1 2013-11-19 Peter Bergner <bergner@vnet.ibm.com>
3 * loop-doloop.c (doloop_optimize_loops): Remove unused
4 loop iterator argument from FOR_EACH_LOOP.
6 2013-11-19 David Malcolm <dmalcolm@redhat.com>
8 Convert gimple types from a union to C++ inheritance.
9 * Makefile.in (GIMPLE_H): Add dep on is-a.h.
10 * coretypes.h (union gimple_statement_d): Remove declaration.
11 (gimple): Convert from being a "union gimple_statement_d *"
12 to a "struct gimple_statement_base *".
13 (const_gimple): Likewise (with "const").
14 * ggc.h (ggc_alloc_cleared_gimple_statement_d_stat): Replace
16 (ggc_alloc_cleared_gimple_statement_stat): ...this.
17 * gimple-pretty-print.c (debug): Change parameter from a
18 "gimple_statement_d &" to a "gimple_statement_base &".
19 (debug): Change parameter from a "gimple_statement_d *" to
20 a "gimple_statement_base *".
21 * gimple-pretty-print.h (debug): Update declarations as above.
22 * gimple.c (gimple_alloc_stat): Update for renaming of
23 ggc_alloc_cleared_gimple_statement_d_stat to
24 ggc_alloc_cleared_gimple_statement_stat.
25 * gimple.h: Include "is-a.h" for use by is_a_helper
26 specializations in followup autogenerated patch.
27 (struct gimple statement_base): Make this type usable as a base
28 class by adding "desc", "tag" and "variable_size" to GTY, thus
29 using opting-in to gengtype's support for simple inheritance.
30 (gimple_statement_with_ops_base): Convert to a subclass of
31 gimple_statement_base, dropping initial "gsbase" field. Note
32 that this type is abstract, with no GSS_ value, and thus no GTY
34 (gimple_statement_with_ops): Convert to a subclass of
35 gimple_statement_with_ops_base, dropping initial "opbase" field.
36 Add tag value to GTY marking. Update marking of op field to
37 reflect how num_ops field is accessed via inheritance.
38 (gimple_statement_with_memory_ops_base): Convert to a subclass of
39 gimple_statement_with_ops_base, dropping initial "opbase" field.
40 Add tag value to GTY marking.
41 (gimple_statement_with_memory_ops): Convert to a subclass of
42 public gimple_statement_with_memory_ops_base, dropping initial
43 "membase" field. Add tag value to GTY marking. Update marking
44 of op field to reflect how num_ops field is accessed via
46 (gimple_statement_call): Analogous changes that also update the
47 marking of the "u" union.
48 (gimple_statement_omp): Convert to a subclass of
49 gimple_statement_base, dropping initial "gsbase" field, adding
50 tag value to GTY marking.
51 (gimple_statement_bind): Likewise.
52 (gimple_statement_catch): Likewise.
53 (gimple_statement_eh_filter): Likewise.
54 (gimple_statement_eh_else): Likewise.
55 (gimple_statement_eh_mnt): Likewise.
56 (gimple_statement_phi): Likewise.
57 (gimple_statement_eh_ctrl): Likewise.
58 (gimple_statement_try): Likewise.
59 (gimple_statement_wce): Likewise.
60 (gimple_statement_asm): Convert to a subclass of
61 gimple_statement_with_memory_ops_base, dropping initial
62 "membase" field, adding tag value to GTY marking, and updating
64 (gimple_statement_omp_critical): Convert to a subclass of
65 gimple_statement_omp, dropping initial "omp" field, adding tag
67 (gimple_statement_omp_for): Likewise.
68 (gimple_statement_omp_parallel): Likewise.
69 (gimple_statement_omp_task): Convert to a subclass of
70 gimple_statement_omp_parallel, dropping initial "par" field,
71 adding tag value to GTY marking.
72 (gimple_statement_omp_sections): Convert to a subclass of
73 gimple_statement_omp, dropping initial "omp" field, adding
74 tag value to GTY marking.
75 (gimple_statement_omp_continue): Convert to a subclass of
76 gimple_statement_base, dropping initial "gsbase" field, adding
77 tag value to GTY marking.
78 (gimple_statement_omp_single): Convert to a subclass of
79 gimple_statement_omp, dropping initial "omp" field, adding
80 tag value to GTY marking.
81 (gimple_statement_omp_atomic_load): Convert to a subclass of
82 gimple_statement_base, dropping initial "gsbase" field, adding
83 tag value to GTY marking.
84 (gimple_statement_omp_atomic_store): Convert to a subclass of
85 gimple_statement_base, dropping initial "gsbase" field, adding
86 tag value to GTY marking.
87 (gimple_statement_transaction): Convert to a subclass of
88 gimple_statement_with_memory_ops_base, dropping initial "gsbase"
89 field, adding tag value to GTY marking.
90 (union gimple_statement_d): Remove.
91 * system.h (CONST_CAST_GIMPLE): Update to use
92 "struct gimple_statement_base *" rather than
93 "union gimple_statement_d *".
94 * tree-ssa-ccp.c (gimple_htab): Convert underlying type from
95 gimple_statement_d to gimple_statement_base.
97 * gimple.h (gimple_use_ops): Port from union to usage of
99 (gimple_set_use_ops): Port from union to usage of as_a.
100 (gimple_set_vuse): Likewise.
101 (gimple_set_vdef): Likewise.
102 (gimple_call_internal_fn): Port from union to a static_cast,
103 given that the type has already been asserted.
104 (gimple_omp_body_ptr): Port from unchecked union usage to
106 (gimple_omp_set_body): Likewise.
108 * gimple-iterator.c (update_bb_for_stmts): Update for conversion of
109 gimple types to a true class hierarchy.
110 (update_call_edge_frequencies): Likewise.
111 (gsi_insert_seq_nodes_before): Likewise.
112 (gsi_insert_seq_nodes_after): Likewise.
113 (gsi_split_seq_after): Likewise.
114 (gsi_set_stmt): Likewise.
115 (gsi_split_seq_before): Likewise.
116 (gsi_remove): Likewise.
117 * gimple-iterator.h (gsi_one_before_end_p): Likewise.
118 (gsi_next): Likewise.
119 (gsi_prev): Likewise.
120 * gimple-pretty-print.c (dump_gimple_debug): Likewise.
121 * gimple-ssa.h (gimple_vuse_op): Likewise.
122 (gimple_vdef_op): Likewise.
123 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
124 * gimple-streamer-out.c (output_gimple_stmt): Likewise.
125 * gimple.c (gimple_set_code): Likewise.
126 (gimple_alloc_stat): Likewise.
127 (gimple_set_subcode): Likewise.
128 (gimple_build_call_internal_1): Likewise.
129 (gimple_check_failed): Likewise.
130 (gimple_call_flags): Likewise.
131 (gimple_set_bb): Likewise.
132 * gimple.h (is_a_helper <gimple_statement_asm> (gimple)): New.
133 (is_a_helper <gimple_statement_bind> (gimple)): Likewise.
134 (is_a_helper <gimple_statement_call> (gimple)): Likewise.
135 (is_a_helper <gimple_statement_catch> (gimple)): Likewise.
136 (is_a_helper <gimple_statement_eh_ctrl> (gimple)): Likewise.
137 (is_a_helper <gimple_statement_eh_else> (gimple)): Likewise.
138 (is_a_helper <gimple_statement_eh_filter> (gimple)): Likewise.
139 (is_a_helper <gimple_statement_eh_mnt> (gimple)): Likewise.
140 (is_a_helper <gimple_statement_omp_atomic_load> (gimple)): Likewise.
141 (is_a_helper <gimple_statement_omp_atomic_store> (gimple)): Likewise.
142 (is_a_helper <gimple_statement_omp_continue> (gimple)): Likewise.
143 (is_a_helper <gimple_statement_omp_critical> (gimple)): Likewise.
144 (is_a_helper <gimple_statement_omp_for> (gimple)): Likewise.
145 (is_a_helper <gimple_statement_omp_parallel> (gimple)): Likewise.
146 (is_a_helper <gimple_statement_omp_sections> (gimple)): Likewise.
147 (is_a_helper <gimple_statement_omp_single> (gimple)): Likewise.
148 (is_a_helper <gimple_statement_omp_task> (gimple)): Likewise.
149 (is_a_helper <gimple_statement_phi> (gimple)): Likewise.
150 (is_a_helper <gimple_statement_transaction> (gimple)): Likewise.
151 (is_a_helper <gimple_statement_try> (gimple)): Likewise.
152 (is_a_helper <gimple_statement_wce> (gimple)): Likewise.
153 (is_a_helper <const gimple_statement_asm> (const_gimple)): Likewise.
154 (is_a_helper <const gimple_statement_bind> (const_gimple)): Likewise.
155 (is_a_helper <const gimple_statement_call> (const_gimple)): Likewise.
156 (is_a_helper <const gimple_statement_catch> (const_gimple)): Likewise.
157 (is_a_helper <const gimple_statement_eh_ctrl> (const_gimple)):
159 (is_a_helper <const gimple_statement_eh_filter> (const_gimple)):
161 (is_a_helper <const gimple_statement_omp_atomic_load> (const_gimple)):
163 (is_a_helper <const gimple_statement_omp_atomic_store>
164 (const_gimple)): Likewise.
165 (is_a_helper <const gimple_statement_omp_continue> (const_gimple)):
167 (is_a_helper <const gimple_statement_omp_critical> (const_gimple)):
169 (is_a_helper <const gimple_statement_omp_for> (const_gimple)):
171 (is_a_helper <const gimple_statement_omp_parallel> (const_gimple)):
173 (is_a_helper <const gimple_statement_omp_sections> (const_gimple)):
175 (is_a_helper <const gimple_statement_omp_single> (const_gimple)):
177 (is_a_helper <const gimple_statement_omp_task> (const_gimple)):
179 (is_a_helper <const gimple_statement_phi> (const_gimple)): Likewise.
180 (is_a_helper <const gimple_statement_transaction> (const_gimple)):
182 (gimple_seq_last): Update for conversion of gimple types to a true
184 (gimple_seq_set_last): Likewise.
185 (gimple_code): Likewise.
186 (gimple_bb): Likewise.
187 (gimple_block): Likewise.
188 (gimple_set_block): Likewise.
189 (gimple_location): Likewise.
190 (gimple_location_ptr): Likewise.
191 (gimple_set_location): Likewise.
192 (gimple_no_warning_p): Likewise.
193 (gimple_set_no_warning): Likewise.
194 (gimple_set_visited): Likewise.
195 (gimple_visited_p): Likewise.
196 (gimple_set_plf): Likewise.
197 (gimple_plf): Likewise.
198 (gimple_set_uid): Likewise.
199 (gimple_uid): Likewise.
200 (gimple_init_singleton): Likewise.
201 (gimple_modified_p): Likewise.
202 (gimple_set_modified): Likewise.
203 (gimple_expr_code): Likewise.
204 (gimple_has_volatile_ops): Likewise.
205 (gimple_set_has_volatile_ops): Likewise.
206 (gimple_omp_subcode): Likewise.
207 (gimple_omp_set_subcode): Likewise.
208 (gimple_omp_return_set_nowait): Likewise.
209 (gimple_omp_section_set_last): Likewise.
210 (gimple_omp_parallel_set_combined_p): Likewise.
211 (gimple_omp_atomic_set_need_value): Likewise.
212 (gimple_omp_atomic_set_seq_cst): Likewise.
213 (gimple_num_ops): Likewise.
214 (gimple_set_num_ops): Likewise.
215 (gimple_assign_nontemporal_move_p): Likewise.
216 (gimple_assign_set_nontemporal_move): Likewise.
217 (gimple_assign_rhs_code): Likewise.
218 (gimple_assign_set_rhs_code): Likewise.
219 (gimple_call_internal_p): Likewise.
220 (gimple_call_with_bounds_p): Likewise.
221 (gimple_call_set_with_bounds): Likewise.
222 (gimple_call_set_tail): Likewise.
223 (gimple_call_tail_p): Likewise.
224 (gimple_call_set_return_slot_opt): Likewise.
225 (gimple_call_return_slot_opt_p): Likewise.
226 (gimple_call_set_from_thunk): Likewise.
227 (gimple_call_from_thunk_p): Likewise.
228 (gimple_call_set_va_arg_pack): Likewise.
229 (gimple_call_va_arg_pack_p): Likewise.
230 (gimple_call_set_nothrow): Likewise.
231 (gimple_call_set_alloca_for_var): Likewise.
232 (gimple_call_alloca_for_var_p): Likewise.
233 (gimple_call_copy_flags): Likewise.
234 (gimple_cond_code): Likewise.
235 (gimple_cond_set_code): Likewise.
236 (gimple_cond_make_false): Likewise.
237 (gimple_cond_make_true): Likewise.
238 (gimple_asm_volatile_p): Likewise.
239 (gimple_asm_set_volatile): Likewise.
240 (gimple_asm_set_input): Likewise.
241 (gimple_asm_input_p): Likewise.
242 (gimple_try_kind): Likewise.
243 (gimple_try_set_kind): Likewise.
244 (gimple_try_catch_is_cleanup): Likewise.
245 (gimple_try_set_catch_is_cleanup): Likewise.
246 (gimple_wce_cleanup_eh_only): Likewise.
247 (gimple_wce_set_cleanup_eh_only): Likewise.
248 (gimple_debug_bind_p): Likewise.
249 (gimple_debug_source_bind_p): Likewise.
250 (gimple_omp_for_set_kind): Likewise.
251 (gimple_omp_for_set_combined_p): Likewise.
252 (gimple_omp_for_set_combined_into_p): Likewise.
253 (gimple_omp_target_set_kind): Likewise.
254 (gimple_transaction_subcode): Likewise.
255 (gimple_transaction_set_subcode): Likewise.
256 (gimple_predict_predictor): Likewise.
257 (gimple_predict_set_predictor): Likewise.
258 (gimple_predict_outcome): Likewise.
259 (gimple_predict_set_outcome): Likewise.
260 (gimple_transaction_set_label): Likewise.
261 (gimple_transaction_set_body): Likewise.
262 (gimple_transaction_label_ptr): Likewise.
263 (gimple_transaction_label): Likewise.
264 (gimple_transaction_body_ptr): Likewise.
265 (gimple_omp_continue_set_control_use): Likewise.
266 (gimple_omp_continue_control_use_ptr): Likewise.
267 (gimple_omp_continue_control_use): Likewise.
268 (gimple_omp_continue_set_control_def): Likewise.
269 (gimple_omp_continue_control_def_ptr): Likewise.
270 (gimple_omp_continue_control_def): Likewise.
271 (gimple_omp_atomic_load_rhs_ptr): Likewise.
272 (gimple_omp_atomic_load_rhs): Likewise.
273 (gimple_omp_atomic_load_set_rhs): Likewise.
274 (gimple_omp_atomic_load_lhs_ptr): Likewise.
275 (gimple_omp_atomic_load_lhs): Likewise.
276 (gimple_omp_atomic_load_set_lhs): Likewise.
277 (gimple_omp_atomic_store_val_ptr): Likewise.
278 (gimple_omp_atomic_store_val): Likewise.
279 (gimple_omp_atomic_store_set_val): Likewise.
280 (gimple_omp_for_cond): Likewise.
281 (gimple_omp_for_set_cond): Likewise.
282 (gimple_omp_sections_set_control): Likewise.
283 (gimple_omp_sections_control_ptr): Likewise.
284 (gimple_omp_sections_control): Likewise.
285 (gimple_omp_sections_set_clauses): Likewise.
286 (gimple_omp_sections_clauses_ptr): Likewise.
287 (gimple_omp_sections_clauses): Likewise.
288 (gimple_omp_teams_set_clauses): Likewise.
289 (gimple_omp_teams_clauses_ptr): Likewise.
290 (gimple_omp_teams_clauses): Likewise.
291 (gimple_omp_target_set_data_arg): Likewise.
292 (gimple_omp_target_data_arg_ptr): Likewise.
293 (gimple_omp_target_data_arg): Likewise.
294 (gimple_omp_target_set_child_fn): Likewise.
295 (gimple_omp_target_child_fn_ptr): Likewise.
296 (gimple_omp_target_child_fn): Likewise.
297 (gimple_omp_target_set_clauses): Likewise.
298 (gimple_omp_target_clauses_ptr): Likewise.
299 (gimple_omp_target_clauses): Likewise.
300 (gimple_omp_single_set_clauses): Likewise.
301 (gimple_omp_single_clauses_ptr): Likewise.
302 (gimple_omp_single_clauses): Likewise.
303 (gimple_omp_task_set_arg_align): Likewise.
304 (gimple_omp_task_arg_align_ptr): Likewise.
305 (gimple_omp_task_arg_align): Likewise.
306 (gimple_omp_task_set_arg_size): Likewise.
307 (gimple_omp_task_arg_size_ptr): Likewise.
308 (gimple_omp_task_arg_size): Likewise.
309 (gimple_omp_task_set_copy_fn): Likewise.
310 (gimple_omp_task_copy_fn_ptr): Likewise.
311 (gimple_omp_task_copy_fn): Likewise.
312 (gimple_omp_task_set_data_arg): Likewise.
313 (gimple_omp_task_data_arg_ptr): Likewise.
314 (gimple_omp_task_data_arg): Likewise.
315 (gimple_omp_task_set_child_fn): Likewise.
316 (gimple_omp_task_child_fn_ptr): Likewise.
317 (gimple_omp_task_child_fn): Likewise.
318 (gimple_omp_task_set_clauses): Likewise.
319 (gimple_omp_task_clauses_ptr): Likewise.
320 (gimple_omp_task_clauses): Likewise.
321 (gimple_omp_parallel_set_data_arg): Likewise.
322 (gimple_omp_parallel_data_arg_ptr): Likewise.
323 (gimple_omp_parallel_data_arg): Likewise.
324 (gimple_omp_parallel_set_child_fn): Likewise.
325 (gimple_omp_parallel_child_fn_ptr): Likewise.
326 (gimple_omp_parallel_child_fn): Likewise.
327 (gimple_omp_parallel_set_clauses): Likewise.
328 (gimple_omp_parallel_clauses_ptr): Likewise.
329 (gimple_omp_parallel_clauses): Likewise.
330 (gimple_omp_for_set_pre_body): Likewise.
331 (gimple_omp_for_pre_body_ptr): Likewise.
332 (gimple_omp_for_set_incr): Likewise.
333 (gimple_omp_for_incr_ptr): Likewise.
334 (gimple_omp_for_incr): Likewise.
335 (gimple_omp_for_set_final): Likewise.
336 (gimple_omp_for_final_ptr): Likewise.
337 (gimple_omp_for_final): Likewise.
338 (gimple_omp_for_set_initial): Likewise.
339 (gimple_omp_for_initial_ptr): Likewise.
340 (gimple_omp_for_initial): Likewise.
341 (gimple_omp_for_set_index): Likewise.
342 (gimple_omp_for_index_ptr): Likewise.
343 (gimple_omp_for_index): Likewise.
344 (gimple_omp_for_collapse): Likewise.
345 (gimple_omp_for_set_clauses): Likewise.
346 (gimple_omp_for_clauses_ptr): Likewise.
347 (gimple_omp_for_clauses): Likewise.
348 (gimple_omp_critical_set_name): Likewise.
349 (gimple_omp_critical_name_ptr): Likewise.
350 (gimple_omp_critical_name): Likewise.
351 (gimple_eh_dispatch_set_region): Likewise.
352 (gimple_eh_dispatch_region): Likewise.
353 (gimple_resx_set_region): Likewise.
354 (gimple_resx_region): Likewise.
355 (gimple_phi_set_arg): Likewise.
356 (gimple_phi_arg): Likewise.
357 (gimple_phi_set_result): Likewise.
358 (gimple_phi_result_ptr): Likewise.
359 (gimple_phi_result): Likewise.
360 (gimple_phi_num_args): Likewise.
361 (gimple_phi_capacity): Likewise.
362 (gimple_wce_set_cleanup): Likewise.
363 (gimple_wce_cleanup_ptr): Likewise.
364 (gimple_try_set_cleanup): Likewise.
365 (gimple_try_set_eval): Likewise.
366 (gimple_try_cleanup_ptr): Likewise.
367 (gimple_try_eval_ptr): Likewise.
368 (gimple_eh_else_set_e_body): Likewise.
369 (gimple_eh_else_set_n_body): Likewise.
370 (gimple_eh_else_e_body_ptr): Likewise.
371 (gimple_eh_else_n_body_ptr): Likewise.
372 (gimple_eh_must_not_throw_set_fndecl): Likewise.
373 (gimple_eh_must_not_throw_fndecl): Likewise.
374 (gimple_eh_filter_set_failure): Likewise.
375 (gimple_eh_filter_set_types): Likewise.
376 (gimple_eh_filter_failure_ptr): Likewise.
377 (gimple_eh_filter_types_ptr): Likewise.
378 (gimple_eh_filter_types): Likewise.
379 (gimple_catch_set_handler): Likewise.
380 (gimple_catch_set_types): Likewise.
381 (gimple_catch_handler_ptr): Likewise.
382 (gimple_catch_types_ptr): Likewise.
383 (gimple_catch_types): Likewise.
384 (gimple_asm_string): Likewise.
385 (gimple_asm_set_label_op): Likewise.
386 (gimple_asm_label_op): Likewise.
387 (gimple_asm_set_clobber_op): Likewise.
388 (gimple_asm_clobber_op): Likewise.
389 (gimple_asm_set_output_op): Likewise.
390 (gimple_asm_output_op_ptr): Likewise.
391 (gimple_asm_output_op): Likewise.
392 (gimple_asm_set_input_op): Likewise.
393 (gimple_asm_input_op_ptr): Likewise.
394 (gimple_asm_input_op): Likewise.
395 (gimple_asm_nlabels): Likewise.
396 (gimple_asm_nclobbers): Likewise.
397 (gimple_asm_noutputs): Likewise.
398 (gimple_asm_ninputs): Likewise.
399 (gimple_bind_set_block): Likewise.
400 (gimple_bind_block): Likewise.
401 (gimple_bind_add_seq): Likewise.
402 (gimple_bind_add_stmt): Likewise.
403 (gimple_bind_set_body): Likewise.
404 (gimple_bind_body_ptr): Likewise.
405 (gimple_bind_append_vars): Likewise.
406 (gimple_bind_set_vars): Likewise.
407 (gimple_bind_vars): Likewise.
408 (gimple_call_clobber_set): Likewise.
409 (gimple_call_use_set): Likewise.
410 (gimple_call_set_internal_fn): Likewise.
411 (gimple_call_set_fntype): Likewise.
412 (gimple_call_fntype): Likewise.
413 (gimple_omp_return_lhs_ptr): Likewise.
414 (gimple_omp_return_lhs): Likewise.
415 (gimple_omp_return_set_lhs): Likewise.
416 (gimple_omp_taskreg_set_data_arg): Likewise.
417 (gimple_omp_taskreg_data_arg_ptr): Likewise.
418 (gimple_omp_taskreg_data_arg): Likewise.
419 (gimple_omp_taskreg_set_child_fn): Likewise.
420 (gimple_omp_taskreg_child_fn_ptr): Likewise.
421 (gimple_omp_taskreg_child_fn): Likewise.
422 (gimple_omp_taskreg_set_clauses): Likewise.
423 (gimple_omp_taskreg_clauses_ptr): Likewise.
424 (gimple_omp_taskreg_clauses): Likewise.
425 (gimple_vuse): Likewise.
426 (gimple_vdef): Likewise.
427 (gimple_vuse_ptr): Likewise.
428 (gimple_vdef_ptr): Likewise.
429 * tree-inline.c (copy_debug_stmt): Likewise.
430 * tree-phinodes.c (make_phi_node): Likewise.
432 * gimple.h (is_a_helper <const gimple_statement_with_ops>::test): New.
433 (is_a_helper <gimple_statement_with_ops>::test): New.
434 (is_a_helper <const gimple_statement_with_memory_ops>::test): New.
435 (is_a_helper <gimple_statement_with_memory_ops>::test): New.
437 * gimple-streamer-in.c (input_gimple_stmt): Port from union
438 access to use of as_a.
439 * gimple.c (gimple_build_asm_1): Likewise.
440 (gimple_build_try): Likewise. Also, return a specific subclass
441 rather than just gimple.
442 (gimple_build_resx): Port from union access to use of as_a.
443 (gimple_build_eh_dispatch): Likewise.
444 (gimple_build_omp_for): Likewise. Also, convert allocation of iter
445 now that gengtype no longer provides a typed allocator function.
446 (gimple_copy): Likewise.
447 * gimple.h (gimple_build_try): Return a specific subclass rather
449 * gimplify.c (gimplify_cleanup_point_expr): Replace union access
450 with subclass access by making use of new return type of
452 * tree-phinodes.c: (allocate_phi_node): Return a
453 "gimple_statement_phi *" rather than just a gimple.
454 (resize_phi_node): Likewise.
455 (make_phi_node): Replace union access with subclass access by
456 making use of new return type of allocate_phi_node.
457 (reserve_phi_args_for_new_edge): Replace union access with as_a.
458 (remove_phi_arg_num): Accept a "gimple_statement_phi *" rather
460 (remove_phi_args): Update for change to remove_phi_arg_num.
462 * gdbhooks.py (GimplePrinter.to_string): Update lookup of
463 code field to reflect inheritance, rather than embedding of
464 the base gimple type.
466 2013-11-19 Richard Biener <rguenther@suse.de>
468 * cfgloop.h (struct loop_iterator): C++-ify, add constructor
469 and destructor and make fel_next a member function.
470 (fel_next): Transform into ...
471 (loop_iterator::next): ... this.
472 (fel_init): Transform into ...
473 (loop_iterator::loop_iterator): ... this.
474 (loop_iterator::~loop_iterator): New.
475 (FOR_EACH_LOOP): Remove loop-iterator argument.
476 (FOR_EACH_LOOP_BREAK): Remove no longer necessary macro.
477 * cfgloop.c, cfgloopmanip.c, config/mn10300/mn10300.c,
478 graphite-clast-to-gimple.c, graphite-scop-detection.c,
479 graphite-sese-to-poly.c, ipa-inline-analysis.c, ipa-pure-const.c,
480 loop-init.c, loop-invariant.c, loop-unroll.c, loop-unswitch.c,
481 modulo-sched.c, predict.c, sel-sched-ir.c, tree-cfg.c, tree-data-ref.c,
482 tree-if-conv.c, tree-loop-distribution.c, tree-parloops.c,
483 tree-predcom.c, tree-scalar-evolution.c, tree-ssa-dce.c,
484 tree-ssa-loop-ch.c, tree-ssa-loop-im.c, tree-ssa-loop-ivcanon.c,
485 tree-ssa-loop-ivopts.c, tree-ssa-loop-manip.c, tree-ssa-loop-niter.c,
486 tree-ssa-loop-prefetch.c, tree-ssa-loop-unswitch.c,
487 tree-ssa-threadupdate.c, tree-vectorizer.c, tree-vrp.c: Adjust
488 uses of FOR_EACH_LOOP and remove loop_iterator variables. Replace
489 FOR_EACH_LOOP_BREAK with break.
491 2013-11-19 Richard Biener <rguenther@suse.de>
493 PR tree-optimization/59164
494 * tree-vect-loop-manip.c (vect_update_ivs_after_vectorizer):
496 * tree-vect-loop.c (vect_analyze_loop_operations): Adjust
497 check whether we can create an epilogue loop to reflect the
498 cases where we create one.
500 2013-11-19 Andrew MacLeod <amacleod@redhat.com>
502 * graphite-sese-to-poly.c: Include expr.h.
504 2013-11-19 Richard Biener <rguenther@suse.de>
507 * tree-ssa-ter.c (find_replaceable_in_bb): Avoid forwarding
508 loads into stmts that may clobber it.
510 2013-11-19 Bernd Schmidt <bernds@codesourcery.com>
512 * cgraphunit.c (symtab_terminator): New variable.
513 (queued_nodes): Renamed from first. Use symtab_terminator as
515 (analyze_functions): Adjust accordingly.
516 (cgraph_process_new_functions): Return void.
517 * cgraph.h (cgraph_process_new_functions): Adjust declaration.
519 2013-11-19 Marek Polacek <polacek@redhat.com>
521 * opts.c (common_handle_option): Add -fsanitize=null option.
522 Turn off -fdelete-null-pointer-checks option when doing the
523 NULL pointer checking.
524 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH): Add.
525 * tree-pass.h (make_pass_ubsan): Declare.
526 (make_pass_sanopt): Declare.
527 * timevar.def (TV_TREE_UBSAN): New timevar.
528 * passes.def: Add pass_sanopt and pass_ubsan.
529 * ubsan.h (ubsan_null_ckind): New enum.
530 (ubsan_mismatch_data): New struct.
531 (ubsan_expand_null_ifn): Declare.
532 (ubsan_create_data): Adjust declaration.
533 (ubsan_type_descriptor): Likewise.
534 * asan.c: Include "ubsan.h".
535 (pass_data_sanopt): New pass.
536 (execute_sanopt): New function.
537 (gate_sanopt): Likewise.
538 (make_pass_sanopt): Likewise.
539 (class pass_sanopt): New class.
540 * ubsan.c: Include tree-pass.h, gimple-ssa.h, gimple-walk.h,
541 gimple-iterator.h and cfgloop.h.
542 (PROB_VERY_UNLIKELY): Define.
543 (tree_type_map_hash): New function.
544 (ubsan_type_descriptor): Add new parameter.
545 Improve type name generation.
546 (ubsan_create_data): Add new parameter. Add pointer data into
548 (ubsan_expand_null_ifn): New function.
549 (instrument_member_call): Likewise.
550 (instrument_mem_ref): Likewise.
551 (instrument_null): Likewise.
552 (ubsan_pass): Likewise.
553 (gate_ubsan): Likewise.
554 (make_pass_ubsan): Likewise.
555 (ubsan_instrument_unreachable): Adjust ubsan_create_data call.
556 (class pass_ubsan): New class.
557 (pass_data_ubsan): New pass.
558 * flag-types.h (enum sanitize_code): Add SANITIZE_NULL.
559 * internal-fn.c (expand_UBSAN_NULL): New function.
560 * cgraphunit.c (varpool_finalize_decl): Call varpool_assemble_decl
561 even when !flag_toplevel_reorder.
562 * internal-fn.def (UBSAN_NULL): New.
564 2013-11-19 Jan Hubicka <jh@suse.cz>
566 * cgraph.c (cgraph_create_indirect_edge): Use get_polymorphic_call_info.
567 * cgraph.h (cgraph_indirect_call_info): Add outer_type, maybe_in_construction
568 and maybe_derived_type.
569 * ipa-utils.h (ipa_polymorphic_call_context): New structure.
570 (ipa_dummy_polymorphic_call_context): New global var.
571 (possible_polymorphic_call_targets): Add context paramter.
572 (dump_possible_polymorphic_call_targets): Likewise; update
574 (possible_polymorphic_call_target_p): Likewise.
575 (get_polymorphic_call_info): New function.
576 * ipa-devirt.c (ipa_dummy_polymorphic_call_context): New function.
577 (add_type_duplicate): Remove forgotten debug output.
578 (method_class_type): Add sanity check.
579 (maybe_record_node): Add FINALP parameter.
580 (record_binfo): Add OUTER_TYPE and OFFSET; walk the inner
581 by info by get_binfo_at_offset.
582 (possible_polymorphic_call_targets_1): Add OUTER_TYPE/OFFSET parameters;
583 pass them to record-binfo.
584 (polymorphic_call_target_d): Add context and FINAL.
585 (polymorphic_call_target_hasher::hash): Hash context.
586 (polymorphic_call_target_hasher::equal): Compare context.
587 (free_polymorphic_call_targets_hash):
588 (get_class_context): New function.
589 (contains_type_p): New function.
590 (get_polymorphic_call_info): New function.
591 (walk_bases): New function.
592 (possible_polymorphic_call_targets): Add context parameter; honnor it.
593 (dump_possible_polymorphic_call_targets): Dump context.
594 (possible_polymorphic_call_target_p): Add context.
595 (update_type_inheritance_graph): Update comment.s
596 (ipa_set_jf_known_type): Assert that compoentn type is known.
597 (ipa_note_param_call): Do not tamper with offsets.
598 (ipa_analyze_indirect_call_uses): When offset is being changed; clear
600 (update_indirect_edges_after_inlining): Likewise.
601 (ipa_write_indirect_edge_info): Stream new fields.
602 (ipa_read_indirect_edge_info): Stream in new fields.
604 2013-11-19 Jan Hubicka <jh@suse.cz>
606 * tree-pretty-print.c (dump_generic_node): Print class type of
609 2013-11-19 Joey Ye <joey.ye@arm.com>
611 * config/arm/arm.opt (-marm-pic-data-is-text-relative): New option.
612 * doc/invoke.texi (-marm-pic-data-is-text-relative): Documentation
614 * config/arm/arm.c (arm_option_override): By default disable
615 -marm-pic-data-is-text-relative.
616 (legitimize_pic_address): Use arm_pic_data_is_text_relative.
617 (arm_assemble_integer): Likewise.
618 * config/arm/arm.h (TARGET_DEFAULT_PIC_DATA_IS_TEXT_RELATIVE):
619 New macro to initialize -marm-pic-data-is-text-relative.
621 2013-11-19 Bin Cheng <bin.cheng@arm.com>
623 * tree-ssa-loop-ivopts.c (enum ainc_type): New.
624 (address_cost_data): New field.
625 (get_address_cost): Compute auto-increment rtx cost in ainc_costs.
626 Use ainc_costs for auto-increment rtx patterns.
629 2013-11-19 James Greenhalgh <james.greenhalgh@arm.com>
631 * config/aarch64/aarch64.md: Remove v8type from all insns.
633 2013-11-19 Richard Biener <rguenther@suse.de>
635 PR tree-optimization/57517
636 * tree-predcom.c (combinable_refs_p): Verify the combination
637 is always executed when the refs are.
639 2013-11-19 Jeff Law <law@redhat.com>
641 * tree-ssa-threadupdate.c: Include ssa-iterators.h
642 (copy_phi_arg_into_existing_phi): New function.
643 (any_remaining_duplicated_blocks): Likewise.
644 (ssa_fix_duplicate_block_edges): Handle multiple duplicated
645 blocks on a jump threading path.
647 * tree-ssa-threadupdate.c (thread_through_loop_header): Do not
648 thread through a joiner which has the latch edge.
650 2013-11-19 Jan Hubicka <jh@suse.cz>
652 * md.texi (setmem): Document new parameter.
653 * optabs.c (maybe_gen_insn): Support 9 operands.
654 * builtins.c (determine_block_size): Add probable_max_size;
656 (expand_builtin_memcpy. expand_builtin_memset_args): Pass around
658 * expr.c (emit_block_move_via_movmem, emit_block_move_hints,
659 emit_block_move, clear_storage_hints, set_storage_via_setmem):
661 * expr.h (emit_block_move_hints, clear_storage_hints,
662 set_storage_via_setmem): Update prototype.
663 * i386.md (setmem, movmem patterns): Add 9th operand.
664 * i386-protos.h (ix86_expand_set_or_movmem): Update prototype.
665 * i386.c (ix86_expand_set_or_movmem): Take probable_max_size_exp
666 argument; pass it to decide_alg.
668 2013-11-19 David Malcolm <dmalcolm@redhat.com>
670 * basic-block.h (n_basic_blocks_for_function): Rename macro to...
671 (n_basic_blocks_for_fn): ...this.
673 (n_basic_blocks): Eliminate macro as work towards making uses of
676 * cfgloop.c (init_loops_structure): Update for renaming of
677 "n_basic_blocks_for_function" to "n_basic_blocks_for_fn".
678 * graph.c (draw_cfg_nodes_no_loops): Likewise.
679 * ipa-utils.c (ipa_merge_profiles): Likewise.
680 * lto-streamer-in.c (make_new_block): Likewise.
681 * tree-cfg.c (init_empty_tree_cfg_for_function): Likewise.
682 (dump_function_to_file): Likewise.
684 * alias.c (init_alias_analysis): Replace usage of "n_basic_blocks"
685 macro with "n_basic_blocks_for_fn (cfun)".
686 * bb-reorder.c (partition_hot_cold_basic_blocks): Likewise.
687 (duplicate_computed_gotos): Likewise.
688 (reorder_basic_blocks): Likewise.
689 * bt-load.c (augment_live_range): Likewise.
690 * cfg.c (expunge_block): Likewise.
691 (compact_blocks): Likewise.
692 * cfganal.c (single_pred_before_succ_order): Likewise.
693 (compute_idf): Likewise.
694 (flow_dfs_compute_reverse_init): Likewise.
695 (pre_and_rev_post_order_compute): Likewise.
696 (pre_and_rev_post_order_compute_fn): Likewise.
697 (inverted_post_order_compute): Likewise.
698 (post_order_compute): Likewise.
699 (print_edge_list): Likewise.
700 (find_unreachable_blocks): Likewise.
701 (mark_dfs_back_edges): Likewise.
702 * cfgcleanup.c (try_optimize_cfg): Likewise.
703 (try_forward_edges): Likewise.
704 * cfghooks.c (dump_flow_info): Likewise.
705 * cfgloop.c (verify_loop_structure): Likewise.
706 (get_loop_body): Likewise.
707 (flow_loops_find): Likewise.
708 * cfgloopmanip.c (add_loop): Likewise.
709 (remove_path): Likewise.
710 (find_path): Likewise.
711 * cfgrtl.c (rtl_flow_call_edges_add): Likewise.
712 (rtl_verify_bb_layout): Likewise.
713 (entry_of_function): Likewise.
714 (rtl_create_basic_block): Likewise.
715 * coverage.c (coverage_compute_cfg_checksum): Likewise.
716 * cprop.c (one_cprop_pass): Likewise.
717 (is_too_expensive): Likewise.
718 * df-core.c (df_compute_cfg_image): Likewise.
719 (df_compact_blocks): Likewise.
720 (df_worklist_dataflow_doublequeue): Likewise.
721 * dominance.c (calculate_dominance_info): Likewise.
722 (calc_dfs_tree): Likewise.
723 (calc_dfs_tree_nonrec): Likewise.
724 (init_dom_info): Likewise.
725 * domwalk.c (cmp_bb_postorder): Likewise.
726 * function.c (thread_prologue_and_epilogue_insns): Likewise.
727 (generate_setjmp_warnings): Likewise.
728 * fwprop.c (build_single_def_use_links): Likewise.
729 * gcse.c (is_too_expensive): Likewise.
730 (one_code_hoisting_pass): Likewise.
731 (one_pre_gcse_pass): Likewise.
732 * graphite.c (graphite_initialize): Likewise.
733 * haifa-sched.c (haifa_sched_init): Likewise.
734 * ipa-inline-analysis.c (estimate_function_body_sizes): Likewise.
735 * ira.c (split_live_ranges_for_shrink_wrap): Likewise.
736 * ira-build.c (ira_build): Likewise.
737 * lcm.c (compute_nearerout): Likewise.
738 (compute_available): Likewise.
739 (compute_laterin): Likewise.
740 (compute_antinout_edge): Likewise.
741 * lra-lives.c (lra_create_live_ranges): Likewise.
742 * lra.c (has_nonexceptional_receiver): Likewise.
743 * mcf.c (create_fixup_graph): Likewise.
744 * profile.c (branch_prob): Likewise.
745 * reg-stack.c (convert_regs_2): Likewise.
746 * regrename.c (regrename_analyze): Likewise.
747 * reload1.c (has_nonexceptional_receiver): Likewise.
748 * reorg.c (dbr_schedule): Likewise.
749 * sched-deps.c (sched_deps_init): Likewise.
750 * sched-ebb.c (schedule_ebbs): Likewise.
751 * sched-rgn.c (extend_regions): Likewise.
752 (schedule_insns): Likewise.
753 (sched_rgn_init): Likewise.
754 (extend_rgns): Likewise.
755 (haifa_find_rgns): Likewise.
756 * sel-sched-ir.c (recompute_rev_top_order): Likewise.
757 (sel_recompute_toporder): Likewise.
758 * sel-sched.c (run_selective_scheduling): Likewise.
759 * store-motion.c (one_store_motion_pass): Likewise.
760 (remove_reachable_equiv_notes): Likewise.
761 * tracer.c (tracer): Likewise.
762 (tail_duplicate): Likewise.
763 * tree-cfg.c (gimple_flow_call_edges_add): Likewise.
764 (dump_cfg_stats): Likewise.
765 (gimple_dump_cfg): Likewise.
766 (create_bb): Likewise.
767 (build_gimple_cfg): Likewise.
768 * tree-cfgcleanup.c (merge_phi_nodes): Likewise.
769 * tree-inline.c (optimize_inline_calls): Likewise.
770 (fold_marked_statements): Likewise.
771 * tree-ssa-ifcombine.c (tree_ssa_ifcombine): Likewise.
772 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
773 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
774 * tree-ssa-loop-manip.c (compute_live_loop_exits): Likewise.
775 * tree-ssa-math-opts.c (execute_cse_reciprocals): Likewise.
776 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
777 * tree-ssa-pre.c (do_pre): Likewise.
778 (init_pre): Likewise.
779 (compute_avail): Likewise.
780 * tree-ssa-reassoc.c (init_reassoc): Likewise.
781 * tree-ssa-sccvn.c (init_scc_vn): Likewise.
782 * tree-ssa-tail-merge.c (alloc_cluster_vectors): Likewise.
783 (init_worklist): Likewise.
784 * tree-ssa-uncprop.c (associate_equivalences_with_edges): Likewise.
785 * var-tracking.c (variable_tracking_main_1): Likewise.
786 (vt_find_locations): Likewise.
787 (vt_stack_adjustments): Likewise.
788 * config/s390/s390.c (s390_optimize_nonescaping_tx): Likewise.
789 * config/spu/spu.c (spu_machine_dependent_reorg): Likewise.
791 2013-11-18 Jan Hubicka <jh@suse.cz>
793 * profile.c (compute_branch_probabilities): Do not sanity check run_max.
795 2013-11-18 Kenneth Zadeck <zadeck@naturalbridge.com>
797 * tree.c (int_fits_type_p): Change GET_MODE_BITSIZE to
799 * fold-const.c (fold_single_bit_test_into_sign_test)
800 (fold_binary_loc): Change GET_MODE_BITSIZE to
803 2013-11-18 Teresa Johnson <tejohnson@google.com>
805 * gcc/cfgrtl.c (cfg_layout_initialize): Assert if we
806 try to go into cfglayout after bb reordering.
807 * gcc/passes.def: Move compgotos before bb reordering
808 since it goes into cfglayout.
810 2013-11-18 Bernd Schmidt <bernds@codesourcery.com>
812 * cgraphunit.c (ipa_passes): Don't execute all_lto_gen_passes.
813 * lto-streamer-out.c (lto_output, produce_asm_for_decls): No longer
815 (pass_data_ipa_lto_gimple_out, pass_ipa_lto_gimple_out,
816 make_pass_ipa_lto_gimple_out, pass_data_ipa_lto_finish_out,
817 pass_ipa_lto_finish_out, make_pass_ipa_lto_finish_out): Remove.
818 * lto-streamer.h (lto_output, produce_asm_for_decls): Declare.
819 * pass-manager.h (GCC_PASS_LISTS, class pass_manager):
820 Remove all_lto_gen_passes.
821 * passes.c (pass_manager::dump_passes): Remove its use.
822 (pass_manager::register_pass): Likewise.
823 (ipa_read_summaries, ipa_read_optimization_summaries): Likewise.
824 (pass_manager::pass_manager): Don't initialize or use it.
825 (write_lto): New static function.
826 (ipa_write_summaries_1, ipa_write_optimization_summaries): Use it
827 instead of using all_lto_gen_passes.
828 * passes.def (all_lto_gen_passes, pass_ipa_lto_gimple_out,
829 pass_ipa_lto_finish_out): Delete.
830 * tree-pass.h (make_pass_ipa_lto_gimple_out,
831 make_pass_ipa_lto_finish_out): Don't declare.
833 2013-11-18 Jeff Law <law@redhat.com>
835 * tree-ssa-threadupdate.c (redirection_data): Record two
836 duplicated blocks instead of just one.
837 (local_info): Explain why we don't create a template for the
838 second duplicated block in a thread path.
839 (create_block_for_threading): Accept argument indicating array
840 index into redirection_data to store its result.
841 (lookup_redirection_data): Initialize both duplicate blocks.
842 (ssa_create_duplicates): If a jump threading path needs multiple
843 blocks duplicated, then duplicate them.
844 (ssa_fix_duplicate_block_edges): Corresponding changes.
845 (ssa_fixup_template_block, thread_single_edge): Likewise.
847 2013-11-18 Marek Polacek <polacek@redhat.com>
849 * doc/invoke.texi: Extend -fsanitize=undefined documentation.
851 2013-11-18 Andrew Pinski <apinski@cavium.com>
852 Steve Ellcey <sellcey@mips.com>
855 * config/mips/mips.md (*mov<GPR:mode>_on_<MOVECC:mode>): Remove
856 type restriction from equality_operator on conditonal move.
857 (*mov<SCALARF:mode>_on_<MOVECC:mode>): Ditto.
858 (*mov<GPR:mode>_on_<GPR2:mode>_ne): New.
860 2013-11-18 Jeff Law <law@redhat.com>
862 * tree-ssa-threadupdate.c: Fix file block comment.
863 Fix minor indention issue.
865 2013-11-18 Uros Bizjak <ubizjak@gmail.com>
867 * config/i386/i386.c (ix86_decompose_address): Use REG_P instead of
868 ix86_address_subreg_operand. Move subreg checks to
869 ix86_validate_address_register. Move address override check to
870 ix86_legitimate_address_p.
871 (ix86_validate_address_register): New function.
872 (ix86_legitimate_address_p): Call ix86_validate_address_register
873 to validate base and index registers. Add address override check
874 from ix86_decompose_address.
875 (ix86_decompose_address): Remove.
877 2013-11-18 Richard Biener <rguenther@suse.de>
879 PR tree-optimization/59125
880 PR tree-optimization/54570
881 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): When inlining
882 is not complete do not treat component-references with offset zero
883 but different fields as equal.
884 * tree-object-size.c: Include tree-phinodes.h and ssa-iterators.h.
885 (compute_object_sizes): Apply TLC. Propagate the constant
886 results into all uses and fold their stmts.
887 * passes.def (pass_all_optimizations): Move pass_object_sizes
888 after the first pass_forwprop and before pass_fre.
890 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
892 * tree.h (tree_to_uhwi): Return an unsigned HOST_WIDE_INT.
893 * tree.c (tree_to_uhwi): Return an unsigned HOST_WIDE_INT.
894 (tree_ctz): Remove cast to unsigned type.
895 * builtins.c (fold_builtin_memory_op): Likewise.
896 * dwarf2out.c (descr_info_loc): Likewise.
897 * godump.c (go_output_typedef): Likewise.
898 * omp-low.c (expand_omp_simd): Likewise.
899 * stor-layout.c (excess_unit_span): Likewise.
900 * tree-object-size.c (addr_object_size): Likewise.
901 * tree-sra.c (analyze_all_variable_accesses): Likewise.
902 * tree-ssa-forwprop.c (simplify_builtin_call): Likewise.
903 (simplify_rotate): Likewise.
904 * tree-ssa-strlen.c (adjust_last_stmt, handle_builtin_memcpy)
905 (handle_pointer_plus): Likewise.
906 * tree-switch-conversion.c (check_range): Likewise.
907 * tree-vect-patterns.c (vect_recog_rotate_pattern): Likewise.
908 * tsan.c (instrument_builtin_call): Likewise.
909 * cfgexpand.c (defer_stack_allocation): Add cast to HOST_WIDE_INT.
910 * trans-mem.c (tm_log_add): Likewise.
911 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Likewise.
912 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
913 * config/rs6000/rs6000.c (rs6000_aggregate_candidate): Likewise.
914 * config/mips/mips.c (r10k_safe_mem_expr_p): Make offset unsigned.
916 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
918 * tree.h (host_integerp, tree_low_cst): Delete.
919 * tree.c (host_integerp, tree_low_cst): Delete.
921 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
923 * expr.h: Update comments to refer to tree_to_[su]hwi rather
925 * fold-const.c (fold_binary_loc): Likewise.
926 * expr.c (store_constructor): Use tree_to_uhwi rather than
928 * ipa-utils.h (possible_polymorphic_call_target_p): Likewise.
929 * stmt.c (emit_case_dispatch_table): Likewise.
930 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
932 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
934 * alias.c, asan.c, builtins.c, cfgexpand.c, cgraph.c,
935 config/aarch64/aarch64.c, config/alpha/predicates.md,
936 config/arm/arm.c, config/darwin.c, config/epiphany/epiphany.c,
937 config/i386/i386.c, config/iq2000/iq2000.c, config/m32c/m32c-pragma.c,
938 config/mep/mep-pragma.c, config/mips/mips.c,
939 config/picochip/picochip.c, config/rs6000/rs6000.c, cppbuiltin.c,
940 dbxout.c, dwarf2out.c, emit-rtl.c, except.c, expr.c, fold-const.c,
941 function.c, gimple-fold.c, godump.c, ipa-cp.c, ipa-prop.c, omp-low.c,
942 predict.c, sdbout.c, stor-layout.c, trans-mem.c, tree-object-size.c,
943 tree-sra.c, tree-ssa-ccp.c, tree-ssa-forwprop.c,
944 tree-ssa-loop-ivcanon.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-niter.c,
945 tree-ssa-loop-prefetch.c, tree-ssa-strlen.c, tree-stdarg.c,
946 tree-switch-conversion.c, tree-vect-generic.c, tree-vect-loop.c,
947 tree-vect-patterns.c, tree-vrp.c, tree.c, tsan.c, ubsan.c, varasm.c:
948 Replace tree_low_cst (..., 1) with tree_to_uhwi throughout.
950 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
952 * builtins.c, cilk-common.c, config/aarch64/aarch64.c,
953 config/alpha/alpha.c, config/arm/arm.c, config/c6x/predicates.md,
954 config/i386/i386.c, config/ia64/predicates.md, config/s390/s390.c,
955 coverage.c, dbxout.c, dwarf2out.c, except.c, explow.c, expr.c, expr.h,
956 fold-const.c, gimple-fold.c, godump.c, ipa-prop.c, omp-low.c,
957 predict.c, rtlanal.c, sdbout.c, stmt.c, stor-layout.c, targhooks.c,
958 tree-cfg.c, tree-data-ref.c, tree-inline.c, tree-ssa-forwprop.c,
959 tree-ssa-loop-prefetch.c, tree-ssa-phiopt.c, tree-ssa-sccvn.c,
960 tree-ssa-strlen.c, tree-stdarg.c, tree-vect-data-refs.c,
961 tree-vect-patterns.c, tree.c, tree.h, var-tracking.c, varasm.c:
962 Replace tree_low_cst (..., 0) with tree_to_shwi throughout.
964 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
966 * tree.h (tree_to_shwi, tree_to_uhwi): Declare, with inline expansions.
967 * tree.c (tree_to_shwi, tree_to_uhwi): New functions.
969 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
971 * expr.h: Update comments to refer to tree_fits_[su]hwi_p rather
974 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
976 * builtins.c, config/alpha/alpha.c, config/iq2000/iq2000.c,
977 config/mips/mips.c, dbxout.c, dwarf2out.c, expr.c, fold-const.c,
978 gimple-fold.c, godump.c, omp-low.c, predict.c, sdbout.c, stor-layout.c,
979 tree-dfa.c, tree-sra.c, tree-ssa-forwprop.c, tree-ssa-loop-prefetch.c,
980 tree-ssa-phiopt.c, tree-ssa-sccvn.c, tree-ssa-strlen.c,
981 tree-ssa-structalias.c, tree-vect-data-refs.c, tree-vect-patterns.c,
982 tree.c, varasm.c, alias.c, cfgexpand.c, config/aarch64/aarch64.c,
983 config/arm/arm.c, config/epiphany/epiphany.c, config/i386/i386.c,
984 config/m32c/m32c-pragma.c, config/mep/mep-pragma.c,
985 config/rs6000/rs6000.c, config/sparc/sparc.c, emit-rtl.c, function.c,
986 gimplify.c, ipa-prop.c, stmt.c, trans-mem.c, tree-cfg.c,
987 tree-object-size.c, tree-ssa-ccp.c, tree-ssa-loop-ivcanon.c,
988 tree-stdarg.c, tree-switch-conversion.c, tree-vect-generic.c,
989 tree-vrp.c, tsan.c, ubsan.c: Replace host_integerp (..., 1) with
990 tree_fits_uhwi_p throughout.
992 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
994 * builtins.c, config/alpha/alpha.c, config/c6x/predicates.md,
995 config/ia64/predicates.md, config/iq2000/iq2000.c, config/mips/mips.c,
996 config/s390/s390.c, dbxout.c, dwarf2out.c, except.c, explow.c, expr.c,
997 expr.h, fold-const.c, gimple-fold.c, gimple-ssa-strength-reduction.c,
998 gimple.c, godump.c, graphite-scop-detection.c, graphite-sese-to-poly.c,
999 omp-low.c, predict.c, rtlanal.c, sdbout.c, simplify-rtx.c,
1000 stor-layout.c, tree-data-ref.c, tree-dfa.c, tree-pretty-print.c,
1001 tree-sra.c, tree-ssa-alias.c, tree-ssa-forwprop.c,
1002 tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c, tree-ssa-math-opts.c,
1003 tree-ssa-phiopt.c, tree-ssa-reassoc.c, tree-ssa-sccvn.c,
1004 tree-ssa-strlen.c, tree-ssa-structalias.c, tree-vect-data-refs.c,
1005 tree-vect-patterns.c, tree-vectorizer.h, tree.c, var-tracking.c,
1006 varasm.c: Replace host_integerp (..., 0) with tree_fits_shwi_p
1009 2013-11-18 Richard Sandiford <rdsandiford@googlemail.com>
1011 * tree.h (tree_fits_shwi_p, tree_fits_uhwi_p): Declare.
1012 * tree.c (tree_fits_shwi_p, tree_fits_uhwi_p): Define.
1014 2013-11-18 Kirill Yukhin <kirill.yukhin@intel.com>
1016 * gcc/config/ia64/ia64.c (ia64_split_tmode_move): Mark load with
1017 `dead' flag if it kills address, not its post-increment.
1019 2013-11-18 Ilya Enkovich <ilya.enkovich@intel.com>
1021 * builtin-types.def (BT_FN_PTR_CONST_PTR_VAR): New.
1022 * chkp-builtins.def (BUILT_IN_CHKP_BIND_BOUNDS): New.
1023 * cfgexpand.c (expand_call_stmt): Expand BUILT_IN_CHKP_BIND_BOUNDS.
1024 * gimple.c (gimple_call_get_nobnd_arg_index): Remove.
1025 * gimple.h (gf_mask): Add GF_CALL_WITH_BOUNDS.
1026 (gimple_call_with_bounds_p): New.
1027 (gimple_call_set_with_bounds): New.
1028 (gimple_call_num_nobnd_args): Remove.
1029 (gimple_call_nobnd_arg): Remove.
1030 * tree.h (CALL_WITH_BOUNDS_P): New.
1031 * rtl.h (CALL_EXPR_WITH_BOUNDS_P): New.
1033 2013-11-18 Trevor Saunders <tsaunders@mozilla.com>
1035 * cgraph.h (symtab_node_asm_name): Rename to symtab_node::asm_name.
1036 (symtab_node_name): Rename to symtab_node::name.
1037 (cgraph_node_asm_name): Remove.
1038 (varpool_node_asm_name): Remove.
1039 * cgraph.c cgraphclones.c cgraphunit.c ipa-cp.c ipa-devirt.c
1040 ipa-inline-analysis.c ipa-inline-transform.c ipa-inline.c
1041 ipa-profile.c ipa-prop.c ipa-pure-const.c ipa-ref.c ipa-reference.c
1042 ipa-utils.c ipa.c symtab.c tree-inline.c tree-sra.c
1043 tree-ssa-structalias.c value-prof.c varpool.c Adjust.
1045 2013-11-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1047 * config/arm/aarch-cost-tables.h (cortexa53_extra_costs): New table.
1048 * config/arm/arm.c (arm_cortex_a53_tune): New.
1049 * config/arm/arm-cores.def (cortex-a53): Use cortex_a53 tuning struct.
1051 2013-11-12 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
1053 * config.gcc (i[34567]86-*-linux* | ...): Add bdver4.
1054 (case ${target}): Add bdver4.
1055 * config/i386/bdver3.md: Add bdver4.
1056 * config/i386/driver-i386.c: (host_detect_local_cpu): Let
1057 -march=native recognize bdver4 processors.
1058 * config/i386/i386-c.c (ix86_target_macros_internal): Add
1059 bdver4 def_and_undef
1060 * config/i386/i386.c (struct processor_costs bdver4_cost): New.
1061 (m_BDVER4): New definition.
1062 (m_AMD_MULTIPLE): Includes m_BDVER4.
1063 (processor_target_table): Add bdver4 entry.
1064 (static const char *const cpu_names): Add bdver4 entry.
1065 (software_prefetching_beneficial_p): Add bdver3.
1066 (ix86_option_override_internal): Add bdver4 instruction sets.
1067 (ix86_issue_rate): Add bdver4.
1068 (ix86_adjust_cost): Add bdver4.
1069 (ia32_multipass_dfa_lookahead): Add bdver4.
1070 (enum processor_model): Add M_AMDFAM15H_BDVER4.
1071 (struct _arch_names_table): Add M_AMDFAM15H_BDVER4.
1072 (has_dispatch): Add bdver4.
1073 * config/i386/i386.h (TARGET_BDVER4): New definition.
1074 (enum target_cpu_default): Add TARGET_CPU_DEFAULT_bdver4.
1075 (enum processor_type): Add PROCESSOR_BDVER4.
1076 * config/i386/i386.md (define_attr "cpu"): Add bdver4.
1077 * config/i386/i386.opt (flag_dispatch_scheduler): Add bdver4.
1078 * gcc/doc/extend.texi: Add details about bdver4.
1079 * gcc/doc/invoke.texi: Add details about bdver4. Add
1080 fma4 and fsgsbase for bdver3. Add fma4 for bdver2.
1082 2013-11-17 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
1084 * config/rs6000/rs6000.c (rs6000_emit_move): Use low word of
1085 sdmode_stack_slot also in little-endian mode.
1087 2013-11-17 Jan Hubicka <jh@suse.cz>
1089 * doc/md.texi (setmem, movstr): Update documentation.
1090 * builtins.c (determine_block_size): New function.
1091 (expand_builtin_memcpy): Use it and pass it to emit_block_move_hints.
1092 (expand_builtin_memset_args): Use it and pass it to
1093 set_storage_via_setmem.
1094 * expr.c (emit_block_move_via_movmem): Add min_size/max_size
1095 parameters; update call to expander.
1096 (emit_block_move_hints): Add min_size/max_size parameters.
1097 (clear_storage_hints): Likewise.
1098 (set_storage_via_setmem): Likewise.
1099 (clear_storage): Update.
1100 * expr.h (emit_block_move_hints, clear_storage_hints,
1101 set_storage_via_setmem): Update prototypes.
1102 * i386.c (ix86_expand_set_or_movmem): Add bounds; export.
1103 (ix86_expand_movmem, ix86_expand_setmem): Remove.
1104 (ix86_expand_movmem, ix86_expand_setmem): Remove.
1105 * i386.md (movmem, setmem): Pass parameters.
1107 2013-11-17 Uros Bizjak <ubizjak@gmail.com>
1110 * config/i386/i386.c (ix86_address_subreg_operand): Do not
1111 reject non-integer subregs.
1112 (ix86_decompose_address): Do not reject invalid CONST_INT RTXes.
1113 Move check for invalid x32 constant addresses ...
1114 (ix86_legitimate_address_p): ... here.
1116 2011-11-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1118 * config/rs6000/rs6000.c (rs6000_frame_related): Add split_reg
1119 parameter and use it in REG_FRAME_RELATED_EXPR note.
1120 (emit_frame_save): Call rs6000_frame_related with extra NULL_RTX
1122 (rs6000_emit_prologue): Likewise, but for little endian VSX
1123 stores, pass the source register of the store instead.
1125 2013-11-17 Andrew MacLeod <amacleod@redhat.com>
1127 * gimple.h: Reorder prototypes to match .c declaration order, and
1128 remove protyotypes for functions not in gimple.c.
1129 (LABEL): Move to tree-into-ssa.c.
1130 * gimple.c: Remove unused prototypes.
1131 (get_base_address): Move to tree.c.
1132 * tree.c (get_base_address): Relocate from gimple.c.
1133 * builtins.h (validate_gimple_arglist): Add prototype.
1134 * trans-mem.h (compute_transaction_bits, is_tm_ending): Add prototype.
1135 * cfgexpand.h: New File.
1136 (gimple_assign_rhs_to_tree, estimated_stack_frame_size): Add protoype.
1137 * tree.h (build_addr): Move to tree-nested.h.
1138 * tree-nested.h: New File.
1139 (build_addr, lower_nested_functions, insert_field_into_struct): Add
1141 * tree-inline.h (estimated_stack_frame_size): Remove prototype.
1142 * ipa-inline-analysis.c: Include cfgexpand.h.
1143 * cgraphunit.c: Include tree-nested.h.
1144 * omp-low.c: Likewise.
1145 * tree-parloops.c: Likewise.
1146 * gimple-low.h: Likewise.
1147 * tree-profile.h: Likewise.
1148 * expr.c: Include cfgexpand.h.
1149 * tree-affine.c: Likewise.
1150 * tree-ssa.c: Likewise.
1151 * tree-ssa-loop-im.c: Include trans-mem.h.
1152 * tree-ssa-tail-merge.c: Likewise.
1153 * value-prof.c: Include builtins.h and tree-nested.h.
1154 * tree-into-ssa.c (LABEL): Define here.
1156 2013-11-16 Joern Rennecke <joern.rennecke@embecosm.com>
1158 * config/arc/arc.c (arc_predicate_delay_insns): New function.
1159 (pass_data_arc_predicate_delay_insns): New pass_data instance.
1160 (pass_arc_predicate_delay_insns): New subclass of rtl_opt_class.
1161 (make_pass_arc_predicate_delay_insns): New function.
1162 (arc_init): Register pass_arc_predicate_delay_insns if
1163 flag_delayed_branch is active.
1165 2013-11-16 Joern Rennecke <joern.rennecke@embecosm.com>
1167 * config/arc/constraints.md (Rcq): Simplify register number test.
1169 2013-11-15 Aldy Hernandez <aldyh@redhat.com>
1171 * gimple.h (enum gf_mask): Change the ordering of GF_OMP_* bits.
1173 2013-11-15 Kaz Kojima <kkojima@gcc.gnu.org>
1175 * config/sh/sh.c (barrier_align): Return 0 when barrier_or_label
1178 2013-11-15 Aldy Hernandez <aldyh@redhat.com>
1180 * Makefile.in (C_COMMON_OBJS): Depend on c-cilkplus.o.
1181 * gimple-pretty-print.c (dump_omp_for): Add case for
1182 GF_OMP_FOR_KIND_CILKSIMD.
1183 * gimple.h (enum gf_mask): Restructure entries to add
1184 GF_OMP_FOR_KIND_CILKSIMD.
1185 * gimplify.c (is_gimple_stmt): Add case for CILK_SIMD.
1186 (gimplify_omp_for): Handle CILK_SIMD.
1187 (gimplify_expr): Add ccase for CILK_SIMD.
1188 * omp-low.c (extract_omp_for_data): Handle CILK_SIMD.
1189 (build_outer_var_ref): Same.
1190 (check_omp_nesting_restrictions): Same.
1191 (lower_rec_input_clauses): Same.
1192 (lower_lastprivate_clauses): Same.
1193 (expand_omp_for): Same.
1194 (execute_expand_omp): Check flag_enable_cilkplus.
1195 (execute_lower_omp): Same.
1196 (diagnose_sb_0): Handle CILK_SIMD.
1197 (diagnose_omp_structured_block_errors): Check flag_enable_cilkplus.
1198 (setjmp_or_longjmp_p): New.
1199 (scan_omp_1_stmt): Error on setjmp/longjmp in a simd construct.
1200 * tree-pretty-print.c (dump_generic_node): Add case for CILK_SIMD.
1201 * tree.def: Add tree code for CILK_SIMD.
1203 2013-11-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1205 * config/rs6000/altivec.md (UNSPEC_VPERM_X, UNSPEC_VPERM_UNS_X):
1207 (altivec_vperm_<mode>): Revert earlier little endian change.
1208 (*altivec_vperm_<mode>_internal): Remove.
1209 (altivec_vperm_<mode>_uns): Revert earlier little endian change.
1210 (*altivec_vperm_<mode>_uns_internal): Remove.
1211 * config/rs6000/vector.md (vec_realign_load_<mode>): Revise commentary.
1213 2013-11-15 Jeff Law <law@redhat.com>
1215 * basic-block.h (has_abnormal_or_eh_outgoing_edge): Renamed from
1216 has_abnormal_or_outgoing_edge. Check for EH edges as well.
1217 * gimple-ssa-isolate-paths.c
1218 (find_implicit_erroneous_behaviour): Corresponding changes.
1219 Do not check stmt_ends_bb_p or GIMPLE_RETURN anymore.
1220 (find_explicit_erroneous_behaviour): Likewise.
1222 2013-11-15 Jeff Law <law@redhat.com>
1224 * ifcvt.c (find_cond_trap): Properly handle case where
1227 2013-11-15 Andreas Schwab <schwab@linux-m68k.org>
1229 * configure: Regenerate.
1231 2013-11-15 James Greenhalgh <james.greenhalgh@arm.com>
1233 * config/aarch64/aarch64-simd.md: Remove simd_type from all patterns.
1234 * config/aarch64/aarch64.md: Likewise, correct "type" attribute
1235 where it is incorrect or missing.
1237 2013-11-15 Richard Sandiford <rdsandiford@googlemail.com>
1239 * dwarf2out.c (gen_enumeration_type_die): Remove unnecessary
1241 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
1242 Use TREE_INT_CST_LOW rather than tree_low_cst when reading the
1244 * fold-const.c (fold_binary_loc): Replace a host_integerp/tree_low_cst
1245 pair with a TREE_CODE test and TREE_INT_CST_LOW.
1246 * tree-vect-generic.c (expand_vector_divmod): Likewise.
1248 2013-11-15 Richard Biener <rguenther@suse.de>
1250 PR tree-optimization/50262
1251 * tree-ssa-alias.h (struct pt_solution): Split
1252 vars_contains_global into vars_contains_nonlocal,
1253 vars_contains_escaped and vars_contains_escaped_heap.
1254 * tree-ssa-structalias.c (label_visit): Expand comment.
1255 (handle_lhs_call): Adjust comment.
1256 (set_uids_in_ptset): Set the new flags appropriately.
1257 (pt_solution_set): Adjust.
1258 (pt_solution_set_var): Likewise.
1259 (pt_solution_ior_into): Likewise.
1260 (pt_solution_includes_global): Likewise.
1261 (pt_solutions_intersect_1): Optimize escaped handling.
1262 (compute_points_to_sets): Remove heap variable globalization.
1263 (ipa_escaped_pt): Adjust initializer.
1264 (pass_data_ipa_pta): Do not run TODO_update_ssa.
1265 * gimple-pretty-print.c (pp_points_to_solution): Print split flags.
1266 * tree-ssa-alias.c (dump_points_to_solution): Likewise.
1268 2013-11-15 Richard Biener <rguenther@suse.de>
1270 * tree-loop-distribution.c (tree_loop_distribution): Make sure
1271 to distribute all stores.
1273 2013-11-15 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
1275 * doc/invoke.texi (-mabi=elfv1, -mabi=elfv2): Document.
1277 2013-11-15 Joseph Myers <joseph@codesourcery.com>
1279 * acinclude.m4 (GCC_GLIBC_VERSION_GTE_IFELSE): New configure macro.
1280 * configure.ac: Determine target_header_dir earlier.
1281 (--with-glibc-version): New configure option.
1282 Use GCC_GLIBC_VERSION_GTE_IFELSE in enable_gnu_unique_object,
1283 gcc_cv_libc_provides_ssp and gcc_cv_target_ldbl128 tests.
1284 * configure: Regenerate.
1285 * doc/install.texi (--enable-gnu-unique-object): Don't refer to
1286 native toolchains for default.
1287 (--with-glibc-version): Document.
1289 2013-11-15 Eric Botcazou <ebotcazou@adacore.com>
1291 * fold-const.c (fold_binary_loc) <comparisons>: Reuse local variable.
1293 2013-11-15 Uros Bizjak <ubizjak@gmail.com>
1295 * lto-streamer-in.c (input function): Call cgraph_create_node if
1296 cgraph_get_node failed.
1298 2013-11-14 Olivier Hainque <hainque@adacore.com>
1300 * cfgexpand.c (defer_stack_allocation): When optimization is enabled,
1301 defer allocation of DECL_IGNORED_P variables at toplevel unless really
1302 small. Factorize size threshold computation from the existing one.
1303 (expand_used_vars): Refine comment.
1305 2013-11-14 Cong Hou <congh@google.com>
1307 * tree-vectorizer.h (struct dr_with_seg_len): Remove the base
1308 address field as it can be obtained from dr. Rename the struct.
1309 * tree-vect-data-refs.c (comp_dr_with_seg_len_pair): Consider
1310 steps of data references during sort.
1311 (vect_prune_runtime_alias_test_list): Adjust with the change to
1312 struct dr_with_seg_len.
1313 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
1314 Adjust with the change to struct dr_with_seg_len.
1316 2013-11-14 Jeff Law <law@redhat.com>
1319 * basic-block.h (has_abnormal_outgoing_edge_p): Moved here from...
1320 * tree-inline.c (has_abnormal_outgoing_edge_p): Remove.
1321 * gimple-ssa-isolate-paths.c: Include tree-cfg.h.
1322 (find_implicit_erroneous_behaviour): If a block has abnormal outgoing
1323 edges, then ignore it. If the statement exhibiting erroneous
1324 behaviour ends basic blocks, with the exception of GIMPLE_RETURNs,
1325 then we can not optimize.
1326 (find_explicit_erroneous_behaviour): Likewise.
1328 2013-11-14 Andrew MacLeod <amacleod@redhat.com>
1330 * gimplify-me.h: New file. Add prototypes.
1331 * gimplify.h: Don't include gimple.h.
1332 (struct gimplify_hasher, struct gimplify_ctx, is_gimple_sizepos):
1333 Relocate from gimple.h.
1334 * gimple.h (struct gimplify_hasher, struct gimplify_ctx,
1335 is_gimple_sizepos): Move to gimplify.h.
1336 (gimplify_hasher::hash, gimplify_hasher::equal): Move to gimplify.c.
1337 (enum gsi_iterator_update): Move to gimple-iterator.h.
1338 * gimple-iterator.h (enum gsi_iterator_update): Relocate from gimple.h.
1339 * gimplify-me.c: New File.
1340 (force_gimple_operand_1, force_gimple_operand,
1341 force_gimple_operand_gsi_1, force_gimple_operand_gsi,
1342 gimple_regimplify_operands): Relocate from gimplify.c.
1343 * gimplify.c (force_gimple_operand_1, force_gimple_operand,
1344 force_gimple_operand_gsi_1, force_gimple_operand_gsi,
1345 gimple_regimplify_operands): Move to gimplify-me.c.
1346 (gimplify_hasher::hash, gimplify_hasher::equal): Relocate
1348 * Makefile.in (OBJS): Add gimplify-me.o
1349 * asan.c: Include only gimplify.h, gimplify-me.h, and/or gimple.h as
1351 * cfgloopmanip.c: Likewise.
1352 * cgraphunit.c: Likewise.
1353 * cilk-common.c: Likewise.
1354 * fold-const.c: Likewise.
1355 * function.c: Likewise.
1356 * gimple-expr.c: Likewise.
1357 * gimple-fold.c: Likewise.
1358 * gimple-ssa-strength-reduction.c: Likewise.
1359 * gimple.c: Likewise.
1360 * graphite-clast-to-gimple.c: Likewise.
1361 * graphite-sese-to-poly.c: Likewise.
1362 * ipa-prop.c: Likewise.
1363 * ipa-split.c: Likewise.
1365 * langhooks.c: Likewise.
1366 * omp-low.c: Likewise.
1368 * stor-layout.c: Likewise.
1369 * targhooks.c: Likewise.
1370 * trans-mem.c: Likewise.
1371 * tree-affine.c: Likewise.
1372 * tree-cfg.c: Likewise.
1373 * tree-cfgcleanup.c: Likewise.
1374 * tree-complex.c: Likewise.
1375 * tree-if-conv.c: Likewise.
1376 * tree-inline.c: Likewise.
1377 * tree-loop-distribution.c: Likewise.
1378 * tree-nested.c: Likewise.
1379 * tree-parloops.c: Likewise.
1380 * tree-predcom.c: Likewise.
1381 * tree-profile.c: Likewise.
1382 * tree-scalar-evolution.c: Likewise.
1383 * tree-sra.c: Likewise.
1384 * tree-ssa-address.c: Likewise.
1385 * tree-ssa-ccp.c: Likewise.
1386 * tree-ssa-dce.c: Likewise.
1387 * tree-ssa-forwprop.c: Likewise.
1388 * tree-ssa-ifcombine.c: Likewise.
1389 * tree-ssa-loop-im.c: Likewise.
1390 * tree-ssa-loop-ivopts.c: Likewise.
1391 * tree-ssa-loop-manip.c: Likewise.
1392 * tree-ssa-loop-niter.c: Likewise.
1393 * tree-ssa-loop-prefetch.c: Likewise.
1394 * tree-ssa-loop-unswitch.c: Likewise.
1395 * tree-ssa-math-opts.c: Likewise.
1396 * tree-ssa-phiopt.c: Likewise.
1397 * tree-ssa-phiprop.c: Likewise.
1398 * tree-ssa-pre.c: Likewise.
1399 * tree-ssa-propagate.c: Likewise.
1400 * tree-ssa-reassoc.c: Likewise.
1401 * tree-ssa-sccvn.c: Likewise.
1402 * tree-ssa-strlen.c: Likewise.
1403 * tree-ssa.c: Likewise.
1404 * tree-switch-conversion.c: Likewise.
1405 * tree-tailcall.c: Likewise.
1406 * tree-vect-data-refs.c: Likewise.
1407 * tree-vect-generic.c: Likewise.
1408 * tree-vect-loop-manip.c: Likewise.
1409 * tree-vect-loop.c: Likewise.
1410 * tree-vect-patterns.c: Likewise.
1411 * tree-vect-stmts.c: Likewise.
1414 * value-prof.c: Likewise.
1415 * config/aarch64/aarch64.c: Likewise.
1416 * config/alpha/alpha.c: Likewise.
1417 * config/darwin.c: Likewise.
1418 * config/i386/i386.c: Likewise.
1419 * config/ia64/ia64.c: Likewise.
1420 * config/mep/mep.c: Likewise.
1421 * config/mips/mips.c: Likewise.
1422 * config/rs6000/rs6000.c: Likewise.
1423 * config/s390/s390.c: Likewise.
1424 * config/sh/sh.c: Likewise.
1425 * config/sparc/sparc.c: Likewise.
1426 * config/spu/spu.c: Likewise.
1427 * config/stormy16/stormy16.c: Likewise.
1428 * config/tilegx/tilegx.c: Likewise.
1429 * config/tilepro/tilepro.c: Likewise.
1430 * config/xtensa/xtensa.c: Likewise.
1432 2013-11-14 Diego Novillo <dnovillo@google.com>
1434 * Makefile.in (PLUGIN_HEADERS): Add stringpool.h.
1436 2013-11-14 Diego Novillo <dnovillo@google.com>
1438 * tree.h: Include fold-const.h.
1439 (aggregate_value_p): Moved to function.h.
1440 (alloca_call_p): Moved to calls.h.
1441 (allocate_struct_function): Moved to function.h.
1442 (apply_tm_attr): Moved to attribs.h.
1443 (array_at_struct_end_p): Moved to expr.h.
1444 (array_ref_element_size): Moved to tree-dfa.h.
1445 (array_ref_low_bound): Moved to tree-dfa.h.
1446 (array_ref_up_bound): Moved to tree.h.
1447 (assemble_alias): Moved to cgraph.h.
1448 (bit_from_pos): Moved to stor-layout.h.
1449 (build_addr): Moved to tree-nested.h.
1450 (build_duplicate_type): Moved to tree-inline.h.
1451 (build_fold_addr_expr): Moved to fold-const.h.
1452 (build_fold_addr_expr_with_type): Moved to fold-const.h.
1453 (build_fold_addr_expr_with_type_loc): Moved to fold-const.h.
1454 (build_fold_indirect_ref): Moved to fold-const.h.
1455 (build_fold_indirect_ref_loc): Moved to fold-const.h.
1456 (build_personality_function): Moved to tree.h.
1457 (build_range_check): Moved to fold-const.h.
1458 (build_simple_mem_ref): Moved to fold-const.h.
1459 (build_simple_mem_ref_loc): Moved to fold-const.h.
1460 (build_tm_abort_call): Moved to trans-mem.h.
1461 (byte_from_pos): Moved to stor-layout.h.
1462 (call_expr_flags): Moved to calls.h.
1463 (can_move_by_pieces): Moved to expr.h.
1464 (categorize_ctor_elements): Moved to expr.h.
1465 (change_decl_assembler_name): Moved to gcc-symtab.h.
1466 (combine_comparisons): Moved to fold-const.h.
1467 (complete_ctor_at_level_p): Moved to tree.h.
1468 (component_ref_field_offset): Moved to tree-dfa.h.
1469 (compute_builtin_object_size): Moved to tree-object-size.h.
1470 (compute_record_mode): Moved to stor-layout.h.
1471 (constant_boolean_node): Moved to fold-const.h.
1472 (constructor_static_from_elts_p): Moved to varasm.h.
1473 (cxx11_attribute_p): Moved to attribs.h.
1474 (debug_body): Moved to print-tree.h.
1475 (debug_find_tree): Moved to tree-inline.h.
1476 (debug_fold_checksum): Moved to fold-const.h.
1477 (debug_head): Moved to print-tree.h.
1478 (debug_head): Moved to print-tree.h.
1479 (debug_raw): Moved to print-tree.h.
1480 (debug_tree): Moved to print-tree.h.
1481 (debug_vec_tree): Moved to print-tree.h.
1482 (debug_verbose): Moved to print-tree.h.
1483 (debug_verbose): Moved to print-tree.h.
1484 (decl_attributes): Moved to attribs.h.
1485 (decl_binds_to_current_def_p): Moved to varasm.h.
1486 (decl_default_tls_model): Moved to varasm.h.
1487 (decl_replaceable_p): Moved to varasm.h.
1488 (div_if_zero_remainder): Moved to fold-const.h.
1489 (double_int mem_ref_offset): Moved to fold-const.h.
1490 (dump_addr): Moved to print-tree.h.
1491 (element_precision): Moved to machmode.h.
1492 (expand_dummy_function_end): Moved to function.h.
1493 (expand_function_end): Moved to function.h.
1494 (expand_function_start): Moved to function.h.
1495 (expand_label): Moved to stmt.h.
1496 (expr_first): Moved to tree-iterator.h.
1497 (expr_last): Moved to tree-iterator.h.
1498 (finalize_size_functions): Moved to stor-layout.h.
1499 (finish_builtin_struct): Moved to stor-layout.h.
1500 (finish_record_layout): Moved to stor-layout.h.
1501 (fixup_signed_type): Moved to stor-layout.h.
1502 (fixup_unsigned_type): Moved to stor-layout.h.
1503 (flags_from_decl_or_type): Moved to calls.h.
1504 (fold): Moved to fold-const.h.
1505 (fold_abs_const): Moved to fold-const.h.
1506 (fold_binary): Moved to fold-const.h.
1507 (fold_binary_loc): Moved to fold-const.h.
1508 (fold_binary_to_constant): Moved to fold-const.h.
1509 (fold_build1): Moved to fold-const.h.
1510 (fold_build1_initializer_loc): Moved to fold-const.h.
1511 (fold_build1_loc): Moved to fold-const.h.
1512 (fold_build1_stat_loc): Moved to fold-const.h.
1513 (fold_build2): Moved to fold-const.h.
1514 (fold_build2_initializer_loc): Moved to fold-const.h.
1515 (fold_build2_loc): Moved to fold-const.h.
1516 (fold_build2_stat_loc): Moved to fold-const.h.
1517 (fold_build3): Moved to fold-const.h.
1518 (fold_build3_loc): Moved to fold-const.h.
1519 (fold_build3_stat_loc): Moved to fold-const.h.
1520 (fold_build_call_array): Moved to fold-const.h.
1521 (fold_build_call_array_initializer): Moved to fold-const.h.
1522 (fold_build_call_array_initializer_loc): Moved to fold-const.h.
1523 (fold_build_call_array_loc): Moved to fold-const.h.
1524 (fold_build_cleanup_point_expr): Moved to fold-const.h.
1525 (fold_convert): Moved to fold-const.h.
1526 (fold_convert_loc): Moved to fold-const.h.
1527 (fold_convertible_p): Moved to fold-const.h.
1528 (fold_defer_overflow_warnings): Moved to fold-const.h.
1529 (fold_deferring_overflow_warnings_p): Moved to fold-const.h.
1530 (fold_fma): Moved to fold-const.h.
1531 (fold_ignored_result): Moved to fold-const.h.
1532 (fold_indirect_ref): Moved to fold-const.h.
1533 (fold_indirect_ref_1): Moved to fold-const.h.
1534 (fold_indirect_ref_loc): Moved to fold-const.h.
1535 (fold_read_from_constant_string): Moved to fold-const.h.
1536 (fold_real_zero_addition_p): Moved to fold-const.h.
1537 (fold_single_bit_test): Moved to fold-const.h.
1538 (fold_strip_sign_ops): Moved to fold-const.h.
1539 (fold_ternary): Moved to fold-const.h.
1540 (fold_ternary_loc): Moved to fold-const.h.
1541 (fold_unary): Moved to tree-data-ref.h.
1542 (fold_unary_ignore_overflow): Moved to fold-const.h.
1543 (fold_unary_ignore_overflow_loc): Moved to fold-const.h.
1544 (fold_unary_loc): Moved to fold-const.h.
1545 (fold_unary_to_constant): Moved to fold-const.h.
1546 (fold_undefer_and_ignore_overflow_warnings): Moved to fold-const.h.
1547 (fold_undefer_overflow_warnings): Moved to fold-const.h.
1548 (folding_initializer): Moved to fold-const.h.
1549 (free_temp_slots): Moved to function.h.
1550 (generate_setjmp_warnings): Moved to function.h.
1551 (get_attribute_name): Moved to attribs.h.
1552 (get_identifier): Moved to stringpool.h.
1553 (get_identifier_with_length): Moved to stringpool.h.
1554 (get_inner_reference): Moved to tree.h.
1555 (gimple_alloca_call_p): Moved to calls.h.
1556 (gimplify_parameters): Moved to function.h.
1557 (highest_pow2_factor): Moved to expr.h.
1558 (indent_to): Moved to print-tree.h.
1559 (init_attributes): Moved to attribs.h.
1560 (init_dummy_function_start): Moved to function.h.
1561 (init_function_start): Moved to function.h.
1562 (init_inline_once): Moved to tree-inline.h.
1563 (init_object_sizes): Moved to tree-object-size.h.
1564 (init_temp_slots): Moved to function.h.
1565 (init_tree_optimization_optabs): Moved to optabs.h.
1566 (initialize_sizetypes): Moved to stor-layout.h.
1567 (initializer_constant_valid_for_bitfield_p): Moved to varasm.h.
1568 (initializer_constant_valid_p): Moved to varasm.h.
1569 (int_const_binop): Moved to fold-const.h.
1570 (internal_reference_types): Moved to stor-layout.h.
1571 (invert_tree_comparison): Moved to fold-const.h.
1572 (invert_truthvalue): Moved to fold-const.h.
1573 (invert_truthvalue_loc): Moved to fold-const.h.
1574 (is_tm_ending_fndecl): Moved to trans-mem.h.
1575 (is_tm_may_cancel_outer): Moved to trans-mem.h.
1576 (is_tm_pure): Moved to trans-mem.h.
1577 (is_tm_safe): Moved to trans-mem.h.
1578 (layout_decl): Moved to stor-layout.h.
1579 (layout_type): Moved to stor-layout.h.
1580 (lookup_attribute_spec): Moved to attribs.h.
1581 (make_accum_type): Moved to stor-layout.h.
1582 (make_decl_one_only): Moved to varasm.h.
1583 (make_decl_rtl): Moved to tree.h.
1584 (make_decl_rtl_for_debug): Moved to varasm.h.
1585 (make_fract_type): Moved to stor-layout.h.
1586 (make_or_reuse_sat_signed_accum_type): Moved to stor-layout.h.
1587 (make_or_reuse_sat_signed_fract_type): Moved to stor-layout.h.
1588 (make_or_reuse_sat_unsigned_accum_type): Moved to stor-layout.h.
1589 (make_or_reuse_sat_unsigned_fract_type): Moved to stor-layout.h.
1590 (make_or_reuse_signed_accum_type): Moved to stor-layout.h.
1591 (make_or_reuse_signed_fract_type): Moved to stor-layout.h.
1592 (make_or_reuse_unsigned_accum_type): Moved to stor-layout.h.
1593 (make_or_reuse_unsigned_fract_type): Moved to stor-layout.h.
1594 (make_range): Moved to fold-const.h.
1595 (make_range_step): Moved to fold-const.h.
1596 (make_sat_signed_accum_type): Moved to stor-layout.h.
1597 (make_sat_signed_fract_type): Moved to stor-layout.h.
1598 (make_sat_unsigned_accum_type): Moved to stor-layout.h.
1599 (make_sat_unsigned_fract_type): Moved to stor-layout.h.
1600 (make_signed_accum_type): Moved to stor-layout.h.
1601 (make_signed_fract_type): Moved to stor-layout.h.
1602 (make_signed_type): Moved to stor-layout.h.
1603 (make_unsigned_accum_type): Moved to stor-layout.h.
1604 (make_unsigned_fract_type): Moved to stor-layout.h.
1605 (make_unsigned_type): Moved to stor-layout.h.
1606 (mark_decl_referenced): Moved to varasm.h.
1607 (mark_referenced): Moved to varasm.h.
1608 (may_negate_without_overflow_p): Moved to fold-const.h.
1609 (maybe_get_identifier): Moved to stringpool.h.
1610 (merge_ranges): Moved to fold-const.h.
1611 (merge_weak): Moved to varasm.h.
1612 (mode_for_size_tree): Moved to stor-layout.h.
1613 (multiple_of_p): Moved to fold-const.h.
1614 (must_pass_in_stack_var_size): Moved to calls.h.
1615 (must_pass_in_stack_var_size_or_pad): Moved to calls.h.
1616 (native_encode_expr): Moved to fold-const.h.
1617 (native_interpret_expr): Moved to fold-const.h.
1618 (non_lvalue): Moved to fold-const.h.
1619 (non_lvalue_loc): Moved to fold-const.h.
1620 (normalize_offset): Moved to stor-layout.h.
1621 (normalize_rli): Moved to stor-layout.h.
1622 (notice_global_symbol): Moved to varasm.h.
1623 (omit_one_operand): Moved to fold-const.h.
1624 (omit_one_operand_loc): Moved to fold-const.h.
1625 (omit_two_operands): Moved to fold-const.h.
1626 (omit_two_operands_loc): Moved to fold-const.h.
1627 (operand_equal_p): Moved to tree-data-ref.h.
1628 (parse_input_constraint): Moved to stmt.h.
1629 (parse_output_constraint): Moved to stmt.h.
1630 (place_field): Moved to stor-layout.h.
1631 (pop_function_context): Moved to function.h.
1632 (pop_temp_slots): Moved to function.h.
1633 (pos_from_bit): Moved to stor-layout.h.
1634 (preserve_temp_slots): Moved to function.h.
1635 (print_node): Moved to print-tree.h.
1636 (print_node_brief): Moved to print-tree.h.
1637 (print_rtl): Moved to rtl.h.
1638 (process_pending_assemble_externals): Moved to varasm.h.
1639 (ptr_difference_const): Moved to fold-const.h.
1640 (push_function_context): Moved to function.h.
1641 (push_struct_function): Moved to function.h.
1642 (push_temp_slots): Moved to function.h.
1643 (record_tm_replacement): Moved to trans-mem.h.
1644 (relayout_decl): Moved to stor-layout.h.
1645 (resolve_asm_operand_names): Moved to stmt.h.
1646 (resolve_unique_section): Moved to varasm.h.
1647 (rli_size_so_far): Moved to stor-layout.h.
1648 (rli_size_unit_so_far): Moved to stor-layout.h.
1649 (round_down): Moved to fold-const.h.
1650 (round_down_loc): Moved to fold-const.h.
1651 (round_up): Moved to fold-const.h.
1652 (round_up_loc): Moved to fold-const.h.
1653 (set_decl_incoming_rtl): Moved to emit-rtl.h.
1654 (set_decl_rtl): Moved to tree.h.
1655 (set_min_and_max_values_for_integral_type): Moved to stor-layout.h.
1656 (set_user_assembler_name): Moved to varasm.h.
1657 (setjmp_call_p): Moved to calls.h.
1658 (size_binop): Moved to fold-const.h.
1659 (size_binop_loc): Moved to fold-const.h.
1660 (size_diffop): Moved to fold-const.h.
1661 (size_diffop_loc): Moved to fold-const.h.
1662 (size_int_kind): Moved to fold-const.h.
1663 (stack_protect_epilogue): Moved to function.h.
1664 (start_record_layout): Moved to stor-layout.h.
1665 (supports_one_only): Moved to varasm.h.
1666 (swap_tree_comparison): Moved to fold-const.h.
1667 (tm_malloc_replacement): Moved to trans-mem.h.
1668 (tree build_fold_addr_expr_loc): Moved to fold-const.h.
1669 (tree build_invariant_address): Moved to fold-const.h.
1670 (tree_binary_nonnegative_warnv_p): Moved to fold-const.h.
1671 (tree_binary_nonzero_warnv_p): Moved to fold-const.h.
1672 (tree_call_nonnegative_warnv_p): Moved to fold-const.h.
1673 (tree_expr_nonnegative_p): Moved to fold-const.h.
1674 (tree_expr_nonnegative_warnv_p): Moved to fold-const.h.
1675 (tree_output_constant_def): Moved to varasm.h.
1676 (tree_overlaps_hard_reg_set): Moved to stmt.h.
1677 (tree_single_nonnegative_warnv_p): Moved to fold-const.h.
1678 (tree_single_nonzero_warnv_p): Moved to fold-const.h.
1679 (tree_swap_operands_p): Moved to fold-const.h.
1680 (tree_unary_nonnegative_warnv_p): Moved to fold-const.h.
1681 (tree_unary_nonzero_warnv_p): Moved to fold-const.h.
1682 (update_alignment_for_field): Moved to stor-layout.h.
1683 (use_register_for_decl): Moved to function.h.
1684 (variable_size): Moved to rtl.h.
1685 (vector_type_mode): Moved to stor-layout.h.
1686 * cgraph.h: Corresponding changes.
1687 * emit-rtl.h: Corresponding changes.
1688 * expr.h: Corresponding changes.
1689 * function.h: Corresponding changes.
1690 * optabs.h: Corresponding changes.
1691 * trans-mem.h: Corresponding changes.
1692 Protect against multiple inclusion.
1693 * tree-inline.h: Corresponding changes.
1694 * tree-iterator.h: Corresponding changes.
1695 * tree-dfa.h: Include expr.h.
1696 * tree-ssanames.h: Include stringpool.h.
1697 * attribs.h: New file.
1698 * calls.h: New file.
1699 * fold-const.h: New file.
1700 * gcc-symtab.h: New file.
1701 * print-rtl.h: New file.
1702 * print-tree.h: New file.
1704 * stor-layout.h: New file.
1705 * strinpool.h: New file.
1706 * tree-nested.h: New file
1707 * tree-object-size.h: New file.
1708 * varasm.h: New file.
1710 2013-11-14 Diego Novillo <dnovillo@google.com>
1712 * alias.c: Include varasm.h.
1714 * asan.c: Include calls.h.
1715 Include stor-layout.h.
1717 * attribs.c: Include stringpool.h.
1719 Include stor-layout.h.
1720 * builtins.c: Include stringpool.h.
1721 Include stor-layout.h.
1724 Include tree-object-size.h.
1725 * calls.c: Include stor-layout.h.
1727 Include stringpool.h.
1729 * cfgexpand.c: Include stringpool.h.
1731 Include stor-layout.h.
1733 Include print-tree.h.
1734 * cgraph.c: Include varasm.h.
1736 Include print-tree.h.
1737 * cgraphclones.c: Include stringpool.h.
1740 Move inclusion of rtl.h earlier in the file.
1741 * cgraphunit.c: Include varasm.h.
1742 Include stor-layout.h.
1743 Include stringpool.h.
1744 * cilk-common.c: Include stringpool.h.
1745 Include stor-layout.h.
1746 * combine.c: Include stor-layout.h.
1747 * config/aarch64/aarch64-builtins.c: Include stor-layout.h.
1748 Include stringpool.h.
1750 * config/aarch64/aarch64.c: Include stringpool.h.
1751 Include stor-layout.h.
1754 * config/alpha/alpha.c: Include stor-layout.h.
1757 * config/arc/arc.c: Include varasm.h.
1758 Include stor-layout.h.
1759 Include stringpool.h.
1761 * config/arm/arm.c: Include stringpool.h.
1762 Include stor-layout.h.
1765 * config/avr/avr-c.c: Include stor-layout.h.
1766 * config/avr/avr-log.c: Include print-tree.h.
1767 * config/avr/avr.c: Include print-tree.h.
1769 Include stor-layout.h.
1770 Include stringpool.h.
1771 * config/bfin/bfin.c: Include varasm.h.
1773 * config/c6x/c6x.c: Include stor-layout.h.
1776 Include stringpool.h.
1777 * config/cr16/cr16.c: Include stor-layout.h.
1779 * config/cris/cris.c: Include varasm.h.
1780 Include stor-layout.h.
1783 * config/darwin.c: Include stringpool.h.
1785 Include stor-layout.h.
1786 * config/epiphany/epiphany.c: Include stor-layout.h.
1789 Include stringpool.h.
1790 * config/fr30/fr30.c: Include stor-layout.h.
1792 * config/frv/frv.c: Include varasm.h.
1793 Include stor-layout.h.
1794 Include stringpool.h.
1795 * config/h8300/h8300.c: Include stor-layout.h.
1798 Include stringpool.h.
1799 * config/i386/i386.c: Include stringpool.h.
1802 Include stor-layout.h.
1804 * config/i386/winnt-cxx.c: Include stringpool.h.
1806 * config/i386/winnt.c: Include stringpool.h.
1808 * config/ia64/ia64-c.c: Include stringpool.h.
1809 * config/ia64/ia64.c: Include stringpool.h.
1810 Include stor-layout.h.
1813 * config/iq2000/iq2000.c: Include stor-layout.h.
1816 * config/lm32/lm32.c: Include calls.h.
1817 * config/m32c/m32c.c: Include stor-layout.h.
1820 * config/m32r/m32r.c: Include stor-layout.h.
1822 Include stringpool.h.
1824 * config/m68k/m68k.c: Include calls.h.
1825 Include stor-layout.h.
1827 * config/mcore/mcore.c: Include stor-layout.h.
1829 Include stringpool.h.
1831 * config/mep/mep.c: Include varasm.h.
1833 Include stringpool.h.
1834 Include stor-layout.h.
1835 * config/microblaze/microblaze.c: Include varasm.h.
1836 Include stor-layout.h.
1838 * config/mips/mips.c: Include varasm.h.
1839 Include stringpool.h.
1840 Include stor-layout.h.
1842 * config/mmix/mmix.c: Include varasm.h.
1843 Include stor-layout.h.
1845 * config/mn10300/mn10300.c: Include stor-layout.h.
1848 * config/moxie/moxie.c: Include stor-layout.h.
1851 * config/msp430/msp430.c: Include stor-layout.h.
1853 * config/nds32/nds32.c: Include stor-layout.h.
1856 * config/pa/pa.c: Include stor-layout.h.
1857 Include stringpool.h.
1860 * config/pdp11/pdp11.c: Include stor-layout.h.
1863 * config/picochip/picochip.c: Include calls.h.
1864 Include stor-layout.h.
1865 Include stringpool.h.
1867 * config/rl78/rl78.c: Include varasm.h.
1868 Include stor-layout.h.
1870 * config/rs6000/rs6000-c.c: Include stor-layout.h.
1871 Include stringpool.h.
1872 * config/rs6000/rs6000.c: Include stringpool.h.
1873 Include stor-layout.h.
1875 Include print-tree.h.
1877 * config/rx/rx.c: Include varasm.h.
1878 Include stor-layout.h.
1880 * config/s390/s390.c: Include print-tree.h.
1881 Include stringpool.h.
1882 Include stor-layout.h.
1885 * config/score/score.c: Include stringpool.h.
1888 Include stor-layout.h.
1889 * config/sh/sh-c.c: Include stringpool.h.
1890 Include attribs.h.h.
1891 * config/sh/sh.c: Include stringpool.h.
1892 Include stor-layout.h.
1895 * config/sol2-c.c: Include stringpool.h.
1897 * config/sol2-cxx.c: Include stringpool.h.
1898 * config/sol2.c: Include stringpool.h.
1900 * config/sparc/sparc.c: Include stringpool.h.
1901 Include stor-layout.h.
1904 * config/spu/spu-c.c: Include stringpool.h.
1905 * config/spu/spu.c: Include stringpool.h.
1906 Include stor-layout.h.
1909 * config/stormy16/stormy16.c: Include stringpool.h.
1910 Include stor-layout.h.
1913 * config/tilegx/tilegx.c: Include stringpool.h.
1914 Include stor-layout.h.
1917 * config/tilepro/tilepro.c: Include stringpool.h.
1918 Include stor-layout.h.
1921 * config/v850/v850-c.c: Include stringpool.h.
1923 * config/v850/v850.c: Include stringpool.h.
1924 Include stor-layout.h.
1927 * config/vax/vax.c: Include calls.h.
1929 * config/vms/vms.c: Include stringpool.h.
1930 * config/vxworks.c: Include stringpool.h.
1931 * config/xtensa/xtensa.c: Include stringpool.h.
1932 Include stor-layout.h.
1935 * convert.c: Include stor-layout.h.
1936 * coverage.c: Include stringpool.h.
1937 Include stor-layout.h.
1938 * dbxout.c: Include varasm.h.
1939 Include stor-layout.h.
1940 * dojump.c: Include stor-layout.h.
1941 * dse.c: Include stor-layout.h.
1942 * dwarf2asm.c: Include stringpool.h.
1944 * dwarf2cfi.c: Include stor-layout.h.
1945 * dwarf2out.c: Include rtl.h.
1946 Include stringpool.h.
1947 Include stor-layout.h.
1951 Move inclusion of rtl.h earlier in the file.
1952 * emit-rtl.c: Include varasm.h.
1953 * except.c: Include stringpool.h.
