satisfies use_register_for_decl, just take into account type
10950 alignment, rather than decl alignment.
10952 PR tree-optimization/59622
10953 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
10954 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
10955 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
10956 Don't devirtualize for inplace at all. For targets.length () == 1,
10957 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
10959 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10961 * config/i386/i386.md (cpu): Remove the unused btver1.
10963 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
10965 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
10967 2014-01-09 Jakub Jelinek <jakub@redhat.com>
10970 * tree-core.h (struct target_globals): New forward declaration.
10971 (struct tree_target_option): Add globals field.
10972 * tree.h (TREE_TARGET_GLOBALS): Define.
10973 (prepare_target_option_nodes_for_pch): New prototype.
10974 * target-globals.h (struct target_globals): Define even if
10975 !SWITCHABLE_TARGET.
10976 * tree.c (prepare_target_option_node_for_pch,
10977 prepare_target_option_nodes_for_pch): New functions.
10978 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
10979 * config/i386/i386.c: Include target-globals.h.
10980 (ix86_set_current_function): Instead of doing target_reinit
10981 unconditionally, use save_target_globals_default_opts and
10982 restore_target_globals.
10984 2014-01-09 Richard Biener <rguenther@suse.de>
10986 PR tree-optimization/59715
10987 * tree-cfg.h (split_critical_edges): Declare.
10988 * tree-cfg.c (split_critical_edges): Export.
10989 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
10991 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
10993 * cfgexpand.c (expand_stack_vars): Optionally disable
10994 asan stack protection.
10995 (expand_used_vars): Likewise.
10996 (partition_stack_vars): Likewise.
10997 * asan.c (asan_emit_stack_protection): Optionally disable
10998 after return stack usage.
10999 (instrument_derefs): Optionally disable memory access instrumentation.
11000 (instrument_builtin_call): Likewise.
11001 (instrument_strlen_call): Likewise.
11002 (asan_protect_global): Optionally disable global variables protection.
11003 * doc/invoke.texi: Added doc for new options.
11004 * params.def: Added new options.
11005 * params.h: Likewise.
11007 2014-01-09 Jakub Jelinek <jakub@redhat.com>
11009 PR rtl-optimization/59724
11010 * ifcvt.c (cond_exec_process_if_block): Don't call
11011 flow_find_head_matching_sequence with 0 longest_match.
11012 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
11013 non-active insns if !stop_after.
11014 (try_head_merge_bb): Revert 2014-01-07 changes.
11016 2014-01-08 Jeff Law <law@redhat.com>
11018 * ree.c (get_sub_rtx): New function, extracted from...
11019 (merge_def_and_ext): Here.
11020 (combine_reaching_defs): Use get_sub_rtx.
11022 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
11024 * cgraph.h (varpool_variable_node): Do not choke on null node.
11026 2014-01-08 Catherine Moore <clm@codesourcery.com>
11028 * config/mips/mips.md (simple_return): Attempt to use JRC
11030 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
11032 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
11034 PR rtl-optimization/59137
11035 * reorg.c (steal_delay_list_from_target): Call update_block for
11037 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
11039 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11041 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
11042 two duplicate entries.
11044 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
11047 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
11049 * config/mips/mips.c (mips_truncated_op_cost): New function.
11050 (mips_rtx_costs): Adjust test for BADDU.
11051 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
11053 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
11055 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
11056 (*baddu_si): ...this new pattern.
11058 2014-01-08 Jakub Jelinek <jakub@redhat.com>
11061 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
11063 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
11065 PR middle-end/57748
11066 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
11068 (expand_expr, expand_normal): Adjust.
11069 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
11070 inner_reference_p. Use inner_reference_p to expand inner references.
11071 (store_expr): Adjust.
11072 * cfgexpand.c (expand_call_stmt): Adjust.
11074 2014-01-08 Rong Xu <xur@google.com>
11076 * gcov-io.c (gcov_var): Move from gcov-io.h.
11077 (gcov_position): Ditto.
11078 (gcov_is_error): Ditto.
11079 (gcov_rewrite): Ditto.
11080 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
11081 only part to libgcc/libgcov.h.
11083 2014-01-08 Marek Polacek <polacek@redhat.com>
11085 PR middle-end/59669
11086 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
11088 2014-01-08 Marek Polacek <polacek@redhat.com>
11091 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
11093 2014-01-08 Jakub Jelinek <jakub@redhat.com>
11095 PR rtl-optimization/59649
11096 * stor-layout.c (get_mode_bounds): For BImode return
11097 0 and STORE_FLAG_VALUE.
11099 2014-01-08 Richard Biener <rguenther@suse.de>
11101 PR middle-end/59630
11102 * gimple.h (is_gimple_builtin_call): Remove.
11103 (gimple_builtin_call_types_compatible_p): New.
11104 (gimple_call_builtin_p): New overload.
11105 * gimple.c (is_gimple_builtin_call): Remove.
11106 (validate_call): Rename to ...
11107 (gimple_builtin_call_types_compatible_p): ... this and export. Also
11108 check return types.
11109 (validate_type): New static function.
11110 (gimple_call_builtin_p): New overload and adjust.
11111 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
11112 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
11113 (gimple_fold_stmt_to_constant_1): Likewise.
11114 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
11116 2014-01-08 Richard Biener <rguenther@suse.de>
11118 PR middle-end/59471
11119 * gimplify.c (gimplify_expr): Gimplify register-register type
11120 VIEW_CONVERT_EXPRs to separate stmts.
11122 2014-01-07 Jeff Law <law@redhat.com>
11124 PR middle-end/53623
11125 * ree.c (combine_set_extension): Handle case where source
11126 and destination registers in an extension insn are different.
11127 (combine_reaching_defs): Allow source and destination registers
11128 in extension to be different under limited circumstances.
11129 (add_removable_extension): Remove restriction that the
11130 source and destination registers in the extension are the same.
11131 (find_and_remove_re): Emit a copy from the extension's
11132 destination to its source after the defining insn if
11133 the source and destination registers are different.
11135 PR middle-end/59285
11136 * ifcvt.c (merge_if_block): If we are merging a block with more than
11137 one successor with a block with no successors, remove any BARRIER
11138 after the second block.
11140 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
11142 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
11144 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
11147 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
11148 for 14-bit register offsets when INT14_OK_STRICT is false.
11150 2014-01-07 Roland Stigge <stigge@antcom.de>
11151 Michael Meissner <meissner@linux.vnet.ibm.com>
11154 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
11155 Only check TFmode for SPE constants. Don't check TImode or TDmode.
11157 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
11159 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
11162 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
11164 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
11165 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
11166 rtx is const0_rtx or not.
11168 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
11171 * target-globals.c (save_target_globals): Remove this_fn_optab
11173 * toplev.c: Include optabs.h.
11174 (target_reinit): Temporarily restore the global options if another
11175 set of options are in force.
11177 2014-01-07 Jakub Jelinek <jakub@redhat.com>
11179 PR rtl-optimization/58668
11180 * cfgcleanup.c (flow_find_cross_jump): Don't count
11181 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
11182 to determine what is counted.
11183 (flow_find_head_matching_sequence): Use active_insn_p to determine
11185 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
11187 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
11188 determine what is counted.
11190 PR tree-optimization/59643
11191 * tree-predcom.c (split_data_refs_to_components): If one dr is
11192 read and one write, determine_offset fails and the write isn't
11193 in the bad component, just put the read into the bad component.
11195 2014-01-07 Mike Stump <mikestump@comcast.net>
11196 Jakub Jelinek <jakub@redhat.com>
11199 * tree-core.h (struct tree_optimization_option): Change optabs
11200 type from unsigned char * to void *.
11201 * optabs.c (init_tree_optimization_optabs): Adjust
11202 TREE_OPTIMIZATION_OPTABS initialization.
11204 2014-01-06 Jakub Jelinek <jakub@redhat.com>
11207 * config/i386/i386.h (struct machine_function): Add
11208 no_drap_save_restore field.
11209 * config/i386/i386.c (ix86_save_reg): Use
11210 !cfun->machine->no_drap_save_restore instead of
11211 crtl->stack_realign_needed.
11212 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
11213 this function clears frame_pointer_needed. Set
11214 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
11215 and DRAP reg is needed.
11217 2014-01-06 Marek Polacek <polacek@redhat.com>
11220 * doc/implement-c.texi: Mention that other integer types are
11221 permitted as bit-field types in strictly conforming mode.
11223 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
11225 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
11226 is newly allocated.
11228 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
11230 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
11232 2014-01-06 Martin Jambor <mjambor@suse.cz>
11235 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
11237 * ipa-prop.c (ipa_print_node_params): Fix indentation.
11239 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
11243 * var-tracking.c (add_stores): Preserve the value of the source even if
11244 we don't record the store.
11246 2014-01-06 Terry Guo <terry.guo@arm.com>
11248 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
11250 2014-01-05 Iain Sandoe <iain@codesourcery.com>
11253 * config/darwin.c (darwin_function_section): Adjust return values to
11254 correspond to optimisation changes made in r206070.
11256 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
11258 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
11259 from prefetch_block tune setting.
11260 (nocona_cost): Correct size of prefetch block to 64.
11262 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
11264 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
11265 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
11266 used to save the static chain register in the computation of the offset
11267 from which the FP registers need to be restored.
11269 2014-01-04 Jakub Jelinek <jakub@redhat.com>
11271 PR tree-optimization/59519
11272 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
11273 ICE if get_current_def (current_new_name) is already non-NULL, as long
11274 as it is a phi result of some other phi in *new_exit_bb that has
11277 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
11278 or vmovdqu* for misaligned_operand.