1954 Include stor-layout.h.
1955 * explow.c: Include stor-layout.h.
1956 * expmed.c: Include stor-layout.h.
1957 * expr.c: Include stringpool.h.
1958 Include stor-layout.h.
1961 * final.c: Include varasm.h.
1962 * fold-const.c: Include stor-layout.h.
1964 Include tree-iterator.h.
1965 * function.c: Include stor-layout.h.
1967 Include stringpool.h.
1968 * genattrtab.c (write_header): Emit includes for varasm.h,
1969 stor-layout.h and calls.h.
1970 * genautomata.c (main): Likewise.
1971 * genemit.c: Likewise.
1972 * genopinit.c: Likewise.
1973 * genoutput.c (output_prologue): Likewise.
1974 * genpeep.c: Likewise.
1975 * genpreds.c (write_insn_preds_c): Likewise.
1976 * gengtype.c (open_base_files): Add stringpool.h.
1977 * gimple-expr.c: Include stringpool.h.
1978 Include stor-layout.h.
1979 * gimple-fold.c: Include stringpool.h.
1982 Include stor-layout.h.
1983 * gimple-low.c: Include tree-nested.h.
1985 * gimple-pretty-print.c: Include stringpool.h.
1986 * gimple-ssa-strength-reduction.c: Include stor-layout.h.
1988 * gimple-walk.c: Include stmt.h.
1989 * gimple.c: Include calls.h.
1991 Include stor-layout.h.
1992 * gimplify.c: Include stringpool.h.
1995 Include stor-layout.h.
1997 Include print-tree.h.
1999 * gimplify-me.c: Include stmt.h
2000 Include stor-layout.h
2001 * internal-fn.c: Include stor-layout.h.
2002 * ipa-devirt.c: Include print-tree.h.
2004 * ipa-inline-analysis.c: Include stor-layout.h.
2005 Include stringpool.h.
2006 Include print-tree.h.
2007 * ipa-inline.c: Include trans-mem.h.
2009 * ipa-prop.c: Include expr.h.
2010 Include stor-layout.h.
2011 Include print-tree.h.
2012 * ipa-pure-const.c: Include print-tree.h.
2014 * ipa-reference.c: Include calls.h.
2015 * ipa-split.c: Include stringpool.h.
2018 * ipa.c: Include calls.h.
2019 Include stringpool.h.
2020 * langhooks.c: Include stringpool.h.
2022 * lto-cgraph.c: Include stringpool.h.
2023 * lto-streamer-in.c: Include stringpool.h.
2024 * lto-streamer-out.c: Include stor-layout.h.
2025 Include stringpool.h.
2026 * omp-low.c: Include stringpool.h.
2027 Include stor-layout.h.
2029 * optabs.c: Include stor-layout.h.
2030 Include stringpool.h.
2032 * passes.c: Include varasm.h.
2033 * predict.c: Include calls.h.
2034 * print-rtl.c: Include print-tree.h.
2035 * print-tree.c: Include varasm.h.
2036 Include print-rtl.h.
2037 Include stor-layout.h.
2038 * realmpfr.c: Include stor-layout.h.
2039 * reg-stack.c: Include varasm.h.
2040 * sdbout.c: Include varasm.h.
2041 Include stor-layout.h.
2042 * simplify-rtx.c: Include varasm.h.
2043 * stmt.c: Include varasm.h.
2044 Include stor-layout.h.
2045 * stor-layout.c: Include stor-layout.h.
2046 Include stringpool.h.
2048 Include print-tree.h.
2049 * symtab.c: Include rtl.h.
2050 Include print-tree.h.
2054 * targhooks.c: Include stor-layout.h.
2056 * toplev.c: Include varasm.h.
2057 Include tree-inline.h.
2058 * trans-mem.c: Include calls.h.
2062 * tree-affine.c: Include expr.h.
2063 * tree-browser.c: Include print-tree.h.
2064 * tree-call-cdce.c: Include stor-layout.h.
2065 * tree-cfg.c: Include trans-mem.h.
2066 Include stor-layout.h.
2067 Include print-tree.h.
2068 * tree-complex.c: Include stor-layout.h.
2069 * tree-data-ref.c: Include expr.h.
2070 * tree-dfa.c: Include stor-layout.h.
2071 * tree-eh.c: Include expr.h.
2073 * tree-emutls.c: Include stor-layout.h.
2075 * tree-if-conv.c: Include stor-layout.h.
2076 * tree-inline.c: Include stor-layout.h.
2078 * tree-loop-distribution.c: Include stor-layout.h.
2079 * tree-nested.c: Include stringpool.h.
2080 Include stor-layout.h.
2081 * tree-object-size.c: Include tree-object-size.h.
2082 * tree-outof-ssa.c: Include stor-layout.h.
2083 * tree-parloops.c: Include stor-layout.h.
2084 Include tree-nested.h.
2085 * tree-pretty-print.c: Include stor-layout.h.
2087 * tree-profile.c: Include varasm.h.
2088 Include tree-nested.h.
2089 * tree-scalar-evolution.c: Include expr.h.
2090 * tree-sra.c: Include stor-layout.h.
2091 * tree-ssa-address.c: Include stor-layout.h.
2092 * tree-ssa-ccp.c: Include stor-layout.h.
2093 * tree-ssa-dce.c: Include calls.h.
2094 * tree-ssa-dom.c: Include stor-layout.h.
2095 * tree-ssa-forwprop.c: Include stor-layout.h.
2096 * tree-ssa-ifcombine.c: Include stor-layout.h.
2097 * tree-ssa-loop-ivopts.c: Include stor-layout.h.
2098 * tree-ssa-loop-niter.c: Include calls.h.
2100 * tree-ssa-loop-prefetch.c: Include stor-layout.h.
2101 * tree-ssa-math-opts.c: Include stor-layout.h.
2102 * tree-ssa-operands.c: Include stmt.h.
2103 Include print-tree.h.
2104 * tree-ssa-phiopt.c: Include stor-layout.h.
2105 * tree-ssa-reassoc.c: Include stor-layout.h.
2106 * tree-ssa-sccvn.c: Include stor-layout.h.
2107 * tree-ssa-sink.c: Include stor-layout.h.
2108 * tree-ssa-strlen.c: Include stor-layout.h.
2109 * tree-ssa-structalias.c: Include stor-layout.h.
2111 * tree-ssa-tail-merge.c: Include stor-layout.h.
2112 Include trans-mem.h.
2113 * tree-ssa-uncprop.c: Include stor-layout.h.
2114 * tree-ssa.c: Include stor-layout.h.
2115 * tree-ssanames.c: Include stor-layout.h.
2116 * tree-streamer-in.c: Include stringpool.h.
2117 * tree-streamer-out.c: Include stor-layout.h.
2118 * tree-switch-conversion.c: Include varasm.h.
2119 Include stor-layout.h.
2120 * tree-tailcall.c: Include stor-layout.h.
2121 * tree-vect-data-refs.c: Include stor-layout.h.
2122 * tree-vect-generic.c: Include stor-layout.h.
2123 * tree-vect-loop.c: Include stor-layout.h.
2124 * tree-vect-patterns.c: Include stor-layout.h.
2125 * tree-vect-slp.c: Include stor-layout.h.
2126 * tree-vect-stmts.c: Include stor-layout.h.
2127 * tree-vectorizer.c: Include stor-layout.h.
2128 * tree-vrp.c: Include stor-layout.h.
2130 * tree.c: Include stor-layout.h.
2134 * tsan.c: Include expr.h.
2135 * ubsan.c: Include stor-layout.h.
2136 Include stringpool.h.
2137 * value-prof.c: Include tree-nested.h.
2139 * var-tracking.c: Include varasm.h.
2140 Include stor-layout.h.
2141 * varasm.c: Include stor-layout.h.
2142 Include stringpool.h.
2143 Include gcc-symtab.h.
2145 * varpool.c: Include varasm.h.
2146 * vmsdbgout.c: Include varasm.h.
2147 * xcoffout.c: Include varasm.h.
2149 2013-11-14 Joern Rennecke <joern.rennecke@embecosm.com>
2151 * config/arc/arc.md (doloop_begin_i): Remove extra alignment;
2154 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2156 * config/rs6000/sysv4le.h (LINUX64_DEFAULT_ABI_ELFv2): Define.
2158 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2159 Alan Modra <amodra@gmail.com>
2161 * config/rs6000/rs6000.h (RS6000_SAVE_AREA): Handle ABI_ELFv2.
2162 (RS6000_SAVE_TOC): Remove.
2163 (RS6000_TOC_SAVE_SLOT): New macro.
2164 * config/rs6000/rs6000.c (rs6000_parm_offset): New function.
2165 (rs6000_parm_start): Use it.
2166 (rs6000_function_arg_advance_1): Likewise.
2167 (rs6000_emit_prologue): Use RS6000_TOC_SAVE_SLOT.
2168 (rs6000_emit_epilogue): Likewise.
2169 (rs6000_call_aix): Likewise.
2170 (rs6000_output_function_prologue): Do not save/restore r11
2171 around calling _mcount for ABI_ELFv2.
2173 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2174 Alan Modra <amodra@gmail.com>
2176 * config/rs6000/rs6000-protos.h (rs6000_reg_parm_stack_space):
2178 * config/rs6000/rs6000.h (RS6000_REG_SAVE): Remove.
2179 (REG_PARM_STACK_SPACE): Call rs6000_reg_parm_stack_space.
2180 * config/rs6000/rs6000.c (rs6000_parm_needs_stack): New function.
2181 (rs6000_function_parms_need_stack): Likewise.
2182 (rs6000_reg_parm_stack_space): Likewise.
2183 (rs6000_function_arg): Do not replace BLKmode by Pmode when
2184 returning a register argument.
2186 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2187 Michael Gschwind <mkg@us.ibm.com>
2189 * config/rs6000/rs6000.h (FP_ARG_MAX_RETURN): New macro.
2190 (ALTIVEC_ARG_MAX_RETURN): Likewise.
2191 (FUNCTION_VALUE_REGNO_P): Use them.
2192 * config/rs6000/rs6000.c (TARGET_RETURN_IN_MSB): Define.
2193 (rs6000_return_in_msb): New function.
2194 (rs6000_return_in_memory): Handle ELFv2 homogeneous aggregates.
2195 Handle aggregates of up to 16 bytes for ELFv2.
2196 (rs6000_function_value): Handle ELFv2 homogeneous aggregates.
2198 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2199 Michael Gschwind <mkg@us.ibm.com>
2201 * config/rs6000/rs6000.h (AGGR_ARG_NUM_REG): Define.
2202 * config/rs6000/rs6000.c (rs6000_aggregate_candidate): New function.
2203 (rs6000_discover_homogeneous_aggregate): Likewise.
2204 (rs6000_function_arg_boundary): Handle homogeneous aggregates.
2205 (rs6000_function_arg_advance_1): Likewise.
2206 (rs6000_function_arg): Likewise.
2207 (rs6000_arg_partial_bytes): Likewise.
2208 (rs6000_psave_function_arg): Handle BLKmode arguments.
2210 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2211 Michael Gschwind <mkg@us.ibm.com>
2213 * config/rs6000/rs6000.h (AGGR_ARG_NUM_REG): Define.
2214 * config/rs6000/rs6000.c (rs6000_aggregate_candidate): New function.
2215 (rs6000_discover_homogeneous_aggregate): Likewise.
2216 (rs6000_function_arg_boundary): Handle homogeneous aggregates.
2217 (rs6000_function_arg_advance_1): Likewise.
2218 (rs6000_function_arg): Likewise.
2219 (rs6000_arg_partial_bytes): Likewise.
2220 (rs6000_psave_function_arg): Handle BLKmode arguments.
2222 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2224 * config/rs6000/rs6000.c (machine_function): New member
2226 (rs6000_emit_prologue): Set r2_setup_needed if necessary.
2227 (rs6000_output_mi_thunk): Set r2_setup_needed.
2228 (rs6000_output_function_prologue): Output global entry point
2229 prologue and local entry point marker if needed for ABI_ELFv2.
2230 Output -mprofile-kernel code here.
2231 (output_function_profiler): Do not output -mprofile-kernel
2232 code here; moved to rs6000_output_function_prologue.
2233 (rs6000_file_start): Output ".abiversion 2" for ABI_ELFv2.
2235 (rs6000_emit_move): Do not handle dot symbols for ABI_ELFv2.
2236 (rs6000_output_function_entry): Likewise.
2237 (rs6000_assemble_integer): Likewise.
2238 (rs6000_elf_encode_section_info): Likewise.
2239 (rs6000_elf_declare_function_name): Do not create dot symbols
2240 or .opd section for ABI_ELFv2.
2242 (rs6000_trampoline_size): Update for ABI_ELFv2 trampolines.
2243 (rs6000_trampoline_init): Likewise.
2244 (rs6000_elf_file_end): Call file_end_indicate_exec_stack for ABI_ELFv2.
2246 (rs6000_call_aix): Handle ELFv2 indirect calls. Do not check
2247 for function descriptors in ABI_ELFv2.
2249 * config/rs6000/rs6000.md ("*call_indirect_aix<mode>"): Support
2250 on ABI_AIX only, not ABI_ELFv2.
2251 ("*call_value_indirect_aix<mode>"): Likewise.
2252 ("*call_indirect_elfv2<mode>"): New pattern.
2253 ("*call_value_indirect_elfv2<mode>"): Likewise.
2255 * config/rs6000/predicates.md ("symbol_ref_operand"): Do not
2256 check for function descriptors in ABI_ELFv2.
2257 ("current_file_function_operand"): Likewise.
2259 * config/rs6000/ppc-asm.h [__powerpc64__ && _CALL_ELF == 2]:
2261 (FUNC_NAME): Define ELFv2 variant.
2262 (JUMP_TARGET): Likewise.
2263 (FUNC_START): Likewise.
2264 (HIDDEN_FUNC): Likewise.
2265 (FUNC_END): Likeiwse.
2267 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2269 * config.gcc [powerpc*-*-* | rs6000-*-*]: Support --with-abi=elfv1
2270 and --with-abi=elfv2.
2271 * config/rs6000/option-defaults.h (OPTION_DEFAULT_SPECS): Add "abi".
2272 * config/rs6000/rs6000.opt (mabi=elfv1): New option.
2273 (mabi=elfv2): Likewise.
2274 * config/rs6000/rs6000-opts.h (enum rs6000_abi): Add ABI_ELFv2.
2275 * config/rs6000/linux64.h (DEFAULT_ABI): Do not hard-code to AIX_ABI
2277 (ELFv2_ABI_CHECK): New macro.
2278 (SUBSUBTARGET_OVERRIDE_OPTIONS): Use it to decide whether to set
2279 rs6000_current_abi to ABI_AIX or ABI_ELFv2.
2280 (GLIBC_DYNAMIC_LINKER64): Support ELFv2 ld.so version.
2281 * config/rs6000/rs6000-c.c (rs6000_cpu_cpp_builtins): Predefine
2282 _CALL_ELF and __STRUCT_PARM_ALIGN__ if appropriate.
2284 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Handle ABI_ELFv2.
2285 (debug_stack_info): Likewise.
2286 (rs6000_file_start): Treat ABI_ELFv2 the same as ABI_AIX.
2287 (rs6000_legitimize_tls_address): Likewise.
2288 (rs6000_conditional_register_usage): Likewise.
2289 (rs6000_emit_move): Likewise.
2290 (init_cumulative_args): Likewise.
2291 (rs6000_function_arg_advance_1): Likewise.
2292 (rs6000_function_arg): Likewise.
2293 (rs6000_arg_partial_bytes): Likewise.
2294 (rs6000_output_function_entry): Likewise.
2295 (rs6000_assemble_integer): Likewise.
2296 (rs6000_savres_strategy): Likewise.
2297 (rs6000_stack_info): Likewise.
2298 (rs6000_function_ok_for_sibcall): Likewise.
2299 (rs6000_emit_load_toc_table): Likewise.
2300 (rs6000_savres_routine_name): Likewise.
2301 (ptr_regno_for_savres): Likewise.
2302 (rs6000_emit_prologue): Likewise.
2303 (rs6000_emit_epilogue): Likewise.
2304 (rs6000_output_function_epilogue): Likewise.
2305 (output_profile_hook): Likewise.
2306 (output_function_profiler): Likewise.
2307 (rs6000_trampoline_size): Likewise.
2308 (rs6000_trampoline_init): Likewise.
2309 (rs6000_elf_output_toc_section_asm_op): Likewise.
2310 (rs6000_elf_encode_section_info): Likewise.
2311 (rs6000_elf_reloc_rw_mask): Likewise.
2312 (rs6000_elf_declare_function_name): Likewise.
2313 (rs6000_function_arg_boundary): Treat ABI_ELFv2 the same as ABI_AIX,
2314 except that rs6000_compat_align_parm is always assumed false.
2315 (rs6000_gimplify_va_arg): Likewise.
2316 (rs6000_call_aix): Update comment.
2317 (rs6000_sibcall_aix): Likewise.
2318 * config/rs6000/rs6000.md ("tls_gd_aix<TLSmode:tls_abi_suffix>"):
2319 Treat ABI_ELFv2 the same as ABI_AIX.
2320 ("*tls_gd_call_aix<TLSmode:tls_abi_suffix>"): Likewise.
2321 ("tls_ld_aix<TLSmode:tls_abi_suffix>"): Likewise.
2322 ("*tls_ld_call_aix<TLSmode:tls_abi_suffix>"): Likewise.
2323 ("load_toc_aix_si"): Likewise.
2324 ("load_toc_aix_di"): Likewise.
2326 ("call_value"): Likewise.
2327 ("*call_local_aix<mode>"): Likewise.
2328 ("*call_value_local_aix<mode>"): Likewise.
2329 ("*call_nonlocal_aix<mode>"): Likewise.
2330 ("*call_value_nonlocal_aix<mode>"): Likewise.
2331 ("*call_indirect_aix<mode>"): Likewise.
2332 ("*call_value_indirect_aix<mode>"): Likewise.
2333 ("sibcall"): Likewise.
2334 ("sibcall_value"): Likewise.
2335 ("*sibcall_aix<mode>"): Likewise.
2336 ("*sibcall_value_aix<mode>"): Likewise.
2337 * config/rs6000/predicates.md ("symbol_ref_operand"): Likewise.
2338 ("current_file_function_operand"): Likewise.
2340 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2342 * config/rs6000/rs6000.c (rs6000_arg_partial_bytes): Simplify logic
2343 by making use of the fact that for vector / floating point arguments
2344 passed both in VRs/FPRs and in the fixed parameter area, the partial
2345 bytes mechanism is in fact not used.
2347 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2349 * config/rs6000/rs6000.c (rs6000_psave_function_arg): New function.
2350 (rs6000_finish_function_arg): Likewise.
2351 (rs6000_function_arg): Use rs6000_psave_function_arg and
2352 rs6000_finish_function_arg to handle both vector and floating
2353 point arguments that are also passed in GPRs / the stack.
2355 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2357 * config/rs6000/rs6000.c (USE_FP_FOR_ARG_P): Remove TYPE argument.
2358 (USE_ALTIVEC_FOR_ARG_P): Likewise.
2359 (rs6000_darwin64_record_arg_advance_recurse): Update uses.
2360 (rs6000_function_arg_advance_1):Likewise.
2361 (rs6000_darwin64_record_arg_recurse): Likewise.
2362 (rs6000_function_arg): Likewise.
2363 (rs6000_arg_partial_bytes): Likewise.
2365 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2367 * config/rs6000/rs6000.c (rs6000_option_override_internal): Replace
2368 "DEFAULT_ABI != ABI_AIX" test by testing for ABI_V4 or ABI_DARWIN.
2369 (rs6000_savres_strategy): Likewise.
2370 (rs6000_return_addr): Likewise.
2371 (rs6000_emit_load_toc_table): Replace "DEFAULT_ABI != ABI_AIX" by
2372 testing for ABI_V4 (since ABI_DARWIN is impossible here).
2373 (rs6000_emit_prologue): Likewise.
2374 (legitimate_lo_sum_address_p): Simplify DEFAULT_ABI test.
2375 (rs6000_elf_declare_function_name): Remove duplicated test.
2376 * config/rs6000/rs6000.md ("load_toc_v4_PIC_1"): Explicitly test
2377 for ABI_V4 (instead of "DEFAULT_ABI != ABI_AIX" test).
2378 ("load_toc_v4_PIC_1_normal"): Likewise.
2379 ("load_toc_v4_PIC_1_476"): Likewise.
2380 ("load_toc_v4_PIC_1b"): Likewise.
2381 ("load_toc_v4_PIC_1b_normal"): Likewise.
2382 ("load_toc_v4_PIC_1b_476"): Likewise.
2383 ("load_toc_v4_PIC_2"): Likewise.
2384 ("load_toc_v4_PIC_3b"): Likewise.
2385 ("load_toc_v4_PIC_3c"): Likewise.
2386 * config/rs6000/rs6000.h (RS6000_REG_SAVE): Simplify DEFAULT_ABI test.
2387 (RS6000_SAVE_AREA): Likewise.
2388 (FP_ARG_MAX_REG): Likewise.
2389 (RETURN_ADDRESS_OFFSET): Likewise.
2390 * config/rs6000/sysv.h (TARGET_TOC): Test for ABI_V4 instead
2392 (SUBTARGET_OVERRIDE_OPTIONS): Likewise.
2393 (MINIMAL_TOC_SECTION_ASM_OP): Likewise.
2395 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2397 * config/rs6000/rs6000.c (rs6000_call_indirect_aix): Rename to ...
2398 (rs6000_call_aix): ... this. Handle both direct and indirect calls.
2399 Create call insn directly instead of via various gen_... routines.
2400 Mention special registers used by the call in CALL_INSN_FUNCTION_USAGE.
2401 (rs6000_sibcall_aix): New function.
2402 * config/rs6000/rs6000.md (TOC_SAVE_OFFSET_32BIT): Remove.
2403 (TOC_SAVE_OFFSET_64BIT): Likewise.
2404 (AIX_FUNC_DESC_TOC_32BIT): Likewise.
2405 (AIX_FUNC_DESC_TOC_64BIT): Likewise.
2406 (AIX_FUNC_DESC_SC_32BIT): Likewise.
2407 (AIX_FUNC_DESC_SC_64BIT): Likewise.
2408 ("call" expander): Call rs6000_call_aix.
2409 ("call_value" expander): Likewise.
2410 ("call_indirect_aix<ptrsize>"): Replace this pattern ...
2411 ("call_indirect_aix<ptrsize>_nor11"): ... and this pattern ...
2412 ("*call_indirect_aix<mode>"): ... by this insn pattern.
2413 ("call_value_indirect_aix<ptrsize>"): Replace this pattern ...
2414 ("call_value_indirect_aix<ptrsize>_nor11"): ... and this pattern ...
2415 ("*call_value_indirect_aix<mode>"): ... by this insn pattern.
2416 ("*call_nonlocal_aix32", "*call_nonlocal_aix64"): Replace by ...
2417 ("*call_nonlocal_aix<mode>"): ... this pattern.
2418 ("*call_value_nonlocal_aix32", "*call_value_nonlocal_aix64"): Replace
2419 ("*call_value_nonlocal_aix<mode>"): ... by this pattern.
2420 ("*call_local_aix<mode>"): New insn pattern.
2421 ("*call_value_local_aix<mode>"): Likewise.
2422 ("sibcall" expander): Call rs6000_sibcall_aix.
2423 ("sibcall_value" expander): Likewise. Move earlier in file.
2424 ("*sibcall_nonlocal_aix<mode>"): Replace by ...
2425 ("*sibcall_aix<mode>"): ... this pattern.
2426 ("*sibcall_value_nonlocal_aix<mode>"): Replace by ...
2427 ("*sibcall_value_aix<mode>"): ... this pattern.
2428 * config/rs6000/rs6000-protos.h (rs6000_call_indirect_aix): Remove.
2429 (rs6000_call_aix): Add prototype.
2430 (rs6000_sibcall_aix): Likewise.
2432 2013-11-14 Jakub Jelinek <jakub@redhat.com>
2435 * asan.c (asan_emit_stack_protection): Ensure -fsection-anchors
2436 isn't confused by the artificial decl.
2438 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2440 * config/rs6000/rs6000.c (rs6000_emit_prologue): Do not place a
2441 RTX_FRAME_RELATED_P marker on the UNSPEC_MOVESI_FROM_CR insn.
2442 Instead, add USEs of all modified call-saved CR fields to the
2443 insn storing the result to the stack slot, and provide an
2444 appropriate REG_FRAME_RELATED_EXPR for that insn.
2445 * config/rs6000/rs6000.md ("*crsave"): New insn pattern.
2446 * config/rs6000/predicates.md ("crsave_operation"): New predicate.
2448 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2449 Alan Modra <amodra@gmail.com>
2451 * function.c (assign_parms): Use all.reg_parm_stack_space instead
2452 of re-evaluating REG_PARM_STACK_SPACE target macro.
2453 (locate_and_pad_parm): New parameter REG_PARM_STACK_SPACE. Use it
2454 instead of evaluating target macro REG_PARM_STACK_SPACE every time.
2455 (assign_parm_find_entry_rtl): Update call.
2456 * calls.c (initialize_argument_information): Update call.
2457 (emit_library_call_value_1): Likewise.
2458 * expr.h (locate_and_pad_parm): Update prototype.
2460 2013-11-14 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2462 * calls.c (store_unaligned_arguments_into_pseudos): Skip PARALLEL
2465 2013-11-14 DJ Delorie <dj@redhat.com>
2467 * config/rx/rx.c (rx_mode_dependent_address_p): Allow offsets up
2470 2013-11-14 Jeff Law <law@redhat.com>
2472 * tree-ssa-threadedge.c (thread_through_normal_block): Only push the
2473 EDGE_START_JUMP_THREAD marker if the jump threading path is empty.
2475 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
2477 * doc/invoke.texi: Update documentation for AArch64's -mcpu
2480 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
2482 * config/aarch64/aarch64-cores.def (example-1): Remove.
2483 (example-2): Likewise.
2484 * config/aarch64/aarch64-tune.md: Regenerate.
2485 * config/aarch64/aarch64.md: Do not include "large.md" or "small.md".
2486 (generic_sched): Remove "large", "small".
2487 * config/aarch64/large.md: Delete.
2488 * config/aarch64/small.md: Delete.
2490 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
2492 * config/aarch64/aarch64-cores.def (cortex-a57): Tune for cortexa15.
2493 * config/aarch64/aarch64-tune.md: Regenerate.
2494 * config/aarch64/aarch64.md: Include cortex-a15 pipeline model.
2495 (generic_sched): "no" if we are tuning for cortexa15.
2496 * config/arm/cortex-a15.md: Include cortex-a15-neon.md by
2499 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
2501 * config/aarch64/aarch64-arches.def (armv8-a): Tune for cortex-a53.
2502 * config/aarch64/aarch64.md: Do not include aarch64-generic.md.
2503 * config/aarch64/aarch64.c (aarch64_tune): Initialize to cortexa53.
2504 (all_cores): Use cortexa53 when tuning for "generic".
2505 (aarch64_override_options): Fix comment.
2506 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Set to cortexa53.
2507 * config/aarch64/aarch64-generic.md: Delete.
2509 2013-11-14 James Greenhalgh <james.greenhalgh@arm.com>
2511 * config/aarch64/aarch64.c (all_architectures): Remove "generic".
2513 2013-11-14 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2515 * config/aarch64/aarch64.c: Include aarch-cost-tables.h.
2516 (generic_rtx_cost_table): Remove.
2517 (aarch64_rtx_costs): Use fields from cpu_cost_table.
2518 * config/aarch64/aarch64-protos.h (tune_params): Use cpu_cost_table for
2520 (cpu_rtx_cost_table): Remove.
2522 2013-11-14 Julian Brown <julian@codesourcery.com>
2523 Joey Ye <joey.ye@arm.com>
2525 * config/arm/arm.c (arm_cortex_m_branch_cost): New.
2526 (arm_v7m_tune): New.
2527 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune, arm_9e_tune,
2528 arm_v6t2_tune, arm_cortex_tune, arm_cortex_a15_tune,
2529 arm_cortex_a5_tune, arm_v6m_tune): Add comments for Sched adj cost.
2530 * config/arm/arm-cores.def (cortex-m4, cortex-m3): Use arm_v7m_tune.
2532 2013-11-14 Kirill Yukhin <kirill.yukhin@intel.com>
2535 * config/ia64/ia64.c (ia64_split_tmode_move): Relax `dead'
2538 2013-11-14 Jakub Jelinek <jakub@redhat.com>
2539 Uros Bizjak <ubizjak@gmail.com>
2542 * config/i386/i386.md (*anddi_2): Only allow CCZmode if
2543 operands[2] satisfies_constraint_Z that might have bit 31 set.
2545 2013-11-13 Jeff Law <law@redhat.com>
2547 PR tree-optimization/59102
2548 * gimple-ssa-isolate-paths.c
2549 (insert_trap_and_remove_trailing_statments): Ensure STMT is a
2550 gimple assignment before looking at gimple_assign_lhs.
2552 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
2554 * ira.c: Add comment about threads at the top of file.
2556 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
2558 * ira-color.c (coalesce_allocnos): Don't allocate and free
2561 2013-11-14 Tom de Vries <tom@codesourcery.com>
2563 * tree-ssa-tail-merge.c (gimple_equal_p): Add test for structural
2564 equality for GIMPLE_ASSIGN.
2566 2013-11-14 Tom de Vries <tom@codesourcery.com>
2568 * tree-ssa-tail-merge.c (gimple_operand_equal_value_p): Factor new
2570 (gimple_equal_p): ... here.
2572 2013-11-14 Tom de Vries <tom@codesourcery.com>
2574 * trans-mem.c (is_tm_ending): New function.
2575 * gimple.h (is_tm_ending): Declare.
2576 * tree-ssa-tail-merge.c (gimple_equal_p): Remove test on
2578 (find_duplicate): Use is_tm_ending instead of is_tm_ending_fndecl.
2580 2013-11-14 Tom de Vries <tom@codesourcery.com>
2582 * tree-ssa-tail-merge.c (gimple_equal_p): Remove equal variable.
2584 2013-11-13 Andrew MacLeod <amacleod@redhat.com>
2586 * gimple-walk.h: New File. Relocate prototypes from gimple.h.
2587 (struct walk_stmt_info): Relocate here from gimple.h.
2588 * gimple-iterator.h: New File. Relocate prototypes from gimple.h.
2589 (struct gimple_stmt_iterator_d): Relocate here from gimple.h.
2590 (gsi_start_1, gsi_none, gsi_start_bb, gsi_last_1, gsi_last_bb,
2591 gsi_end_p, gsi_one_before_end_p, gsi_next, gsi_prev, gsi_stmt,
2592 gsi_after_labels, gsi_next_nondebug, gsi_prev_nondebug,
2593 gsi_start_nondebug_bb, gsi_start_nondebug_after_labels_bb,
2594 gsi_last_nondebug_bb, gsi_bb, gsi_seq): Relocate here from gimple.h.
2595 * gimple.h (struct gimple_stmt_iterator_d): Move to gimple-iterator.h.
2596 (gsi_start_1, gsi_none, gsi_start_bb, gsi_last_1, gsi_last_bb,
2597 gsi_end_p, gsi_one_before_end_p, gsi_next, gsi_prev, gsi_stmt,
2598 gsi_after_labels, gsi_next_nondebug, gsi_prev_nondebug,
2599 gsi_start_nondebug_bb, gsi_start_nondebug_after_labels_bb,
2600 gsi_last_nondebug_bb, gsi_bb, gsi_seq): Move to gimple-iterator.h.
2601 (struct walk_stmt_info): Move to gimple-walk.h.
2602 (gimple_seq_set_location): Move to gimple.c
2603 * gimple-walk.c: New File.
2604 (walk_gimple_seq_mod, walk_gimple_seq, walk_gimple_asm, walk_gimple_op,
2605 walk_gimple_stmt, get_base_loadstore, walk_stmt_load_store_addr_ops,
2606 walk_stmt_load_store_ops): Relocate here from gimple.c.
2607 * gimple-iterator.c: Include gimple-iterator.h.
2608 * gimple.c (walk_gimple_seq_mod, walk_gimple_seq, walk_gimple_asm,
2609 walk_gimple_op, walk_gimple_stmt, get_base_loadstore,
2610 walk_stmt_load_store_addr_ops, walk_stmt_load_store_ops): Move to
2612 (gimple_seq_set_location): Relocate from gimple.h.
2613 * tree-phinodes.h (set_phi_nodes): Move to tree-phinodes.c.
2614 * tree-phinodes.c (set_phi_nodes): Relocate from tree-phinodes.h.
2615 * gengtype.c (open_base_files): Add gimple-iterator.h to include list.
2616 * Makefile.in (OBJS): Add gimple-walk.o
2617 * asan.c: Update Include list as required for gimple-iterator.h and
2619 * cfgexpand.c: Likewise.
2620 * cfgloop.c: Likewise.
2621 * cfgloopmanip.c: Likewise.
2622 * cgraph.c: Likewise.
2623 * cgraphbuild.c: Likewise.
2624 * cgraphunit.c: Likewise.
2625 * gimple-fold.c: Likewise.
2626 * gimple-low.c: Likewise.
2627 * gimple-pretty-print.c: Likewise.
2628 * gimple-ssa-isolate-paths.c: Likewise.
2629 * gimple-ssa-strength-reduction.c: Likewise.
2630 * gimple-streamer-in.c: Likewise.
2631 * gimple-streamer-out.c: Likewise.
2632 * gimplify.c: Likewise.
2633 * graphite-blocking.c: Likewise.
2634 * graphite-clast-to-gimple.c: Likewise.
2635 * graphite-dependences.c: Likewise.
2636 * graphite-interchange.c: Likewise.
2637 * graphite-optimize-isl.c: Likewise.
2638 * graphite-poly.c: Likewise.
2639 * graphite-scop-detection.c: Likewise.
2640 * graphite-sese-to-poly.c: Likewise.
2641 * graphite.c: Likewise.
2642 * ipa-inline-analysis.c: Likewise.
2643 * ipa-profile.c: Likewise.
2644 * ipa-prop.c: Likewise.
2645 * ipa-pure-const.c: Likewise.
2646 * ipa-split.c: Likewise.
2647 * lto-streamer-in.c: Likewise.
2648 * lto-streamer-out.c: Likewise.
2649 * omp-low.c: Likewise.
2650 * predict.c: Likewise.
2651 * profile.c: Likewise.
2653 * tracer.c: Likewise.
2654 * trans-mem.c: Likewise.
2655 * tree-call-cdce.c: Likewise.
2656 * tree-cfg.c: Likewise.
2657 * tree-cfgcleanup.c: Likewise.
2658 * tree-complex.c: Likewise.
2659 * tree-data-ref.c: Likewise.
2660 * tree-dfa.c: Likewise.
2661 * tree-eh.c: Likewise.
2662 * tree-emutls.c: Likewise.
2663 * tree-if-conv.c: Likewise.
2664 * tree-inline.c: Likewise.
2665 * tree-into-ssa.c: Likewise.
2666 * tree-loop-distribution.c: Likewise.
2667 * tree-nested.c: Likewise.
2668 * tree-nrv.c: Likewise.
2669 * tree-object-size.c: Likewise.
2670 * tree-outof-ssa.c: Likewise.
2671 * tree-parloops.c: Likewise.
2672 * tree-predcom.c: Likewise.
2673 * tree-profile.c: Likewise.
2674 * tree-scalar-evolution.c: Likewise.
2675 * tree-sra.c: Likewise.
2676 * tree-ssa-ccp.c: Likewise.
2677 * tree-ssa-coalesce.c: Likewise.
2678 * tree-ssa-copy.c: Likewise.
2679 * tree-ssa-copyrename.c: Likewise.
2680 * tree-ssa-dce.c: Likewise.
2681 * tree-ssa-dom.c: Likewise.
2682 * tree-ssa-dse.c: Likewise.
2683 * tree-ssa-forwprop.c: Likewise.
2684 * tree-ssa-ifcombine.c: Likewise.
2685 * tree-ssa-live.c: Likewise.
2686 * tree-ssa-loop-ch.c: Likewise.
2687 * tree-ssa-loop-im.c: Likewise.
2688 * tree-ssa-loop-ivcanon.c: Likewise.
2689 * tree-ssa-loop-ivopts.c: Likewise.
2690 * tree-ssa-loop-manip.c: Likewise.
2691 * tree-ssa-loop-niter.c: Likewise.
2692 * tree-ssa-loop-prefetch.c: Likewise.
2693 * tree-ssa-loop.c: Likewise.
2694 * tree-ssa-math-opts.c: Likewise.
2695 * tree-ssa-phiopt.c: Likewise.
2696 * tree-ssa-phiprop.c: Likewise.
2697 * tree-ssa-pre.c: Likewise.
2698 * tree-ssa-propagate.c: Likewise.
2699 * tree-ssa-reassoc.c: Likewise.
2700 * tree-ssa-sink.c: Likewise.
2701 * tree-ssa-strlen.c: Likewise.
2702 * tree-ssa-structalias.c: Likewise.
2703 * tree-ssa-tail-merge.c: Likewise.
2704 * tree-ssa-ter.c: Likewise.
2705 * tree-ssa-threadedge.c: Likewise.
2706 * tree-ssa-threadupdate.c: Likewise.
2707 * tree-ssa-uncprop.c: Likewise.
2708 * tree-ssa-uninit.c: Likewise.
2709 * tree-ssa.c: Likewise.
2710 * tree-stdarg.c: Likewise.
2711 * tree-switch-conversion.c: Likewise.
2712 * tree-tailcall.c: Likewise.
2713 * tree-vect-data-refs.c: Likewise.
2714 * tree-vect-generic.c: Likewise.
2715 * tree-vect-loop-manip.c: Likewise.
2716 * tree-vect-loop.c: Likewise.
2717 * tree-vect-patterns.c: Likewise.
2718 * tree-vect-slp.c: Likewise.
2719 * tree-vect-stmts.c: Likewise.
2720 * tree-vectorizer.c: Likewise.
2721 * tree-vrp.c: Likewise.
2724 * value-prof.c: Likewise.
2725 * vtable-verify.c: Likewise.
2727 2013-11-13 Steven Bosscher <steven@gcc.gnu.org>
2729 * gimple-ssa-isolate-paths.c (pass_isolate_erroneous_paths): Comment
2732 2013-11-13 Jeff Law <law@redhat.com>
2734 * PR middle-end/59119
2735 * gimple-ssa-isolate-paths.c (find_implicit_erroneous_behaviour): New
2736 function, extracted from gimple_ssa_isolate_erroneous_paths.
2737 (find_explicit_erroneous_behaviour): Similarly.
2738 (insert_trap_and_remove_trailing_statements): Remove statements
2741 2013-11-13 Steven Bosscher <steven@gcc.gnu.org>
2743 * cfgrtl.c (can_fallthru): Reorder code to move tablejump check up.
2744 Make that check explicit. BB_HEAD cannot be NULL, remove check for it.
2745 * haifa-sched.c (ready_remove_first_dispatch): Check INSN_P before
2746 looking at INSN_CODE.
2747 * reload1.c (delete_dead_insn) Do not expect JUMP_TABLE_DATA to be an
2748 active_insn_p object, respect basic block boundaries.
2749 * reorg.c (follow_jumps): Use invariant that JUMP_TABLE_DATA always
2750 follows immediately after the jump table data label.
2751 * config/nds32/nds32.c (nds32_output_casesi_pc_relative): Likewise.
2752 * config/sh/sh.c (barrier_align): Likewise. Rearrange code such
2753 that JUMP_TABLE_DATA is not expected to be an active_insn_p object.
2755 2013-11-13 Teresa Johnson <tejohnson@google.com>
2758 * predict.c (drop_profile): Error is currently too strict.
2759 (handle_missing_profiles): Pass call_count to drop_profile.
2761 2013-11-13 Teresa Johnson <tejohnson@google.com>
2764 * ipa-inline.c (edge_badness): Fix overflow.
2766 2013-11-13 Vladimir Makarov <vmakarov@redhat.com>
2768 PR rtl-optimization/59036
2769 * ira-color.c (struct allocno_color_data): Add new members
2770 first_thread_allocno, next_thread_allocno, thread_freq.
2771 (sorted_copies): New static var.
2772 (allocnos_conflict_by_live_ranges_p, copy_freq_compare_func): Move up.
2773 (allocno_thread_conflict_p, merge_threads)
2774 (form_threads_from_copies, form_threads_from_bucket)
2775 (form_threads_from_colorable_allocno, init_allocno_threads): New
2777 (bucket_allocno_compare_func): Add comparison by thread frequency
2779 (add_allocno_to_ordered_bucket): Rename to
2780 add_allocno_to_ordered_colorable_bucket. Remove parameter.
2781 (push_only_colorable): Call form_threads_from_bucket.
2782 (color_pass): Call init_allocno_threads. Use
2783 consideration_allocno_bitmap instead of coloring_allocno_bitmap
2784 for nuillify allocno color data.
2785 (ira_initiate_assign, ira_finish_assign): Allocate/free sorted_copies.
2786 (coalesce_allocnos): Use static sorted copies.
2788 2013-11-13 Jakub Jelinek <jakub@redhat.com>
2790 * passes.c (execute_todo): Don't call do_per_function if
2792 (execute_one_ipa_transform_pass, execute_one_pass): Don't call
2793 execute_function_dump if dump_file is NULL.
2795 2013-11-13 Martin Jambor <mjambor@suse.cz>
2797 * cgraph.c (cgraph_get_create_node): Do what
2798 cgraph_get_create_real_symbol_node used to do.
2799 (cgraph_get_create_real_symbol_node): Removed. Changed all users to
2800 call cgraph_get_create_node.
2801 * cgraph.h (cgraph_get_create_real_symbol_node): Removed.
2802 * lto-streamer-in.c (input_function): Call cgraph_get_node instead of
2803 cgraph_get_create_node. Assert we get a node.
2805 2013-11-13 Tejas Belagod <tejas.belagod@arm.com>
2807 * config/aarch64/aarch64-simd.md (vec_extract): New.
2809 2013-11-13 Tejas Belagod <tejas.belagod@arm.com>
2811 * config/aarch64/aarch64-simd.md (vec_set<mode>): Add w -> w option to
2814 2013-11-13 Eric Botcazou <ebotcazou@adacore.com>
2816 * cfgexpand.c (expand_used_vars): Allocate space for partitions based
2817 on PARM_DECLs or RESULT_DECLs only if they are ignored for debug info
2818 or if optimization is enabled.
2819 * tree-ssa-coalesce.c (coalesce_ssa_name): If optimization is disabled,
2820 require that all the names based on a PARM_DECL or a RESULT_DECL that
2821 isn't ignored for debug info be coalesced.
2823 2013-11-13 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2825 * config/c6x/c6x.c: Include "gimple-expr.h".
2827 2013-11-13 Richard Biener <rguenther@suse.de>
2829 * gimple-streamer-out.c (output_gimple_stmt): Also wrap
2830 decls in ADDR_EXPR operands inside a MEM_REF and optimize that.
2831 * gimple-streamer-in.c (input_gimple_stmt): Remove now dead code
2832 dealing with type mismatches inside component reference chains.
2834 2013-11-13 Marc Glisse <marc.glisse@inria.fr>
2836 PR tree-optimization/59077
2837 * ipa-pure-const.c (better_state): Update *state.
2839 2013-11-13 Christophe Lyon <christophe.lyon@linaro.org>
2841 * config/aarch64/aarch64.h (FRAME_GROWS_DOWNWARD): Define to 1.
2842 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
2843 Update offset calculations.
2845 2013-11-13 Eric Botcazou <ebotcazou@adacore.com>
2848 * dwarf2out.c (add_byte_size_attribute): Also use int_size_in_bytes
2849 for fields. Do not add the attribute if the size is negative.
2851 2013-11-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2853 * config/arm/arm.c: Include aarch-cost-tables.h.
2854 (generic_extra_costs): Move from here...
2855 * config/arm/aarch-cost-tables.h: ... To here. New file.
2857 2013-11-13 Alexander Ivchenko <alexander.ivchenko@intel.com>
2858 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2859 Sergey Lega <sergey.s.lega@intel.com>
2860 Anna Tikhonova <anna.tikhonova@intel.com>
2861 Ilya Tocar <ilya.tocar@intel.com>
2862 Andrey Turetskiy <andrey.turetskiy@intel.com>
2863 Ilya Verbin <ilya.verbin@intel.com>
2864 Kirill Yukhin <kirill.yukhin@intel.com>
2865 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2867 * config/i386/i386.c (ix86_print_operand): Support z-masking.
2868 * config/i386/predicate.md (const_0_to_4_operand): New.
2869 (const_0_to_5_operand): Ditto.
2870 * config/i386/sse.md (UNSPEC_COMPRESS): New.
2871 (UNSPEC_COMPRESS_STORE): Ditto.
2872 (UNSPEC_EXPAND): Ditto.
2873 (UNSPEC_EMBEDDED_ROUNDING): Ditto.
2874 (define_mode_attr ssescalarsize): Ditto.
2875 (avx512f_load<mode>_mask): Ditto.
2876 (avx512f_store<mode>_mask): Ditto.
2877 (avx512f_storedqu<mode>_mask): Ditto.
2878 (avx512f_vmcmp<mode>3_mask): Ditto.
2879 (avx512f_fmadd_<mode>_mask): Ditto.
2880 (avx512f_fmadd_<mode>_mask3): Ditto.
2881 (avx512f_fmsub_<mode>_mask): Ditto.
2882 (avx512f_fmsub_<mode>_mask3): Ditto.
2883 (avx512f_fnmadd_<mode>_mask): Ditto.
2884 (avx512f_fnmadd_<mode>_mask3): Ditto.
2885 (avx512f_fnmsub_<mode>_mask): Ditto.
2886 (avx512f_fnmsub_<mode>_mask3): Ditto.
2887 (avx512f_fmaddsub_<mode>_mask): Ditto.
2888 (avx512f_fmaddsub_<mode>_mask3): Ditto.
2889 (avx512f_fmsubadd_<mode>_mask): Ditto.
2890 (avx512f_fmsubadd_<mode>_mask3): Ditto.
2891 (vec_unpacku_float_lo_v16si): Ditto.
2892 (avx512f_vextract<shuffletype>32x4_mask): Ditto.
2893 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
2894 (avx512f_vextract<shuffletype>64x4_mask): Ditto.
2895 (vec_extract_lo_<mode>_maskm): Ditto.
2896 (vec_extract_hi_<mode>_maskm): Ditto.
2897 (avx512f_vternlog<mode>_mask): Ditto.
2898 (avx512f_shufps512_mask): Ditto.
2899 (avx512f_fixupimm<mode>_mask): Ditto.
2900 (avx512f_shufpd512_mask): Ditto.
2901 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
2902 (avx512f_<code>v8div16qi2_mask/trunc): Ditto.
2903 (*avx512f_<code>v8div16qi2_store_mask): Ditto.
2904 (ashr<mode>3<mask_name>): Ditto.
2905 (avx512f_vinsert<shuffletype>32x4_mask): Ditto.
2906 (avx512f_vinsert<shuffletype>64x4_mask): Ditto.
2907 (avx512f_shuf_<shuffletype>64x2_mask): Ditto.
2908 (avx512f_shuf_<shuffletype>32x4_mask): Ditto.
2909 (avx512f_pshufdv3_mask): Ditto.
2910 (avx512f_perm<mode>_mask): Ditto.
2911 (avx512f_vpermi2var<mode>3_mask): Ditto.
2912 (avx512f_vpermt2var<mode>3_mask): Ditto.
2913 (avx512f_compress<mode>_mask): Ditto.
2914 (avx512f_compressstore<mode>_mask): Ditto.
2915 (avx512f_expand<mode>_mask): Ditto.
2916 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>): Extend
2918 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
2919 (<plusminus_insn><mode>3<mask_name>): Ditto.
2920 (*<plusminus_insn><mode>3<mask_name>): Ditto.
2921 (mul<mode>3<mask_name>): Ditto.
2922 (*mul<mode>3<mask_name>): Ditto.
2923 (<sse>_div<mode>3<mask_name>): Ditto.
2924 (<mask_codefor>rcp14<mode><mask_name>): Ditto.
2925 (<sse>_sqrt<mode>2<mask_name>): Ditto.
2926 (<mask_codefor>rsqrt14<mode><mask_name>): Ditto.
2927 (<code><mode>3<mask_name>/smaxmin): Ditto.
2928 (*<code><mode>3_finite<mask_name>/smaxmin): Ditto.
2929 (*<code><mode>3<mask_name>/smaxmin): Ditto.
2930 (float<sseintvecmodelower><mode>2<mask_name>): Ditto.
2931 (ufloatv16siv16sf2<mask_name>): Ditto.
2932 (<mask_codefor>avx512f_fix_notruncv16sfv16si<mask_name>): Ditto.
2933 (<mask_codefor>avx512f_ufix_notruncv16sfv16si<mask_name>): Ditto.
2934 (<fixsuffix>fix_truncv16sfv16si2<mask_name>): Ditto.
2935 (float<si2dfmodelower><mode>2<mask_name>): Ditto.
2936 (ufloatv8siv8df<mask_name>): Ditto.
2937 (<mask_codefor>avx512f_cvtpd2dq512<mask_name>): Ditto.
2938 (avx512f_ufix_notruncv8dfv8si<mask_name>): Ditto.
2939 (<fixsuffix>fix_truncv8dfv8si2<mask_name>): Ditto.
2940 (<mask_codefor>avx512f_cvtpd2ps512<mask_name>): Ditto.
2941 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix><mask_name>): Ditto.
2942 (<mask_codefor>avx512f_unpckhps512<mask_name>): Ditto.
2943 (<mask_codefor>avx512f_unpcklps512<mask_name>): Ditto.
2944 (<mask_codefor>avx512f_movshdup512<mask_name>): Ditto.
2945 (<mask_codefor>avx512f_movsldup512<mask_name>): Ditto.
2946 (<mask_codefor>avx512f_vextract<shuffletype>32x4_1<mask_name>): Ditto.
2947 (vec_extract_lo_<mode><mask_name>): Ditto.
2948 (vec_extract_hi_<mode><mask_name>): Ditto.
2949 (<mask_codefor>avx512f_unpckhpd512<mask_name>): Ditto.
2950 (avx512f_movddup512<mask_name>): Ditto.
2951 (avx512f_unpcklpd512<mask_name>): Ditto.
2952 (*avx512f_unpcklpd512<mask_name>): Ditto.
2953 (*avx512f_vmscalef<mode>): Ditto.
2954 (avx512f_scalef<mode><mask_name>): Ditto.
2955 (avx512f_getexp<mode><mask_name>): Ditto.
2956 (<mask_codefor>avx512f_align<mode><mask_name>): Ditto.
2957 (avx512f_rndscale<mode><mask_name>): Ditto.
2958 (avx512f_shufps512_1<mask_name>): Ditto.
2959 (avx512f_shufpd512_1<mask_name>): Ditto.
2960 (<plusminus_insn><mode>3<mask_name>): Ditto.
2961 (*<plusminus_insn><mode>3<mask_name>): Ditto.
2962 (vec_widen_umult_even_v16si<mask_name>): Ditto.
2963 (*vec_widen_umult_even_v16si<mask_name>): Ditto.
2964 (vec_widen_smult_even_v16si<mask_name>): Ditto.
2965 (*vec_widen_smult_even_v16si<mask_name>): Ditto.
2966 (mul<mode>3<mask_name>): Ditto.
2967 (*<sse4_1_avx2>_mul<mode>3<mask_name>): Ditto.
2968 (<shift_insn><mode>3<mask_name>): Ditto.
2969 (avx512f_<rotate>v<mode><mask_name>/rotate): Ditto.
2970 (avx512f_<rotate><mode><mask_name>): Ditto.
2971 (<code><mode>3<mask_name>/maxmin): Ditto.
2972 (*avx2_<code><mode>3<mask_name>/maxmin): Ditto.
2973 (<sse2_avx2>_andnot<mode>3<mask_name>): Ditto.
2974 (*andnot<mode>3<mask_name>): Ditto.
2975 (<mask_codefor><code><mode>3<mask_name>/any_logic): Ditto.
2976 (<mask_codefor>avx512f_interleave_highv16si<mask_name>): Ditto.
2977 (<mask_codefor>avx512f_interleave_lowv16si<mask_name>): Ditto.
2978 (<mask_codefor>avx512f_vinsert<shuffletype>32x4_1<mask_name>): Ditto.
2979 (vec_set_lo_<mode><mask_name>): Ditto.
2980 (vec_set_hi_<mode><mask_name>): Ditto.
2981 (avx512f_shuf_<shuffletype>64x2_1<mask_name>): Ditto.
2982 (avx512f_shuf_<shuffletype>32x4_1<mask_name>): Ditto.
2983 (avx512f_pshufd_1<mask_name>): Ditto.
2984 (<mask_codefor>abs<mode>2<mask_name>): Ditto.
2985 (<mask_codefor>avx512f_<code>v16qiv16si2<mask_name>): Ditto.
2986 (avx512f_<code>v16hiv16si2<mask_name>/any_extend): Ditto.
2987 (avx512f_<code>v8qiv8di2<mask_name>/any_extend): Ditto.
2988 (avx512f_<code>v8hiv8di2<mask_name>/any_extend): Ditto.
2989 (avx512f_<code>v8siv8di2<mask_name>/any_extend): Ditto.
2990 (avx512er_exp2<mode><mask_name>): Ditto.
2991 (<mask_codefor>avx512er_rcp28<mode><mask_name>): Ditto.
2992 (<mask_codefor>avx512er_rsqrt28<mode><mask_name>): Ditto.
2993 (<avx2_avx512f>_permvar<mode><mask_name>): Ditto.
2994 (<avx2_avx512f>_perm<mode>_1<mask_name>): Ditto.
2995 (<mask_codefor>avx512f_vec_dup<mode><mask_name>): Ditto.
2996 (<mask_codefor>avx512f_broadcast<mode><mask_name>/V16FI): Ditto.
2997 (<mask_codefor>avx512f_broadcast<mode><mask_name>/V8FI): Ditto.
2998 (<mask_codefor>avx512f_vec_dup_gpr<mode><mask_name>): Ditto.
2999 (<mask_codefor>avx512f_vec_dup_mem<mode><mask_name>): Ditto.
3000 (<sse2_avx_avx512f>_vpermil<mode><mask_name>/VF2): Ditto.
3001 (<sse2_avx_avx512f>_vpermil<mode><mask_name>/VF1): Ditto.
3002 (*<sse2_avx_avx512f>_vpermilp<mode><mask_name>): Ditto.
3003 (<sse2_avx_avx512f>_vpermilvar<mode>3<mask_name>): Ditto.
3004 (<avx2_avx512f>_ashrv<mode><mask_name>): Ditto.
3005 (<avx2_avx512f>_<shift_insn>v<mode><mask_name>): Ditto.
3006 (<mask_codefor>avx512f_vcvtph2ps512<mask_name>): Ditto.
3007 (<mask_codefor>avx512f_vcvtps2ph512<mask_name>): Ditto.
3008 (avx512f_getmant<mode><mask_name>): Ditto.
3009 (clz<mode>2<mask_name>): Ditto.
3010 (<mask_codefor>conflict<mode><mask_name>): Ditto.
3011 (*srcp14<mode>): Remove visibility.
3012 (*rsqrt14<mode>): Ditto.
3013 (*fma_fmsub_<mode>): Ditto.
3014 (*fma_fnmadd_<mode>): Ditto.
3015 (*avx512f_rndscale<mode>): Ditto.
3016 * config/i386/subst.md: New file.
3018 2013-11-13 Joseph Myers <joseph@codesourcery.com>
3020 * doc/extend.texi (Statement Exprs, Typeof): Discuss __auto_type.
3021 * ginclude/stdatomic.h (kill_dependency, atomic_store_explicit)
3022 (atomic_load_explicit, atomic_exchange_explicit)
3023 (atomic_compare_exchange_strong_explicit)
3024 (atomic_compare_exchange_weak_explicit): Use __auto_type to
3025 declare variable initialized with PTR argument.
3027 2013-11-12 Jeff Law <law@redhat.com>
3029 * tree-ssa-threadedge.c (thread_around_empty_blocks): New argument
3030 backedge_seen_p. Set, use and pass it to children appropriately.
3031 (thread_through_normal_block): Similarly.
3032 (thread_across_edge): Similarly.
3034 * gimple-ssa-isolate-paths.c (check_loadstore): Mark discovered
3035 memory references as volatile.
3036 (insert_trap_and_remove_trailing_statements): Fix comment.
3038 2013-11-12 Vladimir Makarov <vmakarov@redhat.com>
3041 * ira-costs.c (record_operand_costs): Check operands number for
3044 2013-11-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3047 * config/rs6000/rs6000.md (movdi_internal32): Eliminate
3048 constraints that would allow DImode into the traditional Altivec
3049 registers, but cause undesirable code generation when loading 0 as
3051 (movdi_internal64): Likewise.
3052 (cmp<mode>_fpr): Do not use %x for CR register output.
3053 (extendsfdf2_fpr): Fix constraints when -mallow-upper-df and
3054 -mallow-upper-sf debug switches are used.
3056 2013-11-12 Andrew MacLeod <amacleod@redhat.com>
3058 * gimple-expr.h (create_tmp_var_name, create_tmp_var_raw,
3059 create_tmp_var, create_tmp_reg, mark_addressable, is_gimple_reg_rhs):
3060 Relocate prototypes from gimple.h.
3061 * gimplify.h: New File. Relocate some prototypes from gimple.h here.
3062 (gimple_predicate, enum fallback, enum gimplify_status): Relocate
3064 * gimple.h: Move some prototypes to gimplify.h.
3065 (gimple_predicate, enum fallback, enum gimplify_status): Move to
3067 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location):
3068 Relocate from gimpify.c.
3069 * gimple-expr.c (remove_suffix, tmp_var_id_num, create_tmp_var_name,
3070 create_tmp_var_raw, create_tmp_var, create_tmp_reg, mark_addressable,
3071 is_gimple_reg_rhs) Relocate from gimplify.c.
3072 * gimplify.c (mark_addressable): Move to gimple-expr.c.
3073 (gimple_seq_add_stmt_without_update): Move to gimple.c.
3074 (remove_suffix, tmp_var_id_num, create_tmp_var_name,
3075 create_tmp_var_raw, create_tmp_var, create_tmp_reg,
3076 is_gimple_reg_rhs): Move to gimple-expr.c.
3077 (should_carry_location_p): Move to gimple.c.
3078 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location): Move
3080 (annotate_one_with_location, annotate_all_with_location_after,
3081 annotate_all_with_location): Move to gimple.c.
3082 (compare_case_labels, sort_case_labels,
3083 preprocess_case_label_vec_for_gimple): Move to gimple.c.
3084 (rhs_predicate_for): Make static.
3085 (gimplify_assign): Relocate from gimple.c.
3086 * gimple.c (gimplify_assign): Move to gimplify.c.
3087 (gimple_seq_add_stmt_without_update, should_carry_location_p,
3088 annotate_one_with_location, annotate_all_with_location_after,
3089 annotate_all_with_location, compare_case_labels, sort_case_labels,
3090 preprocess_case_label_vec_for_gimple): Relocate from gimplify.c.
3091 * tree.h (unshare_expr, unshare_expr_without_location,
3092 mark_addressable): Move prototypes to gimplify.h.
3093 * Makefile.in (GTFILES): gimple-expr.c now has the GTY tag for
3095 * asan.c: Include gimplify.h rather than gimple.h.
3096 * cfgloopmanip.c: Likewise.
3097 * cgraphunit.c: Likewise.
3098 * cilk-common.c: Likewise.
3099 * dwarf2out.c: Dont include gimple.h.
3100 * fold-const.c: Include gimplify.h rather than gimple.h.
3101 * function.c: Likewise.
3102 * gimple-fold.c: Likewise.
3103 * gimple-ssa-strength-reduction.c: Likewise.
3104 * graphite-clast-to-gimple.c: Likewise.
3105 * graphite-sese-to-poly.c: Likewise.
3106 * ipa-prop.c: Likewise.
3107 * ipa-split.c: Likewise.
3109 * langhooks.c: Dont include gimple.h.
3110 * loop-init.c: Include gimplify.h rather than gimple.h.
3111 * omp-low.c: Likewise.
3113 * stor-layout.c: Likewise.
3114 * targhooks.c: Likewise.
3115 * trans-mem.c: Likewise.
3116 * tree-affine.c: Likewise.
3117 * tree-cfg.c: Likewise.
3118 * tree-cfgcleanup.c: Likewise.
3119 * tree-complex.c: Likewise.
3120 * tree-if-conv.c: Likewise.
3121 * tree-inline.c: Likewise.
3122 * tree-iterator.c: Likewise.
3123 * tree-loop-distribution.c: Likewise.
3124 * tree-nested.c: Likewise.
3125 * tree-parloops.c: Likewise.
3126 * tree-predcom.c: Likewise.
3127 * tree-profile.c: Likewise.
3128 * tree-scalar-evolution.c: Likewise.
3129 * tree-sra.c: Likewise.
3130 * tree-ssa-address.c: Likewise.
3131 * tree-ssa-ccp.c: Likewise.
3132 * tree-ssa-dce.c: Likewise.
3133 * tree-ssa-forwprop.c: Likewise.
3134 * tree-ssa-ifcombine.c: Likewise.
3135 * tree-ssa-loop-im.c: Likewise.
3136 * tree-ssa-loop-ivopts.c: Likewise.
3137 * tree-ssa-loop-manip.c: Likewise.
3138 * tree-ssa-loop-niter.c: Likewise.
3139 * tree-ssa-loop-prefetch.c: Likewise.
3140 * tree-ssa-loop-unswitch.c: Likewise.
3141 * tree-ssa-math-opts.c: Likewise.
3142 * tree-ssa-phiopt.c: Likewise.
3143 * tree-ssa-phiprop.c: Likewise.
3144 * tree-ssa-pre.c: Likewise.
3145 * tree-ssa-propagate.c: Likewise.
3146 * tree-ssa-reassoc.c: Likewise.
3147 * tree-ssa-sccvn.c: Likewise.
3148 * tree-ssa-strlen.c: Likewise.
3149 * tree-ssa.c: Likewise.
3150 * tree-switch-conversio: Likewise.n.c
3151 * tree-tailcall.c: Likewise.
3152 * tree-vect-data-refs.c: Likewise.
3153 * tree-vect-generic.c: Likewise.
3154 * tree-vect-loop-manip.c: Likewise.
3155 * tree-vect-loop.c: Likewise.
3156 * tree-vect-patterns.c: Likewise.
3157 * tree-vect-stmts.c: Likewise.
3159 * value-prof.c: Likewise.
3160 * config/aarch64/aarch64.c: Include gimplify.h instead of gimple.h.
3161 * config/alpha/alpha.c: Likewise.
3162 * config/darwin.c: Likewise.
3163 * config/i386/i386.c: Likewise.
3164 * config/ia64/ia64.c: Likewise.
3165 * config/mep/mep.c: Likewise.
3166 * config/mips/mips.c: Likewise.
3167 * config/rs6000/rs6000.c: Likewise.
3168 * config/s390/s390.c: Likewise.
3169 * config/sh/sh.c: Likewise.
3170 * config/sparc/sparc.c: Likewise.
3171 * config/spu/spu.c: Likewise.
3172 * config/stormy16/stormy16.c: Likewise.
3173 * config/tilegx/tilegx.c: Likewise.
3174 * config/tilepro/tilepro.c: Likewise.
3175 * config/xtensa/xtensa.c: Likewise.
3177 2013-11-12 Adam Butcher <adam@jessamine.co.uk>
3179 * tree.c (grow_tree_vec_stat): New function ...
3180 * tree.h (grow_tree_vec_stat) (grow_tree_vec): ... and its declaration
3181 and macro front-end.
3183 2013-11-12 Marek Polacek <polacek@redhat.com>
3185 * final.c (update_alignments): Initialize label to NULL_RTX.
3187 2013-11-12 Jeff Law <law@redhat.com>
3189 * gimple-ssa-isolate-paths.c (check_loadstore): New function.
3190 (insert_trap_and_remove_trailing_statements): New argument OP which
3191 is the NULL pointer. Emit the trap after the load/store through
3192 the NULL pointer. Simplify the RHS of a store through a NULL pointer
3193 when trivial to do so.
3194 (isolate_path): Corresponding changes.
3195 (gimple_ssa_isolate_erroneous_path): Likewise.
3197 2013-11-12 Teresa Johnson <tejohnson@google.com>
3198 Jan Hubicka <jh@suse.cz>
3200 * predict.c (drop_profile): New function.
3201 (handle_missing_profiles): Ditto.
3202 (counts_to_freqs): Don't overwrite estimated frequencies
3203 when function has no profile counts.
3204 * predict.h (handle_missing_profiles): Declare.
3205 * tree-inline.c (freqs_to_counts): New function.
3206 (copy_cfg_body): Invoke freqs_to_counts as needed.
3207 * tree-profile.c (tree_profiling): Invoke handle_missing_profiles.
3209 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
3212 * config/i386/x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL):
3214 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Set for m_HASWELL.
3216 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
3219 * config/i386/i386.c (ix86_option_override_internal): Check
3220 X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL and
3221 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL for
3222 MASK_AVX256_SPLIT_UNALIGNED_LOAD and
3223 MASK_AVX256_SPLIT_UNALIGNED_STORE.
3225 * config/i386/x86-tune.def (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL):
3226 Clear m_COREI7_AVX and update comments.
3227 (X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): Likewise.
3229 2013-11-12 Martin Jambor <mjambor@suse.cz>
3231 PR rtl-optimization/10474
3232 * ira.c (interesting_dest_for_shprep): New function.
3233 (split_live_ranges_for_shrink_wrap): Likewise.
3234 (find_moveable_pseudos): Move calculation of dominance info,
3235 df_analysios and the final anlyses to...
3236 (ira): ...here, call split_live_ranges_for_shrink_wrap.
3238 2013-11-12 Bin Cheng <bin.cheng@arm.com>
3240 * tree-ssa-loop-ivopts.c (force_expr_to_var_cost): Refactor the code.
3241 Handle type conversion.
3243 2013-11-11 Martin Liska <marxin.liska@gmail.com>
3244 Jan Hubicka <jh@suse.cz>
3246 * cgraph.c (dump_cgraph_node): Profile dump added.
3247 * cgraph.h (struct cgraph_node): New time profile variable added.
3248 * cgraphclones.c (cgraph_clone_node): Time profile is cloned.
3249 * gcov-io.h (gcov_type): New profiler type introduced.
3250 * ipa-profile.c (lto_output_node): Streaming for time profile added.
3251 (input_node): Time profiler is read from LTO stream.
3252 * predict.c (maybe_hot_count_p): Hot prediction changed.
3253 * profile.c (instrument_values): New case for time profiler added.
3254 (compute_value_histograms): Read of time profile.
3255 * tree-pretty-print.c (dump_function_header): Time profiler is dumped.
3256 * tree-profile.c (init_ic_make_global_vars): Time profiler
3258 (gimple_init_edge_profiler): TP function instrumentation.
3259 (gimple_gen_time_profiler): New.
3260 * value-prof.c (gimple_add_histogram_value): Support for time profiler
3262 (dump_histogram_value): TP type added to dumps.
3263 (visit_hist): More sensitive check that takes TP into account.
3264 (gimple_find_values_to_profile): TP instrumentation.
3265 * value-prof.h (hist_type): New histogram type added.
3266 (struct histogram_value_t): Pointer to struct function added.
3267 * libgcc/Makefile.in: New GCOV merge function for TP added.
3268 * libgcov.c: function_counter variable introduced.
3269 (_gcov_merge_time_profile): New.
3270 (_gcov_time_profiler): New.
3272 2013-11-11 Marc Glisse <marc.glisse@inria.fr>
3273 Jeff Law <law@redhat.com>
3275 * tree-ssa-alias.c (stmt_kills_ref_p_1): Use
3276 ao_ref_init_from_ptr_and_size for builtins.
3278 2013-11-11 Uros Bizjak <ubizjak@gmail.com>
3279 H.J. Lu <hongjiu.lu@intel.com>
3282 * config/i386/x86-tune.def
3283 (X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
3284 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES.
3285 * config/i386/i386.h
3286 (TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
3287 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES. Update for renamed
3288 X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES.
3289 * config/i386/i386.c (ix86_expand_set_or_movmem): Use
3290 TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES to calculate
3291 misaligned_prologue_used. Check that
3292 desired_aling <= epilogue_size_needed.
3294 2013-11-11 Cong Hou <congh@google.com>
3296 PR tree-optimization/59050
3297 * tree-vect-data-refs.c (comp_dr_addr_with_seg_len_pair): Bug fix.
3299 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
3302 * expmed.c (emit_store_flag): Fail for const-const comparison.
3304 2013-11-11 Tristan Gingold <gingold@adacore.com>
3305 Eric Botcazou <ebotcazou@adacore.com>
3307 * tree.h (CONSTRUCTOR_NO_CLEARING): Define.
3308 * tree-core.h (CONSTRUCTOR_NO_CLEARING): Document it.
3309 * tree.def (CONSTRUCTOR): Likewise.
3310 * doc/generic.texi (CONSTRUCTOR): Likewise. Update description.
3311 * gimplify.c (gimplify_init_constructor): Do not clear the object when
3312 the constructor is incomplete and CONSTRUCTOR_NO_CLEARING is set.
3314 2013-11-11 Basile Starynkevitch <basile@starynkevitch.net>
3316 * toplev.c (toplev_main): Move PLUGIN_FINISH invocation before
3319 2013-11-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3321 * config/arm/arm.c (arm_new_rtx_costs): Return after handling
3324 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
3326 * config/arc/arc.h (LOGICAL_OP_NON_SHORT_CIRCUIT): Define.
3328 2013-11-08 Jeff Law <law@redhat.com>
3330 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
3331 threading paths first, then perform PHI node checks if applicable.
3333 2013-11-10 Karlson2k <k2k@narod.ru>
3334 Kai Tietz <ktietz@redhat.com>
3337 * configure.ac: Adding for exported symbols check
3338 and for rdynamic-check executable-extension.
3339 * configure: Regenerated.
3341 2013-11-10 Uros Bizjak <ubizjak@gmail.com>
3343 * mode-switching.c (optimize_mode_switching): Mark block as
3344 nontransparent, if last_mode at block exit is different from no_mode.
3346 2013-11-09 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3348 * function.c (NAME__MAIN): Move to...
3349 * cfgexpand.c (NAME__MAIN): ...here.
3351 2013-11-09 Richard Sandiford <rdsandiford@googlemail.com>
3353 * target.def (can_use_doloop_p): New hook.
3354 * doc/tm.texi.in (TARGET_CAN_USE_DOLOOP_P): Add.
3355 * doc/tm.texi: Regenerate.
3356 * doc/md.texi (doloop_begin, doloop_end): Update documentation.
3357 * hooks.h (hook_bool_dint_dint_uint_true): Declare.
3358 * hooks.c (hook_bool_dint_dint_uint_true): New function.
3359 * targhooks.h (can_use_doloop_if_innermost): Declare.
3360 * targhooks.c (can_use_doloop_if_innermost): New function.
3361 * target.h: Include double-int.h.
3362 * loop-doloop.c (doloop_optimize): Call targetm.can_use_doloop_p.
3363 Remove iteration count, maximum iteration count, loop depth and
3364 enter-at-top inputs from doloop_begin and doloop_end.
3365 * config/arc/arc.md (doloop_begin, doloop_end): Update for new
3367 * config/arc/arc.c (arc_can_use_doloop_p): New function.
3368 (TARGET_CAN_USE_DOLOOP_P): Define.
3369 * config/arm/thumb2.md (doloop_end): Update for new interface.
3370 * config/arm/arm.c (TARGET_CAN_USE_DOLOOP_P): Define.
3371 * config/bfin/bfin.md (doloop_end): Update for new interface.
3372 * config/bfin/bfin.c (bfin_can_use_doloop_p): New function.
3373 (TARGET_CAN_USE_DOLOOP_P): Define.
3374 * config/c6x/c6x.md (doloop_end): Update for new interface.
3375 * config/ia64/ia64.md (doloop_end): Update for new interface.
3376 * config/ia64/ia64.c (TARGET_CAN_USE_DOLOOP_P): Define.
3377 * config/mep/mep.md (doloop_begin, doloop_end): Update for new
3379 * config/mep/mep.c (mep_emit_doloop): Likewise.
3380 (TARGET_CAN_USE_DOLOOP_P): Define.
3381 * config/rs6000/rs6000.md (doloop_end): Update for new interface.
3382 * config/rs6000/rs6000.c (TARGET_CAN_USE_DOLOOP_P): Define.
3383 * config/s390/s390.md (doloop_end): Update for new interface.
3384 * config/sh/sh.md (doloop_end): Likewise.
3385 * config/spu/spu.md (doloop_end): Likewise.
3386 * config/spu/spu.c (TARGET_CAN_USE_DOLOOP_P): Define.
3387 * config/tilegx/tilegx.md (doloop_end): Update for new interface.
3388 * config/tilegx/tilegx.c (TARGET_CAN_USE_DOLOOP_P): Define.
3389 * config/tilepro/tilepro.md (doloop_end): Update for new interface.
3390 * config/tilepro/tilepro.c (TARGET_CAN_USE_DOLOOP_P): Define.
3391 * config/v850/v850.md (doloop_begin, doloop_end): Update for new
3393 * config/v850/v850.c (TARGET_CAN_USE_DOLOOP_P): Define.
3395 2013-11-08 H.J. Lu <hongjiu.lu@intel.com>
3398 * doc/extend.texi: Move Cilk Plus Builtins node before Other
3401 2013-11-08 Andrew MacLeod <amacleod@redhat.com>
3402 Joseph Myers <joseph@codesourcery.com>
3404 * ginclude/stdatomic.h: New file.
3405 * Makefile.in (USER_H): Add stdatomic.h.
3407 2013-11-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3409 * config/arm/arm.c (arm_new_rtx_costs): Break after handling
3412 2013-11-08 Jeff Law <law@redhat.com>
3414 * tree-ssa-threadupdate.h (delete_thread_path): Declare.
3415 * tree-ssa-threadupdate.c (delete_thread_path): New function.
3416 (ssa_redirect_edges, thread_block_1): Use it.
3417 (thread_through_loop_header, mark_threaded_blocks): Likewise.
3418 (thread_through_all_blocks, register_jump_thread): Likewise.
3419 * tree-ssa-threadedge.c (thread_across_edge): Likewise.
3421 2013-11-08 James Greenhalgh <james.greenhalgh@arm.com>
3423 * config/arm/aarch-common.c
3424 (search_term): New typedef.
3425 (shift_rtx_costs): New array.
3426 (arm_rtx_shift_left_p): New.
3427 (arm_find_sub_rtx_with_search_term): Likewise.
3428 (arm_find_sub_rtx_with_code): Likewise.
3429 (arm_early_load_addr_dep): Add sanity checking.
3430 (arm_no_early_alu_shift_dep): Likewise.
3431 (arm_no_early_alu_shift_value_dep): Likewise.
3432 (arm_no_early_mul_dep): Likewise.
3433 (arm_no_early_store_addr_dep): Likewise.
3435 2013-11-08 Richard Biener <rguenther@suse.de>
3437 PR tree-optimization/59047
3438 * tree-predcom.c (ref_at_iteration): Handle bitfield accesses properly.
3440 2013-11-08 Ilya Enkovich <ilya.enkovich@intel.com>
3442 * common.opt (fcheck-pointer-bounds): Move to ...
3443 * c-family/c.opt: ... here.
3444 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): Remove.
3445 (LANG_HOOKS_INITIALIZER): Remove LANG_HOOKS_CHKP_SUPPORTED.
3446 * langhooks.h (lang_hooks): Remove chkp_supported field.
3447 * toplev.c (process_options): Remove chkp_supported check.
3449 2013-11-08 Richard Biener <rguenther@suse.de>
3451 PR tree-optimization/59038
3452 PR tree-optimization/58955
3453 * tree-loop-distribution.c (pg_add_dependence_edges): Revert
3454 previous change. Handle known dependences correctly.
3456 2013-11-08 Tom de Vries <tom@codesourcery.com>
3458 * config/rs6000/t-xilinx: Remove duplicate contents.
3460 2013-11-07 Andrew MacLeod <amacleod@redhat.com>
3461 Joseph Myers <joseph@codesourcery.com>
3463 * tree-core.h (enum cv_qualifier): Add TYPE_QUAL_ATOMIC.
3464 (enum tree_index): Add TI_ATOMICQI_TYPE, TI_ATOMICHI_TYPE,
3465 TI_ATOMICSI_TYPE, TI_ATOMICDI_TYPE and TI_ATOMICTI_TYPE.
3466 (struct tree_base): Add atomic_flag field.
3467 * tree.h (TYPE_ATOMIC): New accessor macro.
3468 (TYPE_QUALS, TYPE_QUALS_NO_ADDR_SPACE): Add TYPE_QUAL_ATOMIC.
3469 (TYPE_QUALS_NO_ADDR_SPACE_NO_ATOMIC): New macro.
3470 (atomicQI_type_node, atomicHI_type_node, atomicSI_type_node)
3471 (atomicDI_type_node, atomicTI_type_node): New macros for type nodes.
3472 * tree.c (set_type_quals): Set TYPE_ATOMIC.
3473 (find_atomic_core_type): New function.
3474 (build_qualified_type): Adjust alignment for qualified types.
3475 (build_atomic_base): New function
3476 (build_common_tree_nodes): Build atomicQI_type_node,
3477 atomicHI_type_node, atomicSI_type_node, atomicDI_type_node and
3479 * print-tree.c (print_node): Print atomic qualifier.
3480 * tree-pretty-print.c (dump_generic_node): Print atomic type attribute.
3481 * target.def (atomic_assign_expand_fenv): New hook.
3482 * doc/tm.texi.in (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New @hook.
3483 * doc/tm.texi: Regenerate.
3484 * targhooks.c (default_atomic_assign_expand_fenv): New function.
3485 * targhooks.h (default_atomic_assign_expand_fenv): Declare.
3486 * sync-builtins.def (__atomic_feraiseexcept): New built-in function.
3487 * config/i386/i386-builtin-types.def (VOID_FTYPE_PUSHORT): New
3489 * config/i386/i386.c (enum ix86_builtins): Add
3490 IX86_BUILTIN_FNSTENV, IX86_BUILTIN_FLDENV, IX86_BUILTIN_FNSTSW and
3491 IX86_BUILTIN_FNCLEX.
3492 (bdesc_special_args): Add __builtin_ia32_fnstenv,
3493 __builtin_ia32_fldenv, __builtin_ia32_fnstsw and __builtin_ia32_fnclex.
3494 (ix86_expand_builtin): Handle the new built-in functions.
3495 (ix86_atomic_assign_expand_fenv): New function.
3496 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New macro.
3497 * config/i386/i386.md (UNSPECV_FNSTENV, UNSPECV_FLDENV)
3498 (UNSPECV_FNSTSW, UNSPECV_FNCLEX): New unspecs.
3499 (fnstenv, fldenv, fnstsw, fnclex): New insns.
3501 2013-11-07 Steve Ellcey <sellcey@mips.com>
3503 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add fp64 directory.
3504 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add -mfp64 flag.
3505 (MULTILIB_DIRNAMES): Add fp64 directory.
3506 (MULTILIB_EXCEPTIONS): Add new exclusions.
3508 2013-11-07 Aldy Hernandez <aldyh@redhat.com>
3510 * gimplify.c (gimple_regimplify_operands): Do not set
3512 * graphite-sese-to-poly.c (remove_simple_copy_phi): Same.
3513 (rewrite_close_phi_out_of_ssa): Same.
3514 (rewrite_phi_out_of_ssa): Same.
3515 (rewrite_degenerate_phi): Same.
3516 (handle_scalar_deps_crossing_scop_limits): Same.
3517 * tree-if-conv.c (predicate_scalar_phi): Same.
3518 * tree-parloops.c (create_loads_for_reductions): Same.
3519 (create_final_loads_for_reduction): Same.
3520 (create_loads_and_stores_for_name): Same.
3521 (transform_to_exit_first_loop): Same.
3522 (create_parallel_loop): Same.
3523 * tree-ssa-loop-im.c
3524 (move_computations_dom_walker::before_dom_children): Same.
3525 * tree-ssa-loop-manip.c (rewrite_phi_with_iv): Same.
3526 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Same.
3527 * tree-ssa-propagate.c (substitute_and_fold): Same.
3528 * tree-vect-loop.c (vect_finalize_reduction): Same.
3529 * tree-vect-stmts.c (vectorizable_call): Same.
3531 2013-11-07 Mike Stump <mikestump@comcast.net>
3533 * config/pdp11/pdp11.c: Include dbxout.h.
3534 * config/picochip/picochip.c: Likewise.
3536 2013-11-07 Cong Hou <congh@google.com>
3538 PR tree-optimization/56764
3539 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
3540 Combine alias checks if it is possible to amortize the runtime
3541 overhead. Return the number of alias checks after merging.
3542 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
3543 Use the function vect_create_cond_for_alias_checks () to check
3544 the number of alias checks.
3546 2013-11-07 Jeff Law <law@redhat.com>
3548 * varpool.c (ctor_for_folding): Fix typo in comment.
3550 2013-11-07 Joern Rennecke <joern.rennecke@embecosm.com>
3552 * config/arc/arc.c (arc_ifcvt): Use commutativity, e.g.:
3553 reg_a := reg_b + reg_a ==> reg_a := reg_a + reg_b
3555 2013-11-07 Jeff Law <law@redhat.com>
3557 * doc/invoke.texi (-fisolate-erroneous-paths): Document.
3559 * gimple-ssa-isolate-paths.c (gate_isolate_erroneous_paths):
3560 No longer check if we have __builtin_trap, assume it's available.
3562 2013-11-07 Diego Novillo <dnovillo@google.com>
3564 * attribs.c (lookup_scoped_attribute_spec): Make static.
3565 (get_attribute_namespace): Likewise.
3566 * builtins.c (more_const_call_expr_args_p): Move from tree.h.
3567 (validate_arglist): Move earlier in the file. Make static.
3568 (expand_stack_restore): Move from stmt.c
3569 (expand_stack_save): Move from stmt.c
3570 (rewrite_call_expr_array): Move earlier in the file.
3571 (rewrite_call_expr_valist): Likewise.
3572 * cfgexpand.c: Include hard-reg-set.h before tree.h
3575 (expand_asm_loc): Move from stmt.c.
3576 (n_occurrences): Move from stmt.c.
3577 (check_operand_nalternatives): Move from stmt.c.
3578 (tree_conflicts_with_clobbers_p): Move from stmt.c.
3579 (expand_asm_operands): Move from stmt.c
3580 (expand_asm_stmt): Move from stmt.c
3581 (expand_computed_goto): Move from stmt.c
3582 (expand_goto): Move from stmt.c
3583 (expand_null_return_1): Move from stmt.c
3584 (expand_null_return): Move from stmt.c
3585 (expand_value_return): Move from stmt.c
3586 (expand_return): Move from stmt.c
3587 (expand_main_function): Move from function.c
3588 (stack_protect_prologue): Move from function.c
3589 * cgraphclones.c (build_function_type_skip_args): Move from tree.c.
3590 (build_function_decl_skip_args): Move from tree.c.
3591 * explow.c (tree_expr_size): Move from tree.c.
3592 * expr.c (addr_expr_of_non_mem_decl_p): Remove.
3593 (fields_length): Move from tree.c.
3594 * fold-const.c (size_low_cst): Move from tree.c.
3595 (tree_expr_nonzero_warnv_p): Make static. Move earlier in the file.
3596 (tree_expr_nonzero_p): Make static. Move earlier in the file.
3597 (fold_build3_initializer_loc): Remove.
3598 (tree_invalid_nonnegative_warnv_p): Make static.
3599 * function.c (expand_main_function): Move to cfgexpand.c.
3600 (stack_protect_prologue): Move to cfgexpand.c.
3601 (set_insn_locations): Move earlier in the file.
3602 * gimple-fold.c: Include langhooks.h.
3603 (truth_type_for): Move from tree.c.
3604 * print-tree.c (print_vec_tree): Remove.
3605 * stmt.c (expand_computed_goto): Move to cfgexpand.c.
3606 (expand_goto): Move to cfgexpand.c.
3607 (n_occurrences): Move to cfgexpand.c.
3608 (expand_asm_loc): Move to cfgexpand.c
3609 (tree_conflicts_with_clobbers_p): Move to cfgexpand.c.
3610 (expand_asm_operands): Move to cfgexpand.c.
3611 (expand_asm_stmt): Move to cfgexpand.c.
3612 (check_operand_nalternatives): Move to cfgexpand.c
3613 (expand_null_return): Move to cfgexpand.c.
3614 (expand_value_return): Move to cfgexpand.c.
3615 (expand_null_return_1): Move to cfgexpand.c.
3616 (expand_return): Move to cfgexpand.c.
3617 (expand_stack_save): Move to builtins.c.
3618 (expand_stack_restore): Move to builtins.c
3619 * symtab.c: Include output.h.
3620 (decl_assembler_name_hash): Move from tree.c.
3621 (decl_assembler_name_equal): Move from tree.c.
3622 * trans-mem.c (is_tm_safe_or_pure): Move from tree.h.
3623 * tree-eh.c (in_array_bounds_p): Move from tree.c.
3624 (range_in_array_bounds_p): Move from tree.c.
3625 * tree-object-size.c (fini_object_sizes): Make static.
3626 * tree-ssa-dom.c (iterative_hash_exprs_commutative): Move from tree.h.
3627 * tree-vrp.c (ssa_name_nonnegative_p): Remove.
3628 * tree.c (decl_assembler_name_equal): Move to symtab.c.
3629 (tree_expr_size): Move to explow.c.
3630 (decl_assembler_name_hash): Move to symtab.c.
3631 (real_twop): Remove.
3632 (tree_expr_size): Move to explow.c.
3633 (stabilize_reference_1): Move earlier in the file. Make static.
3634 (omp_remove_redundant_declare_simd_attrs): Remove.
3635 (simple_cst_list_equal): Move earlier in the file. Make static.
3636 (size_low_cst): Move to fold-const.c.
3637 (build_type_no_quals): Remove.
3638 (build_function_type_skip_args): Move to cgraphclones.c.
3639 (build_function_decl_skip_args): Move to cgraphclones.c.
3640 (in_array_bounds_p): Move to tree-eh.c.
3641 (range_in_array_bounds_p): Move to tree-eh.c.
3642 (truth_type_for): Move to gimple-fold.c.
3643 (list_equal_p): Remove.
3644 * tree.h (decl_assembler_name_equal): Remove.
3645 (decl_assembler_name_hash): Remove.
3646 (truth_type_for): Remove.
3647 (build_type_no_quals): Remove.
3648 (build_function_decl_skip_args): Remove.
3649 (in_array_bounds_p): Remove.
3650 (range_in_array_bounds_p): Remove.
3651 (size_low_cst): Remove.
3652 (omp_remove_redundant_declare_simd_attrs): Remove.
3653 (tree_expr_size): Remove.
3654 (fields_length): Remove.
3655 (stabilize_reference_1): Remove.
3656 (expand_goto): Remove.
3657 (expand_stack_save): Remove.
3658 (expand_stack_restore): Remove.
3659 (expand_return): Remove.
3660 (fold_build3_initializer_loc): Remove.
3661 (tree_expr_nonzero_p): Remove.
3662 (tree_invalid_nonnegative_warnv_p): Remove.
3663 (tree_expr_nonzero_warnv_p): Remove.
3664 (fold_builtin_snprintf_chk): Remove.
3665 (validate_arglist): Remove.
3666 (iterative_hash_exprs_commutative): Move to tree-ssa-dom.c.
3667 (simple_cst_list_equal): Remove.
3668 (real_twop): Remove.
3669 (expand_main_function): Remove.
3670 (stack_protect_prologue): Remove.
3671 (print_vec_tree): Remove.
3672 (lookup_scoped_attribute_spec): Remove.
3673 (get_attribute_namespace): Remove.
3674 (expand_computed_goto): Remove.
3675 (expand_asm_stmt): Remove.
3676 (list_equal_p): Remove.
3677 (ssa_name_nonnegative_p): Remove.
3678 (fini_object_sizes): Remove.
3679 (addr_expr_of_non_mem_decl_p): Remove.
3680 (is_tm_safe_or_pure): Move to trans-mem.c.
3681 (more_const_call_expr_args_p): Remove.
3682 (save_vtable_map_decl): Remove.
3684 2013-11-07 Thomas Schwinge <thomas@codesourcery.com>
3686 * doc/sourcebuild.texi (Top Level) <lto-plugin>: GNU ld can use
3687 linker plugins, too.
3689 * config/arc/arc.h (LINK_COMMAND_SPEC): For -ftree-parallelize-loops=*,
3690 link to libgomp and its dependencies.
3691 * config/ia64/hpux.h (LIB_SPEC): Likewise.
3692 * config/pa/pa-hpux11.h (LIB_SPEC): Likewise.
3693 * config/pa/pa64-hpux.h (LIB_SPEC): Likewise.
3694 * gcc.c (GOMP_SELF_SPECS): Update comment about libgomp's dependencies.
3696 2013-11-07 Jakub Jelinek <jakub@redhat.com>
3698 * tree-ssa-loop-niter.c: Include tree-ssanames.h.
3699 (determine_value_range): Add loop argument. Use get_range_info to
3701 (bound_difference): Adjust caller.
3703 2013-11-07 Richard Biener <rguenther@suse.de>
3704 Jakub Jelinek <jakub@redhat.com>
3706 * tree-vrp.c (find_assert_locations): Pre-seed live bitmaps for loop
3707 latches from header PHI arguments from the latch edge.
3709 2013-11-07 Paolo Carlini <paolo.carlini@oracle.com>
3712 * varasm.c (output_constant): Handle NULLPTR_TYPE.
3714 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
3716 * config/i386/i386.c (ix86_expand_set_or_movmem): Don't set
3717 misaligned_prologue_used when it has been set.
3719 2013-11-07 Yury Gribov <y.gribov@samsung.com>
3720 Jakub Jelinek <jakub@redhat.com>
3723 * asan.c (get_mem_refs_of_builtin_call): Allow
3724 integer literals as addresses in instrumented builtins.
3726 2013-11-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3728 * config/aarch64/aarch64.c (aarch64_legitimize_reload_address):
3729 Explain why plus_constant is not used.
3731 2013-11-07 Richard Biener <rguenther@suse.de>
3733 * tree-ssa-ccp.c (canonicalize_float_value): Rename to ...
3734 (canonicalize_value): ... this. Also handle stripping of
3736 (get_value, set_lattice_value, get_value_for_expr): Adjust.
3737 * gimple-fold.c (canonicalize_constructor_val): Strip TREE_OVERFLOW.
3738 * tree-ssa-threadedge.c (set_ssa_name_value): Likewise.
3740 2013-11-07 Richard Biener <rguenther@suse.de>
3742 * tree-dfa.c (get_ref_base_and_extent): Fix casting.
3744 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
3747 * config/i386/i386.md (push peepholer/splitter): Use Pmode
3748 with stack_pointer_rtx.
3750 2013-11-07 Bin Cheng <bin.cheng@arm.com>
3752 * tree-ssa-loop-ivopts.c (get_shiftadd_cost): Check equality
3753 using operand_equal_p.
3755 2013-11-07 Bin Cheng <bin.cheng@arm.com>
3757 * tree-ssa-loop-ivopts.c (alloc_iv): Lower address expressions.
3758 * tree-affine.c (get_inner_reference_aff): Return base.
3759 * tree-affine.h (get_inner_reference_aff): Change prototype.
3761 2013-11-06 Tobias Burnus <burnus@net-b.de>
3763 * doc/invoke.texi (Wdate-time): Fix typo.
3765 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
3767 * config/sh/sh.md (addsf3, divsf3, divsf3_i, rsqrtsf2, cmpgtdf_t,
3768 cmpeqdf_t, *ieee_ccmpeqdf_t, negdf2, sqrtdf2, absdf2): Use
3769 fp_arith_reg_operand instead of arith_reg_operand.
3771 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
3773 * config/sh/sh.md (adddi3): Remove empty constraints.
3774 Remove can_create_pseudo_p and arith_reg_operand check.
3775 (adddi3_compact, subdi3_compact, *negdi2): Remove constraints.
3776 Split before reload.
3778 2013-11-06 Jeff Law <law@redhat.com>
3779 Tom Tromey <tromey@redhat.com>
3781 * gdbinit.in: Disable strict type checking.
3783 2013-11-06 Vladimir Makarov <vmakarov@redhat.com>
3785 * tree-pass.h (make_pass_live_range_shrinkage): New external.
3786 * timevar.def (TV_LIVE_RANGE_SHRINKAGE): New.
3787 * sched-rgn.c (gate_handle_live_range_shrinkage): New.
3788 (rest_of_handle_live_range_shrinkage): Ditto
3789 (class pass_live_range_shrinkage): Ditto.
3790 (pass_data_live_range_shrinkage): Ditto.
3791 (make_pass_live_range_shrinkage): Ditto.
3792 * sched-int.h (initialize_live_range_shrinkage): New prototype.
3793 (finish_live_range_shrinkage): Ditto.
3794 * sched-deps.c (create_insn_reg_set): Make void return value.
3795 * passes.def: Add pass_live_range_shrinkage.
3796 * ira.c (update_equiv_regs): Don't move if flag_live_range_shrinkage.
3797 * haifa-sched.c (live_range_shrinkage_p): New.
3798 (initialize_live_range_shrinkage, finish_live_range_shrinkage):
3800 (rank_for_schedule): Add code for pressure relief through live
3802 (schedule_insn): Print more debug info.
3803 (sched_init): Setup SCHED_PRESSURE_WEIGHTED for pressure relief
3804 through live range shrinkage.
3805 * doc/invoke.texi (-flive-range-shrinkage): New.
3806 * common.opt (flive-range-shrinkage): New.
3808 2013-11-06 Uros Bizjak <ubizjak@gmail.com>
3811 * config/i386/i386.c (ix86_avx_u128_mode_needed): Require
3812 AVX_U128_DIRTY mode for call_insn RTXes that use AVX256 registers.
3813 (ix86_avx_u128_mode_needed): Return AVX_U128_DIRTY mode for call_insn
3814 RTXes that return in AVX256 register.
3816 2013-11-06 Richard Biener <rguenther@suse.de>
3818 PR tree-optimization/58653
3819 * tree-predcom.c (ref_at_iteration): Rewrite to generate a MEM_REF.
3820 (prepare_initializers_chain): Adjust.
3822 2013-11-06 Andrew MacLeod <amacleod@redhat.com>
3824 * gimple.h (block_in_transaction): Move to basic-block.h and rename.
3825 (gimple_in_transaction): Use bb_in_transaction.
3826 * basic-block.h (bb_in_transaction): Relocate here and rename.
3827 * tree-ssa-loop-im.c (execute_sm): Use bb_in_transaction.
3829 2013-11-06 Richard Biener <rguenther@suse.de>
3831 * tree.c (drop_tree_overflow): New function.
3832 * tree.h (drop_tree_overflow): Declare.
3833 * gimplify.c (gimplify_expr): Drop TREE_OVERFLOW.
3834 * tree-vrp.c (range_int_cst_singleton_p): Use
3835 is_overflow_infinity instead of testing TREE_OVERFLOW.
3836 (extract_range_from_assert): Likewise.
3837 (zero_nonzero_bits_from_vr): Likewise.
3838 (extract_range_basic): Likewise.
3839 (register_new_assert_for): Use drop_tree_overflow.
3840 (vrp_visit_phi_node): Likewise.
3842 2013-11-06 Eric Botcazou <ebotcazou@adacore.com>
3844 * config/i386/i386.c (ix86_expand_prologue): Optimize stack
3845 checking for leaf functions without dynamic stack allocation.
3846 * config/ia64/ia64.c (ia64_emit_probe_stack_range): Adjust.
3847 (ia64_expand_prologue): Likewise.
3848 * config/mips/mips.c (mips_expand_prologue): Likewise.
3849 * config/rs6000/rs6000.c (rs6000_emit_prologue): Likewise.
3850 * config/sparc/sparc.c (sparc_expand_prologue): Likewise.
3851 (sparc_flat_expand_prologue): Likewise.
3853 2013-11-06 James Greenhalgh <james.greenhalgh@arm.com>
3855 * config/aarch64/arm_neon.h (__ST2_LANE_FUNC): Better model data size.
3856 (__ST3_LANE_FUNC): Likewise.
3857 (__ST4_LANE_FUNC): Likewise.
3859 2013-11-06 Nick Clifton <nickc@redhat.com>
3861 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define the
3862 name returned by msp430_mcu_name.
3863 (LIB_SPEC): If a -T option has not been specified then set a
3864 default, mcu-specific, linker script.
3865 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more mcu names.
3866 * config/msp430/msp430.c (msp430x_names): Likewise.
3867 Alpha sort the names for ease of comparison.
3868 (msp430_mcu_name): New function: Returns a string suitable for
3869 use as a C preprocessor symbol based upon the name of the MCU
3871 (msp430_option_override): Accept msp430x and msp430xv2 as generic
3873 * config/msp430/msp430-protos.h (msp430_mcu_name): Prototype.
3875 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
3876 unless it is the last part of a spec substring.
3877 * doc/invoke.texi (Spec Files): Document space insertion
3880 2013-11-06 Christian Bruel <christian.bruel@st.com>
3882 * config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
3883 Factorize probabilities, Use adjust_address instead of
3884 adjust_automodify_address when possible. Enable for optimize.
3885 (sh_expand_strlen): New function.
3886 * config/sh/sh-protos.h (sh_expand_strlen): Declare.
3887 * config/sh/sh.md (strlensi): New pattern.
3888 (UNSPEC_BUILTIN_STRLEN): Define.
3890 2013-11-06 Jakub Jelinek <jakub@redhat.com>
3893 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
3894 (expand_assignment): If *bitpos is negative, set *offset
3895 and adjust *bitpos, so that it is not negative.
3897 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
3899 * config/i386/bdver3.md : Added two additional decoder units
3900 to support issue rate of 4 and remodeled vector unit.
3901 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
3902 architectures is set to 4.
3903 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
3904 lookahead is set to 4 for BD architectures.
3906 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3908 * config/rs6000/rs6000.c (rs6000_option_override_internal):
3909 Remove restriction against use of VSX instructions when generating
3910 code for little endian mode.
3912 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3914 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
3915 for both big and little endian.
3916 (mulv8hi3): Swap input operands for merge high and merge low
3917 instructions for little endian.
3919 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3921 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
3922 define_insn to define_expand that uses even patterns for big
3923 endian and odd patterns for little endian.
3924 (vec_widen_smult_even_v16qi): Likewise.
3925 (vec_widen_umult_even_v8hi): Likewise.
3926 (vec_widen_smult_even_v8hi): Likewise.
3927 (vec_widen_umult_odd_v16qi): Likewise.
3928 (vec_widen_smult_odd_v16qi): Likewise.
3929 (vec_widen_umult_odd_v8hi): Likewise.
3930 (vec_widen_smult_odd_v8hi): Likewise.
3931 (altivec_vmuleub): New define_insn.
3932 (altivec_vmuloub): Likewise.
3933 (altivec_vmulesb): Likewise.
3934 (altivec_vmulosb): Likewise.
3935 (altivec_vmuleuh): Likewise.
3936 (altivec_vmulouh): Likewise.
3937 (altivec_vmulesh): Likewise.
3938 (altivec_vmulosh): Likewise.
3940 2013-11-05 Mike Stump <mikestump@comcast.net>
3942 * Makefile.in (mostlyclean): Remove c-family objects.
3944 2013-11-05 Ian Lance Taylor <iant@google.com>
3946 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
3947 If possible, add .cfi directives to record change to bx.
3948 * config/i386/i386.c (ix86_emit_cfi): New function.
3949 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
3951 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
3953 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
3954 immediately after a label for a tablejump pattern.
3956 * config/arm/arm.c (is_jump_table): Remove.
3957 (create_fix_barrier): Use tablejump_p instead.
3958 (arm_reorg): Likewise.
3959 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
3960 (thumb2_output_casesi): Likewise.
3961 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
3962 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
3963 casesi_shift_media, casesi_load_media): Likewise.
3964 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
3965 * config/microblaze/microblaze.md: Likewise.
3967 2013-11-05 Tobias Burnus <burnus@net-b.de>
3969 * doc/invoke.texi (-Wdate-time): Document.
3971 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
3973 * double-int.c (lshift_double, rshift_double): Remove
3974 SHIFT_COUNT_TRUNCATED handling.
3976 2013-11-05 Jeff Law <law@redhat.com>
3978 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
3979 * common.opt (-fisolate-erroneous-paths): Add option and documentation.
3980 * gimple-ssa-isolate-paths.c: New file.
3981 * gimple.c (check_loadstore): New function.
3982 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
3983 Verify OP is in the argument list and the argument corresponding
3984 to OP is a pointer type. Use operand_equal_p rather than
3985 pointer equality when testing if OP is on the nonnull list.
3986 Use check_loadstore rather than count_ptr_derefs. Handle
3987 GIMPLE_RETURN statements.
3988 * tree-vrp.c (infer_nonnull_range): Remove.
3989 * gimple.h (infer_nonnull_range): Declare.
3990 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
3991 * passes.def: Add pass_isolate_erroneous_paths.
3992 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
3993 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
3994 * tree-ssa.c (struct count_ptr_d): Remove.
3995 (count_ptr_derefs, count_uses_and_derefs): Remove.
3996 * tree-ssa.h (count_uses_and_derefs): Remove.
3998 2013-11-05 Jakub Jelinek <jakub@redhat.com>
4000 PR rtl-optimization/58997
4001 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
4002 get_iv_value to be in iv->mode rather than iv->extend_mode.
4003 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
4004 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
4005 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
4008 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
4010 * gimple.h: Move some prototypes to gimple-expr.h and add to include
4012 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
4013 Move to gimple-expr.h.
4014 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
4015 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
4016 is_gimple_id, virtual_operand_p, is_gimple_addressable,
4017 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
4019 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
4020 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
4021 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
4022 is_gimple_address, is_gimple_invariant_address,
4023 is_gimple_ip_invariant_address, is_gimple_min_invariant,
4024 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
4025 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
4026 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
4027 gimple_decl_printable_name, useless_type_conversion_p,
4028 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
4030 * gimple-expr.c: New File.
4031 (useless_type_conversion_p, gimple_set_body, gimple_body,
4032 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
4033 gimple_can_coalesce_p, extract_ops_from_tree_1,
4034 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
4035 is_gimple_address, is_gimple_invariant_address,
4036 is_gimple_ip_invariant_address, is_gimple_min_invariant,
4037 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
4038 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
4039 is_gimple_mem_ref_addr): Relocate here.
4040 * Makefile.in (OBJS): Add gimple-expr.o.
4042 2013-11-05 David Malcolm <dmalcolm@redhat.com>
4044 * gengtype-parse.c (struct_field_seq): Support empty structs.
4046 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
4048 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
4050 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
4052 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
4054 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
4056 (LONG_DOUBLE_TYPE_SIZE): New define.
4057 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
4059 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
4062 * doc/extend.texi [visibility ("visibility_type")]: Add example
4063 about visibility attribute on namespace declaration.
4065 2013-11-05 Richard Biener <rguenther@suse.de>
4068 * passes.def (all_passes): Start with pass_fixup_cfg again.
4070 2013-11-05 Richard Biener <rguenther@suse.de>
4072 PR tree-optimization/58955
4073 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
4076 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4078 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
4080 (vec_pack_ufix_trunc_v2df): Likewise.
4082 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
4085 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
4086 pattern, instead of word_mode.
4088 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
4090 (set_storage_via_setmem): Likewise.
4092 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
4094 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
4095 gimple.h and the rest of the condition in eliminate_build.
4096 (eliminate_build): Call new routine.
4097 * gimple.h (phi_ssa_name_p): Delete.
4099 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
4101 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
4102 case of no prefix and reserving zero slots, because when that's the
4103 case we'll never get here.
4104 * vec.h (va_heap::reserve): Don't try and handle
4105 vec_prefix::calculate_allocation returning zero because that should
4108 2013-11-05 Richard Biener <rguenther@suse.de>
4111 * tree-dfa.c (get_ref_base_and_extent): Merge common code
4112 in MEM_REF and TARGET_MEM_REF handling. Make sure to
4113 process trailing array detection before diving into the
4114 view-converted object (and possibly apply some extra offset).
4116 2013-11-05 Joseph Myers <joseph@codesourcery.com>
4118 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
4120 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
4122 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
4124 PR tree-optimization/58958
4125 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
4126 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
4128 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
4130 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
4131 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
4133 2013-11-05 Jakub Jelinek <jakub@redhat.com>
4135 PR tree-optimization/58984
4136 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
4137 set *SIZE_P if non-NULL on success.
4138 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
4140 (ipcp_transform_function): Likewise. Punt if size of access
4141 is different from TYPE_SIZE on v->value's type.
4143 2013-11-05 Tobias Burnus <burnus@net-b.de>
4145 * doc/invoke.texi (-fopenmp-simd): Document new option.
4146 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
4147 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
4148 * tree.c (attribute_value_equal): Ditto.
4150 2013-11-04 Wei Mi <wmi@google.com>
4152 * sched-rgn.c (add_branch_dependences): Keep insns in
4153 a SCHED_GROUP at the end of BB to remain their location.
4155 2013-11-04 Wei Mi <wmi@google.com>
4157 * config/i386/i386.c (memory_address_length): Extract a part
4158 of code to rip_relative_addr_p.
4159 (rip_relative_addr_p): New Function.
4160 (ix86_macro_fusion_p): Ditto.
4161 (ix86_macro_fusion_pair_p): Ditto.
4162 * config/i386/i386.h: Add new tune features about macro-fusion.
4163 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
4164 * doc/tm.texi: Generated.
4165 * doc/tm.texi.in: Ditto.
4166 * haifa-sched.c (try_group_insn): New Function.
4167 (group_insns_for_macro_fusion): Ditto.
4168 (sched_init): Call group_insns_for_macro_fusion.
4169 * target.def: Add two hooks: macro_fusion_p and
4170 macro_fusion_pair_p.
4172 2013-11-04 Kostya Serebryany <kcc@google.com>
4174 Update to match the changed asan API.
4175 * asan.c (asan_function_start): New function.
4176 (asan_emit_stack_protection): Update the string stored in the
4177 stack red zone to match new API. Store the PC of the current
4178 function in the red zone.
4179 (asan_global_struct): Update the __asan_global definition to match
4181 (asan_add_global): Ditto.
4182 * asan.h (asan_function_start): New prototype.
4183 * final.c (final_start_function): Call asan_function_start.
4184 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
4187 2013-11-04 Wei Mi <wmi@google.com>
4189 * config/i386/i386-c.c (ix86_target_macros_internal): Separate
4190 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
4191 * config/i386/i386.c (ix86_option_override_internal): Ditto.
4192 (ix86_issue_rate): Ditto.
4193 (ix86_adjust_cost): Ditto.
4194 (ia32_multipass_dfa_lookahead): Ditto.
4195 (ix86_sched_init_global): Ditto.
4196 (get_builtin_code_for_version): Ditto.
4197 * config/i386/i386.h (enum target_cpu_default): Ditto.
4198 (enum processor_type): Ditto.
4199 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
4201 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
4203 PR rtl-optimization/58967
4204 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
4205 !lra_in_progress for mode sizes bigger word.
4207 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4209 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
4210 arguments to merge instruction for little endian.
4211 (vec_widen_umult_lo_v16qi): Likewise.
4212 (vec_widen_smult_hi_v16qi): Likewise.
4213 (vec_widen_smult_lo_v16qi): Likewise.
4214 (vec_widen_umult_hi_v8hi): Likewise.
4215 (vec_widen_umult_lo_v8hi): Likewise.
4216 (vec_widen_smult_hi_v8hi): Likewise.
4217 (vec_widen_smult_lo_v8hi): Likewise.
4219 2013-11-04 Ian Lance Taylor <iant@google.com>
4221 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
4222 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
4223 builtins that take pointers.
4224 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
4225 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
4226 (run_gcc): Pass -fnon-call-exceptions.
4228 2013-11-04 Jakub Jelinek <jakub@redhat.com>
4230 * optabs.c (expand_vec_perm): Revert one incorrect line from
4233 PR tree-optimization/58978
4234 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
4235 use_stmt by single_imm_use directly. Only call single_imm_use
4238 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
4240 PR rtl-optimization/58968
4241 * lra-spills.c (return_regno_p): New function.
4242 (lra_final_code_change): Use it.
4244 2013-11-04 Joseph Myers <joseph@codesourcery.com>
4246 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
4247 * target.def (float_exceptions_rounding_supported_p): New hook.
4248 * targhooks.c (default_float_exceptions_rounding_supported_p): New
4250 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
4251 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
4253 * doc/tm.texi: Regenerate.
4254 * config.gcc (powerpc*-*-linux*): Set extra_objs.
4255 * config/rs6000/rs6000-linux.c: New file.
4256 * config/rs6000/rs6000-protos.h
4257 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
4258 * config/rs6000/linux.h
4259 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
4260 * config/rs6000/linux64.h
4261 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
4262 * config/rs6000/t-linux (rs6000-linux.o): New rule.
4263 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
4265 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4267 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
4268 Replace the define_insn_and_split with a define_insn and two
4269 define_splits, with the split after reload re-permuting the source
4270 register to its original value.
4271 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
4272 (*vsx_le_perm_store_v8hi): Likewise.
4273 (*vsx_le_perm_store_v16qi): Likewise.
4275 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4277 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
4280 2013-11-04 Jakub Jelinek <jakub@redhat.com>
4282 PR tree-optimization/58946
4283 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
4284 bbs with bbinfo[idx].op != NULL before all blocks with
4285 bbinfo[idx].op == NULL.
4287 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
4289 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
4290 (avr_dump_double_int_hex): Likewise.
4291 (avr_log_vadump): Remove %D and %X handling.
4292 * config/avr/avr.c (avr_double_int_push_digit): Delete.
4293 (avr_map_op_t): Change map from double_int to unsigned int.
4294 (avr_map_op): Update accordingly.
4295 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
4296 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
4297 unsigned ints rather than double_ints.
4299 2013-11-03 Marek Polacek <polacek@redhat.com>
4301 Implement -fsanitize=vla-bound.
4302 * opts.c (common_handle_option): Handle vla-bound.
4303 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
4304 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
4305 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
4307 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4309 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
4312 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
4314 * config/i386/constraints.md (Ts, Tv): New address constrains.
4315 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
4316 constraint for address_no_seg_operand.
4317 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
4318 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
4319 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
4320 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
4321 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
4322 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
4323 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
4324 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
4325 constraint for vsib_address_operand.
4327 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
4329 * gcse.c (pre_delete): Remove references to regmove from comments.
4330 * recog.c: (validate_replace_rtx_1): Likewise.
4331 * config/rl78/rl78.c: Likewise.
4332 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
4333 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
4334 * common/config/mmix/mmix-common.c: Likewise.
4336 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
4338 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
4339 * gimplify.c (gimplify_compound_lval): Likewise.
4340 * graphite-clast-to-gimple.c (gloog): Likewise.
4341 * graphite-dependences.c (loop_is_parallel_p): Likewise.
4342 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
4343 (limit_scop); Likewise.
4344 (build_scops): Likewise.
4345 (dot_scop): Likewise.
4346 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
4347 (build_scop_drs): Likewise.
4348 (insert_stmts): Likewise.
4349 (insert_out_of_ssa_copy): Likewise.
4350 (remove_phi): Likewise.
4351 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
4352 * hw-doloop.c (discover_loop): Likewise.
4353 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
4354 * tree-dfa.c (dump_enumerated_decls): Likewise.
4355 * tree-if-conv.c (if_convertable_loop_p): Likewise.
4356 * tree-inline.c (tree_function_versioning): Likewise.
4357 * tree-loop-distribution.c (build_rdg): Likewise.
4358 (rdg_flag_vertex_and_dependent): Likewise.
4359 (distribute_loop): Likewise.
4360 * tree-parloops.c (loop_parallel_p): Likewise.
4361 (eliminate_local_variables): Likewise.
4362 (separate_decls_in_region): Likewise.
4363 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
4364 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
4365 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
4366 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
4367 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
4368 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
4369 (vectorizable_condition): Likewise.
4371 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
4373 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
4374 * configure: Regenerate.
4375 * config/i386/i386.md (*movdi_internal): Change
4376 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
4377 (*movdf_internal): Ditto.
4378 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
4379 * config/i386/sse.md (vec_concatv2di): Output interunit movq
4380 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
4382 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
4384 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
4387 2013-10-31 David Malcolm <dmalcolm@redhat.com>
4389 Automated part of renaming of symtab_node_base to symtab_node.
4391 Patch autogenerated by rename_symtab.py from
4392 https://github.com/davidmalcolm/gcc-refactoring-scripts
4393 revision 58bb219cc090b2f4516a9297d868c245495ee622
4394 with ChangeLog entry fixed up by hand.
4396 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
4398 (cgraph_node_for_asm): Likewise.
4399 * cgraph.h (symtab_node_base): Likewise.
4400 (cgraph_node): Likewise.
4401 (varpool_node): Likewise.
4402 (is_a_helper <cgraph_node>::test): Likewise.
4403 (is_a_helper <varpool_node>::test): Likewise.
4404 (symtab_nodes): Likewise.
4405 (symtab_register_node): Likewise.
4406 (symtab_unregister_node): Likewise.
4407 (symtab_remove_node): Likewise.
4408 (symtab_get_node): Likewise.
4409 (symtab_node_for_asm): Likewise.
4410 (symtab_node_asm_name): Likewise.
4411 (symtab_node_name): Likewise.
4412 (symtab_insert_node_to_hashtable): Likewise.
4413 (symtab_add_to_same_comdat_group): Likewise.
4414 (symtab_dissolve_same_comdat_group_list): Likewise.
4415 (dump_symtab_node): Likewise.
4416 (debug_symtab_node): Likewise.
4417 (dump_symtab_base): Likewise.
4418 (verify_symtab_node): Likewise.
4419 (verify_symtab_base): Likewise.
4420 (symtab_used_from_object_file_p): Likewise.
4421 (symtab_alias_ultimate_target): Likewise.
4422 (symtab_resolve_alias): Likewise.
4423 (fixup_same_cpp_alias_visibility): Likewise.
4424 (symtab_for_node_and_aliases): Likewise.
4425 (symtab_nonoverwritable_alias): Likewise.
4426 (availability symtab_node_availability): Likewise.
4427 (symtab_semantically_equivalent_p): Likewise.
4428 (fixup_same_cpp_alias_visibility): Likewise.
4429 (symtab_prevail_in_asm_name_hash): Likewise.
4431 (varpool): Likewise.
4432 (varpool_first_variable): Likewise.
4433 (varpool_next_variable): Likewise.
4434 (varpool_first_static_initializer): Likewise.
4435 (varpool_next_static_initializer): Likewise.
4436 (varpool_first_defined_variable): Likewise.
4437 (varpool_next_defined_variable): Likewise.
4438 (cgraph_first_defined_function): Likewise.
4439 (cgraph_next_defined_function): Likewise.
4440 (cgraph_first_function): Likewise.
4441 (cgraph_next_function): Likewise.
4442 (cgraph_first_function_with_gimple_body): Likewise.
4443 (cgraph_next_function_with_gimple_body): Likewise.
4444 (symtab_alias_target): Likewise.
4445 (symtab_real_symbol_p): Likewise.
4446 (symtab_can_be_discarded): Likewise.
4447 * cgraphbuild.c (mark_address): Likewise.
4448 (mark_load): Likewise.
4449 (mark_store): Likewise.
4450 * cgraphunit.c (decide_is_symbol_needed): Likewise.
4452 (enqueue_node): Likewise.
4453 (referred_to_p): Likewise.
4454 (cgraph_process_same_body_aliases): Likewise.
4455 (analyze_functions): Likewise.
4456 (handle_alias_pairs): Likewise.
4457 (output_weakrefs): Likewise.
4458 (compile): Likewise.
4459 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
4460 * ipa-inline-analysis.c (inline_write_summary): Likewise.
4461 * ipa-prop.c (remove_described_reference): Likewise.
4462 (try_decrement_rdesc_refcount): Likewise.
4463 (ipa_edge_duplication_hook): Likewise.
4464 * ipa-ref.c (ipa_record_reference): Likewise.
4465 (ipa_maybe_record_reference): Likewise.
4466 (ipa_clone_ref): Likewise.
4467 (ipa_clone_references): Likewise.
4468 (ipa_clone_referring): Likewise.
4469 (ipa_find_reference): Likewise.
4470 (ipa_remove_stmt_references): Likewise.
4471 (ipa_clear_stmts_in_references): Likewise.
4472 * ipa-ref.h (symtab_node_base): Likewise.
4473 (ipa_ref): Likewise.
4474 (ipa_record_reference): Likewise.
4475 (ipa_maybe_record_reference): Likewise.
4476 (ipa_clone_references): Likewise.
4477 (ipa_clone_referring): Likewise.
4478 (ipa_clone_ref): Likewise.
4479 (ipa_find_reference): Likewise.
4480 (ipa_remove_stmt_references): Likewise.
4481 (ipa_clear_stmts_in_references): Likewise.
4482 * ipa-reference.c (ipa_reference_write_optimization_summary):
4484 * ipa.c (enqueue_node): Likewise.
4485 (process_references): Likewise.
4486 (walk_polymorphic_call_targets): Likewise.
4487 (symtab_remove_unreachable_nodes): Likewise.
4488 (address_taken_from_non_vtable_p): Likewise.
4489 (comdat_can_be_unshared_p_1): Likewise.
4490 (comdat_can_be_unshared_p): Likewise.
4491 (can_replace_by_local_alias): Likewise.
4492 (function_and_variable_visibility): Likewise.
4493 * is-a.h: Likewise (within example in comment).
4494 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
4495 (lto_symtab_encoder_encode): Likewise.
4496 (lto_symtab_encoder_delete_node): Likewise.
4497 (lto_symtab_encoder_in_partition_p): Likewise.
4498 (lto_set_symtab_encoder_in_partition): Likewise.
4499 (output_refs): Likewise.
4500 (compute_ltrans_boundary): Likewise.
4501 (output_symtab): Likewise.
4502 (input_node): Likewise.
4503 (input_ref): Likewise.
4504 (input_edge): Likewise.
4505 (input_cgraph_1): Likewise.
4506 (input_refs): Likewise.
4507 (output_cgraph_opt_summary): Likewise.
4508 (input_node_opt_summary): Likewise.
4509 (input_cgraph_opt_section): Likewise.
4510 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
4512 * lto-streamer-out.c (lto_output): Likewise.
4513 (output_symbol_p): Likewise.
4514 (produce_symtab): Likewise.
4515 * lto-streamer.h (lto_encoder_entry): Likewise.
4516 (lto_free_function_in_decl_state_for_node): Likewise.
4517 (lto_symtab_encoder_encode): Likewise.
4518 (lto_symtab_encoder_delete_node): Likewise.
4519 (lto_symtab_encoder_in_partition_p): Likewise.
4520 (lto_set_symtab_encoder_in_partition): Likewise.
4521 (lto_symtab_encoder_lookup): Likewise.
4522 (lsei_node): Likewise.
4523 (lto_symtab_encoder_deref): Likewise.
4524 * symtab.c (symtab_hash): Likewise.
4525 (assembler_name_hash): Likewise.
4526 (symtab_nodes): Likewise.
4527 (hash_node): Likewise.
4528 (eq_node): Likewise.
4529 (hash_node_by_assembler_name): Likewise.
4530 (eq_assembler_name): Likewise.
4531 (insert_to_assembler_name_hash): Likewise.
4532 (unlink_from_assembler_name_hash): Likewise.
4533 (symtab_prevail_in_asm_name_hash): Likewise.
4534 (symtab_register_node): Likewise.
4535 (symtab_insert_node_to_hashtable): Likewise.
4536 (symtab_unregister_node): Likewise.
4537 (symtab_get_node): Likewise.
4538 (symtab_remove_node): Likewise.
4539 (symtab_initialize_asm_name_hash): Likewise.
4540 (symtab_node_for_asm): Likewise.
4541 (symtab_add_to_same_comdat_group): Likewise.
4542 (symtab_dissolve_same_comdat_group_list): Likewise.
4543 (symtab_node_asm_name): Likewise.
4544 (symtab_node_name): Likewise.
4545 (dump_symtab_base): Likewise.
4546 (dump_symtab_node): Likewise.
4547 (dump_symtab): Likewise.
4548 (debug_symtab_node): Likewise.
4549 (verify_symtab_base): Likewise.
4550 (verify_symtab_node): Likewise.
4551 (verify_symtab): Likewise.
4552 (symtab_used_from_object_file_p): Likewise.
4553 (symtab_node_availability): Likewise.
4554 (symtab_alias_ultimate_target): Likewise.
4555 (fixup_same_cpp_alias_visibility): Likewise.
4556 (symtab_resolve_alias): Likewise.
4557 (symtab_for_node_and_aliases): Likewise.
4558 (symtab_for_node_and_aliases): Likewise.
4559 (symtab_nonoverwritable_alias_1): Likewise.
4560 (symtab_nonoverwritable_alias): Likewise.
4561 (symtab_semantically_equivalent_p): Likewise.
4562 * value-prof.c (init_node_map): Likewise.
4563 * varasm.c (find_decl): Likewise.
4564 * varpool.c (varpool_node_for_asm): Likewise.
4565 (varpool_remove_unreferenced_decls): Likewise.
4567 2013-10-31 David Malcolm <dmalcolm@redhat.com>
4569 Manual part of renaming of symtab_node_base to symtab_node.
4571 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
4572 clashes with the preferred name for the base class.
4573 (const_symtab_node): Remove redundant typedef.
4575 2013-10-31 Jakub Jelinek <jakub@redhat.com>
4577 * optabs.c (expand_vec_perm): Avoid vector mode punning
4578 SUBREGs in SET_DEST.
4579 * expmed.c (store_bit_field_1): Likewise.
4580 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
4581 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
4582 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
4583 * config/i386/i386.c (ix86_expand_vector_move_misalign,
4584 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
4585 ix86_expand_sse_unpack, ix86_expand_args_builtin,
4586 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
4587 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
4588 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
4589 expand_vec_perm_vpshufb2_vpermq,
4590 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
4591 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
4592 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
4593 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
4596 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
4598 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
4599 Use gcc_unreachable for unhandled modes. Do not check results of
4600 expand_simple_binop. If not expanded to target, move the result.
4602 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
4603 Shiva Chen <shiva0217@gmail.com>
4605 * config.gcc (nds32*-*-*): Add nds32 target.
4606 * config/nds32/nds32.c: New file.
4607 * config/nds32/nds32.h: New file.
4608 * config/nds32/nds32.md: New file.
4609 * config/nds32/constants.md: New file.
4610 * config/nds32/constraints.md: New file.
4611 * config/nds32/iterators.md: New file.
4612 * config/nds32/nds32-doubleword.md: New file.
4613 * config/nds32/nds32-intrinsic.md: New file.
4614 * config/nds32/nds32_intrinsic.h: New file.
4615 * config/nds32/nds32-modes.def: New file.
4616 * config/nds32/nds32-multiple.md: New file.
4617 * config/nds32/nds32.opt: New file.
4618 * config/nds32/nds32-opts.h: New file.
4619 * config/nds32/nds32-protos.h: New file.
4620 * config/nds32/nds32-peephole2.md: New file.
4621 * config/nds32/pipelines.md: New file.
4622 * config/nds32/predicates.md: New file.
4623 * config/nds32/t-mlibs: New file.
4624 * common/config/nds32: New directory and files.
4626 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
4627 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
4628 * doc/install.texi (Cross-Compiler-Specific Options): Document
4629 --with-nds32-lib for nds32 target.
4630 * doc/extend.texi (Function Attributes, Target Builtins): Document
4631 nds32 specific attributes.
4633 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
4635 * lra-constraints (process_alt_operands): Use the result
4636 elimination register for operand when matching constraints.
4638 2013-10-31 Jakub Jelinek <jakub@redhat.com>
4640 * tree-vrp.c (maybe_set_nonzero_bits): New function.
4641 (remove_range_assertions): Call it.
4643 * tree.c (tree_ctz): New function.
4644 * tree.h (tree_ctz): New prototype.
4645 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
4646 first argument from tree to const_tree.
4647 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
4648 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
4649 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
4651 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
4652 * tree-vect-loop.c (vect_analyze_loop_operations,
4653 vect_transform_loop): Don't force scalar loop for bound just because
4654 number of iterations is unknown, only do it if it is not known to be
4655 a multiple of vectorization_factor.
4656 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
4658 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
4659 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
4660 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
4661 (set_nonzero_bits, get_nonzero_bits): New prototypes.
4662 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
4663 a default def isn't partially constant.
4664 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
4665 is known to be partially zero.
4666 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
4667 to see if a default def isn't partially constant.
4668 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
4669 creation of a range, if VR_RANGE, try to improve nonzero_bits from
4671 (set_nonzero_bits, get_nonzero_bits): New functions.
4673 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
4674 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
4675 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
4676 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
4677 uses but in the condition and ASSERT_EXPR and the other successor of
4678 the predecessor bb is __builtin_unreachable (), set_range_info of the
4679 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
4681 2013-10-31 Martin Jambor <mjambor@suse.cz>
4683 PR rtl-optimization/58934
4685 2013-10-30 Martin Jambor <mjambor@suse.cz>
4686 PR rtl-optimization/10474
4687 * ira.c (find_moveable_pseudos): Do not calculate dominance info
4689 (interesting_dest_for_shprep): New function.
4690 (split_live_ranges_for_shrink_wrap): Likewise.
4691 (ira): Calculate dominance info and df analysis. Call
4692 split_live_ranges_for_shrink_wrap.
4694 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
4695 Yury Gribov <y.gribov@samsung.com>
4698 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
4699 shadow pointer to avoid clobbering the main one.
4701 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4703 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
4705 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
4708 * ira-color.c (update_costs_from_copies): Add new parameter. Use
4709 it for calling update_costs_from_allocno.
4710 (assign_hard_reg): Call restore_costs_from_copies only for
4711 !retry_p. Pass new argument to update_costs_from_copies.
4712 (color_pass): Pass new argument to update_costs_from_copies.
4713 (ira_mark_allocation_change): Ditto.
4715 2013-10-30 Sharad Singhai <singhai@google.com>
4718 * opts.c (common_handle_option): Remove deprecated option
4719 -ftree-vectorizer-verbose.
4720 * doc/invoke.texi (Debugging Options): Ditto.
4721 * opts-global.c (handle_common_deferred_options): Ditto.
4722 (dump_remap_tree_vectorizer_verbose): Delete.
4723 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
4725 2013-10-30 DJ Delorie <dj@redhat.com>
4727 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
4728 that take no arguments.
4730 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
4733 * reload1.c (update_eliminables_and_spill): New function, broken
4735 (reload): Use it. Check for frame size change after frame size
4736 alignment, and call update_eliminables_and_spill first if continue-ing.
4738 2013-10-30 Cong Hou <congh@google.com>
4741 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
4742 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
4743 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
4745 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
4747 PR rtl-optimization/58369
4748 * reload1.c (compute_reload_subreg_offset): New function.
4749 (choose_reload_regs): Use it to pass endian-correct
4750 offset to subreg_regno_offset.
4752 2013-10-30 Tobias Burnus <burnus@net-b.de>
4755 * tree-cfg.c (replace_loop_annotate): Replace warning by
4758 2013-10-30 Jason Merrill <jason@redhat.com>
4760 * configure.ac (loose_warn): Add -Wno-format if
4761 --disable-build-format-warnings.
4763 2013-10-30 David Malcolm <dmalcolm@redhat.com>
4765 * cgraphunit.c (analyze_functions): Split symtab_node declarations
4766 onto multiple lines to make things easier for rename_symtab.py.
4768 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
4769 (symtab_semantically_equivalent_p): Likewise.
4771 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
4774 * lra.c (check_rtl): Remove address check before LRA work.
4776 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
4778 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
4779 POINTER_PLUS_EXPR in the defining statement.
4781 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
4783 * regmove.c: Remove.
4784 * tree-pass.h (make_pass_regmove): Remove.
4785 * timevar.def (TV_REGMOVE): Remove.
4786 * passes.def (pass_regmove): Remove.
4787 * opts.c (default_options_table): Remove entry for regmove.
4788 * doc/passes.texi: Remove regmove pass description.
4789 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
4791 (-fdump-rtl-regmove): Ditto.
4792 * common.opt (foptimize-register-move, fregmove): Ignore.
4793 * Makefile.in (OBJS): Remove regmove.o.
4794 * regmove.c: Remove.
4795 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
4797 (struct ira_allocno) New member allocno_prefs.
4798 (ALLOCNO_PREFS): New macro.
4799 (ira_prefs, ira_prefs_num): New external vars.
4800 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
4801 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
4802 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
4804 (ira_add_allocno_copy_to_list): Remove prototype.
4805 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
4806 (ira_pref_iterator): New type.
4807 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
4808 (FOR_EACH_PREF): New macro.
4809 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
4810 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
4812 (ira_setup_alts): New function.
4813 (decrease_live_ranges_number): New function.
4814 (ira): Call the above function.
4815 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
4816 (ira_create_allocno): Initialize allocno prefs.
4817 (pref_pool, pref_vec): New static vars.
4818 (initiate_prefs, find_allocno_pref, ira_create_pref): New
4820 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
4821 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
4822 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
4823 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
4824 (ira_add_allocno_copy_to_list): Make static. Rename to
4825 add_allocno_copy_to_list.
4826 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
4827 swap_allocno_copy_ends_if_necessary.
4828 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
4829 ira_remove_allocno_prefs.
4830 (ira_flattening): Ditto.
4831 (ira_build): Call initiate_prefs, print_prefs.
4832 (ira_destroy): Call finish_prefs.
4833 * ira-color.c (struct update_cost_record): New.
4834 (struct allocno_color_data): Add new member update_cost_records.
4835 (update_cost_record_pool): New static var.
4836 (init_update_cost_records, get_update_cost_record): New functions.
4837 (free_update_cost_record_list, finish_update_cost_records): Ditto.
4838 (struct update_cost_queue_elem): Add member from.
4839 (initiate_cost_update): Call init_update_cost_records.
4840 (finish_cost_update): Call finish_update_cost_records.
4841 (queue_update_cost, get_next_update_cost): Add new param from.
4842 (Update_allocno_cost, update_costs_from_allocno): New functions.
4843 (update_costs_from_prefs): Ditto.
4844 (update_copy_costs): Rename to update_costs_from_copies.
4845 (restore_costs_from_copies): New function.
4846 (update_conflict_hard_regno_costs): Don't go back.
4847 (assign_hard_reg): Call restore_costs_from_copies. Add printing
4849 (pop_allocnos): Add priniting more debug info.
4850 (color_allocnos): Remove prefs for conflicting hard regs.
4851 Call update_costs_from_prefs.
4852 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
4853 (get_dup_num): Rename, modify, and move to ira.c
4854 (process_regs_for_copy): Add prefs.
4855 (add_insn_allocno_copies): Put src as first arg of
4856 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
4857 * ira-costs.c (record_reg_classes): Modify and move code into
4858 record_operands_costs.
4859 (find_costs_and_classes): Create prefs for the hard reg of small
4861 (process_bb_node_for_hard_reg_moves): Add prefs.
4863 2013-10-30 Richard Biener <rguenther@suse.de>
4866 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
4867 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
4869 (pre_and_rev_post_order_compute): ... which now wraps it.
4870 * graph.c (draw_cfg_nodes_no_loops): Use
4871 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
4874 2013-10-30 Christian Bruel <christian.bruel@st.com>
4876 * config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
4877 (sh_expand_cmpstr): Handle known align and schedule improvements.
4878 * config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
4879 * config/sh/sh.md (cmpstrnsi): New pattern.
4881 2013-10-30 Martin Jambor <mjambor@suse.cz>
4883 PR rtl-optimization/10474
4884 * ira.c (find_moveable_pseudos): Do not calculate dominance info
4886 (interesting_dest_for_shprep): New function.
4887 (split_live_ranges_for_shrink_wrap): Likewise.
4888 (ira): Calculate dominance info and df analysis. Call
4889 split_live_ranges_for_shrink_wrap.
4891 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4894 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
4896 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
4898 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
4899 * tree.h (POINTER_BOUNDS_P): New.
4900 (BOUNDED_TYPE_P): New.
4902 (pointer_bounds_type_node): New.
4903 * tree.c (build_common_tree_nodes): Initialize
4904 pointer_bounds_type_node.
4905 * gimple.h (gimple_call_get_nobnd_arg_index): New.
4906 (gimple_call_num_nobnd_args): New.
4907 (gimple_call_nobnd_arg): New.
4908 (gimple_return_retbnd): New.
4909 (gimple_return_set_retbnd): New
4910 * gimple.c (gimple_build_return): Increase number of ops
4911 for return statement.
4912 (gimple_call_get_nobnd_arg_index): New.
4913 * gimple-pretty-print.c (dump_gimple_return): Print second op.
4915 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
4917 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
4918 with "chkp ctor" and "bnd_legacy" attributes.
4919 * gimplify.c (gimplify_init_constructor): Avoid infinite
4920 loop during gimplification of bounds initializer.
4922 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
4924 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
4925 (handle_bnd_legacy): New.
4926 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
4927 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
4930 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
4932 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
4933 (BT_FN_PTR_CONST_PTR): New.
4934 (BT_FN_CONST_PTR_CONST_PTR): New.
4935 (BT_FN_PTR_CONST_PTR_SIZE): New.
4936 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
4937 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
4938 (BT_FN_VOID_CONST_PTR_SIZE): New.
4939 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
4940 * chkp-builtins.def: New.
4941 * builtins.def: include chkp-builtins.def.
4942 (DEF_CHKP_BUILTIN): New.
4943 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
4944 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
4945 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
4946 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
4947 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
4948 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
4949 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
4950 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
4951 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
4952 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
4953 * common.opt (fcheck-pointer-bounds): New.
4954 * toplev.c (process_options): Check Pointer Bounds Checker is
4956 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
4958 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
4960 * target.def (builtin_chkp_function): New.
4961 (chkp_bound_type): New.
4962 (chkp_bound_mode): New.
4963 (fn_abi_va_list_bounds_size): New.
4964 (load_bounds_for_arg): New.
4965 (store_bounds_for_arg): New.
4966 * targhooks.h (default_load_bounds_for_arg): New.
4967 (default_store_bounds_for_arg): New.
4968 (default_fn_abi_va_list_bounds_size): New.
4969 (default_chkp_bound_type): New.
4970 (default_chkp_bound_mode): New.
4971 (default_builtin_chkp_function): New.
4972 * targhooks.c (default_load_bounds_for_arg): New.
4973 (default_store_bounds_for_arg): New.
4974 (default_fn_abi_va_list_bounds_size): New.
4975 (default_chkp_bound_type): New.
4976 (default_chkp_bound_mode); New.
4977 (default_builtin_chkp_function): New.
4978 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
4979 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
4980 (TARGET_STORE_BOUNDS_FOR_ARG): New.
4981 (TARGET_BUILTIN_CHKP_FUNCTION): New.
4982 (TARGET_CHKP_BOUND_TYPE): New.
4983 (TARGET_CHKP_BOUND_MODE): New.
4984 * doc/tm.texi: Regenerated.
4985 * langhooks.h (lang_hooks): Add chkp_supported field.
4986 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
4987 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
4989 2013-10-29 Andrew Pinski <apinski@cavium.com>
4991 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
4992 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
4993 fails, combining the branches anyways.
4994 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
4995 increases the number of combinings.
4996 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
4998 2013-10-29 Mike Stump <mikestump@comcast.net>
5000 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
5001 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
5002 (make_vector_mode): Increase namebuf to 16.
5003 (emit_insn_modes_h): When processing BImode, don't
5004 also match partial int modes.
5005 (emit_class_narrowest_mode): Likewise.
5007 * config/bfin/bfin-modes.def: Add precision to PDI.
5008 * config/m32c/m32c-modes.def: Add precision to PSI.
5009 * config/msp430/msp430-modes.def: Add precision to PSI.
5010 * config/rs6000/rs6000-modes.def: Add precision to PTI.
5011 * config/sh/sh-modes.def: Add precision to PSI and PDI.
5013 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
5016 * config/sh/sh.md (*addc): Rename existing variations to ...
5017 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
5018 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
5019 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
5020 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
5022 2013-10-29 Teresa Johnson <tejohnson@google.com>
5025 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
5026 insanities when updating probabilities.
5028 2013-10-29 David Malcolm <dmalcolm@redhat.com>
5030 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
5031 prettyprinter for cgraph_node to reflect the conversion of the
5032 symtable types to a C++ class hierarchy: it now *is* a
5033 symtab_node_base, rather than having one (named "symbol").
5035 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
5037 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
5038 __cilkrts_pop_frame. If matched, then return true for built-in
5040 (expand_builtin): Added BUILT_IN_CILK_DETACH and
5041 BUILT_IN_CILK_POP_FRAME case.
5042 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
5043 (lhs_cilk_detect_spawn): Likewise.
5044 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
5045 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
5046 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
5047 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
5048 (LANG_HOOKS_CILKPLUS): Likewise.
5049 * tree.h (CILK_SPAWN_FN): Likewise.
5050 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
5051 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
5052 (OBJS): Added cilk-common.o.
5053 (BUILTINS_DEF): Added cilk-builtins.def.
5054 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
5055 (lhd_cilk_detect_spawn): Likewise.
5056 * langhooks.h (lang_hooks_for_cilkplus): New struct.
5057 (struct lang_hooks): Added new field called "cilkplus."
5058 * cilk-common.c: New file.
5060 * cilk-builtins.def: Likewise.
5061 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
5062 "__cilk" macro and set it to 200.
5063 * function.h (struct function::cilk_frame_decl): New field.
5064 (struct function::is_cilk_function): Likewise.
5065 (struct function::calls_cilk_spawn): Likewise.
5066 * gimplify.c (gimplify_call_expr): Added a check if the function call
5067 being gimplified is a spawn detach point. If so, then add pop_frame
5068 and detach function calls.
5069 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
5070 for gimplifying _Cilk_spawn and _Cilk_sync statements.
5071 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
5072 function. If so, added a _Cilk_sync and gimplified it.
5073 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
5074 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
5075 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
5077 (can_inline_edge_p): Prevent inling of spawnee function.
5078 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
5079 for functions that use Cilk keywords.
5080 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
5081 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
5082 CILK_SYNC_STMT cases.
5083 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
5085 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
5086 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
5087 * passes.texi (Cilk Keywords): New section that describes the compiler
5088 code changes for handling Cilk Keywords.
5090 2013-10-29 David Malcolm <dmalcolm@redhat.com>
5092 Patch autogenerated by refactor_symtab.py from
5093 https://github.com/davidmalcolm/gcc-refactoring-scripts
5094 revision 58bb219cc090b2f4516a9297d868c245495ee622
5096 * asan.c (asan_finish_file): Update for conversion of symtab types to
5097 a true class hierarchy.
5098 * cfgexpand.c (estimated_stack_frame_size): Likewise.
5099 * cgraph.c (cgraph_get_body): Likewise.
5100 (cgraph_get_create_real_symbol_node): Likewise.
5101 (verify_cgraph_node): Likewise.
5102 (verify_edge_corresponds_to_fndecl): Likewise.
5103 (verify_edge_count_and_frequency): Likewise.
5104 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
5105 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
5106 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
5107 (cgraph_node_cannot_return): Likewise.
5108 (cgraph_set_pure_flag_1): Likewise.
5109 (cgraph_set_const_flag_1): Likewise.
5110 (cgraph_set_nothrow_flag_1): Likewise.
5111 (cgraph_make_node_local_1): Likewise.
5112 (cgraph_for_node_and_aliases): Likewise.
5113 (cgraph_for_node_thunks_and_aliases): Likewise.
5114 (cgraph_node_can_be_local_p): Likewise.
5115 (cgraph_node_cannot_be_local_p_1): Likewise.
5116 (cgraph_function_body_availability): Likewise.
5117 (dump_cgraph_node): Likewise.
5118 (cgraph_rtl_info): Likewise.
5119 (cgraph_mark_address_taken_node): Likewise.
5120 (cgraph_remove_node): Likewise.
5121 (cgraph_release_function_body): Likewise.
5122 (cgraph_update_edges_for_call_stmt_node): Likewise.
5123 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
5124 (cgraph_make_edge_direct): Likewise.
5125 (cgraph_resolve_speculation): Likewise.
5126 (cgraph_speculative_call_info): Likewise.
5127 (cgraph_turn_edge_to_speculative): Likewise.
5128 (cgraph_create_edge_1): Likewise.
5129 (cgraph_set_call_stmt): Likewise.
5130 (cgraph_node_for_asm): Likewise.
5131 (cgraph_add_thunk): Likewise.
5132 (cgraph_same_body_alias): Likewise.
5133 (cgraph_create_function_alias): Likewise.
5134 (cgraph_create_node): Likewise.
5135 (cgraph_create_empty_node): Likewise.
5136 (record_function_versions): Likewise.
5137 (used_from_object_file_p): Likewise.
5138 * cgraph.h (symtab_can_be_discarded): Likewise.
5139 (symtab_real_symbol_p): Likewise.
5140 (cgraph_mark_force_output_node): Likewise.
5141 (cgraph_edge_recursive_p): Likewise.
5142 (symtab_alias_target): Likewise.
5143 (varpool_all_refs_explicit_p): Likewise.
5144 (varpool_can_remove_if_no_refs): Likewise.
5145 (cgraph_only_called_directly_or_aliased_p): Likewise.
5146 (cgraph_next_function_with_gimple_body): Likewise.
5147 (cgraph_first_function_with_gimple_body): Likewise.
5148 (cgraph_function_with_gimple_body_p): Likewise.
5149 (cgraph_next_function): Likewise.
5150 (cgraph_first_function): Likewise.
5151 (cgraph_next_defined_function): Likewise.
5152 (cgraph_first_defined_function): Likewise.
5153 (varpool_next_defined_variable): Likewise.
5154 (varpool_first_defined_variable): Likewise.
5155 (varpool_next_static_initializer): Likewise.
5156 (varpool_first_static_initializer): Likewise.
5157 (varpool_next_variable): Likewise.
5158 (varpool_first_variable): Likewise.
5159 (varpool_node_name): Likewise.
5160 (varpool): Likewise.
5162 (is_a_helper <varpool_node>::test): Likewise.
5163 (is_a_helper <cgraph_node>::test): Likewise.
5164 (varpool_variable_node): Likewise.
5165 (cgraph_function_or_thunk_node): Likewise.
5166 (varpool_alias_target): Likewise.
5167 (cgraph_alias_target): Likewise.
5168 (cgraph_node_name): Likewise.
5169 (varpool_node_asm_name): Likewise.
5170 (cgraph_node_asm_name): Likewise.
5171 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
5172 (cgraph_rebuild_references): Likewise.
5173 (rebuild_cgraph_edges): Likewise.
5174 (record_eh_tables): Likewise.
5175 (build_cgraph_edges): Likewise.
5176 (mark_store): Likewise.
5177 (mark_load): Likewise.
5178 (mark_address): Likewise.
5179 (record_type_list): Likewise.
5180 (record_reference): Likewise.
5181 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
5182 (cgraph_materialize_clone): Likewise.
5183 (cgraph_function_versioning): Likewise.
5184 (cgraph_copy_node_for_versioning): Likewise.
5185 (update_call_expr): Likewise.
5186 (cgraph_find_replacement_node): Likewise.
5187 (cgraph_create_virtual_clone): Likewise.
5188 (cgraph_clone_node): Likewise.
5189 * cgraphunit.c (compile): Likewise.
5190 (output_weakrefs): Likewise.
5191 (output_in_order): Likewise.
5192 (expand_function): Likewise.
5193 (assemble_thunks_and_aliases): Likewise.
5194 (expand_thunk): Likewise.
5195 (mark_functions_to_output): Likewise.
5196 (handle_alias_pairs): Likewise.
5197 (analyze_functions): Likewise.
5198 (walk_polymorphic_call_targets): Likewise.
5199 (varpool_finalize_decl): Likewise.
5200 (process_function_and_variable_attributes): Likewise.
5201 (cgraph_process_same_body_aliases): Likewise.
5202 (analyze_function): Likewise.
5203 (cgraph_add_new_function): Likewise.
5204 (cgraph_finalize_function): Likewise.
5205 (referred_to_p): Likewise.
5206 (cgraph_reset_node): Likewise.
5207 (cgraph_process_new_functions): Likewise.
5208 (enqueue_node): Likewise.
5209 (decide_is_symbol_needed): Likewise.
5210 * coverage.c (coverage_compute_profile_id): Likewise.
5211 * dbxout.c (dbxout_expand_expr): Likewise.
5212 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
5213 (reference_to_unused): Likewise.
5214 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
5215 * gimplify.c (unvisit_body): Likewise.
5216 (unshare_body): Likewise.
5217 * ipa-cp.c (ipcp_generate_summary): Likewise.
5218 (ipcp_decision_stage): Likewise.
5219 (identify_dead_nodes): Likewise.
5220 (decide_whether_version_node): Likewise.
5221 (decide_about_value): Likewise.
5222 (perhaps_add_new_callers): Likewise.
5223 (create_specialized_node): Likewise.
5224 (update_profiling_info): Likewise.
5225 (ipcp_propagate_stage): Likewise.
5226 (estimate_local_effects): Likewise.
5227 (good_cloning_opportunity_p): Likewise.
5228 (devirtualization_time_bonus): Likewise.
5229 (propagate_constants_accross_call): Likewise.
5230 (initialize_node_lattices): Likewise.
5231 (ipcp_cloning_candidate_p): Likewise.
5232 (determine_versionability): Likewise.
5233 (print_all_lattices): Likewise.
5234 (print_lattice): Likewise.
5235 (ipcp_discover_new_direct_edges): Likewise.
5236 * ipa-devirt.c (ipa_devirt): Likewise.
5237 (likely_target_p): Likewise.
5238 (update_type_inheritance_graph): Likewise.
5239 (possible_polymorphic_call_target_p): Likewise.
5240 (dump_possible_polymorphic_call_targets): Likewise.
5241 (devirt_variable_node_removal_hook): Likewise.
5242 (record_binfo): Likewise.
5243 (maybe_record_node): Likewise.
5244 (build_type_inheritance_graph): Likewise.
5245 * ipa-inline-analysis.c (inline_write_summary): Likewise.
5246 (inline_generate_summary): Likewise.
5247 (inline_analyze_function): Likewise.
5248 (do_estimate_growth): Likewise.
5249 (simple_edge_hints): Likewise.
5250 (estimate_node_size_and_time): Likewise.
5251 (estimate_edge_devirt_benefit): Likewise.
5252 (compute_inline_parameters): Likewise.
5253 (estimate_function_body_sizes): Likewise.
5254 (compute_bb_predicates): Likewise.
5255 (initialize_inline_failed): Likewise.
5256 (dump_inline_summary): Likewise.
5257 (dump_inline_edge_summary): Likewise.
5258 * ipa-inline-transform.c (inline_transform): Likewise.
5259 (preserve_function_body_p): Likewise.
5260 (save_inline_function_body): Likewise.
5261 (inline_call): Likewise.
5262 (clone_inlined_nodes): Likewise.
5263 (can_remove_node_now_p): Likewise.
5264 (can_remove_node_now_p_1): Likewise.
5265 * ipa-inline.c (early_inliner): Likewise.
5266 (early_inline_small_functions): Likewise.
5267 (inline_always_inline_functions): Likewise.
5268 (ipa_inline): Likewise.
5269 (flatten_function): Likewise.
5270 (inline_small_functions): Likewise.
5271 (speculation_useful_p): Likewise.
5272 (recursive_inlining): Likewise.
5273 (update_caller_keys): Likewise.
5274 (reset_edge_caches): Likewise.
5275 (update_edge_key): Likewise.
5276 (edge_badness): Likewise.
5277 (relative_time_benefit): Likewise.
5278 (want_inline_self_recursive_call_p): Likewise.
5279 (want_inline_small_function_p): Likewise.
5280 (want_early_inline_function_p): Likewise.
5281 (num_calls): Likewise.
5282 (can_early_inline_edge_p): Likewise.
5283 (can_inline_edge_p): Likewise.
5284 (report_inline_failed_reason): Likewise.
5285 * ipa-profile.c (ipa_profile): Likewise.
5286 (ipa_propagate_frequency): Likewise.
5287 (ipa_propagate_frequency_1): Likewise.
5288 (ipa_profile_generate_summary): Likewise.
5289 * ipa-prop.c (ipcp_transform_function): Likewise.
5290 (read_replacements_section): Likewise.
5291 (ipa_prop_read_section): Likewise.
5292 (ipa_modify_call_arguments): Likewise.
5293 (ipa_print_node_params): Likewise.
5294 (propagate_controlled_uses): Likewise.
5295 (update_indirect_edges_after_inlining): Likewise.
5296 (remove_described_reference): Likewise.
5297 (ipa_make_edge_direct_to_target): Likewise.
5298 (ipa_analyze_node): Likewise.
5299 (ipa_analyze_params_uses): Likewise.
5300 (ipa_compute_jump_functions): Likewise.
5301 (ipa_get_callee_param_type): Likewise.
5302 (ipa_print_node_jump_functions): Likewise.
5303 (ipa_initialize_node_params): Likewise.
5304 (ipa_populate_param_decls): Likewise.
5305 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
5306 (write_agg_replacement_chain): Likewise.
5307 (ipa_write_node_info): Likewise.
5308 (ipa_edge_duplication_hook): Likewise.
5309 (try_decrement_rdesc_refcount): Likewise.
5310 * ipa-pure-const.c (propagate_nothrow): Likewise.
5311 (propagate_pure_const): Likewise.
5312 (pure_const_read_summary): Likewise.
5313 (pure_const_write_summary): Likewise.
5314 (analyze_function): Likewise.
5315 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
5316 (ipa_ref_referring_ref_list): Likewise.
5317 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
5318 (ipa_remove_stmt_references): Likewise.
5319 (ipa_find_reference): Likewise.
5320 (ipa_dump_referring): Likewise.
5321 (ipa_dump_references): Likewise.
5322 (ipa_record_reference): Likewise.
5323 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
5324 (ipa_reference_write_optimization_summary): Likewise.
5325 (write_node_summary_p): Likewise.
5326 (propagate): Likewise.
5327 (read_write_all_from_decl): Likewise.
5328 (generate_summary): Likewise.
5329 (analyze_function): Likewise.
5330 (propagate_bits): Likewise.
5331 (ipa_reference_get_not_written_global): Likewise.
5332 (ipa_reference_get_not_read_global): Likewise.
5333 * ipa-split.c (execute_split_functions): Likewise.
5334 (split_function): Likewise.
5335 * ipa-utils.c (ipa_merge_profiles): Likewise.
5336 (dump_cgraph_node_set): Likewise.
5337 (ipa_reverse_postorder): Likewise.
5338 (ipa_edge_within_scc): Likewise.
5339 (ipa_get_nodes_in_cycle): Likewise.
5340 (ipa_free_postorder_info): Likewise.
5341 (ipa_reduced_postorder): Likewise.
5342 (searchc): Likewise.
5343 (recursive_call_p): Likewise.
5344 * ipa.c (ipa_cdtor_merge): Likewise.
5345 (record_cdtor_fn): Likewise.
5346 (function_and_variable_visibility): Likewise.
5347 (varpool_externally_visible_p): Likewise.
5348 (cgraph_externally_visible_p): Likewise.
5349 (comdat_can_be_unshared_p): Likewise.
5350 (comdat_can_be_unshared_p_1): Likewise.
5351 (address_taken_from_non_vtable_p): Likewise.
5352 (ipa_discover_readonly_nonaddressable_vars): Likewise.
5353 (symtab_remove_unreachable_nodes): Likewise.
5354 (walk_polymorphic_call_targets): Likewise.
5355 (process_references): Likewise.
5356 (enqueue_node): Likewise.
5357 (has_addr_references_p): Likewise.
5358 (cgraph_non_local_node_p_1): Likewise.
5359 * is-a.h (varpool_analyze_node): Likewise.
5360 * lto-cgraph.c (input_symtab): Likewise.
5361 (merge_profile_summaries): Likewise.
5362 (input_cgraph_1): Likewise.
5363 (input_edge): Likewise.
5364 (input_varpool_node): Likewise.
5365 (input_node): Likewise.
5366 (input_overwrite_node): Likewise.
5367 (compute_ltrans_boundary): Likewise.
5368 (output_refs): Likewise.
5369 (lto_output_varpool_node): Likewise.
5370 (lto_output_node): Likewise.
5371 (reachable_from_other_partition_p): Likewise.
5372 (referenced_from_other_partition_p): Likewise.
5373 (lto_output_edge): Likewise.
5374 (output_node_opt_summary): Likewise.
5375 (add_node_to): Likewise.
5376 (reachable_from_this_partition_p): Likewise.
5377 (lto_set_symtab_encoder_in_partition): Likewise.
5378 (lto_symtab_encoder_in_partition_p): Likewise.
5379 (lto_set_symtab_encoder_encode_initializer): Likewise.
5380 (lto_symtab_encoder_encode_initializer_p): Likewise.
5381 (lto_set_symtab_encoder_encode_body): Likewise.
5382 (lto_symtab_encoder_encode_body_p): Likewise.
5383 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
5385 * lto-streamer-in.c (lto_read_body): Likewise.
5386 (fixup_call_stmt_edges): Likewise.
5387 (fixup_call_stmt_edges_1): Likewise.
5388 * lto-streamer-out.c (produce_symtab): Likewise.
5389 (output_symbol_p): Likewise.
5390 (write_symbol): Likewise.
5391 (lto_output): Likewise.
5392 (copy_function): Likewise.
5393 (output_function): Likewise.
5394 * passes.c (function_called_by_processed_nodes_p): Likewise.
5395 (ipa_write_optimization_summaries): Likewise.
5396 (ipa_write_summaries): Likewise.
5397 (do_per_function_toporder): Likewise.
5398 (do_per_function): Likewise.
5399 (dump_passes): Likewise.
5400 * symtab.c (symtab_semantically_equivalent_p): Likewise.
5401 (symtab_nonoverwritable_alias): Likewise.
5402 (symtab_nonoverwritable_alias_1): Likewise.
5403 (symtab_for_node_and_aliases): Likewise.
5404 (symtab_resolve_alias): Likewise.
5405 (fixup_same_cpp_alias_visibility): Likewise.
5406 (symtab_alias_ultimate_target): Likewise.
5407 (symtab_used_from_object_file_p): Likewise.
5408 (verify_symtab_base): Likewise.
5409 (dump_symtab_base): Likewise.
5410 (symtab_node_name): Likewise.
5411 (symtab_node_asm_name): Likewise.
5412 (symtab_dissolve_same_comdat_group_list): Likewise.
5413 (symtab_add_to_same_comdat_group): Likewise.
5414 (symtab_unregister_node): Likewise.
5415 (symtab_insert_node_to_hashtable): Likewise.
5416 (symtab_register_node): Likewise.
5417 (unlink_from_assembler_name_hash): Likewise.
5418 (insert_to_assembler_name_hash): Likewise.
5419 (eq_assembler_name): Likewise.
5420 (hash_node_by_assembler_name): Likewise.
5421 (eq_node): Likewise.
5422 (hash_node): Likewise.
5423 * toplev.c (wrapup_global_declaration_2): Likewise.
5424 * trans-mem.c (ipa_tm_execute): Likewise.
5425 (ipa_tm_transform_clone): Likewise.
5426 (ipa_tm_transform_transaction): Likewise.
5427 (ipa_tm_transform_calls_redirect): Likewise.
5428 (ipa_tm_insert_gettmclone_call): Likewise.
5429 (ipa_tm_insert_irr_call): Likewise.
5430 (ipa_tm_create_version): Likewise.
5431 (ipa_tm_create_version_alias): Likewise.
5432 (ipa_tm_mark_forced_by_abi_node): Likewise.
5433 (ipa_tm_mark_force_output_node): Likewise.
5434 (ipa_tm_diagnose_tm_safe): Likewise.
5435 (ipa_tm_mayenterirr_function): Likewise.
5436 (ipa_tm_scan_irr_function): Likewise.
5437 (ipa_tm_note_irrevocable): Likewise.
5438 (ipa_tm_scan_calls_clone): Likewise.
5439 (get_cg_data): Likewise.
5440 * tree-eh.c (tree_could_trap_p): Likewise.
5441 * tree-emutls.c (ipa_lower_emutls): Likewise.
5442 (create_emultls_var): Likewise.
5443 (lower_emutls_function_body): Likewise.
5444 (gen_emutls_addr): Likewise.
5445 (emutls_decl): Likewise.
5446 (new_emutls_decl): Likewise.
5447 * tree-inline.c (tree_function_versioning): Likewise.
5448 (optimize_inline_calls): Likewise.
5449 (expand_call_inline): Likewise.
5450 (estimate_num_insns): Likewise.
5451 (copy_bb): Likewise.
5452 (delete_unreachable_blocks_update_callgraph): Likewise.
5453 * tree-nested.c (gimplify_all_functions): Likewise.
5454 (create_nesting_tree): Likewise.
5455 (check_for_nested_with_variably_modified): Likewise.
5456 * tree-pretty-print.c (dump_function_header): Likewise.
5457 * tree-profile.c (tree_profiling): Likewise.
5458 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
5459 (modify_function): Likewise.
5460 (convert_callers): Likewise.
5461 (convert_callers_for_node): Likewise.
5462 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
5463 (associate_varinfo_to_alias): Likewise.
5464 (create_variable_info_for): Likewise.
5465 (get_constraint_for_ssa_var): Likewise.
5466 * tree-vectorizer.c (increase_alignment): Likewise.
5467 * tree.c (find_decls_types_in_var): Likewise.
5468 (find_decls_types_in_node): Likewise.
5469 (free_lang_data_in_decl): Likewise.
5470 * value-prof.c (gimple_ic_transform): Likewise.
5471 (gimple_ic): Likewise.
5472 (check_ic_target): Likewise.
5473 (init_node_map): Likewise.
5474 * varasm.c (decl_binds_to_current_def_p): Likewise.
5475 (default_binds_local_p_1): Likewise.
5476 (dump_tm_clone_pairs): Likewise.
5477 (assemble_alias): Likewise.
5478 (find_decl): Likewise.
5479 (mark_decl_referenced): Likewise.
5480 * varpool.c (varpool_for_node_and_aliases): Likewise.
5481 (varpool_extra_name_alias): Likewise.
5482 (varpool_create_variable_alias): Likewise.
5483 (add_new_static_var): Likewise.
5484 (varpool_finalize_named_section_flags): Likewise.
5485 (varpool_remove_unreferenced_decls): Likewise.
5486 (enqueue_node): Likewise.
5487 (varpool_assemble_decl): Likewise.
5488 (assemble_aliases): Likewise.
5489 (varpool_analyze_node): Likewise.
5490 (cgraph_variable_initializer_availability): Likewise.
5491 (varpool_add_new_variable): Likewise.
5492 (ctor_for_folding): Likewise.
5493 (dump_varpool_node): Likewise.
5494 (varpool_remove_initializer): Likewise.
5495 (varpool_remove_node): Likewise.
5496 (varpool_node_for_decl): Likewise.
5497 (varpool_create_empty_node): Likewise.
5498 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
5499 (ix86_get_function_versions_dispatcher): Likewise.
5501 2013-10-29 David Malcolm <dmalcolm@redhat.com>
5503 * cgraph.h (symtab_node_base): Convert to a class;
5504 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
5505 chain_next/prev from symtab_node_def.
5506 (cgraph_node): Inherit from symtab_node; add GTY option
5507 tag ("SYMTAB_FUNCTION").
5508 (varpool_node): Inherit from symtab_node; add GTY option
5509 tag ("SYMTAB_VARIABLE").
5510 (symtab_node_def): Remove.
5511 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
5512 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
5513 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
5514 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
5516 * ipa-ref.h (symtab_node_def): Drop.
5517 (symtab_node): Change underlying type from symtab_node_def to
5519 (const_symtab_node): Likwise.
5521 * is-a.h: Update examples in comment.
5523 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
5524 (assembler_name_hash): Likewise.
5526 2013-10-29 Martin Liska <marxin.liska@gmail.com>
5528 * doc/tree-ssa.texi (gimple_phi_result): Document.
5529 (gimple_phi_num_args, gimple_phi_arg): Likewise.
5530 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
5531 (PHI_RESULT, PHI_NUM_ARGS): Remove.
5532 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
5534 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
5536 * expr.h: Revert change and include tree-core.h.
5537 * rtl.h: Revert change and don't include tree-core.h.
5539 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
5541 * config/darwin.c: Include gimple.h.
5542 * config/i386/winnt.c: Likewise.
5544 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
5546 PR tree-optimization/19831
5547 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
5549 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
5551 * tree-outof-ssa.h: Remove include files.
5552 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
5554 * tree-ssa-coalesce.c: Likewise.
5555 * cfgexpand.c: Likewise.
5556 * tree-ssa-ter.c: Likewise.
5557 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
5558 * lto-streamer.h: Likewise.
5559 * cgraphbuild.c: Add gimple.h to include list.
5560 * data-streamer-in.c: Likewise.
5561 * ipa-cp.c: Likewise.
5562 * tree-streamer.c: Likewise.
5563 * lto-compress.c: Likewise.
5564 * ipa-reference.c: Likewise.
5565 * data-streamer-out.c: Likewise.
5566 * lto-cgraph.c: Likewise.
5567 * cgraphclones.c: Likewise.
5568 * ipa-utils.c: Likewise.
5569 * data-streamer.c: Likewise.
5570 * ipa-split.c: Likewise.
5571 * lto-section-in.c: Likewise.
5572 * tree-streamer-out.c: Likewise.
5573 * ipa-prop.c: Likewise.
5574 * tree-streamer-in.c: Likewise.
5575 * symtab.c: Likewise.
5576 * opts-global.c: Likewise.
5577 * lto-opts.c: Likewise.
5578 * lto-section-out.c: Likewise.
5579 * lto-streamer.c: Likewise.
5580 * rtl.h: Add tree-core.h to include list.
5581 * expr.h: Remove tree-core.h from include list.
5582 * gimple.h: Likewise.
5583 * ipa-utils.h: Likewise.
5584 * streamer-hooks.h: Likewise.
5585 * streamer-hooks.c: Include input.h.
5587 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5589 * config/arm/arm.c (cortexa7_extra_costs): New table.
5590 (arm_cortex_a7_tune): New.
5591 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
5593 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
5595 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
5597 2013-10-29 David Malcolm <dmalcolm@redhat.com>
5599 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
5600 to use autogenerated markers for a class-hierarchy, every class
5601 must have a GTY marker.
5602 * gengtype.h (struct type): Add linked list of subclasses to
5603 the "s" member of the union.
5604 (add_subclass): New decl.
5605 * gengtype-state.c (read_state_struct_type): Set up subclass
5607 * gengtype.c (get_ultimate_base_class): New.
5608 (add_subclass): New.
5609 (new_structure): Set up subclass linked list.
5610 (set_gc_used_type): Propagate usage information to subclasses.
5611 (output_mangled_typename): Use get_ultimate_base_class.
5612 (walk_subclasses): Use the subclass linked list, avoiding an
5613 O(N^2) when writing out all types.
5614 (walk_type): Issue an error if the base class is missing a tag,
5615 rather than generating bogus C code. Add a gcc_unreachable
5616 default case, in case people omit tags from concrete subclasses,
5617 or get the values wrong.
5618 (write_func_for_structure): Issue an error for subclasses for
5619 which the base doesn't have a "desc", since otherwise the
5620 autogenerated routines for the base would silently fail to visit
5621 any subclass fields.
5622 (write_root): Use get_ultimate_base_class, tweaking constness of
5623 tp to match that function's signature.
5625 2013-10-29 David Malcolm <dmalcolm@redhat.com>
5627 * doc/gty.texi (GTY Options): Add note about inheritance to
5628 description of desc and tag.
5629 (Inheritance and GTY): New.
5631 2013-10-29 David Malcolm <dmalcolm@redhat.com>
5633 * gengtype-parse.c (opts_have): Drop "static" so that
5634 we can use this from gengtype.c.
5635 * gengtype.c (set_gc_used_type): Mark any base class as used;
5636 update field traversal to visit inherited fields.
5637 (output_mangled_typename): Convert references to classes within
5638 an inheritance hierarchy to reference the ultimate base class,
5639 since only it will have gt_ functions.
5640 (get_string_option): New.
5641 (walk_subclasses): New.
5642 (walk_type): Treat GTY structs that have a "desc" as being the
5643 root of an inheritance hierarchy. Generate a switch on it
5644 within the marking function which walks all subclasses, adding
5645 cases for them via walk_subclasses. For subclasses, visit all
5646 fields of the type (including inherited ones).
5647 (write_func_for_structure): Don't write fns for subclasses, only
5648 for the ultimate base class within an inheritance hierarchy.
5649 Subclasses-marking will be handled by the base class marking functions.
5650 (write_types): Likewise.
5651 (write_local_func_for_structure): Likewise.
5652 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
5653 a "tag" option (and are thus concrete subclasses).
5654 (write_root): Use the marker function for the ultimate base class.
5655 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
5656 (opts_have): Add declaration.
5658 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
5660 * lra-spills.c (lra_final_code_change): Remove useless move insns
5661 originated from moves of pseudos.
5663 2013-10-28 Jeff Law <law@redhat.com>
5665 * lower-subreg.c (resolve_simple_move): Fix comment typo.
5667 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
5669 * df-scan.c (df_collection_rec): Adjust.
5670 (copy_defs): New constant.
5671 (copy_uses): Likewise.
5672 (copy_eq_uses): Likewise.
5673 (copy_mw): Likewise.
5674 (copy_all): Likewise.
5675 (df_insn_rescan): Adjust.
5676 (df_notes_rescan): Likewise.
5677 (df_swap_refs): Likewise.
5678 (df_sort_and_compress_refs): Likewise.
5679 (df_sort_and_compress_mws): Likewise.
5680 (df_install_refs): Likewise.
5681 (df_install_mws): Likewise.
5682 (df_refs_add_to_chains): Add flags parameter controlling which vectors
5684 (df_bb_refs_record): Adjust.
5685 (df_record_entry_block_defs): Likewise.
5686 (df_record_exit_block_defs): Likewise.
5687 (df_refs_verify): Likewise.
5688 (df_mws_verify): Likewise.
5689 (df_insn_refs_verify): Likewise.
5690 (df_bb_verify): Likewise.
5691 * ipa-pure-const.c (finish_state): Remove.
5692 (propagate): Adjust.
5693 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
5694 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
5695 var-tracking.c: Adjust.
5696 * vec.c (stack_vecs): Remove.
5697 (register_stack_vec): Likewise.
5698 (stack_vec_register_index): Likewise.
5699 (unregister_stack_vec): Likewise.
5700 * vec.h (struct va_stack): Remove.
5701 (struct vec<T, A, vl_ptr>): Specialize as
5702 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
5703 allocation strategy compatable with the vl_ptr layout.
5704 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
5705 specialization anyway.
5706 (class stack_vec): New class.
5707 (vec_stack_alloc): Remove.
5708 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
5710 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
5711 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5712 Sergey Lega <sergey.s.lega@intel.com>
5713 Anna Tikhonova <anna.tikhonova@intel.com>
5714 Ilya Tocar <ilya.tocar@intel.com>
5715 Andrey Turetskiy <andrey.turetskiy@intel.com>
5716 Ilya Verbin <ilya.verbin@intel.com>
5717 Kirill Yukhin <kirill.yukhin@intel.com>
5718 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5720 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
5721 (*prefetch_avx512pf_<mode>): New.
5722 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
5723 (avx512f_maskcmp<mode>3): Ditto.
5724 (vashrv16si3): Ditto.
5726 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
5727 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5728 Sergey Lega <sergey.s.lega@intel.com>
5729 Anna Tikhonova <anna.tikhonova@intel.com>
5730 Ilya Tocar <ilya.tocar@intel.com>
5731 Andrey Turetskiy <andrey.turetskiy@intel.com>
5732 Ilya Verbin <ilya.verbin@intel.com>
5733 Kirill Yukhin <kirill.yukhin@intel.com>
5734 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5736 * config/i386/i386.md (any_truncate): New.
5737 (trunsuffix): Ditto.
5738 * config/i386/predicates.md (const_8_to_9_operand): New.
5739 (const_10_to_11_operand): Ditto.
5740 (const_12_to_13_operand): Ditto.
5741 (const_14_to_15_operand): Ditto.
5742 (const_16_to_19_operand): Ditto.
5743 (const_20_to_23_operand): Ditto.
5744 (const_24_to_27_operand): Ditto.
5745 (const_28_to_31_operand): Ditto.
5746 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
5747 (cvtusi2<ssescalarmodesuffix>32): New.
5748 (cvtusi2<ssescalarmodesuffix>64): Ditto.
5749 (ufloatv16siv16sf2): Ditto.
5750 (avx512f_fix_notruncv16sfv16si): Ditto.
5751 (avx512f_ufix_notruncv16sfv16si): Ditto.
5752 (avx512f_vcvtss2usi): Ditto.
5753 (avx512f_vcvtss2usiq): Ditto.
5754 (avx512f_vcvttss2usi): Ditto.
5755 (avx512f_vcvttss2usiq): Ditto.
5756 (avx512f_vcvtsd2usi): Ditto.
5757 (avx512f_vcvtsd2usiq): Ditto.
5758 (avx512f_vcvttsd2usi): Ditto.
5759 (avx512f_vcvttsd2usiq): Ditto.
5760 (ufloatv8siv8df): Ditto.
5761 (avx512f_cvtdq2pd512_2): Ditto.
5762 (avx512f_cvtpd2dq512): Ditto.
5763 (avx512f_ufix_notruncv8dfv8si): Ditto.
5764 (avx512f_cvtpd2ps512): Ditto.
5765 (vec_unpacks_lo_v16sf): Ditto.
5766 (vec_unpacks_hi_v16sf): Ditto.
5767 (vec_unpacks_float_hi_v16si): Ditto.
5768 (vec_unpacks_float_lo_v16si): Ditto.
5769 (avx512f_unpckhps512): Ditto.
5770 (avx512f_unpcklps512): Ditto.
5771 (avx512f_movshdup512): Ditto.
5772 (avx512f_movsldup512): Ditto.
5773 (vec_extract_lo_v32hi): Ditto.
5774 (vec_extract_hi_v32hi): Ditto.
5775 (vec_extract_lo_v64qi): Ditto.
5776 (vec_extract_hi_v64qi): Ditto.
5777 (avx512f_unpckhpd512): Ditto.
5778 (avx512f_movddup512): Ditto.
5779 (avx512f_unpcklpd512): Ditto.
5780 (*avx512f_unpcklpd512): Ditto.
5781 (avx512f_shufps512_1): Ditto.
5782 (avx512f_shufpd512_1): Ditto.
5783 (avx512f_interleave_highv8di): Ditto.
5784 (avx512f_interleave_lowv8di): Ditto.
5785 (PMOV_DST_MODE): Ditto.
5786 (pmov_src_mode): Ditto.
5787 (pmov_src_lower): Ditto.
5789 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
5790 (*avx512f_<code>v8div16qi2): Ditto.
5791 (*avx512f_<code>v8div16qi2_store): Ditto.
5792 (vec_widen_umult_even_v16si): Ditto.
5793 (*vec_widen_umult_even_v16si): Ditto.
5794 (vec_widen_smult_even_v16si): Ditto.
5795 (*vec_widen_smult_even_v16si): Ditto.
5796 (avx512f_interleave_highv16si): Ditto.
5797 (avx512f_interleave_lowv16si): Ditto.
5798 (avx512f_<code>v16qiv16si2): Ditto.
5799 (avx512f_<code>v16hiv16si2): Ditto.
5800 (avx512f_<code>v8qiv8di2): Ditto.
5801 (avx512f_<code>v8hiv8di2): Ditto.
5802 (avx512f_<code>v8siv8di2): Ditto.
5803 (avx512cd_maskb_vec_dupv8di): Ditto.
5804 (avx512cd_maskw_vec_dupv16si): Ditto.
5805 (avx512f_vcvtph2ps512): Ditto.
5806 (avx512f_vcvtps2ph512): Ditto.
5807 (VEC_EXTRACT_MODE): Extened with wider modes.
5808 (VEC_PERM_AVX2): Ditto.
5809 (VEC_PERM_CONST): Ditto.
5811 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
5813 * config/arc/arc.c (arc_ccfsm_post_advance):
5814 Add comment about TYPE_RETURN.
5816 2013-10-28 Bin Cheng <bin.cheng@arm.com>
5818 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
5819 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
5820 (strip_offset): Convert offset to unsigned number.
5822 2013-10-27 Tom de Vries <tom@codesourcery.com>
5824 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
5825 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
5827 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
5829 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
5830 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
5831 (TARGET_BIG_ENDIAN): New macro.
5832 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
5833 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
5834 * config/sh/sh.c: Likewise.
5835 * config/sh/sh.h: Likewise.
5837 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
5839 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
5840 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
5842 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
5844 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
5846 (cmpstr_t, cmpstrsi): Fix formatting.
5848 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
5851 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
5852 addressing, do not use general_operand for memory operands.
5854 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
5857 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
5858 * lra-spills.c (lra_final_code_change): Remove useless move insns.
5860 2013-10-26 Jeff Law <law@redhat.com>
5862 * predict.c (PRED_MUDFLAP): Remove.
5863 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
5865 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
5866 (OBJS): Remove tree-nomudflap.o
5867 (GTFILES): Remove tree-mudflap.c
5868 * builtins.c (expand_builtin_alloc): Remove mudflap support.
5869 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
5870 (mfwrap_spec, mflib_spec): Likewise.
5871 (cpp_unique_options, cc1_options, static_specs): Likewise.
5872 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
5873 * passes.def: Likewise.
5874 * toplev.c (compile_file, process_options): Likewise.
5875 * tree-inline.c (copy_tree_r): Likewise.
5876 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
5877 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
5878 (build_constant_desc, output_constant_def_contents): Likewise.
5879 (categorize_decl_for_section): Likewise.
5880 * tree-mudflap.c: Removed.
5881 * tree-mudflap.h: Removed.
5882 * tree-nomudflap.c: Removed.
5883 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
5884 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
5885 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
5886 * config/sol2.h (MFLIB_SPEC): Likewise.
5887 * doc/install.texi: Remove mudflap references.
5888 * doc/passes.texi: Similarly.
5889 * doc/sourcebuild.texi: Similarly.
5890 * doc/invoke.texi: Remove mudlfap related options.
5892 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
5894 PR rtl-optimization/58759
5895 * lra-constraints.c (lra_constraints): Remove wrong condition to
5896 remove insn setting up an equivalent pseudo.
5898 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
5900 * config/rs6000/rs6000-protos.h
5901 (rs6000_secondary_memory_needed_mode): New prototype.
5902 * config/rs6000/rs6000.c: Include ira.h.
5903 (TARGET_LRA_P): Redefine.
5904 (rs6000_legitimate_offset_address_p): Call
5905 legitimate_constant_pool_address_p in strict mode for LRA.
5906 (rs6000_legitimate_address_p): Ditto.
5907 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
5908 (rs6000_emit_move): Add LRA version of code to generate load/store
5910 (rs6000_secondary_memory_needed_mode): New.
5911 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
5912 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
5913 constants, memory, and FP registers.
5914 (rs6000_lra_p): New.
5915 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
5916 * config/rs6000/rs6000.opt (mlra): New option.
5917 * lra-spills.c (lra_final_code_change): Remove useless move insns.
5919 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
5921 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
5922 has_single_use () and not do the conversion if has_single_use ()
5923 returns false for the multiplication result.
5925 2013-10-25 David Malcolm <dmalcolm@redhat.com>
5927 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
5928 to "NODE", since this works on a "tree", not an
5930 (CONSTANT_CLASS_P): Likewise.
5933 (INDIRECT_REF_P): Likewise.
5934 (REFERENCE_CLASS_P): Likewise.
5935 (COMPARISON_CLASS_P): Likewise.
5936 (UNARY_CLASS_P): Likewise.
5937 (BINARY_CLASS_P): Likewise.
5938 (STATEMENT_CLASS_P): Likewise.
5939 (VL_EXP_CLASS_P): Likewise.
5940 (EXPRESSION_CLASS_P): Likewise.
5941 (IS_TYPE_OR_DECL_P): Likewise.
5943 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
5945 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
5946 ADDR_EXPR in the defining statement.
5948 2013-10-25 Richard Biener <rguenther@suse.de>
5950 PR tree-optimization/58626
5951 * tree-loop-distribution.c (enum rdg_dep_type): Remove
5952 anti_dd, output_dd and input_dd.
5953 (struct rdg_edge): Remove level and relation members.
5954 (RDGE_LEVEL, RDGE_RELATION): Remove.
5955 (dot_rdg_1): Adjust.
5956 (create_rdg_edge_for_ddr): Remove.
5957 (create_rdg_edges_for_scalar): Adjust.
5958 (create_edge_for_control_dependence): Likewise.
5959 (create_rdg_edges): Split into ...
5960 (create_rdg_flow_edges): ... this
5961 (create_rdg_cd_edges): ... and this.
5963 (build_rdg): Likewise, do not compute data dependences or
5965 (pg_add_dependence_edges): New function.
5967 (distribute_loop): First apply all non-dependence based
5968 partition mergings. Then compute dependences between partitions
5969 and merge and order partitions according to them.
5971 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
5973 PR rtl-optimization/58831
5974 * alias.c (init_alias_analysis): At the beginning of each iteration,
5975 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
5977 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
5979 * recog.c (search_ofs): New static variable moved from...
5980 (peep2_find_free_register): ...here.
5981 (peephole2_optimize): Initialize it.
5983 2013-10-25 Tobias Burnus <burnus@net-b.de>
5985 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
5987 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
5989 * config/i386/i386.h (TARGET_MPX): New define.
5990 (TARGET_MPX_P): Ditto.
5992 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
5994 * config/i386/constraints.md (B): New.
5997 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
5998 * config/i386/i386-modes.def (BND32): New.
6000 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
6001 * config/i386/i386.c (isa_opts): Add mmpx.
6002 (regclass_map): Add bound registers.
6003 (dbx_register_map): Likewise.
6004 (dbx64_register_map): Likewise.
6005 (svr4_dbx_register_map): Likewise.
6007 (ix86_option_override_internal): Support MPX ISA.
6008 (ix86_conditional_register_usage): Support bound registers.
6009 (print_reg): Likewise.
6010 (ix86_code_end): Add MPX bnd prefix.
6011 (output_set_got): Likewise.
6012 (ix86_output_call_insn): Likewise.
6013 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
6014 (ix86_print_operand_punct_valid_p): Likewise.
6015 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
6017 (ix86_class_likely_spilled_p): Add bound regs support.
6018 (ix86_hard_regno_mode_ok): Likewise.
6019 (x86_order_regs_for_local_alloc): Likewise.
6020 (ix86_bnd_prefixed_insn_p): New.
6021 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
6022 (FIXED_REGISTERS): Add bound registers.
6023 (CALL_USED_REGISTERS): Likewise.
6024 (REG_ALLOC_ORDER): Likewise.
6025 (HARD_REGNO_NREGS): Likewise.
6027 (VALID_BND_REG_MODE): New.
6028 (FIRST_BND_REG): New.
6029 (LAST_BND_REG): New.
6030 (reg_class): Add BND_REGS.
6031 (REG_CLASS_NAMES): Likewise.
6032 (REG_CLASS_CONTENTS): Likewise.
6034 (ANY_BND_REG_P): New.
6036 (HI_REGISTER_NAMES): Add bound registers.
6037 * config/i386/i386.md (UNSPEC_BNDMK): New.
6038 (UNSPEC_BNDMK_ADDR): New.
6039 (UNSPEC_BNDSTX): New.
6040 (UNSPEC_BNDLDX): New.
6041 (UNSPEC_BNDLDX_ADDR): New.
6042 (UNSPEC_BNDCL): New.
6043 (UNSPEC_BNDCU): New.
6044 (UNSPEC_BNDCN): New.
6045 (UNSPEC_MPX_FENCE): New.
6048 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
6049 (length_immediate): Likewise.
6050 (prefix_0f): Likewise.
6052 (prefix_rep): Check for bnd prefix.
6053 (length_nobnd): New.
6054 (length): Use length_nobnd if specified.
6059 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
6062 (simple_return_internal): Likewise.
6063 (simple_return_pop_internal): Likewise.
6064 (*indirect_jump): Add MPX bnd prefix.
6065 (*tablejump_1): Likewise.
6066 (simple_return_internal_long): Likewise.
6067 (simple_return_indirect_internal): Likewise.
6071 (*mov<mode>_internal_mpx): New.
6072 (<mode>_<bndcheck>): New.
6073 (*<mode>_<bndcheck>): New.
6078 * config/i386/predicates.md (lea_address_operand): Rename to...
6079 (address_no_seg_operand): ... this.
6080 (address_mpx_no_base_operand): New.
6081 (address_mpx_no_index_operand): New.
6082 (bnd_mem_operator): New.
6083 * config/i386/i386.opt (mmpx): New.
6084 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
6085 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
6087 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
6089 * mode-classes.def (MODE_POINTER_BOUNDS): New.
6090 * tree.def (POINTER_BOUNDS_TYPE): New.
6091 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
6092 (POINTER_BOUNDS_MODE): New.
6093 (make_pointer_bounds_mode): New.
6094 * machmode.h (POINTER_BOUNDS_MODE_P): New.
6095 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
6096 (layout_type): Support POINTER_BOUNDS_TYPE.
6097 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
6098 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
6099 (type_contains_placeholder_1): Likewise.
6100 * tree.h (POINTER_BOUNDS_TYPE_P): New.
6101 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
6102 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
6104 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
6106 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
6107 mode of address computation when calling memory_address_addr_space.
6109 2013-08-24 Richard Henderson <rth@twiddle.net>
6112 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
6113 instead of create_convert_operand_to.
6114 (maybe_emit_sync_lock_test_and_set): Likewise.
6115 (expand_atomic_compare_and_swap): Likewise.
6116 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
6118 2013-08-24 Sriraman Tallam <tmsriram@google.com>
6120 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
6121 Change param_is to use the struct and not the pointer to the struct.
6123 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
6125 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
6126 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
6127 Move to targhooks.c.
6128 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
6129 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
6130 Relocate from gimplify.c.
6131 * targhooks.h: Add 2 prototypes.
6132 * tree.h: Delete 2 prototypes.
6134 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
6136 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
6137 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
6139 2013-10-24 Cong Hou <congh@google.com>
6141 * convert.c (convert_to_real): Guard those unsafe math function
6142 convertions with flag_unsafe_math_optimizations. Handle sqrt()
6145 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
6148 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
6150 (cgraph_create_edge): Use the new argument.
6151 (cgraph_create_indirect_edge): Likewise.
6153 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
6155 * config/arc/arc.c (arc_ccfsm_post_advance): Also handle
6157 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
6158 changes statep->state.
6160 2013-10-24 Tobias Burnus <burnus@net-b.de>
6163 * tree-cfg.c (replace_loop_annotate): New function.
6164 (execute_build_cfg): Call it.
6165 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
6166 * internal-fn.c (expand_ANNOTATE): New function.
6167 * internal-fn.def (ANNOTATE): Define as new internal function.
6168 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
6169 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
6170 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
6171 * doc/extend.texi (Pragmas): Document #pragma ivdep.
6172 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
6174 2013-10-17 Ian Bolton <ian.bolton@arm.com>
6175 Marcus Shawcroft <marcus.shawcroft@arm.com>
6177 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
6178 Special case reload SP+C into none GENERAL_REGS.
6180 2013-10-24 Michael Matz <matz@suse.de>
6182 * gengtype.c (is_file_equal): Check that files will be same length.
6184 2013-10-25 Christian Bruel <christian.bruel@st.com>
6186 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
6187 * config/sh/t-sh (sh-mem.o): New rule.
6188 * config/sh/sh-mem.cc (expand_block_move): Moved here.
6189 (sh_expand_cmpstr): New function.
6190 * config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
6191 * config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
6192 * config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
6193 (rotlhi3_8): Rename.
6195 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6197 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
6198 * Makefile.in (CXXDEPMODE): Assign and change users.
6199 (CCDEPMODE): Delete.
6200 * configure: Regenerate.
6202 2013-10-23 David Malcolm <dmalcolm@redhat.com>
6204 * gengtype-parse.c (require_without_advance): New.
6205 (type): For GTY-marked types that are not GTY((user)), parse any
6206 base classes, requiring them to be single-inheritance, and not
6207 be templates. For non-GTY-marked types and GTY((user)),
6208 continue to skip over any C++ inheritance specification.
6209 * gengtype-state.c (state_writer::write_state_struct_type):
6210 Write base class of type (if any).
6211 (read_state_struct_type): Read base class of type (if any).
6212 * gengtype.c (new_structure): Add a "base_class" parameter.
6213 (create_optional_field_): Update for new parameter to new_structure.
6214 (adjust_field_rtx_def): Likewise.
6215 (adjust_field_tree_exp): Likewise.
6216 * gengtype.h (struct type): Add "base_class" field to the s
6218 (new_structure): Add "base" parameter.
6220 2013-10-23 Sriraman Tallam <tmsriram@google.com>
6223 * config/i386/i386.c (ix86_option_override_internal):
6224 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
6225 (ix86_valid_target_attribute_tree):
6226 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
6227 Change TARGET_SSE to TARGET_SSE_P (opts->...)
6229 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
6231 * tree-ssa-loop.h: Remove include files.
6232 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
6233 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
6234 * graphite-clast-to-gimple.c: Likewise.
6235 * graphite-scop-detection.c: Likewise.
6236 * graphite-sese-to-poly.c: Likewise.
6237 * ipa-inline-analysis.c: Likewise.
6238 * ipa-pure-const.c: Likewise.
6239 * loop-init.c: Likewise.
6240 * passes.c: Likewise.
6241 * predict.c: Likewise.
6242 * tree-cfg.c: Likewise.
6243 * tree-cfgcleanup.c: Likewise.
6244 * tree-chrec.c: Likewise.
6245 * tree-data-ref.c: Likewise.
6246 * tree-loop-distribution.c: Likewise.
6247 * tree-parloops.c: Likewise.
6248 * tree-predcom.c: Likewise.
6249 * tree-scalar-evolution.c: Likewise.
6250 * tree-ssa-address.c: Likewise.
6251 * tree-ssa.c: Likewise.
6252 * tree-ssa-dce.c: Likewise.
6253 * tree-ssa-loop.c: Likewise.
6254 * tree-ssa-loop-im.c: Likewise.
6255 * tree-ssa-loop-ivcanon.c: Likewise.
6256 * tree-ssa-loop-ivopts.c: Likewise.
6257 * tree-ssa-loop-manip.c: Likewise.
6258 * tree-ssa-loop-niter.c: Likewise.
6259 * tree-ssa-loop-prefetch.c: Likewise.
6260 * tree-ssa-loop-unswitch.c: Likewise.
6261 * tree-ssa-reassoc.c: Likewise.
6262 * tree-vect-data-refs.c: Likewise.
6263 * tree-vect-loop.c: Likewise.
6264 * tree-vect-loop-manip.c: Likewise.
6265 * tree-vectorizer.c: Likewise.
6266 * tree-vect-stmts.c: Likewise.
6267 * tree-vrp.c: Likewise.
6269 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6271 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
6273 2013-10-23 Jakub Jelinek <jakub@redhat.com>
6275 PR tree-optimization/58775
6276 PR tree-optimization/58791
6277 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
6278 (insert_stmt_after): Rewritten, don't move the stmt, but really
6280 (get_stmt_uid_with_default): Remove.
6281 (build_and_add_sum): Use insert_stmt_after and
6282 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
6283 (update_range_test): Set uid on stmts added by
6284 force_gimple_operand_gsi. Don't immediately modify statements
6285 in inter-bb optimization, just update oe->op values.
6286 (optimize_range_tests): Return bool whether any changed have been made.
6287 (update_ops): New function.
6288 (struct inter_bb_range_test_entry): New type.
6289 (maybe_optimize_range_tests): Perform statement changes here.
6290 (not_dominated_by, appears_later_in_bb, get_def_stmt,
6291 ensure_ops_are_available): Remove.
6292 (find_insert_point): Rewritten.
6293 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
6294 return LHS of the (new resp. old) stmt. Don't call
6295 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
6296 instead of last, move new stmt at the right place.
6297 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
6298 (negate_value): Likewise. Set uids.
6299 (break_up_subtract_bb): Initialize uids.
6300 (reassociate_bb): Adjust rewrite_expr_tree caller.
6301 (do_reassoc): Don't call renumber_gimple_stmt_uids.
6303 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
6306 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
6307 TARGET_32BIT final condition.
6308 (mulsi3_internal2 and splitter): Same.
6310 2013-10-23 Jeff Law <law@redhat.com>
6312 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
6313 through joiner blocks with abnormal outgoing edges.
6315 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
6316 Add parameter JOINERS, to allow/disallow threading through joiner
6318 (thread_block): New. Call thread_block_1.
6319 (mark_threaded_blocks): Remove code to filter out certain cases
6320 of threading through joiner blocks.
6321 (thread_through_all_blocks): Document how we can have a dangling
6322 edge AUX field and clear it.
6324 2013-10-23 Ian Lance Taylor <iant@google.com>
6326 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
6327 (C++ Dialect Options): Likewise.
6328 (Optimize Options): Likewise.
6330 2013-10-23 Tom de Vries <tom@codesourcery.com>
6332 PR tree-optimization/58805
6333 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
6335 2013-10-23 Jakub Jelinek <jakub@redhat.com>
6337 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
6338 sequence based on get_range_info returned range.
6340 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
6342 * tree-ssa.h: Remove all #include's
6343 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
6344 * alias.c: Move required includes from tree-ssa.h.
6346 * builtins.c: Likewise.
6347 * calls.c: Likewise.
6348 * cfgexpand.c: Likewise.
6349 * cfghooks.c: Likewise.
6350 * cfgloop.c: Likewise.
6351 * cfgloopmanip.c: Likewise.
6352 * cgraph.c: Likewise.
6353 * cgraphbuild.c: Likewise.
6354 * cgraphclones.c: Likewise.
6355 * cgraphunit.c: Likewise.
6357 * except.c: Likewise.
6359 * final.c: Likewise.
6360 * fold-const.c: Likewise.
6361 * ggc-page.c: Likewise.
6362 * gimple-builder.c: Likewise.
6363 * gimple-fold.c: Likewise.
6364 * gimple-iterator.c: Likewise.
6365 * gimple-low.c: Likewise.
6366 * gimple-pretty-print.c: Likewise.
6367 * gimple-ssa-strength-reduction.c: Likewise.
6368 * gimple-streamer-in.c: Likewise.
6369 * gimple-streamer-out.c: Likewise.
6370 * gimplify.c: Likewise.
6371 * graphite-blocking.c: Likewise.
6372 * graphite-clast-to-gimple.c: Likewise.
6373 * graphite-dependences.c: Likewise.
6374 * graphite-interchange.c: Likewise.
6375 * graphite-optimize-isl.c: Likewise.
6376 * graphite-poly.c: Likewise.
6377 * graphite-scop-detection.c: Likewise.
6378 * graphite-sese-to-poly.c: Likewise.
6379 * graphite.c: Likewise.
6380 * ipa-cp.c: Likewise.
6381 * ipa-inline-analysis.c: Likewise.
6382 * ipa-inline-transform.c: Likewise.
6383 * ipa-inline.c: Likewise.
6384 * ipa-prop.c: Likewise.
6385 * ipa-pure-const.c: Likewise.
6386 * ipa-reference.c: Likewise.
6387 * ipa-split.c: Likewise.
6388 * ipa-utils.c: Likewise.
6389 * loop-init.c: Likewise.
6390 * lto-cgraph.c: Likewise.
6391 * lto-section-in.c: Likewise.
6392 * lto-section-out.c: Likewise.
6393 * lto-streamer-in.c: Likewise.
6394 * lto-streamer-out.c: Likewise.
6395 * lto-streamer.c: Likewise.
6396 * omp-low.c: Likewise.
6397 * passes.c: Likewise.
6398 * predict.c: Likewise.
6399 * print-tree.c: Likewise.
6400 * profile.c: Likewise.
6402 * targhooks.c: Likewise.
6403 * tracer.c: Likewise.
6404 * trans-mem.c: Likewise.
6405 * tree-call-cdce.c: Likewise.
6406 * tree-cfg.c: Likewise.
6407 * tree-cfgcleanup.c: Likewise.
6408 * tree-chrec.c: Likewise.
6409 * tree-complex.c: Likewise.
6410 * tree-data-ref.c: Likewise.
6411 * tree-dfa.c: Likewise.
6412 * tree-eh.c: Likewise.
6413 * tree-emutls.c: Likewise.
6414 * tree-if-conv.c: Likewise.
6415 * tree-inline.c: Likewise.
6416 * tree-into-ssa.c: Likewise.
6417 * tree-loop-distribution.c: Likewise.
6418 * tree-mudflap.c: Likewise.
6419 * tree-nested.c: Likewise.
6420 * tree-nrv.c: Likewise.
6421 * tree-object-size.c: Likewise.
6422 * tree-outof-ssa.c: Likewise.
6423 * tree-parloops.c: Likewise.
6424 * tree-phinodes.c: Likewise.
6425 * tree-predcom.c: Likewise.
6426 * tree-pretty-print.c: Likewise.
6427 * tree-profile.c: Likewise.
6428 * tree-scalar-evolution.c: Likewise.
6429 * tree-sra.c: Likewise.
6430 * tree-ssa-address.c: Likewise.
6431 * tree-ssa-alias.c: Likewise.
6432 * tree-ssa-ccp.c: Likewise.
6433 * tree-ssa-coalesce.c: Likewise.
6434 * tree-ssa-copy.c: Likewise.
6435 * tree-ssa-copyrename.c: Likewise.
6436 * tree-ssa-dce.c: Likewise.
6437 * tree-ssa-dom.c: Likewise.
6438 * tree-ssa-dse.c: Likewise.
6439 * tree-ssa-forwprop.c: Likewise.
6440 * tree-ssa-ifcombine.c: Likewise.
6441 * tree-ssa-live.c: Likewise.
6442 * tree-ssa-loop-ch.c: Likewise.
6443 * tree-ssa-loop-im.c: Likewise.
6444 * tree-ssa-loop-ivcanon.c: Likewise.
6445 * tree-ssa-loop-ivopts.c: Likewise.
6446 * tree-ssa-loop-manip.c: Likewise.
6447 * tree-ssa-loop-niter.c: Likewise.
6448 * tree-ssa-loop-prefetch.c: Likewise.
6449 * tree-ssa-loop-unswitch.c: Likewise.
6450 * tree-ssa-loop.c: Likewise.
6451 * tree-ssa-math-opts.c: Likewise.
6452 * tree-ssa-operands.c: Likewise.
6453 * tree-ssa-phiopt.c: Likewise.
6454 * tree-ssa-phiprop.c: Likewise.
6455 * tree-ssa-pre.c: Likewise.
6456 * tree-ssa-propagate.c: Likewise.
6457 * tree-ssa-reassoc.c: Likewise.
6458 * tree-ssa-sccvn.c: Likewise.
6459 * tree-ssa-sink.c: Likewise.
6460 * tree-ssa-strlen.c: Likewise.
6461 * tree-ssa-structalias.c: Likewise.
6462 * tree-ssa-tail-merge.c: Likewise.
6463 * tree-ssa-ter.c: Likewise.
6464 * tree-ssa-threadedge.c: Likewise.
6465 * tree-ssa-threadupdate.c: Likewise.
6466 * tree-ssa-uncprop.c: Likewise.
6467 * tree-ssa-uninit.c: Likewise.
6468 * tree-ssa.c: Likewise.
6469 * tree-ssanames.c: Likewise.
6470 * tree-stdarg.c: Likewise.
6471 * tree-streamer-in.c: Likewise.
6472 * tree-switch-conversion.c: Likewise.
6473 * tree-tailcall.c: Likewise.
6474 * tree-vect-data-refs.c: Likewise.
6475 * tree-vect-generic.c: Likewise.
6476 * tree-vect-loop-manip.c: Likewise.
6477 * tree-vect-loop.c: Likewise.
6478 * tree-vect-patterns.c: Likewise.
6479 * tree-vect-slp.c: Likewise.
6480 * tree-vect-stmts.c: Likewise.
6481 * tree-vectorizer.c: Likewise.
6482 * tree-vrp.c: Likewise.
6485 * value-prof.c: Likewise.
6486 * var-tracking.c: Likewise.
6487 * varpool.c: Likewise.
6488 * vtable-verify.c: Likewise.
6490 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6492 * config/tilegx/tilegx.c: Include "tree.h".
6494 2013-10-23 Jakub Jelinek <jakub@redhat.com>
6496 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
6497 the info, not after it.
6498 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
6499 dump_ssaname_info call.
6500 (pp_gimple_stmt_1): Adjust caller.
6501 (dump_phi_nodes): Likewise. Don't print "# " here.
6503 2013-10-22 Jan Hubicka <jh@suse.cz>
6505 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
6507 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
6508 * i386.c (expand_small_movmem_or_setmem): New function.
6509 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
6511 (alg_usable_p): Add support for value ranges; cleanup.
6512 (ix86_expand_set_or_movmem): Add support for misaligned moves.
6514 2013-10-22 Sterling Augustine <saugustine@google.com>
6516 * doc/invoke.texi: Document -ggnu-pubnames.
6517 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
6519 * dwarf2out.c: Include gdb/gdb-index.h.
6520 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
6521 debug_generate_pub_sections.
6522 (is_java, output_pubtables, output_pubname): New functions.
6523 (include_pubname_in_output): Handle debug_generate_pub_sections at
6525 (size_of_pubnames): Use new local space_for_flags based on
6526 debug_generate_pub_sections.