11279 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
11280 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
11281 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
11282 aligned_mem for AVX512F masked aligned load and store builtins and for
11283 non-temporal moves.
11285 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
11287 PR tree-optimization/59651
11288 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
11289 Address range for negative step should be added by TYPE_SIZE_UNIT.
11291 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
11293 * config/m68k/m68k.c (handle_move_double): Handle pushes with
11294 overlapping registers also for registers other than the stack pointer.
11296 2014-01-03 Marek Polacek <polacek@redhat.com>
11299 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
11302 2014-01-03 Jakub Jelinek <jakub@redhat.com>
11305 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
11308 * config/i386/i386.md (MODE_SIZE): New mode attribute.
11309 (push splitter): Use <P:MODE_SIZE> instead of
11310 GET_MODE_SIZE (<P:MODE>mode).
11311 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
11312 (mov -1, reg peephole2): Likewise.
11313 * config/i386/sse.md (*mov<mode>_internal,
11314 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
11315 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
11316 *<code><mode>3, *andnot<mode>3<mask_name>,
11317 <mask_codefor><code><mode>3<mask_name>): Likewise.
11318 * config/i386/subst.md (mask_mode512bit_condition,
11319 sd_mask_mode512bit_condition): Likewise.
11321 2014-01-02 Xinliang David Li <davidxl@google.com>
11323 PR tree-optimization/59303
11324 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
11325 (dump_predicates): Better output format.
11326 (pred_equal_p): New function.
11327 (is_neq_relop_p): Ditto.
11328 (is_neq_zero_form_p): Ditto.
11329 (pred_expr_equal_p): Ditto.
11330 (pred_neg_p): Ditto.
11331 (simplify_pred): Ditto.
11332 (simplify_preds_2): Ditto.
11333 (simplify_preds_3): Ditto.
11334 (simplify_preds_4): Ditto.
11335 (simplify_preds): Ditto.
11336 (push_pred): Ditto.
11337 (push_to_worklist): Ditto.
11338 (get_pred_info_from_cmp): Ditto.
11339 (is_degenerated_phi): Ditto.
11340 (normalize_one_pred_1): Ditto.
11341 (normalize_one_pred): Ditto.
11342 (normalize_one_pred_chain): Ditto.
11343 (normalize_preds): Ditto.
11344 (normalize_cond_1): Remove function.
11345 (normalize_cond): Ditto.
11346 (is_gcond_subset_of): Ditto.
11347 (is_subset_of_any): Ditto.
11348 (is_or_set_subset_of): Ditto.
11349 (is_and_set_subset_of): Ditto.
11350 (is_norm_cond_subset_of): Ditto.
11351 (pred_chain_length_cmp): Ditto.
11352 (convert_control_dep_chain_into_preds): Type change.
11353 (find_predicates): Ditto.
11354 (find_def_preds): Ditto.
11355 (destroy_predicates_vecs): Ditto.
11356 (find_matching_predicates_in_rest_chains): Ditto.
11357 (use_pred_not_overlap_with_undef_path_pred): Ditto.
11358 (is_pred_expr_subset): Ditto.
11359 (is_pred_chain_subset_of): Ditto.
11360 (is_included_in): Ditto.
11361 (is_superset_of): Ditto.
11363 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
11365 Update copyright years.
11367 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
11369 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
11370 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
11371 config/arc/arc.md, config/arc/arc.opt,
11372 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
11373 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
11374 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
11375 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
11376 config/linux-protos.h, config/linux.c, config/winnt-c.c,
11377 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
11378 vtable-verify.c, vtable-verify.h: Use the standard form for the
11381 2014-01-02 Tobias Burnus <burnus@net-b.de>
11383 * gcc.c (process_command): Update copyright notice dates.
11384 * gcov-dump.c: Ditto.
11386 * doc/cpp.texi: Bump @copying's copyright year.
11387 * doc/cppinternals.texi: Ditto.
11388 * doc/gcc.texi: Ditto.
11389 * doc/gccint.texi: Ditto.
11390 * doc/gcov.texi: Ditto.
11391 * doc/install.texi: Ditto.
11392 * doc/invoke.texi: Ditto.
11394 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11396 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
11398 2014-01-01 Jakub Jelinek <jakub@redhat.com>
11400 * config/i386/sse.md (*mov<mode>_internal): Guard
11401 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
11403 PR rtl-optimization/59647
11404 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
11405 new_rtx into UNSIGNED_FLOAT rtxes.
11407 Copyright (C) 2014 Free Software Foundation, Inc.
11409 Copying and distribution of this file, with or without modification,
11410 are permitted in any medium without royalty provided the copyright
11411 notice and this notice are preserved.