6527 (output_pubnames): Unify pubnames and pubtypes output logic.
6528 Genericize comments. Call output_pubname.
6529 (dwarf2out_finish): Move logic to output_pubnames and call it.
6531 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
6534 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
6535 Remove CCCmode handling.
6536 <case LTU>: Return 'c' suffix for CCCmode.
6537 <case GEU>: Return 'nc' suffix for CCCmode.
6538 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
6539 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
6540 (*sub<mode>3_cc_overflow): Ditto.
6541 (*subsi3_zext_cc_overflow): Ditto.
6543 2013-10-22 Steve Ellcey <sellcey@mips.com>
6545 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
6546 (AND (NOT OP1) (NOT OP2)).
6548 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
6550 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
6551 meaning of merge-high and merge-low masks for little endian; avoid
6552 use of vector-pack masks for little endian for mismatched modes.
6554 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6556 * config/tilepro/tilepro.c: Include "tree.h".
6558 2013-10-22 Andreas Schwab <schwab@suse.de>
6560 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
6563 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
6565 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
6566 [Ed Smith-Rowland]): New entries.
6567 ([Stephen M. Webb]): Update.
6569 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
6571 * tree-ssa-ter.h: Remove duplicate copy of file contents.
6573 2013-10-21 Marek Polacek <polacek@redhat.com>
6576 * fold-const.c (fold_range_test): Return 0 if the type is not
6579 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
6581 * system.h: Move hwint.h include further down.
6582 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
6583 gcc_checking_asserts.
6584 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
6586 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
6588 Fix volatile issues in optimize_bit_field_compare.
6589 * fold-const.c (optimize_bit_field_compare): Bail out if
6590 lvolatilep or rvolatilep.
6592 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
6594 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
6595 and get_inner_reference returning different pmode for non-volatile
6596 bit-field members dependent on flag_strict_volatile_bitfields.
6597 * stor-layout.c (layout_decl): Remove special handling of
6598 flag_strict_volatile_bitfields.
6599 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
6600 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
6602 2013-10-21 Paulo Matos <pmatos@broadcom.com>
6604 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
6607 2013-10-21 Jeff Law <law@redhat.com>
6609 * tree-ssa-threadedge.c (thread_through_normal_block): New
6610 argument VISITED. Remove VISISTED as a local variable. When we
6611 have a threadable jump, verify the destination of the jump has not
6613 (thread_across_edge): Allocate VISITED bitmap once at function
6614 scope and use it throughout. Make sure to set appropriate bits in
6615 VISITED for E (start of jump thread path).
6616 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
6617 through a joiner if any edge on the path has a recorded jump thread.
6619 2013-10-21 Ian Lance Taylor <iant@google.com>
6621 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
6622 don't imply that attributes can solve all problems.
6623 (Directory Options): Fix typo.
6625 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6627 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
6628 extend and extend_add.
6630 2013-10-21 Richard Biener <rguenther@suse.de>
6632 PR tree-optimization/58794
6633 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
6634 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
6636 2013-10-21 Richard Biener <rguenther@suse.de>
6639 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
6640 to (T) X for sign-changing conversions (or no conversion).
6642 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
6644 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
6646 2013-10-20 Jan Hubicka <jh@suse.cz>
6648 * config/i386/i386-tune.def: Add comment; organize into categories
6650 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
6652 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
6653 argument. Update function comment.
6654 (expand_set_or_movmem_via_rep): New function combining
6655 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
6656 (expand_movmem_via_rep_mov): Remove.
6657 expand_setmem_via_rep_stos): Remove.
6658 (expand_movmem_epilogue): Update calls correspondingly.
6659 (expand_setmem_epilogue_via_loop): Likewise.
6661 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
6662 (expand_set_or_movmem_prologue): New function combining
6663 expand_movmem_prologue and expand_setmem_prologue.
6664 (expand_movmem_prologue): Remove.
6665 (expand_setmem_prologue): Remove.
6666 (expand_set_or_movmem_constant_prologue): New function combining
6667 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
6668 (expand_constant_movmem_prologue): Remove.
6669 (expand_constant_setmem_prologue): Remove.
6670 (promote_duplicated_reg): Allow vector-const0 value.
6671 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
6672 and ix86_expand_setmem.
6673 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
6674 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
6676 2013-10-21 Diego Novillo <dnovillo@google.com>
6678 * asan.c: Include tree.h
6679 * bb-reorder.c: Likewise.
6680 * cfgcleanup.c: Likewise.
6681 * cfgloopmanip.c: Likewise.
6682 * data-streamer-in.c: Likewise.
6683 * data-streamer-out.c: Likewise.
6684 * data-streamer.c: Likewise.
6685 * dwarf2cfi.c: Likewise.
6686 * graphite-blocking.c: Likewise.
6687 * graphite-clast-to-gimple.c: Likewise.
6688 * graphite-dependences.c: Likewise.
6689 * graphite-interchange.c: Likewise.
6690 * graphite-optimize-isl.c: Likewise.
6691 * graphite-poly.c: Likewise.
6692 * graphite-scop-detection.c: Likewise.
6693 * graphite-sese-to-poly.c: Likewise.
6694 * graphite.c: Likewise.
6695 * ipa-devirt.c: Likewise.
6696 * ipa-profile.c: Likewise.
6699 * loop-init.c: Likewise.
6700 * loop-unroll.c: Likewise.
6701 * lower-subreg.c: Likewise.
6702 * lto/lto-object.c: Likewise.
6703 * recog.c: Likewise.
6704 * reginfo.c: Likewise.
6705 * tree-loop-distribution.c: Likewise.
6706 * tree-parloops.c: Likewise.
6707 * tree-ssa-strlen.c: Likewise.
6708 * tree-streamer.c: Likewise.
6709 * value-prof.c: Likewise.
6710 * target-globals.c: Likewise.
6711 * expr.h: Include tree-core.h instead of tree.h.
6712 * gimple.h: Likewise.
6713 * ipa-prop.h: Likewise.
6714 * ipa-utils.h: Likewise.
6715 * lto-streamer.h: Likewise.
6716 * streamer-hooks.h: Likewise.
6717 * ipa-reference.h: Include cgraph.h instead of tree.h.
6718 * cgraph.h: Include basic-block.h instead of tree.h.
6719 * tree-streamer.h: Do not include tree.h.
6720 * genattrtab.c (write_header): Generate inclusion of tree.h.
6721 * genautomata.c (main): Likewise.
6722 * genemit.c: Likewise.
6723 * genopinit.c: Likewise.
6724 * genoutput.c (output_prologue): Likewise.
6725 * genpeep.c: Likewise.
6727 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6729 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
6731 (vec_unpacku_hi_v8hi): Likewise.
6732 (vec_unpacku_lo_v16qi): Likewise.
6733 (vec_unpacku_lo_v8hi): Likewise.
6735 2013-10-20 Jan Hubicka <jh@suse.cz>
6737 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
6738 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
6739 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
6741 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
6743 * config/mips/mips.h (ISA_HAS_WSBH): Define.
6744 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
6746 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
6748 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
6751 * system.h: Undef m_slot.
6753 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6755 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
6756 all elements for both endian flavors.
6758 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
6761 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
6762 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
6763 and SI_REG for 64bit SYSV ABI targets.
6765 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
6767 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
6768 to multi_reg_return. Clarify that we are skipping USEs of multiple
6769 return registers. Use bool type where appropriate.
6771 2013-10-18 Jan Hubicka <jh@suse.cz>
6773 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
6775 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
6776 (X86_TUNE_PUSH_MEMORY): Likewise.
6777 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
6778 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
6779 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
6781 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
6782 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
6784 (ix86_option_override_internal): Update to use tune features instead
6787 2013-10-18 Cong Hou <congh@google.com>
6789 PR tree-optimization/58508
6790 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
6791 statement that contains data refs with zero-step.
6793 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
6795 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
6796 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
6797 * gimple-low.c (gimple_check_call_arg,
6798 gimple_check_call_matching_types): Move to cgraph.c.
6799 * gimple-low.h: Remove prototype.
6800 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
6801 Relocate from gimple-low.c.
6802 * cgraph.h: Add prototype. Don't include basic-block.h.
6803 * gimplify.c: Add gimple-low to include list.
6804 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
6805 * tree-eh.c: Add gimple-low to include list.
6806 * tree-nested.c: Likewise.
6807 * cfgexpand.c: Add tree-ssa-address.h to include list.
6809 * gimple-fold.c: Likewise.
6810 * gimple-ssa-strength-reduction.c: Likewise.
6811 * trans-mem.c: Likewise.
6812 * tree-mudflap.c: Likewise.
6813 * tree-ssa-loop-ivopts.c: Likewise.
6814 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
6815 (degenerate_phi_result): Move to tree-phinodes.c.
6816 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
6817 * tree-ssa-threadedge.c: Likewise.
6818 * tree-vrp.c: Likewise.
6819 * tree-phinodes.c (degenerate_phi_result): Relocate here.
6820 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
6821 * tree-phinodes.h (degenerate_phi_result): Add prototype.
6822 * tree-ssa-copy.c: Include tree-ssa-dom.h.
6823 * tree-ssa-forwprop.c: Likewise.
6824 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
6825 pass_data_cleanup_cfg_post_optimizing,
6826 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
6827 * tree-optimize.c: Delete File.
6828 * graphite.c: Include tree-cfgcleanup.h.
6829 * passes.c: Likewise.
6830 * tree-cfg.c: Likewise.
6831 * tree-profile.c: Likewise.
6832 * tree-ssa-dse.c: Likewise.
6833 * tree-ssa-loop-ivcanon.c: Likewise.
6834 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
6835 * tree-outof-ssa.c: Include sbitmap.h.
6836 * tree-ssa-live.c: Likewise.
6837 * tree-ssa-propagate.c: Likewise.
6838 * tree-ssa-structalias.c: Likewise.
6839 * tree-stdarg.c: Likewise.
6840 * Makefile.in (OBJS): Delete tree-optimize.o.
6841 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
6842 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
6843 * varasm.c: Include basic-block.h.
6844 * cfgloop.h: Include function.h instead of basic-block.h
6845 (bb_loop_depth): Move to cfgloop.c.
6846 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
6848 2013-10-18 Teresa Johnson <tejohnson@google.com>
6850 * predict.c (probably_never_executed): Compare frequency-based
6851 count to number of training runs.
6852 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
6854 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6856 * config/arm/arm.c (cortexa9_extra_costs): New table.
6857 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
6859 2013-10-18 Jeff Law <law@redhat.com>
6861 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
6863 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
6864 (register_jump_thread): Add "registered_jump_thread" debug
6866 * dbgcnt.def (registered_jump_thread): New debug counter.
6868 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
6870 * config/rs6000/rs6000.c: Include cgraph.h.
6872 2013-10-18 Teresa Johnson <tejohnson@google.com>
6874 * tree-ssa-tail-merge.c (replace_block_by): Update edge
6875 weights during merging.
6877 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
6879 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
6880 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
6881 * tree-ssa-operands.h: Remove prototype.
6882 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
6883 * gimple.c (virtual_operand_p): Relocate from gimple.c.
6884 * gimple.h: Add prototype.
6885 * gimple-ssa.h: Include tree-ssa-operands.h.
6886 * tree-dump.c: Add tree-cfg.h to include list.
6887 * tree-ssa-alias.c: Add ipa-reference.h to include list.
6888 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
6889 * config/i386/i386.c: Don't include tree-flow.h.
6890 * config/rs6000/rs6000.c: Likewise.
6892 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6894 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
6896 2013-10-18 Richard Biener <rguenther@suse.de>
6898 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
6899 or TYPE_UNSIGNED of integral types.
6900 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
6901 NULL_TREE for zero-precision integral types.
6903 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
6905 * config/aarch64/arm_neon.h
6906 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
6908 2013-10-17 Sriraman Tallam <tmsriram@google.com>
6911 * opth-gen.awk: Define target_flags_explicit.
6913 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
6915 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
6916 fields to the reg_addr array that describes the valid addressing
6917 mode for any register, general purpose registers, floating point
6918 registers, and Altivec registers.
6919 (FIRST_RELOAD_REG_CLASS): Likewise.
6920 (LAST_RELOAD_REG_CLASS): Likewise.
6921 (struct reload_reg_map_type): Likewise.
6922 (reload_reg_map_type): Likewise.
6923 (RELOAD_REG_VALID): Likewise.
6924 (RELOAD_REG_MULTIPLE): Likewise.
6925 (RELOAD_REG_INDEXED): Likewise.
6926 (RELOAD_REG_OFFSET): Likewise.
6927 (RELOAD_REG_PRE_INCDEC): Likewise.
6928 (RELOAD_REG_PRE_MODIFY): Likewise.
6929 (reg_addr): Likewise.
6930 (mode_supports_pre_incdec_p): New helper functions to say whether
6931 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
6932 (mode_supports_pre_modify_p): Likewise.
6933 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
6934 print the valid address mode bits for each mode.
6935 (rs6000_debug_print_mode): Likewise.
6936 (rs6000_debug_reg_global): Likewise.
6937 (rs6000_setup_reg_addr_masks): New function to set up the address
6938 mask bits for each type.
6939 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
6940 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
6941 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
6942 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
6943 PRE_MODIFY are supported.
6944 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
6945 registers, instead of {src,dest}_av_p.
6946 (rs6000_print_options_internal): Tweak the debug output slightly.
6948 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
6950 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
6953 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
6955 * tree-flow.h (struct omp_region): Move to omp-low.c.
6956 Remove omp_ prototypes and variables.
6957 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
6958 (copy_var_decl): Relocate prototype from tree-flow.h.
6959 * gimple.c (copy_var_decl): Relocate from omp-low.c.
6960 * tree.h: Move prototype to omp-low.h.
6961 * omp-low.h: New File. Relocate prototypes here.
6962 * omp-low.c (struct omp_region): Make local here.
6963 (root_omp_region): Make static.
6964 (copy_var_decl) Move to gimple.c.
6965 (new_omp_region): Make static.
6966 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
6967 * tree-cfg.c: Include omp-low.h.
6968 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
6969 * gimplify.c: Include omp-low.h.
6970 * tree-parloops.c: Likewise.
6972 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
6974 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
6975 operands are in memory, prefer to force non-matched operand 1 to
6978 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
6981 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
6982 restrict TImode addresses to single indirect registers if both
6983 -mquad-memory and -mvsx-timode are used.
6984 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
6985 we should emit load/store quad. Remove using %y for quad memory
6988 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
6989 constraints to allow load/store quad on machines where TImode is
6990 not allowed in VSX registers. Use 'n' instead of 'F' constraint
6991 for TImode to load integer constants.
6993 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6995 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
6997 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
6999 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
7000 handling of STACK_REG.
7002 2013-10-17 Richard Biener <rguenther@suse.de>
7004 PR tree-optimization/58143
7005 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
7007 (rewrite_to_defined_overflow): Likewise.
7008 (move_computations_dom_walker::before_dom): Rewrite stmts
7009 with undefined signed overflow that are not always executed
7010 into unsigned arithmetic.
7012 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7015 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
7016 explicit isa flag to be an options variable, instead of using
7017 global_options_set. Remove define from rs6000.h.
7018 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
7020 * config/rs6000/rs6000.c (rs6000_option_override_internal):
7021 Initialize rs6000_isa_flags_explicit.
7022 (rs6000_function_specific_save): Add gcc_options* parameter, so
7023 that the powerpc builds after the 2013-10-15 changes.
7024 (rs6000_function_specific_restore): Likewise.
7026 2013-10-16 DJ Delorie <dj@redhat.com>
7028 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
7029 op is a REG before checking REGNO.
7030 (rl78_alloc_physical_registers): Verify pattern is a SET before
7033 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7035 * config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
7037 (vec_unpacks_lo_v4sf): Likewise.
7038 (vec_unpacks_float_hi_v4si): Likewise.
7039 (vec_unpacks_float_lo_v4si): Likewise.
7040 (vec_unpacku_float_hi_v4si): Likewise.
7041 (vec_unpacku_float_lo_v4si): Likewise.
7043 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7045 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
7046 (vsx_concat_v2sf): Likewise.
7048 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
7050 * config/aarch64/aarch64.md
7051 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
7053 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
7055 PR tree-optimization/58697
7056 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
7057 estimated_loop_iterations_int.
7058 (max_stmt_executions_int): Call get_max_loop_iterations_int.
7059 (get_max_loop_iterations_int): New. HWINT version of
7060 get_max_loop_iterations.
7061 * cfgloop.h: Add prototypes.
7062 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
7063 * loop-unroll.c (decide_peel_once_rolling): Call
7064 get_estimated_loop_iterations_int.
7065 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
7066 * tree-ssa-loop-niter.h: Tweak prototypes.
7068 2013-10-16 David Malcolm <dmalcolm@redhat.com>
7070 * gengtype-parse.c (struct_field_seq): Ignore access-control
7071 keywords ("public:" etc).
7073 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
7075 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
7076 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
7078 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
7080 * config/arm/arm.opt (mlra): New option.
7081 * config/arm/arm.c (arm_lra_p): New function.
7082 (TARGET_LRA_P): Define.
7084 2013-10-16 Paulo Matos <pmatos@broadcom.com>
7086 * tree-core.h (tree_code_name): Remove.
7087 * tree.h (get_tree_code_name): New prototype.
7088 * tree.c (tree_code_name): Make static.
7089 (get_tree_code_name): New function.
7090 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
7091 tree_class_check_failed, tree_range_check_failed,
7092 tree_not_class_check_failed, omp_clause_check_failed,
7093 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
7094 wrapper get_tree_code_name instead of calling tree_code_name directly.
7095 * tree-vrp.c (dump_asserts_for): Likewise.
7096 * tree-dump.c (dequeue_and_dump): Likewise.
7097 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
7098 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
7099 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
7100 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
7101 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
7102 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
7103 dump_gimple_omp_for): Likewise.
7104 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
7105 * tree-ssa-pre.c (print_pre_expr): Likewise.
7106 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
7107 * print-tree.c (print_node_brief, print_node): Likewise.
7108 * gimple.c (gimple_check_failed): Likewise.
7109 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
7110 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
7111 * config/mep/mep.c (mep_validate_vliw): Likewise.
7112 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
7113 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
7115 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7117 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
7120 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
7122 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
7123 (MULTILIB_MATCHES): Add multilib for -march=v8.
7125 2013-10-15 Sriraman Tallam <tmsriram@google.com>
7128 * optc-save-gen.awk: Add extra parameter to the save and restore
7130 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
7131 * tree.c (build_optimization_node): New parameter. Add extra parameter
7132 to call to cl_optimization_save.
7133 (build_target_option_node): New parameter. Add extra parameter
7134 to call to cl_target_option_save.
7135 * tree.h (build_optimization_node): New parameter.
7136 (build_target_option_node): New parameter.
7137 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
7138 build_optimization_node and build_target_option_node.
7139 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
7140 (handle_pragma_push_options): Ditto.
7141 * toplev.c (process_options): Ditto.
7142 * opts.c (init_options_struct): Check for opts_set non-null.
7143 * target.def (target_option.save): New parameter.
7144 (target_option.restore): New parameter.
7145 * tm.texi: Generate.
7146 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
7147 (ix86_pragma_target_parse): Ditto.
7148 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
7150 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
7151 to build_optimization_node and build_target_option_node.
7152 (rs6000_valid_attribute_p): Ditto.
7153 (rs6000_pragma_target_parse): Ditto.
7154 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
7156 * config/i386/i386.h:
7157 TARGET_64BIT_P: New Macro
7158 TARGET_MMX_P: New Macro.
7159 TARGET_3DNOW_P: New Macro.
7160 TARGET_3DNOW_A_P: New Macro.
7161 TARGET_SSE_P: New Macro.
7162 TARGET_SSE2_P: New Macro.
7163 TARGET_SSE3_P: New Macro.
7164 TARGET_SSSE3_P: New Macro.
7165 TARGET_SSE4_1_P: New Macro.
7166 TARGET_SSE4_2_P: New Macro.
7167 TARGET_AVX_P: New Macro.
7168 TARGET_AVX2_P: New Macro.
7169 TARGET_AVX512F_P: New Macro.
7170 TARGET_AVX512PF_P: New Macro.
7171 TARGET_AVX512ER_P: New Macro.
7172 TARGET_AVX512CD_P: New Macro.
7173 TARGET_FMA_P: New Macro.
7174 TARGET_SSE4A_P: New Macro.
7175 TARGET_FMA4_P: New Macro.
7176 TARGET_XOP_P: New Macro.
7177 TARGET_LWP_P: New Macro.
7178 TARGET_ABM_P: New Macro.
7179 TARGET_BMI_P: New Macro.
7180 TARGET_BMI2_P: New Macro.
7181 TARGET_LZCNT_P: New Macro.
7182 TARGET_TBM_P: New Macro.
7183 TARGET_POPCNT_P: New Macro.
7184 TARGET_SAHF_P: New Macro.
7185 TARGET_MOVBE_P: New Macro.
7186 TARGET_CRC32_P: New Macro.
7187 TARGET_AES_P: New Macro.
7188 TARGET_PCLMUL_P: New Macro.
7189 TARGET_CMPXCHG16B_P: New Macro.
7190 TARGET_FSGSBASE_P: New Macro.
7191 TARGET_RDRND_P: New Macro.
7192 TARGET_F16C_P: New Macro.
7193 TARGET_RTM_P: New Macro.
7194 TARGET_HLE_P: New Macro.
7195 TARGET_RDSEED_P: New Macro.
7196 TARGET_PRFCHW_P: New Macro.
7197 TARGET_ADX_P: New Macro.
7198 TARGET_FXSR_P: New Macro.
7199 TARGET_XSAVE_P: New Macro.
7200 TARGET_XSAVEOPT_P: New Macro.
7201 TARGET_LP64_P: New Macro.
7202 TARGET_X32_P: New Macro.
7203 TARGET_FPMATH_DEFAULT_P: New Macro.
7204 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
7205 * config/i386/i386.c (ix86_option_override_internal): New parameters.
7207 Change ix86_tune_string to access opts->x_ix86_tune_string.
7208 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
7209 Change ix86_arch_string to access opts->x_ix86_arch_string.
7210 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
7211 Change ix86_pmode to access opts->x_ix86_pmode.
7212 Change ix86_abi to access opts->x_ix86_abi.
7213 Change ix86_cmodel to access opts->x_ix86_cmodel.
7214 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
7215 Change ix86_isa_flags_explicit to access
7216 opts->x_ix86_isa_flags_explicit.
7217 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
7218 Change ix86_regparm to access opts->x_ix86_regparm.
7219 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
7220 Change ix86_preferred_stack_boundary_arg to access
7221 opts->x_ix86_preferred_stack_boundary_arg.
7222 Change ix86_force_align_arg_pointer to access
7223 opts->x_ix86_force_align_arg_pointer.
7224 Change ix86_incoming_stack_boundar_arg to access
7225 opts->x_ix86_incoming_stack_boundar_arg.
7226 Change ix86_fpmath to access opts->x_ix86_fpmath.
7227 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
7228 Change ix86_recip_name to access opts->x_ix86_recip_name.
7229 Change ix86_stack_protector_guard to access
7230 opts->x_ix86_stack_protector_guard.
7231 Change ix86_tune_memcpy_strategy to access
7232 opts->x_ix86_tune_memcpy_strategy.
7233 Change ix86_tune_memset_strategy to access
7234 opts->x_ix86_tune_memset_strategy.
7235 Change global_options to access opts.
7236 Change global_options_set to access opts_set.
7237 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
7238 Change TARGET_MMX to TARGET_MMX_P (opts->...).
7239 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
7240 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
7241 Change TARGET_SSE to TARGET_SSE_P (opts->...).
7242 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
7243 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
7244 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
7245 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
7246 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
7247 Change TARGET_AVX to TARGET_AVX_P (opts->...).
7248 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
7249 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
7250 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
7251 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
7252 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
7253 Change TARGET_FMA to TARGET_FMA_P (opts->...).
7254 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
7255 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
7256 Change TARGET_XOP to TARGET_XOP_P (opts->...).
7257 Change TARGET_LWP to TARGET_LWP_P (opts->...).
7258 Change TARGET_ABM to TARGET_ABM_P (opts->...).
7259 Change TARGET_BMI to TARGET_BMI_P (opts->...).
7260 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
7261 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
7262 Change TARGET_TBM to TARGET_TBM_P (opts->...).
7263 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
7264 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
7265 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
7266 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
7267 Change TARGET_AES to TARGET_AES_P (opts->...).
7268 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
7269 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
7270 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
7271 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
7272 Change TARGET_F16C to TARGET_F16C_P (opts->...).
7273 Change TARGET_RTM to TARGET_RTM_P (opts->...).
7274 Change TARGET_HLE to TARGET_HLE_P (opts->...).
7275 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
7276 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
7277 Change TARGET_ADX to TARGET_ADX_P (opts->...).
7278 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
7279 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
7280 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
7281 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
7282 Change TARGET_X32 to TARGET_X32_P (opts->...).
7283 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
7284 Change TARGET_FLOAT_RETURNS_IN_80387 to
7285 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
7286 (ix86_function_specific_save): New parameter. Use opts-> fields
7287 to replace global fields.
7288 (ix86_function_specific_restore): Ditto.
7289 (ix86_valid_target_attribute_inner_p): New parameters.
7291 Fix call to ix86_handle_option and set_option.
7292 (ix86_valid_target_attribute_tree): New parameters.
7293 Change global_options to access opts.
7294 Change global_options_set to access opts_set.
7295 Fix call to ix86_valid_target_attribute_inner_p.
7296 Change ix86_tune_string to access opts->x_ix86_tune_string.
7297 Change ix86_arch_string to access opts->x_ix86_arch_string.
7298 Change ix86_fpmath to access opts->x_ix86_fpmath
7299 Fix call to ix86_option_override_internal.
7300 Fix call to ix86_add_new_builtins.
7301 Fix calls to build_optimization_node and build_target_option_node.
7302 (ix86_valid_target_attribute_p): Remove access to global_options.
7303 Use new gcc_options structure func_options.
7304 Fix call to ix86_valid_target_attribute_tree.
7305 Fix call to build_optimization_node.
7306 (get_builtin_code_for_version): Fix call to
7307 ix86_valid_target_attribute_tree.
7309 2013-10-15 David Malcolm <dmalcolm@redhat.com>
7311 * Makefile.in (PICFLAG): New.
7312 (enable_host_shared): New.
7313 (INTERNAL_CFLAGS): Use PICFLAG.
7314 (LIBIBERTY): Use pic build of libiberty.a if configured with
7315 --enable-host-shared.
7316 * configure.ac: Add --enable-host-shared, setting up new
7318 * configure: Regenerate.
7319 * doc/install.texi (--enable-shared): Add note contrasting it with ...
7320 (--enable-host-shared): New option.
7322 2013-10-15 Richard Biener <rguenther@suse.de>
7324 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
7325 for built-in functions.
7327 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
7329 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
7330 (optimize_range_tests_1): New function,
7331 extracted from optimize_range_tests.
7332 (optimize_range_tests_xor): Similarly.
7333 (optimize_range_tests_diff): New function.
7334 (optimize_range_tests): Use optimize_range_tests_1.
7336 2013-10-15 Cong Hou <congh@google.com>
7338 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
7339 requirement of the reduction pattern so that one operand of the
7340 reduction operation can come from outside of the loop.
7342 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
7344 * config/arm/neon-schedgen.ml: Remove.
7345 * config/arm/cortex-a9-neon.md: Remove comment regarding
7348 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
7350 * config/arm/types: Remove old neon types.
7352 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
7354 * config/arm/cortex-a7.md
7355 (cortex_a7_neon_type): New.
7356 (cortex_a7_neon_mul): Update for new types.
7357 (cortex_a7_neon_mla): Likewise.
7358 (cortex_a7_neon): Likewise.
7360 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
7362 * config/arm/cortex-a15-neon.md
7363 (cortex_a15_neon_type): New,
7365 (cortex_a15_neon_int_1): Remove.
7366 (cortex_a15_neon_int_2): Likewise.
7367 (cortex_a15_neon_int_3): Likewise.
7368 (cortex_a15_neon_int_4): Likewise.
7369 (cortex_a15_neon_int_5): Likewise.
7370 (cortex_a15_neon_vqneg_vqabs): Likewise.
7371 (cortex_a15_neon_vmov): Likewise.
7372 (cortex_a15_neon_vaba): Likewise.
7373 (cortex_a15_neon_vaba_qqq): Likewise.
7374 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7375 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7376 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
7378 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7379 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7380 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
7381 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7382 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7383 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7384 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7385 (cortex_a15_neon_shift_1): Likewise.
7386 (cortex_a15_neon_shift_2): Likewise.
7387 (cortex_a15_neon_shift_3): Likewise.
7388 (cortex_a15_neon_vshl_ddd): Likewise.
7389 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7390 (cortex_a15_neon_vsra_vrsra): Likewise.
7391 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7392 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7393 (cortex_a15_neon_bp_3cycle): Likewise.
7394 (cortex_a15_neon_ldm_2): Likewise.
7395 (cortex_a15_neon_stm_2): Likewise.
7396 (cortex_a15_neon_mcr): Likewise.
7397 (cortex_a15_neon_mrc): Likewise.
7398 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7399 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7400 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7401 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7402 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7403 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7404 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7405 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7406 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7407 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7408 (cortex_a15_neon_bp_simple): Likewise.
7409 (cortex_a15_neon_bp_2cycle): Likewise.
7410 (cortex_a15_neon_bp_3cycle): Likewise.
7411 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7412 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7413 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7414 (cortex_a15_neon_vld2_4_regs): Likewise.
7415 (cortex_a15_neon_vld3_vld4): Likewise.
7416 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7417 (cortex_a15_neon_vst1_3_4_regs): Likewise.
7418 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
7419 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
7420 (cortex_a15_neon_vst3_vst4): Rename to...
7421 (cortex_a15_neon_vst4): This, update for new attributes.
7422 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
7423 (cortex_a15_neon_vld3_vld4_lane): Likewise.
7424 (cortex_a15_neon_vst1_vst2_lane): Likewise.
7425 (cortex_a15_neon_vst3_vst4_lane): Likewise.
7426 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
7427 (cortex_a15_neon_ldm_2): Likewise.
7428 (cortex_a15_neon_stm_2): Likewise.
7429 (cortex_a15_neon_mcr): Likewise.
7430 (cortex_a15_neon_mcr_2_mcrr): Likewise.
7431 (cortex_a15_neon_mrc): Likewise.
7432 (cortex_a15_neon_mrrc): Likewise.
7434 (cortex_a15_neon_abd): New.
7435 (cortex_a15_neon_abd_q): Likewise.
7436 (cortex_a15_neon_aba): Likewise.
7437 (cortex_a15_neon_aba_q): Likewise.
7438 (cortex_a15_neon_acc): Likewise.
7439 (cortex_a15_neon_acc_q): Likewise.
7440 (cortex_a15_neon_arith_basic): Likewise.
7441 (cortex_a15_neon_arith_complex): Likewise.
7442 (cortex_a15_neon_multiply): Likewise.
7443 (cortex_a15_neon_multiply_q): Likewise.
7444 (cortex_a15_neon_mla): Likewise.
7445 (cortex_a15_neon_mla_q): Likewise.
7446 (cortex_a15_neon_sat_mla_long): Likewise.
7447 (cortex_a15_neon_shift_acc): Likewise.
7448 (cortex_a15_neon_shift_imm_basic): Likewise.
7449 (cortex_a15_neon_shift_imm_complex): Likewise.
7450 (cortex_a15_neon_shift_reg_basic): Likewise.
7451 (cortex_a15_neon_shift_reg_basic_q): Likewise.
7452 (cortex_a15_neon_shift_reg_complex): Likewise.
7453 (cortex_a15_neon_shift_reg_complex_q): Likewise.
7454 (cortex_a15_neon_fp_negabs): Likewise
7455 (cortex_a15_neon_fp_arith): Likewise
7456 (cortex_a15_neon_fp_arith_q): Likewise
7457 (cortex_a15_neon_fp_cvt_int): Likewise
7458 (cortex_a15_neon_fp_cvt_int_q): Likewise
7459 (cortex_a15_neon_fp_cvt_16): Likewise
7460 (cortex_a15_neon_fp_mul): Likewise
7461 (cortex_a15_neon_fp_mul_q): Likewise
7462 (cortex_a15_neon_fp_mla): Likewise
7463 (cortex_a15_neon_fp_mla_q): Likewise
7464 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
7465 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
7466 (cortex_a15_neon_bitops): Likewise.
7467 (cortex_a15_neon_bitops_q): Likewise.
7468 (cortex_a15_neon_from_gp): Likewise.
7469 (cortex_a15_neon_from_gp_q): Likewise.
7470 (cortex_a15_neon_tbl3_tbl4): Likewise.
7471 (cortex_a15_neon_zip_q): Likewise.
7472 (cortex_a15_neon_to_gp): Likewise.
7473 (cortex_a15_neon_load_a): Likewise.
7474 (cortex_a15_neon_load_b): Likewise.
7475 (cortex_a15_neon_load_c): Likewise.
7476 (cortex_a15_neon_load_d): Likewise.
7477 (cortex_a15_neon_load_e): Likewise.
7478 (cortex_a15_neon_load_f): Likewise.
7479 (cortex_a15_neon_store_a): Likewise.
7480 (cortex_a15_neon_store_b): Likewise.
7481 (cortex_a15_neon_store_c): Likewise.
7482 (cortex_a15_neon_store_d): Likewise.
7483 (cortex_a15_neon_store_e): Likewise.
7484 (cortex_a15_neon_store_f): Likewise.
7485 (cortex_a15_neon_store_g): Likewise.
7486 (cortex_a15_neon_store_h): Likewise.
7487 (cortex_a15_vfp_to_from_gp): Likewise.
7489 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
7491 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
7493 (cortex_a9_neon_vshl_ddd): Remove.
7494 (cortex_a9_neon_vst3_vst4): Likewise.
7495 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7497 (cortex_a9_neon_bit_ops_q): New.
7499 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
7500 (cortex_a9_neon_int_2): Likewise.
7501 (cortex_a9_neon_int_3): Likewise.
7502 (cortex_a9_neon_int_4): Likewise.
7503 (cortex_a9_neon_int_5): Likewise.
7504 (cortex_a9_neon_vqneg_vqabs): Likewise.
7505 (cortex_a9_neon_vmov): Likewise.
7506 (cortex_a9_neon_vaba): Likewise.
7507 (cortex_a9_neon_vaba_qqq): Likewise.
7508 (cortex_a9_neon_shift_1): Likewise.
7509 (cortex_a9_neon_shift_2): Likewise.
7510 (cortex_a9_neon_shift_3): Likewise.
7511 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7512 (cortex_a9_neon_vsra_vrsra): Likewise.
7513 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7514 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7515 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
7517 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7518 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7519 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
7521 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7522 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7523 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7524 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7525 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7526 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7527 (cortex_a9_neon_fp_vsum): Likewise.
7528 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7529 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7530 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7531 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7532 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7533 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7534 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7535 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7536 (cortex_a9_neon_bp_simple): Likewise.
7537 (cortex_a9_neon_bp_2cycle): Likewise.
7538 (cortex_a9_neon_bp_3cycle): Likewise.
7539 (cortex_a9_neon_ldr): Likewise.
7540 (cortex_a9_neon_str): Likewise.
7541 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7542 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7543 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7544 (cortex_a9_neon_vld2_4_regs): Likewise.
7545 (cortex_a9_neon_vld3_vld4): Likewise.
7546 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7547 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7548 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7549 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7550 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7551 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7552 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7553 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7554 (cortex_a9_neon_mcr): Likewise.
7555 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7556 (cortex_a9_neon_mrc): Likewise.
7557 (cortex_a9_neon_mrrc): Likewise.
7559 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
7561 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
7563 (cortex_a8_neon_vshl_ddd): Remove.
7564 (cortex_a8_neon_vst3_vst4): Likewise.
7565 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7567 (cortex_a8_neon_bit_ops_q): New.
7569 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
7570 (cortex_a8_neon_int_2): Likewise..
7571 (cortex_a8_neon_int_3): Likewise.
7572 (cortex_a8_neon_int_5): Likewise.
7573 (cortex_a8_neon_vqneg_vqabs): Likewise.
7574 (cortex_a8_neon_int_4): Likewise.
7575 (cortex_a8_neon_vaba): Likewise.
7576 (cortex_a8_neon_vaba_qqq): Likewise.
7577 (cortex_a8_neon_shift_1): Likewise.
7578 (cortex_a8_neon_shift_2): Likewise.
7579 (cortex_a8_neon_shift_3): Likewise.
7580 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7581 (cortex_a8_neon_vsra_vrsra): Likewise.
7582 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7583 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7584 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
7586 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7587 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7588 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
7590 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7591 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7592 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7593 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7594 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7595 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7596 (cortex_a8_neon_fp_vsum): Likewise.
7597 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7598 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7599 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7600 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7601 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7602 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7603 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7604 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7605 (cortex_a8_neon_bp_simple): Likewise.
7606 (cortex_a8_neon_bp_2cycle): Likewise.
7607 (cortex_a8_neon_bp_3cycle): Likewise.
7608 (cortex_a8_neon_ldr): Likewise.
7609 (cortex_a8_neon_str): Likewise.
7610 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7611 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7612 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7613 (cortex_a8_neon_vld2_4_regs): Likewise.
7614 (cortex_a8_neon_vld3_vld4): Likewise.
7615 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7616 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7617 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7618 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7619 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7620 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7621 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7622 (cortex_a8_neon_mcr): Likewise.
7623 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7624 (cortex_a8_neon_mrc): Likewise.
7625 (cortex_a8_neon_mrrc): Likewise.
7627 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
7629 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
7631 * config/aarch64/aarch64-simd.md (neon_type): Remove.
7632 (aarch64_simd_dup<mode>): Add "type" attribute.
7633 (aarch64_dup_lane<mode>): Likewise.
7634 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
7635 (*aarch64_simd_mov<mode>): Likewise.
7636 (aarch64_simd_mov_from_<mode>low): Likewise.
7637 (aarch64_simd_mov_from_<mode>high): Likewise.
7638 (orn<mode>3): Likewise.
7639 (bic<mode>3): Likewise.
7640 (add<mode>3): Likewise.
7641 (sub<mode>3): Likewise.
7642 (mul<mode>3): Likewise.
7643 (*aarch64_mul3_elt<mode>): Likewise.
7644 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
7645 (*aarch64_mul3_elt_to_128df): Likewise.
7646 (*aarch64_mul3_elt_to_64v2df): Likewise.
7647 (neg<mode>2): Likewise.
7648 (abs<mode>2): Likewise.
7649 (abd<mode>_3): Likewise.
7650 (aba<mode>_3): Likewise.
7651 (fabd<mode>_3): Likewise.
7652 (*fabd_scalar<mode>3): Likewise.
7653 (and<mode>3): Likewise.
7654 (ior<mode>3): Likewise.
7655 (xor<mode>3): Likewise.
7656 (one_cmpl<mode>2): Likewise.
7657 (aarch64_simd_vec_set<mode>): Likewise.
7658 (aarch64_simd_lshr<mode>): Likewise.
7659 (aarch64_simd_ashr<mode>): Likewise.
7660 (aarch64_simd_imm_shl<mode>): Likewise.
7661 (aarch64_simd_reg_sshl<mode): Likewise.
7662 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
7663 (aarch64_simd_reg_shl<mode>_signed): Likewise.
7664 (aarch64_simd_vec_setv2di): Likewise.
7665 (aarch64_simd_vec_set<mode>): Likewise.
7666 (aarch64_mla<mode>): Likewise.
7667 (*aarch64_mla_elt<mode>): Likewise.
7668 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7669 (aarch64_mls<mode>): Likewise.
7670 (*aarch64_mls_elt<mode>): Likewise.
7671 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7672 (<su><maxmin><mode>3): Likewise.
7673 (move_lo_quad_<mode>): Likewise.
7674 (aarch64_simd_move_hi_quad_<mode>): Likewise.
7675 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
7676 (vec_pack_trunc_<mode>): Likewise.
7677 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
7678 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
7679 (*aarch64_<su>mlal_lo<mode>): Likewise.
7680 (*aarch64_<su>mlal_hi<mode>): Likewise.
7681 (*aarch64_<su>mlsl_lo<mode>): Likewise.
7682 (*aarch64_<su>mlsl_hi<mode>): Likewise.
7683 (*aarch64_<su>mlal<mode>): Likewise.
7684 (*aarch64_<su>mlsl<mode>): Likewise.
7685 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
7686 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
7687 (add<mode>3): Likewise.
7688 (sub<mode>3): Likewise.
7689 (mul<mode>3): Likewise.
7690 (div<mode>3): Likewise.
7691 (neg<mode>2): Likewise.
7692 (abs<mode>2): Likewise.
7693 (fma<mode>4): Likewise.
7694 (*aarch64_fma4_elt<mode>): Likewise.
7695 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7696 (*aarch64_fma4_elt_to_128df): Likewise.
7697 (*aarch64_fma4_elt_to_64v2df): Likewise.
7698 (fnma<mode>4): Likewise.
7699 (*aarch64_fnma4_elt<mode>): Likewise.
7700 (*aarch64_fnma4_elt_<vswap_width_name><mode>
7701 (*aarch64_fnma4_elt_to_128df): Likewise.
7702 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7703 (<frint_pattern><mode>2): Likewise.
7704 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
7705 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
7706 (vec_unpacks_lo_v4sf): Likewise.
7707 (aarch64_float_extend_lo_v2df): Likewise.
7708 (vec_unpacks_hi_v4sf): Likewise.
7709 (aarch64_float_truncate_lo_v2sf): Likewise.
7710 (aarch64_float_truncate_hi_v4sf): Likewise.
7711 (aarch64_vmls<mode>): Likewise.
7712 (<su><maxmin><mode>3): Likewise.
7713 (<maxmin_uns><mode>3): Likewise.
7714 (reduc_<sur>plus_<mode>): Likewise.
7715 (reduc_<sur>plus_v2di): Likewise.
7716 (reduc_<sur>plus_v2si): Likewise.
7717 (reduc_<sur>plus_<mode>): Likewise.
7718 (aarch64_addpv4sf): Likewise.
7719 (clz<mode>2): Likewise.
7720 (reduc_<maxmin_uns>_<mode>): Likewise.
7721 (reduc_<maxmin_uns>_v2di): Likewise.
7722 (reduc_<maxmin_uns>_v2si): Likewise.
7723 (reduc_<maxmin_uns>_<mode>): Likewise.
7724 (reduc_<maxmin_uns>_v4sf): Likewise.
7725 (aarch64_simd_bsl<mode>_internal): Likewise.
7726 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
7727 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
7728 (aarch64_get_lane<mode>): Likewise.
7729 (*aarch64_combinez<mode>): Likewise.
7730 (aarch64_combine<mode>): Likewise.
7731 (aarch64_simd_combine<mode>): Likewise.
7732 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
7733 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
7734 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
7735 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
7736 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
7737 (aarch64_<sur>h<addsub><mode>): Likewise.
7738 (aarch64_<sur><addsub>hn<mode>): Likewise.
7739 (aarch64_<sur><addsub>hn2<mode>): Likewise.
7740 (aarch64_pmul<mode>): Likewise.
7741 (aarch64_<su_optab><optab><mode>): Likewise.
7742 (aarch64_<sur>qadd<mode>): Likewise.
7743 (aarch64_sqmovun<mode>): Likewise.
7744 (aarch64_<sur>qmovn<mode>): Likewise.
7745 (aarch64_s<optab><mode>): Likewise.
7746 (aarch64_sq<r>dmulh<mode>): Likewise.
7747 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7748 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7749 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7750 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
7751 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7752 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7753 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
7754 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
7755 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7756 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
7757 (aarch64_sqdmull<mode>): Likewise.
7758 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7759 (aarch64_sqdmull_n<mode>): Likewise.
7760 (aarch64_sqdmull2<mode>_internal): Likewise.
7761 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7762 (aarch64_sqdmull2_n<mode>_internal): Likewise.
7763 (aarch64_<sur>shl<mode>): Likewise.
7764 (aarch64_<sur>q<r>shl<mode>
7765 (aarch64_<sur>shll_n<mode>): Likewise.
7766 (aarch64_<sur>shll2_n<mode>): Likewise.
7767 (aarch64_<sur>shr_n<mode>): Likewise.
7768 (aarch64_<sur>sra_n<mode>): Likewise.
7769 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
7770 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
7771 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
7772 (aarch64_cm<optab><mode>): Likewise.
7773 (aarch64_cm<optab>di): Likewise.
7774 (aarch64_cm<optab><mode>): Likewise.
7775 (aarch64_cm<optab>di): Likewise.
7776 (aarch64_cmtst<mode>): Likewise.
7777 (aarch64_cmtstdi): Likewise.
7778 (aarch64_cm<optab><mode>): Likewise.
7779 (*aarch64_fac<optab><mode>): Likewise.
7780 (aarch64_addp<mode>): Likewise.
7781 (aarch64_addpdi): Likewise.
7782 (sqrt<mode>2): Likewise.
7783 (vec_load_lanesoi<mode>): Likewise.
7784 (vec_store_lanesoi<mode>): Likewise.
7785 (vec_load_lanesci<mode>): Likewise.
7786 (vec_store_lanesci<mode>): Likewise.
7787 (vec_load_lanesxi<mode>): Likewise.
7788 (vec_store_lanesxi<mode>): Likewise.
7789 (*aarch64_mov<mode>): Likewise.
7790 (aarch64_ld2<mode>_dreg): Likewise.
7791 (aarch64_ld2<mode>_dreg): Likewise.
7792 (aarch64_ld3<mode>_dreg): Likewise.
7793 (aarch64_ld3<mode>_dreg): Likewise.
7794 (aarch64_ld4<mode>_dreg): Likewise.
7795 (aarch64_ld4<mode>_dreg): Likewise.
7796 (aarch64_tbl1<mode>): Likewise.
7797 (aarch64_tbl2v16qi): Likewise.
7798 (aarch64_combinev16qi): Likewise.
7799 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
7800 (aarch64_st2<mode>_dreg): Likewise.
7801 (aarch64_st2<mode>_dreg): Likewise.
7802 (aarch64_st3<mode>_dreg): Likewise.
7803 (aarch64_st3<mode>_dreg): Likewise.
7804 (aarch64_st4<mode>_dreg): Likewise.
7805 (aarch64_st4<mode>_dreg): Likewise.
7806 (*aarch64_simd_ld1r<mode>): Likewise.
7807 (aarch64_frecpe<mode>): Likewise.
7808 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
7809 (aarch64_frecps<mode>): Likewise.
7811 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
7813 * config/arm/iterators.md (V_elem_ch): New.
7815 (VQH_type): Likewise.
7816 * config/arm/arm.md (is_neon_type): New.
7817 (conds): Use is_neon_type.
7818 (anddi3_insn): Update type attribute.
7819 (xordi3_insn): Likewise.
7820 (one_cmpldi2): Likewise.
7821 * config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
7822 * config/arm/neon.md (neon_mov): Update type attribute.
7823 (*movmisalign<mode>_neon_store): Likewise.
7824 (*movmisalign<mode>_neon_load): Likewise.
7825 (vec_set<mode>_internal): Likewise.
7826 (vec_set<mode>_internal): Likewise.
7827 (vec_setv2di_internal): Likewise.
7828 (vec_extract<mode>): Likewise.
7829 (vec_extract<mode>): Likewise.
7830 (vec_extractv2di): Likewise.
7831 (*add<mode>3_neon): Likewise.
7832 (adddi3_neon): Likewise.
7833 (*sub<mode>3_neon): Likewise.
7834 (subdi3_neon): Likewise.
7835 (fma<VCVTF:mode>4): Likewise.
7836 (fma<VCVTF:mode>4_intrinsic): Likewise.
7837 (*fmsub<VCVTF:mode>4): Likewise.
7838 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
7839 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
7840 (ior<mode>3): Likewise.
7841 (and<mode>3): Likewise.
7842 (orn<mode>3_neon): Likewise.
7843 (orndi3_neon): Likewise.
7844 (bic<mode>3_neon): Likewise.
7845 (bicdi3_neon): Likewise.
7846 (xor<mode>3): Likewise.
7847 (one_cmpl<mode>2): Likewise.
7848 (abs<mode>2): Likewise.
7849 (neg<mode>2): Likewise.
7850 (negdi2_neon): Likewise.
7851 (*umin<mode>3_neon): Likewise.
7852 (*umax<mode>3_neon): Likewise.
7853 (*smin<mode>3_neon): Likewise.
7854 (*smax<mode>3_neon): Likewise.
7855 (vashl<mode>3): Likewise.
7856 (vashr<mode>3_imm): Likewise.
7857 (vlshr<mode>3_imm): Likewise.
7858 (ashl<mode>3_signed): Likewise.
7859 (ashl<mode>3_unsigned): Likewise.
7860 (neon_load_count): Likewise.
7861 (ashldi3_neon_noclobber): Likewise.
7862 (ashldi3_neon): Likewise.
7863 (signed_shift_di3_neon): Likewise.
7864 (unsigned_shift_di3_neon): Likewise.
7865 (ashrdi3_neon_imm_noclobber): Likewise.
7866 (lshrdi3_neon_imm_noclobber): Likewise.
7867 (<shift>di3_neon): Likewise.
7868 (widen_ssum<mode>3): Likewise.
7869 (widen_usum<mode>3): Likewise.
7870 (quad_halves_<code>v4si): Likewise.
7871 (quad_halves_<code>v4sf): Likewise.
7872 (quad_halves_<code>v8hi): Likewise.
7873 (quad_halves_<code>v16qi): Likewise.
7874 (reduc_splus_v2di): Likewise.
7875 (neon_vpadd_internal<mode>): Likewise.
7876 (neon_vpsmin<mode>): Likewise.
7877 (neon_vpsmax<mode>): Likewise.
7878 (neon_vpumin<mode>): Likewise.
7879 (neon_vpumax<mode>): Likewise.
7880 (*ss_add<mode>_neon): Likewise.
7881 (*us_add<mode>_neon): Likewise.
7882 (*ss_sub<mode>_neon): Likewise.
7883 (*us_sub<mode>_neon): Likewise.
7884 (neon_vadd<mode>_unspec): Likewise.
7885 (neon_vaddl<mode>): Likewise.
7886 (neon_vaddw<mode>): Likewise.
7887 (neon_vhadd<mode>): Likewise.
7888 (neon_vqadd<mode>): Likewise.
7889 (neon_vaddhn<mode>): Likewise.
7890 (neon_vmul<mode>): Likewise.
7891 (neon_vfms<VCVTF:mode>): Likewise.
7892 (neon_vmlal<mode>): Likewise.
7893 (neon_vmls<mode>): Likewise.
7894 (neon_vmlsl<mode>): Likewise.
7895 (neon_vqdmulh<mode>): Likewise.
7896 (neon_vqdmlal<mode>): Likewise.
7897 (neon_vqdmlsl<mode>): Likewise.
7898 (neon_vmull<mode>): Likewise.
7899 (neon_vqdmull<mode>): Likewise.
7900 (neon_vsub<mode>_unspec): Likewise.
7901 (neon_vsubl<mode>): Likewise.
7902 (neon_vsubw<mode>): Likewise.
7903 (neon_vqsub<mode>): Likewise.
7904 (neon_vhsub<mode>): Likewise.
7905 (neon_vsubhn<mode>): Likewise.
7906 (neon_vceq<mode>): Likewise.
7907 (neon_vcge<mode>): Likewise.
7908 (neon_vcgeu<mode>): Likewise.
7909 (neon_vcgt<mode>): Likewise.
7910 (neon_vcgtu<mode>): Likewise.
7911 (neon_vcle<mode>): Likewise.
7912 (neon_vclt<mode>): Likewise.
7913 (neon_vcage<mode>): Likewise.
7914 (neon_vcagt<mode>): Likewise.
7915 (neon_vtst<mode>): Likewise.
7916 (neon_vabd<mode>): Likewise.
7917 (neon_vabdl<mode>): Likewise.
7918 (neon_vaba<mode>): Likewise.
7919 (neon_vabal<mode>): Likewise.
7920 (neon_vmax<mode>): Likewise.
7921 (neon_vmin<mode>): Likewise.
7922 (neon_vpaddl<mode>): Likewise.
7923 (neon_vpadal<mode>): Likewise.
7924 (neon_vpmax<mode>): Likewise.
7925 (neon_vpmin<mode>): Likewise.
7926 (neon_vrecps<mode>): Likewise.
7927 (neon_vrsqrts<mode>): Likewise.
7928 (neon_vqabs<mode>): Likewise.
7929 (neon_vqneg<mode>): Likewise.
7930 (neon_vcls<mode>): Likewise.
7931 (clz<mode>2): Likewise.
7932 (popcount<mode>2): Likewise.
7933 (neon_vrecpe<mode>): Likewise.
7934 (neon_vrsqrte<mode>): Likewise.
7935 (neon_vget_lane<mode>_sext_internal): Likewise.
7936 (neon_vget_lane<mode>_zext_internal): Likewise.
7937 (neon_vdup_n<mode>): Likewise.
7938 (neon_vdup_n<mode>): Likewise.
7939 (neon_vdup_nv2di): Likewise.
7940 (neon_vdup_lane<mode>_interal): Likewise.
7941 (*neon_vswp<mode>): Likewise.
7942 (neon_vcombine<mode>): Likewise.
7943 (float<mode><V_cvtto>2): Likewise.
7944 (floatuns<mode><V_cvtto>2): Likewise.
7945 (fix_trunc<mode><V_cvtto>2): Likewise.
7946 (fixuns_trunc<mode><V_cvtto>2
7947 (neon_vcvt<mode>): Likewise.
7948 (neon_vcvt<mode>): Likewise.
7949 (neon_vcvtv4sfv4hf): Likewise.
7950 (neon_vcvtv4hfv4sf): Likewise.
7951 (neon_vcvt_n<mode>): Likewise.
7952 (neon_vcvt_n<mode>): Likewise.
7953 (neon_vmovn<mode>): Likewise.
7954 (neon_vqmovn<mode>): Likewise.
7955 (neon_vqmovun<mode>): Likewise.
7956 (neon_vmovl<mode>): Likewise.
7957 (neon_vmul_lane<mode>): Likewise.
7958 (neon_vmul_lane<mode>): Likewise.
7959 (neon_vmull_lane<mode>): Likewise.
7960 (neon_vqdmull_lane<mode>): Likewise.
7961 (neon_vqdmulh_lane<mode>): Likewise.
7962 (neon_vqdmulh_lane<mode>): Likewise.
7963 (neon_vmla_lane<mode>): Likewise.
7964 (neon_vmla_lane<mode>): Likewise.
7965 (neon_vmlal_lane<mode>): Likewise.
7966 (neon_vqdmlal_lane<mode>): Likewise.
7967 (neon_vmls_lane<mode>): Likewise.
7968 (neon_vmls_lane<mode>): Likewise.
7969 (neon_vmlsl_lane<mode>): Likewise.
7970 (neon_vqdmlsl_lane<mode>): Likewise.
7971 (neon_vext<mode>): Likewise.
7972 (neon_vrev64<mode>): Likewise.
7973 (neon_vrev32<mode>): Likewise.
7974 (neon_vrev16<mode>): Likewise.
7975 (neon_vbsl<mode>_internal): Likewise.
7976 (neon_vshl<mode>): Likewise.
7977 (neon_vqshl<mode>): Likewise.
7978 (neon_vshr_n<mode>): Likewise.
7979 (neon_vshrn_n<mode>): Likewise.
7980 (neon_vqshrn_n<mode>): Likewise.
7981 (neon_vqshrun_n<mode>): Likewise.
7982 (neon_vshl_n<mode>): Likewise.
7983 (neon_vqshl_n<mode>): Likewise.
7984 (neon_vqshlu_n<mode>): Likewise.
7985 (neon_vshll_n<mode>): Likewise.
7986 (neon_vsra_n<mode>): Likewise.
7987 (neon_vsri_n<mode>): Likewise.
7988 (neon_vsli_n<mode>): Likewise.
7989 (neon_vtbl1v8qi): Likewise.
7990 (neon_vtbl2v8qi): Likewise.
7991 (neon_vtbl3v8qi): Likewise.
7992 (neon_vtbl4v8qi): Likewise.
7993 (neon_vtbl1v16qi): Likewise.
7994 (neon_vtbl2v16qi): Likewise.
7995 (neon_vcombinev16qi): Likewise.
7996 (neon_vtbx1v8qi): Likewise.
7997 (neon_vtbx2v8qi): Likewise.
7998 (neon_vtbx3v8qi): Likewise.
7999 (neon_vtbx4v8qi): Likewise.
8000 (*neon_vtrn<mode>_insn): Likewise.
8001 (*neon_vzip<mode>_insn): Likewise.
8002 (*neon_vuzp<mode>_insn): Likewise.
8003 (neon_vld1<mode>): Likewise.
8004 (neon_vld1_lane<mode>): Likewise.
8005 (neon_vld1_lane<mode>): Likewise.
8006 (neon_vld1_dup<mode>): Likewise.
8007 (neon_vld1_dup<mode>): Likewise.
8008 (neon_vld1_dupv2di): Likewise.
8009 (neon_vst1<mode>): Likewise.
8010 (neon_vst1_lane<mode>): Likewise.
8011 (neon_vst1_lane<mode>): Likewise.
8012 (neon_vld2<mode>): Likewise.
8013 (neon_vld2<mode>): Likewise.
8014 (neon_vld2_lane<mode>): Likewise.
8015 (neon_vld2_lane<mode>): Likewise.
8016 (neon_vld2_dup<mode>): Likewise.
8017 (neon_vst2<mode>): Likewise.
8018 (neon_vst2<mode>): Likewise.
8019 (neon_vst2_lane<mode>): Likewise.
8020 (neon_vst2_lane<mode>): Likewise.
8021 (neon_vld3<mode>): Likewise.
8022 (neon_vld3qa<mode>): Likewise.
8023 (neon_vld3qb<mode>): Likewise.
8024 (neon_vld3_lane<mode>): Likewise.
8025 (neon_vld3_lane<mode>): Likewise.
8026 (neon_vld3_dup<mode>): Likewise.
8027 (neon_vst3<mode>): Likewise.
8028 (neon_vst3qa<mode>): Likewise.
8029 (neon_vst3qb<mode>): Likewise.
8030 (neon_vst3_lane<mode>): Likewise.
8031 (neon_vst3_lane<mode>): Likewise.
8032 (neon_vld4<mode>): Likewise.
8033 (neon_vld4qa<mode>): Likewise.
8034 (neon_vld4qb<mode>): Likewise.
8035 (neon_vld4_lane<mode>): Likewise.
8036 (neon_vld4_lane<mode>): Likewise.
8037 (neon_vld4_dup<mode>): Likewise.
8038 (neon_vst4<mode>): Likewise.
8039 (neon_vst4qa<mode>): Likewise.
8040 (neon_vst4qb<mode>): Likewise.
8041 (neon_vst4_lane<mode>): Likewise.
8042 (neon_vst4_lane<mode>): Likewise.
8043 (neon_vec_unpack<US>_lo_<mode>): Likewise.
8044 (neon_vec_unpack<US>_hi_<mode>): Likewise.
8045 (neon_vec_<US>mult_lo_<mode>): Likewise.
8046 (neon_vec_<US>mult_hi_<mode>): Likewise.
8047 (neon_vec_<US>shiftl_<mode>): Likewise.
8048 (neon_unpack<US>_<mode>): Likewise.
8049 (neon_vec_<US>mult_<mode>): Likewise.
8050 (vec_pack_trunc_<mode>): Likewise.
8051 (neon_vec_pack_trunc_<mode>): Likewise.
8052 (neon_vabd<mode>_2): Likewise.
8053 (neon_vabd<mode>_3): Likewise.
8055 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
8057 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
8058 (load_pair): Update type attribute.
8059 (store_pair): Update type attribute.
8060 * config/aarch64/iterators.md (q): New.
8062 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
8064 * config/arm/types.md: Add new types for Neon insns.
8066 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
8067 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8068 Sergey Lega <sergey.s.lega@intel.com>
8069 Anna Tikhonova <anna.tikhonova@intel.com>
8070 Ilya Tocar <ilya.tocar@intel.com>
8071 Andrey Turetskiy <andrey.turetskiy@intel.com>
8072 Ilya Verbin <ilya.verbin@intel.com>
8073 Kirill Yukhin <kirill.yukhin@intel.com>
8074 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8076 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
8077 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
8078 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
8080 (srcp14<mode>): Ditto.
8081 (rsqrt14<mode>): Ditto.
8082 (rsqrt14<mode>): Ditto.
8083 (avx512f_vmscalef<mode>): Ditto.
8084 (avx512f_scalef<mode>): Ditto.
8085 (avx512f_getexp<mode>): Ditto.
8086 (avx512f_sgetexp<mode>): Ditto.
8087 (avx512f_fixupimm<mode>): Ditto.
8088 (avx512f_sfixupimm<mode>): Ditto.
8089 (avx512f_rndscale<mode>): Ditto.
8090 (*avx512er_exp2<mode>): Ditto.
8091 (*avx512er_rcp28<mode>): Ditto.
8092 (avx512er_rsqrt28<mode>): Ditto.
8093 (avx512f_getmant<mode>): Ditto.
8094 (avx512f_getmant<mode>): Ditto.
8095 (avx512f_rndscale<mode>): Fix formatting.
8097 2013-10-15 Martin Jambor <mjambor@suse.cz>
8099 * ipa-utils.h (ipa_edge_within_scc): Declare.
8100 * ipa-cp.c (edge_within_scc): Moved...
8101 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
8103 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
8104 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8105 Sergey Lega <sergey.s.lega@intel.com>
8106 Anna Tikhonova <anna.tikhonova@intel.com>
8107 Ilya Tocar <ilya.tocar@intel.com>
8108 Andrey Turetskiy <andrey.turetskiy@intel.com>
8109 Ilya Verbin <ilya.verbin@intel.com>
8110 Kirill Yukhin <kirill.yukhin@intel.com>
8111 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8113 * config/i386/predicates.md (const_8_to_15_operand): New.
8114 (const_16_to_31_operand): Ditto.
8115 * config/i386/sse.md (V8FI): New.
8117 (reduc_splus_v8df): Ditto.
8118 (reduc_splus_v16sf): Ditto.
8119 (avx512f_vextract<shuffletype>32x4_1): Ditto.
8120 (vec_extract_hi_<mode>): Ditto.
8121 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
8122 (vec_set_lo_<mode>): Ditto.
8123 (vec_set_hi_<mode>): Ditto.
8124 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
8125 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
8126 (avx512f_pshufd_1): Ditto.
8127 (avx512f_broadcast<mode>): Ditto.
8128 (avx512f_broadcast<mode>): Ditto.
8129 (define_split): Split vec_extract_lo into move.
8130 (ssequartermode): Ditto.
8131 (ssedoublemode): Extened with wider modes.
8132 (vec_extract_lo_<mode>): Ditto.
8134 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
8135 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8136 Sergey Lega <sergey.s.lega@intel.com>
8137 Anna Tikhonova <anna.tikhonova@intel.com>
8138 Ilya Tocar <ilya.tocar@intel.com>
8139 Andrey Turetskiy <andrey.turetskiy@intel.com>
8140 Ilya Verbin <ilya.verbin@intel.com>
8141 Kirill Yukhin <kirill.yukhin@intel.com>
8142 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8144 * config/i386/predicates.md (register_or_constm1_operand): New.
8145 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
8146 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
8147 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
8148 UNSPEC_SCATTER_PREFETCH
8150 (avx512f_ucmp<mode>3): Ditto.
8151 (avx512f_vternlog<mode>): Ditto.
8152 (avx512f_align<mode>): Ditto.
8153 (<shift_insn><mode>3): Ditto.
8154 (avx512f_<rotate>v<mode>): Ditto.
8155 (avx512f_<rotate><mode>): Ditto.
8156 (avx512f_eq<mode>3): Ditto.
8157 (avx512f_eq<mode>3_1): Ditto.
8158 (avx512f_gt<mode>3): Ditto.
8159 (avx512f_testm<mode>3): Ditto.
8160 (avx512f_testnm<mode>3): Ditto.
8161 (avx512pf_gatherpf<mode>): Ditto.
8162 (*avx512pf_gatherpf<mode>_mask): Ditto.
8163 (*avx512pf_gatherpf<mode>): Ditto.
8164 (avx512pf_scatterpf<mode>): Ditto.
8165 (*avx512pf_scatterpf<mode>_mask): Ditto.
8166 (*avx512pf_scatterpf<mode>): Ditto.
8167 (avx512f_vec_dup_gpr<mode>): Ditto.
8168 (clz<mode>2): Ditto.
8169 (conflict<mode>): Ditto.
8170 (REDUC_SMINMAX_MODE): Extened with wider modes.
8171 (reduc_<code>_<mode>): Ditto.
8172 (vlshr<mode>3): Ditto.
8173 (vashl<mode>3): Ditto.
8175 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
8176 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8177 Sergey Lega <sergey.s.lega@intel.com>
8178 Anna Tikhonova <anna.tikhonova@intel.com>
8179 Ilya Tocar <ilya.tocar@intel.com>
8180 Andrey Turetskiy <andrey.turetskiy@intel.com>
8181 Ilya Verbin <ilya.verbin@intel.com>
8182 Kirill Yukhin <kirill.yukhin@intel.com>
8183 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8185 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
8188 (avx512fmaskmode): Ditto.
8189 (bcstscalarsuff): Ditto.
8190 (avx512f_blendm<mode>): Ditto.
8191 (cmp_imm_predicate): Ditto.
8192 (avx512f_cmp<mode>3): Ditto.
8193 (avx512f_vec_dup<mode>): Ditto.
8194 (avx512f_vec_dup_mem<mode>): Ditto.
8195 (avx512f_vpermi2var<mode>3): Ditto.
8196 (avx512f_vpermt2var<mode>3): Ditto.
8197 (vec_init<mode>): Ditto.
8198 (avx512f_gathersi<mode>): Ditto.
8199 (*avx512f_gathersi<mode>): Ditto.
8200 (*avx512f_gathersi<mode>_2): Ditto.
8201 (avx512f_gatherdi<mode>): Ditto.
8202 (*avx512f_gatherdi<mode>): Ditto.
8203 (*avx512f_gatherdi<mode>_2): Ditto.
8204 (avx512f_scattersi<mode>): Ditto.
8205 (*avx512f_scattersi<mode>): Ditto.
8206 (avx512f_scatterdi<mode>): Ditto.
8207 (*avx512f_scatterdi<mode>): Ditto.
8208 (sseintprefix): Extened with wider modes.
8209 (VEC_GATHER_IDXSI): Ditto.
8210 (VEC_GATHER_IDXDI): Ditto.
8211 (VEC_GATHER_SRCDI): Ditto.
8213 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
8214 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8216 * config/arm/t-aprofile: New file.
8217 * config.gcc: Handle --with-multilib-list option.
8219 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
8221 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
8224 2013-10-15 Richard Biener <rguenther@suse.de>
8226 * tree-loop-distribution.c (build_empty_rdg): Inline into
8228 (rdg_flag_vertex): Inline into single user.
8229 (rdg_flag_vertex_and_dependent): Likewise.
8230 (build_rdg_partition_for_vertex): Remove processed bitmap.
8231 (rdg_build_partitions): Simplify.
8233 2013-10-15 Richard Biener <rguenther@suse.de>
8235 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
8236 Restructure forwarding through conversions and copies to
8237 avoid performing copy-propagation the wrong way. Adjust
8238 recursion invocations.
8239 (forward_propagate_addr_expr): Add argument stating if we
8240 are recursing from a single-use.
8241 (ssa_forward_propagate_and_combine): Adjust.
8243 2013-10-14 David Malcolm <dmalcolm@redhat.com>
8245 * dumpfile.h (gcc::dump_manager): New class, to hold state
8246 relating to dumpfile management.
8247 (get_dump_file_name): Remove in favor of method of dump_manager.
8248 (dump_initialized_p): Likewise.
8249 (dump_start): Likewise.
8250 (dump_finish): Likewise.
8251 (dump_switch_p): Likewise.
8252 (dump_register): Likewise.
8253 (get_dump_file_info): Likewise.
8254 * context.c (gcc::context::context): Construct the dump_manager
8256 * context.h (gcc::context::get_dumps): New.
8257 (gcc::context::m_dumps): New.
8258 * coverage.c (coverage_init): Port to dump_manager API.
8259 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
8260 (extra_dump_files_in_use): Likewise.
8261 (extra_dump_files_alloced): Likewise.
8262 (gcc::dump_manager::dump_manager): New.
8263 (dump_register): Convert to...
8264 (gcc::dump_manager::dump_register): ...method, replacing
8265 function-static next_dump with m_next_dump field.
8266 (get_dump_file_info): Convert to...
8267 (gcc::dump_manager::get_dump_file_info): ...method.
8268 (get_dump_file_name): Convert to...
8269 (gcc::dump_manager::get_dump_file_name): ...method.
8270 (dump_start): Convert to...
8271 (gcc::dump_manager::dump_start): ...method.
8272 (dump_finish): Convert to...
8273 (gcc::dump_manager::dump_finish): ...method.
8274 (dump_begin): Replace body with...
8275 (gcc::dump_manager::dump_begin): ...new method.
8276 (dump_phase_enabled_p): Convert to...
8277 (gcc::dump_manager::dump_phase_enabled_p): ...method.
8278 (dump_phase_enabled_p): Convert to...
8279 (gcc::dump_manager::dump_phase_enabled_p): ...method.
8280 (dump_initialized_p): Convert to...
8281 (gcc::dump_manager::dump_initialized_p): ...method.
8282 (dump_flag_name): Replace body with...
8283 (gcc::dump_manager::dump_flag_name): ...new method.
8284 (dump_enable_all): Convert to...
8285 (gcc::dump_manager::dump_enable_all): ...new method.
8286 (opt_info_enable_passes): Convert to...
8287 (gcc::dump_manager::opt_info_enable_passes): ...new method.
8288 (dump_switch_p_1): Convert to...
8289 (gcc::dump_manager::dump_switch_p_1): ...new method.
8290 (dump_switch_p): Convert to...
8291 (gcc::dump_manager::dump_switch_p): ...new method.
8292 (opt_info_switch_p): Port to dump_manager API.
8293 (enable_rtl_dump_file): Likewise.
8294 * opts-global.c (handle_common_deferred_options): Port to new
8296 * passes.c (pass_manager::finish_optimization_passes): Likewise.
8297 (pass_manager::register_one_dump_file): Likewise.
8298 (pass_manager::register_pass): Likewise.
8299 (pass_init_dump_file): Likewise.
8300 (pass_fini_dump_file): Likewise.
8301 * statistics.c (statistics_early_init): Likewise.
8303 2013-10-14 Richard Biener <rguenther@suse.de>
8305 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
8306 iterative_hash_canonical_type, gimple_canonical_type_hash,
8307 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
8308 gimple_register_canonical_type, print_gimple_types_stats,
8309 free_gimple_type_tables): Move to lto/lto.c
8310 (gt-gimple.h): Do not include.
8311 * gimple.h (gimple_register_canonical_type,
8312 print_gimple_types_stats, free_gimple_type_tables): Remove.
8313 * Makefile.in (GTFILES): Remove gimple.c.
8315 2013-10-14 Travis Snoozy <quandary@remstate.com>
8318 * config/msp430/msp430.c (msp430_option_override): Correct thinko
8319 scanning for msp430x targets.
8321 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
8324 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
8325 (registers_ok_for_ldd_peep): Move around.
8326 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
8327 * config/sparc/sparc.md (widening peepholes): Use it.
8329 2013-10-14 Richard Biener <rguenther@suse.de>
8333 * gimple.c (iterative_hash_canonical_type): Make sure to
8334 record the hash into the correct hashtable slot.
8336 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
8338 PR rtl-optimization/58662
8339 * combine.c (try_combine): Take into account death nodes on I2 when
8340 splitting a PARALLEL of two independent SETs. Fix dump message.
8342 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
8345 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
8346 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
8347 * config/sh/t-sh (sh_treg_combine.o): New entry.
8348 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
8349 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
8350 (register_sh_passes): New function. Register sh_treg_combine pass.
8351 (sh_option_override): Invoke it.
8352 (sh_canonicalize_comparison): Handle op0_preserve_value.
8353 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
8354 opportunities. Canonicalize branch condition.
8355 (nott): Allow only if pseudos can be created for non-SH2A.
8357 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
8360 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
8361 (ix86_expand_movmem): Replace copy_addr_to_reg with
8362 ix86_copy_addr_to_reg.
8363 (ix86_expand_setmem): Likewise.
8365 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
8367 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
8368 provided by ix86_fp_compare_mode instead of CCFPUmode.
8370 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
8372 * config/aarch64/arm_neon.h
8373 (vtbx<1,3>_<psu>8): Fix register constriants.
8375 2013-10-11 Jeff Law <law@redhat.com>
8377 PR tree-optimization/58640
8378 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
8379 threading paths that cross over two loop entry points.
8381 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8383 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
8384 handle vector float as well.
8385 (*vsx_le_perm_load_v4si): Likewise.
8386 (*vsx_le_perm_store_v2di): Likewise.
8387 (*vsx_le_perm_store_v4si): Likewise.
8389 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8391 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
8392 directly to circumvent subtract from splat{31} workaround.
8393 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
8395 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
8396 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
8397 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
8398 (altivec_vperm_<mode>): Convert to define_insn_and_split to
8399 separate big and little endian logic.
8400 (*altivec_vperm_<mode>_internal): New define_insn.
8401 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
8402 separate big and little endian logic.
8403 (*altivec_vperm_<mode>_uns_internal): New define_insn.
8404 (vec_permv16qi): Add little endian logic.
8406 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
8408 * doc/extend.texi (returns_nonnull): Remove arguments.
8410 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8411 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8412 Sergey Lega <sergey.s.lega@intel.com>
8413 Anna Tikhonova <anna.tikhonova@intel.com>
8414 Ilya Tocar <ilya.tocar@intel.com>
8415 Andrey Turetskiy <andrey.turetskiy@intel.com>
8416 Ilya Verbin <ilya.verbin@intel.com>
8417 Kirill Yukhin <kirill.yukhin@intel.com>
8418 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8420 * config/i386/sse.md (VI48F_256_512): New.
8421 (avx2_permvar<mode>): Change to ...
8422 (<avx2_avx512f>_permvar<mode>): This.
8424 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8425 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8426 Sergey Lega <sergey.s.lega@intel.com>
8427 Anna Tikhonova <anna.tikhonova@intel.com>
8428 Ilya Tocar <ilya.tocar@intel.com>
8429 Andrey Turetskiy <andrey.turetskiy@intel.com>
8430 Ilya Verbin <ilya.verbin@intel.com>
8431 Kirill Yukhin <kirill.yukhin@intel.com>
8432 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8434 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
8435 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
8436 * config/i386/sse.md (VI4_AVX): New.
8437 (sf2simodelower): Ditto.
8438 (sse2_cvtps2dq): Change to ...
8439 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
8441 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8442 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8443 Sergey Lega <sergey.s.lega@intel.com>
8444 Anna Tikhonova <anna.tikhonova@intel.com>
8445 Ilya Tocar <ilya.tocar@intel.com>
8446 Andrey Turetskiy <andrey.turetskiy@intel.com>
8447 Ilya Verbin <ilya.verbin@intel.com>
8448 Kirill Yukhin <kirill.yukhin@intel.com>
8449 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8451 * config/i386/sse.md (V_512): New.
8453 (vcond<V_512:mode><VF_512:mode>): Ditto.
8454 (vcond<V_512:mode><VI_512:mode>): Ditto.
8455 (vcondu<V_512:mode><VI_512:mode>): Ditto.
8457 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8458 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8459 Sergey Lega <sergey.s.lega@intel.com>
8460 Anna Tikhonova <anna.tikhonova@intel.com>
8461 Ilya Tocar <ilya.tocar@intel.com>
8462 Andrey Turetskiy <andrey.turetskiy@intel.com>
8463 Ilya Verbin <ilya.verbin@intel.com>
8464 Kirill Yukhin <kirill.yukhin@intel.com>
8465 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8467 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
8468 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
8470 (fma<mode>4): Removed condition.
8471 (fms<mode>4): Ditto.
8472 (fnma<mode>4): Ditto.
8473 (fnms<mode>4): Ditto.
8474 (fma4i_fmadd_<mode>): Ditto.
8475 (*fma_fmadd_<mode>): Ditto.
8476 (*fma_fmsub_<mode>): Ditto.
8477 (*fma_fnmadd_<mode>): Ditto.
8478 (*fma_fnmsub_<mode>): Ditto.
8479 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
8480 (*fma_fmaddsub_<mode>): Ditto.
8481 (*fma_fmsubadd_<mode>): Ditto.
8482 (*fmai_fmadd_<mode>): Ditto.
8483 (*fmai_fmsub_<mode>): Ditto.
8484 (*fmai_fnmadd_<mode>): Ditto.
8485 (*fmai_fnmsub_<mode>): Ditto.
8487 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8488 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8489 Sergey Lega <sergey.s.lega@intel.com>
8490 Anna Tikhonova <anna.tikhonova@intel.com>
8491 Ilya Tocar <ilya.tocar@intel.com>
8492 Andrey Turetskiy <andrey.turetskiy@intel.com>
8493 Ilya Verbin <ilya.verbin@intel.com>
8494 Kirill Yukhin <kirill.yukhin@intel.com>
8495 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8497 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
8498 (VI124_256): Changed to ...
8499 (VI124_256_48_512): This.
8500 (ssepackmode): Extended with wider modes.
8501 (<code><mode>3): Changed iterator.
8502 (*avx2_<code><mode>3): Ditto.
8503 (vec_pack_trunc_<mode>): Ditto.
8505 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8506 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8507 Sergey Lega <sergey.s.lega@intel.com>
8508 Anna Tikhonova <anna.tikhonova@intel.com>
8509 Ilya Tocar <ilya.tocar@intel.com>
8510 Andrey Turetskiy <andrey.turetskiy@intel.com>
8511 Ilya Verbin <ilya.verbin@intel.com>
8512 Kirill Yukhin <kirill.yukhin@intel.com>
8513 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8515 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
8516 (VI8F_256_512): Ditto.
8517 (abs<mode>2): Changed iterator.
8518 (avx2_perm<mode>): Changed to ...
8519 (<avx2_avx512f>_perm<mode>): This.
8520 (avx2_perm<mode>_1): Changed to ...
8521 (<avx2_avx512f>_perm<mode>_1): This.
8523 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8524 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8525 Sergey Lega <sergey.s.lega@intel.com>
8526 Anna Tikhonova <anna.tikhonova@intel.com>
8527 Ilya Tocar <ilya.tocar@intel.com>
8528 Andrey Turetskiy <andrey.turetskiy@intel.com>
8529 Ilya Verbin <ilya.verbin@intel.com>
8530 Kirill Yukhin <kirill.yukhin@intel.com>
8531 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8533 * config/i386/sse.md (VI48_AVX512F): New.
8534 (VI48_AVX2): Changed to ...
8535 (VI48_AVX2_48_AVX512F): This.
8536 (avx2_ashrv<mode>): Changed to ...
8537 (<avx2_avx512f>_ashrv<mode>): This.
8538 (avx2_<shift_insn>v<mode>): Changed to ...
8539 (<avx2_avx512f>_<shift_insn>v<mode>): This.
8541 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8542 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8543 Sergey Lega <sergey.s.lega@intel.com>
8544 Anna Tikhonova <anna.tikhonova@intel.com>
8545 Ilya Tocar <ilya.tocar@intel.com>
8546 Andrey Turetskiy <andrey.turetskiy@intel.com>
8547 Ilya Verbin <ilya.verbin@intel.com>
8548 Kirill Yukhin <kirill.yukhin@intel.com>
8549 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8551 * config/i386/sse.md (VI4_AVX512F): New.
8552 (VI8_AVX2_AVX512F): Ditto.
8553 (mul<mode>3): Extended with wider modes.
8554 (*<sse4_1_avx2>_mul<mode>3): Ditto.
8555 (mul<mode>3): Ditto.
8556 (vec_widen_<s>mult_odd_<mode>): Ditto.
8558 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8559 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8560 Sergey Lega <sergey.s.lega@intel.com>
8561 Anna Tikhonova <anna.tikhonova@intel.com>
8562 Ilya Tocar <ilya.tocar@intel.com>
8563 Andrey Turetskiy <andrey.turetskiy@intel.com>
8564 Ilya Verbin <ilya.verbin@intel.com>
8565 Kirill Yukhin <kirill.yukhin@intel.com>
8566 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8568 * config/i386/sse.md (VI2_AVX512F): New.
8569 (VI124_AVX512F): Ditto.
8570 (sseunpackmode): Extended with wider modes.
8571 (sseunpackfltmode): Ditto.
8572 (vec_unpacks_float_hi_<mode>): Ditto.
8573 (vec_unpacks_float_lo_<mode>): Ditto.
8574 (vec_unpacku_float_hi_<mode>): Ditto.
8575 (vec_unpacku_float_lo_<mode>): Ditto.
8576 (vec_unpacks_lo_<mode>): Ditto.
8577 (vec_unpacks_hi_<mode>): Ditto.
8578 (vec_unpacku_lo_<mode>): Ditto.
8579 (vec_unpacku_hi_<mode>): Ditto.
8581 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8582 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8583 Sergey Lega <sergey.s.lega@intel.com>
8584 Anna Tikhonova <anna.tikhonova@intel.com>
8585 Ilya Tocar <ilya.tocar@intel.com>
8586 Andrey Turetskiy <andrey.turetskiy@intel.com>
8587 Ilya Verbin <ilya.verbin@intel.com>
8588 Kirill Yukhin <kirill.yukhin@intel.com>
8589 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8591 * config/i386/i386.md (multdiv): New.
8592 (multdiv_mnemonic): Ditto.
8593 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
8594 (<sse>_vm<multdiv_mnemonic><mode>3): This.
8595 (<sse>_vmdiv<mode>3): Removed.
8597 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8598 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8599 Sergey Lega <sergey.s.lega@intel.com>
8600 Anna Tikhonova <anna.tikhonova@intel.com>
8601 Ilya Tocar <ilya.tocar@intel.com>
8602 Andrey Turetskiy <andrey.turetskiy@intel.com>
8603 Ilya Verbin <ilya.verbin@intel.com>
8604 Kirill Yukhin <kirill.yukhin@intel.com>
8605 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8607 * config/i386/sse.md (V): Extended with wider modes.
8609 (ssehalfvecmode): Ditto.
8611 (ssepackfltmode): Ditto.
8612 (avx_vec_concat<mode>): Ditto.
8613 (V_256_512): New iterator.
8614 (VF2_512_256): Ditto.
8615 (si2dfmode): New attribute.
8616 (si2dfmodelower): Ditto.
8618 (concat_tg_mode): Ditto.
8619 (floatv4siv4df2): Changed to ...
8620 (float<si2dfmodelower><mode>2): This.
8621 (avx_cvtps2pd256): Changed to ...
8622 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
8623 (vec_pack_trunc_v4df): Changed to ...
8624 (vec_pack_trunc_<mode>): This.
8625 (avx_vpermil<mode>): Changed to ...
8626 (<sse2_avx_avx512f>_vpermil<mode>): This.
8627 (<fixsuffix>fix_truncv8dfv8si2): New.
8628 (vec_pack_sfix_trunc_v8df): Ditto.
8629 (avx512f_rndscale<mode>): Ditto.
8630 (avx512f_roundpd512): Ditto.
8631 (vec_pack_ufix_trunc_<mode>): Updated iterator.
8633 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8634 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8635 Sergey Lega <sergey.s.lega@intel.com>
8636 Anna Tikhonova <anna.tikhonova@intel.com>
8637 Ilya Tocar <ilya.tocar@intel.com>
8638 Andrey Turetskiy <andrey.turetskiy@intel.com>
8639 Ilya Verbin <ilya.verbin@intel.com>
8640 Kirill Yukhin <kirill.yukhin@intel.com>
8641 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8643 * config/i386/i386.md (any_fix): New iterator.
8644 (fixsuffix): New attribute.
8645 * config/i386/sse.md (VF1): Extened with wider modes.
8649 (sseintvecmodelower): Ditto.
8650 (ssescalarmode): Ditto.
8651 (ssescalarnum): Ditto.
8653 (ssexmmmode): Ditto.
8654 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
8655 (<sse>_rcp<mode>2): Change iterator.
8656 (rsqrt<mode>2): Ditto.
8657 (<sse>_rsqrt<mode>2): Ditto.
8658 (avx2_vec_dup<mode>): Ditto.
8659 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
8660 (round<mode>2_sfix): Ditto.
8661 (avx2_pbroadcast<mode>): Ditto.
8662 (*andnot<mode>3): Handle XI mode.
8663 (*<code><mode>3): Ditto.
8664 (AVXTOSSEMODE): Removed.
8665 (avx_vpermil<mode>): Changed to ...
8666 (<sse2_avx_avx512f>_vpermil<mode>): This.
8668 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8669 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8670 Sergey Lega <sergey.s.lega@intel.com>
8671 Anna Tikhonova <anna.tikhonova@intel.com>
8672 Ilya Tocar <ilya.tocar@intel.com>
8673 Andrey Turetskiy <andrey.turetskiy@intel.com>
8674 Ilya Verbin <ilya.verbin@intel.com>
8675 Kirill Yukhin <kirill.yukhin@intel.com>
8676 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8678 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
8679 (<sse>_comi): Ditto.
8680 (<sse>_ucomi): Ditto.
8681 (sse_cvtss2siq_2): Ditto.
8682 (sse2_cvtsd2si): Ditto.
8683 (sse2_cvtsd2siq): Ditto.
8684 (sse2_cvttsd2si): Ditto.
8685 (sse2_cvttsd2siq): Ditto.
8686 (<shift_insn><mode>3): Ditto.
8687 (sse2_cvtsi2sdq): Update constraint and prefix.
8688 (sse_cvtsi2ss): Update prefix.
8689 (sse_cvtsi2ssq): Ditto.
8691 2013-10-11 Jakub Jelinek <jakub@redhat.com>
8693 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
8694 ignore internal calls.
8696 2013-10-11 Richard Biener <rguenther@suse.de>
8698 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
8699 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
8700 TREE_OVERFLOW is set.
8702 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
8704 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
8706 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
8707 * gimple.h: Likewise.
8709 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
8711 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
8714 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
8716 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
8717 * configure: Regenerate.
8719 2013-10-11 Jakub Jelinek <jakub@redhat.com>
8721 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
8722 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
8723 formatting fixes, use pp_colon instead of pp_character (..., ':'),
8724 similarly pp_right_paren.
8725 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
8726 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
8727 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
8728 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
8730 (omp_declare_simd_clauses_equal,
8731 omp_remove_redundant_declare_simd_attrs): New functions.
8732 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
8733 (walk_tree_1): Handle new OpenMP 4.0 clauses.
8734 * tree.h (OMP_LOOP_CHECK): Define.
8735 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
8736 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
8737 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
8738 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
8739 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
8740 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
8741 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
8742 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
8743 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
8744 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
8745 OMP_CLAUSE_SIMDLEN_EXPR): Define.
8746 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
8747 (omp_remove_redundant_declare_simd_attrs): New prototype.
8748 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
8749 GIMPLE_OMP_TEAMS): New codes.
8750 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
8751 * omp-low.c (struct omp_context): Add cancel_label and cancellable
8753 (target_nesting_level): New variable.
8754 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
8755 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
8756 for collapse > 1 static schedule unless ordered.
8757 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
8758 (determine_parallel_type): Adjust get_ws_args_for caller.
8759 (install_var_field): Handle mask & 4 for double indirection.
8760 (scan_sharing_clauses): Ignore shared clause on teams construct.
8761 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
8762 (create_omp_child_function): If inside target or declare target
8763 constructs, set "omp declare target" attribute on the child function.
8764 (find_combined_for): New function.
8765 (scan_omp_parallel): Handle combined loops.
8766 (scan_omp_target, scan_omp_teams): New functions.
8767 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
8768 restrictions and set ctx->cancellable for cancellable constructs.
8769 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
8770 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
8771 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
8772 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
8773 (omp_clause_aligned_alignment): New function.
8774 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
8775 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
8776 on teams constructs. Handle user defined reductions and new
8778 (lower_reduction_clauses): Don't set placeholder to address of ref
8779 if it has already the right type.
8780 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
8781 (expand_parallel_call): Use the new non-_start suffixed builtins,
8782 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
8783 and GOMP_parallel_end after the call.
8784 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
8785 (expand_omp_for_init_counts): Handle combined loops.
8786 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
8788 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
8789 end of cancellable loops.
8790 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
8791 Likewise. Handle collapse > 1 loops.
8792 (expand_omp_simd): Handle combined loops.
8793 (expand_omp_for): Add inner_stmt argument, adjust callers of
8794 expand_omp_for* functions, use expand_omp_for_static*chunk even
8795 for collapse > 1 unless ordered.
8796 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
8797 of cancellable sections.
8798 (expand_omp_single): Remove need_barrier variable, just rely on
8799 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
8800 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
8801 (expand_omp_atomic_load, expand_omp_atomic_store,
8802 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
8803 (expand_omp_target): New function.
8804 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
8805 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
8806 (build_omp_regions_1): Immediately close region for
8807 GF_OMP_TARGET_KIND_UPDATE.
8808 (maybe_add_implicit_barrier_cancel): New function.
8809 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
8811 (lower_omp_single): Likewise. Add clobber after the barrier.
8812 (lower_omp_taskgroup): New function.
8813 (lower_omp_for): Handle combined loops. Adjust
8814 lower_rec_input_clauses caller. Handle cancellation.
8815 (lower_depend_clauses): New function.
8816 (lower_omp_taskreg): Lower depend clauses. Adjust
8817 lower_rec_input_clauses caller. Add clobber after the call. Handle
8819 (lower_omp_target, lower_omp_teams): New functions.
8820 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
8821 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
8822 and GOMP_cancellation_point calls.
8823 (lower_omp): Fold stmts inside of target region.
8824 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
8825 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
8826 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
8827 (BT_FN_VOID_OMPFN_PTR_UINT,
8828 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
8829 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
8830 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
8831 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
8832 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
8833 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
8834 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
8835 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
8836 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
8837 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
8838 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
8839 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
8840 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
8841 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
8842 BUILT_IN_GOMP_PARALLEL_END.
8843 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
8844 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
8845 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
8846 GF_OMP_FOR_KIND_DISTRIBUTE.
8847 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
8848 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
8849 (dump_gimple_omp_return): Print lhs if it has any.
8850 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
8851 gimple_omp_atomic_seq_cst_p.
8852 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
8853 and GIMPLE_OMP_TEAMS.
8854 * langhooks.c (lhd_omp_mappable_type): New function.
8855 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
8856 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
8858 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
8859 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
8860 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
8861 (struct gimplify_omp_ctx): Add combined_loop field.
8862 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
8863 on stmts inside of target region.
8864 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
8865 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
8866 ORT_TARGET and ORT_TARGET_DATA.
8867 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
8869 (omp_notice_threadprivate_variable): Complain about threadprivate
8870 variables in target region.
8871 (omp_notice_variable): Complain about vars with non-mappable type
8872 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
8873 (omp_check_private): Ignore ORT_TARGET* regions.
8874 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
8875 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
8876 (find_combined_omp_for): New function.
8877 (gimplify_omp_for): Handle gimplification of combined loops.
8878 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
8880 (gimplify_omp_target_update): New function.
8881 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
8882 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
8883 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
8884 (gimplify_body): If fndecl has "omp declare target" attribute, add
8885 implicit ORT_TARGET context around it.
8886 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
8887 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
8888 * tree-nested.c (convert_nonlocal_reference_stmt,
8889 convert_local_reference_stmt, convert_gimple_call): Handle
8890 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
8891 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
8892 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
8893 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
8894 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
8895 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
8896 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
8897 BUILT_IN_GOMP_LOOP_END_CANCEL,
8898 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
8899 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
8900 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
8901 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
8902 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
8903 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
8904 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
8905 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
8906 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
8907 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
8908 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
8909 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
8910 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
8911 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
8912 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
8913 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
8914 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
8915 gimple_build_omp_teams): New functions.
8916 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
8917 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
8918 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
8919 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
8920 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
8921 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
8922 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
8923 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
8924 GF_OMP_ATOMIC_SEQ_CST): New.
8925 (gimple_build_omp_taskgroup, gimple_build_omp_target,
8926 gimple_build_omp_teams): New prototypes.
8927 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
8928 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
8929 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
8930 GIMPLE_OMP_SINGLE as end of range.
8931 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
8932 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
8933 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
8934 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
8935 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
8936 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
8937 gimple_omp_target_kind, gimple_omp_target_set_kind,
8938 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
8939 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
8940 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
8941 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
8942 gimple_omp_teams_set_clauses): New inlines.
8943 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
8944 and GIMPLE_OMP_TASKGROUP.
8945 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
8946 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
8947 enum omp_clause_proc_bind_kind): New.
8948 (union omp_clause_subcode): Add depend_kind, map_kind and
8949 proc_bind_kind fields.
8950 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
8951 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
8952 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
8953 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
8954 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
8956 2013-10-10 Teresa Johnson <tejohnson@google.com>
8958 * predict.c (tree_estimate_probability): Add new parameter
8959 for estimate_bb_frequencies.
8960 (estimate_bb_frequencies): Add new parameter to force estimation.
8961 (rebuild_frequencies): When max frequency in function is small,
8962 recompute counts from frequencies.
8963 * predict.h (estimate_bb_frequencies): New parameter.
8965 2013-10-10 David Malcolm <dmalcolm@redhat.com>
8967 * ipa-inline.c (ipa_inline): Fix leak of "order" when
8968 optimizations are disabled.
8970 2013-10-10 David Malcolm <dmalcolm@redhat.com>
8972 * coverage.c (coverage_finish): Fix leak of da_file_name.
8974 2013-10-10 Jan Hubicka <jh@suse.cz>
8976 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
8977 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
8980 2013-10-10 Jakub Jelinek <jakub@redhat.com>
8983 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
8984 if any labels are in FALLTHRU_BB, use a special label emitted
8985 immediately after the asm goto insn rather than label_rtx
8987 (expand_asm_stmt): Adjust caller.
8988 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
8989 edge if the last insn in predecessor is a jump with single successor,
8990 but it isn't simplejump_p.
8992 2013-10-10 Richard Biener <rguenther@suse.de>
8994 PR tree-optimization/58656
8995 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
8997 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
8999 * gimplify.c: Include expr.h and tm_p.h for targets with special
9000 va-arg padding requirements.
9002 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
9004 * tree-flow.h: Move some prototypes to gimple.h.
9005 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
9006 * gimple.h: Relocate some prototypes from tree-flow.h
9007 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
9009 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
9010 (build_va_arg_indirect_ref): Relocate and make static.
9011 (std_gimplify_va_arg_expr): Relocate here.
9012 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
9013 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
9015 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9017 * doc/md.texi: Document the mnemonic attribute.
9019 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9022 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
9023 (set_attr_alternative x ...) when searching for user defined
9026 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
9028 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
9030 2013-10-09 Easwaran Raman <eraman@google.com>
9032 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
9033 * cfgexpand.c (defer_stack_allocation): ...use here
9034 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
9036 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
9038 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
9039 (operand_equal_for_value_replacement): New function, extracted from
9040 value_replacement and enhanced to catch more cases.
9041 (value_replacement): Use operand_equal_for_value_replacement.
9043 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
9045 * loop-doloop.c (doloop_modify, doloop_optimize): Use
9046 get_max_loop_iterations.
9048 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9050 * config/arm/aarch-common.c (arm_early_load_addr_dep):
9051 Place comment above function.
9053 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
9055 * tree-flow.h: Remove all remaining prototypes, enums and structs that
9056 are not related to tree-cfg.c.
9057 * tree-ssa-address.h: New file. Relocate prototypes.
9058 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
9059 (addr_for_mem_ref): New. Combine call to get_address_description and
9060 return addr_for_mem_ref.
9061 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
9062 * tree-ssa-live.h: Adjust prototypes.
9063 * passes.c: Include tree-ssa-live.h.
9064 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
9065 * graphite.c (graphite_transform_loops): Make static.
9066 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
9067 make_pass_graphite, pass_data_graphite_transforms,
9068 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
9069 * ipa-pure-const.c (warn_function_noreturn): Make static.
9070 (execute_warn_function_noreturn, gate_warn_function_noreturn,
9071 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
9072 Relocate from tree-cfg.c
9073 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
9075 (execute_warn_function_noreturn, gate_warn_function_noreturn,
9076 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
9077 Move to ipa-pure-const.c.
9078 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
9079 Relocate from tree-optimize.c.
9080 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
9081 make_pass_fixup_cfg): Move to tree-cfg.c.
9082 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
9083 Relocate some prototypes.
9084 * tree-data-ref.h (tree_check_data_deps) Add prototype.
9085 * tree-dump.c (dump_function_to_file): Remove prototype.
9086 Add tree-flow.h to the include file.
9087 * tree-dump.h: Remove prototype.
9088 * tree-parloops.h: New File. Add prototypes.
9089 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
9090 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
9091 from tree-ssa-loop.c.
9092 * tree-predcom.c (run_tree_predictive_commoning,
9093 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
9094 Relocate here from tree-ssa-loop.c.
9095 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
9096 ssa_name_values.release ().
9097 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
9098 (ssa_name_values): Relocate from tree-flow.h.
9099 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
9100 * tree-ssa-loop.c (run_tree_predictive_commoning,
9101 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
9102 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
9103 make_pass_graphite, pass_data_graphite_transforms,
9104 make_pass_graphite_transforms, gate_tree_parallelize_loops,
9105 tree_parallelize_loops, pass_data_parallelize_loops,
9106 make_pass_parallelize_loops): Move to other files.
9107 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
9109 * tree.h: Remove prototypes from tree-address.c.
9111 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
9113 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
9114 (struct int_tree_map): Move to tree-hasher.h
9115 (SCALE, LABEL, PERCENT): Move to gimple.h
9116 * tree-flow-inline.h: Delete. Move functions to other files.
9117 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
9118 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
9119 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
9120 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
9121 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
9122 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
9123 * tree-hasher.h: Don't include tree-flow.h.
9124 (struct int_tree_map): Relocate from tree-flow.h.
9125 * tree-sra.c (contains_view_convert_expr_p): Relocate from
9126 tree-flow-inline.h and make static.
9127 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
9129 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
9130 tree-flow-inline.h and make static.
9131 * tree.h (is_global_var, may_be_aliased): Relocate from
9133 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
9134 * value-prof.c: No longer include tree-flow-inline.h.
9135 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
9137 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
9139 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
9140 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
9141 (enum move_pos): Move to tree-ssa-loop-im.h
9142 * cfgloop.h: Move some prototypes.
9143 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
9144 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
9145 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
9146 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
9147 (loop_containing_stmt): Relocate from tree-flow-inline.h.
9148 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
9149 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
9150 (enum move_pos): Relocate here.
9151 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
9153 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
9154 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
9155 make_pass_lim): Relocate here from tree-ssa-loop.c.
9156 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
9158 (loop_edge_to_cancel, unloop_loops): Make static.
9159 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
9160 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
9161 (tree_complete_unroll, gate_tree_complete_unroll,
9162 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
9163 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
9164 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
9165 * tree-ssa-loop-ivopts.c: Remove local prototypes.
9166 (stmt_invariant_in_loop_p): Remove unused function.
9167 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
9168 * tree-ssa-loop-manip.h: New file. Add prototypes.
9169 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
9170 (gcov_type_to_double_int): Move to cfgloop.h.
9171 (double_int_cmp, bound_index,
9172 estimate_numbers_of_iterations_loop): Make static.
9173 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
9174 (max_loop_iterations): Factor out get_max_loop_iterations.
9175 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
9177 * tree-ssa-loop-niter.h: New file. Add prototypes.
9178 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
9179 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
9180 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
9181 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
9182 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
9183 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
9184 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
9185 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
9186 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
9187 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
9188 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
9189 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
9190 pass_data_complete_unroll, make_pass_complete_unroll,
9191 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
9192 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
9193 tree-ssa-loop-ivcanon.c.
9194 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
9195 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
9196 tree-ssa-loop-prefetch.c.
9197 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
9199 (get_lsm_tmp_name): Relocate and add suffix parameter.
9200 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
9201 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
9202 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
9203 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
9204 (get_estimated_loop_iterations): Factor out accessor from
9205 estimated_loop_iterations in tree-ssa-loop-niter.c.
9206 (get_max_loop_iterations): Factor out accessor from
9207 _max_loop_iterations in tree-ssa-niter.c.
9208 * loop-unroll.c (decide_unroll_constant_iterations,
9209 decide_unroll_runtime_iterations, decide_peel_simple,
9210 decide_unroll_stupid): Use new get_* accessors.
9212 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
9214 PR tree-optimization/20318
9215 * doc/extend.texi (returns_nonnull): New function attribute.
9216 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
9218 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
9219 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
9221 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
9224 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
9225 false if both components are bitfields.
9227 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
9229 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
9230 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
9231 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
9232 * config/aarch64/aarch64.h
9233 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
9234 * config/aarch64/aarch64-simd-builtins.def
9235 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
9237 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
9239 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
9241 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
9243 * config/aarch64/arm_neon.h (vdiv_f64): Added.
9245 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
9247 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
9248 (vneg_f64): New intrinsic.
9249 (vneg_s8): Asm replaced with C.
9250 (vneg_s16): Likewise.
9251 (vneg_s32): Likewise.
9252 (vneg_s64): New intrinsic.
9253 (vnegq_f32): Asm replaced with C.
9254 (vnegq_f64): Likewise.
9255 (vnegq_s8): Likewise.
9256 (vnegq_s16): Likewise.
9257 (vnegq_s32): Likewise.
9258 (vnegq_s64): Likewise.
9260 2013-10-09 Renlin Li <Renlin.Li@arm.com>
9262 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
9264 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9266 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
9267 packed stack special handling.
9268 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
9269 back to fixed stack slots for FPRs saved due to stdarg.
9271 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9273 * config/s390/s390.c (s390_frame_info): Restructure function.
9275 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9277 * config/s390/s390.c (struct s390_frame_layout): New field
9279 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
9280 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
9281 regs_ever_clobbered to char*.
9282 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
9283 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
9284 (s390_register_info): Enable FPR save slots. Move/Copy some
9285 functionality into ...
9286 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
9287 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
9289 (s390_frame_info): Do gpr slot allocation here now. stdarg does
9290 not imply a stack frame.
9291 (s390_init_frame_layout): Remove variable clobbered_regs.
9292 (s390_update_register_info): Remove function.
9293 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
9294 cannot be used for register renaming.
9295 (s390_hard_regno_scratch_ok): New function.
9296 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
9297 (s390_initial_elimination_offset): Change offset calculation of
9298 the return address pointer.
9299 (save_gprs): Deal with only r6 being saved from the call-saved regs.
9300 (restore_gprs): Set frame related flag.
9301 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
9302 (s390_emit_prologue): Call s390_register_info instead of
9303 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
9304 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
9305 (s390_optimize_prologue): Call s390_optimize_register_info.
9306 Try to remove also FPR slot save/restore INSNs. Remove frame
9307 related flags from restore INSNs.
9309 2013-10-08 DJ Delorie <dj@redhat.com>
9311 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
9314 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
9315 avoid conflict with the MI use of %c.
9316 * config/rl78/rl78-real.md: change %c to %C throughout.
9317 * config/rl78/rl78-virt.md: Likewise.
9319 2013-10-08 Jan Hubicka <jh@suse.cz>
9321 * config/i386/i386.c (ix86_option_override_internal): Switch
9322 to SSE math for -ffast-math when target ISA supports SSE2.
9324 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
9326 * tree-flow.h: Remove some prototypes.
9327 * tree.h: Remove some protypes, add a couple.
9328 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
9329 using_eh_for_cleanups_p): Add interface routines for front ends.
9330 * tree-eh.h: New file. Add protoptyes.
9331 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
9332 (add_stmt_to_eh_lp_fn): Make static.
9333 (lower_try_finally): Use new using_eh_for_cleanups_p.
9334 * emit-rtl.c: Include tree-eh.h.
9335 * gimple.h: Include tree-eh.h.
9337 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
9339 PR tree-optimization/58480
9340 * tree-vrp.c (infer_nonnull_range): New function.
9341 (infer_value_range): Call infer_nonnull_range.
9343 2013-10-08 Dehao Chen <dehao@google.com>
9345 PR tree-optimization/58619
9346 * tree-inline.c (copy_phis_for_bb): Combine location data
9349 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9352 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
9353 RTX_FRAME_RELATED_P on INSN.
9355 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9357 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
9358 (altivec_expand_vec_perm_const): Call it.
9360 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9362 * config/rs6000/vector.md (mov<mode>): Emit permuted move
9363 sequences for LE VSX loads and stores at expand time.
9364 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
9366 * config/rs6000/rs6000.c (rs6000_const_vec): New.
9367 (rs6000_gen_le_vsx_permute): New.
9368 (rs6000_gen_le_vsx_load): New.
9369 (rs6000_gen_le_vsx_store): New.
9370 (rs6000_gen_le_vsx_move): New.
9371 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
9372 (*vsx_le_perm_load_v4si): New.
9373 (*vsx_le_perm_load_v8hi): New.
9374 (*vsx_le_perm_load_v16qi): New.
9375 (*vsx_le_perm_store_v2di): New.
9376 (*vsx_le_perm_store_v4si): New.
9377 (*vsx_le_perm_store_v8hi): New.
9378 (*vsx_le_perm_store_v16qi): New.
9379 (*vsx_xxpermdi2_le_<mode>): New.
9380 (*vsx_xxpermdi4_le_<mode>): New.
9381 (*vsx_xxpermdi8_le_V8HI): New.
9382 (*vsx_xxpermdi16_le_V16QI): New.
9383 (*vsx_lxvd2x2_le_<mode>): New.
9384 (*vsx_lxvd2x4_le_<mode>): New.
9385 (*vsx_lxvd2x8_le_V8HI): New.
9386 (*vsx_lxvd2x16_le_V16QI): New.
9387 (*vsx_stxvd2x2_le_<mode>): New.
9388 (*vsx_stxvd2x4_le_<mode>): New.
9389 (*vsx_stxvd2x8_le_V8HI): New.
9390 (*vsx_stxvd2x16_le_V16QI): New.
9392 2013-10-07 Renlin Li <Renlin.Li@arm.com>
9394 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
9396 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9398 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
9399 loop to work also for 31bit ABI.
9400 Save the stack pointer for frame_size > 0.
9402 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9404 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
9405 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
9406 constraint letters from expanders.
9407 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
9408 retry count to general_operand.
9409 ("tabort"): Give operand 0 a mode.
9410 ("tabort_1"): Add mode and constraint letter for operand 0.
9411 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
9413 2013-10-04 Jeff Law <law@redhat.com>
9415 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
9417 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
9419 (thread_across_edge): Here. Call it.
9421 2013-10-04 Cary Coutant <ccoutant@google.com>
9423 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
9424 discarding a location list expression (or a piece of one).
9426 2013-10-03 Jan Hubicka <jh@suse.cz>
9428 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
9429 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
9430 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
9432 2013-10-03 Jan Hubicka <jh@suse.cz>
9434 * config/i386/i386.c (ix86_option_override_internal): Do not enable
9435 accumulate-outgoing-args when producing unwind info.
9437 2013-10-03 Wei Mi <wmi@google.com>
9439 * lra-constraints.c (insert_move_for_subreg): New function
9440 extracted from simplify_operand_subreg.
9441 (simplify_operand_subreg): Add reload for paradoxical subreg.
9443 2013-10-03 Rong Xu <xur@google.com>
9445 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
9446 the candidate of builtin_expect such that we should fix the
9447 size/time estimation.
9448 (estimate_function_body_sizes): Do the acutally size/time fix-up
9451 2013-10-03 Rong Xu <xur@google.com>
9453 * predict.c (tree_predict_by_opcode): Get the probability
9454 for builtin_expect from param builtin_expect_probability.
9455 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
9456 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
9457 * doc/invoke.texi: Add documentation for builtin-expect-probability.
9459 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
9462 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
9463 * calls.c (alloca_call_p): Use get_callee_fndecl.
9464 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
9465 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
9467 (vrp_visit_stmt): Remove duplicated code.
9469 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
9471 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
9472 ceildf2, btruncdf2, instead of vsx_* name.
9474 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
9475 iterators to only do V2DF and V4SF here. Move the DF code to
9476 rs6000.md where it is combined with SF mode. Replace <VSv> with
9477 just 'v' since only vector operations are handled with these insns
9478 after moving the DF support to rs6000.md.
9479 (vsx_sub<mode>3): Likewise.
9480 (vsx_mul<mode>3): Likewise.
9481 (vsx_div<mode>3): Likewise.
9482 (vsx_fre<mode>2): Likewise.
9483 (vsx_neg<mode>2): Likewise.
9484 (vsx_abs<mode>2): Likewise.
9485 (vsx_nabs<mode>2): Likewise.
9486 (vsx_smax<mode>3): Likewise.
9487 (vsx_smin<mode>3): Likewise.
9488 (vsx_sqrt<mode>2): Likewise.
9489 (vsx_rsqrte<mode>2): Likewise.
9490 (vsx_fms<mode>4): Likewise.
9491 (vsx_nfma<mode>4): Likewise.
9492 (vsx_copysign<mode>3): Likewise.
9493 (vsx_btrunc<mode>2): Likewise.
9494 (vsx_floor<mode>2): Likewise.
9495 (vsx_ceil<mode>2): Likewise.
9496 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
9497 (vsx_sminsf3): Likewise.
9498 (vsx_fmadf4): Likewise.
9499 (vsx_fmsdf4): Likewise.
9500 (vsx_nfmadf4): Likewise.
9501 (vsx_nfmsdf4): Likewise.
9502 (vsx_cmpdf_internal1): Likewise.
9504 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
9505 simpler to select whether a target has SPE or traditional floating
9506 point support in iterators.
9507 (TARGET_DF_SPE): Likewise.
9508 (TARGET_SF_FPR): Likewise.
9509 (TARGET_DF_FPR): Likewise.
9510 (TARGET_SF_INSN): Macros to say whether floating point support
9511 exists for a given operation for expanders.
9512 (TARGET_DF_INSN): Likewise.
9514 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
9515 combining of SF/DF mode operations, using both traditional and VSX
9523 (abs<mode>2): Combine SF/DF modes using traditional floating point
9524 instructions. Add support for using the upper DF registers with
9525 VSX support, and SF registers with power8-vector support. Update
9526 expanders for operations supported by both the SPE and traditional
9527 floating point units.
9528 (abs<mode>2_fpr): Likewise.
9529 (nabs<mode>2): Likewise.
9530 (nabs<mode>2_fpr): Likewise.
9531 (neg<mode>2): Likewise.
9532 (neg<mode>2_fpr): Likewise.
9533 (add<mode>3): Likewise.
9534 (add<mode>3_fpr): Likewise.
9535 (sub<mode>3): Likewise.
9536 (sub<mode>3_fpr): Likewise.
9537 (mul<mode>3): Likewise.
9538 (mul<mode>3_fpr): Likewise.
9539 (div<mode>3): Likewise.
9540 (div<mode>3_fpr): Likewise.
9541 (sqrt<mode>3): Likewise.
9542 (sqrt<mode>3_fpr): Likewise.
9543 (fre<Fs>): Likewise.
9544 (rsqrt<mode>2): Likewise.
9545 (cmp<mode>_fpr): Likewise.
9546 (smax<mode>3): Likewise.
9547 (smin<mode>3): Likewise.
9548 (smax<mode>3_vsx): Likewise.
9549 (smin<mode>3_vsx): Likewise.
9550 (negsf2): Delete SF operations that are merged with DF.
9557 (fmasf4_fpr): Likewise.
9558 (fmssf4_fpr): Likewise.
9559 (nfmasf4_fpr): Likewise.
9560 (nfmssf4_fpr): Likewise.
9561 (sqrtsf2): Likewise.
9562 (rsqrtsf_internal1): Likewise.
9563 (smaxsf3): Likewise.
9564 (sminsf3): Likewise.
9565 (cmpsf_internal1): Likewise.
9566 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
9567 (negdf2): Delete DF operations that are merged with SF.
9569 (nabsdf2): Likewise.
9575 (rsqrtdf_internal1): Likewise.
9576 (fmadf4_fpr): Likewise.
9577 (fmsdf4_fpr): Likewise.
9578 (nfmadf4_fpr): Likewise.
9579 (nfmsdf4_fpr): Likewise.
9580 (sqrtdf2): Likewise.
9581 (smaxdf3): Likewise.
9582 (smindf3): Likewise.
9583 (cmpdf_internal1): Likewise.
9584 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
9585 (btrunc<mode>2): Delete separate expander, and combine with the
9586 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
9587 (btrunc<mode>2_fpr): Likewise.
9588 (ceil<mode>2): Likewise.
9589 (ceil<mode>2_fpr): Likewise.
9590 (floor<mode>2): Likewise.
9591 (floor<mode>2_fpr): Likewise.
9592 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
9593 Add support for using the upper registers with VSX and
9594 power8-vector. Move insns to be closer to the define_expands. On
9595 VSX systems, prefer the traditional form of FMA over the VSX
9596 version, since the traditional form allows the target not to
9597 overlap with the inputs.
9598 (fms<mode>4_fpr): Likewise.
9599 (nfma<mode>4_fpr): Likewise.
9600 (nfms<mode>4_fpr): Likewise.
9602 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9603 Richard Earnshaw <richard.earnshaw@arm.com>
9605 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
9606 (struct mult_cost_table): Likewise.
9607 (struct mem_cost_table): Likewise.
9608 (struct fp_cost_table): Likewise.
9609 (struct vector_cost_table): Likewise.
9610 (cpu_cost_table): Likewise.
9611 * config/arm/arm.opt (mold-rts-costs): New option.
9612 (mnew-generic-costs): Likewise.
9613 * config/arm/arm.c (generic_extra_costs): New table.
9614 (cortexa15_extra_costs): Likewise.
9615 (arm_slowmul_tune): Use NULL as new costs.
9616 (arm_fastmul_tune): Likewise.
9617 (arm_strongarm_tune): Likewise.
9618 (arm_xscale_tune): Likewise.
9619 (arm_9e_tune): Likewise.
9620 (arm_v6t2_tune): Likewise.
9621 (arm_cortex_a5_tune): Likewise.
9622 (arm_cortex_a9_tune): Likewise.
9623 (arm_v6m_tune): Likewise.
9624 (arm_fa726te_tune): Likewise.
9625 (arm_cortex_a15_tune): Use cortex15_extra_costs.
9626 (arm_cortex_tune): Use generict_extra_costs.
9627 (shifter_op_p): New function.
9628 (arm_unspec_cost): Likewise.
9629 (LIBCALL_COST): Define.
9630 (arm_new_rtx_costs): New function.
9631 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
9632 table is available. Use old costs otherwise unless mnew-generic-costs
9634 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
9635 (cpu_cost_table): Declare.
9637 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
9640 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
9641 (*subs_mul_imm_<mode>)
9642 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
9643 (*sub_<shift>_<mode>)
9644 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
9645 Remove k constraint.
9647 2013-10-03 Ian Bolton <ian.bolton@arm.com>
9649 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
9651 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
9653 2013-10-02 Teresa Johnson <tejohnson@google.com>
9655 * predict.c (probably_never_executed): New function.
9656 (probably_never_executed_bb_p): Invoke probably_never_executed.
9657 (probably_never_executed_edge_p): Ditto.
9658 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
9659 Treat profile insanities conservatively.
9661 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
9663 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
9665 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
9667 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
9668 use it. Use smaller increase for scratch. Don't increase reject
9669 for early clobber scratch.
9670 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
9671 setting eliminated regs except setting fp from hfp.
9672 (lra_eliminate): Check lra_insn_recog_data on NULL.
9674 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
9677 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
9678 setting -mvsx-timode by default until the underlying problem is fixed.
9679 (RS6000_CPU, power7 defaults): Likewise.
9681 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
9683 * config/x-linux (host-linux.o): Remove header dependencies.
9684 Use $(COMPILE) and $(POSTCOMPILE).
9685 * config/t-linux-android (linux-android.o): Ditto.
9687 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
9689 * Makefile.in (expmed.o-warn): Remove.
9691 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
9693 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
9694 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
9696 2013-10-02 Teresa Johnson <tejohnson@google.com>
9698 * dojump.c (do_jump_1): Divide probability between
9699 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
9701 2013-10-02 Tom Tromey <tromey@redhat.com>
9703 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
9705 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
9707 * tree-flow.h: Remove some prototypes.
9708 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
9709 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
9710 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
9711 mark_virtual_phi_result_for_renaming): Relocate here.
9712 * tree-into-ssa.h: Add prototypes.
9713 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
9714 single_pred_before_succ_order.
9715 (blocks_in_phiopt_order): Rename and move to cfganal.c.
9716 (nonfreeing_call_p) Move to gimple.c.
9717 * cfganal.c (single_pred_before_succ_order): Move and renamed from
9719 * basic-block.h (single_pred_before_succ_order): Add prototype.
9720 * gimple.c (nonfreeing_call_p): Relocate here.
9721 * gimple.h: Add prototype.
9722 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
9723 * tree-ssa-dom.h: New file. Relocate prototypes here.
9724 * tree-ssa.h: Include tree-ssa-dom.h.
9726 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
9728 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
9729 Use $(COMPILE) and $(POSTCOMPILE).
9731 * config/alpha/x-alpha (driver-alpha.o): Ditto.
9733 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
9735 * tree-flow.h: Remove some prototypes.
9736 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
9737 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
9738 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
9739 propagate_tree_value*): Move from here to...
9740 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
9741 propagate_tree_value*): Relocate here.
9742 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
9743 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
9744 * gimple-fold.c: Remove gimple-fold.h from include list.
9745 * tree-vrp.c: Remove gimple-fold.h from include list.
9746 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
9747 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
9748 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
9749 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
9750 * sese.c: Add tree-ssa-propagate.h to include list.
9752 2013-10-02 Richard Biener <rguenther@suse.de>
9754 * tree-loop-distribution.c: Include tree-vectorizer.h for
9756 (enum partition_kind): Remove PKIND_REDUCTION.
9757 (struct partition_s): Remove has_writes member, add reduction_p member.
9758 (partition_alloc): Adjust.
9759 (partition_builtin_p): Likewise.
9760 (partition_has_writes): Remove.
9761 (partition_reduction_p): New function.
9762 (partition_merge_into): Likewise.
9763 (generate_code_for_partition): Commonize builtin partition
9765 (rdg_cannot_recompute_vertex_p): Remove.
9766 (already_processed_vertex_p): Likewise.
9767 (rdg_flag_vertex): Do not set has_writes.
9768 (classify_partition): Adjust.
9769 (rdg_build_partitions): Do not set has_writes, treat all
9770 partitions as useful.
9771 (distribute_loop): Record number of library calls generated. Adjust.
9772 (tree_loop_distribution): Report number of loops and library
9773 calls generated as opt-info.
9775 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
9777 * tree-flow.h: Include new .h files. Move prototypes.
9778 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
9779 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
9780 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
9781 * tree-pretty-print.h: Add prototypes from tree-flow.h.
9782 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
9783 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
9784 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
9785 * tree.h (get_ref_base_and_extent): Move prototype out.
9786 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
9788 * gimple-low.h: New File. Add prototypes from tree-flow.h.
9789 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
9790 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
9791 * tree-scalar-evolution.c: Include tree.h.
9792 * sese.c: Include tree.h.
9793 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
9794 * dwarf2out.c: Include tree-dfa.h.
9795 * tree-chrec.c: Include tree.h.
9796 * tree-data-ref.c: Include tree.h.
9798 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
9800 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
9803 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9805 * config/t-sol2 (sol2-c.o): Remove header dependencies.
9806 Use $(COMPILE) and $(POSTCOMPILE).
9807 (sol2-cxx.o): Likewise.
9808 (sol2-stubs.o): Likewise.
9810 * config/x-solaris (host-solaris.o): Likewise.
9812 * config/sparc/t-sparc (sparc.o): Remove.
9813 (sparc-c.o): Remove header dependencies.
9814 Use $(COMPILE) and $(POSTCOMPILE).
9815 * config/sparc/x-sparc: Likewise.
9817 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
9819 * config/arc/arc-opts.h: Add 2013 to Copyright years.
9820 * config/arc/arc700.md: Likewise.
9821 * config/arc/arc-modes.def: Likewise.
9822 * config/arc/arc-simd.h: Likewise.
9823 * config/arc/t-arc-uClibc: Likewise.
9824 * config/arc/t-arc-newlib: Likewise.
9826 2013-10-02 Renlin Li <renlin.li@arm.com>
9828 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
9830 (aarch64_expand_epilogue): Likewise.
9832 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9833 Yufeng Zhang <yufeng.zhang@arm.com>
9835 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
9837 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
9838 'base_in' represent a conversion and legal_cast_p_1 holds; set
9839 'base_in' with the returned value from get_unwidened.
9841 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9843 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
9844 plus_constant is not used.
9846 2013-10-01 Wei Mi <wmi@google.com>
9848 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
9849 * config/i386/i386.md: Add define_peephole2 to
9850 break partial reg stall for cvtss2sd/cvtsd2ss.
9852 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
9854 * config/arc/arc.c (pass_arc_ifcvt::clone):
9855 Update for ctxt_ -> m_ctxt change.
9857 2013-10-01 Jeff Law <law@redhat.com>
9859 * tree-ssa-threadupdate.c (struct redirection_data): Delete
9860 outgoing_edge and intermediate_edge fields. Instead store the path.
9861 (redirection_data::hash): Hash on the last edge's destination index.
9862 (redirection_data::equal): Check the entire thread path.
9863 (lookup_redirectio_data): Corresponding changes.
9864 (create_edge_and_update_destination_phis): Likewise.
9865 (thread_single_edge): Likewise.
9867 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
9868 Diego Novillo <dnovillo@google.com>
9870 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
9871 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
9872 (vld32wh_insn, vld32wl_insn): Delete commented-out old
9873 versions of these patterns.
9875 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
9876 (__builtin_arc_aligned): Likewise.
9878 * config/arc/arc.md: Expand adc_0 comment stating the intended
9879 purpose and why it isn't ready.
9880 Replace commented out call_value_via_label_mixed with a
9881 plain comment about bl_s.
9883 * config/arc/arc.c (stdio.h): Don't include directly.
9884 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
9885 Remove if (1) condition.
9886 (arc_encode_section_info): Fix comment.
9888 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
9890 * config/arc/arc.c (arc_conditional_register_usage):
9891 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
9892 Also set reg_alloc_order for DMA config regs.
9894 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
9895 Jeremy Bennett <jeremy.bennett@embecosm.com>
9897 * doc/install.texi (--with-cpu): Mention ARC.
9898 (arc-*-elf32): New paragraph.
9899 (arc-linux-uclibc): Likewise.
9900 * doc/md.texi (Machine Constraints): Add ARC part.
9901 * doc/invoke.texi: (menu): Add ARC Options.
9902 (Machine Dependent Options) <ARC Options>: Add synopsis.
9903 (node ARC Options): Add.
9904 * doc/extend.texi (long_call / short_call attribute): Add ARC.
9905 (ARC Built-in Functions): New section defining
9906 generic ARC built-in functions.
9907 (ARC SIMD Built-in Functions): New section defining SIMD specific
9909 (Declaring Attributes of Functions): Extended
9910 description of short_call and long_call attributes for ARC and
9911 added index entries.
9913 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
9914 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
9915 Joern Rennecke <joern.rennecke@embecosm.com>
9916 Muhammad Khurram Riaz <khurram.riaz@arc.com>
9917 Brendan Kehoe <brendan@zen.org>
9918 Michael Eager <eager@eagercon.com>
9919 Simon Cook <simon.cook@embecosm.com>
9920 Jeremy Bennett <jeremy.bennett@embecosm.com>
9922 * config/arc, common/config/arc: New directories.
9924 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
9925 Brendan Kehoe <brendan@zen.org>
9926 Simon Cook <simon.cook@embecosm.com>
9928 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
9930 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
9932 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
9933 * tree-ssa-coalesce.h: New. Move prototype to here.
9934 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
9935 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
9936 (gimple_can_coalesce_p): Move to...
9937 * gimple.c (gimple_can_coalesce_p): Here.
9939 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
9941 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
9942 (dump_decl_set): Move to gimple.c.
9943 * gimple.h: Don't include tree-ssa-operands.h.
9944 (dump_decl_set): Add prototype.
9945 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
9946 Move to gimple-ssa.h.
9947 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
9948 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
9949 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
9950 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
9951 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
9953 (dump_decl_set): Relocate here.
9954 * gimple-ssa.h: New file.
9955 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
9956 Relocate from gimple.h.
9957 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
9958 * tree-ssa-operands.c (swap_ssa_operands): Rename from
9959 swap_tree_operands and remove non-ssa path.
9960 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
9961 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
9963 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
9964 vect_is_simple_reduction_1): Use swap_ssa_operands.
9965 * tree-flow.h: Move various prototypes to tree-phinodes.h.
9966 (enum need_phi_state): Move to tree-into-ssa.c.
9967 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
9968 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
9969 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
9970 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
9971 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
9972 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
9973 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
9974 num_imm_uses): Move to ssa-iterators.h.
9975 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
9976 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
9978 (op_iter_done, op_iter_next_def, op_iter_next_tree,
9979 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
9980 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
9981 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
9982 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
9983 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
9984 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
9985 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
9986 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
9987 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
9988 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
9989 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
9990 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
9991 (set_phi_nodes): Move to tree-phinodes.h.
9992 * tree-ssa-operands.h (enum ssa_op_iter_type,
9993 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
9994 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
9995 (dump_decl_set): Remove prototype.
9996 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
9997 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
9998 (set_phi_nodes): Relocate from tree-flow-inline.h.
9999 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
10001 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
10002 include list. Temporarily add gimple.h to include list.
10003 * ssa-iterators.h: New file.
10004 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
10005 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
10006 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
10007 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
10008 Relocate from tree-ssa-operands.h.
10009 (delink_imm_use, link_imm_use_to_list, link_imm_use,
10010 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
10011 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
10012 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
10013 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
10014 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
10015 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
10016 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
10017 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
10018 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
10019 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
10020 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
10021 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
10022 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
10023 Relocate from tree-flow-inline.h.
10024 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
10026 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
10029 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
10030 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
10031 Insert '\t' to output template.
10032 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
10033 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
10034 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
10035 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
10037 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
10039 * doc/install.texi (Host/target specific installation notes for GCC):
10040 Put @anchor before @heading.
10041 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
10042 Use @email for email addresses.
10044 2013-10-01 Jeff Law <law@redhat.com>
10046 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
10047 a vec. Only delete the path if we create one without successfully
10048 registering a jump thread.
10049 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
10051 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
10052 (paths): New vector of jump threading paths.
10053 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
10054 (THREAD_PATH): New accessor macro for the entire thread path.
10055 (lookup_redirection_data): Get intermediate and final outgoing edge
10056 from the thread path.
10057 (create_edge_and_update_destination_phis): Copy the threading path.
10058 (ssa_fix_duplicate_block_edges): Get edges and block types from the
10059 jump threading path.
10060 (ssa_redirect_edges): Get edges and block types from the jump threading
10061 path. Free the path vector.
10062 (thread_block): Get edges from the jump threading path. Look at the
10063 entire path to see if we thread to a loop exit. If we cancel a jump
10064 thread request, then free the path vector.
10065 (thread_single_edge): Get edges and block types from the jump threading
10066 path. Free the path vector.
10067 (thread_through_loop_header): Get edges and block types from the jump
10068 threading path. Free the path vector.
10069 (mark_threaded_blocks): Iterate over the vector of paths and store
10070 the path on the appropriate edge. Get edges and block types from the
10071 jump threading path.
10072 (mark_threaded_blocks): Get edges and block types from the jump
10073 threading path. Free the path vector.
10074 (thread_through_all_blocks): Use the vector of paths rather than
10075 a vector of 3-edge sets.
10076 (register_jump_thread): Accept pointer to a path vector rather
10077 than the path vector itself. Store the path vector for later use.
10080 2013-10-01 Jakub Jelinek <jakub@redhat.com>
10081 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10084 * config/s390/s390.c (s390_split_branches): Modify check for table
10086 (s390_chunkify_start): Rearrange table jump insn check in order to
10087 deal with compare and branch insns correctly.
10089 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
10093 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
10094 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
10095 define_insn_and_split.
10096 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
10097 (shiftsi3_compare): New pattern.
10098 (rrx): New pattern.
10099 * config/arm/unspecs.md (UNSPEC_RRX): New.
10101 2013-10-01 Alan Modra <amodra@gmail.com>
10103 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
10104 casing inout operands.
10106 2013-10-01 Richard Biener <rguenther@suse.de>
10108 PR tree-optimization/58553
10109 * tree-loop-distribution.c (struct partition_s): Add niter member.
10110 (classify_partition): Populate niter member for the partition
10111 and properly identify whether the relevant store happens before
10112 or after the loop exit.
10113 (generate_memset_builtin): Use niter member from the partition.
10114 (generate_memcpy_builtin): Likewise.
10116 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
10118 * vec.h (vec_prefix, vec): Prefix member names with "m_".
10119 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
10121 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
10123 * basic-block.h (edge_list): Prefix member names with "m_".
10124 * context.h (context): Likewise.
10125 * domwalk.h (dom_walker): Likewise.
10126 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
10127 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
10128 * hash-table.h (hash_table): Likewise.
10129 * machmode.h (bit_field_mode_iterator): Likewise.
10130 * pass_manager.h (pass_list): Likewise.
10131 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
10132 * tree-pass.h (pass_data): Likewise.
10133 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
10134 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
10135 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
10136 * asan.c (pass_data_asan): Update accordingly.
10137 * cfganal.c (control_dependences::find_control_dependence): Likewise.
10138 (control_dependences::control_dependences): Likewise.
10139 (control_dependences::~control_dependences): Likewise.
10140 (control_dependences::~control_dependences): Likewise.
10141 (control_dependences::get_edges_dependent_on): Likewise.
10142 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
10143 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
10144 * context.c (gcc::context::context): Likewise.
10145 * cprop.c (pass_rtl_cprop::clone): Likewise.
10146 * domwalk.c (dom_walker::walk): Likewise.
10147 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
10148 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
10149 * mode-switching.c (pass_mode_switching::clone): Likewise.
10150 * passes.c (opt_pass::opt_pass): Likewise.
10151 (pass_manager::pass_manager): Likewise.
10152 * predict.c (pass_strip_predict_hints::clone): Likewise.
10153 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
10154 (pass_split_all_insns::clone): Likewise.
10155 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
10157 (bit_field_mode_iterator::next_mode): Likewise.
10158 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
10159 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
10160 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
10161 * tree-complex.c (pass_lower_complex::clone): Likewise.
10162 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
10163 * tree-object-size.c (pass_object_sizes::clone): Likewise.
10164 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
10165 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
10166 (pass_fold_builtins::clone): Likewise.
10167 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
10168 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
10169 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
10170 (pass_cd_dce::clone): Likewise.
10171 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
10172 (pass_phi_only_cprop::clone): Likewise.
10173 * tree-ssa-dse.c (pass_dse::clone): Likewise.
10174 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
10175 * tree-ssa-loop.c (pass_lim::clone): Likewise.
10176 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
10177 * tree-ssa-pre.c (pass_fre::clone): Likewise.
10178 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
10179 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
10180 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
10181 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
10182 * tree-vrp.c (pass_vrp::clone): Likewise.
10183 * tsan.c (pass_tsan::clone): Likewise.
10185 2013-09-30 Jakub Jelinek <jakub@redhat.com>
10187 PR middle-end/58564
10188 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
10189 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
10191 PR middle-end/58564
10192 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
10193 optimization, punt if sign_bit_p looked through any zero extension.
10195 2013-09-30 Teresa Johnson <tejohnson@google.com>
10197 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
10198 Update redirected out edge count in joiner case.
10199 (ssa_redirect_edges): Common the joiner and non-joiner cases
10200 so that joiner case gets profile updates.
10202 2013-09-30 Richard Biener <rguenther@suse.de>
10204 PR tree-optimization/58554
10205 * tree-loop-distribution.c (classify_partition): Require
10206 unconditionally executed stores for memcpy and memset recognition.
10207 (tree_loop_distribution): Calculate dominance info.
10209 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
10211 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
10212 (NO_PROFILE_COUNTERS): Likewise.
10213 (PROFILE_HOOK): Likewise.
10214 (FUNCTION_PROFILER): Likewise.
10215 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
10217 2013-09-30 Iain Sandoe <iain@codesourcery.com>
10219 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
10220 calls and defines in TARGET_MACHO conditional.
10221 (load_macho_picbase_di): Likewise.
10222 (reload_macho_picbase): Likewise.
10223 (reload_macho_picbase_si): Likewise.
10224 (reload_macho_picbase_di): Likewise.
10225 (nonlocal_goto_receiver): Likewise.
10227 2013-09-30 Nick Clifton <nickc@redhat.com>
10229 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
10230 that use the MSP430X ISA.
10231 (msp430_option_override): Scan -mmcu command line option for any
10232 MCU name that supports the MSP430X ISA.
10233 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
10234 -mmcu options which enable the MSP430X ISA.
10236 2013-09-30 Richard Biener <rguenther@suse.de>
10238 PR middle-end/58532
10239 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
10240 before looking for setjmp-like calls.
10242 2013-09-29 Iain Sandoe <iain@codesourcery.com>
10245 * config/darwin-protos.h (machopic_get_function_picbase): New.
10246 * config/darwin.c (machopic_get_function_picbase): New.
10247 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
10248 label for a new func. (load_macho_picbase_di): Likewise.
10249 (reload_macho_picbase): New expand.
10250 (reload_macho_picbase_si): New insn.
10251 (reload_macho_picbase_di): New insn.
10252 (nonlocal_goto_receiver): New define and split.
10253 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
10254 (unspecv enum): Add UNSPECV_NLGR.
10256 2013-09-29 Iain Sandoe <iain@codesourcery.com>
10258 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
10259 that altivec registers are correctly sized on Darwin.
10261 2013-09-29 Iain Sandoe <iain@codesourcery.com>
10263 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
10264 darwin-driver.o): Use COMPILE and POSTCOMPILE.
10265 * config/x-darwin (host-darwin.o): Likewise.
10266 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
10267 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
10268 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
10270 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
10272 * doc/invoke.texi: Fix usage of @tie{} command.
10274 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
10276 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
10278 2013-09-28 Jan Hubicka <jh@suse.cz>
10280 * config/i386/x86-tune.def: Add documentation for each of the options;
10283 2013-09-28 Jan Hubicka <jh@suse.cz>
10285 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
10287 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
10288 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
10289 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
10290 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
10291 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
10293 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
10295 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
10296 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
10297 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
10298 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
10299 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
10300 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
10301 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
10302 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
10303 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
10304 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
10305 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
10306 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
10307 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
10308 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
10309 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
10310 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
10311 gimple.h, godump.c, graphite-clast-to-gimple.c,
10312 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
10313 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
10314 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
10315 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
10316 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
10317 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
10318 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
10319 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
10320 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
10321 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
10322 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
10323 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
10324 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
10325 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
10326 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
10327 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
10328 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
10329 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
10330 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
10331 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
10332 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
10333 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
10334 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
10335 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
10336 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
10337 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
10338 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
10339 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
10340 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
10341 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
10342 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
10343 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
10344 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
10345 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
10346 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
10347 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
10348 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
10349 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
10350 whitespace before "(".
10352 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
10354 * expr.h (extract_bit_field): Remove packedp parameter.
10355 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
10356 from forward declaration.
10357 (store_split_bit_field): Remove packedp arg from calls to
10358 extract_fixed_bit_field.
10359 (extract_bit_field_1): Remove packedp parameter and packedp
10360 argument from recursive calls and calls to extract_fixed_bit_field.
10361 (extract_bit_field): Remove packedp parameter and corresponding
10362 arg to extract_bit_field_1.
10363 (extract_fixed_bit_field): Remove packedp parameter. Remove code
10365 (extract_split_bit_field): Remove packedp arg from call to
10366 extract_fixed_bit_field.
10367 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
10368 (copy_blkmode_from_reg): Likewise.
10369 (copy_blkmode_to_reg): Likewise.
10370 (read_complex_part): Likewise.
10371 (store_field): Likewise.
10372 (expand_expr_real_1): Likewise.
10373 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
10374 to extract_bit_field.
10375 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
10376 call to extract_bit_field.
10377 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
10378 call to extract_bit_field.
10379 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
10380 and special packedp behavior from -fstrict-volatile-bitfields
10383 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10385 * lra-eliminations.c (init_elim_table): Guard value_p.
10387 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
10389 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
10390 DFmode, DImode, and SFmode in the upper VSX registers based on the
10391 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
10392 if -mpower8-vector. Combine -mvsx-timode handling with the rest
10393 of the VSX register handling.
10395 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
10396 (f32_sv): Likewise.
10397 (zero_extendsidi2_lfiwzx): Add support for loading into the
10398 Altivec registers with -mpower8-vector. Use wu/wv constraints to
10399 only do VSX memory options on Altivec registers.
10400 (extendsidi2_lfiwax): Likewise.
10401 (extendsfdf2_fpr): Likewise.
10402 (mov<mode>_hardfloat, SF/SD modes): Likewise.
10403 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
10404 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
10405 (movdi_internal64): Likewise.
10407 2013-09-27 Xinliang David Li <davidxl@google.com>
10409 * opts.c (finish_options): Adjust parameters
10410 according to vect cost model.
10411 (common_handle_option): Set dynamic vect cost
10413 targhooks.c (default_add_stmt_cost): Compute stmt cost
10415 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
10416 Use helper function.
10417 * tree-vectorizer.h (unlimited_cost_model): New function.
10418 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
10419 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
10421 (vect_enhance_data_refs_alignment): Ditto.
10422 * flag-types.h: New enum.
10423 * common/config/i386/i386-common.c (ix86_option_init_struct):
10424 No need to initialize vect_cost_model flag.
10425 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
10428 2013-09-27 Diego Novillo <dnovillo@google.com>
10430 * gimple.h (enum ssa_mode): Remove.
10432 2013-09-27 Paulo Matos <pmatos@broadcom.com>
10434 * cfgloop.h (number_of_loops): Fix typo in check for null.
10436 2013-09-27 Jakub Jelinek <jakub@redhat.com>
10438 PR middle-end/58551
10439 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
10440 are children of outermost saved_cfun's loop, and set it up to
10441 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
10442 if loop != loop0 and SESE region contains bbs that belong to loop0.
10444 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
10446 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
10447 (binary_scale_code_p, get_base_term, get_index_term): New functions.
10448 (set_address_segment, set_address_base, set_address_index)
10449 (set_address_disp): Accept the argument unconditionally.
10450 (baseness): Remove must_be_base_p and must_be_index_p checks.
10451 (decompose_normal_address): Classify as much as possible in the
10454 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
10456 * cse.c (count_reg_usage): Handle INT_LIST.
10457 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
10458 * reginfo.c (reg_scan_mark_refs): Likewise.
10459 * reload1.c (eliminate_regs_1): Likewise.
10461 2013-09-27 Iain Sandoe <iain@codesourcery.com>
10463 PR middle-end/58547
10464 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
10467 2013-09-27 Richard Biener <rguenther@suse.de>
10469 PR tree-optimization/58459
10470 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
10471 restriction not propagating into loops.
10473 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
10475 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
10476 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
10477 * doc/tree-ssa.texi (Walking use-def chains): Delete.
10479 2013-09-26 Richard Biener <rguenther@suse.de>
10481 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
10483 2013-09-26 Richard Biener <rguenther@suse.de>
10485 * alias.h (component_uses_parent_alias_set): Rename to ...
10486 (component_uses_parent_alias_set_from): ... this.
10487 * alias.c (component_uses_parent_alias_set): Rename to ...
10488 (component_uses_parent_alias_set_from): ... this and return
10489 the desired parent.
10490 (reference_alias_ptr_type_1): Use the result from
10491 component_uses_parent_alias_set_from instead of stripping
10492 components one at a time.
10493 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
10495 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
10497 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
10498 Move prototypes to...
10499 * tree-ssa-ter.h: New File. Move prototypes here.
10500 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
10501 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
10502 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
10503 from is_replaceable_p.
10504 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
10505 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
10506 (process_replaceable): Use ter_is_replaceable_p.
10507 (find_replaceable_in_bb): Use ter_is_replaceable_p.
10508 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
10509 newly refactored ssa_is_replaceable_p.
10510 * cfgexpand.c: Include tree-outof-ssa.h.
10511 * ssaexpand.h: Delete.
10513 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
10515 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
10516 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
10518 (create_gimple_tmp): Delete.
10519 (get_expr_type, build_assign, build_type_cast): Move to...
10520 * gimple-builder.c: New File.
10521 (get_expr_type): Relocate from gimple.c.
10522 (build_assign, build_type_cast): Change to only create ssanames.
10523 * gimple.h: Move prototypes to...
10524 * gimple-builder.h: New File. Here.
10525 * tree-ssa.h: And here.
10526 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
10527 count_uses_and_derefs): Relocate from gimple.c.
10528 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
10529 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
10530 * tree-ssa-math-opts (execute_cse_reciprocals): Use
10531 gimple_replace_ssa_lhs.
10532 * asan.c: Include gimple-builder.h.
10533 * Makefile.in: Add gimple-builder.o.
10535 2013-09-26 Richard Biener <rguenther@suse.de>
10537 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
10538 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
10539 (loe_visit_block): Use gcc_checking_assert.
10540 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
10541 gimple_assign_ssa_name_copy_p.
10542 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
10545 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
10547 * config/rs6000/t-rs6000 (rs6000.o): Remove.
10548 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
10550 2013-09-26 Richard Biener <rguenther@suse.de>
10552 PR tree-optimization/58539
10553 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
10554 the fact that debug statements are not taking part in loop-closed
10557 2013-09-26 Nick Clifton <nickc@redhat.com>
10559 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
10560 time warning message.
10561 (msp430_print_operand_raw): Delete unused letter parameter.
10562 (TARGET_PRINT_OPERAND_ADDRESS): Define.
10563 (msp430_print_operand_address): New function.
10564 (msp430_print_operand): Move address printing code from here to
10566 * config/msp430/msp430.md (movsipsi2): Add comment in generated
10568 (zero_extendpsisi2): Likewise.
10569 (extendpsisi2): New pattern.
10570 (andneghi3): New pattern.
10572 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
10574 * config/aarch64/aarch64.opt (mlra): New option.
10575 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
10576 (TARGET_LRA_P): Define.
10578 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
10580 * expr.c (expand_assignment): Remove obsolete comment.
10582 2013-09-25 Jeff Law <law@redhat.com>
10584 * tree-flow.h (thread_through_all_blocks): Prototype moved into
10585 tree-ssa-threadupdate.h.
10586 (register_jump_thread): Similarly.
10587 * tree-ssa-threadupdate.h: New header file.
10588 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
10589 * tree-vrp.c: Likewise.
10590 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
10591 (thread_around_empty_blocks): Change type of path vector argument to
10592 an edge,type pair from just an edge. Initialize both elements when
10593 appending to a jump threading path. Tweak references to elements
10595 (thread_across_edge): Similarly. Release memory for the elements
10597 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
10598 (dump_jump_thread_path): New function broken out from
10599 register_jump_thread.
10600 (register_jump_thread): Use dump_jump_thread_path. Change type of
10601 path vector entries. Search the path for NULL edges and dump
10602 the path if one is found. Tweak the conversion of path to 3-edge
10603 form to use the block copy type information embedded in the path.
10605 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
10607 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
10609 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
10610 Vladimir Makarov <vmakarov@redhat.com>
10612 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
10613 from the least significant bit.
10614 (strip_address_mutations): Add bitfield operations handling.
10615 (must_be_index_p): Add shifting and rotate operations handling.
10616 (set_address_base): Use must_be_base_p predicate.
10617 (set_address_index): Use must_be_index_p predicate.
10619 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
10620 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
10621 Sergey Lega <sergey.s.lega@intel.com>
10622 Anna Tikhonova <anna.tikhonova@intel.com>
10623 Ilya Tocar <ilya.tocar@intel.com>
10624 Andrey Turetskiy <andrey.turetskiy@intel.com>
10625 Ilya Verbin <ilya.verbin@intel.com>
10626 Kirill Yukhin <kirill.yukhin@intel.com>
10627 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
10629 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
10631 (ix86_expand_vector_move_misalign): Support new unaligned load and
10632 stores and use new names.
10633 (CODE_FOR_sse2_storedqu): Rename to ...
10634 (CODE_FOR_sse2_storedquv16qi): ... this.
10635 (CODE_FOR_sse2_loaddqu): Rename to ...
10636 (CODE_FOR_sse2_loaddquv16qi): ... this.
10637 (CODE_FOR_avx_loaddqu256): Rename to ...
10638 (CODE_FOR_avx_loaddquv32qi): ... this.
10639 (CODE_FOR_avx_storedqu256): Rename to ...
10640 (CODE_FOR_avx_storedquv32qi): ... this.
10641 * config/i386/i386.md (fpint_logic): New.
10642 * config/i386/sse.md (VMOVE): Extend for AVX512.
10646 (VI_UNALIGNED_LOADSTORE): Ditto.
10647 (sse2_avx_avx512f): Ditto.
10648 (sse2_avx2): Extend for AVX512.
10649 (sse4_1_avx2): Ditto.
10650 (avx2_avx512f): New.
10651 (sse): Extend for AVX512.
10654 (avxsizesuffix): Ditto.
10655 (sseintvecmode): Ditto.
10656 (ssePSmode): Ditto.
10657 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
10658 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
10659 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
10660 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
10661 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
10662 (<sse2_avx_avx512f>_storedqu<mode): ... this.
10663 (<sse>_movnt<mode>): Replace constraint "x" with "v".
10664 (STORENT_MODE): Extend for AVX512.
10665 (*absneg<mode>2): Replace constraint "x" with "v".
10666 (*mul<mode>3): Ditto.
10667 (*ieee_smin<mode>3): Ditto.
10668 (*ieee_smax<mode>3): Ditto.
10669 (avx_cmp<mode>3): Replace VF with VF_128_256.
10670 (*<sse>_maskcmp<mode>3_comm): Ditto.
10671 (<sse>_maskcmp<mode>3): Ditto.
10672 (<sse>_andnot<mode>3): Extend for AVX512.
10673 (<code><mode>3, anylogic): Replace VF with VF_128_256.
10674 (<code><mode>3, fpint_logic): New.
10675 (*<code><mode>3): Extend for AVX512.
10676 (avx512flogicsuff): New.
10677 (avx512f_<logic><mode>): Ditto.
10678 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
10680 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
10681 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
10682 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
10683 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
10684 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
10685 (xop_vpermil2<mode>3): Ditto.
10686 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
10687 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
10688 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
10689 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
10691 2013-09-25 Tom Tromey <tromey@redhat.com>
10693 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
10694 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
10695 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
10696 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
10697 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
10698 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
10699 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
10700 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
10701 (GRAPHITE_HTAB_H): Remove.
10703 2013-09-25 Tom Tromey <tromey@redhat.com>
10705 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
10707 2013-09-25 Tom Tromey <tromey@redhat.com>
10709 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
10711 2013-09-25 Tom Tromey <tromey@redhat.com>
10713 * config/i386/t-i386 (i386.o): Remove.
10714 (i386-c.o): Use COMPILE and POSTCOMPILE.
10716 2013-09-25 Tom Tromey <tromey@redhat.com>
10718 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
10720 2013-09-25 Tom Tromey <tromey@redhat.com>
10722 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
10723 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
10724 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
10725 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
10726 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
10727 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
10728 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
10729 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
10730 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
10731 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
10732 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
10733 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
10734 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
10735 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
10736 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
10737 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
10738 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
10739 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
10740 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
10741 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
10742 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
10743 (tree-ssa-pre.o, tree-ssa-sccvn.o)
10744 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
10745 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
10746 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
10747 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
10748 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
10749 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
10750 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
10751 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
10752 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
10753 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
10754 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
10755 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
10756 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
10757 (graphite-clast-to-gimple.o, graphite-dependences.o)
10758 (graphite-interchange.o, graphite-poly.o)
10759 (graphite-scop-detection.o, graphite-sese-to-poly.o)
10760 (graphite-optimize-isl.o, tree-vect-loop.o)
10761 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
10762 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
10763 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
10764 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
10765 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
10766 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
10767 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
10768 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
10769 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
10770 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
10771 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
10772 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
10773 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
10774 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
10775 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
10776 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
10777 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
10778 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
10779 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
10780 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
10781 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
10782 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
10783 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
10784 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
10785 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
10786 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
10787 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
10788 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
10789 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
10790 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
10791 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
10792 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
10793 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
10794 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
10795 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
10796 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
10797 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
10798 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
10799 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
10800 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
10801 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
10802 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
10803 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
10804 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
10805 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
10806 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
10807 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
10808 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
10809 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
10810 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
10811 (gcov-dump.o): Remove.
10812 (default-c.o): Use COMPILE and POSTCOMPILE.
10813 (CFLAGS-gcc.o): New variable.
10814 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
10816 2013-09-25 Tom Tromey <tromey@redhat.com>
10818 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
10819 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
10820 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
10821 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
10822 (c-family/c-pragma.o, c-family/c-pretty-print.o)
10823 (c-family/c-semantics.o, c-family/c-ada-spec.o)
10824 (c-family/array-notation-common.o, c-family/stub-objc.o)
10825 (c-family/c-ubsan.o): Remove.
10827 2013-09-25 Tom Tromey <tromey@redhat.com>
10829 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
10831 2013-09-25 Tom Tromey <tromey@redhat.com>
10833 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
10834 to add -DENABLE_SHARED_LIBGCC.
10835 (gcc.o): Don't use subshell.
10837 2013-09-25 Tom Tromey <tromey@redhat.com>
10839 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
10840 * configure.ac: Don't invoke AM_PROG_CC_C_O.
10841 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
10842 * configure, config.in: Rebuild.
10844 2013-09-25 Tom Tromey <tromey@redhat.com>
10846 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
10847 (COMPILE, POSTCOMPILE): New variables.
10848 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
10849 (DEPFILES): New variable.
10850 Include ".Po" files.
10851 * configure.ac: Add checks for dependency checking.
10852 * configure, aclocal.m4: Regenerate.
10854 2013-09-25 Tom Tromey <tromey@redhat.com>
10856 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
10857 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
10859 2013-09-25 Tom Tromey <tromey@redhat.com>
10861 * Makefile.in (generated_files): Add options.h,
10862 target-hooks-def.h, insn-opinit.h,
10863 common/common-target-hooks-def.h, pass-instances.def,
10864 c-family/c-target-hooks-def.h.
10866 2013-09-25 Jeff Law <law@redhat.com>
10868 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
10869 than foo[foo.length () - 1] to access last member in a vec.
10870 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
10872 2013-09-25 Richard Biener <rguenther@suse.de>
10874 PR middle-end/58521
10875 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
10877 2013-09-25 Jan Hubicka <jh@suse.cz>
10879 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
10882 2013-09-25 Marek Polacek <polacek@redhat.com>
10885 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
10886 when determining the type name.
10888 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
10890 * config/sh/sh.md: Fix formatting.
10892 2013-09-24 Xinliang David Li <davidxl@google.com>
10894 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
10895 max peel iterations parameter.
10896 * param.def: New parameter.
10897 * doc/invoke.texi: Document New parameter.
10899 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
10901 * gimple-pretty-print.c: Various whitespace tweaks.
10902 * tree-core.h: Likewise.
10903 * tree-pretty-print.c: Likewise.
10904 * tree-ssa-alias.c: Likewise.
10905 * tree-ssa-copy.c: Likewise.
10906 * tree-ssanames.c: Likewise.
10907 * tree-ssanames.h: Likewise.
10908 * tree-vrp.c: Likewise.
10910 2013-09-24 Alan Modra <amodra@gmail.com>
10912 PR middle-end/57134
10913 PR middle-end/57586
10914 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
10915 for output operands that disallow regs. Don't use EXPAND_WRITE on
10918 2013-09-24 Richard Biener <rguenther@suse.de>
10920 PR middle-end/58513
10921 * tree.c (reference_alias_ptr_type): Move ...
10922 * alias.c (reference_alias_ptr_type): ... here and implement
10923 in terms of the new reference_alias_ptr_type_1.
10924 (ref_all_alias_ptr_type_p): New helper.
10925 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
10926 use ref_all_alias_ptr_type_p.
10927 (get_deref_alias_set): Add flag_strict_aliasing check here.
10928 (reference_alias_ptr_type_1): New function, split out from ...
10929 (get_alias_set): ... here.
10930 (alias_ptr_types_compatible_p): New function.
10931 * alias.h (reference_alias_ptr_type): Declare.
10932 (alias_ptr_types_compatible_p): Likewise.
10933 * tree.h (reference_alias_ptr_type): Remove.
10934 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
10935 to compare MEM_REF alias types.
10937 2013-09-24 Richard Biener <rguenther@suse.de>
10939 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
10941 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
10943 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
10944 reload helper function arrays into a single array reg_addr.
10945 (reload_fpr_gpr): Likewise.
10946 (reload_gpr_vsx): Likewise.
10947 (reload_vsx_gpr): Likewise.
10948 (struct rs6000_reg_addr): Likewise.
10949 (reg_addr): Likewise.
10950 (rs6000_debug_reg_global): Change rs6000_vector_reload,
10951 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
10952 (rs6000_init_hard_regno_mode_ok): Likewise.
10953 (rs6000_secondary_reload_direct_move): Likewise.
10954 (rs6000_secondary_reload): Likewise.
10956 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
10957 constraints: wu, ww, and wy. Repurpose wv constraint added during
10958 power8 changes. Put wg constraint in alphabetical order.
10960 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
10961 for future work to add ISA 2.07 VSX single precision support.
10962 (-mvsx-scalar-double): Change default from -1 to 1, update
10963 documentation comment.
10964 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
10965 (-mupper-regs-df): New debug switch to control whether DF values
10966 can go in the traditional Altivec registers.
10967 (-mupper-regs-sf): New debug switch to control whether SF values
10968 can go in the traditional Altivec registers.
10970 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
10971 and wy constraints.
10972 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
10973 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
10974 Add new constraints, wu/ww/wy. Repurpose wv constraint.
10975 (rs6000_debug_legitimate_address_p): Print if we are running
10976 before, during, or after reload.
10977 (rs6000_secondary_reload): Add a comment.
10978 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
10980 * config/rs6000/constraints.md (wa constraint): Sort w<x>
10981 constraints. Update documentation string.
10982 (wd constraint): Likewise.
10983 (wf constraint): Likewise.
10984 (wg constraint): Likewise.
10985 (wn constraint): Likewise.
10986 (ws constraint): Likewise.
10987 (wt constraint): Likewise.
10988 (wx constraint): Likewise.
10989 (wz constraint): Likewise.
10990 (wu constraint): New constraint for ISA 2.07 SFmode scalar
10992 (ww constraint): Likewise.
10993 (wy constraint): Likewise.
10994 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
10995 in the previous submissions.
10996 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
10998 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
11000 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
11001 (REG_BR_PROB): Say that the probability is stored in an int_list.
11002 * reg-notes.def: Update commentary to mention INT_LIST.
11003 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
11004 (INT_LIST): New rtx.
11005 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
11006 * rtlanal.c (int_reg_note_p): New function.
11007 (alloc_reg_note): Assert that the note does not have an int argument.
11008 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
11009 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
11010 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
11011 rather than an INSN_LIST. Handle INT_LIST.
11012 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
11013 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
11014 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
11015 Manipulate them as ints rather than rtxes.
11016 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
11017 * regmove.c (copy_src_to_dest): Likewise.
11018 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
11020 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
11021 into the cases that need it.
11022 * config/arm/arm.c (arm_unwind_emit): Likewise.
11024 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
11025 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
11026 * loop-doloop.c (add_test, doloop_modify): Likewise.
11027 * loop-unswitch.c (compare_and_jump_seq): Likewise.
11028 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
11029 * predict.c (combine_predictions_for_insn): Likewise.
11030 * print-rtl.c (print_rtx): Handle INT_LIST.
11031 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
11032 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
11033 * config/arm/arm.c (emit_unlikely_jump): Likewise.
11034 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
11035 (ix86_split_fp_branch, predict_jump): Likewise.
11036 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
11037 * config/sh/sh.c (expand_cbranchsi4): Likewise.
11038 * config/spu/spu.c (ea_load_store_inline): Likewise.
11040 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
11041 value of a REG_BR_PROB note.
11042 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
11043 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
11044 * emit-rtl.c (try_split): Likewise.
11045 * predict.c (br_prob_note_reliable_p): Likewise.
11046 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
11047 * reorg.c (mostly_true_jump): Likewise.
11048 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
11049 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
11050 * config/i386/i386.c (ix86_print_operand): Likewise.
11051 * config/ia64/ia64.c (ia64_print_operand): Likewise.
11052 * config/mmix/mmix.c (mmix_print_operand): Likewise.
11053 * config/rs6000/rs6000.c (output_cbranch): Likewise.
11054 * config/s390/s390.c (s390_expand_tbegin): Likewise.
11055 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
11056 * config/sparc/sparc.c (output_cbranch): Likewise.
11057 * config/spu/spu.c (get_branch_target): Likewise.
11058 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
11059 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
11061 2013-09-23 Jan Hubicka <jh@suse.cz>
11063 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
11065 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
11066 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
11068 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
11069 (gimple_fold_call): Dump inconsistent devirtualizations; add
11070 sanity check for type based devirtualizations.
11071 * ipa-prop.c: Include ipa-utils.h
11072 (ipa_intraprocedural_devirtualization): Add sanity check.
11073 (try_make_edge_direct_virtual_call): Likewise.
11075 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
11077 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
11078 assignment statements.
11080 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
11082 * gimple-pretty-print.c (dump_ssaname_info): New function.
11083 (dump_gimple_phi): Call it.
11084 (pp_gimple_stmt_1): Likewise.
11085 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
11086 (range_info_def): Declare.
11087 * tree-pretty-print.c (pp_double_int): New function.
11088 (dump_generic_node): Call it.
11089 * tree-pretty-print.h (pp_double_int): Declare.
11090 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
11091 * tree-ssanames.h (range_info_def): New structure.
11092 (value_range_type): Move definition here.
11093 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
11095 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
11097 (set_range_info): New function.
11098 (get_range_info): Likewise.
11099 (duplicate_ssa_name_range_info): Likewise.
11100 (duplicate_ssa_name_fn): Check pointer type and call
11101 duplicate_ssa_name_range_info.
11102 * tree-ssa-copy.c (fini_copy_prop): Likewise.
11103 * tree-vrp.c (value_range_type): Remove definition, now in
11105 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
11106 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
11107 (SSA_NAME_RANGE_INFO): New macro.
11109 2013-09-23 Richard Biener <rguenther@suse.de>
11111 PR tree-optimization/58464
11112 * tree-ssa-pre.c (phi_trans_lookup): Remove.
11113 (phi_trans_add): Change to add conditionally on being not present.
11114 (phi_translate_1): Remove recursion detection here.
11115 (phi_translate): Pre-seed the cache with NULL to catch
11116 recursion here in a more generic way.
11117 (bitmap_find_leader): Adjust comment.
11118 (get_representative_for): Dump value-numbers.
11119 (create_expression_by_pieces): Likewise.
11120 (insert_into_preds_of_block): Likewise.
11122 2013-09-23 Christian Bruel <christian.bruel@st.com>
11125 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
11126 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
11128 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
11131 2013-09-20 Renlin Li <renlin.li@arm.com>
11133 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
11135 (aarch64_expand_epilogue): Likewise.
11136 (aarch64_legitimize_reload_address): Likewise.
11138 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
11140 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
11141 NULL_TREE before pushing them onto the vector. Likewise for labels.
11143 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
11145 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
11146 comparison operators when -fno-trapping-math is in effect.
11147 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
11148 comparison operators in TFmode and assert that unsupported operators
11150 (ia64_print_operand): Likewise.
11152 2013-09-21 Jan Hubicka <jh@suse.cz>
11154 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
11155 (sse_typeless_stores): Enable for core
11156 (sse_load0_by_pxor): Likewise.
11157 (four_jump_limit): Disable for core.
11158 (pad_returns): Likewise.
11159 (avoid_vector_decode): Likewise.
11160 (fuse_cmp_and_branch): Enable for cores.
11161 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
11163 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
11165 PR middle-end/56791
11166 * config/pa/pa.c (pa_option_override): Disable auto increment and
11167 decrement instructions until reload is completed.
11169 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
11170 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
11171 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
11173 2013-09-20 DJ Delorie <dj@redhat.com>
11174 Nick Clifton <nickc@redhat.com>
11176 * config/rl78/rl78.c: Various whitespace and comment tweaks.
11177 (need_to_save): Save bank 0 on interrupts.
11178 (characterize_address): Strip far address wrappers.
11179 (rl78_as_legitimate_address): Likewise.
11180 (transcode_memory_rtx): Likewise.
11181 (rl78_peep_movhi_p): Disable this peephole after devirt.
11182 (rl78_propogate_register_origins): Forget all origins when a
11184 * config/rl78/rl78-virt.md: Various whitespace tweaks.
11185 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
11187 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
11188 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
11189 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
11191 * config/rl78/constraints.md: Various whitespace and paren tweaks.
11193 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
11195 * config/pa/pa.md: In "scc" insn patterns, change output template to
11196 handle const0_rtx in reg_or_0_operand operands.
11198 2013-09-20 Martin Husemann <martin@NetBSD.org>
11201 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
11202 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
11204 2013-09-20 Richard Biener <rguenther@suse.de>
11206 PR middle-end/58484
11207 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
11208 remembering SSA name version and block index.
11209 (new_scev_info_str): Adjust.
11210 (hash_scev_info): Likewise. Also hash the block index.
11211 (eq_scev_info): Adjust.
11212 (find_var_scev_info): Likewise.
11213 (struct instantiate_cache_entry): Remove.
11214 (struct instantiate_cache_type): Use a htab to map name, block
11216 (instantiate_cache_type::~instantiate_cache_type): Adjust.
11217 (get_instantiated_value_entry): Likewise.
11218 (hash_idx_scev_info, eq_idx_scev_info): New functions.
11219 (instantiate_scev_name): Adjust.
11221 2013-09-20 Jeff Law <law@redhat.com>
11223 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
11225 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
11227 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
11228 Call aarch64_simd_expand_args to update op[argc].
11230 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
11232 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
11235 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
11237 * gengtype.c (file_rules): Added rule for *.cc files.
11238 (get_output_file_with_visibility): Give fatal message when no
11241 2013-09-20 Renlin Li <renlin.li@arm.com>
11243 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
11244 (aarch64_expand_epilogue): Likewise.
11245 (aarch64_legitimize_reload_address): Likewise.
11247 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
11249 PR middle-end/57748
11250 * expr.c (expand_assignment): Remove misalignp code path.
11252 2013-09-20 Marek Polacek <polacek@redhat.com>
11255 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
11256 TYPE_PRECISION. Add asserts.
11258 2013-09-20 Richard Biener <rguenther@suse.de>
11260 PR tree-optimization/58453
11261 * tree-loop-distribution.c (distribute_loop): Apply the cost
11262 model for -ftree-loop-distribute-patterns, too.
11264 2013-09-20 Richard Biener <rguenther@suse.de>
11266 PR middle-end/58473
11267 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
11268 make type comparison less strict.
11270 2013-09-20 Alan Modra <amodra@gmail.com>
11272 * configure: Regenerate.
11273 * aclocal.m4: Regenerate.
11275 2013-09-20 Marek Polacek <polacek@redhat.com>
11278 * doc/extend.texi: Document that attribute used is meant to be used
11279 on variables with static storage duration.
11281 2013-09-19 Jakub Jelinek <jakub@redhat.com>
11283 PR tree-optimization/58472
11284 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
11285 simd_lane_access set inv_p = false.
11286 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
11287 the simduid magic VAR_DECL.
11289 2013-09-19 Jan Hubicka <jh@suse.cz>
11291 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
11293 2013-09-17 Jeff Law <law@redhat.com>
11295 * tree-ssa-dom.c (record_temporary_equivalences): New function
11296 split out of dom_opt_dom_walker::after_dom_children.
11297 (dom_opt_dom_walker::thread_across_edge): Move common code
11298 in here from dom_opt_dom_walker::after_dom_children.
11299 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
11301 2013-09-19 Jan Hubicka <jh@suse.cz>
11303 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
11304 (TARGET_GENERIC): Use PROCESOR_GENERIC
11305 (enum processor_type): Unify generic32 and 64.
11306 * i386.md (cpu): Likewise.
11307 * x86-tune.def (use_leave): Enable for generic32.
11308 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
11309 * athlon.md: Change generic64 to generic in all occurences.
11310 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
11311 (ix86_target_macros_internal): Likewise.
11312 * driver-i386.c (host_detect_local_cpu): Likewise.
11313 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
11315 (generic_memcpy, generic_memset, generic_cost): This one.
11316 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
11317 (m_GENERIC32, m_GENERIC64): Remove.
11318 (m_GENERIC): Turn into one flag.
11319 (processor_target): Unify generic tunnings.
11320 (ix86_option_override_internal): Replace generic32/64 by generic.
11321 (ix86_issue_rate): Likewise.
11322 (ix86_adjust_cost): Likewise.
11324 2013-09-19 Jan Hubicka <jh@suse.cz>
11326 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
11327 of speculative flag.
11329 2013-09-19 Jakub Jelinek <jakub@redhat.com>
11331 * omp-low.c (expand_omp_sections): Always pass len - 1 to
11332 GOMP_sections_start, even if !exit_reachable.
11334 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
11336 * lra-constraints.c (need_for_all_save_p): Use macro
11337 HARD_REGNO_CALL_PART_CLOBBERED.
11338 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
11339 set up pseudo conflict hard regs.
11341 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
11344 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
11347 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
11349 PR rtl-optimization/58438
11350 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
11351 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
11352 from previous subpasses.
11354 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
11356 * arm.c (arm_get_frame_offsets): Validate architecture supports
11357 LDRD/STRD before accepting the tuning preference.
11358 (arm_expand_prologue): Likewise.
11359 (arm_expand_epilogue): Likewise.
11361 2013-09-18 Richard Biener <rguenther@suse.de>
11363 PR tree-optimization/58417
11364 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
11365 have chrecs with symbols defined in the loop as operands.
11366 (chrec_fold_multiply): Likewise.
11367 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
11368 parameters before folding binary operations.
11369 (struct instantiate_cache_entry_hasher): Remove.
11370 (struct instantiate_cache_type): Use a pointer-map.
11371 (instantiate_cache_type::instantiate_cache_type): New function.
11372 (instantiate_cache_type::get): Likewise.
11373 (instantiate_cache_type::set): Likewise.
11374 (instantiate_cache_type::~instantiate_cache_type): Adjust.
11375 (get_instantiated_value_entry): Likewise.
11376 (global_cache): New global.
11377 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
11378 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
11379 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
11380 (instantiate_scev_name): Adjust.
11381 (instantiate_scev): Construct global instead of local cache.
11382 (resolve_mixers): Likewise.
11384 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
11385 Paolo Carlini <paolo.carlini@oracle.com>
11388 * doc/implement-cxx.texi: Fix references to the C++ standards.
11390 2013-09-18 Jakub Jelinek <jakub@redhat.com>
11392 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
11393 * tree-vect-data-refs.c (vect_analyze_data_refs): For
11394 simd_lane_access drs, update also DR_ALIGNED_TO.
11396 2013-09-18 Marek Polacek <polacek@redhat.com>
11399 * doc/extend.texi: Document no_sanitize_undefined attribute.
11400 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
11401 no_sanitize_undefined attribute.
11403 2013-09-18 Nick Clifton <nickc@redhat.com>
11405 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
11406 (ASM_DECLARE_FUNCTION_NAME): Define.
11408 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
11410 * compare-elim.c (find_comparison_dom_walker): New class
11411 (find_comparisons_in_bb): Rename to
11412 find_comparison_dom_walker::before_dom_children
11413 (find_comparisons): Adjust
11414 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
11416 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
11417 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
11418 (init_walk_dominator_tree): Remove declaration.
11419 (fini_walk_dominator_tree): Remove declaration.
11420 * fwprop.c (single_def_use_dom_walker): New class
11421 (single_def_use_enter_block): Convert to
11422 single_def_use_dom_walker::before_dom_children.
11423 (single_def_use_leave_block): Convert to
11424 single_def_use_dom_walker::after_dom_children.
11425 (build_single_def_use_links): Adjust.
11426 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
11428 (find_candidates_in_block): Convert to
11429 find_candidates_dom_walker::before_dom_children.
11430 (execute_strength_reduction): Adjust.
11431 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
11432 (sese_dom_walker): New class.
11433 (sese_dom_walker::sese_dom_walker): New constructor.
11434 (sese_dom_walker::~sese_dom_walker): New destructor.
11435 (build_sese_conditions_before): Convert to
11436 sese_dom_walker::before_dom_children.
11437 (build_sese_conditions_after): Convert to
11438 sese_dom_walker::after_dom_children.
11439 (build_poly_scop): Adjust
11440 * tree-into-ssa.c (rewrite_dom_walker): New class
11441 (rewrite_enter_block): Convert to
11442 rewrite_dom_walker::before_dom_children.
11443 (rewrite_leave_block): Convert to
11444 rewrite_dom_walker::after_dom_children.
11445 (rewrite_update_dom_walker): New class.
11446 (rewrite_update_enter_block): Convert to
11447 rewrite_update_dom_walker::before_dom_children.
11448 (rewrite_update_leave_block): Convert to
11449 rewrite_update_dom_walker::after_dom_children.
11450 (rewrite_blocks, rewrite_into_ssa): Adjust.
11451 (mark_def_dom_walker): New class.
11452 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
11453 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
11454 (mark_def_sites_blocks): Convert to
11455 mark_def_dom_walker::before_dom_children.
11456 (mark_def_site_blocks): Remove.
11457 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
11458 (tree_ssa_dominator_optimize): Adjust.
11459 (dom_thread_across_edge): Convert to method
11460 dom_opt_dom_walker::thread_across_edge.
11461 (dom_opt_enter_block): Convert to member function
11462 dom_opt_dom_walker::before_dom_children.
11463 (dom_opt_leave_block): Convert to member function
11464 dom_opt_dom_walker::after_dom_children.
11465 * tree-ssa-dse.c (dse_dom_walker): New class.
11466 (dse_enter_block): Convert to member function
11467 dse_dom_walker::before_dom_children.
11468 (tree_ssa_dse): Adjust.
11469 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
11470 (determine_invariantness_stmt): Convert to method
11471 invariantness_dom_walker::before_dom_children.
11472 (determine_invariantness): Remove
11473 (move_computations_dom_walker): New class.
11474 (move_computations_stmt): Convert to method
11475 move_computations_dom_walker::before_dom_children.
11476 (move_computations, tree_ssa_lim): Adjust.
11477 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
11478 (nt_init_block): Convert to method
11479 notrappping_dom_walker::before_dom_children.
11480 (nt_fini_block): Convert to method
11481 method nontrapping_dom_walker::after_dom_children.
11482 (get_non_trapping): Adjust.
11483 * tree-ssa-pre.c (eliminate_dom_walker): New class.
11484 (eliminate_bb): Convert to method
11485 eliminate_dom_walker::before_dom_children.
11486 (eliminate_leave_block): Convert to method
11487 eliminate_dom_walker::after_dom_children.
11488 (eliminate): Adjust.
11489 * tree-ssa-strlen.c (strlen_dom_walker): New class.
11490 (strlen_enter_block): Convert to method
11491 strlen_dom_walker::before_dom_children.
11492 (strlen_leave_block): Convert to method
11493 method strlen_dom_walker::after_dom_children.
11494 (tree_ssa_strlen): Adjust.
11495 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
11496 (tree_ssa_uncprop): Adjust.
11497 (uncprop_leave_block): Convert to method
11498 uncprop_dom_walker::after_dom_children.
11499 (uncprop_leave_block): Convert to method
11500 uncprop_dom_walker::before_dom_children.
11502 2013-09-18 Bin Cheng <bin.cheng@arm.com>
11504 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
11505 branch in same basic block. Check both src and dest of the move insn.
11507 2013-09-17 Nick Clifton <nickc@redhat.com>
11509 * config/rl78/rl78-real.md (bf): New pattern.
11511 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
11512 (rl78_print_operand): Do not put a # before a %B.
11513 * config/rl78/rl78.opt: Tweak doc strings.
11515 2013-09-17 DJ Delorie <dj@redhat.com>
11517 * config/rl78/constraints.md (Wcv): Allow up to $r31.
11518 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
11519 (rl78_option_override): Likewise, if -mallregs.
11520 (is_virtual_register): Likewise.
11521 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
11522 (REGNO_OK_FOR_BASE_P): Likewise.
11523 * config/rl78/rl78.opt (-mallregs): New.
11525 2013-09-17 Nick Clifton <nickc@redhat.com>
11527 * config/rl78/rl78.c (need_to_save): Change return type to bool.
11528 For interrupt functions: save all call clobbered registers if the
11529 interrupt handler is not a leaf function.
11530 (rl78_expand_prologue): Always recompute the frame information.
11531 For interrupt functions: only select bank 0 if one of the bank 0
11532 registers is going to be psuhed.
11534 2013-09-17 DJ Delorie <dj@redhat.com>
11536 * config/rl78/constraints.md: For each W* constraint, rename to C*
11537 and create a W* constraint that checks for an optional ES: prefix
11539 * config/rl78/rl78.md (UNS_ES_ADDR): New.
11540 (es_addr): New. Used to wrap far addresses.
11541 * config/rl78/rl78-protos.h (rl78_es_addr): New.
11542 (rl78_es_base): New.
11543 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
11544 wrapped far addresses.
11545 (rl78_print_operand_1): Unwrap far addresses before processing.
11546 (rl78_lo16): Wrap far addresses in unspecs.
11547 (rl78_es_addr): New.
11548 (rl78_es_base): New.
11549 (insn_ok_now): Check for not-yet-wrapped far addresses.
11550 (transcode_memory_rtx): Properly re-wrap far addresses.
11552 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
11554 * config/sparc/t-rtems: Add leon3 multilibs.
11556 2013-09-17 Cong Hou <congh@google.com>
11558 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
11559 when checking the dot production pattern. The type of rhs operand
11560 of multiply is now checked correctly.
11562 2013-09-17 Jeff Law <law@redhat.com>
11564 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
11565 edge implied equivalences into successor phis.
11566 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
11567 here from tree-ssa-threadedge.c.
11568 (mark_threaded_blocks): When threading through a joiner, if both
11569 successors of the joiner's clone reach the same block, verify the
11570 PHI arguments are equal. If not, cancel the jump threading request.
11571 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
11572 tree-ssa-threadupdate.c
11573 (thread_across_edge): Don't check PHI argument equality when
11574 threading through joiner block here.
11576 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
11578 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
11579 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
11580 (warn_uninit, warn_uninitialized_vars,
11581 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
11582 Move to tree-ssa-uninit.c.
11583 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
11584 (has_undefined_value_p): New. Pass dependant parts of
11585 ssa_undefined_value_p.
11586 (uninit_undefined_value_p): Use has_undefined_value_p.
11587 (warn_uninit, warn_uninitialized_vars,
11588 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
11589 Move from tree-ssa.c.
11590 * tree-ssa.h: Adjust prototypes.
11592 2013-09-17 Jan Hubicka <jh@suse.cz>
11594 PR middle-end/58332
11595 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
11596 * ipa-inline.c (can_inline_edge_p): Do not downgrade
11597 FUNCTION_NOT_OPTIMIZED.
11598 * ipa-inline-analysis.c (compute_inline_parameters): Function
11599 not optimized is not inlinable unless it is alwaysinline.
11600 (inline_analyze_function): Force calls in not optimized
11601 function not inlinable.
11603 2013-09-17 Jan Hubicka <jh@suse.cz>
11605 PR middle-end/58329
11606 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
11608 * ipa.c (function_and_variable_visibility): Likewise.
11609 * ipa-profile.c (ipa_profile): Likewise.
11611 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
11614 * cgraph.c (cgraph_function_body_availability): Check for ifunc
11615 attribute, and don't inline the resolver in this case.
11617 2013-09-17 Teresa Johnson <tejohnson@google.com>
11619 * coverage.c (get_coverage_counts): Add missing newline.
11621 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11623 PR tree-optimization/58088
11624 * fold-const.c (mask_with_trailing_zeros): New function.
11625 (fold_binary_loc): Make sure we don't recurse infinitely
11626 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
11627 Use mask_with_trailing_zeros where appropriate.
11629 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
11631 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
11632 of 'prev' var to get better distance estimation.
11634 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
11636 * tree-inline.h (struct copy_body_data): Add transform_parameter.
11637 * tree-inline.c (is_parameter_of): New predicate.
11638 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
11639 a parameter has been remapped.
11640 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
11641 (optimize_inline_calls): Initialize transform_parameter.
11642 (copy_gimple_seq_and_replace_locals): Likewise.
11643 (tree_function_versioning): Likewise.
11644 (maybe_inline_call_in_expr): Likewise.
11646 2013-09-17 Nick Clifton <nickc@redhat.com>
11648 * config/msp430/msp430-protos.h: Add prototypes for new functions.
11649 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
11650 interrupt handlers.
11651 (is_attr_func): New function.
11652 (msp430_is_interrupt_func): New function.
11653 (is_naked_func): New function.
11654 (is_reentrant_func): New function.
11655 (is_critical_func): New function.
11656 (msp430_start_function): Add annotations for function attributes.
11657 (msp430_attr): New function.
11658 (msp430_attribute_table): New.
11659 (msp430_function_section): New function.
11660 (TARGET_ASM_FUNCTION_SECTION): Define.
11661 (msp430_builtin): New enum.
11662 (msp430_init_builtins): New function.
11663 (msp430_builtin_devl): New function.
11664 (msp430_expand_builtin): New function.
11665 (TARGET_INIT_BUILTINS): Define.
11666 (TARGET_EXPAND_BUILTINS): Define.
11667 (TARGET_BUILTIN_DECL): Define.
11668 (msp430_expand_prologue): Add support for naked, interrupt,
11669 critical and reentrant functions.
11670 (msp430_expand_epilogue): Likewise.
11671 (msp430_print_operand): Handle 'O' character.
11672 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
11674 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
11675 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
11676 (pushm): Use a 'n' rather than an 'i' constraint.
11677 (msp_return): Add generation of the interrupt return instruction.
11678 (disable_interrupts): New pattern.
11679 (enable_interrupts): New pattern.
11680 (push_intr_state): New pattern.
11681 (pop_intr_state): New pattern.
11682 (bic_SR): New pattern.
11683 (bis_SR): New pattern.
11684 * doc/extend.texi: Document MSP430 function attributes and builtin
11687 2013-09-17 Richard Biener <rguenther@suse.de>
11689 PR tree-optimization/58432
11690 * tree-loop-distribution.c (tree_loop_distribution): Also
11691 scan PHIs for outside loop uses and seed a partition from them.
11693 2013-09-17 Bin Cheng <bin.cheng@arm.com>
11695 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
11696 (restructure_reference): Call backtrace_base_for_ref.
11698 2013-09-17 Alan Modra <amodra@gmail.com>
11701 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
11704 2013-09-16 DJ Delorie <dj@redhat.com>
11706 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
11707 vregs location for RL78/G10.
11708 (rl78_expand_prologue): Avoid SEL on G10.
11709 (rl78_expand_epilogue): Likewise.
11710 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
11711 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
11712 __RL78_G10__ when appropriate.
11713 (ASM_SPEC): Pass -mg10 along to the assembler.
11714 * config/rl78/rl78.md (sel_rb): Disable for G10.
11715 * config/rl78/rl78.opt: Add -mg10 option.
11716 * config/rl78/t-rl78: Add -mg10 multilib.
11718 2013-09-16 Xinliang David Li <davidxl@google.com>
11720 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
11721 * omp-low.c (omp_max_vf): Ditto.
11722 (expand_omp_simd): Ditto.
11723 * tree-vectorizer.c (vectorize_loops): Ditto.
11724 (gate_vect_slp): Ditto.
11725 (gate_increase_alignment): Ditto.
11726 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
11727 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
11728 (gate_tree_vectorize): Name change.
11729 (tree_vectorize): Ditto.
11730 (pass_vectorize::gate): Call new function.
11731 (pass_vectorize::execute): Ditto.
11732 * opts.c: O3 default setting change.
11733 (finish_options): Check new flag.
11734 * doc/invoke.texi: Document new flags.
11735 * common.opt: New flags.
11737 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
11739 * doc/tm.texi.in (Cond Exec Macros): Remove node.
11740 (Condition Code): Don't reference it.
11741 * doc/tm.texi: Regenerate.
11743 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
11745 PR middle-end/58418
11746 * lra-constraints.c (undo_optional_reloads): Consider all optional
11747 reload even if it did not get a hard reg.
11749 2013-09-16 Teresa Johnson <tejohnson@google.com>
11751 * dumpfile.c (dump_loc): Remove newline emission.
11752 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
11753 emission to dump_printf_loc calls where missing.
11754 (vect_mark_for_runtime_alias_test): Ditto.
11755 (vect_analyze_data_ref_dependence): Ditto.
11756 (vect_analyze_data_ref_dependences): Ditto.
11757 (vect_slp_analyze_data_ref_dependence): Ditto.
11758 (vect_slp_analyze_data_ref_dependences): Ditto.
11759 (vect_compute_data_ref_alignment): Ditto.
11760 (vect_update_misalignment_for_peel): Ditto.
11761 (vect_verify_datarefs_alignment): Ditto.
11762 (vector_alignment_reachable_p): Ditto.
11763 (vect_get_data_access_cost): Ditto.
11764 (vect_enhance_data_refs_alignment): Ditto.
11765 (vect_find_same_alignment_drs): Ditto.
11766 (vect_analyze_data_refs_alignment): Ditto.
11767 (vect_analyze_group_access): Ditto.
11768 (vect_analyze_data_ref_access): Ditto.
11769 (vect_analyze_data_ref_accesses): Ditto.
11770 (vect_prune_runtime_alias_test_list): Ditto.
11771 (vect_analyze_data_refs): Ditto.
11772 (vect_create_addr_base_for_vector_ref): Ditto.
11773 (vect_create_data_ref_ptr): Ditto.
11774 (vect_grouped_store_supported): Ditto.
11775 (vect_grouped_load_supported): Ditto.
11776 * value-prof.c (check_counter): Ditto.
11777 (check_ic_target): Ditto.
11778 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
11779 (vect_recog_widen_mult_pattern): Ditto.
11780 (vect_recog_widen_sum_pattern): Ditto.
11781 (vect_recog_over_widening_pattern): Ditto.
11782 (vect_recog_widen_shift_pattern): Ditto.
11783 (vect_recog_rotate_pattern): Ditto.
11784 (vect_recog_vector_vector_shift_pattern): Ditto.
11785 (vect_recog_divmod_pattern): Ditto.
11786 (vect_recog_mixed_size_cond_pattern): Ditto.
11787 (vect_recog_bool_pattern): Ditto.
11788 (vect_pattern_recog_1): Ditto.
11789 (vect_pattern_recog): Ditto.
11790 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
11791 (vect_is_simple_iv_evolution): Ditto.
11792 (vect_analyze_scalar_cycles_1): Ditto.
11793 (vect_get_loop_niters): Ditto.
11794 (vect_analyze_loop_1): Ditto.
11795 (vect_analyze_loop_form): Ditto.
11796 (vect_analyze_loop_operations): Ditto.
11797 (vect_analyze_loop_2): Ditto.
11798 (vect_analyze_loop): Ditto.
11799 (report_vect_op): Ditto.
11800 (vect_is_slp_reduction): Ditto.
11801 (vect_is_simple_reduction_1): Ditto.
11802 (vect_get_known_peeling_cost): Ditto.
11803 (vect_estimate_min_profitable_iters): Ditto.
11804 (vect_model_reduction_cost): Ditto.
11805 (vect_model_induction_cost): Ditto.
11806 (get_initial_def_for_induction): Ditto.
11807 (vect_create_epilog_for_reduction): Ditto.
11808 (vectorizable_reduction): Ditto.
11809 (vectorizable_induction): Ditto.
11810 (vectorizable_live_operation): Ditto.
11811 (vect_loop_kill_debug_uses): Ditto.
11812 (vect_transform_loop): Ditto.
11813 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
11814 (vect_stmt_relevant_p): Ditto.
11815 (process_use): Ditto.
11816 (vect_mark_stmts_to_be_vectorized): Ditto.
11817 (vect_model_simple_cost): Ditto.
11818 (vect_model_promotion_demotion_cost): Ditto.
11819 (vect_model_store_cost): Ditto.
11820 (vect_get_store_cost): Ditto.
11821 (vect_model_load_cost): Ditto.
11822 (vect_get_load_cost): Ditto.
11823 (vect_init_vector_1): Ditto.
11824 (vect_get_vec_def_for_operand): Ditto.
11825 (vect_finish_stmt_generation): Ditto.
11826 (vectorizable_call): Ditto.
11827 (vectorizable_conversion): Ditto.
11828 (vectorizable_assignment): Ditto.
11829 (vectorizable_shift): Ditto.
11830 (vectorizable_operation): Ditto.
11831 (vectorizable_store): Ditto.
11832 (vectorizable_load): Ditto.
11833 (vectorizable_condition): Ditto.
11834 (vect_analyze_stmt): Ditto.
11835 (vect_transform_stmt): Ditto.
11836 (vect_is_simple_use): Ditto.
11837 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
11838 (vect_can_advance_ivs_p): Ditto.
11839 (vect_update_ivs_after_vectorizer): Ditto.
11840 (vect_do_peeling_for_loop_bound): Ditto.
11841 (vect_gen_niters_for_prolog_loop): Ditto.
11842 (vect_update_inits_of_drs): Ditto.
11843 (vect_create_cond_for_alias_checks): Ditto.
11844 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
11845 (vect_build_slp_tree_1): Ditto.
11846 (vect_supported_load_permutation_p): Ditto.
11847 (vect_analyze_slp_instance): Ditto.
11848 (vect_analyze_slp): Ditto.
11849 (vect_make_slp_decision): Ditto.
11850 (vect_detect_hybrid_slp): Ditto.
11851 (vect_bb_vectorization_profitable_p): Ditto.
11852 (vect_slp_analyze_bb_1): Ditto.
11853 (vect_update_slp_costs_according_to_vf): Ditto.
11854 (vect_get_mask_element): Ditto.
11855 (vect_transform_slp_perm_load): Ditto.
11856 (vect_schedule_slp_instance): Ditto.
11857 (vect_schedule_slp): Ditto.
11858 (vect_slp_transform_bb): Ditto.
11859 * profile.c (read_profile_edge_counts): Ditto.
11860 (compute_branch_probabilities): Ditto.
11861 * coverage.c (get_coverage_counts): Ditto.
11863 2013-09-16 Diego Novillo <dnovillo@google.com>
11865 * tree-core.h: Add missing comment lines from refactoring of tree.h.
11867 2013-09-16 Jan Hubicka <jh@suse.cz>
11869 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
11870 abstract functions; for static functions check the presence of body.
11872 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
11874 * config/aarch64/aarch64-simd-builtins.def (fma): New.
11875 * config/aarch64/aarch64-simd.md
11876 (aarch64_mla_elt<mode>): New.
11877 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
11878 (aarch64_mls_elt<mode>): Likewise.
11879 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
11880 (aarch64_fma4_elt<mode>): Likewise.
11881 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
11882 (aarch64_fma4_elt_to_128v2df): Likewise.
11883 (aarch64_fma4_elt_to_64df): Likewise.
11884 (fnma<mode>4): Likewise.
11885 (aarch64_fnma4_elt<mode>): Likewise.
11886 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
11887 (aarch64_fnma4_elt_to_128v2df): Likewise.
11888 (aarch64_fnma4_elt_to_64df): Likewise.
11889 * config/aarch64/iterators.md (VDQSF): New.
11890 * config/aarch64/arm_neon.h
11891 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
11892 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
11894 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
11896 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
11897 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
11898 (aarch64_mul3_elt_to_128df): Likewise.
11899 (aarch64_mul3_elt_to_64v2df): Likewise.
11900 * config/aarch64/iterators.md (VEL): Also handle DFmode.
11902 (VMUL_CHANGE_NLANES) Likewise.
11905 * config/aarch64/arm_neon.h
11906 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
11908 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
11910 * config/aarch64/arm_neon.h
11911 (vcvtx_high_f32_f64): Fix parameters.
11913 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11914 Uros Bizjak <ubizjak@gmail.com>
11916 * config/alpha.c: Include tree-ssanames.h.
11918 2013-09-16 Richard Biener <rguenther@suse.de>
11920 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
11921 (dot_rdg_1): Handle control_dd.
11922 (create_edge_for_control_dependence): New function.
11923 (create_rdg_edges): Add control dependences if asked for.
11924 (build_rdg): Likewise.
11925 (generate_loops_for_partition): If there are not necessary
11926 control stmts remove all their dependencies.
11927 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
11928 (distribute_loop): Pass on control dependences.
11929 (tree_loop_distribution): Compute control dependences and remove
11930 restriction on number of loop nodes.
11932 2013-09-16 Jakub Jelinek <jakub@redhat.com>
11934 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
11935 for internal calls.
11937 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
11939 * cse.c (try_const_anchors): Punt on CC modes.
11941 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11943 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
11945 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
11948 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
11949 calls to word_mode.
11951 2013-09-14 Iain Sandoe <iain@codesourcery.com>
11954 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
11956 (darwin_objc1_section): Likewise.
11957 (darwin_file_end): Emit Image Info section when required.
11959 2013-09-14 Jan Hubicka <jh@suse.cz>
11961 * tree-into-ssa.c (gate_into_ssa): New.
11962 (pass_data_build_ssa): Use it.
11963 * cgraph.h (expand_thunk): Update prototype.
11964 * cgraphunit.c (analyze_function): Expand thunks early.
11965 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
11966 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
11967 set lowered flag; do not add new function.
11968 (assemble_thunks_and_aliases): Update.
11969 * tree-ssa.c (gate_init_datastructures): New gate.
11970 (pass_data_init_datastructures): Use it.
11972 2013-09-14 Iain Sandoe <iain@codesourcery.com>
11975 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
11976 xmm register set described in the psABI.
11978 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
11980 * dwarf2out.c (should_emit_struct_debug): Add check
11981 for type_decl variable is not NULL.
11983 2013-09-13 Jacek Caban <jacek@codeweavers.com>
11985 * config.gcc: Use new winnt-c.c target hooks
11986 * config/t-winnt: New file
11987 * config/winnt-c.c: New file
11988 * doc/tm.texi.in: Document new hook
11989 * doc/tm.texi: Regenerated
11991 2013-09-13 Jan Hubicka <jh@suse.cz>
11993 PR middle-end/58094
11994 * ipa-inline.c (check_callers): New function.
11995 (check_caller_edge): Remove.
11996 (want_inline_function_to_all_callers_p): Also permit alises that are
11998 (inline_to_all_callers): Terminate the walk when devirtualization
11999 introduce new calls.
12001 2013-09-13 Jan Hubicka <jh@suse.cz>
12003 * ipa-inline-analysis.c (struct growth_data): Add node.
12004 (do_estimate_growth_1): Fix detection of recursion.
12006 2013-09-13 Jakub Jelinek <jakub@redhat.com>
12008 PR tree-optimization/58392
12009 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
12010 to avoid shadowing of outer loop variable. If
12011 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
12012 replace_by_duplicate_decl simduid of loops that have it set and
12013 set dest_cfun->has_simduid_loops and/or
12014 dest_cfun->has_force_vect_loops.
12015 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
12016 instead of maybe_lookup_decl.
12017 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
12018 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
12019 call. Copy over force_vect and copy and remap simduid. Set
12020 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
12021 (copy_cfg_body): Remove blocks_to_copy argument. Use
12022 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
12023 caller. Don't set cfun->has_simduid_loops and/or
12024 cfun->has_force_vect_loops here.
12025 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
12027 (expand_call_inline, tree_function_versioning): Adjust copy_body
12030 2013-09-13 Martin Jambor <mjambor@suse.cz>
12033 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
12034 the assert if the edge was a speculative one.
12036 2013-09-13 Richard Biener <rguenther@suse.de>
12038 * tree-data-ref.h (known_dependences_p): Move here ...
12039 * tree-loop-distribution.c (known_dependences_p): ... from here.
12040 (dump_rdg_component, debug_rdg_component): Remove.
12041 (dump_rdg): Adjust.
12042 (generate_loops_for_partition): Use gimple_uid instead of
12043 relying on matching stmt visit order.
12044 (rdg_build_partitions): Take starting stmt vector.
12045 (ldist_gen): Merge into ...
12046 (distribute_loop): ... this function. Do not compute starting
12048 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
12050 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12052 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
12053 Set type attribute correctly. Set predicable_short_it attribute.
12054 (cmpsi_shiftsi): Remove %? from output template.
12056 2013-09-13 Richard Biener <rguenther@suse.de>
12058 * tree-loop-distribution.c (struct rdg_component,
12059 rdg_defs_used_in_other_loops_p, free_rdg_components,
12060 rdg_build_components): Remove.
12061 (stmts_from_loop): Do not record virtual PHIs.
12062 (generate_loops_for_partition): Skip virtual PHIs.
12063 (build_rdg_partition_for_component): Rename to ...
12064 (build_rdg_partition_for_vertex): ... this and adjust.
12065 (rdg_build_partitions): Take a vector of starting vertices
12066 instead of components. Remove unnecessary leftover handling.
12067 (ldist_gen): Do not build components or record other stores.
12068 (distribute_loop): Do not distribute loops containing stmts
12071 2013-09-13 Christian Bruel <christian.bruel@st.com>
12074 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
12076 2013-09-13 Kai Tietz <ktietz@redhat.com>
12078 * config.gcc: Separate cases for mingw and cygwin targets,
12079 and add 64-bit cygwin target case.
12081 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
12082 dll-export inline-functions.
12083 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
12085 2013-09-13 Jeff Law <law@redhat.com>
12087 PR middle-end/58387
12089 2013-09-06 Jeff Law <law@redhat.com>
12091 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
12092 edge implied equivalences into successor phis.
12094 2013-09-12 DJ Delorie <dj@redhat.com>
12096 * config/rl78/rl78-virt.md: Change from | to \; for asm line
12099 2013-09-12 Brooks Moses <bmoses@google.com>
12102 * Makefile.in: Do not install driver binaries in $(target)/bin.
12104 2013-09-12 DJ Delorie <dj@redhat.com>
12106 * config/rl78/rl78.opt (mrelax): New.
12107 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
12108 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
12110 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
12112 (rl78_expand_epilogue): Likewise.
12114 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
12116 PR middle-end/58335
12117 * lra-eliminations.c (remove_reg_equal_offset_note): New.
12118 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
12119 pointer elimination with using remove_reg_equal_offset_note.
12121 2013-09-12 DJ Delorie <dj@redhat.com>
12123 * config/msp430/: New port.
12124 * config.gcc (msp430): Added.
12125 * doc/invoke.texi: Document MSP430 options.
12126 * doc/install.texi: Document msp430-elf
12127 * doc/md.texi: Document msp430-elf
12128 * doc/contrib.texi: Document msp430-elf
12130 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
12133 2013-09-12 Martin Jambor <mjambor@suse.cz>
12136 * ipa-prop.c (remove_described_reference): Give up if the edge in the
12137 reference descriptor is NULL.
12138 (ipa_edge_removal_hook): If owning a reference descriptor, set its
12141 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
12143 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
12144 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
12145 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
12146 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
12147 make_temp_ssa_name): move to tree-ssanames.h
12148 * tree-ssa-alias.h: Move prototype.
12149 * tree-ssa.h: Include tree-ssanames.h.
12150 * tree-ssanames.c (FREE_SSANAMES): Move to here.
12151 * tree-ssanames.h: New. Move items from tree-flow*.h
12152 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
12154 2013-09-12 Richard Biener <rguenther@suse.de>
12156 PR tree-optimization/58404
12157 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
12158 propagate non-invariant addresses into dereferences wrapped
12159 in component references.
12161 2013-09-12 Richard Biener <rguenther@suse.de>
12163 PR tree-optimization/58402
12164 * passes.def: Move pass_late_warn_uninitialized later.
12166 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
12168 * tree-ssa.h: New. Move content from tree-flow.h and
12169 tree-flow-inline.h.
12170 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
12171 Move prototypes belonging to tree-ssa.c.
12172 * tree-flow-inline.h (redirect_edge_var_map_def,
12173 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
12175 * gimple.h: Adjust prototypes.
12176 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
12178 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
12179 * tree.h: Move prototype to tree-ssa.h.
12180 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
12181 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
12182 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
12183 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
12184 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
12185 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
12186 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
12187 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
12188 graphite-blocking.c, graphite-clast-to-gimple.c,
12189 graphite-dependences.c, graphite-interchange.c,
12190 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
12191 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
12192 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
12193 ipa-reference.c, ipa-split.c, ipa-utils.c,
12194 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
12195 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
12196 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
12197 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
12198 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
12199 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
12200 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
12201 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
12202 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
12203 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
12204 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
12205 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
12206 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
12207 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
12208 value-prof.c, var-tracking.c,
12209 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
12211 2013-09-12 Richard Biener <rguenther@suse.de>
12213 PR tree-optimization/58396
12214 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
12215 (build_rdg): Take a loop-nest parameter, fix memleaks.
12216 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
12218 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
12220 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
12223 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
12225 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
12226 The FMA instruction names should have a 'v' prefix.
12228 2013-09-12 Richard Biener <rguenther@suse.de>
12230 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
12231 (dot_rdg): Use popen instead of system in optional code.
12232 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
12233 (already_processed_vertex_p): Adjust.
12234 (has_anti_or_output_dependence, predecessor_has_mem_write,
12235 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
12236 rdg_flag_uses): Remove.
12237 (rdg_flag_vertex): Simplify.
12238 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
12240 (build_rdg_partition_for_component): Process the first vertex
12241 of a component only.
12242 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
12244 2013-09-12 Alan Modra <amodra@gmail.com>
12246 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
12248 2013-09-11 DJ Delorie <dj@redhat.com>
12249 Nick Clifton <nickc@redhat.com>
12251 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
12252 (rl78_stack_based_mem): New.
12253 * config/rl78/constraints.md (Iv08): New.
12262 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
12264 (movsi): Change from expand to insn-and-split.
12265 (ashrsi3): Clobber AX.
12269 * config/rl78/rl78.md (CC_REG): Fix.
12270 (addsi3): Allow memory and immediate operands.
12271 (addsi3_internal): Split into...
12272 (addsi3_internal_virt): ...new, and ...
12273 (addsi3_internal_real): ...new.
12275 (subsi3_internal_virt): New.
12276 (subsi3_internal_real): New.
12277 (mulsi3): Add memory operand.
12278 (mulsi3_rl78): Likewise.
12279 (mulsi3_g13): Likewise.
12280 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
12281 (cbranchqi4_real): Add more constraint options.
12282 (cbranchhi4_real): Expand pattern.
12283 (cbranchhi4_real_signed): New.
12284 (cbranchhi4_real_inverted): New.
12285 (cbranchsi4_real_lt): New.
12286 (cbranchsi4_real_ge): New.
12287 (cbranchsi4_real_signed): New.
12288 (cbranchsi4_real): New.
12290 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
12292 (lshrsi3_virt): Likewise.
12293 (ashlsi3_virt): Likewise.
12294 (cbranchqi4_virt_signed): New.
12295 (cbranchhi4_virt_signed): New.
12296 (cbranchsi4_virt): New.
12297 * config/rl78/rl78.c: Whitespace fixes throughout.
12298 (move_elim_pass): New.
12299 (pass_data_rl78_move_elim): New.
12300 (pass_rl78_move_elim): New.
12301 (make_pass_rl78_move_elim): New.
12302 (rl78_devirt_info): Run devirt earlier.
12303 (rl78_move_elim_info): New.
12304 (rl78_asm_file_start): Register it.
12305 (rl78_split_movsi): New.
12306 (rl78_as_legitimate_address): Allow virtual base registers when
12308 (rl78_addr_space_convert): Remove spurious debug stuff.
12309 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
12310 (rl78_print_operand): More cases for not printing '#'.
12311 (rl78_expand_compare): Remove most of the logic.
12312 (content_memory): New.
12313 (clear_content_memory): New.
12314 (get_content_index): New.
12315 (get_content_name): New.
12316 (display_content_memory): New.
12317 (update_content): New.
12318 (record_content): New.
12319 (already_contains): New.
12320 (insn_ok_now): Re-recog insns with virtual registers.
12321 (add_postponed_content_update): New.
12322 (process_postponed_content_update): New.
12323 (gen_and_emit_move): New.
12324 (transcode_memory_rtx): Record new location content.
12325 Use gen_and_emit_move.
12326 (force_into_acc): New.
12327 (move_to_acc): Use gen_and_emit_move.
12328 (move_from_acc): Likewise.
12329 (move_acc_to_reg): Likewise.
12330 (move_to_x): Likewise.
12331 (move_to_hl): Likewise.
12332 (move_to_de): Likewise.
12333 (rl78_alloc_physical_registers_op1): Record location content.
12334 (has_constraint): New.
12335 (rl78_alloc_physical_registers_op2): Record location content.
12336 Optimize use of HL.
12337 (rl78_alloc_physical_registers_ro1): Likewise.
12338 (rl78_alloc_physical_registers_cmp): Likewise.
12339 (rl78_alloc_physical_registers_umul): Likewise.
12340 (rl78_alloc_address_registers_macax): New.
12341 (rl78_alloc_physical_registers): Initialize and set location
12342 content memory as needed.
12343 (rl78_reorg): Make sure split2 is called.
12344 (rl78_rtx_costs): New.
12346 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
12348 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
12349 for (not (neg ...)) and (neg (not ...)) cases.
12351 2013-09-11 Richard Biener <rguenther@suse.de>
12353 PR middle-end/58377
12354 * passes.def: Split critical edges before late uninit warning passes.
12355 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
12357 2013-09-11 Jakub Jelinek <jakub@redhat.com>
12359 PR tree-optimization/58385
12360 * fold-const.c (build_range_check): If both low and high are NULL,
12361 use omit_one_operand_loc to preserve exp side-effects.
12363 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12365 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
12367 2013-09-11 Richard Biener <rguenther@suse.de>
12369 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
12370 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
12371 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
12372 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
12373 stmts_from_loop, known_dependences_p, build_empty_rdg,
12374 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
12375 * tree-loop-distribution.c: ... here.
12376 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
12377 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
12378 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
12379 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
12380 * tree-loop-distribution.c: ... here.
12381 * tree-loop-distribution.c: Include gimple-pretty-print.h.
12382 (struct partition_s): Add loops member.
12383 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
12384 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
12385 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
12387 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
12388 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
12389 Sergey Lega <sergey.s.lega@intel.com>
12390 Anna Tikhonova <anna.tikhonova@intel.com>
12391 Ilya Tocar <ilya.tocar@intel.com>
12392 Andrey Turetskiy <andrey.turetskiy@intel.com>
12393 Ilya Verbin <ilya.verbin@intel.com>
12394 Kirill Yukhin <kirill.yukhin@intel.com>
12395 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
12397 * config/i386/constraints.md (k): New.
12399 * config/i386/i386.c (const regclass_map): Add new mask registers.
12400 (dbx_register_map): Ditto.
12401 (dbx64_register_map): Ditto.
12402 (svr4_dbx_register_map): Ditto.
12403 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
12405 (ix86_preferred_reload_class): Disable constants for mask registers.
12406 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
12407 (ix86_hard_regno_mode_ok): Support new mask registers.
12408 (x86_order_regs_for_local_alloc): Ditto.
12409 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
12410 (FIXED_REGISTERS): Add new mask registers.
12411 (CALL_USED_REGISTERS): Ditto.
12412 (REG_ALLOC_ORDER): Ditto.
12413 (VALID_MASK_REG_MODE): New.
12414 (FIRST_MASK_REG): Ditto.
12415 (LAST_MASK_REG): Ditto.
12416 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
12417 (MAYBE_MASK_CLASS_P): New.
12418 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
12419 (REG_CLASS_CONTENTS): Ditto.
12420 (MASK_REGNO_P): New.
12421 (ANY_MASK_REG_P): Ditto.
12422 (HI_REGISTER_NAMES): Add new mask registers.
12423 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
12424 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
12426 (attribute "type"): Add mskmov, msklog.
12427 (attribute "length_immediate"): Support them.
12428 (attribute "memory"): Ditto.
12429 (attribute "prefix_0f"): Ditto.
12430 (*movhi_internal): Support new mask registers.
12431 (*movqi_internal): Ditto.
12432 (define_split): Split out clobber pattern is a logic
12433 insn on mask registers.
12434 (*k<logic><mode>): New.
12435 (*andhi_1): Extend to support mask regs.
12436 (*andqi_1): Extend to support mask regs.
12437 (kandn<mode>): New.
12438 (define_split): Split and-not to and and not if operands
12440 (*<code><mode>_1): Separate HI mode to new pattern...
12441 (*<code>hi_1): This.
12442 (*<code>qi_1): Extend to support mask regs.
12443 (kxnor<mode>): New.
12444 (kortestzhi): Ditto.
12445 (kortestchi): Ditto.
12447 (*one_cmpl<mode>2_1): Remove HImode and handle it...
12448 (*one_cmplhi2_1): ...Here, now with mask registers support.
12449 (*one_cmplqi2_1): Support new mask registers.
12450 (HI/QImode arithmetics splitter): Don't split if mask registers
12452 (HI/QImode not splitter): Ditto.
12453 * config/i386/predicated.md (mask_reg_operand): New.
12454 (general_reg_operand): Ditto.
12456 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
12458 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
12459 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
12461 2013-09-10 Jeff Law <law@redhat.com>
12463 PR tree-optimization/58380
12464 * tree-ssa-threadupdate.c (thread_block): Recognize another case
12465 of threading through a buried loop header.
12467 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
12468 return value for single successor case.
12470 2013-09-10 Jan Hubicka <jh@suse.cz>
12472 * ipa-devirt.c (ipa_devirt): Enable with LTO.
12474 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
12477 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
12478 support conditional execution.
12479 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
12481 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
12483 * lra.c (lra): Clear lra_optional_reload_pseudos before every
12485 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
12486 Check destination too to check move insn.
12487 (undo_optional_reloads): Add check that the original peudo did not
12488 changed its allocation and the optional reload was inherited on last
12489 inheritance pass. Break loop after deciding to keep optional reload.
12490 (lra_undo_inheritance): Add check that inherited pseudo still in
12493 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
12495 * config/aarch64/aarch64.md (generic_sched): New.
12496 * config/aarch64/aarch64-generic.md (load): Make conditional
12497 on generic_sched attribute.
12498 (nonload): Likewise.
12500 2013-09-10 Jan Hubicka <jh@suse.cz>
12502 * lto-cgraph.c: Include ipa-utils.h.
12503 (compute_ltrans_boundary): Also add possible targets into the boundary.
12505 2013-09-10 Jan Hubicka <jh@suse.cz>
12507 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
12508 VAR_DECL of vtable rather than full expression.
12510 2013-09-10 Jan Hubicka <jh@suse.cz>
12511 Paolo Carlini <paolo.carlini@oracle.com>
12513 * cgraphunit.c (analyze_functions): Save input_location, set it
12514 to UNKNOWN_LOCATION and restore it at the end.
12516 2013-09-10 Martin Jambor <mjambor@suse.cz>
12518 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
12519 represented by a thunk.
12521 2013-09-10 Jeff Law <law@redhat.com>
12523 PR tree-optimization/58343
12524 * tree-ssa-threadupdate.c (thread_block): Identify and disable
12525 jump threading requests through loop headers buried in the middle
12526 of a jump threading path.
12528 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
12529 in return value/type.
12531 2013-09-10 Jakub Jelinek <jakub@redhat.com>
12533 PR rtl-optimization/58365
12534 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
12535 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
12538 2013-09-10 Richard Biener <rguenther@suse.de>
12540 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
12541 * tree-data-ref.c (create_rdg_vertices): Collect all data
12542 references, signal failure to the caller, use data-ref API.
12543 (build_rdg): Compute data references only once. Maintain lifetime
12544 of data references and data dependences from within RDG.
12545 (free_rdg): Free dependence relations.
12546 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
12547 inventing extra dependences.
12548 (distribute_loop): Update for RDG API changes.
12550 2013-09-10 Kai Tietz <ktietz@redhat.com>
12552 * doc/invoke.texi (fms-extensions): Document changed
12553 behavior for ms-abi targets.
12554 * config/i386/i386.c (ix86_option_override_internal):
12555 Set default value of option -fms-extension for ms-abi targets.
12557 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12559 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
12561 2013-09-10 Alan Modra <amodra@gmail.com>
12564 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
12566 2013-09-10 Alan Modra <amodra@gmail.com>
12568 * config/rs6000/predicates.md (add_cint_operand): New.
12569 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
12570 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
12571 using add_cint_operand.
12572 (largetoc_high_plus_aix): Likewise.
12574 2013-09-09 Jakub Jelinek <jakub@redhat.com>
12576 PR tree-optimization/58364
12577 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
12578 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
12579 the current range can't be an unconditional true or false.
12581 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
12583 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
12585 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
12587 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
12589 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
12592 * doc/invoke.texi (-Wdelete-incomplete): Document it.
12594 2013-09-09 Ian Bolton <ian.bolton@arm.com>
12596 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
12597 NO_REGS for immediate that can't be moved directly into FP_REGS.
12599 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12601 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
12602 comparison with negated operand.
12603 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
12606 2013-09-09 Richard Biener <rguenther@suse.de>
12608 PR middle-end/58326
12609 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
12610 of a subloop record all its block as affecting loop-closed SSA form.
12612 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
12614 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
12615 of an rtx/bitpos pair.
12616 (store_fixed_bit_field): Update accordingly.
12618 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
12620 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
12622 * builtins.c (expand_errno_check): Likewise.
12623 * dwarf2cfi.c (init_return_column_size): Likewise.
12624 * except.c (sjlj_mark_call_sites): Likewise.
12625 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
12626 * lra-constraints.c (emit_inc): Likewise.
12627 * ree.c (combine_set_extension): Likewise.
12628 * regmove.c (fixup_match_2): Likewise.
12629 * reload1.c (inc_for_reload): Likewise.
12631 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
12633 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
12634 (force_to_mode, simplify_shift_const_1, simplify_comparison):
12635 Use gen_int_mode with the mode of the associated simplify_* call.
12636 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
12637 * expmed.c (expand_shift_1): Likewise.
12638 * function.c (instantiate_virtual_regs_in_insn): Likewise.
12639 * loop-iv.c (iv_number_of_iterations): Likewise.
12640 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
12641 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
12643 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
12645 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
12646 of the associated expand_* call.
12647 (asan_emit_stack_protection): Likewise.
12648 * builtins.c (round_trampoline_addr): Likewise.
12649 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
12650 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
12651 (emit_store_flag): Likewise.
12652 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
12653 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
12655 * function.c (instantiate_virtual_regs_in_insn): Likewise.
12656 * ifcvt.c (noce_try_store_flag_constants): Likewise.
12657 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
12658 * modulo-sched.c (generate_prolog_epilog): Likewise.
12659 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
12660 (expand_ctz, expand_ffs, expand_unop): Likewise.
12662 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
12664 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
12665 of the associated gen_rtx_* call.
12666 * caller-save.c (init_caller_save): Likewise.
12667 * combine.c (find_split_point, make_extraction): Likewise.
12668 (make_compound_operation): Likewise.
12669 * dwarf2out.c (mem_loc_descriptor): Likewise.
12670 * explow.c (plus_constant, probe_stack_range): Likewise.
12671 * expmed.c (expand_mult_const): Likewise.
12672 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
12673 * reload1.c (init_reload): Likewise.
12674 * valtrack.c (cleanup_auto_inc_dec): Likewise.
12675 * var-tracking.c (adjust_mems): Likewise.
12676 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
12677 rather than gen_rtx_fmt_ee.
12679 2013-09-09 Jan Hubicka <jh@suse.cz>
12681 PR middle-end/58294
12682 * value-prof.c (gimple_ic): Copy also abnormal edges.
12684 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
12686 * asan.c (asan_shadow_cst): Use gen_int_mode.
12688 2013-09-08 Jan Hubicka <jh@suse.cz>
12690 * ipa-profile.c: Add toplevel comment.
12691 (ipa_propagate_frequency_1): Be more conservative when profile is read.
12692 (contains_hot_call_p): New function.
12693 (ipa_propagate_frequency): Set frequencies based on counts when
12695 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
12696 profile; do not tamper with profile after inlining if it is read.
12698 2013-09-08 Jan Hubicka <jh@suse.cz>
12700 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
12703 2013-09-08 Jan Hubicka <jh@suse.cz>
12705 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
12706 summary generation.
12708 2013-09-08 Jeff Law <law@redhat.com>
12711 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
12714 2013-09-08 Andi Kleen <ak@linux.intel.com>
12716 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
12718 2013-09-08 Jan Hubicka <jh@suse.cz>
12720 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
12722 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
12724 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
12725 for non-debug insns.
12726 * lra.c (new_insn_reg): Take the containing insn as a parameter.
12727 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
12728 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
12731 2013-09-08 Jan Hubicka <jh@suse.cz>
12733 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
12734 targets and devirtualize to BUILT_IN_UNREACHABLE.
12735 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
12736 * ipa.c (walk_polymorphic_call_targets): New function.
12737 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
12738 functions; use the new timevar.
12739 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
12740 was removed from the program.
12741 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
12742 not consider it in the walk when its vtable is dead.
12743 (possible_polymorphic_call_targets_1): Pass anonymous flag to
12745 (devirt_variable_node_removal_hook): New function.
12746 (possible_polymorphic_call_targets): Also register
12747 devirt_variable_node_removal_hook.
12748 (ipa_devirt): Do not do non-speculative devirtualization.
12749 (gate_ipa_devirt): One execute if devirtualizing speculatively.
12751 2013-09-08 Jan Hubicka <jh@suse.cz>
12753 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
12754 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
12755 varpool_remove_variable_insertion_hook): Declare.
12756 * varpool.c (varpool_node_hook_list): New structure.
12757 (first_varpool_node_removal_hook,
12758 first_varpool_variable_insertion_hook): New variables.
12759 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
12760 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
12761 varpool_remove_variable_insertion_hook,
12762 varpool_call_variable_insertion_hooks): New functions.
12763 (varpool_remove_node): Use it.
12765 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
12768 * diagnostic.c (diagnostic_build_prefix): When s.file is
12769 "<built-in>" don't output line and column numbers.
12771 2013-09-06 Jan Hubicka <jh@suse.cz>
12773 * cgraphunit.c (expand_thunk): Get body before touching arguments.
12774 * lto-streamer-out.c: Stream thunks, too.
12775 * lto-streamer-in.c (input_function): Pop cfun here
12776 (lto_read_body): Instead of here.
12778 2013-09-06 Caroline Tice <cmtice@google.com>
12780 * doc/install.texi: Add documentation for the --enable-vtable-verify
12781 and the --disable-libvtv configure options.
12783 2013-09-06 Jeff Law <law@redhat.com>
12785 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
12786 edge implied equivalences into successor phis.
12788 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
12790 * resource.c (mark_referenced_resources): Handle COND_EXEC.
12792 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
12794 * resource.c (mark_target_live_regs): Compute resources taking
12795 into account if a call is predicated or not.
12797 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
12799 * toplev.c (output_stack_usage): Be prepared for suffixes created by
12800 the compiler in the function names.
12802 2013-09-06 Jan Hubicka <jh@suse.cz>
12804 PR middle-end/58094
12805 * ipa-inline.c (has_caller_p): New function.
12806 (want_inline_function_to_all_callers_p): Use it.
12807 (sum_callers, inline_to_all_callers): Break out from ...
12808 (ipa_inline): ... here.
12810 2013-09-06 Jan Hubicka <jh@suse.cz>
12812 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
12813 only when AVX is enabled.
12815 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
12817 * config/aarch64/aarch64.md
12818 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
12819 is fpsimd_<load/store>2.
12820 (load_pair<mode>): Likewise.
12821 (store_pair<mode>): Likewise.
12823 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
12825 * config/arm/types.md (type): Add "mrs" type.
12826 * config/aarch64/aarch64.md
12827 (aarch64_load_tp_hard): Make type "mrs".
12828 * config/arm/arm.md
12829 (load_tp_hard): Make type "mrs".
12830 * config/arm/cortex-a15.md: Update with new attributes.
12831 * config/arm/cortex-a5.md: Update with new attributes.
12832 * config/arm/cortex-a53.md: Update with new attributes.
12833 * config/arm/cortex-a7.md: Update with new attributes.
12834 * config/arm/cortex-a8.md: Update with new attributes.
12835 * config/arm/cortex-a9.md: Update with new attributes.
12836 * config/arm/cortex-m4.md: Update with new attributes.
12837 * config/arm/cortex-r4.md: Update with new attributes.
12838 * config/arm/fa526.md: Update with new attributes.
12839 * config/arm/fa606te.md: Update with new attributes.
12840 * config/arm/fa626te.md: Update with new attributes.
12841 * config/arm/fa726te.md: Update with new attributes.
12843 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
12845 * config/aarch64/aarch64.md
12846 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
12847 (*movtf_aarch64): Likewise.
12848 * config/arm/arm.md
12849 (thumb1_movdi_insn): Use "multiple" for type where more than one
12850 instruction is used for a move.
12851 (*arm32_movhf): Likewise.
12852 (*thumb_movdf_insn): Likewise.
12854 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
12856 * config/arm/types.md (type): Rename fcpys to fmov.
12857 * config/arm/vfp.md
12858 (*arm_movsi_vfp): Rename type fcpys as fmov.
12859 (*thumb2_movsi_vfp): Likewise
12860 (*movhf_vfp_neon): Likewise
12861 (*movhf_vfp): Likewise
12862 (*movsf_vfp): Likewise
12863 (*thumb2_movsf_vfp): Likewise
12864 (*movsfcc_vfp): Likewise
12865 (*thumb2_movsfcc_vfp): Likewise
12866 * config/aarch64/aarch64-simd.md
12867 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
12868 * config/aarch64/aarch64.md
12869 (*movsi_aarch64): Replace type mov_reg with fmovs.
12870 (*movdi_aarch64): Likewise
12871 (*movsf_aarch64): Likewise
12872 (*movdf_aarch64): Likewise
12874 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
12875 * config/arm/iwmmxt.md
12876 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
12877 * config/arm/arm1020e.md: Update with new attributes.
12878 * config/arm/cortex-a15-neon.md: Update with new attributes.
12879 * config/arm/cortex-a5.md: Update with new attributes.
12880 * config/arm/cortex-a53.md: Update with new attributes.
12881 * config/arm/cortex-a7.md: Update with new attributes.
12882 * config/arm/cortex-a8-neon.md: Update with new attributes.
12883 * config/arm/cortex-a9.md: Update with new attributes.
12884 * config/arm/cortex-m4-fpu.md: Update with new attributes.
12885 * config/arm/cortex-r4f.md: Update with new attributes.
12886 * config/arm/marvell-pj4.md: Update with new attributes.
12887 * config/arm/vfp11.md: Update with new attributes.
12889 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
12891 * config/aarch64/aarch64.md
12892 (*madd<mode>): Fix type attribute.
12893 (*maddsi_uxtw): Likewise.
12894 (*msub<mode>): Likewise.
12895 (*msubsi_uxtw): Likewise.
12896 (<su_optab>maddsidi4): Likewise.
12897 (<su_optab>msubsidi4): Likewise.
12899 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
12901 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
12902 * config/arm/arm.md (core_cycles): Remove fdiv.
12903 * config/arm/vfp.md:
12904 (*sqrtsf2_vfp): Update for attribute changes.
12905 (*sqrtdf2_vfp): Likewise.
12906 * config/aarch64/aarch64.md:
12907 (sqrt<mode>2): Update for attribute changes.
12908 * config/arm/arm1020e.md: Update with new attributes.
12909 * config/arm/cortex-a15-neon.md: Update with new attributes.
12910 * config/arm/cortex-a5.md: Update with new attributes.
12911 * config/arm/cortex-a53.md: Update with new attributes.
12912 * config/arm/cortex-a7.md: Update with new attributes.
12913 * config/arm/cortex-a8-neon.md: Update with new attributes.
12914 * config/arm/cortex-a9.md: Update with new attributes.
12915 * config/arm/cortex-m4-fpu.md: Update with new attributes.
12916 * config/arm/cortex-r4f.md: Update with new attributes.
12917 * config/arm/marvell-pj4.md: Update with new attributes.
12918 * config/arm/vfp11.md: Update with new attributes.
12920 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
12922 * config/arm/types.md
12923 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
12924 * config/aarch64/aarch64.md
12925 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
12927 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
12928 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
12929 (float<GPI:mode><GPF:mode>2): Likewise.
12930 * config/arm/vfp.md
12931 (*truncsisf2_vfp): Update with new attributes.
12932 (*truncsidf2_vfp): Likewise.
12933 (fixuns_truncsfsi2): Likewise.
12934 (fixuns_truncdfsi2): Likewise.
12935 (*floatsisf2_vfp): Likewise.
12936 (*floatsidf2_vfp): Likewise.
12937 (floatunssisf2): Likewise.
12938 (floatunssidf2): Likewise.
12939 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
12940 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
12941 * config/arm/arm1020e.md: Update with new attributes.
12942 * config/arm/cortex-a15-neon.md: Update with new attributes.
12943 * config/arm/cortex-a5.md: Update with new attributes.
12944 * config/arm/cortex-a53.md: Update with new attributes.
12945 * config/arm/cortex-a7.md: Update with new attributes.
12946 * config/arm/cortex-a8-neon.md: Update with new attributes.
12947 * config/arm/cortex-a9.md: Update with new attributes.
12948 * config/arm/cortex-m4-fpu.md: Update with new attributes.
12949 * config/arm/cortex-r4f.md: Update with new attributes.
12950 * config/arm/marvell-pj4.md: Update with new attributes.
12951 * config/arm/vfp11.md: Update with new attributes.
12953 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
12955 * config/aarch64/arm_neon.h
12956 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
12957 (vqtbx<1,2,3,4><q>_s8): Likewise.
12959 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
12961 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
12962 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
12963 (add<mode>3): Add type attribute.
12964 (add<mode>3): Likewise.
12965 (usadd<mode>3): Likewise.
12966 (ssadd<mode>3): Likewise.
12967 (sub<mode>3): Likewise.
12968 (sub<mode>3): Likewise.
12969 (ussub<mode>3): Likewise.
12970 (sssub<mode>3): Likewise.
12971 (ssmulsa3): Likewise.
12972 (usmulusa3): Likewise.
12973 (arm_usatsihi): Likewise.
12974 * config/arm/vfp.md
12975 (*movdi_vfp): Add types for all instructions.
12976 (*movdi_vfp_cortexa8): Likewise.
12977 (*movhf_vfp_neon): Likewise.
12978 (*movhf_vfp): Likewise.
12979 (*movdf_vfp): Likewise.
12980 (*thumb2_movdf_vfp): Likewise.
12981 (*thumb2_movdfcc_vfp): Likewise.
12982 * config/arm/arm.md: Add type attribute to all insn patterns.
12983 (*thumb1_adddi3): Add type attribute.
12984 (*arm_adddi3): Likewise.
12985 (*adddi_sesidi_di): Likewise.
12986 (*adddi_zesidi_di): Likewise.
12987 (*thumb1_addsi3): Likewise.
12988 (addsi3_compare0): Likewise.
12989 (*addsi3_compare0_scratch): Likewise.
12990 (*compare_negsi_si): Likewise.
12991 (cmpsi2_addneg): Likewise.
12992 (*addsi3_carryin_<optab>): Likewise.
12993 (*addsi3_carryin_alt2_<optab>): Likewise.
12994 (*addsi3_carryin_clobercc_<optab>): Likewise.
12995 (*subsi3_carryin): Likewise.
12996 (*subsi3_carryin_const): Likewise.
12997 (*subsi3_carryin_compare): Likewise.
12998 (*subsi3_carryin_compare_const): Likewise.
12999 (*arm_subdi3): Likewise.
13000 (*thumb_subdi3): Likewise.
13001 (*subdi_di_zesidi): Likewise.
13002 (*subdi_di_sesidi): Likewise.
13003 (*subdi_zesidi_di): Likewise.
13004 (*subdi_sesidi_di): Likewise.
13005 (*subdi_zesidi_ze): Likewise.
13006 (thumb1_subsi3_insn): Likewise.
13007 (*arm_subsi3_insn): Likewise.
13008 (*anddi3_insn): Likewise.
13009 (*anddi_zesidi_di): Likewise.
13010 (*anddi_sesdi_di): Likewise.
13011 (*ne_zeroextracts): Likewise.
13012 (*ne_zeroextracts): Likewise.
13013 (*ite_ne_zeroextr): Likewise.
13014 (*ite_ne_zeroextr): Likewise.
13015 (*anddi_notdi_di): Likewise.
13016 (*anddi_notzesidi): Likewise.
13017 (*anddi_notsesidi): Likewise.
13018 (andsi_notsi_si): Likewise.
13019 (thumb1_bicsi3): Likewise.
13020 (*iordi3_insn): Likewise.
13021 (*iordi_zesidi_di): Likewise.
13022 (*iordi_sesidi_di): Likewise.
13023 (*thumb1_iorsi3_insn): Likewise.
13024 (*xordi3_insn): Likewise.
13025 (*xordi_zesidi_di): Likewise.
13026 (*xordi_sesidi_di): Likewise.
13027 (*arm_xorsi3): Likewise.
13028 (*andsi_iorsi3_no): Likewise.
13029 (*smax_0): Likewise.
13030 (*smax_m1): Likewise.
13031 (*arm_smax_insn): Likewise.
13032 (*smin_0): Likewise.
13033 (*arm_smin_insn): Likewise.
13034 (*arm_umaxsi3): Likewise.
13035 (*arm_uminsi3): Likewise.
13036 (*minmax_arithsi): Likewise.
13037 (*minmax_arithsi_): Likewise.
13038 (*satsi_<SAT:code>): Likewise.
13039 (arm_ashldi3_1bit): Likewise.
13040 (arm_ashrdi3_1bit): Likewise.
13041 (arm_lshrdi3_1bit): Likewise.
13042 (*arm_negdi2): Likewise.
13043 (*thumb1_negdi2): Likewise.
13044 (*arm_negsi2): Likewise.
13045 (*thumb1_negsi2): Likewise.
13046 (*negdi_extendsid): Likewise.
13047 (*negdi_zero_extend): Likewise.
13048 (*arm_abssi2): Likewise.
13049 (*thumb1_abssi2): Likewise.
13050 (*arm_neg_abssi2): Likewise.
13051 (*thumb1_neg_abss): Likewise.
13052 (one_cmpldi2): Likewise.
13053 (extend<mode>di2): Likewise.
13054 (*compareqi_eq0): Likewise.
13055 (*arm_extendhisi2addsi): Likewise.
13056 (*arm_movdi): Likewise.
13057 (*thumb1_movdi_insn): Likewise.
13058 (*arm_movt): Likewise.
13059 (*thumb1_movsi_insn): Likewise.
13060 (pic_add_dot_plus_four): Likewise.
13061 (pic_add_dot_plus_eight): Likewise.
13062 (tls_load_dot_plus_eight): Likewise.
13063 (*thumb1_movhi_insn): Likewise.
13064 (*thumb1_movsf_insn): Likewise.
13065 (*movdf_soft_insn): Likewise.
13066 (*thumb_movdf_insn): Likewise.
13067 (cbranchsi4_insn): Likewise.
13068 (cbranchsi4_scratch): Likewise.
13069 (*negated_cbranchsi4): Likewise.
13070 (*tbit_cbranch): Likewise.
13071 (*tlobits_cbranch): Likewise.
13072 (*tstsi3_cbranch): Likewise.
13073 (*cbranchne_decr1): Likewise.
13074 (*addsi3_cbranch): Likewise.
13075 (*addsi3_cbranch_scratch): Likewise.
13076 (*arm_cmpdi_insn): Likewise.
13077 (*arm_cmpdi_unsig): Likewise.
13078 (*arm_cmpdi_zero): Likewise.
13079 (*thumb_cmpdi_zero): Likewise.
13080 (*deleted_compare): Likewise.
13081 (*mov_scc): Likewise.
13082 (*mov_negscc): Likewise.
13083 (*mov_notscc): Likewise.
13084 (*cstoresi_eq0_thumb1_insn): Likewise.
13085 (cstoresi_nltu_thumb1): Likewise.
13086 (cstoresi_ltu_thu): Likewise.
13087 (thumb1_addsi3_addgeu): Likewise.
13088 (*arm_jump): Likewise.
13089 (*thumb_jump): Likewise.
13090 (*check_arch2): Likewise.
13091 (arm_casesi_internal): Likewise.
13092 (thumb1_casesi_dispatch): Likewise.
13093 (*arm_indirect_jump): Likewise.
13094 (*thumb1_indirect_jump): Likewise.
13096 (*and_scc): Likewise.
13097 (*ior_scc): Likewise.
13098 (*compare_scc): Likewise.
13099 (*cond_move): Likewise.
13100 (*cond_arith): Likewise.
13101 (*cond_sub): Likewise.
13102 (*cmp_ite0): Likewise.
13103 (*cmp_ite1): Likewise.
13104 (*cmp_and): Likewise.
13105 (*cmp_ior): Likewise.
13106 (*ior_scc_scc): Likewise.
13107 (*ior_scc_scc_cmp): Likewise.
13108 (*and_scc_scc): Likewise.
13109 (*and_scc_scc_cmp): Likewise.
13110 (*and_scc_scc_nod): Likewise.
13111 (*negscc): Likewise.
13112 (movcond_addsi): Likewise.
13113 (movcond): Likewise.
13114 (*ifcompare_plus_move): Likewise.
13115 (*if_plus_move): Likewise.
13116 (*ifcompare_move_plus): Likewise.
13117 (*if_move_plus): Likewise.
13118 (*ifcompare_arith_arith): Likewise.
13119 (*if_arith_arith): Likewise.
13120 (*ifcompare_arith_move): Likewise.
13121 (*if_arith_move): Likewise.
13122 (*ifcompare_move_arith): Likewise.
13123 (*if_move_arith): Likewise.
13124 (*ifcompare_move_not): Likewise.
13125 (*if_move_not): Likewise.
13126 (*ifcompare_not_move): Likewise.
13127 (*if_not_move): Likewise.
13128 (*ifcompare_shift_move): Likewise.
13129 (*if_shift_move): Likewise.
13130 (*ifcompare_move_shift): Likewise.
13131 (*if_move_shift): Likewise.
13132 (*ifcompare_shift_shift): Likewise.
13133 (*ifcompare_not_arith): Likewise.
13134 (*ifcompare_arith_not): Likewise.
13135 (*if_arith_not): Likewise.
13136 (*ifcompare_neg_move): Likewise.
13137 (*if_neg_move): Likewise.
13138 (*ifcompare_move_neg): Likewise.
13139 (*if_move_neg): Likewise.
13140 (prologue_thumb1_interwork): Likewise.
13141 (*cond_move_not): Likewise.
13142 (*sign_extract_onebit): Likewise.
13143 (*not_signextract_onebit): Likewise.
13144 (stack_tie): Likewise.
13145 (align_4): Likewise.
13146 (align_8): Likewise.
13147 (consttable_end): Likewise.
13148 (consttable_1): Likewise.
13149 (consttable_2): Likewise.
13150 (consttable_4): Likewise.
13151 (consttable_8): Likewise.
13152 (consttable_16): Likewise.
13153 (*thumb1_tablejump): Likewise.
13154 (prefetch): Likewise.
13155 (force_register_use): Likewise.
13156 (thumb_eh_return): Likewise.
13157 (load_tp_hard): Likewise.
13158 (load_tp_soft): Likewise.
13159 (tlscall): Likewise.
13160 (*arm_movtas_ze): Likewise.
13161 (*arm_rev): Likewise.
13162 (*arm_revsh): Likewise.
13163 (*arm_rev16): Likewise.
13164 * config/arm/thumb2.md
13165 (*thumb2_smaxsi3): Likewise.
13166 (*thumb2_sminsi3): Likewise.
13167 (*thumb32_umaxsi3): Likewise.
13168 (*thumb2_uminsi3): Likewise.
13169 (*thumb2_negdi2): Likewise.
13170 (*thumb2_abssi2): Likewise.
13171 (*thumb2_neg_abss): Likewise.
13172 (*thumb2_movsi_insn): Likewise.
13173 (tls_load_dot_plus_four): Likewise.
13174 (*thumb2_movhi_insn): Likewise.
13175 (*thumb2_mov_scc): Likewise.
13176 (*thumb2_mov_negs): Likewise.
13177 (*thumb2_mov_negs): Likewise.
13178 (*thumb2_mov_nots): Likewise.
13179 (*thumb2_mov_nots): Likewise.
13180 (*thumb2_movsicc_): Likewise.
13181 (*thumb2_movsfcc_soft_insn): Likewise.
13182 (*thumb2_indirect_jump): Likewise.
13183 (*thumb2_and_scc): Likewise.
13184 (*thumb2_ior_scc): Likewise.
13185 (*thumb2_ior_scc_strict_it): Likewise.
13186 (*thumb2_cond_move): Likewise.
13187 (*thumb2_cond_arith): Likewise.
13188 (*thumb2_cond_ari): Likewise.
13189 (*thumb2_cond_sub): Likewise.
13190 (*thumb2_negscc): Likewise.
13191 (*thumb2_movcond): Likewise.
13192 (thumb2_casesi_internal): Likewise.
13193 (thumb2_casesi_internal_pic): Likewise.
13194 (*thumb2_alusi3_short): Likewise.
13195 (*thumb2_mov<mode>_shortim): Likewise.
13196 (*thumb2_addsi_short): Likewise.
13197 (*thumb2_subsi_short): Likewise.
13198 (thumb2_addsi3_compare0): Likewise.
13199 (*thumb2_cbz): Likewise.
13200 (*thumb2_cbnz): Likewise.
13201 (*thumb2_one_cmplsi2_short): Likewise.
13202 (*thumb2_negsi2_short): Likewise.
13203 (*orsi_notsi_si): Likewise.
13204 * config/arm/arm1020e.md: Update with new attributes.
13205 * config/arm/arm1026ejs.md: Update with new attributes.
13206 * config/arm/arm1136jfs.md: Update with new attributes.
13207 * config/arm/arm926ejs.md: Update with new attributes.
13208 * config/arm/cortex-a15.md: Update with new attributes.
13209 * config/arm/cortex-a5.md: Update with new attributes.
13210 * config/arm/cortex-a53.md: Update with new attributes.
13211 * config/arm/cortex-a7.md: Update with new attributes.
13212 * config/arm/cortex-a8.md: Update with new attributes.
13213 * config/arm/cortex-a9.md: Update with new attributes.
13214 * config/arm/cortex-m4.md: Update with new attributes.
13215 * config/arm/cortex-r4.md: Update with new attributes.
13216 * config/arm/fa526.md: Update with new attributes.
13217 * config/arm/fa606te.md: Update with new attributes.
13218 * config/arm/fa626te.md: Update with new attributes.
13219 * config/arm/fa726te.md: Update with new attributes.
13221 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
13223 * config/aarch64/aarch64-simd.md
13224 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
13225 <vwx> iterator to ensure correct register choice.
13226 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
13227 (aarch64_sqdmull_n<mode>): Likewise.
13228 (aarch64_sqdmull2_n<mode>_internal): Likewise.
13229 * config/aarch64/arm_neon.h
13230 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
13231 (vml<as><q>_n_<su>16): Likewise.
13232 (vml<as>l_high_lane<q>_<su>16): Likewise.
13233 (vml<as>l_high_n_<su>16): Likewise.
13234 (vml<as>l_lane<q>_<su>16): Likewise.
13235 (vml<as>l_n_<su>16): Likewise.
13236 (vmul<q>_lane<q>_<su>16): Likewise.
13237 (vmul<q>_n_<su>16): Likewise.
13238 (vmull_lane<q>_<su>16): Likewise.
13239 (vmull_n_<su>16): Likewise.
13240 (vmull_high_lane<q>_<su>16): Likewise.
13241 (vmull_high_n_<su>16): Likewise.
13242 (vqrdmulh<q>_n_s16): Likewise.
13244 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
13246 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
13247 have the correct lane parameter.
13249 2013-09-06 Richard Biener <rguenther@suse.de>
13251 * cfganal.c (control_dependences::~control_dependences):
13252 Properly free all of the vector.
13254 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
13257 * config/i386/i386.c (ix86_conditional_register_usage):
13258 Proper initialize extended SSE registers.
13260 2013-09-06 Jan Hubicka <jh@suse.cz>
13262 PR tree-optimization/58311
13263 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
13265 2013-09-06 Jan Hubicka <jh@suse.cz>
13267 * Makefile.in (tree-sra.o): Update dependencies.
13268 * tree-sra.c: Include ipa-utils.h
13269 (scan_function): Use recursive_call_p.
13270 (has_caller_p): New function.
13271 (cgraph_for_node_and_aliases): Count also callers of aliases.
13273 2013-09-06 Jan Hubicka <jh@suse.cz>
13275 PR middle-end/58094
13276 * cgraph.h (symtab_semantically_equivalent_p): Declare.
13277 * tree-tailcall.c: Include ipa-utils.h.
13278 (find_tail_calls): Use it.
13279 * ipa-pure-const.c (check_call): Likewise.
13280 * ipa-utils.c (recursive_call_p): New function.
13281 * ipa-utils.h (recursive_call_p): Dclare.
13282 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
13283 (symtab_semantically_equivalent_p): New function.
13284 * Makefile.in (tree-tailcall.o): Update dependencies.
13286 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
13288 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
13289 non-inlinable part.
13291 2013-09-06 Richard Biener <rguenther@suse.de>
13293 * lto-streamer.h (lto_global_var_decls): Remove.
13294 * Makefile.in (OBJS): Remove lto-symtab.o.
13295 (lto-symtab.o): Remove.
13296 (GTFILES): Remove lto-symtab.c
13297 * lto-symtab.c: Move to lto/
13299 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13301 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
13302 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
13303 (UNSPEC_FPINT_RINT): New constant definitions.
13304 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
13305 definition with 2 attributes.
13306 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
13307 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
13310 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13312 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
13313 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
13314 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
13315 attribute to "z196_alone".
13316 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
13319 2013-09-06 Richard Biener <rguenther@suse.de>
13321 * basic-block.h (class control_dependences): New.
13322 * tree-ssa-dce.c (control_dependence_map): Remove.
13324 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
13325 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
13326 find_pdom, find_control_dependence, find_all_control_dependences):
13328 (mark_control_dependent_edges_necessary,
13329 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
13330 tree_dce_done, perform_tree_ssa_dce): Adjust.
13331 * cfganal.c (set_control_dependence_map_bit,
13332 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
13333 find_all_control_dependences): Move from tree-ssa-dce.c and
13334 implement as methods of control_dependences class.
13335 (control_dependences::control_dependences): New.
13336 (control_dependences::~control_dependences): Likewise.
13337 (control_dependences::get_edges_dependent_on): Likewise.
13338 (control_dependences::get_edge): Likewise.
13340 2013-09-04 Jan Hubicka <jh@suse.cz>
13342 * tree.c (types_same_for_odr): Drop overactive check.
13343 * ipa-devirt.c (hash_type_name): Likewise.
13345 2013-09-04 Jan Hubicka <jh@suse.cz>
13347 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
13348 (analyze_functions): ... here.
13350 2013-09-04 Jan Hubicka <jh@suse.cz>
13352 PR middle-end/58201
13353 * cgraphunit.c (analyze_functions): Clear AUX fields
13354 after processing; initialize assembler name has.
13356 2013-09-05 Jeff Law <law@redhat.com>
13358 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
13359 from thread_around_empty_block. Record threading path into PATH.
13360 Recurse if threading through the initial block is successful.
13361 (thread_across_edge): Corresponding changes to slightly simplify.
13363 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
13365 * config/aarch64/aarch64.md
13366 (type): Remove frecpe, frecps, frecpx.
13367 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
13368 fix to be a TARGET_SIMD instruction.
13369 (aarch64_frecps): Remove.
13370 * config/aarch64/aarch64-simd.md
13371 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
13372 (aarch64_frecps<mode>): Handle all float/vector of float modes.
13374 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
13375 Sofiane Naci <sofiane.naci@arm.com>
13377 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
13378 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
13379 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
13380 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
13381 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
13382 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
13383 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
13384 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
13385 into "clz, "rbit". Rename "shift" to "shift_imm".
13386 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
13387 changes. Update for attribute changes all occurrences of arlo_* and
13389 * config/arm/arm-fixed.md: Update for attribute changes
13390 all occurrences of arlo_* types.
13391 * config/arm/thumb2.md: Update for attribute changes all occurrences
13393 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
13394 (cortexa7_older_only): Likewise.
13395 (cortexa7_younger): Likewise.
13396 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
13397 (1020alu_shift_op): Likewise.
13398 (1020alu_shift_reg_op): Likewise.
13399 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
13400 (alu_shift_op): Likewise.
13401 (alu_shift_reg_op): Likewise.
13402 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
13403 (11_alu_shift_op): Likewise.
13404 (11_alu_shift_reg_op): Likewise.
13405 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
13406 (9_alu_shift_reg_op): Likewise.
13407 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
13409 (cortex_a15_alu_shift): Likewise.
13410 (cortex_a15_alu_shift_reg): Likewise.
13411 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
13413 (cortex_a5_alu_shift): Likewise.
13414 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
13416 (cortex_a53_alu_shift): Likewise.
13417 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
13419 (cortex_a7_alu_reg): Likewise.
13420 (cortex_a7_alu_shift): Likewise.
13421 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
13423 (cortex_a8_alu_shift): Likewise.
13424 (cortex_a8_alu_shift_reg): Likewise.
13425 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
13426 (cortex_a9_dp_shift): Likewise.
13427 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
13429 * config/arm/cortex-r4.md
13430 (cortex_r4_alu): Update for attribute changes.
13431 (cortex_r4_mov): Likewise.
13432 (cortex_r4_alu_shift_reg): Likewise.
13433 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
13434 (526_alu_shift_op): Likewise.
13435 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
13436 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
13437 (626te_alu_shift_op): Likewise.
13438 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
13439 (726te_alu_shift_op): Likewise.
13440 (726te_alu_shift_reg_op): Likewise.
13441 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
13442 (mp626_alu_shift_op): Likewise.
13443 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
13444 (pj4_alu_conds): Likewise.
13445 (pj4_shift): Likewise.
13446 (pj4_shift_conds): Likewise.
13447 (pj4_alu_shift): Likewise.
13448 (pj4_alu_shift_conds): Likewise.
13449 * config/aarch64/aarch64.md: Update for attribute change
13450 all occurrences of arlo_* and shift* types.
13452 2013-09-05 Mike Stump <mikestump@comcast.net>
13454 * tree.h: Move documentation for tree_function_decl to tree-core.h
13455 with the declaration.
13457 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
13460 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
13461 looking for widest mode.
13463 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
13465 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
13467 2013-09-05 Richard Biener <rguenther@suse.de>
13469 PR tree-optimization/58137
13470 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
13471 Do not create vectors of pointers.
13472 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
13473 types for the components of the vector initializer.
13474 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
13475 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
13477 2013-09-05 Martin Jambor <mjambor@suse.cz>
13479 * ipa-prop.c (remove_described_reference): Accept missing references,
13480 return false if that hppens, otherwise return true.
13481 (cgraph_node_for_jfunc): New function.
13482 (try_decrement_rdesc_refcount): Likewise.
13483 (try_make_edge_direct_simple_call): Use them.
13484 (ipa_edge_removal_hook): Remove references from rdescs.
13485 (ipa_edge_duplication_hook): Clone rdescs and their references
13486 when the new edge has the same caller as the old one.
13487 * cgraph.c (cgraph_resolve_speculation): Remove speculative
13488 reference before removing any edges.
13490 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
13492 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
13494 * thumb2.md (thumb2_storewb_parisi): New pattern.
13496 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
13498 * config/aarch64/aarch64-option-extensions.def: Add
13499 AARCH64_OPT_EXTENSION of 'crc'.
13500 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
13501 (AARCH64_ISA_CRC): Ditto.
13502 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
13503 description of the CRC extension.
13505 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
13507 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
13508 * config/rs6000/linux.h: Ditto.
13509 * alpha/linux.h: Ditto.
13510 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
13511 no_c99_libc_has_function.
13512 * config/c6x/uclinux-elf.h: Ditto.
13513 * config/lm32/uclinux-elf.h: Ditto.
13514 * config/m68k/uclinux.h: Ditto.
13515 * config/moxie/uclinux.h: Ditto.
13516 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
13517 (crisv32-*-linux*, cris-*-linux*): Ditto.
13518 * config/bfin/bfin.c: Include "tm_p.h".
13520 2013-09-05 Richard Biener <rguenther@suse.de>
13522 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
13523 check for a definition without a basic-block.
13525 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
13526 Sofiane Naci <sofiane.naci@arm.com>
13528 * config/aarch64/aarch64.md
13529 (*movti_aarch64): Rename r_2_f and f_2_r.
13530 (*movsf_aarch64): Likewise.
13531 (*movdf_aarch64): Likewise.
13532 (*movtf_aarch64): Likewise.
13533 (aarch64_movdi_<mode>low): Likewise.
13534 (aarch64_movdi_<mode>high): Likewise.
13535 (aarch64_mov<mode>high_di): Likewise.
13536 (aarch64_mov<mode>low_di): Likewise.
13537 (aarch64_movtilow_tilow): Likewise.
13538 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
13539 values to config/arm/types.md
13540 (attribute "conds"): Update for attribute change.
13541 (anddi3_insn): Likewise.
13542 (iordi3_insn): Likewise.
13543 (xordi3_insn): Likewise.
13544 (one_cmpldi2): Likewise.
13545 * config/arm/types.md (type): Add Neon types.
13546 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
13547 use "type" attribute.
13548 (movmisalign<mode>_neon_store): Likewise.
13549 (movmisalign<mode>_neon_load): Likewise.
13550 (vec_set<mode>_internal): Likewise.
13551 (vec_setv2di_internal): Likewise.
13552 (vec_extract<mode>): Likewise.
13553 (vec_extractv2di): Likewise.
13554 (add<mode>3_neon): Likewise.
13555 (adddi3_neon): Likewise.
13556 (sub<mode>3_neon): Likewise.
13557 (subdi3_neon): Likewise.
13558 (mul<mode>3_neon): Likewise.
13559 (mul<mode>3add<mode>_neon): Likewise.
13560 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
13561 (fma<VCVTF:mode>4)): Likewise.
13562 (fma<VCVTF:mode>4_intrinsic): Likewise.
13563 (fmsub<VCVTF:mode>4)): Likewise.
13564 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
13565 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
13566 (ior<mode>3): Likewise.
13567 (and<mode>3): Likewise.
13568 (anddi3_neon): Likewise.
13569 (orn<mode>3_neon): Likewise.
13570 (orndi3_neon): Likewise.
13571 (bic<mode>3_neon): Likewise.
13572 (bicdi3_neon): Likewise.
13573 (xor<mode>3): Likewise.
13574 (one_cmpl<mode>2): Likewise.
13575 (abs<mode>2): Likewise.
13576 (neg<mode>2): Likewise.
13577 (umin<mode>3_neon): Likewise.
13578 (umax<mode>3_neon): Likewise.
13579 (smin<mode>3_neon): Likewise.
13580 (smax<mode>3_neon): Likewise.
13581 (vashl<mode>3): Likewise.
13582 (vashr<mode>3_imm): Likewise.
13583 (vlshr<mode>3_imm): Likewise.
13584 (ashl<mode>3_signed): Likewise.
13585 (ashl<mode>3_unsigned): Likewise.
13586 (neon_load_count): Likewise.
13587 (ashldi3_neon_noclobber): Likewise.
13588 (signed_shift_di3_neon): Likewise.
13589 (unsigned_shift_di3_neon): Likewise.
13590 (ashrdi3_neon_imm_noclobber): Likewise.
13591 (lshrdi3_neon_imm_noclobber): Likewise.
13592 (widen_ssum<mode>3): Likewise.
13593 (widen_usum<mode>3): Likewise.
13594 (quad_halves_<code>v4si): Likewise.
13595 (quad_halves_<code>v4sf): Likewise.
13596 (quad_halves_<code>v8hi): Likewise.
13597 (quad_halves_<code>v16qi): Likewise.
13598 (reduc_splus_v2di): Likewise.
13599 (neon_vpadd_internal<mode>): Likewise.
13600 (neon_vpsmin<mode>): Likewise.
13601 (neon_vpsmax<mode>): Likewise.
13602 (neon_vpumin<mode>): Likewise.
13603 (neon_vpumax<mode>): Likewise.
13604 (ss_add<mode>_neon): Likewise.
13605 (us_add<mode>_neon): Likewise.
13606 (ss_sub<mode>_neon): Likewise.
13607 (us_sub<mode>_neon): Likewise.
13608 (neon_vadd<mode>_unspec): Likewise.
13609 (neon_vaddl<mode>): Likewise.
13610 (neon_vaddw<mode>): Likewise.
13611 (neon_vhadd<mode>): Likewise.
13612 (neon_vqadd<mode>): Likewise.
13613 (neon_vaddhn<mode>): Likewise.
13614 (neon_vmul<mode>): Likewise.
13615 (neon_vmla<mode>): Likewise.
13616 (neon_vmlal<mode>): Likewise.
13617 (neon_vmls<mode>): Likewise.
13618 (neon_vmlsl<mode>): Likewise.
13619 (neon_vqdmulh<mode>): Likewise.
13620 (neon_vqdmlal<mode>): Likewise.
13621 (neon_vqdmlsl<mode>): Likewise.
13622 (neon_vmull<mode>): Likewise.
13623 (neon_vqdmull<mode>): Likewise.
13624 (neon_vsub<mode>_unspec): Likewise.
13625 (neon_vsubl<mode>): Likewise.
13626 (neon_vsubw<mode>): Likewise.
13627 (neon_vqsub<mode>): Likewise.
13628 (neon_vhsub<mode>): Likewise.
13629 (neon_vsubhn<mode>): Likewise.
13630 (neon_vceq<mode>): Likewise.
13631 (neon_vcge<mode>): Likewise.
13632 (neon_vcgeu<mode>): Likewise.
13633 (neon_vcgt<mode>): Likewise.
13634 (neon_vcgtu<mode>): Likewise.
13635 (neon_vcle<mode>): Likewise.
13636 (neon_vclt<mode>): Likewise.
13637 (neon_vcage<mode>): Likewise.
13638 (neon_vcagt<mode>): Likewise.
13639 (neon_vtst<mode>): Likewise.
13640 (neon_vabd<mode>): Likewise.
13641 (neon_vabdl<mode>): Likewise.
13642 (neon_vaba<mode>): Likewise.
13643 (neon_vabal<mode>): Likewise.
13644 (neon_vmax<mode>): Likewise.
13645 (neon_vmin<mode>): Likewise.
13646 (neon_vpaddl<mode>): Likewise.
13647 (neon_vpadal<mode>): Likewise.
13648 (neon_vpmax<mode>): Likewise.
13649 (neon_vpmin<mode>): Likewise.
13650 (neon_vrecps<mode>): Likewise.
13651 (neon_vrsqrts<mode>): Likewise.
13652 (neon_vqabs<mode>): Likewise.
13653 (neon_vqneg<mode>): Likewise.
13654 (neon_vcls<mode>): Likewise.
13655 (clz<mode>2): Likewise.
13656 (popcount<mode>2): Likewise.
13657 (neon_vrecpe): Likewise.
13658 (neon_vrsqrte): Likewise.
13659 (neon_vget_lane<mode>_sext_internal): Likewise.
13660 (neon_vget_lane<mode>_zext_internal): Likewise.
13661 (neon_vdup_n<mode>): Likewise.
13662 (neon_vdup_nv2di): Likewise.
13663 (neon_vdpu_lane<mode>_internal): Likewise.
13664 (neon_vswp<mode>): Likewise.
13665 (float<mode><V_cvtto>2): Likewise.
13666 (floatuns<mode><V_cvtto>2): Likewise.
13667 (fix_trunc<mode><V_cvtto>)2): Likewise
13668 (fixuns_trunc<mode><V_cvtto)2): Likewise.
13669 (neon_vcvt<mode>): Likewise.
13670 (neon_vcvtv4sfv4hf): Likewise.
13671 (neon_vcvtv4hfv4sf): Likewise.
13672 (neon_vcvt_n<mode>): Likewise.
13673 (neon_vmovn<mode>): Likewise.
13674 (neon_vqmovn<mode>): Likewise.
13675 (neon_vqmovun<mode>): Likewise.
13676 (neon_vmovl<mode>): Likewise.
13677 (neon_vmul_lane<mode>): Likewise.
13678 (neon_vmull_lane<mode>): Likewise.
13679 (neon_vqdmull_lane<mode>): Likewise.
13680 (neon_vqdmulh_lane<mode>): Likewise.
13681 (neon_vmla_lane<mode>): Likewise.
13682 (neon_vmlal_lane<mode>): Likewise.
13683 (neon_vqdmlal_lane<mode>): Likewise.
13684 (neon_vmls_lane<mode>): Likewise.
13685 (neon_vmlsl_lane<mode>): Likewise.
13686 (neon_vqdmlsl_lane<mode>): Likewise.
13687 (neon_vext<mode>): Likewise.
13688 (neon_vrev64<mode>): Likewise.
13689 (neon_vrev32<mode>): Likewise.
13690 (neon_vrev16<mode>): Likewise.
13691 (neon_vbsl<mode>_internal): Likewise.
13692 (neon_vshl<mode>): Likewise.
13693 (neon_vqshl<mode>): Likewise.
13694 (neon_vshr_n<mode>): Likewise.
13695 (neon_vshrn_n<mode>): Likewise.
13696 (neon_vqshrn_n<mode>): Likewise.
13697 (neon_vqshrun_n<mode>): Likewise.
13698 (neon_vshl_n<mode>): Likewise.
13699 (neon_vqshl_n<mode>): Likewise.
13700 (neon_vqshlu_n<mode>): Likewise.
13701 (neon_vshll_n<mode>): Likewise.
13702 (neon_vsra_n<mode>): Likewise.
13703 (neon_vsri_n<mode>): Likewise.
13704 (neon_vsli_n<mode>): Likewise.
13705 (neon_vtbl1v8qi): Likewise.
13706 (neon_vtbl2v8qi): Likewise.
13707 (neon_vtbl3v8qi): Likewise.
13708 (neon_vtbl4v8qi): Likewise.
13709 (neon_vtbx1v8qi): Likewise.
13710 (neon_vtbx2v8qi): Likewise.
13711 (neon_vtbx3v8qi): Likewise.
13712 (neon_vtbx4v8qi): Likewise.
13713 (neon_vtrn<mode>_internal): Likewise.
13714 (neon_vzip<mode>_internal): Likewise.
13715 (neon_vuzp<mode>_internal): Likewise.
13716 (neon_vld1<mode>): Likewise.
13717 (neon_vld1_lane<mode>): Likewise.
13718 (neon_vld1_dup<mode>): Likewise.
13719 (neon_vld1_dupv2di): Likewise.
13720 (neon_vst1<mode>): Likewise.
13721 (neon_vst1_lane<mode>): Likewise.
13722 (neon_vld2<mode>): Likewise.
13723 (neon_vld2_lane<mode>): Likewise.
13724 (neon_vld2_dup<mode>): Likewise.
13725 (neon_vst2<mode>): Likewise.
13726 (neon_vst2_lane<mode>): Likewise.
13727 (neon_vld3<mode>): Likewise.
13728 (neon_vld3qa<mode>): Likewise.
13729 (neon_vld3qb<mode>): Likewise.
13730 (neon_vld3_lane<mode>): Likewise.
13731 (neon_vld3_dup<mode>): Likewise.
13732 (neon_vst3<mode>): Likewise.
13733 (neon_vst3qa<mode>): Likewise.
13734 (neon_vst3qb<mode>): Likewise.
13735 (neon_vst3_lane<mode>): Likewise.
13736 (neon_vld4<mode>): Likewise.
13737 (neon_vld4qa<mode>): Likewise.
13738 (neon_vld4qb<mode>): Likewise.
13739 (neon_vld4_lane<mode>): Likewise.
13740 (neon_vld4_dup<mode>): Likewise.
13741 (neon_vst4<mode>): Likewise.
13742 (neon_vst4qa<mode>): Likewise.
13743 (neon_vst4qb<mode>): Likewise.
13744 (neon_vst4_lane<mode>): Likewise.
13745 (neon_vec_unpack<US>_lo_<mode>): Likewise.
13746 (neon_vec_unpack<US>_hi_<mode>): Likewise.
13747 (neon_vec_<US>mult_lo_<mode>): Likewise.
13748 (neon_vec_<US>mult_hi_<mode>): Likewise.
13749 (neon_vec_<US>shiftl_<mode>): Likewise.
13750 (neon_unpack<US>_<mode>): Likewise.
13751 (neon_vec_<US>mult_<mode>): Likewise.
13752 (vec_pack_trunc_<mode>): Likewise.
13753 (neon_vec_pack_trunk_<mode>): Likewise.
13754 (neon_vabd<mode>_2): Likewise.
13755 (neon_vabd<mode>_3): Likewise.
13756 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
13757 (thumb2_movsi_vfp): Likewise.
13758 (movdi_vfp): Likewise.
13759 (movdi_vfp_cortexa8): Likewise.
13760 (movhf_vfp_neon): Likewise.
13761 (movhf_vfp): Likewiwse.
13762 (movsf_vfp): Likewiwse.
13763 (thumb2_movsf_vfp): Likewiwse.
13764 (movdf_vfp): Likewise.
13765 (thumb2_movdf_vfp): Likewise.
13766 (movsfcc_vfp): Likewise.
13767 (thumb2_movsfcc_vfp): Likewise.
13768 (movdfcc_vfp): Likewise.
13769 (thumb2_movdfcc_vfp): Likewise.
13770 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
13771 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
13772 (v10_v2c): Likewise.
13773 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
13775 (cortex_a15_neon_int_2): Likewise.
13776 (cortex_a15_neon_int_3): Likewise.
13777 (cortex_a15_neon_int_4): Likewise.
13778 (cortex_a15_neon_int_5): Likewise.
13779 (cortex_a15_neon_vqneg_vqabs): Likewise.
13780 (cortex_a15_neon_vmov): Likewise.
13781 (cortex_a15_neon_vaba): Likewise.
13782 (cortex_a15_neon_vaba_qqq): Likewise.
13783 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
13784 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
13785 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
13786 scalar_64_32_long_scalar): Likewise.
13787 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
13788 (cortex_a15_neon_mla_qqq_8_16): Likewise.
13789 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
13790 lotype_qdd_64_32_long): Likewise.
13791 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
13792 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
13793 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
13794 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
13795 (cortex_a15_neon_shift_1): Likewise.
13796 (cortex_a15_neon_shift_2): Likewise.
13797 (cortex_a15_neon_shift_3): Likewise.
13798 (cortex_a15_neon_vshl_ddd): Likewise.
13799 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
13800 (cortex_a15_neon_vsra_vrsra): Likewise.
13801 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
13802 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
13803 (cortex_a15_neon_fp_vmul_ddd): Likewise.
13804 (cortex_a15_neon_fp_vmul_qqd): Likewise.
13805 (cortex_a15_neon_fp_vmla_ddd): Likewise.
13806 (cortex_a15_neon_fp_vmla_qqq): Likewise.
13807 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
13808 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
13809 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
13810 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
13811 (cortex_a15_neon_bp_simple): Likewise.
13812 (cortex_a15_neon_bp_2cycle): Likewise.
13813 (cortex_a15_neon_bp_3cycle): Likewise.
13814 (cortex_a15_neon_vld1_1_2_regs): Likewise.
13815 (cortex_a15_neon_vld1_3_4_regs): Likewise.
13816 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
13817 (cortex_a15_neon_vld2_4_regs): Likewise.
13818 (cortex_a15_neon_vld3_vld4): Likewise.
13819 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
13820 (cortex_a15_neon_vst1_3_4_regs): Likewise.
13821 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
13822 (cortex_a15_neon_vst3_vst4): Likewise.
13823 (cortex_a15_neon_vld1_vld2_lane): Likewise.
13824 (cortex_a15_neon_vld3_vld4_lane" 10
13825 (cortex_a15_neon_vst1_vst2_lane): Likewise.
13826 (cortex_a15_neon_vst3_vst4_lane): Likewise.
13827 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
13828 (cortex_a15_neon_ldm_2): Likewise.0
13829 (cortex_a15_neon_stm_2): Likewise.
13830 (cortex_a15_neon_mcr): Likewise.
13831 (cortex_a15_neon_mcr_2_mcrr): Likewise.
13832 (cortex_a15_neon_mrc): Likewise.
13833 (cortex_a15_neon_mrrc): Likewise.
13834 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
13836 (cortex_a15_alu_shift): Likewise.
13837 (cortex_a15_alu_shift_reg): Likewise.
13838 (cortex_a15_mult32): Likewise.
13839 (cortex_a15_mult64): Likewise.
13840 (cortex_a15_block): Likewise.
13841 (cortex_a15_branch): Likewise.
13842 (cortex_a15_load1): Likewise.
13843 (cortex_a15_load3): Likewise.
13844 (cortex_a15_store1): Likewise.
13845 (cortex_a15_store3): Likewise.
13846 (cortex_a15_call): Likewise.
13847 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
13848 (cortex_a5_f2r): Likewise.
13849 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
13851 (cortex_a53_f2r): Likewise.
13852 * config/arm/cortex-a7.md
13853 (cortex_a7_branch): Update for attribute change.
13854 (cortex_a7_call): Likewise.
13855 (cortex_a7_alu_imm): Likewise.
13856 (cortex_a7_alu_reg): Likewise.
13857 (cortex_a7_alu_shift): Likewise.
13858 (cortex_a7_mul): Likewise.
13859 (cortex_a7_load1): Likewise.
13860 (cortex_a7_store1): Likewise.
13861 (cortex_a7_load2): Likewise.
13862 (cortex_a7_store2): Likewise.
13863 (cortex_a7_load3): Likewise.
13864 (cortex_a7_store3): Likewise.
13865 (cortex_a7_load4): Likewise.
13866 (cortex_a7_store4): Likewise.
13867 (cortex_a7_fpalu): Likewise.
13868 (cortex_a7_fconst): Likewise.
13869 (cortex_a7_fpmuls): Likewise.
13870 (cortex_a7_neon_mul): Likewise.
13871 (cortex_a7_fpmacs): Likewise.
13872 (cortex_a7_neon_mla: Likewise.
13873 (cortex_a7_fpmuld: Likewise.
13874 (cortex_a7_fpmacd: Likewise.
13875 (cortex_a7_fpfmad: Likewise.
13876 (cortex_a7_fdivs: Likewise.
13877 (cortex_a7_fdivd: Likewise.
13878 (cortex_a7_r2f: Likewise.
13879 (cortex_a7_f2r: Likewise.
13880 (cortex_a7_f_flags: Likewise.
13881 (cortex_a7_f_loads: Likewise.
13882 (cortex_a7_f_loadd: Likewise.
13883 (cortex_a7_f_stores: Likewise.
13884 (cortex_a7_f_stored: Likewise.
13885 (cortex_a7_neon): Likewise.
13886 * config/arm/cortex-a8-neon.md
13887 (cortex_a8_neon_mrc): Update for attribute change.
13888 (cortex_a8_neon_mrrc): Likewise.
13889 (cortex_a8_neon_int_1): Likewise.
13890 (cortex_a8_neon_int_2): Likewise.
13891 (cortex_a8_neon_int_3): Likewise.
13892 (cortex_a8_neon_int_4): Likewise.
13893 (cortex_a8_neon_int_5): Likewise.
13894 (cortex_a8_neon_vqneg_vqabs): Likewise.
13895 (cortex_a8_neon_vmov): Likewise.
13896 (cortex_a8_neon_vaba): Likewise.
13897 (cortex_a8_neon_vaba_qqq): Likewise.
13898 (cortex_a8_neon_vsma): Likewise.
13899 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
13900 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
13901 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
13902 long_scalar): Likewise.
13903 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
13904 (cortex_a8_neon_mla_qqq_8_16): Likewise.
13905 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
13906 long_scalar_qdd_64_32_long): Likewise.
13907 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
13908 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
13909 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
13910 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
13911 (cortex_a8_neon_shift_1): Likewise.
13912 (cortex_a8_neon_shift_2): Likewise.
13913 (cortex_a8_neon_shift_3): Likewise.
13914 (cortex_a8_neon_vshl_ddd): Likewise.
13915 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
13916 (cortex_a8_neon_vsra_vrsra): Likewise.
13917 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
13918 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
13919 (cortex_a8_neon_fp_vsum): Likewise.
13920 (cortex_a8_neon_fp_vmul_ddd): Likewise.
13921 (cortex_a8_neon_fp_vmul_qqd): Likewise.
13922 (cortex_a8_neon_fp_vmla_ddd): Likewise.
13923 (cortex_a8_neon_fp_vmla_qqq): Likewise.
13924 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
13925 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
13926 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
13927 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
13928 (cortex_a8_neon_bp_simple): Likewise.
13929 (cortex_a8_neon_bp_2cycle): Likewise.
13930 (cortex_a8_neon_bp_3cycle): Likewise.
13931 (cortex_a8_neon_ldr): Likewise.
13932 (cortex_a8_neon_str): Likewise.
13933 (cortex_a8_neon_vld1_1_2_regs): Likewise.
13934 (cortex_a8_neon_vld1_3_4_regs): Likewise.
13935 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
13936 (cortex_a8_neon_vld2_4_regs): Likewise.
13937 (cortex_a8_neon_vld3_vld4): Likewise.
13938 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
13939 (cortex_a8_neon_vst1_3_4_regs): Likewise.
13940 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
13941 (cortex_a8_neon_vst3_vst4): Likewise.
13942 (cortex_a8_neon_vld1_vld2_lane): Likewise.
13943 (cortex_a8_neon_vld3_vld4_lane): Likewise.
13944 (cortex_a8_neon_vst1_vst2_lane): Likewise.
13945 (cortex_a8_neon_vst3_vst4_lane): Likewise.
13946 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
13947 (cortex_a8_neon_mcr): Likewise.
13948 (cortex_a8_neon_mcr_2_mcrr): Likewise.
13949 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
13950 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
13952 (ca9_neon_mrrc): Likewise.
13953 (cortex_a9_neon_int_1): Likewise.
13954 (cortex_a9_neon_int_2): Likewise.
13955 (cortex_a9_neon_int_3): Likewise.
13956 (cortex_a9_neon_int_4): Likewise.
13957 (cortex_a9_neon_int_5): Likewise.
13958 (cortex_a9_neon_vqneg_vqabs): Likewise.
13959 (cortex_a9_neon_vmov): Likewise.
13960 (cortex_a9_neon_vaba): Likewise.
13961 (cortex_a9_neon_vaba_qqq): Likewise.
13962 (cortex_a9_neon_vsma): Likewise.
13963 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
13964 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
13965 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
13966 long_scalar): Likewise.
13967 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
13968 (cortex_a9_neon_mla_qqq_8_16): Likewise.
13969 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
13970 long_scalar_qdd_64_32_long): Likewise.
13971 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
13972 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
13973 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
13974 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
13975 (cortex_a9_neon_shift_1): Likewise.
13976 (cortex_a9_neon_shift_2): Likewise.
13977 (cortex_a9_neon_shift_3): Likewise.
13978 (cortex_a9_neon_vshl_ddd): Likewise.
13979 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
13980 (cortex_a9_neon_vsra_vrsra): Likewise.
13981 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
13982 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
13983 (cortex_a9_neon_fp_vsum): Likewise.
13984 (cortex_a9_neon_fp_vmul_ddd): Likewise.
13985 (cortex_a9_neon_fp_vmul_qqd): Likewise.
13986 (cortex_a9_neon_fp_vmla_ddd): Likewise.
13987 (cortex_a9_neon_fp_vmla_qqq): Likewise.
13988 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
13989 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
13990 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
13991 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
13992 (cortex_a9_neon_bp_simple): Likewise.
13993 (cortex_a9_neon_bp_2cycle): Likewise.
13994 (cortex_a9_neon_bp_3cycle): Likewise.
13995 (cortex_a9_neon_ldr): Likewise.
13996 (cortex_a9_neon_str): Likewise.
13997 (cortex_a9_neon_vld1_1_2_regs): Likewise.
13998 (cortex_a9_neon_vld1_3_4_regs): Likewise.
13999 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
14000 (cortex_a9_neon_vld2_4_regs): Likewise.
14001 (cortex_a9_neon_vld3_vld4): Likewise.
14002 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
14003 (cortex_a9_neon_vst1_3_4_regs): Likewise.
14004 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
14005 (cortex_a9_neon_vst3_vst4): Likewise.
14006 (cortex_a9_neon_vld1_vld2_lane): Likewise.
14007 (cortex_a9_neon_vld3_vld4_lane): Likewise.
14008 (cortex_a9_neon_vst1_vst2_lane): Likewise.
14009 (cortex_a9_neon_vst3_vst4_lane): Likewise.
14010 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
14011 (cortex_a9_neon_mcr): Likewise.
14012 (cortex_a9_neon_mcr_2_mcrr): Likewise.
14013 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
14014 (cortex_a9_fps): Likewise.
14015 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
14017 (cortex_m4_fmuls): Likewise.
14018 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
14020 (cortex_r4_mrc): Likewise.
14021 * config/arm/iterators.md: Update comment referring to neon_type.
14022 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
14023 (iwmmxt_movsi_insn): Likewise.
14024 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
14026 (pj4_core_to_vfp): Likewise.
14027 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
14029 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
14030 (vfp_fstore): Likewise.
14031 * doc/md.texi: Change references to neon_type to refer to type.
14033 2013-09-04 Dodji Seketeli <dodji@redhat.com>
14035 * tree.h (DECL_BUILT_IN): Fix typo in comment.
14037 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
14039 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
14040 lglobl if not weak.
14042 2013-09-04 Easwaran Raman <eraman@google.com>
14044 PR middle-end/57370
14045 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
14046 (build_and_add_sum): Use it.
14047 (appears_later_in_bb): Simplify code.
14049 2013-09-04 Teresa Johnson <tejohnson@google.com>
14051 * dumpfile.c (dump_finish): Don't close stderr/stdout.
14053 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
14055 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
14057 2013-09-04 Jan Hubicka <jh@suse.cz>
14059 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
14060 * ipa-devirt.c: Include diganostic.h
14061 (odr_type_d): Add types and types_set.
14062 (hash_type_name): Work for types with vtables during LTO.
14063 (odr_hasher::remove): Fix comment; destroy types_set.
14064 (add_type_duplicate): New function,
14065 (get_odr_type): Use it.
14066 (dump_type_inheritance_graph): Dump type duplicates.
14067 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
14069 * tree.c (types_same_for_odr): Give exact answers on types with
14072 2013-09-04 Dodji Seketeli <dodji@redhat.com>
14074 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
14075 explaining their differences.
14077 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
14079 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
14081 2013-09-03 Jeff Law <law@redhat.com>
14083 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
14084 when not threading through a joiner block. Pass joiner/no joiner
14085 state to register_jump_thread.
14086 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
14087 state from argument rather than implying on path length.
14088 Dump the entire jump thread path into debugging dump.
14089 * tree-flow.h (register_jump_thread): Update prototype.
14091 2013-08-29 Xinliang David Li <davidxl@google.com>
14093 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
14094 Remove a trivial gcc_assert.
14096 2013-08-29 Xinliang David Li <davidxl@google.com>
14098 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
14099 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
14100 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
14101 Delay base decl alignment adjustment.
14102 * tree-vectorizer.c (vect_destroy_datarefs): New function.
14103 * tree-vectorizer.h: New data structure.
14104 (set_dr_misalignment): New function.
14105 (dr_misalignment): Ditto.
14106 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
14107 (vectorizable_load): Ditto.
14108 (ensure_base_align): New function.
14109 (vectorize_loops): Add dbg_cnt support.
14110 (execute_vect_slp): Ditto.
14111 * dbgcnt.def: New debug counter.
14112 * Makefile: New dependency.
14114 2013-09-03 Meador Inge <meadori@codesourcery.com>
14118 2013-08-30 Meador Inge <meadori@codesourcery.com>
14120 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
14122 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
14124 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
14125 function descriptor.
14127 2013-09-03 Richard Biener <rguenther@suse.de>
14129 * tree-affine.c (add_elt_to_tree): Fix association issue,
14130 avoid useless converts and make sure to always return a
14131 properly typed result.
14133 2013-09-03 Richard Biener <rguenther@suse.de>
14135 PR middle-end/57656
14136 * fold-const.c (negate_expr_p): Fix division case.
14137 (negate_expr): Likewise.
14139 2013-09-03 Richard Biener <rguenther@suse.de>
14142 * tree-streamer-out.c: Include tm.h.
14143 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
14145 2013-09-03 Jan Hubicka <jh@suse.cz>
14147 * tree-profile.c (tree_profiling): Cleanup CFG when done.
14149 2013-09-03 Alan Modra <amodra@gmail.com>
14151 * config.gcc (powerpc*-*-linux*): Add support for little-endian
14152 multilibs to big-endian target and vice versa.
14153 * config/rs6000/t-linux64: Use := assignment on all vars.
14154 (MULTILIB_EXTRA_OPTS): Remove fPIC.
14155 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
14156 * config/rs6000/t-linux64le: New file.
14157 * config/rs6000/t-linux64bele: New file.
14158 * config/rs6000/t-linux64lebe: New file.
14160 2013-09-02 Jan Hubicka <jh@suse.cz>
14162 * ipa-inline-transform.c (inline_transform): Do not
14163 optimize_inline_calls when not optimizing.
14165 2013-09-02 Jan Hubicka <jh@suse.cz>
14167 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
14168 duplicated nodes for assembler names.
14169 * symtab.c (symtab_unregister_node): Do not attempt to unlink
14170 hard registers from assembler name hash.
14172 2013-09-02 Jan Hubicka <jh@suse.cz>
14174 * ipa-split.c (execute_split_functions): Split externally visible
14175 functions called once.
14177 2013-09-02 Martin Jambor <mjambor@suse.cz>
14180 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
14181 linked list. When finding the correct duplicate, also consider also
14182 the caller in additon to its inlined_to node.
14184 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
14186 * config/aarch64/aarch64-simd-builtins.def
14187 (dup_lane_scalar): Remove.
14188 * config/aarch64/aarch64-simd.md
14189 (aarch64_simd_dup): Add 'w->w' alternative.
14190 (aarch64_dup_lane<mode>): Allow for VALL.
14191 (aarch64_dup_lane_scalar<mode>): Remove.
14192 (aarch64_dup_lane_<vswap_width_name><mode>): New.
14193 (aarch64_get_lane_signed<mode>): Add w->w altenative.
14194 (aarch64_get_lane_unsigned<mode>): Likewise.
14195 (aarch64_get_lane<mode>): Likewise.
14196 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
14197 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
14198 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
14199 (VCON): Change container of V2SF.
14200 (vswap_width_name): Likewise.
14201 * config/aarch64/arm_neon.h
14202 (__aarch64_vdup_lane_any): New.
14203 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
14204 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
14205 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
14207 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
14209 PR middle-end/56382
14210 * expr.c (emit_move_complex): Do not move complex FP values as parts if
14211 the source or the destination is a single hard register.
14213 2013-09-02 Richard Biener <rguenther@suse.de>
14215 PR middle-end/57511
14216 * tree-scalar-evolution.c (instantiate_scev_name): Allow
14219 2013-09-02 Richard Biener <rguenther@suse.de>
14221 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
14222 arithmetic to sizetype.
14224 2013-09-02 Bin Cheng <bin.cheng@arm.com>
14226 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
14227 Find auto-increment use both before and after candidate.
14229 2013-09-02 Marek Polacek <polacek@redhat.com>
14231 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
14233 2013-09-01 Jan Hubicka <jh@suse.cz>
14235 * Makefile.in: Add ipa-profile.o
14236 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
14237 * cgraph.c (struct cgraph_propagate_frequency_data,
14238 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
14239 ipa-profile.c; replace cgraph_ by ipa_ prefix.
14240 * cgraph.h (cgraph_propagate_frequency): Remove.
14241 * ipa-inline-analysis.c: Include ipa-utils.h;
14242 drop duplicated cfgloop.h.
14243 (inline_update_callee_summaries): Update.
14244 * ipa-profile.c: New file.
14245 * ipa-utils.h (ipa_propagate_frequency): Declare.
14246 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
14247 data-streamer.h, value-prof.h.
14248 (symtab_remove_unreachable_nodes): Update profile.
14249 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
14250 account_time_size, cmp_counts, dump_histogram,
14251 ipa_profile_generate_summary, ipa_profile_write_summary,
14252 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
14253 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
14254 Move to ipa-profile.c.
14256 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
14258 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
14260 2013-09-01 Jan Hubicka <jh@suse.cz>
14262 * common.opt (fdevirtualize-speculatively): New function.
14263 * invoke.texi (fdevirtualize-speculatively): Document.
14264 * ipa-devirt.c: Include ipa-inline.h
14265 (likely_target_p): New function.
14266 (ipa_devirt): New function.
14267 (gate_ipa_devirt): New function.
14268 (pass_data_ipa_devirt): New static var.
14269 (pass_ipa_devirt): Likewise.
14270 (make_pass_ipa_devirt): New function.
14271 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
14272 (common_handle_option): Disable devirtualization when
14273 value range profiling is available.
14274 * passes.def (pass_ipa_devirt): Add.
14275 * timever.def (TV_IPA_DEVIRT): New timevar.
14276 * tree-pass.h (make_pass_ipa_devirt):
14278 2013-09-01 Iain Sandoe <iain@codesourcery.com>
14280 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
14281 include sanitize(undefined).
14283 2013-08-31 Diego Novillo <dnovillo@google.com>
14285 * Makefile.in (TREE_CORE_H): Define.
14287 (GTFILES): Add tree-core.h.
14288 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
14290 * tree-core.h: New file.
14291 Move all data structures, enum, typedefs, global
14292 declarations and constants from ...
14293 * tree.h: ... here.
14295 2013-08-31 Jan Hubicka <jh@suse.cz>
14297 * bulitins.c (expand_builtin): Do not early exit for gcov
14298 instrumented functions.
14300 2013-08-31 Marek Polacek <polacek@redhat.com>
14302 * ubsan.c: Include tm_p.h.
14304 2013-08-31 Jan Hubicka <jh@suse.cz>
14306 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
14309 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
14310 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
14311 * tree-cfg.c (verify_gimple_label): ... here.
14312 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
14313 (ipa_merge_profiles): New function.
14314 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
14315 (lto_input_function_body): Likewise.
14316 * ipa-utils.h (ipa_merge_profiles): Declare.
14317 * lto-streamer.h (lto_input_function_body): Update prototype.
14318 (emit_label_in_global_context_p): Remove.
14319 * lto-symtab.c: Include ipa-utils.h
14320 (lto_cgraph_replace_node): Use ipa_merge_profiles.
14322 2013-08-31 Jan Hubicka <jh@suse.cz>
14324 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
14326 2013-08-31 Jan Hubicka <jh@suse.cz>
14328 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
14330 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
14332 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
14333 "cmp" RTX before signed_comparison_operator check to account
14334 for "code" changes.
14336 2013-08-30 Jan Hubicka <jh@suse.cz>
14338 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
14339 (detect_type_change_1): Rename to ...
14340 (detect_type_change): ... this one; early return on non-polymorphic
14342 (detect_type_change_ssa): Add comp_type parameter; update
14343 use of detect_type_change.
14344 (compute_complex_assign_jump_func): Add param_type parameter;
14345 update use of detect_type_change_ssa.
14346 (compute_complex_ancestor_jump_func): Likewise.
14347 (ipa_get_callee_param_type): New function.
14348 (ipa_compute_jump_functions_for_edge): Compute parameter type;
14349 update calls to the jump function computation functions.
14351 2013-08-30 Teresa Johnson <tejohnson@google.com>
14352 Steven Bosscher <steven@gcc.gnu.org>
14354 * cfgrtl.c (fixup_new_cold_bb): New routine.
14355 (commit_edge_insertions): Invoke fixup_partitions.
14356 (find_partition_fixes): New routine.
14357 (fixup_partitions): Ditto.
14358 (verify_hot_cold_block_grouping): Update comments.
14359 (rtl_verify_edges): Invoke find_partition_fixes.
14360 (rtl_verify_bb_pointers): Update comments.
14361 (rtl_verify_bb_layout): Ditto.
14362 * basic-block.h (probably_never_executed_edge_p): Declare.
14363 (fixup_partitions): Ditto.
14364 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
14365 * bb-reorder.c (sanitize_hot_paths): New function.
14366 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
14367 sanitize_hot_paths.
14368 * predict.c (probably_never_executed_edge_p): New routine.
14369 * cfg.c (check_bb_profile): Add partition insanity warnings.
14371 2013-08-30 Meador Inge <meadori@codesourcery.com>
14373 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
14375 2013-08-30 Marek Polacek <polacek@redhat.com>
14377 * Makefile.in (ubsan.o): Add.
14378 (c-family/c-ubsan.o): Add.
14379 (builtins.o): Add ubsan.h dependency.
14380 * ubsan.h: New file.
14381 * ubsan.c: New file.
14382 * common.opt: Add -fsanitize=undefined option.
14383 (flag_sanitize): Add variable.
14384 (fsanitize=): Add option. Add Driver.
14385 (fsanitize=thread): Remove option.
14386 (fsanitize=address): Likewise.
14387 (static-libubsan): New option.
14388 * doc/invoke.texi: Document the new flag and -static-libubsan.
14389 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
14390 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
14391 * builtin-attrs.def (ATTR_COLD): Define.
14392 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
14393 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
14394 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
14395 * flag-types.h (sanitize_code): New enum.
14396 * opts.c (common_handle_option): Parse command line arguments
14397 of -fsanitize=. Add -fsanitize=unreachable option.
14398 * varasm.c (get_variable_section): Adjust.
14399 (assemble_noswitch_variable): Likewise.
14400 (assemble_variable): Likewise.
14401 (output_constant_def_contents): Likewise.
14402 (categorize_decl_for_section): Likewise.
14403 (place_block_symbol): Likewise.
14404 (output_object_block): Likewise.
14405 * builtins.def: Likewise.
14406 * toplev.c (compile_file): Likewise.
14407 (process_options): Likewise.
14408 * cppbuiltin.c: Likewise.
14409 * tsan.c (tsan_pass): Likewise.
14410 (tsan_gate): Likewise.
14411 (tsan_gate_O0): Likewise.
14412 * cfgexpand.c (partition_stack_vars): Likewise.
14413 (expand_stack_vars): Likewise.
14414 (defer_stack_allocation): Likewise.
14415 (expand_used_vars): Likewise.
14416 * cfgcleanup.c (old_insns_match_p): Likewise.
14417 * asan.c (asan_finish_file): Likewise.
14418 (asan_instrument): Likewise.
14419 (gate_asan): Likewise.
14420 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
14421 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
14422 (asan_global_struct): Use pointer_sized_int_node instead
14423 calling build_nonstandard_integer_type.
14424 (initialize_sanitizer_builtins): Likewise.
14425 (asan_finish_file): Likewise.
14426 * gcc.c: Document %{%:function(args):X}.
14427 (static_spec_functions): Add sanitize.
14428 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
14429 store funcval != NULL there.
14430 (do_spec_1): Adjust handle_spec_function caller.
14431 (handle_braces): Allow %:function(args) as condition.
14432 (sanitize_spec_function): New function.
14433 (ADD_STATIC_LIBUBSAN_LIBS): Define.
14434 (LIBUBSAN_SPEC): Likewise.
14435 (LIBUBSAN_EARLY_SPEC): Likewise.
14436 (SANITIZER_SPEC): Handle libubsan.
14437 (SANITIZER_EARLY_SPEC): Likewise.
14438 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
14439 instead of fsanitize=address.
14440 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
14441 instead of fsanitize=address*.
14442 * builtins.c: Include ubsan.h.
14443 (fold_builtin_0): Instrument __builtin_unreachable.
14444 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
14445 instead of flag_asan.
14446 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
14447 (pointer_sized_int_node): Define.
14448 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
14450 2013-08-30 Mike Stump <mikestump@comcast.net>
14452 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
14455 2013-08-29 Jan Hubicka <jh@suse.cz>
14457 * cgraph.c (cgraph_function_body_availability): Handle weakref
14459 * passes.def: Remove pass_fixup_cfg.
14460 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
14461 track when we need to remove functions.
14462 (gate_ipa_inline): Execute inlining always; add comment why.
14463 (pass_data_ipa_inline): Remove TODO_remove_functions.
14464 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
14465 do not produce summaries.
14466 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
14467 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
14468 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
14469 constant pool and vtable.
14471 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
14473 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
14474 New arm_neon.h's internal macros to specify 64-bit constants.
14475 Avoid using stdint.h's macros.
14477 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
14479 * recog.c (verify_changes): Verify that changes[i].old is non-zero
14480 before applying REG_P.
14482 2013-08-30 Jakub Jelinek <jakub@redhat.com>
14484 PR tree-optimization/58277
14485 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
14486 after seeing too many stmts with vdef in between dombb and current
14487 bb, invalidate everything.
14489 2013-08-30 Richard Biener <rguenther@suse.de>
14491 * fold-const.c (fold_single_bit_test): Fix overflow test.
14493 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
14495 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
14496 and which can live in a register, always retrieve the value on entry.
14497 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
14498 passed by invisible reference specially.
14499 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
14500 (vt_add_function_parameter): Correctly deal with a parameter passed by
14501 invisible reference.
14503 2013-08-30 Jan Hubicka <jh@suse.cz>
14505 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
14507 2013-08-30 Richard Biener <rguenther@suse.de>
14509 PR tree-optimization/58228
14510 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
14511 allow invariant loads in nested loop vectorization.
14513 2013-08-30 Richard Biener <rguenther@suse.de>
14515 PR tree-optimization/58223
14516 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
14517 (has_anti_or_output_dependence): ... this and adjust to also
14518 look for output dependences.
14519 (mark_nodes_having_upstream_mem_writes): Adjust.
14520 (rdg_flag_uses): Likewise.
14522 2013-08-30 Richard Biener <rguenther@suse.de>
14524 PR tree-optimization/58010
14525 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
14526 assert that we have a loop-closed PHI.
14528 2013-08-29 Jan Hubicka <jh@suse.cz>
14530 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
14531 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
14532 * lto-section-in.c (lto_free_function_in_decl_state): New function.
14533 (lto_free_function_in_decl_state_for_node): New function.
14535 2013-08-29 Xinliang David Li <davidxl@google.com>
14537 * loop-unroll.c (report_unroll_peel): Minor message change.
14538 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
14539 Emit alignment peeling message with default -fopt-info.
14540 (vect_loop_versioning): Emit loop version info message.
14541 * tree-vectorizer.c (vectorize_loops): Minor message change.
14542 (execute_vect_slp): Ditto.
14544 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
14546 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
14547 of the clone from the DECL_NAME of the original function.
14549 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
14551 * passes.c (register_pass): Add overload.
14552 * tree-pass.h (register_pass): Forward declare it. Add comment.
14554 2013-08-29 Jan Hubicka <jh@suse.cz>
14556 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
14557 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
14558 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
14559 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
14560 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
14561 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
14562 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
14563 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
14565 2013-08-29 Teresa Johnson <tejohnson@google.com>
14567 * dumpfile.c (dump_loc): Output column number.
14568 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
14569 * doc/invoke.texi: Document optall -fopt-info flag.
14570 * profile.c (read_profile_edge_counts): Use new dump framework.
14571 (compute_branch_probabilities): Ditto.
14572 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
14573 when pass not in any opt group.
14574 * pass_manager.h (pass_manager::get_pass_profile): New method.
14575 * value-prof.c (check_counter): Use new dump framework.
14576 (check_ic_target): Ditto.
14577 * coverage.c (get_coverage_counts): Ditto.
14578 (coverage_init): Setup new dump framework.
14580 2013-08-29 Richard Biener <rguenther@suse.de>
14582 PR tree-optimization/58246
14583 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
14584 handle the dominance check inside a basic-block.
14586 2013-08-29 Richard Biener <rguenther@suse.de>
14588 PR middle-end/57287
14589 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
14590 of default defs that appear in abnormal PHI nodes.
14592 2013-08-29 Richard Biener <rguenther@suse.de>
14594 PR tree-optimization/57685
14595 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
14596 single-use operands to avoid exponential complexity.
14598 2013-08-28 Dehao Chen <dehao@google.com>
14600 * ipa-inline.c (edge_badness): Fix integer underflow.
14602 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
14604 * gtm-builtins.def (_ITM_free): Declare leaf.
14606 2013-08-28 Jakub Jelinek <jakub@redhat.com>
14609 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
14610 (*tls_local_dynamic_base_64_largepic): Likewise.
14611 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
14612 Remove predicate from call operand.
14613 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
14614 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
14616 2013-08-28 Jeff Law <law@redhat.com>
14618 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
14619 checks for the number of predecessors and successors allowed.
14620 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
14621 which require copying a joiner block if there is a request which
14622 is a subpath that requires no joiner block copying.
14624 2013-08-28 Jan Hubicka <jh@suse.cz>
14626 * lto-streamer-out.c (DFS_write_tree_body): Drop
14627 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
14628 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
14629 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
14630 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
14631 Do not read DECL_ERROR_ISSUED.
14632 (unpack_ts_decl_with_vis_value_fields): Do not read
14634 (lto_input_ts_binfo_tree_pointers): Do not read
14635 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
14636 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
14637 write DECL_ERROR_ISSUED..
14638 (pack_ts_decl_with_vis_value_fields): Do not write
14640 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
14641 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
14642 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
14643 * tree.h (tree_decl_common): Update comment.
14644 (DECL_ERROR_ISSUED): Remove.
14646 2013-08-28 Jakub Jelinek <jakub@redhat.com>
14648 PR middle-end/58257
14649 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
14651 2013-08-28 Jan Hubicka <jh@suse.cz>
14653 * builtins.def (free): Declare leaf.
14655 2013-08-27 David Malcolm <dmalcolm@redhat.com>
14657 * gdbhooks.py: New.
14658 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
14659 * configure: Regenerate.
14661 2013-08-27 Martin Jambor <mjambor@suse.cz>
14663 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
14664 (ipa_ancestor_jf_data): Likewise.
14665 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
14666 (ipa_get_jf_pass_through_type_preserved): New function.
14667 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
14668 (ipa_get_jf_ancestor_type_preserved): New function.
14669 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
14670 (ipa_get_jf_ancestor_result): Likewise.
14671 (propagate_vals_accross_pass_through): Use
14672 ipa_get_jf_pass_through_result to do all the value mappings.
14673 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
14674 type_preserved flag.
14675 (ipa_set_jf_cst_copy): New function.
14676 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
14677 (ipa_set_jf_arith_pass_through): Likewise.
14678 (ipa_set_ancestor_jf): Likewise.
14679 (compute_complex_assign_jump_func): Set type_preserved instead of
14681 (ipa_compute_jump_functions_for_edge): Likewise.
14682 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
14683 (update_jump_functions_after_inlining): Update type_preserved.
14684 Explicitely create jump functions when combining one with pass_through.
14685 (ipa_write_jump_function): Stream the type_preserved flags.
14686 (ipa_read_jump_function): Likewise.
14688 2013-08-27 Jakub Jelinek <jakub@redhat.com>
14689 Aldy Hernandez <aldyh@redhat.com>
14691 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
14692 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
14693 * function.h (struct function): Add has_force_vect_loops and
14695 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
14696 * gimple.c (gimple_build_omp_critical): Add KIND argument and
14698 * gimple.def: Update CLAUSES comments.
14699 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
14700 (gimple_build_omp_for): Add argument to prototype.
14701 (gimple_omp_for_kind): New.
14702 (gimple_omp_for_set_kind): New.
14703 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
14704 GOVD_DATA_SHARE_CLASS.
14705 (enum omp_region_type): Add ORT_SIMD.
14706 (gimple_add_tmp_var): Handle ORT_SIMD.
14707 (gimplify_var_or_parm_decl): Same.
14708 (is_gimple_stmt): Same.
14709 (omp_firstprivatize_variable): Same.
14710 (omp_add_variable): Only use splay_tree_insert if lookup failed.
14711 (omp_notice_variable): Handle ORT_SIMD.
14712 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
14713 (omp_check_private): Handle ORT_SIMD.
14714 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
14715 OMP_CLAUSE_SAFELEN.
14716 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
14717 Handle OMP_CLAUSE_LASTPRIVATE.
14718 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
14719 OMP_CLAUSE_SAFELEN.
14720 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
14721 (gimplify_expr): Handle OMP_SIMD.
14722 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
14723 (expand_GOMP_SIMD_VF): New.
14724 (expand_GOMP_SIMD_LAST_LANE): New.
14725 * internal-fn.def (GOMP_SIMD_LANE): New.
14726 (GOMP_SIMD_VF): New.
14727 (GOMP_SIMD_LAST_LANE): New.
14728 * omp-low.c: Include target.h.
14729 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
14730 OMP_CLAUSE_SAFELEN.
14731 (check_omp_nesting_restrictions): Same.
14733 (lower_rec_simd_input_clauses): New.
14734 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
14736 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
14737 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
14738 (expand_omp_build_assign): New.
14739 (expand_omp_for_init_counts): New.
14740 (expand_omp_for_init_vars): New.
14741 (extract_omp_for_update_vars): New.
14742 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
14743 and rewrite accordingly.
14744 (expand_omp_simd): New.
14745 (expand_omp_for): Use expand_omp_simd.
14746 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
14747 (lower_omp_for): Do not lower the body.
14748 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
14749 in their own loops.
14750 * tree-flow.h (find_omp_clause): Remove prototype.
14751 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
14752 forcing vectorization of the loop, or if flag_tree_vectorize.
14753 (gate_tree_if_conversion): Similarly.
14754 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
14755 gimple_build_omp_for.
14756 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
14757 * tree-parloops (create_parallel_loop): Pass kind argument to
14758 gimple_build_omp_for.
14759 * tree-pretty-print.c (dump_omp_clause): Add cases for
14760 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
14761 OMP_CLAUSE__SIMDUID_.
14762 (dump_generic_node): Handle OMP_SIMD.
14763 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
14764 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
14765 unroll OMP_SIMD loops here.
14766 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
14767 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
14769 (vect_analyze_data_refs): Handle simd loops.
14770 * tree-vect-loop.c (vectorizable_live_operation): Handle
14772 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
14773 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
14774 (vectorizable_load): Same.
14775 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
14776 (struct simduid_to_vf): New.
14777 (simduid_to_vf::hash): New.
14778 (simduid_to-vf::equal): New.
14779 (struct simd_array_to_simduid): New.
14780 (simd_array_to_simduid::hash): New.
14781 (simd_array_to_simduid::equal): New.
14782 (adjust_simduid_builtins): New.
14783 (struct note_simd_array_uses_struct): New.
14784 (note_simd_array_uses_cb): New.
14785 (note_simd_array_uses): New.
14786 (vectorize_loops): Handle simd hints and adjust simd builtins
14788 * tree-vectorizer.h (struct _stmt_vec_info): Add
14789 simd_lane_access_p field.
14790 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
14791 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
14792 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
14793 (omp_clause_code_name): Same.
14794 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
14795 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
14796 * tree.def (OMP_SIMD): New entry.
14797 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
14798 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
14799 (OMP_CLAUSE_DECL): Adjust range for new clauses.
14800 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
14801 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
14802 (OMP_CLAUSE_LINEAR_STEP): New.
14803 (OMP_CLAUSE_SAFELEN_EXPR): New.
14804 (OMP_CLAUSE__SIMDUID__DECL): New.
14805 (find_omp_clause): New prototype.
14807 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
14809 * config/i386/driver-i386.c (host_detect_local_cpu): Update
14810 Haswell processor detection.
14812 2013-08-27 Christian Widmer <shadow@umbrox.de>
14815 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
14816 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
14817 AVX2 capable processors.
14819 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
14821 * config/aarch64/arm_neon.h: Replace all inline asm implementations
14822 of vget_low_* with implementations in terms of other intrinsics.
14824 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
14826 PR middle-end/57219
14827 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
14828 values to -1, 0 and 1.
14830 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
14832 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
14833 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
14834 (<optab><mode>3_insn): Remove.
14835 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
14836 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
14837 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
14838 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
14839 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
14840 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
14841 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
14842 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
14843 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
14844 (ror<mode>3_insn): Likewise.
14845 * config/aarch64/predicates.md (aarch64_simd_register): New.
14847 2013-08-27 Richard Biener <rguenther@suse.de>
14849 PR tree-optimization/57521
14850 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
14851 one edge is non-critical.
14852 (find_phi_replacement_condition): Make sure to use a non-critical
14853 edge. Cleanup and remove old bug workarounds.
14854 (bb_postdominates_preds): Remove.
14855 (if_convertible_loop_p_1): Do not compute post-dominators.
14856 (combine_blocks): Do not free post-dominators.
14857 (main_tree_if_conversion): Likewise.
14858 (pass_data_if_conversion): Add TODO_verify_ssa.
14860 2013-08-27 DJ Delorie <dj@redhat.com>
14862 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
14864 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
14866 * function.c (assign_parm_find_data_types): Set passed_mode and
14867 nominal_mode to the TYPE_MODE of nominal_type for the built
14868 pointer type in case of the struct-pass-by-reference.
14870 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
14872 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
14873 (UINT16_TYPE): Change default to "unsigned int".
14875 * config/avr/avr.opt (mfract-convert-truncate): New option.
14876 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
14877 is set, round negative fractional integers according to n1169
14878 when converting to integer types.
14880 2013-08-26 Jan Hubicka <jh@suse.cz>
14882 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
14883 methods can not be called indirectly when their address is not taken.
14885 2013-08-26 Jan Hubicka <jh@suse.cz>
14887 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
14890 2013-08-26 Jan Hubicka <jh@suse.cz>
14892 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
14895 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
14897 * reload.c (find_valid_class): Allow classes that do not include
14898 FIRST_PSEUDO_REGISTER - 1.
14900 2013-08-26 Jan Hubicka <jh@suse.cz>
14902 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
14903 fix edge count/frequency when speculation failed; fix type check
14904 for the direct call.
14906 2013-08-26 Jan Hubicka <jh@suse.cz>
14908 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
14910 2013-08-26 Jan Hubicka <jh@suse.cz>
14912 * ipa-inline-transform.c (inline_transform): Be ready for basic block
14913 to be changed by edge redirection.
14915 2013-08-26 Jan Hubicka <jh@suse.cz>
14917 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
14918 formating; add sanity check.
14919 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
14920 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
14921 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
14922 (resolve_noninline_speculation): Update callee keys, too.
14924 2013-08-26 Jan Hubicka <jh@suse.cz>
14926 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
14927 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
14929 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
14931 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
14934 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
14936 * config/i386/i386.c (ix86_debug_options): Remove prototype.
14937 (x86_64_elf_select_section): Ditto.
14938 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
14940 (ix86_pass_by_reference): Ditto.
14941 (output_set_got): Ditto.
14942 (ix86_unary_operator_ok): Ditto.
14943 (ix86_expand_builtin): Ditto.
14945 2013-08-23 Jan Hubicka <jh@suse.cz>
14947 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
14949 2013-08-23 Jan Hubicka <jh@suse.cz>
14951 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
14952 (tree_decl_with_vis): Add FINAL field.
14954 2013-08-23 Jeff Law <law@redhat.com>
14956 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
14957 the debugging dump when the expression is fully redundant.
14959 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
14961 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
14962 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
14963 * pretty-print.c (pp_formatted_text_data): Likewise.
14964 (pp_write_text_to_stream): Likewise.
14965 (pp_write_text_as_dot_label_to_stream): Likewise.
14966 (pp_append_r): Likewise.
14967 (pp_format): Likewise.
14968 (pp_flush): Likewise.
14969 (pp_clear_output_area): Likewise.
14970 (pp_append_text): Likewise.
14971 (pp_formatted_text): Likewise.
14972 (pp_remaining_character_count_for_line): Likewise.
14973 (pp_newline): Likewise.
14974 (pp_character): Likewise.
14975 (output_buffer::~output_buffer): Define.
14976 (pretty_printer::~pretty_printer): Destruct output buffer.
14977 * pretty-print.h (output_buffer::~output_buffer): Declare.
14978 (pretty_printer::~pretty_printer): Declare virtual.
14980 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
14983 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
14984 HOST_WIDE_INT_M1U): New macros.
14985 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
14986 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
14987 unsigned -1 for lshift.
14988 * cse.c (cse_insn): Likewise.
14989 * double-int.c (rshift_double, lshift_double): Likewise.
14990 * builtins.c (fold_builtin_bitop): Likewise.
14991 * combine.c (force_to_mode): Likewise.
14992 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
14993 * simplify-rtx.c (simplify_const_unary_operation,
14994 simplify_const_binary_operation): Likewise.
14995 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
14996 check_va_list_escapes): Likewise.
14997 * rtlanal.c (nonzero_bits1): Likewise.
14998 * expmed.c (expand_smod_pow2): Likewise.
14999 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
15001 2013-08-23 Jan Hubicka <jh@suse.cz>
15003 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
15004 as having address taken.
15006 2013-08-23 Jan Hubicka <jh@suse.cz>
15008 * ipa-utils.h (method_class_type): Declare.
15009 * ipa-devirt.c (method_class_type): Export.
15011 * cgraphunit.c (analyze_functions): Do basic devirtualization;
15012 do not walk base classes of anonymous types.
15014 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
15016 PR rtl-optimization/58220
15017 PR regression/58221
15018 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
15019 handle SEQUENCE insns properly.
15021 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
15023 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
15025 (pp_newline_and_indent): Likewise.
15026 (pp_separate_with): Likewise.
15027 * pretty-print.c (pp_newline_and_flush): Define.
15028 (pp_newline_and_indent): Likewise.
15029 (pp_separate_with): Likewise.
15031 2013-08-23 Jakub Jelinek <jakub@redhat.com>
15034 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
15035 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
15037 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
15039 * config/i386/predicates.md (ext_sse_reg_operand): New.
15040 * config/i386/i386.md (*movti_internal): Use
15041 predicate to determine if EVEX is needed.
15042 (*movsi_internal): Ditto.
15043 (*movdf_internal): Ditto.
15044 (*movsf_internal): Ditto.
15045 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
15047 2013-08-23 Jakub Jelinek <jakub@redhat.com>
15049 PR tree-optimization/58209
15050 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
15051 (find_tail_calls): Give up for pointer result types if m is non-NULL.
15052 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
15053 emit POINTER_PLUS_EXPR.
15054 (create_tailcall_accumulator): For pointer result type accumulate in
15057 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
15059 * configure.ac: Add backslashes missing from the last change.
15060 * configure: Regenerate.
15062 2013-08-22 Jan Hubicka <jh@susue.cz>
15064 * ipa.c (function_and_variable_visibility): First remember function
15065 was global and then make it local.
15067 2013-08-22 Julian Brown <julian@codesourcery.com>
15069 * configure.ac: Add aarch64 to list of arches which use "nop" in
15071 * configure: Regenerate.
15073 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15075 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
15076 gnu_libc_has_function.
15077 * config/s390/tpf.h: Likewise.
15079 2013-08-22 Jan Hubicka <jh@susue.cz>
15081 * timevar.c (validate_phases): Add cast.
15083 2013-08-22 Jan Hubicka <jh@susue.cz>
15085 * timevar.c (validate_phases): Use size_t for memory.
15086 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
15088 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
15090 * pretty-print.h (output_buffer::output_buffer): Declare.
15091 (pretty_printer::pretty_printer): Likewise.
15092 (pp_construct): Remove.
15093 * pretty-print.c (output_buffer::output_buffer): Define.
15094 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
15095 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
15096 (print_gimple_expr): Likewise.
15097 (print_gimple_seq): Likewise.
15098 (gimple_dump_bb): Likewise.
15099 * sched-vis.c (dump_value_slim): Likewise.
15100 (dump_insn_slim): Likewise.
15101 (dump_rtl_slim): Likewise.
15102 (str_pattern_slim): Likewise.
15103 * tree-mudflap.c (mf_varname_tree): Likewise.
15104 * graph.c (print_graph_cfg): Likewise.
15105 (start_graph_dump): Likewise.
15106 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
15108 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
15109 pretty printer initialization.
15110 * coretypes.h (diagnostic_context): Remove superflous type alias
15112 (pretty_printer): Likewise. Declare directly as a class.
15113 (pretty_print_info): Remove declaration as class.
15114 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
15115 and pp_clear_output_area.
15116 (asan_add_global): Likewise.
15118 2013-08-22 Jan Hubicka <jh@suse.cz>
15120 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
15121 * ipa-utils.h (update_type_inheritance_graph): Declare.
15122 (possible_polymorphic_call_target_p): Declare.
15123 (possible_polymorphic_call_target_p): New.
15124 * ipa-devirt.c: Update toplevel comments.
15125 (cached_polymorphic_call_targets): Move up.
15126 (odr_type_d): Move ID down.
15127 (polymorphic_type_binfo_p): Update comment.
15128 (odr_hasher::remove): Likewise;
15129 (get_odr_type): Set anonymous_namespace.
15130 (dump_odr_type): Dump it.
15131 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
15132 (maybe_record_node): Record node in cached_polymorphic_call_targets.
15133 (record_binfo): Add comment.
15134 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
15136 (devirt_node_removal_hook): Do not iCE when cache is freed.
15137 (possible_polymorphic_call_target_p): New predicate.
15138 (update_type_inheritance_graph): New function.
15140 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
15141 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
15142 Sergey Lega <sergey.s.lega@intel.com>
15143 Anna Tikhonova <anna.tikhonova@intel.com>
15144 Ilya Tocar <ilya.tocar@intel.com>
15145 Andrey Turetskiy <andrey.turetskiy@intel.com>
15146 Ilya Verbin <ilya.verbin@intel.com>
15147 Kirill Yukhin <kirill.yukhin@intel.com>
15148 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
15150 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
15151 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
15152 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
15153 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
15154 (OPTION_MASK_ISA_AVX2_UNSET): Update.
15155 (OPTION_MASK_ISA_AVX512F_UNSET): New.
15156 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
15157 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
15158 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
15159 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
15160 OPT_mavx512pf, OPT_mavx512er cases.
15161 * config/i386/constraints.md (v): New constraint.
15162 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
15163 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
15164 (bit_AVX512CD): New.
15165 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
15166 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
15167 * config/i386/i386-c.c (ix86_target_macros_internal):
15168 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
15170 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
15171 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
15172 * config/i386/i386.c (regclass_map, dbx_register_map)
15173 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
15174 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
15175 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
15176 -mavx512pf options.
15177 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
15178 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
15179 -mavx512pf options. Fix formatting.
15180 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
15181 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
15182 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
15183 -mavx512cd, -mavx512pf options.
15184 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
15185 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
15186 (ix86_preferred_output_reload_class): Replace SSE_REGS with
15188 (ix86_hard_regno_mode_ok): Support 512-bit registers.
15189 (ix86_set_reg_reg_cost): Ditto.
15190 (x86_order_regs_for_local_alloc): Ditto.
15191 (MAX_VECT_LEN): Extend to 64-byte.
15192 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
15193 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
15194 (TARGET_AVX512ER, TARGET_AVX512CD): New.
15195 (BIGGEST_ALIGNMENT): Extend to 512-bits.
15196 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
15197 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
15198 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
15199 (SSE_REG_MODE_P): Support new modes.
15200 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
15201 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
15202 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
15203 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
15204 (REG_CLASS_CONTENTS): Add new registers.
15205 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
15206 (EXT_REX_SSE_REGNO_P): New.
15207 (HI_REGISTER_NAMES): Add new registers.
15208 * config/i386/i386.md: Define constants for new registers.
15209 (mode): Add new 512-bit modes.
15210 (prefix): Support evex prefix.
15211 (isa): Support avx512f, noavx512f, fma_avx512f.
15212 (ssemodesuffix): Add new 512-bit modes.
15214 (*movxi_internal_avx512f): Ditto.
15215 (*movdi_internal): Replace constraint "x" with the new constraint "v".
15217 (*movsi_internal): Likewise.
15218 (*movdf_internal): Likewise.
15219 (*movsf_internal): Likewise.
15220 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
15221 (<code><mode>3): Likewise.
15222 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
15224 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
15225 with the new constraint "v".
15226 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
15228 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
15229 with the new constraint "v".
15230 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
15231 (<sse2>_storedqu<avxsizesuffix>): Likewise.
15232 (*<plusminus_insn><mode>3): Likewise.
15233 (<sse>_vm<plusminus_insn><mode>3): Likewise.
15234 (*mul<mode>3): Likewise.
15235 (<sse>_vmmul<mode>3): Likewise.
15236 (<sse>_div<mode>3): Likewise.
15237 (<sse>_vmdiv<mode>3): Likewise.
15238 (<sse>_sqrt<mode>2): Likewise.
15239 (<sse>_vmsqrt<mode>2): Likewise.
15240 (*<code><mode>3_finite): Likewise.
15241 (*<code><mode>3) <smaxmin>: Likewise.
15242 (<sse>_vm<code><mode>3): Likewise.
15243 (*<code><mode>3) <any_logic>: Likewise.
15244 (*fma_fmadd_<mode>): Likewise.
15245 (*fma_fmsub_<mode>): Likewise.
15246 (*fma_fnmadd_<mode>): Likewise.
15247 (*fma_fnmsub_<mode>): Likewise.
15248 (*fma_fmaddsub_<mode>): Likewise.
15249 (*fma_fmsubadd_<mode>): Likewise.
15250 (*fmai_fmadd_<mode>): Likewise.
15251 (*fmai_fmsub_<mode>): Likewise.
15252 (*fmai_fnmadd_<mode>): Likewise.
15253 (*fmai_fnmsub_<mode>): Likewise.
15254 (sse_cvtsi2ss): Likewise.
15255 (sse_cvtsi2ssq): Likewise.
15256 (sse_cvtss2si): Likewise.
15257 (sse_cvtss2si_2): Likewise.
15258 (sse_cvtss2siq): Likewise.
15259 (sse_cvtss2siq_2): Likewise.
15260 (sse_cvttss2si): Likewise.
15261 (sse_cvtss2siq_2): Likewise.
15262 (float<sseintvecmodelower><mode>2): Likewise.
15263 (sse2_cvtsd2si_2): Likewise.
15264 (sse2_cvtsd2siq_2): Likewise.
15265 (*<plusminus_insn><mode>3): Likewise.
15266 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
15267 (*<sse4_1_avx2>_mul<mode>3): Likewise.
15268 (ashr<mode>3): Likewise.
15269 (<shift_insn><mode>3): Likewise.
15270 (avx2_<code><mode>3): Likewise.
15271 (*avx2_<code><mode>3): Likewise.
15272 (*andnot<mode>3): Likewise.
15273 (*<code><mode>3) <any_logic>: Likewise.
15274 (abs<mode>2): Likewise.
15275 (avx2_permvar<mode>): Likewise.
15276 (avx2_perm<mode>_1): Likewise.
15277 (*avx_vpermilp<mode>): Likewise.
15278 (avx_vpermilvar<mode>3): Likewise.
15279 (avx2_ashrv<mode>): Likewise.
15280 (avx2_<shift_insn>v<mode>): Likewise.
15281 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
15283 * doc/rtl.texi: Document XImode.
15285 2013-08-21 Jeff Law <law@redhat.com>
15287 * tree-flow.h (register_jump_thread): Pass vector of edges
15288 instead of each important edge.
15289 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
15290 thread path into a vector and pass that to register_jump_thread.
15291 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
15292 passed in edge vector to the current 3-edge form.
15295 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
15297 * dce.c (fini_dce): Call df_analyze again just in case
15298 delete_unmarked_insns removed anything.
15300 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
15302 * reload.h (struct reg_equivs): Rename to ..
15303 (struct reg_equivs_s): .. this.
15305 2013-08-20 Martin Liska <marxin.liska@gmail.com>
15307 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
15309 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15311 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
15313 2013-08-21 Jeff Law <law@redhat.com>
15315 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
15316 simplify assignments too. If the RHS collapses to a singleton
15317 range, then return the value for the range.
15319 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
15321 * config/i386/sse.md (V16): Rename to...
15323 (mov<mode>): Update iterator name.
15324 (*mov<mode>_internal): Ditto.
15325 (push<mode>1): Ditto.
15326 (movmisalign<mode>): Ditto.
15328 2013-08-20 Jan Hubicka <jh@suse.cz>
15331 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
15332 entry for direct edges.
15333 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
15335 2013-08-20 David Malcolm <dmalcolm@redhat.com>
15337 Revert my last two changes, r201865 and r201864:
15340 2013-08-20 David Malcolm <dmalcolm@redhat.com>
15342 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
15343 instances can own GC refs.
15345 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
15346 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
15347 (gcc::context::gt_pch_nx): Likewise.
15348 (gcc::context::gt_pch_nx): Likewise.
15349 * ggc.h (gt_ggc_mx <T>): New.
15350 (gt_pch_nx_with_op <T>): New.
15351 (gt_pch_nx <T>): New.
15352 * passes.c (opt_pass::gt_ggc_mx): New.
15353 (opt_pass::gt_pch_nx): New.
15354 (opt_pass::gt_pch_nx_with_op): New.
15355 (pass_manager::gt_ggc_mx): New.
15356 (pass_manager::gt_pch_nx): New.
15357 (pass_manager::gt_pch_nx_with_op): New.
15358 (pass_manager::operator new): Use
15359 ggc_internal_cleared_alloc_stat rather than xcalloc.
15360 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
15361 (pass_manager::gt_ggc_mx): New.
15362 (pass_manager::gt_pch_nx): New.
15363 (pass_manager::gt_pch_nx_with_op): New.
15364 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
15365 (opt_pass::operator new): New.
15366 (opt_pass::gt_ggc_mx): New.
15367 (opt_pass::gt_pch_nx): New.
15368 (opt_pass::gt_pch_nx_with_op): New.
15371 2013-08-20 David Malcolm <dmalcolm@redhat.com>
15373 * Makefile.in (GTFILES): Add context.h.
15374 * context.c (gcc::context::operator new): New.
15375 (gcc::context::gt_ggc_mx): New.
15376 (gcc::context::gt_pch_nx): New.
15377 (gcc::context::gt_pch_nx): New.
15378 * context.h (gcc::context): Add GTY((user)) marking.
15379 (gcc::context::operator new): New.
15380 (gcc::context::gt_ggc_mx): New.
15381 (gcc::context::gt_pch_nx): New.
15382 (gcc::context::gt_pch_nx): New.
15383 (g): Add GTY marking.
15384 (gt_ggc_mx (gcc::context *)): New.
15385 (gt_pch_nx (gcc::context *)): New.
15386 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
15387 void *cookie)): New.
15388 * gengtype.c (open_base_files) <ifiles>: Add context.h.
15390 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
15392 * dce.c (fini_dce): Call df_analyze again just in case
15393 delete_unmarked_insns removed anything.
15395 2013-08-20 Teresa Johnson <tejohnson@google.com>
15397 PR rtl-optimizations/57451
15398 * final.c (reemit_insn_block_notes): Prevent lexical blocks
15399 from crossing split section boundaries.
15401 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
15403 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
15404 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
15405 with MULTLIB_DEFAULTS.
15407 2013-08-20 Nick Clifton <nickc@redhat.com>
15409 * target.def (narrow_volatile_bitfield): Note that the default
15410 value is false, not !TARGET_STRICT_ALIGN.
15411 * doc/tm.texi: Regenerate.
15413 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
15415 Fix LIB_SPEC for systems without libpthread.
15417 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
15418 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
15420 * config/i386/linux-common.h: Likewise.
15421 * config/mips/linux-common.h: Likewise.
15423 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
15425 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
15428 2013-08-20 David Malcolm <dmalcolm@redhat.com>
15430 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
15431 instances can own GC refs.
15433 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
15434 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
15435 (gcc::context::gt_pch_nx): Likewise.
15436 (gcc::context::gt_pch_nx): Likewise.
15437 * ggc.h (gt_ggc_mx <T>): New.
15438 (gt_pch_nx_with_op <T>): New.
15439 (gt_pch_nx <T>): New.
15440 * passes.c (opt_pass::gt_ggc_mx): New.
15441 (opt_pass::gt_pch_nx): New.
15442 (opt_pass::gt_pch_nx_with_op): New.
15443 (pass_manager::gt_ggc_mx): New.
15444 (pass_manager::gt_pch_nx): New.
15445 (pass_manager::gt_pch_nx_with_op): New.
15446 (pass_manager::operator new): Use
15447 ggc_internal_cleared_alloc_stat rather than xcalloc.
15448 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
15449 (pass_manager::gt_ggc_mx): New.
15450 (pass_manager::gt_pch_nx): New.
15451 (pass_manager::gt_pch_nx_with_op): New.
15452 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
15453 (opt_pass::operator new): New.
15454 (opt_pass::gt_ggc_mx): New.
15455 (opt_pass::gt_pch_nx): New.
15456 (opt_pass::gt_pch_nx_with_op): New.
15458 2013-08-20 David Malcolm <dmalcolm@redhat.com>
15460 * Makefile.in (GTFILES): Add context.h.
15461 * context.c (gcc::context::operator new): New.
15462 (gcc::context::gt_ggc_mx): New.
15463 (gcc::context::gt_pch_nx): New.
15464 (gcc::context::gt_pch_nx): New.
15465 * context.h (gcc::context): Add GTY((user)) marking.
15466 (gcc::context::operator new): New.
15467 (gcc::context::gt_ggc_mx): New.
15468 (gcc::context::gt_pch_nx): New.
15469 (gcc::context::gt_pch_nx): New.
15470 (g): Add GTY marking.
15471 (gt_ggc_mx (gcc::context *)): New.
15472 (gt_pch_nx (gcc::context *)): New.
15473 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
15474 void *cookie)): New.
15475 * gengtype.c (open_base_files) <ifiles>: Add context.h.
15477 2013-08-20 Alan Modra <amodra@gmail.com>
15480 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
15481 (rs6000_emit_epilogue): Likewise.
15483 2013-08-19 Dehao Chen <dehao@google.com>
15485 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
15487 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
15488 Jakub Jelinek <jakub@redhat.com>
15490 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
15491 (BUILT_IN_FABSD64): Likewise.
15492 (BUILT_IN_FABSD128): Likewise.
15493 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
15494 (fold_builtin_1): Likewise.
15495 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
15496 destination and source operands.
15497 (*abstd2_fpr): Likewise.
15498 (*nabstd2_fpr): Likewise.
15500 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
15502 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
15505 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
15507 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
15510 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
15512 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
15513 * builtins.c (default_libc_has_function): New.
15514 (gnu_libc_has_function): Ditto.
15515 (no_c99_libc_has_function): Ditto.
15516 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
15517 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
15518 (fold_builtin_sincos): Likewise.
15519 (fold_builtin_cexp): Likewise.
15520 * builtins.def (DEF_C94_BUILTIN): Likewise.
15521 (DEF_C99_BUILTIN): Likewise.
15522 (DEF_C99_C90RES_BUILTIN): Likewise.
15523 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
15524 definitions to using this define.
15525 * config/darwin-protos.h (darwin_libc_has_function): New.
15526 * config/darwin.c (darwin_libc_has_function): Ditto.
15527 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
15528 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
15529 * config/darwin.h: Ditto.
15530 * config/elfos.h: Ditto.
15531 * config/freebsd.h: Ditto.
15532 * config/i386/cygming.h: Ditto.
15533 * config/i386/djgpp.h: Ditto.
15534 * config/i386/i386-interix.h: Ditto.
15535 * config/microblaze/microblaze.h: Ditto.
15536 * config/mmix/mmix.h: Ditto.
15537 * config/gnu-user.h: Ditto.
15538 * config/ia64/hpux.h: Ditto.
15539 * config/pa/pa-hpux.h: Ditto.
15540 * config/pdp11/pdp11.h: Ditto.
15541 * config/picochip/picochip.h: Ditto.
15542 * config/linux.h: Ditto.
15543 * config/netbsd.h: Ditto.
15544 * config/openbsd.h: Ditto.
15545 * config/rs6000/aix43.h: Ditto.
15546 * config/rs6000/aix51.h: Ditto.
15547 * config/rs6000/aix52.h: Ditto.
15548 * config/rs6000/aix53.h: Ditto.
15549 * config/rs6000/aix61.h: Ditto.
15550 * config/rs6000/darwin.h: Ditto.
15551 * config/rs6000/linux.h: Ditto.
15552 * config/rs6000/linux64.h: Ditto.
15553 * config/s390/tpf.h: Ditto.
15554 * config/sol2-10.h: Ditto.
15555 * config/sol2.h: Ditto.
15556 * config/vms/vms.h: Ditto.
15557 * config/vxworks.h: Ditto.
15558 * config/linux-android.c (linux_android_libc_has_function):
15559 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
15560 * config/linux-protos.h (linux_android_libc_has_function):
15562 * config/i386/i386.c (ix86_libc_has_function): New.
15563 * config/i386/i386-protos.h
15564 (ix86_libc_has_function): New declaration.
15565 * config/i386/i386.md
15566 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
15567 ("isinf<mode>2): Likewise.
15568 * convert.c (convert_to_integer): Using new target hook
15569 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
15570 TARGET_C99_FUNCTIONS.
15571 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
15572 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
15573 * coretypes.h (function_class): New enum for different
15574 classes of functions.
15575 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
15576 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
15577 (TARGET_HAS_SINCOS): Likewise.
15578 (TARGET_LIBC_HAS_FUNCTION): New.
15579 * doc/tm.texi: Regenerated.
15580 * targhooks.h (default_libc_has_function): New declaration.
15581 (no_c99_libc_has_function): Ditto.
15582 (gnu_libc_has_function): Ditto.
15583 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
15584 and TARGET_HAS_SINCOS.
15586 2013-08-18 Jan Hubicka <jh@suse.cz>
15588 * Makeifle-in (ipa-devirt.o): New.
15589 (GTFILES): Add ipa-utils.h and ipa-devirt.c
15590 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
15591 (analyze_functions): Look into possible targets of polymorphic call.
15592 * dumpfile.c (dump_files): Add type-inheritance dump.
15593 * dumpfile.h (TDI_inheritance): New.
15594 * ipa-devirt.c: New file.
15595 * ipa-utils.h (odr_type_d): Forward declare.
15596 (odr_type): New type.
15597 (build_type_inheritance_graph): Declare.
15598 (possible_polymorphic_call_targets): Declare and introduce inline
15599 variant when only edge is pased.
15600 (dump_possible_polymorphic_call_targets): Likewise.
15601 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
15602 * tree.c (type_in_anonymous_namespace_p): Break out from ...
15603 (types_same_for_odr): ... here.
15604 * tree.h (type_in_anonymous_namespace_p): Declare.
15606 2013-08-18 Jakub Jelinek <jakub@redhat.com>
15608 PR tree-optimization/58006
15609 * tree-parloops.c (take_address_of): Don't ICE if get_name
15611 (eliminate_local_variables_stmt): Remove clobber stmts.
15613 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
15615 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
15616 error message is issued for an alias to undefined symbol.
15618 2013-08-18 Jan Hubicka <jh@suse.cz>
15620 * cgraph.c (cgraph_create_indirect_edge): Discover
15621 polymorphic calls and record basic info into indirect_info.
15622 * gimple-fold.c (gimple_fold_call): When doing BINFO based
15623 devirtualization, ignore objc function calls.
15624 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
15625 call with no parm index info.
15626 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
15627 * tree.c (virtual_method_call_p): New function.
15628 * tree.h (virtual_method_call_p): Declare.
15630 2013-08-16 Jan Hubicka <jh@suse.cz>
15632 PR middle-end/58179
15633 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
15635 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
15637 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
15640 2013-08-16 David Malcolm <dmalcolm@redhat.com>
15642 * gengtype.c (type_for_name): Add special-case support for
15643 locating types within the "gcc::" namespace.
15644 (open_base_files): Emit a "using namespace gcc" directive.
15646 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
15649 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
15650 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
15652 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
15653 array instead of each individual operand as a separate argument.
15654 (emit_fusion_gpr_load): Likewise.
15655 (expand_fusion_gpr_load): Add new function declaration.
15657 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
15658 signature to have the operands passed as an array, instead of as
15659 separate arguments. Allow ZERO_EXTEND to be in the memory
15660 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
15661 depend on the register live/dead flags when peepholes are run.
15662 (expand_fusion_gpr_load): New function to be called from the
15663 peephole2 pass, to change the register that addis sets to be the
15665 (emit_fusion_gpr_load): Change the calling signature to have the
15666 operands passed as an array, instead of as separate arguments.
15667 Allow ZERO_EXTEND to be in the memory address, and also
15668 SIGN_EXTEND if -mpower8-fusion-sign.
15670 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
15671 unspec enumeration.
15672 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
15673 adjust the register addis loads up in the peephole2 pass. Do not
15674 depend on the register live/dead state when the peephole pass is done.
15676 2013-08-16 David Malcolm <dmalcolm@redhat.com>
15678 * gengtype.c (create_user_defined_type): Ensure that the kind
15679 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
15680 declaration is seen before the GTY((user)) marking.
15682 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
15685 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
15687 2013-08-16 Jan Hubicka <jh@suse.cz>
15689 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
15690 arugment expected_type.
15691 (gimple_fold_call): Use it.
15692 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
15693 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
15694 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
15695 (try_make_edge_direct_virtual_call): Likewise.
15696 * tree.c (obj_type_ref_class): New.
15697 * tree.h (obj_type_ref_class): Use it.
15699 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
15701 * sched-vis.c (rtl_slim_pp_initialized): Remove.
15702 (rtl_slim_pp): Likewise.
15703 (init_rtl_slim_pretty_print): Likewise.
15704 (dump_value_slim): Don't call it. Use local pretty printer.
15705 (dump_insn_slim): Likewise.
15706 (dump_rtl_slim): Likewise.
15707 (str_pattern_slim): Likewise.
15708 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
15711 2013-08-16 Jakub Jelinek <jakub@redhat.com>
15713 PR tree-optimization/58164
15714 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
15715 walk gimple_goto_dest of GIMPLE_GOTO.
15717 PR tree-optimization/58165
15718 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
15719 bi_call must be the last stmt in a bb, don't split_block, instead
15720 use fallthru edge from it and give up if there is none.
15721 Release conds vector when returning early.
15723 2013-08-14 Xinliang David Li <davidxl@google.com>
15725 * config/i386/i386.c (ix86_option_override_internal):
15726 Remove unused variable and field.
15728 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15731 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
15732 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
15733 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
15734 and Linux, correct BLKmode alignment when 128-bit alignment is
15735 required and compatibility flag is not set.
15736 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
15737 for zero-size arguments when compatibility flag is not set.
15739 2013-08-14 Jakub Jelinek <jakub@redhat.com>
15741 PR tree-optimization/58145
15742 * tree-sra.c (build_ref_for_offset): If prev_base has
15743 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
15745 2013-08-14 Xinliang David Li <davidxl@google.com>
15747 * config/i386/i386.c (ix86_option_override_internal):
15748 Fix uninitialized variable error.
15750 2013-08-14 Xinliang David Li <davidxl@google.com>
15752 * config/i386/i386.opt: Define two new options.
15753 * config/i386/x86-tune.def: Add arch selector field in macros.
15754 * config/i386/i386.h: Adjust macro definition.
15755 * config/i386/i386.c (ix86_option_override_internal):
15757 (parse_mtune_ctrl_str): New function.
15758 (set_ix86_tune_features): New function.
15759 (ix86_function_specific_restore): Call the new helper function.
15761 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
15763 PR rtl-optimization/57662
15764 * sel-sched.c (code_motion_process_successors): When the current insn
15765 is removed after the recursive traversal, break from the loop.
15766 Add comments and debug printouts.
15768 2013-08-14 Jakub Jelinek <jakub@redhat.com>
15769 Alexandre Oliva <aoliva@redhat.com>
15772 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
15773 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
15774 there also UNSPEC_PLTOFF.
15776 2013-08-14 Marek Polacek <polacek@redhat.com>
15778 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
15779 NUM_CONDITIONS bit positions.
15781 2013-08-13 Cary Coutant <ccoutant@google.com>
15783 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
15784 (attr_checksum): Hash vector contents instead of pointer.
15785 (attr_checksum_ordered): Likewise.
15787 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
15789 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
15790 when Pmode != word_mode. Add length_address attribute.
15791 (sse3_monitor_<mode>): Merge from sse3_monitor and
15792 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
15793 Pmode != word_mode. Update insn length attribute.
15794 * config/i386/i386.c (ix86_option_override_internal): Update
15795 ix86_gen_monitor selection for merged sse3_monitor insn.
15797 2013-08-13 Julian Brown <julian@codesourcery.com>
15799 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
15800 perform invalid legitimization on greater-than-word-size modes for
15801 TARGET_E500_DOUBLE.
15803 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
15805 * ira.c (setup_class_translate_array): Use aclass instead of cl
15806 for classes not fully covered by allocno classes.
15808 2013-08-13 Jakub Jelinek <jakub@redhat.com>
15810 PR tree-optimization/57661
15811 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
15812 * tree-inline.c (tree_function_versioning): Initialize it.
15813 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
15814 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
15815 that is not being copied.
15818 * asan.c (instrument_strlen_call): Fix typo in comment.
15819 Use char * type even for the lhs of POINTER_PLUS_EXPR.
15821 2013-08-13 Steve Ellcey <sellcey@mips.com>
15823 * config/mips/mips.md (prefetch): Use lw instead of ld on
15824 loongson in 32bit mode.
15826 2013-08-13 Nick Clifton <nickc@redhat.com>
15828 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
15830 2013-08-13 Jan Hubicka <jh@suse.cz>
15832 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
15833 introduced edge; fix typo in sanity check.
15834 (cgraph_resolve_speculation): Export; improve diagnostic.
15835 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
15836 speculation at type mismatch.
15837 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
15838 (cgraph_resolve_speculation): Declare.
15839 (symtab_can_be_discarded): New function.
15840 * value-prof.c (gimple_ic_transform): Remove actual transform code.
15841 * ipa-inline-transform.c (speculation_removed): New global var.
15842 (clone_inlined_nodes): See if speculation can be removed.
15843 (inline_call): If speculations was removed, we growths may not match.
15844 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
15845 (speculation_useful_p): New function.
15846 (resolve_noninline_speculation): New function.
15847 (inline_small_functions): Resolve useless speculations.
15848 * ipa-inline.h (speculation_useful_p): Declare
15849 * ipa.c (can_replace_by_local_alias): Simplify.
15850 (ipa_profile): Produce speculative calls in non-lto, too;
15851 add simple cost model; produce local aliases.
15853 2013-08-13 David Malcolm <dmalcolm@redhat.com>
15855 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
15858 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
15860 * config/i386/i386.c (ix86_function_versions): Use error + inform.
15862 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
15864 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
15865 iterator instead of X87MODEF.
15867 2013-08-12 Perez Read <netfirewall@gmail.com>
15870 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
15871 operand 0 for intel asm alternative.
15872 (*movabs<mode>_2): Ditto for operand 1.
15874 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
15876 * config/aarch64/arm_none.h
15877 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
15879 2013-08-12 Nick Clifton <nickc@redhat.com>
15881 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
15883 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
15885 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
15886 expand for QI/HImode operand to produce more effictive code for
15887 unsigned char(short) --> float(double) conversion.
15889 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
15891 * doc/invoke.texi: Mention that -ftls-model does not force the final
15894 2013-08-12 Marek Polacek <polacek@redhat.com>
15895 Marc Glisse <marc.glisse@inria.fr>
15897 PR tree-optimization/57980
15898 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
15899 when creating -1 constant.
15901 2013-08-10 Jan Hubicka <jh@suse.cz>
15903 Workaround binutils PR14342.
15904 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
15905 (gimple_init_edge_profiler): Likewise.
15906 (gimple_gen_ic_func_profiler): Likewise.
15908 2013-08-09 Jan Hubicka <jh@suse.cz>
15910 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
15912 2013-08-09 Xinliang David Li <davidxl@google.com>
15914 * config/i386/stringop.def: New file.
15915 * config/i386/stringop.opt: New file.
15916 * config/i386/i386-opts.h: Include stringopt.def.
15917 * config/i386/i386.opt: Include stringopt.opt.
15918 * config/i386/i386.c (ix86_option_override_internal):
15919 Override default size based stringop inline strategies with options.
15920 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
15923 2013-08-09 Jan Hubicka <jh@suse.cz>
15925 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
15927 2013-08-09 Jan Hubicka <jh@suse.cz>
15929 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
15931 (dump_cgraph_node): Dump profile-id.
15932 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
15933 and common_target_probability.
15934 * lto-cgraph.c (lto_output_edge): Stream common targets.
15935 (lto_output_node): Stream profile ids.
15936 (input_node): Stream profile ids.
15937 (input_edge): Stream common targets.
15938 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
15939 * ipa.c: Include value-prof.h
15940 (ipa_profile_generate_summary): Turn indirect call statement histograms
15941 into common targets.
15942 (ipa_profile): Turn common targets into speculative edges.
15944 2013-08-09 Jan Hubicka <jh@suse.cz>
15946 * cgraph.h (cgraph_node): Add profile_id.
15947 * value-prof.c (cgraph_node_map): Turn into pointer_map.
15948 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
15949 (del_node_map): Update.
15950 (find_func_by_funcdef_no): Replace by ...
15951 (find_func_by_profile_id): ... this one.
15952 (gimple_ic_transform): Do not remove useful histograms when
15953 speculation is not done; dump info when indirect call removal
15955 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
15956 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
15957 (__gcov_indirect_call_profiler_v2): .. this one.
15958 * profile.h (init_node_map): Update.
15959 * coverage.c (coverage_compute_profile_id): New function.
15960 * coverage.h (coverage_compute_profile_id): Declare.
15961 * tree-profile.c (init_ic_make_global_vars): Make
15962 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
15963 (gimple_init_edge_profiler): Update prototype of
15964 __gcov_indirect_call_profiler.
15965 (gimple_gen_ic_func_profiler): Simplify.
15966 (tree_profiling): Use init_node_map
15968 2013-08-09 Jan Hubicka <jh@suse.cz>
15970 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
15971 non-speculative refs.
15972 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
15973 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
15974 (cgraph_set_call_stmt): Likewise.
15975 (cgraph_create_edge_1): Fix release checking compilatoin;
15976 clear lto_stmt_uid.
15977 (cgraph_free_edge): Free indirect info.
15978 (cgraph_turn_edge_to_speculative): New function.
15979 (cgraph_speculative_call_info): New function.
15980 (cgraph_make_edge_direct): Return direct edge; handle speculation.
15981 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
15982 (dump_cgraph_node): Dump speculation.
15983 (verify_edge_count_and_frequency): Accept speculative edges.
15984 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
15985 (verify_cgraph_node): Handle speculation.
15986 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
15987 (cgraph_set_call_stmt): Update prototype.
15988 (cgraph_make_edge_direct): Update prototype.
15989 (cgraph_speculative_call_info): Declare.
15990 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
15991 to change; update call of ipa_find_references.
15992 * ipa-ref.c (ipa_record_reference): Fix return value; clear
15993 lto_stmt_uid and speculative flags.
15994 (ipa_dump_references): Dump speculation.
15995 (ipa_clone_references): Clone speculative flag.
15996 (ipa_clone_referring): Likewise.
15997 (ipa_clone_ref): New function.
15998 (ipa_find_reference): Look into lto_stmt_uids
15999 (ipa_clear_stmts_in_references): Do not clear speculative calls.
16000 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
16001 (ipa_find_reference): Update declaration.
16002 (ipa_clone_ref): Declare.
16003 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
16004 stream speculative flag.
16005 (lto_output_ref): Stream statements uids and speculation.
16006 (input_ref): Likewise.
16007 (input_edge): Stream speuclation.
16008 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
16009 (cgraph_set_call_stmt_including_clones): Handle speculation.
16010 * ipa-inline.c (heap_edge_removal_hook): New function.
16011 (inline_small_functions): Register it.
16012 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
16013 also initialize refs.
16014 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
16016 (try_make_edge_direct_simple_call): Likewise.
16017 (try_make_edge_direct_simple_call): Likewise.
16018 (update_indirect_edges_after_inlining): Likewise.
16019 (remove_described_reference): Look proper lto_stmt_uid.
16020 (propagate_controlled_uses): Likewise.
16021 (propagate_controlled_uses): Liekwise.
16022 * tree-inline.c (copy_bb): Copy speculative edges.
16023 (redirect_all_calls): New function.
16024 (copy_cfg_body): Do redirection after loop info is updated.
16025 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
16027 2013-08-09 Jan Hubicka <jh@suse.cz>
16029 * lto-streamer-out.c (output_function): Renumber PHIs.
16030 * lto-streamer-in.c (input_function): Likewise.
16032 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
16034 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
16035 (get_lane_unsigned): Likewise.
16036 (dup_lane_scalar): Likewise.
16037 (get_lane): enable for VALL.
16038 * config/aarch64/aarch64-simd.md
16039 (aarch64_dup_lane_scalar<mode>): Remove.
16040 (aarch64_get_lane_signed<mode>): Likewise.
16041 (aarch64_get_lane_unsigned<mode>): Likewise.
16042 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
16043 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
16044 (aarch64_get_lane<mode>): Enable for all vector modes.
16045 (aarch64_get_lanedi): Remove misleading constraints.
16046 * config/aarch64/arm_neon.h
16047 (__aarch64_vget_lane_any): Define.
16048 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
16049 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
16050 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
16051 * config/aarch64/iterators.md (VDQQH): New.
16052 (VDQQHS): Likewise.
16053 (vwcore): Likewise.
16055 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
16057 * configure.ac: Add GAS check for LEON instructions on SPARC.
16058 * configure: Regenerate.
16059 * config.in: Likewise.
16060 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
16062 * config/sparc/sparc.opt (LEON, LEON3): New masks.
16063 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
16065 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
16066 (AS_LEON_FLAG): New macro.
16067 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
16068 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
16069 Deal with LEON and LEON3 for the memory model.
16070 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
16071 (atomic_compare_and_swap<mode>_1): Likewise.
16072 (*atomic_compare_and_swap<mode>_1): Likewise.
16074 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16076 * config/arm/neon.md (vcond): Fix floating-point vector
16077 comparisons against 0.
16079 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
16081 * lra-constraints.c (emit_spill_move): Remove assert.
16082 (process_alt_operands): Add more debugging
16083 output. Increase reject for spilling into memory. Decrease
16084 reject for reloading scratch.
16085 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
16087 2013-08-08 Steve Ellcey <sellcey@mips.com>
16089 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
16090 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
16091 micromips incompatible. Add nan2008.
16092 (MULTILIB_DIRNAMES): Add nan2008.
16093 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
16094 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
16095 and micromips incompatible. Add nan2008.
16096 (MULTILIB_DIRNAMES): Add nan2008.
16097 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
16099 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
16101 PR rtl-optimization/58079
16102 * combine.c (combine_simplify_rtx): Avoid using SUBST if
16103 simplify_comparison has widened a comparison with an integer.
16105 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16107 * config/arm/neon.md (movmisalign<mode>): Disable when we
16108 don't allow unaligned accesses.
16109 (*movmisalign<mode>_neon_store): Likewise.
16110 (*movmisalign<mode>_neon_load): Likewise.
16111 (*movmisalign<mode>_neon_store): Likewise.
16112 (*movmisalign<mode>_neon_load): Likewise.
16114 2013-08-08 Jan Hubicka <jh@suse.cz>
16116 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
16117 (make_pass_rebuild_cgraph_edges): Also clear references.
16118 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
16119 * ipa-inline-transform.c (inline_transform): Remove all references
16121 * cgraphunit.c (expand_function): Remove all references after
16123 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
16124 (ipa_find_reference): Rewrite to iterator.
16125 (remove_stmt_references): Likewise.
16126 (ipa_clear_stmts_in_references): New function.
16127 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
16128 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
16130 * ipa-split.c (split_function): Remove references in split function.
16132 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
16135 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
16136 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
16138 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
16141 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
16142 suggested mode for the assignment isn't compatible with the
16143 registers required.
16145 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
16148 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
16150 2013-08-07 Xinliang David Li <davidxl@google.com>
16152 * config/i386/i386.opt: New option -mtune-ctrl=.
16153 * config/i386/x86-tune.def: New file.
16154 * config/i386/i386.h: include x86-tune.def.
16155 * config/i386/i386.c (ix86_option_override_internal):
16156 Parsing -mtune-ctrl= option and set tune features.
16158 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
16161 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
16162 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
16165 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
16167 * rtl.h (update_alignments): Declare.
16168 * final.c (grow_label_align): New function extracted from...
16169 (shorten_branches): ...here. Call it.
16170 (update_alignments): New function.
16171 * reorg.c (sibling_labels): New variable.
16172 (get_label_before): Add SIBLING parameter. If it is non-zero, push
16173 the new label along with it onto the sibling_labels vector.
16174 (fill_simple_delay_slots): Adjust call to get_label_before.
16175 (fill_slots_from_thread): Likewise.
16176 (relax_delay_slots): Likewise.
16177 (make_return_insns): Likewise.
16178 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
16180 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
16182 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
16183 document its semantics.
16184 (diagnostic_report_diagnostic): Adjust accordingly.
16186 2013-08-07 David Malcolm <dmalcolm@redhat.com>
16188 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
16189 (sparc_option_override): ...and port to new C++ pass API.
16190 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
16192 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
16194 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
16196 2013-08-06 Caroline Tice <cmtice@google.com>
16198 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
16199 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
16200 * tree-pass.h: Add pass_vtable_verify.
16201 * varasm.c (assemble_variable): Add code to properly set the comdat
16202 section and name for the .vtable_map_vars section.
16203 (assemble_vtyv_preinit_initializer): New function.
16204 (default_sectin_type_flags): Make sure .vtable_map_vars section has
16206 * output.h: Add function decl for assemble_vtv_preinit_initializer.
16207 * vtable-verify.c: New file.
16208 * vtable-verify.h: New file.
16209 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
16210 initialiation levels.
16211 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
16212 * passes.def: Insert pass_vtable_verify.
16213 * aclocal.m4: Reorder includes.
16214 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
16215 -fvtv-counts options.
16216 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
16217 as appropriate, if -fvtable-verify=... is used.
16218 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
16219 -fvtable-verify=... is used.
16220 * Makefile.in (OBJS): Add vtable-verify.o to list.
16221 (vtable-verify.o): Add new build rule.
16222 (GTFILES): Add vtable-verify.c to list.
16223 * common.opt (fvtable-verify=): New flag.
16224 (vtv_priority): Values for fvtable-verify= flag.
16225 (fvtv-counts): New flag.
16226 (fvtv-debug): New flag.
16227 * tree.h (save_vtable_map_decl): New extern function decl.
16229 2013-08-07 David Malcolm <dmalcolm@redhat.com>
16231 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
16232 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
16233 (pass_data_rl78_devirt): ...new pass_data instance and...
16234 (make_pass_rl78_devirt): ...new function.
16235 (rl78_asm_file_start): Port pass registration to new C++ API.
16237 2013-08-07 David Malcolm <dmalcolm@redhat.com>
16239 * coretypes.h (rtl_opt_pass): Add.
16240 (gcc::context): Add.
16241 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
16242 (epiphany_init): Port to new C++ pass API.
16243 (epiphany_optimize_mode_switching): Likewise.
16244 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
16245 (pass_manager::get_pass_mode_switching): New.
16246 (pass_manager::get_pass_peephole2): New.
16247 * mode-switching.c (pass_mode_switching): Add clone method.
16248 * recog.c (pass_peephole2): Add clone method.
16249 (pass_split_all_insns): Add clone method.
16251 2013-08-06 David Malcolm <dmalcolm@redhat.com>
16253 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
16254 (mips_option_override): ...here, porting to new C++ API for passes.
16256 2013-08-06 Jan Hubicka <jh@suse.cz>
16258 * cgraph.c (cgraph_get_body): New function based on lto.c
16260 * cgraph.h (cgraph_get_body): Declare.
16261 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
16263 * cgraphunit.c (expand_function): Get body prior expanding.
16264 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
16265 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
16267 * passes.c (do_per_function_toporder): Get body.
16268 * tree-inline.c (expand_call_inline): Get body prior inlining it.
16269 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
16271 2013-08-06 Martin Jambor <mjambor@suse.cz>
16274 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
16275 is not re-finalized. Rename second parameter to no_collect.
16277 2013-08-06 Martin Jambor <mjambor@suse.cz>
16279 PR middle-end/58041
16280 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
16281 MEM_REF has proper alignment information.
16283 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
16286 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
16288 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
16289 rtx (*) (rtx, ...) with insn_gen_fn.
16290 * genoutput.c (output_insn_data): Cast gen_? function pointers to
16291 insn_gen_fn::stored_funcptr. Add initializer braces.
16293 2013-08-05 David Malcolm <dmalcolm@redhat.com>
16295 Rewrite how instances of passes are cloned to remove assumptions
16296 about their sizes (thus allowing pass subclasses to have
16297 additional data fields, albeit non-GC-managed ones at this point).
16299 * passes.c (make_pass_instance): Now that passes have clone
16300 methods, rewrite this function to eliminate XNEW and memcpy
16301 calls that used hardcoded sizes. Since this function no longer
16302 creates pass instances, rename it to...
16303 (add_pass_instance): ...this. Document the old way that passes were
16304 numbered and flagged, and rework this function to continue using it.
16305 (next_pass_1): Add an initial_pass argument for use by
16307 (position_pass): When adding multiple instances of a pass, use
16308 the pass's clone method, rather than relying on the XNEW/memcpy
16309 within the former make_pass_instance (now add_pass_instance).
16310 (pass_manager::pass_manager): When invoking next_pass_1, also supply
16311 the initial instance of the current pass within the pass manager.
16313 2013-08-05 David Malcolm <dmalcolm@redhat.com>
16315 This is the automated part of the conversion of passes from C
16316 structs to C++ classes.
16318 Patch autogenerated by refactor_passes.py from
16319 https://github.com/davidmalcolm/gcc-refactoring-scripts
16320 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
16322 * asan.c (pass_asan): Convert from a global struct to a subclass of
16323 gimple_opt_pass along with...
16324 (pass_data_asan): ...new pass_data instance and...
16325 (make_pass_asan): ...new function.
16326 (pass_asan_O0): Convert from a global struct to a subclass of
16327 gimple_opt_pass along with...
16328 (pass_data_asan_O0): ...new pass_data instance and...
16329 (make_pass_asan_O0): ...new function.
16330 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
16331 subclass of rtl_opt_pass along with...
16332 (pass_data_inc_dec): ...new pass_data instance and...
16333 (make_pass_inc_dec): ...new function.
16334 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
16335 a subclass of rtl_opt_pass along with...
16336 (pass_data_reorder_blocks): ...new pass_data instance and...
16337 (make_pass_reorder_blocks): ...new function.
16338 (pass_duplicate_computed_gotos): Convert from a global struct to a
16339 subclass of rtl_opt_pass along with...
16340 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
16341 (make_pass_duplicate_computed_gotos): ...new function.
16342 (pass_partition_blocks): Convert from a global struct to a subclass of
16343 rtl_opt_pass along with...
16344 (pass_data_partition_blocks): ...new pass_data instance and...
16345 (make_pass_partition_blocks): ...new function.
16346 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
16347 struct to a subclass of rtl_opt_pass along with...
16348 (pass_data_branch_target_load_optimize1): ...new pass_data instance
16350 (make_pass_branch_target_load_optimize1): ...new function.
16351 (pass_branch_target_load_optimize2): Convert from a global struct to a
16352 subclass of rtl_opt_pass along with...
16353 (pass_data_branch_target_load_optimize2): ...new pass_data instance
16355 (make_pass_branch_target_load_optimize2): ...new function.
16356 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
16357 of rtl_opt_pass along with...
16358 (pass_data_jump): ...new pass_data instance and...
16359 (make_pass_jump): ...new function.
16360 (pass_jump2): Convert from a global struct to a subclass of
16361 rtl_opt_pass along with...
16362 (pass_data_jump2): ...new pass_data instance and...
16363 (make_pass_jump2): ...new function.
16364 * cfgexpand.c (pass_expand): Convert from a global struct to a
16365 subclass of rtl_opt_pass along with...
16366 (pass_data_expand): ...new pass_data instance and...
16367 (make_pass_expand): ...new function.
16368 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
16369 of rtl_opt_pass along with...
16370 (pass_data_free_cfg): ...new pass_data instance and...
16371 (make_pass_free_cfg): ...new function.
16372 (pass_into_cfg_layout_mode): Convert from a global struct to a
16373 subclass of rtl_opt_pass along with...
16374 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
16375 (make_pass_into_cfg_layout_mode): ...new function.
16376 (pass_outof_cfg_layout_mode): Convert from a global struct to a
16377 subclass of rtl_opt_pass along with...
16378 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
16379 (make_pass_outof_cfg_layout_mode): ...new function.
16380 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
16381 struct to a subclass of gimple_opt_pass along with...
16382 (pass_data_build_cgraph_edges): ...new pass_data instance and...
16383 (make_pass_build_cgraph_edges): ...new function.
16384 (pass_rebuild_cgraph_edges): Convert from a global struct to a
16385 subclass of gimple_opt_pass along with...
16386 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
16387 (make_pass_rebuild_cgraph_edges): ...new function.
16388 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
16389 subclass of gimple_opt_pass along with...
16390 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
16392 (make_pass_remove_cgraph_callee_edges): ...new function.
16393 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
16394 struct to a subclass of rtl_opt_pass along with...
16395 (pass_data_stack_adjustments): ...new pass_data instance and...
16396 (make_pass_stack_adjustments): ...new function.
16397 * combine.c (pass_combine): Convert from a global struct to a subclass
16398 of rtl_opt_pass along with...
16399 (pass_data_combine): ...new pass_data instance and...
16400 (make_pass_combine): ...new function.
16401 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
16402 global struct to a subclass of rtl_opt_pass along with...
16403 (pass_data_compare_elim_after_reload): ...new pass_data instance
16405 (make_pass_compare_elim_after_reload): ...new function.
16406 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
16407 of rtl_opt_pass along with...
16408 (pass_data_rtl_cprop): ...new pass_data instance and...
16409 (make_pass_rtl_cprop): ...new function.
16410 * cse.c (pass_cse): Convert from a global struct to a subclass of
16411 rtl_opt_pass along with...
16412 (pass_data_cse): ...new pass_data instance and...
16413 (make_pass_cse): ...new function.
16414 (pass_cse2): Convert from a global struct to a subclass of
16415 rtl_opt_pass along with...
16416 (pass_data_cse2): ...new pass_data instance and...
16417 (make_pass_cse2): ...new function.
16418 (pass_cse_after_global_opts): Convert from a global struct to a
16419 subclass of rtl_opt_pass along with...
16420 (pass_data_cse_after_global_opts): ...new pass_data instance and...
16421 (make_pass_cse_after_global_opts): ...new function.
16422 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
16423 of rtl_opt_pass along with...
16424 (pass_data_ud_rtl_dce): ...new pass_data instance and...
16425 (make_pass_ud_rtl_dce): ...new function.
16426 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
16427 rtl_opt_pass along with...
16428 (pass_data_fast_rtl_dce): ...new pass_data instance and...
16429 (make_pass_fast_rtl_dce): ...new function.
16430 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
16431 a subclass of rtl_opt_pass along with...
16432 (pass_data_df_initialize_opt): ...new pass_data instance and...
16433 (make_pass_df_initialize_opt): ...new function.
16434 (pass_df_initialize_no_opt): Convert from a global struct to a
16435 subclass of rtl_opt_pass along with...
16436 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
16437 (make_pass_df_initialize_no_opt): ...new function.
16438 (pass_df_finish): Convert from a global struct to a subclass of
16439 rtl_opt_pass along with...
16440 (pass_data_df_finish): ...new pass_data instance and...
16441 (make_pass_df_finish): ...new function.
16442 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
16443 rtl_opt_pass along with...
16444 (pass_data_rtl_dse1): ...new pass_data instance and...
16445 (make_pass_rtl_dse1): ...new function.
16446 (pass_rtl_dse2): Convert from a global struct to a subclass of
16447 rtl_opt_pass along with...
16448 (pass_data_rtl_dse2): ...new pass_data instance and...
16449 (make_pass_rtl_dse2): ...new function.
16450 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
16451 subclass of rtl_opt_pass along with...
16452 (pass_data_dwarf2_frame): ...new pass_data instance and...
16453 (make_pass_dwarf2_frame): ...new function.
16454 * except.c (pass_set_nothrow_function_flags): Convert from a global
16455 struct to a subclass of rtl_opt_pass along with...
16456 (pass_data_set_nothrow_function_flags): ...new pass_data instance
16458 (make_pass_set_nothrow_function_flags): ...new function.
16459 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
16460 subclass of rtl_opt_pass along with...
16461 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
16463 (make_pass_convert_to_eh_region_ranges): ...new function.
16464 * final.c (pass_compute_alignments): Convert from a global struct to a
16465 subclass of rtl_opt_pass along with...
16466 (pass_data_compute_alignments): ...new pass_data instance and...
16467 (make_pass_compute_alignments): ...new function.
16468 (pass_final): Convert from a global struct to a subclass of
16469 rtl_opt_pass along with...
16470 (pass_data_final): ...new pass_data instance and...
16471 (make_pass_final): ...new function.
16472 (pass_shorten_branches): Convert from a global struct to a subclass of
16473 rtl_opt_pass along with...
16474 (pass_data_shorten_branches): ...new pass_data instance and...
16475 (make_pass_shorten_branches): ...new function.
16476 (pass_clean_state): Convert from a global struct to a subclass of
16477 rtl_opt_pass along with...
16478 (pass_data_clean_state): ...new pass_data instance and...
16479 (make_pass_clean_state): ...new function.
16480 * function.c (pass_instantiate_virtual_regs): Convert from a global
16481 struct to a subclass of rtl_opt_pass along with...
16482 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
16483 (make_pass_instantiate_virtual_regs): ...new function.
16484 (pass_leaf_regs): Convert from a global struct to a subclass of
16485 rtl_opt_pass along with...
16486 (pass_data_leaf_regs): ...new pass_data instance and...
16487 (make_pass_leaf_regs): ...new function.
16488 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
16489 subclass of rtl_opt_pass along with...
16490 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
16492 (make_pass_thread_prologue_and_epilogue): ...new function.
16493 (pass_match_asm_constraints): Convert from a global struct to a
16494 subclass of rtl_opt_pass along with...
16495 (pass_data_match_asm_constraints): ...new pass_data instance and...
16496 (make_pass_match_asm_constraints): ...new function.
16497 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
16498 subclass of rtl_opt_pass along with...
16499 (pass_data_rtl_fwprop): ...new pass_data instance and...
16500 (make_pass_rtl_fwprop): ...new function.
16501 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
16502 rtl_opt_pass along with...
16503 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
16504 (make_pass_rtl_fwprop_addr): ...new function.
16505 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
16506 rtl_opt_pass along with...
16507 (pass_data_rtl_pre): ...new pass_data instance and...
16508 (make_pass_rtl_pre): ...new function.
16509 (pass_rtl_hoist): Convert from a global struct to a subclass of
16510 rtl_opt_pass along with...
16511 (pass_data_rtl_hoist): ...new pass_data instance and...
16512 (make_pass_rtl_hoist): ...new function.
16513 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
16514 subclass of gimple_opt_pass along with...
16515 (pass_data_lower_cf): ...new pass_data instance and...
16516 (make_pass_lower_cf): ...new function.
16517 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
16518 from a global struct to a subclass of gimple_opt_pass along with...
16519 (pass_data_strength_reduction): ...new pass_data instance and...
16520 (make_pass_strength_reduction): ...new function.
16521 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
16522 of rtl_opt_pass along with...
16523 (pass_data_rtl_ifcvt): ...new pass_data instance and...
16524 (make_pass_rtl_ifcvt): ...new function.
16525 (pass_if_after_combine): Convert from a global struct to a subclass of
16526 rtl_opt_pass along with...
16527 (pass_data_if_after_combine): ...new pass_data instance and...
16528 (make_pass_if_after_combine): ...new function.
16529 (pass_if_after_reload): Convert from a global struct to a subclass of
16530 rtl_opt_pass along with...
16531 (pass_data_if_after_reload): ...new pass_data instance and...
16532 (make_pass_if_after_reload): ...new function.
16533 * init-regs.c (pass_initialize_regs): Convert from a global struct to
16534 a subclass of rtl_opt_pass along with...
16535 (pass_data_initialize_regs): ...new pass_data instance and...
16536 (make_pass_initialize_regs): ...new function.
16537 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
16538 of ipa_opt_pass_d along with...
16539 (pass_data_ipa_cp): ...new pass_data instance and...
16540 (make_pass_ipa_cp): ...new function.
16541 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
16542 global struct to a subclass of gimple_opt_pass along with...
16543 (pass_data_inline_parameters): ...new pass_data instance and...
16544 (make_pass_inline_parameters): ...new function.
16545 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
16546 subclass of gimple_opt_pass along with...
16547 (pass_data_early_inline): ...new pass_data instance and...
16548 (make_pass_early_inline): ...new function.
16549 (pass_ipa_inline): Convert from a global struct to a subclass of
16550 ipa_opt_pass_d along with...
16551 (pass_data_ipa_inline): ...new pass_data instance and...
16552 (make_pass_ipa_inline): ...new function.
16553 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
16554 struct to a subclass of gimple_opt_pass along with...
16555 (pass_data_local_pure_const): ...new pass_data instance and...
16556 (make_pass_local_pure_const): ...new function.
16557 (pass_ipa_pure_const): Convert from a global struct to a subclass of
16558 ipa_opt_pass_d along with...
16559 (pass_data_ipa_pure_const): ...new pass_data instance and...
16560 (make_pass_ipa_pure_const): ...new function.
16561 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
16562 to a subclass of ipa_opt_pass_d along with...
16563 (pass_data_ipa_reference): ...new pass_data instance and...
16564 (make_pass_ipa_reference): ...new function.
16565 * ipa-split.c (pass_split_functions): Convert from a global struct to
16566 a subclass of gimple_opt_pass along with...
16567 (pass_data_split_functions): ...new pass_data instance and...
16568 (make_pass_split_functions): ...new function.
16569 (pass_feedback_split_functions): Convert from a global struct to a
16570 subclass of gimple_opt_pass along with...
16571 (pass_data_feedback_split_functions): ...new pass_data instance and...
16572 (make_pass_feedback_split_functions): ...new function.
16573 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
16574 global struct to a subclass of simple_ipa_opt_pass along with...
16575 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
16577 (make_pass_ipa_function_and_variable_visibility): ...new function.
16578 (pass_ipa_free_inline_summary): Convert from a global struct to a
16579 subclass of simple_ipa_opt_pass along with...
16580 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
16581 (make_pass_ipa_free_inline_summary): ...new function.
16582 (pass_ipa_whole_program_visibility): Convert from a global struct to a
16583 subclass of ipa_opt_pass_d along with...
16584 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
16586 (make_pass_ipa_whole_program_visibility): ...new function.
16587 (pass_ipa_profile): Convert from a global struct to a subclass of
16588 ipa_opt_pass_d along with...
16589 (pass_data_ipa_profile): ...new pass_data instance and...
16590 (make_pass_ipa_profile): ...new function.
16591 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
16592 ipa_opt_pass_d along with...
16593 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
16594 (make_pass_ipa_cdtor_merge): ...new function.
16595 * ira.c (pass_ira): Convert from a global struct to a subclass of
16596 rtl_opt_pass along with...
16597 (pass_data_ira): ...new pass_data instance and...
16598 (make_pass_ira): ...new function.
16599 (pass_reload): Convert from a global struct to a subclass of
16600 rtl_opt_pass along with...
16601 (pass_data_reload): ...new pass_data instance and...
16602 (make_pass_reload): ...new function.
16603 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
16604 subclass of rtl_opt_pass along with...
16605 (pass_data_cleanup_barriers): ...new pass_data instance and...
16606 (make_pass_cleanup_barriers): ...new function.
16607 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
16608 of rtl_opt_pass along with...
16609 (pass_data_loop2): ...new pass_data instance and...
16610 (make_pass_loop2): ...new function.
16611 (pass_rtl_loop_init): Convert from a global struct to a subclass of
16612 rtl_opt_pass along with...
16613 (pass_data_rtl_loop_init): ...new pass_data instance and...
16614 (make_pass_rtl_loop_init): ...new function.
16615 (pass_rtl_loop_done): Convert from a global struct to a subclass of
16616 rtl_opt_pass along with...
16617 (pass_data_rtl_loop_done): ...new pass_data instance and...
16618 (make_pass_rtl_loop_done): ...new function.
16619 (pass_rtl_move_loop_invariants): Convert from a global struct to a
16620 subclass of rtl_opt_pass along with...
16621 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
16622 (make_pass_rtl_move_loop_invariants): ...new function.
16623 (pass_rtl_unswitch): Convert from a global struct to a subclass of
16624 rtl_opt_pass along with...
16625 (pass_data_rtl_unswitch): ...new pass_data instance and...
16626 (make_pass_rtl_unswitch): ...new function.
16627 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
16628 subclass of rtl_opt_pass along with...
16629 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
16631 (make_pass_rtl_unroll_and_peel_loops): ...new function.
16632 (pass_rtl_doloop): Convert from a global struct to a subclass of
16633 rtl_opt_pass along with...
16634 (pass_data_rtl_doloop): ...new pass_data instance and...
16635 (make_pass_rtl_doloop): ...new function.
16636 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
16637 a subclass of rtl_opt_pass along with...
16638 (pass_data_lower_subreg): ...new pass_data instance and...
16639 (make_pass_lower_subreg): ...new function.
16640 (pass_lower_subreg2): Convert from a global struct to a subclass of
16641 rtl_opt_pass along with...
16642 (pass_data_lower_subreg2): ...new pass_data instance and...
16643 (make_pass_lower_subreg2): ...new function.
16644 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
16645 struct to a subclass of ipa_opt_pass_d along with...
16646 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
16647 (make_pass_ipa_lto_gimple_out): ...new function.
16648 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
16649 of ipa_opt_pass_d along with...
16650 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
16651 (make_pass_ipa_lto_finish_out): ...new function.
16652 * mode-switching.c (pass_mode_switching): Convert from a global struct
16653 to a subclass of rtl_opt_pass along with...
16654 (pass_data_mode_switching): ...new pass_data instance and...
16655 (make_pass_mode_switching): ...new function.
16656 * modulo-sched.c (pass_sms): Convert from a global struct to a
16657 subclass of rtl_opt_pass along with...
16658 (pass_data_sms): ...new pass_data instance and...
16659 (make_pass_sms): ...new function.
16660 * omp-low.c (pass_expand_omp): Convert from a global struct to a
16661 subclass of gimple_opt_pass along with...
16662 (pass_data_expand_omp): ...new pass_data instance and...
16663 (make_pass_expand_omp): ...new function.
16664 (pass_lower_omp): Convert from a global struct to a subclass of
16665 gimple_opt_pass along with...
16666 (pass_data_lower_omp): ...new pass_data instance and...
16667 (make_pass_lower_omp): ...new function.
16668 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
16669 of gimple_opt_pass along with...
16670 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
16671 (make_pass_diagnose_omp_blocks): ...new function.
16672 * passes.c (pass_early_local_passes): Convert from a global struct to
16673 a subclass of simple_ipa_opt_pass along with...
16674 (pass_data_early_local_passes): ...new pass_data instance and...
16675 (make_pass_early_local_passes): ...new function.
16676 (pass_all_early_optimizations): Convert from a global struct to a
16677 subclass of gimple_opt_pass along with...
16678 (pass_data_all_early_optimizations): ...new pass_data instance and...
16679 (make_pass_all_early_optimizations): ...new function.
16680 (pass_all_optimizations): Convert from a global struct to a subclass
16681 of gimple_opt_pass along with...
16682 (pass_data_all_optimizations): ...new pass_data instance and...
16683 (make_pass_all_optimizations): ...new function.
16684 (pass_all_optimizations_g): Convert from a global struct to a subclass
16685 of gimple_opt_pass along with...
16686 (pass_data_all_optimizations_g): ...new pass_data instance and...
16687 (make_pass_all_optimizations_g): ...new function.
16688 (pass_rest_of_compilation): Convert from a global struct to a subclass
16689 of rtl_opt_pass along with...
16690 (pass_data_rest_of_compilation): ...new pass_data instance and...
16691 (make_pass_rest_of_compilation): ...new function.
16692 (pass_postreload): Convert from a global struct to a subclass of
16693 rtl_opt_pass along with...
16694 (pass_data_postreload): ...new pass_data instance and...
16695 (make_pass_postreload): ...new function.
16696 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
16697 subclass of rtl_opt_pass along with...
16698 (pass_data_gcse2): ...new pass_data instance and...
16699 (make_pass_gcse2): ...new function.
16700 * postreload.c (pass_postreload_cse): Convert from a global struct to
16701 a subclass of rtl_opt_pass along with...
16702 (pass_data_postreload_cse): ...new pass_data instance and...
16703 (make_pass_postreload_cse): ...new function.
16704 * predict.c (pass_profile): Convert from a global struct to a subclass
16705 of gimple_opt_pass along with...
16706 (pass_data_profile): ...new pass_data instance and...
16707 (make_pass_profile): ...new function.
16708 (pass_strip_predict_hints): Convert from a global struct to a subclass
16709 of gimple_opt_pass along with...
16710 (pass_data_strip_predict_hints): ...new pass_data instance and...
16711 (make_pass_strip_predict_hints): ...new function.
16712 * recog.c (pass_peephole2): Convert from a global struct to a subclass
16713 of rtl_opt_pass along with...
16714 (pass_data_peephole2): ...new pass_data instance and...
16715 (make_pass_peephole2): ...new function.
16716 (pass_split_all_insns): Convert from a global struct to a subclass of
16717 rtl_opt_pass along with...
16718 (pass_data_split_all_insns): ...new pass_data instance and...
16719 (make_pass_split_all_insns): ...new function.
16720 (pass_split_after_reload): Convert from a global struct to a subclass
16721 of rtl_opt_pass along with...
16722 (pass_data_split_after_reload): ...new pass_data instance and...
16723 (make_pass_split_after_reload): ...new function.
16724 (pass_split_before_regstack): Convert from a global struct to a
16725 subclass of rtl_opt_pass along with...
16726 (pass_data_split_before_regstack): ...new pass_data instance and...
16727 (make_pass_split_before_regstack): ...new function.
16728 (pass_split_before_sched2): Convert from a global struct to a subclass
16729 of rtl_opt_pass along with...
16730 (pass_data_split_before_sched2): ...new pass_data instance and...
16731 (make_pass_split_before_sched2): ...new function.
16732 (pass_split_for_shorten_branches): Convert from a global struct to a
16733 subclass of rtl_opt_pass along with...
16734 (pass_data_split_for_shorten_branches): ...new pass_data instance
16736 (make_pass_split_for_shorten_branches): ...new function.
16737 * ree.c (pass_ree): Convert from a global struct to a subclass of
16738 rtl_opt_pass along with...
16739 (pass_data_ree): ...new pass_data instance and...
16740 (make_pass_ree): ...new function.
16741 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
16742 subclass of rtl_opt_pass along with...
16743 (pass_data_stack_regs): ...new pass_data instance and...
16744 (make_pass_stack_regs): ...new function.
16745 (pass_stack_regs_run): Convert from a global struct to a subclass of
16746 rtl_opt_pass along with...
16747 (pass_data_stack_regs_run): ...new pass_data instance and...
16748 (make_pass_stack_regs_run): ...new function.
16749 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
16750 subclass of rtl_opt_pass along with...
16751 (pass_data_cprop_hardreg): ...new pass_data instance and...
16752 (make_pass_cprop_hardreg): ...new function.
16753 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
16754 subclass of rtl_opt_pass along with...
16755 (pass_data_reginfo_init): ...new pass_data instance and...
16756 (make_pass_reginfo_init): ...new function.
16757 * regmove.c (pass_regmove): Convert from a global struct to a subclass
16758 of rtl_opt_pass along with...
16759 (pass_data_regmove): ...new pass_data instance and...
16760 (make_pass_regmove): ...new function.
16761 * regrename.c (pass_regrename): Convert from a global struct to a
16762 subclass of rtl_opt_pass along with...
16763 (pass_data_regrename): ...new pass_data instance and...
16764 (make_pass_regrename): ...new function.
16765 * reorg.c (pass_delay_slots): Convert from a global struct to a
16766 subclass of rtl_opt_pass along with...
16767 (pass_data_delay_slots): ...new pass_data instance and...
16768 (make_pass_delay_slots): ...new function.
16769 (pass_machine_reorg): Convert from a global struct to a subclass of
16770 rtl_opt_pass along with...
16771 (pass_data_machine_reorg): ...new pass_data instance and...
16772 (make_pass_machine_reorg): ...new function.
16773 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
16774 of rtl_opt_pass along with...
16775 (pass_data_sched): ...new pass_data instance and...
16776 (make_pass_sched): ...new function.
16777 (pass_sched2): Convert from a global struct to a subclass of
16778 rtl_opt_pass along with...
16779 (pass_data_sched2): ...new pass_data instance and...
16780 (make_pass_sched2): ...new function.
16781 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
16782 to a subclass of rtl_opt_pass along with...
16783 (pass_data_stack_ptr_mod): ...new pass_data instance and...
16784 (make_pass_stack_ptr_mod): ...new function.
16785 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
16786 to a subclass of rtl_opt_pass along with...
16787 (pass_data_rtl_store_motion): ...new pass_data instance and...
16788 (make_pass_rtl_store_motion): ...new function.
16789 * tracer.c (pass_tracer): Convert from a global struct to a subclass
16790 of gimple_opt_pass along with...
16791 (pass_data_tracer): ...new pass_data instance and...
16792 (make_pass_tracer): ...new function.
16793 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
16794 to a subclass of gimple_opt_pass along with...
16795 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
16796 (make_pass_diagnose_tm_blocks): ...new function.
16797 (pass_lower_tm): Convert from a global struct to a subclass of
16798 gimple_opt_pass along with...
16799 (pass_data_lower_tm): ...new pass_data instance and...
16800 (make_pass_lower_tm): ...new function.
16801 (pass_tm_init): Convert from a global struct to a subclass of
16802 gimple_opt_pass along with...
16803 (pass_data_tm_init): ...new pass_data instance and...
16804 (make_pass_tm_init): ...new function.
16805 (pass_tm_mark): Convert from a global struct to a subclass of
16806 gimple_opt_pass along with...
16807 (pass_data_tm_mark): ...new pass_data instance and...
16808 (make_pass_tm_mark): ...new function.
16809 (pass_tm_edges): Convert from a global struct to a subclass of
16810 gimple_opt_pass along with...
16811 (pass_data_tm_edges): ...new pass_data instance and...
16812 (make_pass_tm_edges): ...new function.
16813 (pass_tm_memopt): Convert from a global struct to a subclass of
16814 gimple_opt_pass along with...
16815 (pass_data_tm_memopt): ...new pass_data instance and...
16816 (make_pass_tm_memopt): ...new function.
16817 (pass_ipa_tm): Convert from a global struct to a subclass of
16818 simple_ipa_opt_pass along with...
16819 (pass_data_ipa_tm): ...new pass_data instance and...
16820 (make_pass_ipa_tm): ...new function.
16821 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
16822 subclass of gimple_opt_pass along with...
16823 (pass_data_call_cdce): ...new pass_data instance and...
16824 (make_pass_call_cdce): ...new function.
16825 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
16826 subclass of gimple_opt_pass along with...
16827 (pass_data_build_cfg): ...new pass_data instance and...
16828 (make_pass_build_cfg): ...new function.
16829 (pass_split_crit_edges): Convert from a global struct to a subclass of
16830 gimple_opt_pass along with...
16831 (pass_data_split_crit_edges): ...new pass_data instance and...
16832 (make_pass_split_crit_edges): ...new function.
16833 (pass_warn_function_return): Convert from a global struct to a
16834 subclass of gimple_opt_pass along with...
16835 (pass_data_warn_function_return): ...new pass_data instance and...
16836 (make_pass_warn_function_return): ...new function.
16837 (pass_warn_function_noreturn): Convert from a global struct to a
16838 subclass of gimple_opt_pass along with...
16839 (pass_data_warn_function_noreturn): ...new pass_data instance and...
16840 (make_pass_warn_function_noreturn): ...new function.
16841 (pass_warn_unused_result): Convert from a global struct to a subclass
16842 of gimple_opt_pass along with...
16843 (pass_data_warn_unused_result): ...new pass_data instance and...
16844 (make_pass_warn_unused_result): ...new function.
16845 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
16846 a subclass of gimple_opt_pass along with...
16847 (pass_data_merge_phi): ...new pass_data instance and...
16848 (make_pass_merge_phi): ...new function.
16849 * tree-complex.c (pass_lower_complex): Convert from a global struct to
16850 a subclass of gimple_opt_pass along with...
16851 (pass_data_lower_complex): ...new pass_data instance and...
16852 (make_pass_lower_complex): ...new function.
16853 (pass_lower_complex_O0): Convert from a global struct to a subclass of
16854 gimple_opt_pass along with...
16855 (pass_data_lower_complex_O0): ...new pass_data instance and...
16856 (make_pass_lower_complex_O0): ...new function.
16857 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
16858 subclass of gimple_opt_pass along with...
16859 (pass_data_lower_eh): ...new pass_data instance and...
16860 (make_pass_lower_eh): ...new function.
16861 (pass_refactor_eh): Convert from a global struct to a subclass of
16862 gimple_opt_pass along with...
16863 (pass_data_refactor_eh): ...new pass_data instance and...
16864 (make_pass_refactor_eh): ...new function.
16865 (pass_lower_resx): Convert from a global struct to a subclass of
16866 gimple_opt_pass along with...
16867 (pass_data_lower_resx): ...new pass_data instance and...
16868 (make_pass_lower_resx): ...new function.
16869 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
16870 of gimple_opt_pass along with...
16871 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
16872 (make_pass_lower_eh_dispatch): ...new function.
16873 (pass_cleanup_eh): Convert from a global struct to a subclass of
16874 gimple_opt_pass along with...
16875 (pass_data_cleanup_eh): ...new pass_data instance and...
16876 (make_pass_cleanup_eh): ...new function.
16877 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
16878 to a subclass of simple_ipa_opt_pass along with...
16879 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
16880 (make_pass_ipa_lower_emutls): ...new function.
16881 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
16882 a subclass of gimple_opt_pass along with...
16883 (pass_data_if_conversion): ...new pass_data instance and...
16884 (make_pass_if_conversion): ...new function.
16885 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
16886 subclass of gimple_opt_pass along with...
16887 (pass_data_build_ssa): ...new pass_data instance and...
16888 (make_pass_build_ssa): ...new function.
16889 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
16890 global struct to a subclass of gimple_opt_pass along with...
16891 (pass_data_loop_distribution): ...new pass_data instance and...
16892 (make_pass_loop_distribution): ...new function.
16893 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
16894 subclass of gimple_opt_pass along with...
16895 (pass_data_mudflap_1): ...new pass_data instance and...
16896 (make_pass_mudflap_1): ...new function.
16897 (pass_mudflap_2): Convert from a global struct to a subclass of
16898 gimple_opt_pass along with...
16899 (pass_data_mudflap_2): ...new pass_data instance and...
16900 (make_pass_mudflap_2): ...new function.
16901 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
16902 subclass of gimple_opt_pass along with...
16903 (pass_data_mudflap_1): ...new pass_data instance and...
16904 (make_pass_mudflap_1): ...new function.
16905 (pass_mudflap_2): Convert from a global struct to a subclass of
16906 gimple_opt_pass along with...
16907 (pass_data_mudflap_2): ...new pass_data instance and...
16908 (make_pass_mudflap_2): ...new function.
16909 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
16910 gimple_opt_pass along with...
16911 (pass_data_nrv): ...new pass_data instance and...
16912 (make_pass_nrv): ...new function.
16913 (pass_return_slot): Convert from a global struct to a subclass of
16914 gimple_opt_pass along with...
16915 (pass_data_return_slot): ...new pass_data instance and...
16916 (make_pass_return_slot): ...new function.
16917 * tree-object-size.c (pass_object_sizes): Convert from a global struct
16918 to a subclass of gimple_opt_pass along with...
16919 (pass_data_object_sizes): ...new pass_data instance and...
16920 (make_pass_object_sizes): ...new function.
16921 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
16922 global struct to a subclass of gimple_opt_pass along with...
16923 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
16925 (make_pass_cleanup_cfg_post_optimizing): ...new function.
16926 (pass_fixup_cfg): Convert from a global struct to a subclass of
16927 gimple_opt_pass along with...
16928 (pass_data_fixup_cfg): ...new pass_data instance and...
16929 (make_pass_fixup_cfg): ...new function.
16930 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
16931 (make_pass_mudflap_1): ...new function.
16932 (pass_mudflap_2): Replace declaration with that of...
16933 (make_pass_mudflap_2): ...new function.
16934 (pass_asan): Replace declaration with that of...
16935 (make_pass_asan): ...new function.
16936 (pass_asan_O0): Replace declaration with that of...
16937 (make_pass_asan_O0): ...new function.
16938 (pass_tsan): Replace declaration with that of...
16939 (make_pass_tsan): ...new function.
16940 (pass_tsan_O0): Replace declaration with that of...
16941 (make_pass_tsan_O0): ...new function.
16942 (pass_lower_cf): Replace declaration with that of...
16943 (make_pass_lower_cf): ...new function.
16944 (pass_refactor_eh): Replace declaration with that of...
16945 (make_pass_refactor_eh): ...new function.
16946 (pass_lower_eh): Replace declaration with that of...
16947 (make_pass_lower_eh): ...new function.
16948 (pass_lower_eh_dispatch): Replace declaration with that of...
16949 (make_pass_lower_eh_dispatch): ...new function.
16950 (pass_lower_resx): Replace declaration with that of...
16951 (make_pass_lower_resx): ...new function.
16952 (pass_build_cfg): Replace declaration with that of...
16953 (make_pass_build_cfg): ...new function.
16954 (pass_early_tree_profile): Replace declaration with that of...
16955 (make_pass_early_tree_profile): ...new function.
16956 (pass_cleanup_eh): Replace declaration with that of...
16957 (make_pass_cleanup_eh): ...new function.
16958 (pass_sra): Replace declaration with that of...
16959 (make_pass_sra): ...new function.
16960 (pass_sra_early): Replace declaration with that of...
16961 (make_pass_sra_early): ...new function.
16962 (pass_early_ipa_sra): Replace declaration with that of...
16963 (make_pass_early_ipa_sra): ...new function.
16964 (pass_tail_recursion): Replace declaration with that of...
16965 (make_pass_tail_recursion): ...new function.
16966 (pass_tail_calls): Replace declaration with that of...
16967 (make_pass_tail_calls): ...new function.
16968 (pass_tree_loop): Replace declaration with that of...
16969 (make_pass_tree_loop): ...new function.
16970 (pass_tree_loop_init): Replace declaration with that of...
16971 (make_pass_tree_loop_init): ...new function.
16972 (pass_lim): Replace declaration with that of...
16973 (make_pass_lim): ...new function.
16974 (pass_tree_unswitch): Replace declaration with that of...
16975 (make_pass_tree_unswitch): ...new function.
16976 (pass_predcom): Replace declaration with that of...
16977 (make_pass_predcom): ...new function.
16978 (pass_iv_canon): Replace declaration with that of...
16979 (make_pass_iv_canon): ...new function.
16980 (pass_scev_cprop): Replace declaration with that of...
16981 (make_pass_scev_cprop): ...new function.
16982 (pass_empty_loop): Replace declaration with that of...
16983 (make_pass_empty_loop): ...new function.
16984 (pass_record_bounds): Replace declaration with that of...
16985 (make_pass_record_bounds): ...new function.
16986 (pass_graphite): Replace declaration with that of...
16987 (make_pass_graphite): ...new function.
16988 (pass_graphite_transforms): Replace declaration with that of...
16989 (make_pass_graphite_transforms): ...new function.
16990 (pass_if_conversion): Replace declaration with that of...
16991 (make_pass_if_conversion): ...new function.
16992 (pass_loop_distribution): Replace declaration with that of...
16993 (make_pass_loop_distribution): ...new function.
16994 (pass_vectorize): Replace declaration with that of...
16995 (make_pass_vectorize): ...new function.
16996 (pass_slp_vectorize): Replace declaration with that of...
16997 (make_pass_slp_vectorize): ...new function.
16998 (pass_complete_unroll): Replace declaration with that of...
16999 (make_pass_complete_unroll): ...new function.
17000 (pass_complete_unrolli): Replace declaration with that of...
17001 (make_pass_complete_unrolli): ...new function.
17002 (pass_parallelize_loops): Replace declaration with that of...
17003 (make_pass_parallelize_loops): ...new function.
17004 (pass_loop_prefetch): Replace declaration with that of...
17005 (make_pass_loop_prefetch): ...new function.
17006 (pass_iv_optimize): Replace declaration with that of...
17007 (make_pass_iv_optimize): ...new function.
17008 (pass_tree_loop_done): Replace declaration with that of...
17009 (make_pass_tree_loop_done): ...new function.
17010 (pass_ch): Replace declaration with that of...
17011 (make_pass_ch): ...new function.
17012 (pass_ccp): Replace declaration with that of...
17013 (make_pass_ccp): ...new function.
17014 (pass_phi_only_cprop): Replace declaration with that of...
17015 (make_pass_phi_only_cprop): ...new function.
17016 (pass_build_ssa): Replace declaration with that of...
17017 (make_pass_build_ssa): ...new function.
17018 (pass_build_alias): Replace declaration with that of...
17019 (make_pass_build_alias): ...new function.
17020 (pass_build_ealias): Replace declaration with that of...
17021 (make_pass_build_ealias): ...new function.
17022 (pass_dominator): Replace declaration with that of...
17023 (make_pass_dominator): ...new function.
17024 (pass_dce): Replace declaration with that of...
17025 (make_pass_dce): ...new function.
17026 (pass_dce_loop): Replace declaration with that of...
17027 (make_pass_dce_loop): ...new function.
17028 (pass_cd_dce): Replace declaration with that of...
17029 (make_pass_cd_dce): ...new function.
17030 (pass_call_cdce): Replace declaration with that of...
17031 (make_pass_call_cdce): ...new function.
17032 (pass_merge_phi): Replace declaration with that of...
17033 (make_pass_merge_phi): ...new function.
17034 (pass_split_crit_edges): Replace declaration with that of...
17035 (make_pass_split_crit_edges): ...new function.
17036 (pass_pre): Replace declaration with that of...
17037 (make_pass_pre): ...new function.
17038 (pass_profile): Replace declaration with that of...
17039 (make_pass_profile): ...new function.
17040 (pass_strip_predict_hints): Replace declaration with that of...
17041 (make_pass_strip_predict_hints): ...new function.
17042 (pass_lower_complex_O0): Replace declaration with that of...
17043 (make_pass_lower_complex_O0): ...new function.
17044 (pass_lower_complex): Replace declaration with that of...
17045 (make_pass_lower_complex): ...new function.
17046 (pass_lower_vector): Replace declaration with that of...
17047 (make_pass_lower_vector): ...new function.
17048 (pass_lower_vector_ssa): Replace declaration with that of...
17049 (make_pass_lower_vector_ssa): ...new function.
17050 (pass_lower_omp): Replace declaration with that of...
17051 (make_pass_lower_omp): ...new function.
17052 (pass_diagnose_omp_blocks): Replace declaration with that of...
17053 (make_pass_diagnose_omp_blocks): ...new function.
17054 (pass_expand_omp): Replace declaration with that of...
17055 (make_pass_expand_omp): ...new function.
17056 (pass_expand_omp_ssa): Replace declaration with that of...
17057 (make_pass_expand_omp_ssa): ...new function.
17058 (pass_object_sizes): Replace declaration with that of...
17059 (make_pass_object_sizes): ...new function.
17060 (pass_strlen): Replace declaration with that of...
17061 (make_pass_strlen): ...new function.
17062 (pass_fold_builtins): Replace declaration with that of...
17063 (make_pass_fold_builtins): ...new function.
17064 (pass_stdarg): Replace declaration with that of...
17065 (make_pass_stdarg): ...new function.
17066 (pass_early_warn_uninitialized): Replace declaration with that of...
17067 (make_pass_early_warn_uninitialized): ...new function.
17068 (pass_late_warn_uninitialized): Replace declaration with that of...
17069 (make_pass_late_warn_uninitialized): ...new function.
17070 (pass_cse_reciprocals): Replace declaration with that of...
17071 (make_pass_cse_reciprocals): ...new function.
17072 (pass_cse_sincos): Replace declaration with that of...
17073 (make_pass_cse_sincos): ...new function.
17074 (pass_optimize_bswap): Replace declaration with that of...
17075 (make_pass_optimize_bswap): ...new function.
17076 (pass_optimize_widening_mul): Replace declaration with that of...
17077 (make_pass_optimize_widening_mul): ...new function.
17078 (pass_warn_function_return): Replace declaration with that of...
17079 (make_pass_warn_function_return): ...new function.
17080 (pass_warn_function_noreturn): Replace declaration with that of...
17081 (make_pass_warn_function_noreturn): ...new function.
17082 (pass_cselim): Replace declaration with that of...
17083 (make_pass_cselim): ...new function.
17084 (pass_phiopt): Replace declaration with that of...
17085 (make_pass_phiopt): ...new function.
17086 (pass_forwprop): Replace declaration with that of...
17087 (make_pass_forwprop): ...new function.
17088 (pass_phiprop): Replace declaration with that of...
17089 (make_pass_phiprop): ...new function.
17090 (pass_tree_ifcombine): Replace declaration with that of...
17091 (make_pass_tree_ifcombine): ...new function.
17092 (pass_dse): Replace declaration with that of...
17093 (make_pass_dse): ...new function.
17094 (pass_nrv): Replace declaration with that of...
17095 (make_pass_nrv): ...new function.
17096 (pass_rename_ssa_copies): Replace declaration with that of...
17097 (make_pass_rename_ssa_copies): ...new function.
17098 (pass_sink_code): Replace declaration with that of...
17099 (make_pass_sink_code): ...new function.
17100 (pass_fre): Replace declaration with that of...
17101 (make_pass_fre): ...new function.
17102 (pass_check_data_deps): Replace declaration with that of...
17103 (make_pass_check_data_deps): ...new function.
17104 (pass_copy_prop): Replace declaration with that of...
17105 (make_pass_copy_prop): ...new function.
17106 (pass_vrp): Replace declaration with that of...
17107 (make_pass_vrp): ...new function.
17108 (pass_uncprop): Replace declaration with that of...
17109 (make_pass_uncprop): ...new function.
17110 (pass_return_slot): Replace declaration with that of...
17111 (make_pass_return_slot): ...new function.
17112 (pass_reassoc): Replace declaration with that of...
17113 (make_pass_reassoc): ...new function.
17114 (pass_rebuild_cgraph_edges): Replace declaration with that of...
17115 (make_pass_rebuild_cgraph_edges): ...new function.
17116 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
17117 (make_pass_remove_cgraph_callee_edges): ...new function.
17118 (pass_build_cgraph_edges): Replace declaration with that of...
17119 (make_pass_build_cgraph_edges): ...new function.
17120 (pass_local_pure_const): Replace declaration with that of...
17121 (make_pass_local_pure_const): ...new function.
17122 (pass_tracer): Replace declaration with that of...
17123 (make_pass_tracer): ...new function.
17124 (pass_warn_unused_result): Replace declaration with that of...
17125 (make_pass_warn_unused_result): ...new function.
17126 (pass_diagnose_tm_blocks): Replace declaration with that of...
17127 (make_pass_diagnose_tm_blocks): ...new function.
17128 (pass_lower_tm): Replace declaration with that of...
17129 (make_pass_lower_tm): ...new function.
17130 (pass_tm_init): Replace declaration with that of...
17131 (make_pass_tm_init): ...new function.
17132 (pass_tm_mark): Replace declaration with that of...
17133 (make_pass_tm_mark): ...new function.
17134 (pass_tm_memopt): Replace declaration with that of...
17135 (make_pass_tm_memopt): ...new function.
17136 (pass_tm_edges): Replace declaration with that of...
17137 (make_pass_tm_edges): ...new function.
17138 (pass_split_functions): Replace declaration with that of...
17139 (make_pass_split_functions): ...new function.
17140 (pass_feedback_split_functions): Replace declaration with that of...
17141 (make_pass_feedback_split_functions): ...new function.
17142 (pass_strength_reduction): Replace declaration with that of...
17143 (make_pass_strength_reduction): ...new function.
17144 (pass_ipa_lower_emutls): Replace declaration with that of...
17145 (make_pass_ipa_lower_emutls): ...new function.
17146 (pass_ipa_function_and_variable_visibility): Replace declaration with
17148 (make_pass_ipa_function_and_variable_visibility): ...new function.
17149 (pass_ipa_tree_profile): Replace declaration with that of...
17150 (make_pass_ipa_tree_profile): ...new function.
17151 (pass_early_local_passes): Replace declaration with that of...
17152 (make_pass_early_local_passes): ...new function.
17153 (pass_ipa_whole_program_visibility): Replace declaration with that
17155 (make_pass_ipa_whole_program_visibility): ...new function.
17156 (pass_ipa_lto_gimple_out): Replace declaration with that of...
17157 (make_pass_ipa_lto_gimple_out): ...new function.
17158 (pass_ipa_increase_alignment): Replace declaration with that of...
17159 (make_pass_ipa_increase_alignment): ...new function.
17160 (pass_ipa_inline): Replace declaration with that of...
17161 (make_pass_ipa_inline): ...new function.
17162 (pass_ipa_free_lang_data): Replace declaration with that of...
17163 (make_pass_ipa_free_lang_data): ...new function.
17164 (pass_ipa_free_inline_summary): Replace declaration with that of...
17165 (make_pass_ipa_free_inline_summary): ...new function.
17166 (pass_ipa_cp): Replace declaration with that of...
17167 (make_pass_ipa_cp): ...new function.
17168 (pass_ipa_reference): Replace declaration with that of...
17169 (make_pass_ipa_reference): ...new function.
17170 (pass_ipa_pure_const): Replace declaration with that of...
17171 (make_pass_ipa_pure_const): ...new function.
17172 (pass_ipa_pta): Replace declaration with that of...
17173 (make_pass_ipa_pta): ...new function.
17174 (pass_ipa_lto_finish_out): Replace declaration with that of...
17175 (make_pass_ipa_lto_finish_out): ...new function.
17176 (pass_ipa_tm): Replace declaration with that of...
17177 (make_pass_ipa_tm): ...new function.
17178 (pass_ipa_profile): Replace declaration with that of...
17179 (make_pass_ipa_profile): ...new function.
17180 (pass_ipa_cdtor_merge): Replace declaration with that of...
17181 (make_pass_ipa_cdtor_merge): ...new function.
17182 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
17184 (make_pass_cleanup_cfg_post_optimizing): ...new function.
17185 (pass_init_datastructures): Replace declaration with that of...
17186 (make_pass_init_datastructures): ...new function.
17187 (pass_fixup_cfg): Replace declaration with that of...
17188 (make_pass_fixup_cfg): ...new function.
17189 (pass_expand): Replace declaration with that of...
17190 (make_pass_expand): ...new function.
17191 (pass_instantiate_virtual_regs): Replace declaration with that of...
17192 (make_pass_instantiate_virtual_regs): ...new function.
17193 (pass_rtl_fwprop): Replace declaration with that of...
17194 (make_pass_rtl_fwprop): ...new function.
17195 (pass_rtl_fwprop_addr): Replace declaration with that of...
17196 (make_pass_rtl_fwprop_addr): ...new function.
17197 (pass_jump): Replace declaration with that of...
17198 (make_pass_jump): ...new function.
17199 (pass_jump2): Replace declaration with that of...
17200 (make_pass_jump2): ...new function.
17201 (pass_lower_subreg): Replace declaration with that of...
17202 (make_pass_lower_subreg): ...new function.
17203 (pass_cse): Replace declaration with that of...
17204 (make_pass_cse): ...new function.
17205 (pass_fast_rtl_dce): Replace declaration with that of...
17206 (make_pass_fast_rtl_dce): ...new function.
17207 (pass_ud_rtl_dce): Replace declaration with that of...
17208 (make_pass_ud_rtl_dce): ...new function.
17209 (pass_rtl_dce): Replace declaration with that of...
17210 (make_pass_rtl_dce): ...new function.
17211 (pass_rtl_dse1): Replace declaration with that of...
17212 (make_pass_rtl_dse1): ...new function.
17213 (pass_rtl_dse2): Replace declaration with that of...
17214 (make_pass_rtl_dse2): ...new function.
17215 (pass_rtl_dse3): Replace declaration with that of...
17216 (make_pass_rtl_dse3): ...new function.
17217 (pass_rtl_cprop): Replace declaration with that of...
17218 (make_pass_rtl_cprop): ...new function.
17219 (pass_rtl_pre): Replace declaration with that of...
17220 (make_pass_rtl_pre): ...new function.
17221 (pass_rtl_hoist): Replace declaration with that of...
17222 (make_pass_rtl_hoist): ...new function.
17223 (pass_rtl_store_motion): Replace declaration with that of...
17224 (make_pass_rtl_store_motion): ...new function.
17225 (pass_cse_after_global_opts): Replace declaration with that of...
17226 (make_pass_cse_after_global_opts): ...new function.
17227 (pass_rtl_ifcvt): Replace declaration with that of...
17228 (make_pass_rtl_ifcvt): ...new function.
17229 (pass_into_cfg_layout_mode): Replace declaration with that of...
17230 (make_pass_into_cfg_layout_mode): ...new function.
17231 (pass_outof_cfg_layout_mode): Replace declaration with that of...
17232 (make_pass_outof_cfg_layout_mode): ...new function.
17233 (pass_loop2): Replace declaration with that of...
17234 (make_pass_loop2): ...new function.
17235 (pass_rtl_loop_init): Replace declaration with that of...
17236 (make_pass_rtl_loop_init): ...new function.
17237 (pass_rtl_move_loop_invariants): Replace declaration with that of...
17238 (make_pass_rtl_move_loop_invariants): ...new function.
17239 (pass_rtl_unswitch): Replace declaration with that of...
17240 (make_pass_rtl_unswitch): ...new function.
17241 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
17242 (make_pass_rtl_unroll_and_peel_loops): ...new function.
17243 (pass_rtl_doloop): Replace declaration with that of...
17244 (make_pass_rtl_doloop): ...new function.
17245 (pass_rtl_loop_done): Replace declaration with that of...
17246 (make_pass_rtl_loop_done): ...new function.
17247 (pass_web): Replace declaration with that of...
17248 (make_pass_web): ...new function.
17249 (pass_cse2): Replace declaration with that of...
17250 (make_pass_cse2): ...new function.
17251 (pass_df_initialize_opt): Replace declaration with that of...
17252 (make_pass_df_initialize_opt): ...new function.
17253 (pass_df_initialize_no_opt): Replace declaration with that of...
17254 (make_pass_df_initialize_no_opt): ...new function.
17255 (pass_reginfo_init): Replace declaration with that of...
17256 (make_pass_reginfo_init): ...new function.
17257 (pass_inc_dec): Replace declaration with that of...
17258 (make_pass_inc_dec): ...new function.
17259 (pass_stack_ptr_mod): Replace declaration with that of...
17260 (make_pass_stack_ptr_mod): ...new function.
17261 (pass_initialize_regs): Replace declaration with that of...
17262 (make_pass_initialize_regs): ...new function.
17263 (pass_combine): Replace declaration with that of...
17264 (make_pass_combine): ...new function.
17265 (pass_if_after_combine): Replace declaration with that of...
17266 (make_pass_if_after_combine): ...new function.
17267 (pass_ree): Replace declaration with that of...
17268 (make_pass_ree): ...new function.
17269 (pass_partition_blocks): Replace declaration with that of...
17270 (make_pass_partition_blocks): ...new function.
17271 (pass_match_asm_constraints): Replace declaration with that of...
17272 (make_pass_match_asm_constraints): ...new function.
17273 (pass_regmove): Replace declaration with that of...
17274 (make_pass_regmove): ...new function.
17275 (pass_split_all_insns): Replace declaration with that of...
17276 (make_pass_split_all_insns): ...new function.
17277 (pass_fast_rtl_byte_dce): Replace declaration with that of...
17278 (make_pass_fast_rtl_byte_dce): ...new function.
17279 (pass_lower_subreg2): Replace declaration with that of...
17280 (make_pass_lower_subreg2): ...new function.
17281 (pass_mode_switching): Replace declaration with that of...
17282 (make_pass_mode_switching): ...new function.
17283 (pass_sms): Replace declaration with that of...
17284 (make_pass_sms): ...new function.
17285 (pass_sched): Replace declaration with that of...
17286 (make_pass_sched): ...new function.
17287 (pass_ira): Replace declaration with that of...
17288 (make_pass_ira): ...new function.
17289 (pass_reload): Replace declaration with that of...
17290 (make_pass_reload): ...new function.
17291 (pass_clean_state): Replace declaration with that of...
17292 (make_pass_clean_state): ...new function.
17293 (pass_branch_prob): Replace declaration with that of...
17294 (make_pass_branch_prob): ...new function.
17295 (pass_value_profile_transformations): Replace declaration with that
17297 (make_pass_value_profile_transformations): ...new function.
17298 (pass_postreload_cse): Replace declaration with that of...
17299 (make_pass_postreload_cse): ...new function.
17300 (pass_gcse2): Replace declaration with that of...
17301 (make_pass_gcse2): ...new function.
17302 (pass_split_after_reload): Replace declaration with that of...
17303 (make_pass_split_after_reload): ...new function.
17304 (pass_branch_target_load_optimize1): Replace declaration with that
17306 (make_pass_branch_target_load_optimize1): ...new function.
17307 (pass_thread_prologue_and_epilogue): Replace declaration with that
17309 (make_pass_thread_prologue_and_epilogue): ...new function.
17310 (pass_stack_adjustments): Replace declaration with that of...
17311 (make_pass_stack_adjustments): ...new function.
17312 (pass_peephole2): Replace declaration with that of...
17313 (make_pass_peephole2): ...new function.
17314 (pass_if_after_reload): Replace declaration with that of...
17315 (make_pass_if_after_reload): ...new function.
17316 (pass_regrename): Replace declaration with that of...
17317 (make_pass_regrename): ...new function.
17318 (pass_cprop_hardreg): Replace declaration with that of...
17319 (make_pass_cprop_hardreg): ...new function.
17320 (pass_reorder_blocks): Replace declaration with that of...
17321 (make_pass_reorder_blocks): ...new function.
17322 (pass_branch_target_load_optimize2): Replace declaration with that
17324 (make_pass_branch_target_load_optimize2): ...new function.
17325 (pass_leaf_regs): Replace declaration with that of...
17326 (make_pass_leaf_regs): ...new function.
17327 (pass_split_before_sched2): Replace declaration with that of...
17328 (make_pass_split_before_sched2): ...new function.
17329 (pass_compare_elim_after_reload): Replace declaration with that of...
17330 (make_pass_compare_elim_after_reload): ...new function.
17331 (pass_sched2): Replace declaration with that of...
17332 (make_pass_sched2): ...new function.
17333 (pass_stack_regs): Replace declaration with that of...
17334 (make_pass_stack_regs): ...new function.
17335 (pass_stack_regs_run): Replace declaration with that of...
17336 (make_pass_stack_regs_run): ...new function.
17337 (pass_df_finish): Replace declaration with that of...
17338 (make_pass_df_finish): ...new function.
17339 (pass_compute_alignments): Replace declaration with that of...
17340 (make_pass_compute_alignments): ...new function.
17341 (pass_duplicate_computed_gotos): Replace declaration with that of...
17342 (make_pass_duplicate_computed_gotos): ...new function.
17343 (pass_variable_tracking): Replace declaration with that of...
17344 (make_pass_variable_tracking): ...new function.
17345 (pass_free_cfg): Replace declaration with that of...
17346 (make_pass_free_cfg): ...new function.
17347 (pass_machine_reorg): Replace declaration with that of...
17348 (make_pass_machine_reorg): ...new function.
17349 (pass_cleanup_barriers): Replace declaration with that of...
17350 (make_pass_cleanup_barriers): ...new function.
17351 (pass_delay_slots): Replace declaration with that of...
17352 (make_pass_delay_slots): ...new function.
17353 (pass_split_for_shorten_branches): Replace declaration with that of...
17354 (make_pass_split_for_shorten_branches): ...new function.
17355 (pass_split_before_regstack): Replace declaration with that of...
17356 (make_pass_split_before_regstack): ...new function.
17357 (pass_convert_to_eh_region_ranges): Replace declaration with that
17359 (make_pass_convert_to_eh_region_ranges): ...new function.
17360 (pass_shorten_branches): Replace declaration with that of...
17361 (make_pass_shorten_branches): ...new function.
17362 (pass_set_nothrow_function_flags): Replace declaration with that of...
17363 (make_pass_set_nothrow_function_flags): ...new function.
17364 (pass_dwarf2_frame): Replace declaration with that of...
17365 (make_pass_dwarf2_frame): ...new function.
17366 (pass_final): Replace declaration with that of...
17367 (make_pass_final): ...new function.
17368 (pass_rtl_seqabstr): Replace declaration with that of...
17369 (make_pass_rtl_seqabstr): ...new function.
17370 (pass_release_ssa_names): Replace declaration with that of...
17371 (make_pass_release_ssa_names): ...new function.
17372 (pass_early_inline): Replace declaration with that of...
17373 (make_pass_early_inline): ...new function.
17374 (pass_inline_parameters): Replace declaration with that of...
17375 (make_pass_inline_parameters): ...new function.
17376 (pass_update_address_taken): Replace declaration with that of...
17377 (make_pass_update_address_taken): ...new function.
17378 (pass_convert_switch): Replace declaration with that of...
17379 (make_pass_convert_switch): ...new function.
17380 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
17381 to a subclass of simple_ipa_opt_pass along with...
17382 (pass_data_ipa_tree_profile): ...new pass_data instance and...
17383 (make_pass_ipa_tree_profile): ...new function.
17384 * tree-sra.c (pass_sra_early): Convert from a global struct to a
17385 subclass of gimple_opt_pass along with...
17386 (pass_data_sra_early): ...new pass_data instance and...
17387 (make_pass_sra_early): ...new function.
17388 (pass_sra): Convert from a global struct to a subclass of
17389 gimple_opt_pass along with...
17390 (pass_data_sra): ...new pass_data instance and...
17391 (make_pass_sra): ...new function.
17392 (pass_early_ipa_sra): Convert from a global struct to a subclass of
17393 gimple_opt_pass along with...
17394 (pass_data_early_ipa_sra): ...new pass_data instance and...
17395 (make_pass_early_ipa_sra): ...new function.
17396 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
17397 subclass of gimple_opt_pass along with...
17398 (pass_data_ccp): ...new pass_data instance and...
17399 (make_pass_ccp): ...new function.
17400 (pass_fold_builtins): Convert from a global struct to a subclass of
17401 gimple_opt_pass along with...
17402 (pass_data_fold_builtins): ...new pass_data instance and...
17403 (make_pass_fold_builtins): ...new function.
17404 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
17405 subclass of gimple_opt_pass along with...
17406 (pass_data_copy_prop): ...new pass_data instance and...
17407 (make_pass_copy_prop): ...new function.
17408 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
17409 global struct to a subclass of gimple_opt_pass along with...
17410 (pass_data_rename_ssa_copies): ...new pass_data instance and...
17411 (make_pass_rename_ssa_copies): ...new function.
17412 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
17413 subclass of gimple_opt_pass along with...
17414 (pass_data_dce): ...new pass_data instance and...
17415 (make_pass_dce): ...new function.
17416 (pass_dce_loop): Convert from a global struct to a subclass of
17417 gimple_opt_pass along with...
17418 (pass_data_dce_loop): ...new pass_data instance and...
17419 (make_pass_dce_loop): ...new function.
17420 (pass_cd_dce): Convert from a global struct to a subclass of
17421 gimple_opt_pass along with...
17422 (pass_data_cd_dce): ...new pass_data instance and...
17423 (make_pass_cd_dce): ...new function.
17424 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
17425 subclass of gimple_opt_pass along with...
17426 (pass_data_dominator): ...new pass_data instance and...
17427 (make_pass_dominator): ...new function.
17428 (pass_phi_only_cprop): Convert from a global struct to a subclass of
17429 gimple_opt_pass along with...
17430 (pass_data_phi_only_cprop): ...new pass_data instance and...
17431 (make_pass_phi_only_cprop): ...new function.
17432 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
17433 subclass of gimple_opt_pass along with...
17434 (pass_data_dse): ...new pass_data instance and...
17435 (make_pass_dse): ...new function.
17436 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
17437 a subclass of gimple_opt_pass along with...
17438 (pass_data_forwprop): ...new pass_data instance and...
17439 (make_pass_forwprop): ...new function.
17440 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
17441 struct to a subclass of gimple_opt_pass along with...
17442 (pass_data_tree_ifcombine): ...new pass_data instance and...
17443 (make_pass_tree_ifcombine): ...new function.
17444 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
17445 subclass of gimple_opt_pass along with...
17446 (pass_data_ch): ...new pass_data instance and...
17447 (make_pass_ch): ...new function.
17448 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
17449 subclass of gimple_opt_pass along with...
17450 (pass_data_tree_loop): ...new pass_data instance and...
17451 (make_pass_tree_loop): ...new function.
17452 (pass_tree_loop_init): Convert from a global struct to a subclass of
17453 gimple_opt_pass along with...
17454 (pass_data_tree_loop_init): ...new pass_data instance and...
17455 (make_pass_tree_loop_init): ...new function.
17456 (pass_lim): Convert from a global struct to a subclass of
17457 gimple_opt_pass along with...
17458 (pass_data_lim): ...new pass_data instance and...
17459 (make_pass_lim): ...new function.
17460 (pass_tree_unswitch): Convert from a global struct to a subclass of
17461 gimple_opt_pass along with...
17462 (pass_data_tree_unswitch): ...new pass_data instance and...
17463 (make_pass_tree_unswitch): ...new function.
17464 (pass_predcom): Convert from a global struct to a subclass of
17465 gimple_opt_pass along with...
17466 (pass_data_predcom): ...new pass_data instance and...
17467 (make_pass_predcom): ...new function.
17468 (pass_vectorize): Convert from a global struct to a subclass of
17469 gimple_opt_pass along with...
17470 (pass_data_vectorize): ...new pass_data instance and...
17471 (make_pass_vectorize): ...new function.
17472 (pass_graphite): Convert from a global struct to a subclass of
17473 gimple_opt_pass along with...
17474 (pass_data_graphite): ...new pass_data instance and...
17475 (make_pass_graphite): ...new function.
17476 (pass_graphite_transforms): Convert from a global struct to a subclass
17477 of gimple_opt_pass along with...
17478 (pass_data_graphite_transforms): ...new pass_data instance and...
17479 (make_pass_graphite_transforms): ...new function.
17480 (pass_check_data_deps): Convert from a global struct to a subclass of
17481 gimple_opt_pass along with...
17482 (pass_data_check_data_deps): ...new pass_data instance and...
17483 (make_pass_check_data_deps): ...new function.
17484 (pass_iv_canon): Convert from a global struct to a subclass of
17485 gimple_opt_pass along with...
17486 (pass_data_iv_canon): ...new pass_data instance and...
17487 (make_pass_iv_canon): ...new function.
17488 (pass_scev_cprop): Convert from a global struct to a subclass of
17489 gimple_opt_pass along with...
17490 (pass_data_scev_cprop): ...new pass_data instance and...
17491 (make_pass_scev_cprop): ...new function.
17492 (pass_record_bounds): Convert from a global struct to a subclass of
17493 gimple_opt_pass along with...
17494 (pass_data_record_bounds): ...new pass_data instance and...
17495 (make_pass_record_bounds): ...new function.
17496 (pass_complete_unroll): Convert from a global struct to a subclass of
17497 gimple_opt_pass along with...
17498 (pass_data_complete_unroll): ...new pass_data instance and...
17499 (make_pass_complete_unroll): ...new function.
17500 (pass_complete_unrolli): Convert from a global struct to a subclass of
17501 gimple_opt_pass along with...
17502 (pass_data_complete_unrolli): ...new pass_data instance and...
17503 (make_pass_complete_unrolli): ...new function.
17504 (pass_parallelize_loops): Convert from a global struct to a subclass
17505 of gimple_opt_pass along with...
17506 (pass_data_parallelize_loops): ...new pass_data instance and...
17507 (make_pass_parallelize_loops): ...new function.
17508 (pass_loop_prefetch): Convert from a global struct to a subclass of
17509 gimple_opt_pass along with...
17510 (pass_data_loop_prefetch): ...new pass_data instance and...
17511 (make_pass_loop_prefetch): ...new function.
17512 (pass_iv_optimize): Convert from a global struct to a subclass of
17513 gimple_opt_pass along with...
17514 (pass_data_iv_optimize): ...new pass_data instance and...
17515 (make_pass_iv_optimize): ...new function.
17516 (pass_tree_loop_done): Convert from a global struct to a subclass of
17517 gimple_opt_pass along with...
17518 (pass_data_tree_loop_done): ...new pass_data instance and...
17519 (make_pass_tree_loop_done): ...new function.
17520 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
17521 struct to a subclass of gimple_opt_pass along with...
17522 (pass_data_cse_reciprocals): ...new pass_data instance and...
17523 (make_pass_cse_reciprocals): ...new function.
17524 (pass_cse_sincos): Convert from a global struct to a subclass of
17525 gimple_opt_pass along with...
17526 (pass_data_cse_sincos): ...new pass_data instance and...
17527 (make_pass_cse_sincos): ...new function.
17528 (pass_optimize_bswap): Convert from a global struct to a subclass of
17529 gimple_opt_pass along with...
17530 (pass_data_optimize_bswap): ...new pass_data instance and...
17531 (make_pass_optimize_bswap): ...new function.
17532 (pass_optimize_widening_mul): Convert from a global struct to a
17533 subclass of gimple_opt_pass along with...
17534 (pass_data_optimize_widening_mul): ...new pass_data instance and...
17535 (make_pass_optimize_widening_mul): ...new function.
17536 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
17537 subclass of gimple_opt_pass along with...
17538 (pass_data_phiopt): ...new pass_data instance and...
17539 (make_pass_phiopt): ...new function.
17540 (pass_cselim): Convert from a global struct to a subclass of
17541 gimple_opt_pass along with...
17542 (pass_data_cselim): ...new pass_data instance and...
17543 (make_pass_cselim): ...new function.
17544 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
17545 subclass of gimple_opt_pass along with...
17546 (pass_data_phiprop): ...new pass_data instance and...
17547 (make_pass_phiprop): ...new function.
17548 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
17549 subclass of gimple_opt_pass along with...
17550 (pass_data_pre): ...new pass_data instance and...
17551 (make_pass_pre): ...new function.
17552 (pass_fre): Convert from a global struct to a subclass of
17553 gimple_opt_pass along with...
17554 (pass_data_fre): ...new pass_data instance and...
17555 (make_pass_fre): ...new function.
17556 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
17557 subclass of gimple_opt_pass along with...
17558 (pass_data_reassoc): ...new pass_data instance and...
17559 (make_pass_reassoc): ...new function.
17560 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
17561 subclass of gimple_opt_pass along with...
17562 (pass_data_sink_code): ...new pass_data instance and...
17563 (make_pass_sink_code): ...new function.
17564 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
17565 subclass of gimple_opt_pass along with...
17566 (pass_data_strlen): ...new pass_data instance and...
17567 (make_pass_strlen): ...new function.
17568 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
17569 struct to a subclass of gimple_opt_pass along with...
17570 (pass_data_build_alias): ...new pass_data instance and...
17571 (make_pass_build_alias): ...new function.
17572 (pass_build_ealias): Convert from a global struct to a subclass of
17573 gimple_opt_pass along with...
17574 (pass_data_build_ealias): ...new pass_data instance and...
17575 (make_pass_build_ealias): ...new function.
17576 (pass_ipa_pta): Convert from a global struct to a subclass of
17577 simple_ipa_opt_pass along with...
17578 (pass_data_ipa_pta): ...new pass_data instance and...
17579 (make_pass_ipa_pta): ...new function.
17580 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
17581 subclass of gimple_opt_pass along with...
17582 (pass_data_uncprop): ...new pass_data instance and...
17583 (make_pass_uncprop): ...new function.
17584 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
17585 global struct to a subclass of gimple_opt_pass along with...
17586 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
17587 (make_pass_late_warn_uninitialized): ...new function.
17588 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
17589 to a subclass of gimple_opt_pass along with...
17590 (pass_data_init_datastructures): ...new pass_data instance and...
17591 (make_pass_init_datastructures): ...new function.
17592 (pass_early_warn_uninitialized): Convert from a global struct to a
17593 subclass of gimple_opt_pass along with...
17594 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
17595 (make_pass_early_warn_uninitialized): ...new function.
17596 (pass_update_address_taken): Convert from a global struct to a
17597 subclass of gimple_opt_pass along with...
17598 (pass_data_update_address_taken): ...new pass_data instance and...
17599 (make_pass_update_address_taken): ...new function.
17600 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
17601 struct to a subclass of gimple_opt_pass along with...
17602 (pass_data_release_ssa_names): ...new pass_data instance and...
17603 (make_pass_release_ssa_names): ...new function.
17604 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
17605 subclass of gimple_opt_pass along with...
17606 (pass_data_stdarg): ...new pass_data instance and...
17607 (make_pass_stdarg): ...new function.
17608 * tree-switch-conversion.c (pass_convert_switch): Convert from a
17609 global struct to a subclass of gimple_opt_pass along with...
17610 (pass_data_convert_switch): ...new pass_data instance and...
17611 (make_pass_convert_switch): ...new function.
17612 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
17613 to a subclass of gimple_opt_pass along with...
17614 (pass_data_tail_recursion): ...new pass_data instance and...
17615 (make_pass_tail_recursion): ...new function.
17616 (pass_tail_calls): Convert from a global struct to a subclass of
17617 gimple_opt_pass along with...
17618 (pass_data_tail_calls): ...new pass_data instance and...
17619 (make_pass_tail_calls): ...new function.
17620 * tree-vect-generic.c (pass_lower_vector): Convert from a global
17621 struct to a subclass of gimple_opt_pass along with...
17622 (pass_data_lower_vector): ...new pass_data instance and...
17623 (make_pass_lower_vector): ...new function.
17624 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
17625 gimple_opt_pass along with...
17626 (pass_data_lower_vector_ssa): ...new pass_data instance and...
17627 (make_pass_lower_vector_ssa): ...new function.
17628 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
17629 to a subclass of gimple_opt_pass along with...
17630 (pass_data_slp_vectorize): ...new pass_data instance and...
17631 (make_pass_slp_vectorize): ...new function.
17632 (pass_ipa_increase_alignment): Convert from a global struct to a
17633 subclass of simple_ipa_opt_pass along with...
17634 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
17635 (make_pass_ipa_increase_alignment): ...new function.
17636 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
17637 gimple_opt_pass along with...
17638 (pass_data_vrp): ...new pass_data instance and...
17639 (make_pass_vrp): ...new function.
17640 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
17641 subclass of simple_ipa_opt_pass along with...
17642 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
17643 (make_pass_ipa_free_lang_data): ...new function.
17644 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
17645 gimple_opt_pass along with...
17646 (pass_data_tsan): ...new pass_data instance and...
17647 (make_pass_tsan): ...new function.
17648 (pass_tsan_O0): Convert from a global struct to a subclass of
17649 gimple_opt_pass along with...
17650 (pass_data_tsan_O0): ...new pass_data instance and...
17651 (make_pass_tsan_O0): ...new function.
17652 * var-tracking.c (pass_variable_tracking): Convert from a global
17653 struct to a subclass of rtl_opt_pass along with...
17654 (pass_data_variable_tracking): ...new pass_data instance and...
17655 (make_pass_variable_tracking): ...new function.
17656 * web.c (pass_web): Convert from a global struct to a subclass of
17657 rtl_opt_pass along with...
17658 (pass_data_web): ...new pass_data instance and...
17659 (make_pass_web): ...new function.
17660 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
17661 declaration with that of...
17662 (make_pass_mode_switch_use): ...new function.
17663 (pass_resolve_sw_modes): Replace declaration with that of...
17664 (make_pass_resolve_sw_modes): ...new function.
17665 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
17666 from a global struct to a subclass of rtl_opt_pass along with...
17667 (pass_data_mode_switch_use): ...new pass_data instance and...
17668 (make_pass_mode_switch_use): ...new function.
17669 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
17670 from a global struct to a subclass of rtl_opt_pass along with...
17671 (pass_data_resolve_sw_modes): ...new pass_data instance and...
17672 (make_pass_resolve_sw_modes): ...new function.
17673 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
17674 struct to a subclass of rtl_opt_pass along with...
17675 (pass_data_insert_vzeroupper): ...new pass_data instance and...
17676 (make_pass_insert_vzeroupper): ...new function.
17677 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
17678 global struct to a subclass of rtl_opt_pass along with...
17679 (pass_data_work_around_errata): ...new pass_data instance and...
17680 (make_pass_work_around_errata): ...new function.
17681 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
17682 struct to a subclass of rtl_opt_pass along with...
17683 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
17684 (make_pass_mips_machine_reorg2): ...new function.
17686 2013-08-05 David Malcolm <dmalcolm@redhat.com>
17688 * passes.c (pass_manager::operator new): New.
17690 2013-08-05 David Malcolm <dmalcolm@redhat.com>
17692 Handwritten part of conversion of passes to C++ classes.
17694 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
17695 (toplev.o): Add dep on PASS_MANAGER_H.
17696 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
17697 of early local pases to reflect this moving from a global to a
17698 member of gcc::pass_manager.
17699 (cgraph_add_new_function): Likewise.
17700 * lto-cgraph.c (lto_output_node): Update for conversion of
17701 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
17702 * passes.c (opt_pass::clone): New.
17703 (opt_pass::gate): New.
17704 (opt_pass::execute): New.
17705 (opt_pass::opt_pass): New.
17706 (pass_manager::execute_early_local_passes): New.
17707 (pass_manager::execute_pass_mode_switching): new.
17708 (finish_optimization_passes): Convert to...
17709 (pass_manager::finish_optimization_passes): ...this.
17710 (finish_optimization_passes): Update for conversion of passes to
17712 (register_dump_files_1): Use has_gate since we cannot portably
17713 check a vtable entry against NULL.
17714 (dump_one_pass): Likewise.
17715 (ipa_write_summaries_2): Likewise.
17716 (ipa_write_optimization_summaries_1): Likewise.
17717 (ipa_read_summaries_1): Likewise.
17718 (ipa_read_optimization_summaries_1): Likewise.
17719 (execute_ipa_stmt_fixups): Likewise.
17720 (pass_manager::pass_manager): Rewrite pass-creation, invoking
17721 pass-creation functions rather than wiring up globals, and
17722 storing the results in fields of pass_manager generated using
17723 pass-instances.def.
17724 (pass_manager::dump_profile_report): Update for conversion of
17725 passes to C++ classes.
17726 (pass_manager::execute_ipa_summary_passes): Likewise.
17727 (execute_one_ipa_transform_pass): Likewise.
17728 (execute_one_pass): Use has_gate and has_execute since we cannot
17729 portably check a vtable entry against NULL.
17730 * pass_manager.h (pass_manager::finish_optimization_passes): New.
17731 (pass_manager): Use pass-instances.def to add fields for the
17732 various pass instances.
17733 * toplev.c (finalize): Update for move of
17734 finish_optimization_passes to a method of gcc::pass_manager.
17735 * toplev.h (finish_optimization_passes): Move to method of class
17737 * tree-pass.h (struct pass_data): New.
17738 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
17739 (opt_pass::gate): Convert to virtual function.
17740 (opt_pass::~opt_pass): New.
17741 (opt_pass::clone): New.
17742 (opt_pass::execute): Convert to virtual function.
17743 (opt_pass::opt_pass): New.
17744 (opt_pass::ctxt_): new.
17745 (gimple_opt_pass): Convert to subclass of opt_pass.
17746 (gimple_opt_pass::gimple_opt_pass): New.
17747 (rtl_opt_pass): Convert to subclass of opt_pass.
17748 (rtl_opt_pass::rtl_opt_pass): New.
17749 (ipa_opt_pass_d): Convert to subclass of opt_pass.
17750 (ipa_opt_pass_d::ipa_opt_pass_d): New.
17751 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
17752 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
17753 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
17754 invocation of pass_mode_switching to reflect this moving from a
17755 global to a member of gcc::pass_manager.
17756 (ix86_option_override): Rework how pass_insert_vzeroupper is
17757 added to the pass_manager to reflect autogenerated changes.
17758 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
17760 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
17762 PR rtl-optimization/57708
17763 * recog.c (peep2_find_free_register): Validate all regs in a
17766 2013-08-05 Jan Hubicka <jh@suse.cz>
17769 * cgraph.c (verify_cgraph_node): Accept local flags from other
17771 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
17772 (function_and_variable_visibility): Likewise.
17773 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
17775 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
17777 * graph.c (init_graph_slim_pretty_print): Remove.
17778 (print_graph_cfg): Do not call it. Use local pretty printer.
17779 (start_graph_dump): Likewise.
17781 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
17783 * gimple-pretty-print.c (buffer): Remove.
17784 (initialized): Likewise.
17785 (maybe_init_pretty_print): Likewise.
17786 (print_gimple_stmt): Do not call it. Use non-static local
17787 pretty_printer variable.
17788 (print_gimple_expr): Likewise.
17789 (print_gimple_seq): Likewise.
17790 (gimple_dump_bb): Likewise.
17792 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
17794 * asan.c (asan_pp): Remove.
17795 (asan_pp_initialized): Likewise.
17796 (asan_pp_initialize): Likewise.
17797 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
17798 (asan_emit_stack_protection): Tidy. Use local pretty printer.
17799 (asan_add_global): Likewise.
17801 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
17803 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
17804 * diagnostic.h (diagnostic_flush_buffer): Adjust.
17805 * pretty-print.c (pp_formatted_text_data): Likewise.
17806 (pp_indent): Rename from pp_base_indent.
17807 (pp_format): Rename from pp_base_format.
17808 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
17809 (pp_format_verbatim): Rename from pp_base_format_verbatim.
17810 (pp_flush): Rename from pp_base_flush.
17811 (pp_set_line_maximum_length): Rename from
17812 pp_base_set_line_maximum_length.
17813 (pp_clear_output_area): Rename from pp_base_clear_output_area.
17814 (pp_set_prefix): Rename from pp_base_set_prefix.
17815 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
17816 (pp_emit_prefix): Rename from pp_base_emit_prefix.
17817 (pp_append_text): Rename from pp_base_append_text.
17818 (pp_formatted_text): Rename from pp_base_formatted_text.
17819 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
17820 (pp_remaining_character_count_for_line): Rename from
17821 pp_base_remaining_character_count_for_line.
17822 (pp_newline): Rename from pp_base_newline.
17823 (pp_character): Rename from pp_base_character.
17824 (pp_string): Rename from pp_base_string.
17825 (pp_maybe_space): Rename from pp_base_maybe_space.
17826 * asan.c (asan_pp_string): Adjust.
17827 (asan_emit_stack_protection): Likewise.
17828 (asan_add_global): Likewise.
17829 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
17830 * tree-mudflap.c (mf_varname_tree): Likewise.
17831 * tree-pretty-print.c (pp_tree_identifier): Rename from
17832 pp_base_tree_identifier.
17833 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
17834 Declare as function.
17836 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
17838 * pretty-print.h (pp_bar_bar): New.
17839 (pp_ampersand_ampersand): Likewise.
17840 (pp_less_equal): Likewise.
17841 (pp_greater_equal): Likewise.
17842 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
17843 printer functions instead of pp_string or operators and punctuators.
17844 (dump_gimple_call): Likewise.
17845 (dump_gimple_omp_for): Likewise.
17846 (dump_gimple_transaction): Likewise.
17847 (dump_gimple_phi): Likewise.
17848 (pp_gimple_stmt_1): Likewise.
17849 * sched-vis.c (print_insn): Likewise.
17850 * tree-mudflap.c (mf_varname_tree): Likewise.
17851 * tree-pretty-print.c (dump_block_node): Likewise.
17852 (dump_generic_node): Likewise.
17854 2013-08-02 Jan Hubicka <jh@suse.cz>
17856 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
17858 * lto-streamer-out.c (tree_is_indexable): Results decls and
17859 parm decls are not indexable.
17860 (DFS_write_tree_body): Do not follow args and results.
17861 (hash_tree): Likewise.
17862 (output_functions): Rearrange so struct function is needed
17863 only when real body is output; be able to also ouptut abstract
17864 functions; output DECL_ARGUMENTS and DECL_RESULT.
17865 (lto_output): When not in WPA, ale store abstract functions.
17866 (write_symbol): Do not care about RESULT_DECL.
17867 (output_symbol_p): Handle correctly sbtract decls.
17868 * lto-streamer-in.c (input_function): Rearrange so struct
17869 function can be NULL at entry; allow streaming of
17870 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
17871 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
17872 sanity check during LTO.
17873 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
17874 RESULT_DECl and DECL_ARGUMENTS.
17875 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
17878 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
17880 * pretty-print.h (pp_underscore): New.
17882 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
17883 printer functions instead of pp_character.
17884 (dump_binary_rhs): Likewise.
17885 (dump_ternary_rhs): Likewise.
17886 (dump_gimple_call_args): Likewise.
17887 (pp_points_to_solution): Likewise.
17888 (dump_gimple_call): Likewise.
17889 (dump_gimple_switch): Likewise.
17890 (dump_gimple_cond): Likewise.
17891 (dump_gimple_bind): Likewise.
17892 (dump_gimple_try): Likewise.
17893 (dump_gimple_omp_for): Likewise.
17894 (dump_gimple_omp_continue): Likewise.
17895 (dump_gimple_omp_single): Likewise.
17896 (dump_gimple_omp_sections): Likewise.
17897 (dump_gimple_omp_block): Likewise.
17898 (dump_gimple_omp_critical): Likewise.
17899 (dump_gimple_transaction): Likewise.
17900 (dump_gimple_asm): Likewise.
17901 (dump_gimple_phi): Likewise.
17902 (dump_gimple_omp_parallel): Likewise.
17903 (dump_gimple_omp_task): Likewise.
17904 (dump_gimple_omp_atomic_load): Likewise.
17905 (dump_gimple_omp_atomic_store): Likewise.
17906 (dump_gimple_mem_ops): Likewise.
17907 (pp_gimple_stmt_1): Likewise.
17908 (pp_cfg_jump): Likewise.
17909 (dump_implicit_edges): Likewise.
17910 (gimple_dump_bb_for_graph): Likewise.
17911 * graph.c (draw_cfg_node): Likewise.
17912 * langhooks.c (lhd_print_error_function): Likewise.
17913 * sched-vis.c (print_exp): Likewise.
17914 (print_value): Likewise.
17915 (print_pattern): Likewise.
17916 (print_insn): Likewise.
17917 (rtl_dump_bb_for_graph): Likewise.
17918 * tree-pretty-print.c (dump_function_declaration): Likewise.
17919 (dump_array_domain): Likewise.
17920 (dump_omp_clause): Likewise.
17921 (dump_location): Likewise.
17922 (dump_generic_node): Likewise.
17923 (print_struct_decl): Likewise.
17924 * diagnostic.c (diagnostic_show_locus): Use pp_space.
17926 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
17928 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
17929 candidate table when replacing a candidate statement.
17930 (replace_rhs_if_not_dup): Likewise.
17931 (replace_one_candidate): Likewise.
17933 2013-08-02 Jan Hubicka <jh@suse.cz>
17934 Martin Liska <marxin.liska@gmail.com>
17936 * cgraphunit.c (add_new_function): Fix logic when adding from
17938 (assemble_thunk): Rename to ...
17939 (expand_thunk); .. this one; export; get it working with
17940 general functions; make produced gimple valid.
17941 * cgraph.h (expand_thunk): Declare.
17943 2013-08-02 Jan Hubicka <jh@suse.cz>
17945 * ipa-cp.c (gather_context_independent_values): Use
17946 ipa_get_param_move_cost.
17947 (get_replacement_map): Remove PARAM; move parameter folding
17949 (create_specialized_node): Update.
17950 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
17951 assert that we have gimple body; update move_cost.
17952 (count_formal_params): Assert that we have gimple body.
17953 (ipa_dump_param): New function.
17954 (ipa_alloc_node_params): Break out from ...
17955 (ipa_initialize_node_params): ... here.
17956 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
17957 (ipa_write_node_info): Stream move costs.
17958 (ipa_read_node_info): Read move costs.
17959 (ipa_update_after_lto_read): Do not recompute node params.
17960 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
17961 (ipa_get_param): Check we are not in WPA.
17962 (ipa_get_param_move_cost): New.
17963 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
17964 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
17965 parm numbers to be present.
17967 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
17969 PR rtl-optimization/58048
17970 * lra-constraints.c (process_alt_operands): Don't check asm
17971 operand on register.
17973 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
17975 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
17976 the implied StoreLoad barrier for atomic operations if before.
17978 2013-08-02 Jan Hubicka <jh@suse.cz>
17979 Martin Liska <marxin.liska@gmail.com>
17981 * cgraph.c (cgraph_function_body_availability): Do not check
17983 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
17984 symtab_node_availability): Declare.
17985 * ipa.c (can_replace_by_local_alias): New.
17986 (function_and_variable_visibility): Use it.
17987 * symtab.c (symtab_for_node_and_aliases,
17988 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
17990 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
17992 PR rtl-optimization/57963
17993 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
17994 (lra_constraints): Use them.
17996 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
17998 * config/arm/types.md (define_attr "type"): Add "load_acq"
18000 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
18002 (cortex_a53_store1): Likewise.
18004 2013-08-01 Jan Hubicka <jh@suse.cz>
18006 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
18007 partitions are not needed.
18009 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
18011 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
18012 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
18013 MAYBE_NON_Q_CLASS_P where appropriate.
18015 2013-08-01 Jan Hubicka <jh@suse.cz>
18017 * cgraph.h (release_function_body): Declare.
18018 * tree.c (free_lang_data_in_decl): Free, parameters and return values
18019 of unused delcarations.
18021 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18023 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
18024 RTL form when subtracting a constant.
18026 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18028 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
18029 Generate canonical plus rtx with negated immediate instead of minus
18031 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
18033 2013-08-01 Jan Hubicka <jh@suse.cz>
18035 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
18036 (cgraph_release_function_body): Likewise.
18037 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
18038 * cgraph.h (cgrpah_node): Rename abstract_and_needed
18039 to used_as_abstract_origin.
18040 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
18041 symbols used as abstract origins.
18042 * cgraphunit.c (analyze_functions): Update.
18043 * ipa.c (symtab_remove_unreachable_nodes): Recompute
18044 used_as_abstract_origin.
18045 * tree-inline.c (tree_function_versioning): Update
18046 used_as_abstract_origin; be ready for DECL_RESULT and
18047 DECL_ARGUMENTS to be NULL.
18049 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
18050 for abstract functions.
18051 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
18054 2013-08-01 Jan Hubicka <jh@suse.cz>
18056 * profile.c (compute_value_histograms): Fix thinko.
18058 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
18060 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
18061 aarch-common-protos.h to extra_headers.
18062 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
18063 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
18064 * config/aarch64/t-aarch64 (aarch-common.o): Define.
18066 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
18068 * config/aarch64/aarch64.md (define_attr "type"): Delete.
18069 Include "../arm/types.md". Define "type" attribute for all patterns.
18070 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
18073 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
18075 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
18076 to support power8 load fusion.
18077 (fusion_gpr_mem_load): Likewise.
18079 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
18081 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
18082 declarations for power8 load fusion.
18083 (emit_fusion_gpr_load): Likewise.
18085 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
18086 tuning for power8, turn on fusion mode by default. Turn on sign
18087 extending fusion mode if normal fusion mode is on, and we are at
18089 (fusion_gpr_load_p): New function, return true if we can fuse an
18090 addis instruction with a dependent load to a GPR.
18091 (emit_fusion_gpr_load): Emit the instructions for power8 load
18094 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
18095 (VSX load fusion peepholes): New peepholes to fuse together an
18096 addi instruction with a VSX load instruction.
18098 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
18099 peepholes to fuse an addis instruction with a load to a GPR base
18100 register. If we are supporting sign extending fusions, convert
18101 sign extending loads to zero extending loads and add an explicit
18104 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
18106 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
18107 aarch-common-protos.h to extra_headers.
18108 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
18109 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
18110 (arm_early_store_addr_dep): Likewise.
18111 (arm_no_early_alu_shift_dep): Likewise.
18112 (arm_no_early_alu_shift_value_dep): Likewise.
18113 (arm_no_early_mul_dep): Likewise.
18114 (arm_no_early_store_addr_dep): Likewise.
18115 (arm_mac_accumulator_is_mul_result): Likewise.
18116 (arm_mac_accumulator_is_result): Likewise.
18117 * config/arm/aarch-common.c: ... here. New file.
18118 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
18120 (arm_early_store_addr_dep): Likewise.
18121 (arm_no_early_alu_shift_dep): Likewise.
18122 (arm_no_early_alu_shift_value_dep): Likewise.
18123 (arm_no_early_mul_dep): Likewise.
18124 (arm_no_early_store_addr_dep): Likewise.
18125 (arm_mac_accumulator_is_mul_result): Likewise.
18126 (arm_mac_accumulator_is_result): Likewise.
18127 * config/arm/aarch-common-protos.h: ... here. New file.
18128 * config/arm/t-arm (aarch-common.o): Define.
18130 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
18132 * config/arm/arm.md: Include new file "types.md".
18133 (define_attr "type"): Move from here to ...
18134 (define_attr "mul32"): Likewise.
18135 (define_attr "mul64"): Likewise.
18136 * config/arm/types.md: ... here. New file.
18138 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
18140 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
18141 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
18143 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
18145 * gen-pass-instances.awk: Fix offset of substr().
18147 2013-07-31 David Malcolm <dmalcolm@redhat.com>
18149 * Makefile.in (pass-instances.def): New.
18150 (passes.o): Replace dependency on passes.def with one on
18153 * gen-pass-instances.awk: New.
18155 * passes.c (pass_manager::pass_manager): Use pass-instances.def
18156 rather than passes.def, updating local definition of NEXT_PASS
18157 macro to add an extra NUM parameter (currently unused).
18159 2013-07-30 David Malcolm <dmalcolm@redhat.com>
18161 * Makefile.in (PASS_MANAGER_H): New.
18162 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
18163 (passes.o): Likewise.
18164 (statistics.o): Likewise.
18165 (cgraphunit.o): Likewise.
18166 (context.o): Depend on PASS_MANAGER_H.
18168 * pass_manager.h: New.
18170 * cgraphunit.c (cgraph_add_new_function): Update for moves
18171 of globals to fields of pass_manager.
18172 (analyze_function): Likewise.
18173 (expand_function): Likewise.
18174 (ipa_passes): Likewise.
18175 (compile): Likewise.
18177 * context.c (context::context): New.
18178 * context.h (context::context): New.
18179 (context::get_passes): New.
18180 (context::passes_): New.
18182 * lto-cgraph.c (input_node): Update for moves of globals to
18183 fields of pass_manager.
18185 * passes.c (all_passes): Remove, in favor of a field of the
18186 same name within the new class pass_manager.
18187 (all_small_ipa_passes): Likewise.
18188 (all_lowering_passes): Likewise.
18189 (all_regular_ipa_passes): Likewise.
18190 (all_late_ipa_passes): Likewise.
18191 (all_lto_gen_passes): Likewise.
18192 (passes_by_id): Likewise.
18193 (passes_by_id_size): Likewise.
18194 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
18195 the new class pass_manager.
18196 (set_pass_for_id): Convert to...
18197 (pass_manager::set_pass_for_id): ...method.
18198 (get_pass_for_id): Convert to...
18199 (pass_manager::get_pass_for_id): ...method.
18200 (register_one_dump_file): Move body of implementation into...
18201 (pass_manager::register_one_dump_file): ...here.
18202 (register_dump_files_1): Convert to...
18203 (pass_manager::register_dump_files_1): ...method.
18204 (register_dump_files): Convert to...
18205 (pass_manager::register_dump_files): ...method.
18206 (create_pass_tab): Update for moves of globals to fields of
18208 (dump_passes): Move body of implementation into...
18209 (pass_manager::dump_passes): ...here.
18210 (register_pass): Move body of implementation into...
18211 (pass_manager::register_pass): ...here.
18212 (init_optimization_passes): Convert into...
18213 (pass_manager::pass_manager): ...constructor for new
18214 pass_manager class, and initialize the pass_lists array.
18215 (check_profile_consistency): Update for moves of globals to
18216 fields of pass_manager.
18217 (dump_profile_report): Move body of implementation into...
18218 (pass_manager::dump_profile_report): ...here.
18219 (ipa_write_summaries_1): Update for moves of pass lists from
18220 being globals to fields of pass_manager.
18221 (ipa_write_optimization_summaries): Likewise.
18222 (ipa_read_summaries): Likewise.
18223 (ipa_read_optimization_summaries): Likewise.
18224 (execute_all_ipa_stmt_fixups): Likewise.
18226 * statistics.c (statistics_fini): Update for moves of globals to
18227 fields of pass_manager.
18229 * toplev.c (general_init): Replace call to
18230 init_optimization_passes with construction of the pass_manager
18233 * tree-pass.h (all_passes): Remove, in favor of a field of the
18234 same name within the new class pass_manager.
18235 (all_small_ipa_passes): Likewise.
18236 (all_lowering_passes): Likewise.
18237 (all_regular_ipa_passes): Likewise.
18238 (all_lto_gen_passes): Likewise.
18239 (all_late_ipa_passes): Likewise.
18240 (passes_by_id): Likewise.
18241 (passes_by_id_size): Likewise.
18242 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
18243 the new class pass_manager.
18244 (get_pass_for_id): Remove.
18246 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
18248 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
18251 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
18253 * arm.md (mulhi3): New expand pattern.
18255 2013-07-30 Jan Hubicka <jh@suse.cz>
18256 Martin Liska <marxin.liska@gmail.com>
18258 * profile.c (compute_value_histograms): Do not ICE when
18259 there is mismatch only on some counters.
18261 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18263 PR rtl-optimization/57637
18264 * function.c (move_insn_for_shrink_wrap): Also check the
18265 GEN set of the LIVE problem for the liveness analysis
18266 if it exists, otherwise give up.
18268 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
18270 PR tree-optimization/57993
18271 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
18272 replaced statement in the candidate table.
18273 (phi_add_costs): Return infinite cost when the hidden basis does
18274 not dominate all phis on which the candidate is dependent.
18275 (replace_one_candidate): Record replaced statement in the
18278 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
18280 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
18281 (ashlv2si3): New expander.
18282 (*ashlv2si3_i): New define_insn_and_split.
18283 * predicates.md (float_operation): Allow patterns with three
18284 basic sub-patterns.
18286 PR rtl-optimization/58021
18287 * mode-switching.c (create_pre_exit): Always split off preceding
18288 insns if we are not at the basic block head.
18290 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
18292 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
18293 (UCLIBC_DYNAMIC_LINKER): New macro.
18294 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
18296 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
18297 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
18299 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
18300 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
18301 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
18302 for SF and DF modes. Use ieee_quad_format for TF mode.
18303 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
18304 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
18305 (mips_option_override): Handle `-mnan=legacy'.
18306 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
18307 `-mabs=2008' and `-mnan=2008'.
18308 (OPTION_DEFAULT_SPECS): Add "nan" default.
18309 (ASM_SPEC): Handle `-mnan='.
18310 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
18311 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
18312 comment accordingly.
18313 (neg<mode>2): Likewise.
18314 * config/mips/mips.opt (mabs, mnan): New options.
18315 * doc/install.texi (Configuration): Document `--with-nan=' option.
18316 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
18318 (MIPS Options): Document them.
18319 * config.gcc <mips*-*-*>: Handle `--with-nan='.
18320 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
18321 * configure: Regenerate.
18322 * config.in: Regenerate.
18324 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
18326 * config/i386/i386.md (float post-reload splitters): Do not check
18327 for subregs of SSE registers.
18329 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
18330 H.J. Lu <hongjiu.lu@intel.com>
18334 * config/i386/i386.md (post-reload splitter
18335 to avoid partial SSE reg dependency stalls): New pattern.
18337 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
18339 * config/s390/s390.md ("movcc"): Swap load and store instructions.
18341 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
18343 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
18344 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
18346 2013-07-26 Cary Coutant <ccoutant@google.com>
18348 * dwarf2out.c (die_checksum_ordered): Don't include template
18349 instantiations in signature.
18350 (is_template_parameter): New function.
18351 (is_template_instantiation): New function.
18352 (generate_skeleton_bottom_up): Don't include template instantiations
18354 (generate_skeleton): Likewise.
18355 (break_out_comdat_types): Move recursive call to break out nested
18357 (prune_unused_types_mark_generic_parms_dies): Call
18358 is_template_parameter.
18360 2013-07-26 Ian Bolton <ian.bolton@arm.com>
18362 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
18363 uses vector registers.
18364 * config/aarch64/iterators.md: Add attributes rtn and vas.
18366 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18367 Richard Earnshaw <richard.earnshaw@arm.com>
18369 * combine.c (simplify_comparison): Re-canonicalize operands
18371 * config/arm/arm.md (movcond_addsi): New splitter.
18373 2013-07-25 Sterling Augustine <saugustine@google.com>
18375 * dwarf2out.c (size_of_pubnames): Move code to...
18376 (include_pubname_in_output): ...here. New.
18377 (want_pubnames): Rearrange.
18378 (output_pubnames): Call include_pubname_in_output. Move assertion.
18380 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
18382 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
18384 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
18387 * doc/extend.texi: Remove obsolete builtins. Fix
18388 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
18390 2013-07-25 Jan Hubicka <jh@suse.cz>
18392 * cgraph.c (release_function_body): Break out from ...
18393 (cgraph_release_function_body): ... this one; also release DECL_RESULT
18394 and DECL_ARGUMENTS.
18395 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
18396 old_tree in the map.
18397 (create_specialized_node): Update.
18398 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
18400 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
18401 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
18402 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
18403 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
18406 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18408 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
18409 addsi3_carryin_alt2_<optab>): Correct output template.
18411 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18413 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
18414 Adjust for arm_restrict_it.
18415 Remove trailing whitespace.
18417 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
18419 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
18420 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
18422 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
18424 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
18426 PR rtl-optimization/57960
18427 * lra-constraints.c (process_alt_operands): Use the right mode
18428 when checking strict_low.
18430 2013-07-25 Jan Hubicka <jh@suse.cz>
18432 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
18433 * cgraph.c (cgraph_remove_node): Do not release function body
18434 when in cgraph streaming.
18435 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
18436 in other partitions are not considered reachable; fix handling of
18439 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18441 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
18443 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18448 * config/arm/arm.md ("*sibcall_insn): Replace use of
18449 Ss with US. Adjust output for v5 and v4t.
18450 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
18452 * config/arm/constraints.md ("Ss"): Rename to US.
18454 2013-07-25 Terry Guo <terry.guo@arm.com>
18456 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
18457 shift_add/shift_sub0/shift_sub1 RTXs.
18459 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
18460 Anton Blanchard <anton@au1.ibm.com>
18462 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
18463 (altivec_vpks<VI_char>ss): Likewise.
18464 (altivec_vpks<VI_char>us): Likewise.
18465 (altivec_vpku<VI_char>us): Likewise.
18466 (altivec_vpku<VI_char>um): Likewise.
18468 2013-07-24 David Malcolm <dmalcolm@redhat.com>
18470 Introduce context class.
18472 * Makefile.in (CONTEXT_H): New.
18473 (OBJS): Add context.o.
18474 (toplev.o): Add CONTEXT_H to dependencies.
18477 * toplev.c (general_init): Create the singleton gcc::context instance.
18483 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
18485 PR rtl-optimization/57968
18486 * mode-switching.c (create_pre_exit): Allow instructions that
18487 don't set a return register to need a non-exit mode.
18489 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
18490 Anton Blanchard <anton@au1.ibm.com>
18492 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
18493 operands to vperm for little endian.
18494 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
18495 of lvsl to create the control mask for a vperm for little endian.
18497 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18498 Anton Blanchard <anton@au1.ibm.com>
18500 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
18501 two operands for little-endian.
18503 2013-07-23 Steve Ellcey <sellcey@mips.com>
18505 * config/mips/mips.c (mips_case_values_threshold): New.
18506 (TARGET_CASE_VALUES_THRESHOLD): Define.
18508 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18509 Anton Blanchard <anton@au1.ibm.com>
18511 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
18512 selection of field for vector splat in little endian mode.
18514 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
18516 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
18517 expanders to rs6000.md.
18518 (ior<mode>3): Likewise.
18519 (and<mode>3): Likewise.
18520 (one_cmpl<mode>2): Likewise.
18521 (nor<mode>3): Likewise.
18522 (andc<mode>3): Likewise.
18523 (eqv<mode>3): Likewise.
18524 (nand<mode>3): Likewise.
18525 (orc<mode>3): Likewise.
18527 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
18530 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
18531 to split multi-word logical operations.
18532 (rs6000_split_logical_di): Likewise.
18533 (rs6000_split_logical): Likewise.
18535 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
18536 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
18537 and allow TImode operations in 32-bit.
18538 (vsx_and<mode>3_64bit): Likewise.
18539 (vsx_ior<mode>3_32bit): Likewise.
18540 (vsx_ior<mode>3_64bit): Likewise.
18541 (vsx_xor<mode>3_32bit): Likewise.
18542 (vsx_xor<mode>3_64bit): Likewise.
18543 (vsx_one_cmpl<mode>2_32bit): Likewise.
18544 (vsx_one_cmpl<mode>2_64bit): Likewise.
18545 (vsx_nor<mode>3_32bit): Likewise.
18546 (vsx_nor<mode>3_64bit): Likewise.
18547 (vsx_andc<mode>3_32bit): Likewise.
18548 (vsx_andc<mode>3_64bit): Likewise.
18549 (vsx_eqv<mode>3_32bit): Likewise.
18550 (vsx_eqv<mode>3_64bit): Likewise.
18551 (vsx_nand<mode>3_32bit): Likewise.
18552 (vsx_nand<mode>3_64bit): Likewise.
18553 (vsx_orc<mode>3_32bit): Likewise.
18554 (vsx_orc<mode>3_64bit): Likewise.
18556 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
18557 logical types in GPRs.
18559 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
18560 logical insns to rs6000.md, and allow TImode operations in
18562 (altivec_ior<mode>3): Likewise.
18563 (altivec_xor<mode>3): Likewise.
18564 (altivec_one_cmpl<mode>2): Likewise.
18565 (altivec_nor<mode>3): Likewise.
18566 (altivec_andc<mode>3): Likewise.
18568 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
18569 attributes for moving the 128-bit logical operations into
18571 (BOOL_REGS_OUTPUT): Likewise.
18572 (BOOL_REGS_OP1): Likewise.
18573 (BOOL_REGS_OP2): Likewise.
18574 (BOOL_REGS_UNARY): Likewise.
18575 (BOOL_REGS_AND_CR0): Likewise.
18576 (one_cmpl<mode>2): Add support for DI logical operations on
18577 32-bit, splitting the operations to 32-bit.
18578 (anddi3): Likewise.
18579 (iordi3): Likewise.
18580 (xordi3): Likewise.
18581 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
18582 changes to combine the 32/64-bit code, allow logical operations on
18583 TI mode in 32-bit, and to use similar match_operator patterns like
18584 scalar mode uses. Combine the Altivec and VSX code for logical
18585 operations, and move it here.
18586 (ior<mode>3, 128-bit types): Likewise.
18587 (xor<mode>3, 128-bit types): Likewise.
18588 (one_cmpl<mode>3, 128-bit types): Likewise.
18589 (nor<mode>3, 128-bit types): Likewise.
18590 (andc<mode>3, 128-bit types): Likewise.
18591 (eqv<mode>3, 128-bit types): Likewise.
18592 (nand<mode>3, 128-bit types): Likewise.
18593 (orc<mode>3, 128-bit types): Likewise.
18594 (and<mode>3_internal): Likewise.
18595 (bool<mode>3_internal): Likewise.
18596 (boolc<mode>3_internal1): Likewise.
18597 (boolc<mode>3_internal2): Likewise.
18598 (boolcc<mode>3_internal1): Likewise.
18599 (boolcc<mode>3_internal2): Likewise.
18600 (eqv<mode>3_internal1): Likewise.
18601 (eqv<mode>3_internal2): Likewise.
18602 (one_cmpl1<mode>3_internal): Likewise.
18604 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
18606 * config/microblaze/microblaze.c (microblaze_expand_prologue):
18607 Rename flag_stack_usage to flag_stack_usage_info.
18609 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
18611 * config/microblaze/sync.md: New file.
18612 * config/microblaze/microblaze.md: Include sync.md
18613 * config/microblaze/microblaze.c: Add print_operand 'y'.
18614 * config/microblaze/constraints.md: Add memory_contraint
18615 'Q' which is a single register.
18617 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
18619 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
18621 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
18623 * reload.c (find_reloads): Exit loop once we find this operand
18624 cannot be reloaded somehow for this alternative.
18626 * reload.c (find_reloads): Exit loop once we find a hard register.
18628 * rtlanal.c (computed_jump_p): Exit loop once we find label
18631 * i386.c (ix86_pad_returns): Exit loop after setting replace.
18633 * cfgloopmanip.c (remove_path): Exit loop after setting
18636 * gensupport.c (subst_dup): Avoid loop if code is not
18637 MATCH_DUP nor MATCH_OP_DUP.
18639 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
18641 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
18643 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
18645 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
18646 true for SP_REGNUM if mode == ptr_mode.
18647 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
18648 with value R0_REGNUM + 31.
18650 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
18652 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
18653 pad pointer-typed argument downward.
18655 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
18657 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
18658 and __ILP32__ when the ILP32 model is in use.
18660 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
18662 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
18663 (aarch64_load_symref_appropriately): In the case of
18664 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
18665 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
18666 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
18667 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
18668 if 'mode' doesn't equal to 'ptr_mode'.
18669 (aarch64_output_mi_thunk): Add an assertion on the alignment of
18670 'vcall_offset'; change to call aarch64_emit_move differently depending
18671 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
18672 to calculate the upper bound of 'vcall_offset'.
18673 (aarch64_cannot_force_const_mem): Change to also return true if
18675 (aarch64_legitimize_reload_address): In the case of large
18676 displacements, add new local variable 'xmode' and an assertion
18677 based on it; change to use 'xmode' to generate the new rtx and
18679 (aarch64_asm_trampoline_template): Change to generate the template
18680 differently depending on TARGET_ILP32 or not; change to use
18681 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
18682 (aarch64_trampoline_size): Removed.
18683 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
18684 and replace immediate literals with it. Change to use 'ptr_mode'
18685 instead of 'DImode' and call convert_memory_address if the mode
18686 of 'fnaddr' doesn't equal to 'ptr_mode'.
18687 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
18689 (aarch64_elf_asm_destructor): Likewise.
18690 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
18691 on TARGET_ILP32 instead of aarch64_trampoline_size.
18692 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
18693 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
18694 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
18695 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
18696 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
18697 (storewb_pair<GPI:mode>_<P:mode>): ... this.
18698 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
18699 depending on the value of 'mode'.
18700 (add_losym_<mode>): New.
18701 (ldr_got_small_<mode>): New, based on ldr_got_small.
18702 (ldr_got_small): Remove.
18703 (ldr_got_small_sidi): New.
18704 * config/aarch64/iterators.md (P): New.
18705 (PTR): Change to 'ptr_mode' in the condition.
18707 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
18709 * config.gcc (aarch64*-*-*): Support --with-abi.
18710 (aarch64*-*-elf): Support --with-multilib-list.
18711 (aarch64*-*-linux*): Likewise.
18712 (supported_defaults): Add abi to aarch64*-*-*.
18713 * configure.ac: Mention AArch64 for --with-multilib-list.
18714 * configure: Re-generated.
18715 * config/aarch64/biarchilp32.h: New file.
18716 * config/aarch64/biarchlp64.h: New file.
18717 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
18719 (MULTILIB_DEFAULTS): Ditto.
18720 (DRIVER_SELF_SPECS): Ditto.
18721 (ASM_SPEC): Update to also substitute -mabi.
18722 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
18723 file whose name depends on -mabi= and -mbig-endian.
18724 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
18726 (POINTER_SIZE): New define.
18727 (POINTERS_EXTEND_UNSIGNED): Ditto.
18728 (enum aarch64_abi_type): New enumeration tag.
18729 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
18730 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
18731 (TARGET_ILP32): New define.
18732 * config/aarch64/aarch64.opt (mabi): New.
18733 (aarch64_abi): New.
18734 (ilp32, lp64): New values for -mabi.
18735 * config/aarch64/t-aarch64 (comma): New define.
18736 (MULTILIB_OPTIONS): Ditto.
18737 (MULTILIB_DIRNAMES): Ditto.
18738 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
18739 * doc/invoke.texi: Document -mabi for AArch64.
18741 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
18743 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
18745 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
18746 Anton Blanchard <anton@au1.ibm.com>
18748 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
18749 endianness when selecting field to splat.
18751 2013-07-22 Eric Christopher <echristo@gmail.com>
18753 * dwarf2out.c (die_odr_checksum): New function to use
18754 CHECKSUM_ macros and ULEB128 for DIE tag.
18755 (generate_type_signature): Use.
18757 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
18759 * config.gcc (sparc*-*-*): Accept leon3 processor.
18760 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
18761 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
18762 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
18763 * config/sparc/sparc.opt (enum processor_type): Add leon3.
18764 (mfix-ut699): Adjust comment.
18765 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
18766 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
18767 (CPP_CPU_SPEC): Likewise.
18768 (ASM_CPU_SPEC): Likewise.
18769 * config/sparc/sparc.c (leon3_cost): New constant.
18770 (sparc_option_override): Add leon3 support.
18771 (mem_ref): New function.
18772 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
18773 (sparc_do_work_around_errata): Look into the instruction in the delay
18774 slot and adjust accordingly. Add fix for the data cache nullify issues
18775 of the UT699. Change insertion position for the NOP.
18776 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
18777 (leon3_load): New reservation.
18778 (leon_store): Bump latency to 2.
18779 (grfpu): New automaton.
18780 (grfpu_alu): New unit.
18781 (grfpu_ds): Likewise.
18782 (leon_fp_alu): Adjust.
18783 (leon_fp_mult): Delete.
18784 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
18785 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
18786 * config/sparc/sparc.md (cpu): Add leon3.
18787 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
18788 (swapsi): Likewise.
18789 (atomic_test_and_set): Likewise.
18790 (ldstub): Likewise.
18792 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
18794 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
18795 default for R5900 targets.
18796 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
18797 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
18798 * config/mips/mips.c (mips_option_override): Report an error for
18799 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
18800 for -march=r5900 -mhard-float.
18802 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
18804 * df-problems.c (can_move_insns_across): Exit loop once we
18805 find a non-fixed, non-global register.
18807 * ipa-pure-const.c (propagate_nothrow): Exit loop after
18810 * omega.c (omega_eliminate_red): Break after setting red_found.
18811 (omega_problem_has_red_equations): Similarly after setting found.
18812 (omega_query_variable): Similarly after setting coupled.
18814 2013-07-22 Marek Polacek <polacek@redhat.com>
18816 * gimplify.c: Don't include gimple.h twice.
18818 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18820 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
18821 instead of TARGET_THUMB1.
18822 (Pz): New constraint.
18823 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
18825 (compare_negsi_si): Likewise.
18826 (compare_addsi2_op0): Likewise.
18827 (compare_addsi2_op1): Likewise.
18828 (addsi3_carryin_<optab>): Likewise.
18829 (addsi3_carryin_alt2_<optab>): Likewise.
18830 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
18831 for arm_restrict_it.
18832 (subsi3_carryin): Likewise.
18833 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
18834 (minmax_arithsi): Disable for arm_restrict_it.
18835 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
18836 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
18837 (satsi_<SAT:code>_shift): Likewise.
18838 (arm_shiftsi3): Add alternative for 16-bit encoding.
18839 (arm32_movhf): Disable for arm_restrict_it.
18840 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
18841 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
18843 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
18845 * config/arm/arm.md (attribute "insn"): Delete.
18846 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
18847 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
18848 (not_shiftsi): Update for attribute change.
18849 (not_shiftsi_compare0): Likewise.
18850 (not_shiftsi_compare0_scratch): Likewise.
18851 (arm_one_cmplsi2): Likewise.
18852 (thumb1_one_cmplsi2): Likewise.
18853 (notsi_compare0): Likewise.
18854 (notsi_compare0_scratch): Likewise.
18855 (thumb1_movdi_insn): Likewise.
18856 (arm_movsi_insn): Likewise.
18857 (movhi_insn_arch4): Likewise.
18858 (movhi_bytes): Likewise.
18859 (arm_movqi_insn): Likewise.
18860 (thumb1_movqi_insn): Likewise.
18861 (arm32_movhf): Likewise.
18862 (thumb1_movhf): Likewise.
18863 (arm_movsf_soft_insn): Likewise.
18864 (thumb1_movsf_insn): Likewise.
18865 (thumb_movdf_insn): Likewise.
18866 (movsicc_insn): Likewise.
18867 (movsfcc_soft_insn): Likewise.
18868 (and_scc): Likewise.
18869 (cond_move): Likewise.
18870 (if_move_not): Likewise.
18871 (if_not_move): Likewise.
18872 (if_shift_move): Likewise.
18873 (if_move_shift): Likewise.
18874 (if_shift_shift): Likewise.
18875 (if_not_arith): Likewise.
18876 (if_arith_not): Likewise.
18877 (cond_move_not): Likewise.
18878 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
18879 (neon_mov<mode>): Likewise.
18880 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
18881 (thumb2_movsi_vfp): Likewise.
18882 (movsf_vfp): Likewise.
18883 (thumb2_movsf_vfp): Likewise.
18884 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
18886 (cortexa7_older_only): Likewise.
18887 (cortexa7_younger): Likewise.
18888 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
18889 (1020alu_shift_op): Likewise.
18890 (1020alu_shift_reg_op): Likewise.
18891 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
18892 (alu_shift_op): Likewise.
18893 (alu_shift_reg_op): Likewise.
18894 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
18895 (11_alu_shift_op): Likewise.
18896 (11_alu_shift_reg_op): Likewise.
18897 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
18898 (9_alu_shift_reg_op): Likewise.
18899 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
18901 (cortex_a15_alu_shift): Likewise.
18902 (cortex_a15_alu_shift_reg): Likewise.
18903 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
18904 (cortex_a5_alu_shift): Likewise.
18905 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
18907 (cortex_a53_alu_shift): Likewise.
18908 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
18910 (cortex_a7_alu_reg): Likewise.
18911 (cortex_a7_alu_shift): Likewise.
18912 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
18913 (cortex_a8_alu_shift): Likewise.
18914 (cortex_a8_alu_shift_reg): Likewise.
18915 (cortex_a8_mov): Likewise.
18916 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
18917 (cortex_a9_dp_shift): Likewise.
18918 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
18919 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
18920 (cortex_r4_mov): Likewise.
18921 (cortex_r4_alu_shift): Likewise.
18922 (cortex_r4_alu_shift_reg): Likewise.
18923 * config/arm/fa526.md (526_alu_op): Update for attribute change.
18924 (526_alu_shift_op): Likewise.
18925 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
18926 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
18927 (626te_alu_shift_op): Likewise.
18928 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
18929 (726te_alu_op): Likewise.
18930 (726te_alu_shift_op): Likewise.
18931 (726te_alu_shift_reg_op): Likewise.
18932 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
18933 (mp626_alu_shift_op): Likewise.
18934 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
18935 (pj4_alu_e1_conds): Likewise.
18936 (pj4_alu): Likewise.
18937 (pj4_alu_conds): Likewise.
18938 (pj4_shift): Likewise.
18939 (pj4_shift_conds): Likewise.
18940 (pj4_alu_shift): Likewise.
18941 (pj4_alu_shift_conds): Likewise.
18943 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18945 * config/arm/predicates.md (shiftable_operator_strict_it):
18947 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
18948 Disable cond_exec version for arm_restrict_it.
18949 (thumb2_smaxsi3): Convert to generate cond_exec.
18950 (thumb2_sminsi3): Likewise.
18951 (thumb32_umaxsi3): Likewise.
18952 (thumb2_uminsi3): Likewise.
18953 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
18954 (thumb2_neg_abssi2): Likewise.
18955 (thumb2_mov_scc): Add alternative for 16-bit encoding.
18956 (thumb2_movsicc_insn): Adjust alternatives.
18957 (thumb2_mov_negscc): Disable for arm_restrict_it.
18958 (thumb2_mov_negscc_strict_it): New pattern.
18959 (thumb2_mov_notscc_strict_it): New pattern.
18960 (thumb2_mov_notscc): Disable for arm_restrict_it.
18961 (thumb2_ior_scc): Likewise.
18962 (thumb2_ior_scc_strict_it): New pattern.
18963 (thumb2_cond_move): Adjust for arm_restrict_it.
18964 (thumb2_cond_arith): Disable for arm_restrict_it.
18965 (thumb2_cond_arith_strict_it): New pattern.
18966 (thumb2_cond_sub): Adjust for arm_restrict_it.
18967 (thumb2_movcond): Likewise.
18968 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
18969 (thumb2_zero_extendhisi2_v6): Likewise.
18970 (thumb2_zero_extendqisi2_v6): Likewise.
18971 (orsi_notsi_si): Likewise.
18972 (orsi_not_shiftsi_si): Likewise.
18974 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
18976 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
18977 instruction sequence is 1 byte shorter.
18979 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
18981 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
18982 it is not needed after split.
18984 2013-07-20 Iain Sandoe <iain@codesourcery.com>
18987 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
18988 second label for nonlocal goto receivers. Don't output pic base labels
18989 unless we're producing PIC; mark that action unreachable().
18990 (ix86_save_reg): If the function contains a nonlocal label, save the
18992 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
18993 * config/darwin.c (emitted_pic_label_num): New GTY.
18994 (update_pic_label_number_if_needed): New.
18995 (machopic_output_function_base_name): Adjust for nonlocal receiver
18997 (machopic_should_output_picbase_label): New.
18998 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
18999 (nonlocal_goto_receiver): New insn and split.
19001 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
19003 * config/aarch64/aarch64-builtins.c
19004 (aarch64_fold_builtin): Fold abs in all modes.
19005 * config/aarch64/aarch64-simd-builtins.def
19006 (abs): Enable for all modes.
19007 * config/aarch64/arm_neon.h
19008 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
19009 (vabs_f64): Add missing intrinsic.
19011 2013-07-19 Ian Bolton <ian.bolton@arm.com>
19013 * config/aarch64/arm_neon.h (vabs_s64): New function
19015 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
19018 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
19019 * config/avr/avr.md (adjust_len): Add `round'.
19020 * config/avr/avr-protos.h (avr_out_round): New prototype.
19021 (avr_out_plus): Add `out_label' argument.
19022 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
19023 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
19024 Handle the case where `insn' is just a pattern.
19025 (avr_out_bitop): Handle the case where `insn' is just a pattern.
19026 (avr_out_round): New function.
19027 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
19029 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
19031 * config/microblaze/microblaze.c (microblaze_expand_prologue):
19032 Add check for flag_stack_usage to handle -fstack-usage support
19034 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
19036 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
19037 interaction for new Power8 flags and VSX.
19039 2013-07-18 Sriraman Tallam <tmsriram@google.com>
19041 PR middle-end/57698
19042 * tree-inline.c (expand_call_inline): Emit errors during
19043 early_inlining only if optimization is not turned on.
19045 2013-07-18 David Malcolm <dmalcolm@redhat.com>
19049 * passes.c (init_optimization_passes): Move the construction of
19050 the pass hierarchy into a new passes.def file.
19052 * Makefile.in (passes.o): Add dependency on passes.def.
19054 2013-07-18 David Malcolm <dmalcolm@redhat.com>
19056 * passes.c (init_optimization_passes): Introduce macros for
19057 constructing the tree of passes (INSERT_PASSES_AFTER,
19058 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
19059 TERMINATE_PASS_LIST).
19061 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
19062 Wei Mi <wmi@google.com>
19064 PR rtl-optimization/57878
19065 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
19067 (reload_pseudo_compare_func): Check nregs first for reload
19070 2013-07-18 David Malcolm <dmalcolm@redhat.com>
19072 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
19074 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
19076 * read-rtl.c (validate_const_int): Once an invalid character is
19077 seen, quit the loop.
19079 * gengtype.c (write_roots): Similarly once we find the "deletable"
19080 or "if_marked" option.
19082 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
19084 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
19085 "xtab" and "sat". Move value "clz" from here to ...
19086 (attriubte "type"): ... here.
19087 (satsi_<SAT:code>): Delete "insn" attribute.
19088 (satsi_<SAT:code>_shift): Likewise.
19089 (arm_zero_extendqisi2addsi): Likewise.
19090 (arm_extendqisi2addsi): Likewise.
19091 (clzsi2): Update for attribute changes.
19092 (rbitsi2): Likewise.
19093 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
19095 (arm_usatsihi): Likewise.
19096 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
19098 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
19100 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
19101 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
19102 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
19103 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
19104 in alphabetical order.
19105 (attribute "core_cycles"): Update for attribute changes.
19106 (arm_addsi3): Likewise.
19107 (addsi3_compare0): Likewise.
19108 (addsi3_compare0_scratch): Likewise.
19109 (addsi3_compare_op1): Likewise.
19110 (addsi3_compare_op2): Likewise.
19111 (compare_addsi2_op0): Likewise.
19112 (compare_addsi2_op1): Likewise.
19113 (addsi3_carryin_shift_<optab>): Likewise.
19114 (subsi3_carryin_shift): Likewise.
19115 (rsbsi3_carryin_shift): Likewise.
19116 (arm_subsi3_insn): Likewise.
19117 (subsi3_compare0): Likewise.
19118 (subsi3_compare): Likewise.
19119 (arm_andsi3_insn): Likewise.
19120 (thumb1_andsi3_insn): Likewise.
19121 (andsi3_compare0): Likewise.
19122 (andsi3_compare0_scratch): Likewise.
19123 (zeroextractsi_compare0_scratch
19124 (andsi_not_shiftsi_si): Likewise.
19125 (iorsi3_insn): Likewise.
19126 (iorsi3_compare0): Likewise.
19127 (iorsi3_compare0_scratch): Likewise.
19128 (arm_xorsi3): Likewise.
19129 (thumb1_xorsi3_insn): Likewise.
19130 (xorsi3_compare0): Likewise.
19131 (xorsi3_compare0_scratch): Likewise.
19132 (satsi_<SAT:code>_shift): Likewise.
19134 (arm_shiftsi3): Likewise.
19135 (shiftsi3_compare0): Likewise.
19136 (not_shiftsi): Likewise.
19137 (not_shiftsi_compare0): Likewise.
19138 (not_shiftsi_compare0_scratch): Likewise.
19139 (arm_one_cmplsi2): Likewise.
19140 (thumb_one_complsi2): Likewise.
19141 (notsi_compare0): Likewise.
19142 (notsi_compare0_scratch): Likewise.
19143 (thumb1_zero_extendhisi2): Likewise.
19144 (arm_zero_extendhisi2): Likewise.
19145 (arm_zero_extendhisi2_v6): Likewise.
19146 (arm_zero_extendhisi2addsi): Likewise.
19147 (thumb1_zero_extendqisi2): Likewise.
19148 (thumb1_zero_extendqisi2_v6): Likewise.
19149 (arm_zero_extendqisi2): Likewise.
19150 (arm_zero_extendqisi2_v6): Likewise.
19151 (arm_zero_extendqisi2addsi): Likewise.
19152 (thumb1_extendhisi2): Likewise.
19153 (arm_extendhisi2): Likewise.
19154 (arm_extendhisi2_v6): Likewise.
19155 (arm_extendqisi): Likewise.
19156 (arm_extendqisi_v6): Likewise.
19157 (arm_extendqisi2addsi): Likewise.
19158 (thumb1_extendqisi2): Likewise.
19159 (thumb1_movdi_insn): Likewise.
19160 (arm_movsi_insn): Likewise.
19161 (movsi_compare0): Likewise.
19162 (movhi_insn_arch4): Likewise.
19163 (movhi_bytes): Likewise.
19164 (arm_movqi_insn): Likewise.
19165 (thumb1_movqi_insn): Likewise.
19166 (arm32_movhf): Likewise.
19167 (thumb1_movhf): Likewise.
19168 (arm_movsf_soft_insn): Likewise.
19169 (thumb1_movsf_insn): Likewise.
19170 (movdf_soft_insn): Likewise.
19171 (thumb_movdf_insn): Likewise.
19172 (arm_cmpsi_insn): Likewise.
19173 (cmpsi_shiftsi): Likewise.
19174 (cmpsi_shiftsi_swp): Likewise.
19175 (arm_cmpsi_negshiftsi_si): Likewise.
19176 (movsicc_insn): Likewise.
19177 (movsfcc_soft_insn): Likewise.
19178 (arith_shiftsi): Likewise.
19179 (arith_shiftsi_compare0
19180 (arith_shiftsi_compare0_scratch
19181 (sub_shiftsi): Likewise.
19182 (sub_shiftsi_compare0
19183 (sub_shiftsi_compare0_scratch
19184 (and_scc): Likewise.
19185 (cond_move): Likewise.
19186 (if_plus_move): Likewise.
19187 (if_move_plus): Likewise.
19188 (if_move_not): Likewise.
19189 (if_not_move): Likewise.
19190 (if_shift_move): Likewise.
19191 (if_move_shift): Likewise.
19192 (if_shift_shift): Likewise.
19193 (if_not_arith): Likewise.
19194 (if_arith_not): Likewise.
19195 (cond_move_not): Likewise.
19196 (thumb1_ashlsi3): Set type attribute.
19197 (thumb1_ashrsi3): Likewise.
19198 (thumb1_lshrsi3): Likewise.
19199 (thumb1_rotrsi3): Likewise.
19200 (shiftsi3_compare0_scratch): Likewise.
19201 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
19202 (neon_mov<mode>): Likewise.
19203 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
19205 (thumb2_movsi_insn): Likewise.
19206 (thumb2_cmpsi_neg_shiftsi): Likewise.
19207 (thumb2_extendqisi_v6): Likewise.
19208 (thumb2_zero_extendhisi2_v6): Likewise.
19209 (thumb2_zero_extendqisi2_v6): Likewise.
19210 (thumb2_shiftsi3_short): Likewise.
19211 (thumb2_addsi3_compare0_scratch): Likewise.
19212 (orsi_not_shiftsi_si): Likewise.
19213 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
19214 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
19216 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
19217 (1020alu_shift_op): Likewise.
19218 (1020alu_shift_reg_op): Likewise.
19219 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
19220 (alu_shift_op): Likewise.
19221 (alu_shift_reg_op): Likewise.
19222 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
19223 (11_alu_shift_op): Likewise.
19224 (11_alu_shift_reg_op): Likewise.
19225 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
19226 (9_alu_shift_reg_op): Likewise.
19227 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
19229 (cortex_a15_alu_shift): Likewise.
19230 (cortex_a15_alu_shift_reg): Likewise.
19231 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
19233 (cortex_a5_alu_shift): Likewise.
19234 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
19236 (cortex_a53_alu_shift): Likewise.
19237 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
19239 (cortex_a7_alu_reg): Likewise.
19240 (cortex_a7_alu_shift): Likewise.
19241 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
19243 (cortex_a8_alu_shift): Likewise.
19244 (cortex_a8_alu_shift_reg): Likewise.
19245 (cortex_a8_mov): Likewise.
19246 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
19247 (cortex_a9_dp_shift): Likewise.
19248 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
19250 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
19252 (cortex_r4_mov): Likewise.
19253 (cortex_r4_alu_shift): Likewise.
19254 (cortex_r4_alu_shift_reg): Likewise.
19255 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
19256 (526_alu_shift_op): Likewise.
19257 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
19258 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
19259 (626te_alu_shift_op): Likewise.
19260 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
19261 (726te_alu_op): Likewise.
19262 (726te_alu_shift_op): Likewise.
19263 (726te_alu_shift_reg_op): Likewise.
19264 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
19265 (mp626_alu_shift_op): Likewise.
19266 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
19267 (pj4_alu_e1_conds): Likewise.
19268 (pj4_alu): Likewise.
19269 (pj4_alu_conds): Likewise.
19270 (pj4_shift): Likewise.
19271 (pj4_shift_conds): Likewise.
19272 (pj4_alu_shift): Likewise.
19273 (pj4_alu_shift_conds): Likewise.
19274 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
19276 (cortexa7_older_only): Likewise.
19277 (cortexa7_younger): Likewise.
19279 2013-07-18 David Malcolm <dmalcolm@redhat.com>
19281 * ipa-pure-const.c (generate_summary): Rename to...
19282 (pure_const_generate_summary): ... this.
19284 2013-07-17 Iain Sandoe <iain@codesourcery.com>
19286 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
19288 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
19291 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
19294 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19296 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
19297 enabled without -march=zEC12.
19298 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
19301 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
19303 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
19304 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
19305 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
19306 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
19307 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
19308 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
19309 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
19311 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
19312 and ISA_HAS_NMADD3_NMSUB3.
19313 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
19314 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
19315 (nmsub4<mode>, nmsub3<mode>): Likewise.
19316 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
19318 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
19320 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
19321 TARGET_MIPS5400 checking.
19323 2013-07-16 Jakub Jelinek <jakub@redhat.com>
19324 Peter Bergner <bergner@vnet.ibm.com>
19326 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
19327 registers in the comment.
19328 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
19329 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
19330 rather than FIRST_PSEUDO_REGISTERS.
19332 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
19334 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
19335 enable extra ISA flags with TARGET_HTM.
19337 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
19339 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
19342 2013-07-15 Cong Hou <congh@google.com>
19344 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
19345 in compare function for sorting.
19347 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
19349 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
19350 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
19351 * config/rs6000/rs6000.opt: Add -mhtm option.
19352 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
19353 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
19354 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
19355 __HTM__ if the HTM instructions are available.
19356 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
19357 htm_spr_reg_operand): New define_predicates.
19358 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
19359 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
19361 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
19362 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
19363 HTM builtin functions.
19364 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
19365 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
19366 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
19367 (rs6000_builtin_mask_calculate): Likewise.
19368 (rs6000_option_override_internal): Likewise.
19369 (bdesc_htm): Add new HTM builtin support.
19370 (htm_spr_num): New function.
19371 (htm_spr_regno): Likewise.
19372 (rs6000_htm_spr_icode): Likewise.
19373 (htm_expand_builtin): Likewise.
19374 (htm_init_builtins): Likewise.
19375 (rs6000_expand_builtin): Add support for HTM builtin functions.
19376 (rs6000_init_builtins): Likewise.
19377 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
19379 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
19380 (TARGET_HTM, MASK_HTM): Define macros.
19381 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
19382 (FIXED_REGISTERS): Likewise.
19383 (CALL_USED_REGISTERS): Likewise.
19384 (CALL_REALLY_USED_REGISTERS): Likewise.
19385 (REG_ALLOC_ORDER): Likewise.
19386 (enum reg_class): Likewise.
19387 (REG_CLASS_NAMES): Likewise.
19388 (REG_CLASS_CONTENTS): Likewise.
19389 (REGISTER_NAMES): Likewise.
19390 (ADDITIONAL_REGISTER_NAMES): Likewise.
19391 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
19392 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
19393 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
19394 * config/rs6000/htm.md: New file.
19395 * config/rs6000/htmintrin.h: New file.
19396 * config/rs6000/htmxlintrin.h: New file.
19398 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
19400 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
19401 Define SYMBOL_TINY_GOT, update comment.
19402 * config/aarch64/aarch64.c
19403 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
19404 (aarch64_expand_mov_immediate): Likewise.
19405 (aarch64_print_operand): Likewise.
19406 (aarch64_classify_symbol): Likewise.
19407 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
19408 (ldr_got_tiny): Define.
19410 2013-07-13 Tobias Grosser <tobias@grosser.es>
19412 PR tree-optimization/54094
19413 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
19414 scheduling dimension for the parallelism check from the polyhedral
19415 information in the AST.
19416 * graphite-dependences.c (carries_deps): Do not assume the schedule is
19419 2013-07-13 Jason Merrill <jason@redhat.com>
19421 * print-tree.c (debug_vec_tree): Use debug_raw.
19422 (debug_raw (vec<tree, va_gc> &)): New.
19423 (debug_raw (vec<tree, va_gc> *)): New.
19424 * tree.h: Declare them.
19426 2013-07-13 Bin Cheng <bin.cheng@arm.com>
19428 * ifcvt.c (ifcvt_after_combine): New static variable.
19429 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
19431 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
19432 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
19433 rest_of_handle_if_after_reload): Pass new argument for if_convert.
19435 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
19437 * config/mips/mips.c (mips_expand_call): Remove empty statement.
19439 2013-07-12 Michael Matz <matz@suse.de>
19441 PR middle-end/55771
19442 * convert.c (convert_to_real): Reject non-float inner types.
19444 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
19446 * config/aarch64/aarch64-protos.h
19447 (aarch64_simd_immediate_valid_for_move): Remove.
19448 * config/aarch64/aarch64.c (simd_immediate_info): New member.
19449 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
19451 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
19453 2013-07-11 Steve Ellcey <sellcey@mips.com>
19455 * config/mips/mips.c (mips_conditional_register_usage): Do not
19456 use t[0-7] registers in MIPS16 mode when optimizing for size.
19458 2013-07-11 Sriraman Tallam <tmsriram@google.com>
19460 * config/i386/i386.c (dispatch_function_versions): Fix array
19461 indexing of function_version_info to match actual_versions.
19463 2013-07-11 Teresa Johnson <tejohnson@google.com>
19465 * vec.h (struct va_gc): Move release out-of-line.
19466 (va_gc::release): Call ggc_free on released vec.
19468 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
19470 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
19471 Require GOT register as additional operand in UNSPEC.
19472 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
19473 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
19474 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
19475 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
19476 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
19477 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
19478 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
19480 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
19483 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
19484 name seen by assembler/linker rather if available.
19486 2013-07-11 Andreas Schwab <schwab@suse.de>
19488 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
19490 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
19492 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
19494 2013-07-10 Joseph Myers <joseph@codesourcery.com>
19496 * doc/tm.texi.in: Move hook documentation to ....
19497 * target.def: ... here.
19499 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
19500 text on @hook line.
19501 * doc/tm.texi: Regenerate.
19503 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
19506 * doc/invoke.texi: Document -Wconditionally-supported.
19508 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
19511 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
19514 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
19517 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
19518 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
19519 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
19520 Remove duplicate devices.
19521 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
19522 * config/avr/t-multilib: Regenerate.
19523 * config/avr/avr-tables.opt: Regenerate.
19524 * doc/avr-mmcu.texi: Regenerate.
19526 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
19529 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
19531 2013-07-10 Graham Stott <graham.stott@btinternet.com>
19533 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
19534 the cost of MULT when optimizing for size.
19536 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
19538 * config/cr16/cr16-protos.h: Don't include target.h.
19540 2013-07-09 Joseph Myers <joseph@codesourcery.com>
19542 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
19543 adjust register size for TDmode and TFmode for VSX registers.
19545 2013-07-08 Kai Tietz <ktietz@redhat.com>
19548 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
19549 hook_bool_const_tree_true.
19551 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19553 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
19554 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
19555 * config/s390/s390.md: Define FPR*_REGNUM constants.
19556 Fix FPR2_REGNUM constant (18 -> 17).
19557 ("*trunc<BFP:mode><DFP_ALL:mode>2")
19558 ("*trunc<DFP_ALL:mode><BFP:mode>2")
19559 ("trunc<BFP:mode><DFP_ALL:mode>2")
19560 ("trunc<DFP_ALL:mode><BFP:mode>2")
19561 ("*extend<BFP:mode><DFP_ALL:mode>2")
19562 ("*extend<DFP_ALL:mode><BFP:mode>2")
19563 ("extend<BFP:mode><DFP_ALL:mode>2")
19564 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
19567 2013-07-08 Graham Stott <graham.stott@btinternet.com>
19569 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
19571 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19573 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
19574 and cfun_fpr_bit_p to cfun_fpr_save_p.
19575 (s390_frame_area, s390_register_info, s390_frame_info)
19576 (s390_emit_prologue, s390_emit_epilogue)
19577 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
19579 * config/s390/s390.h: Define *_REGNUM macros for floating point
19582 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
19584 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
19586 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
19588 PR rtl-optimization/57786
19589 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
19590 and break out of the loop when it is set to false.
19592 2013-07-08 Jakub Jelinek <jakub@redhat.com>
19595 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
19596 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
19597 (const_int 63)) 0)).
19598 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
19599 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
19600 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
19602 PR rtl-optimization/57829
19603 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
19604 mask bits outside of mode are just sign-extension from mode to HWI.
19606 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
19608 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
19609 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
19610 adjust_address instead of change_address to keep info about alignment.
19611 (emit_strmov): Remove.
19612 (emit_memmov): New function.
19613 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
19614 (expand_movmem_epilogue): Likewise and return updated rtx for
19616 (expand_constant_movmem_prologue): Likewise and return updated rtx for
19617 destination and source.
19618 (decide_alignment): Refactor, handle vector_loop.
19619 (ix86_expand_movmem): Likewise.
19620 (ix86_expand_setmem): Likewise.
19621 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
19623 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
19625 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
19626 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
19628 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
19630 * config/i386/sse.md (sse_movlhps): Change alternative 3
19631 of operand 2 to "m".
19633 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
19636 * config/i386/sse.md (iptr): New mode attribute.
19637 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
19638 (<sse>_vm<plusminus_insn><mode>3): Ditto.
19639 (<sse>_vmmul<mode>3): Ditto.
19640 (<sse>_vmdiv<mode>3): Ditto.
19641 (sse_vmrcpv4sf2): Ditto.
19642 (<sse>_vmsqrt<mode>2): Ditto.
19643 (sse_vmrsqrtv4sf2): Ditto.
19644 (<sse>_vm<code><mode>3): Ditto.
19645 (avx_vmcmp<mode>3): Ditto.
19646 (<sse>_vmmaskcmp<mode>3): Ditto.
19647 (<sse>_comi): Ditto.
19648 (<sse>_ucomi): Ditto.
19649 (*xop_vmfrcz_<mode>): Ditto.
19650 (*fmai_fmadd_<mode>): Ditto.
19651 (*fmai_fmsub_<mode>): Ditto.
19652 (*fmai_fnmadd_<mode>): Ditto.
19653 (*fmai_fnmsub_<mode>): Ditto.
19654 (*fma4i_vmfmadd_<mode>): Ditto.
19655 (*fma4i_vmfmsub_<mode>): Ditto.
19656 (*fma4i_vmfnmadd_<mode>): Ditto.
19657 (*fma4i_vmfnmsub_<mode>): Ditto.
19658 (*xop_vmfrcz_<mode>): Ditto.
19659 (sse_cvtps2pi): Ditto.
19660 (sse_cvttps2pi): Ditto.
19661 (sse_cvtss2si): Ditto.
19662 (sse_cvtss2si_2): Ditto.
19663 (sse_cvtss2siq_2): Ditto.
19664 (sse_cvttss2si): Ditto.
19665 (sse_cvttss2siq): Ditto.
19666 (sse_cvtsd2si): Ditto.
19667 (sse_cvtsd2si_2): Ditto.
19668 (sse_cvtsd2siq_2): Ditto.
19669 (sse_cvttsd2si): Ditto.
19670 (sse_cvttsd2siq): Ditto.
19671 (sse_cvtsd2ss): Ditto.
19672 (sse_cvtss2sd): Ditto.
19673 (avx2_pbroadcast<mode>): Ditto.
19674 (avx2_pbroadcast<mode>_1): Ditto.
19675 (*avx_vperm_broadcast_v4sf): Ditto.
19677 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
19678 (sse_movlhps): Ditto.
19679 (sse_storehps): Ditto.
19680 (sse_loadhps): Ditto.
19681 (sse_storelps): Ditto.
19682 (sse_loadlps): Ditto.
19683 (*vec_concatv4sf): Ditto.
19684 (*vec_interleave_highv2df): Ditto.
19685 (*vec_interleave_lowv2df): Ditto.
19686 (*vec_extractv2df_1_sse): Ditto.
19687 (*vec_extractv2df_0_sse): Ditto.
19688 (sse2_storelpd): Ditto.
19689 (sse2_loadlpd): Ditto.
19690 (sse2_movsd): Ditto.
19691 (*vec_concatv4si): Ditto.
19692 (vec_concatv2di): Ditto.
19694 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
19695 for Intel asm dialect.
19696 (mmx_punpcklwd): Ditto.
19697 (mmx_punpckldq): Ditto.
19699 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
19700 for intel assembler dialect.
19702 2013-07-06 Jakub Jelinek <jakub@redhat.com>
19705 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
19706 for BUILT_IN_C{LZ,LRSB}*.
19707 * tree.h (CASE_INT_FN): Add FN##IMAX case.
19708 * tree-vrp.c (extract_range_basic): Handle
19709 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
19710 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
19711 fall thru to code calling set_value*.
19712 * builtins.c (expand_builtin): Remove *IMAX cases.
19713 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
19714 if width is bigger than 2*HWI.
19716 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
19718 PR rtl-optimization/55342
19719 * lra-int.h (lra_subreg_reload_pseudos): New.
19720 * lra.c: Add undoing optional reloads to the block diagram.
19721 (lra_subreg_reload_pseudos): New.
19722 (lra_optional_reload_pseudos): Change comments.
19723 (lra): Init and clear lra_subreg_reload_pseudos. Clear
19724 lra_optional_reload_pseudos after undo transformations.
19725 * lra-assigns.c (pseudo_prefix_title): New.
19726 (lra_setup_reg_renumber): Use it.
19727 (spill_for): Ditto. Check subreg reload pseudos too.
19728 (assign_by_spills): Consider subreg reload pseudos too.
19729 * lra-constraints.c (simplify_operand_subreg): Use
19730 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
19731 (curr_insn_transform): Recognize and do optional reloads.
19732 (undo_optional_reloads): New.
19733 (lra_undo_inheritance): Call undo_optional_reloads.
19735 2013-07-05 Thomas Quinot <quinot@adacore.com>
19737 * tree-complex.c (expand_complex_operations_1): Fix typo.
19739 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
19741 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
19742 (tune_params): New member 'const vec_costs'.
19743 * config/aarch64/aarch64.c (generic_vector_cost): New.
19744 (generic_tunings): New member 'generic_vector_cost'.
19745 (aarch64_builtin_vectorization_cost): New.
19746 (aarch64_add_stmt_cost): New.
19747 (TARGET_VECTORIZE_ADD_STMT_COST): New.
19748 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
19750 2013-07-03 Jakub Jelinek <jakub@redhat.com>
19753 * config/i386/predicates.md (vsib_address_operand): Disallow
19754 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
19756 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
19758 PR middle-end/55030
19759 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
19760 expand_builtin_setjmp_receiver.
19761 (expand_label): Adjust, call expand_builtin_setjmp_receiver
19762 with NULL for the label parameter.
19763 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
19764 the frame-pointer. Adjust comments.
19765 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
19766 only if LABEL is non-NULL.
19768 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
19770 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
19771 (ARM_ABI_AAPCS64): Ditto.
19773 (ARM_DEFAULT_ABI): Ditto.
19775 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
19777 * config/aarch64/aarch64-builtins.c
19778 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
19779 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
19781 * config/aarch64/aarch64-simd.md
19782 (aarch64_ld1<VALL:mode>): New.
19783 (aarch64_st1<VALL:mode>): Likewise.
19784 * config/aarch64/arm_neon.h
19785 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
19787 2013-07-02 Sriraman Tallam <tmsriram@google.com>
19789 * config/i386/i386.c (gate_insert_vzeroupper): Check if
19791 (ix86_option_override_internal):Turn on all -mavx target flags by
19792 default as they are dependent on AVX anyway.
19794 2013-07-02 Cary Coutant <ccoutant@google.com>
19796 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
19797 deterministic hash.
19798 (loc_checksum_ordered): Likewise.
19799 (hash_loc_operands): Remove inline keyword.
19801 2013-07-02 Jakub Jelinek <jakub@redhat.com>
19803 PR tree-optimization/57741
19804 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
19805 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
19806 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
19807 Allow REAL_CST step_exprs if flag_associative_math.
19808 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
19810 2013-07-02 Ian Bolton <ian.bolton@arm.com>
19812 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
19814 2013-07-02 Ian Bolton <ian.bolton@arm.com>
19816 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
19818 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19820 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
19822 (iorsi3_insn): Likewise.
19823 (arm_xorsi3): Likewise.
19825 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
19827 * arm.md (attribute "wtype"): Delete. Move attribute values from here
19829 (attribute "type"): ... here, and prefix with "wmmx_".
19830 (attribute "core_cycles"): Update for attribute changes.
19831 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
19832 (tbcstv4hi): Likewise.
19833 (tbcstv2si): Likewise.
19834 (iwmmxt_iordi3): Likewise.
19835 (iwmmxt_xordi3): Likewise.
19836 (iwmmxt_anddi3): Likewise.
19837 (iwmmxt_nanddi3): Likewise.
19838 (iwmmxt_arm_movdi): Likewise.
19839 (iwmmxt_movsi_insn): Likewise.
19840 (mov<mode>_internal): Likewise.
19841 (and<mode>3_iwmmxt): Likewise.
19842 (ior<mode>3_iwmmxt): Likewise.
19843 (xor<mode>3_iwmmxt): Likewise.
19844 (add<mode>3_iwmmxt): Likewise.
19845 (ssaddv8qi3): Likewise.
19846 (ssaddv4hi3): Likewise.
19847 (ssaddv2si3): Likewise.
19848 (usaddv8qi3): Likewise.
19849 (usaddv4hi3): Likewise.
19850 (usaddv2si3): Likewise.
19851 (sub<mode>3_iwmmxt): Likewise.
19852 (sssubv8qi3): Likewise.
19853 (sssubv4hi3): Likewise.
19854 (sssubv2si3): Likewise.
19855 (ussubv8qi3): Likewise.
19856 (ussubv4hi3): Likewise.
19857 (ussubv2si3): Likewise.
19858 (mulv4hi3_iwmmxt): Likewise.
19859 (smulv4hi3_highpart): Likewise.
19860 (umulv4hi3_highpart): Likewise.
19861 (iwmmxt_wmacs): Likewise.
19862 (iwmmxt_wmacsz): Likewise.
19863 (iwmmxt_wmacu): Likewise.
19864 (iwmmxt_wmacuz): Likewise.
19865 (iwmmxt_clrdi): Likewise.
19866 (iwmmxt_clrv8qi): Likewise.
19867 (iwmmxt_clr4hi): Likewise.
19868 (iwmmxt_clr2si): Likewise.
19869 (iwmmxt_uavgrndv8qi3): Likewise.
19870 (iwmmxt_uavgrndv4hi3): Likewise.
19871 (iwmmxt_uavgv8qi3): Likewise.
19872 (iwmmxt_uavgv4hi3): Likewise.
19873 (iwmmxt_tinsrb): Likewise.
19874 (iwmmxt_tinsrh): Likewise.
19875 (iwmmxt_tinsrw): Likewise.
19876 (iwmmxt_textrmub): Likewise.
19877 (iwmmxt_textrmsb): Likewise.
19878 (iwmmxt_textrmuh): Likewise.
19879 (iwmmxt_textrmsh): Likewise.
19880 (iwmmxt_textrmw): Likewise.
19881 (iwmxxt_wshufh): Likewise.
19882 (eqv8qi3): Likewise.
19883 (eqv4hi3): Likewise.
19884 (eqv2si3): Likewise.
19885 (gtuv8qi3): Likewise.
19886 (gtuv4hi3): Likewise.
19887 (gtuv2si3): Likewise.
19888 (gtv8qi3): Likewise.
19889 (gtv4hi3): Likewise.
19890 (gtv2si3): Likewise.
19891 (smax<mode>3_iwmmxt): Likewise.
19892 (umax<mode>3_iwmmxt): Likewise.
19893 (smin<mode>3_iwmmxt): Likewise.
19894 (umin<mode>3_iwmmxt): Likewise.
19895 (iwmmxt_wpackhss): Likewise.
19896 (iwmmxt_wpackwss): Likewise.
19897 (iwmmxt_wpackdss): Likewise.
19898 (iwmmxt_wpackhus): Likewise.
19899 (iwmmxt_wpackwus): Likewise.
19900 (iwmmxt_wpackdus): Likewise.
19901 (iwmmxt_wunpckihb): Likewise.
19902 (iwmmxt_wunpckihh): Likewise.
19903 (iwmmxt_wunpckihw): Likewise.
19904 (iwmmxt_wunpckilb): Likewise.
19905 (iwmmxt_wunpckilh): Likewise.
19906 (iwmmxt_wunpckilw): Likewise.
19907 (iwmmxt_wunpckehub): Likewise.
19908 (iwmmxt_wunpckehuh): Likewise.
19909 (iwmmxt_wunpckehuw): Likewise.
19910 (iwmmxt_wunpckehsb): Likewise.
19911 (iwmmxt_wunpckehsh): Likewise.
19912 (iwmmxt_wunpckehsw): Likewise.
19913 (iwmmxt_wunpckelub): Likewise.
19914 (iwmmxt_wunpckeluh): Likewise.
19915 (iwmmxt_wunpckeluw): Likewise.
19916 (iwmmxt_wunpckelsb): Likewise.
19917 (iwmmxt_wunpckelsh): Likewise.
19918 (iwmmxt_wunpckelsw): Likewise.
19919 (ror<mode>3): Likewise.
19920 (ashr<mode>3_iwmmxt): Likewise.
19921 (lshr<mode>3_iwmmxt): Likewise.
19922 (ashl<mode>3_iwmmxt): Likewise.
19923 (ror<mode>3_di): Likewise.
19924 (ashr<mode>3_di): Likewise.
19925 (lshr<mode>3_di): Likewise.
19926 (ashl<mode>3_di): Likewise.
19927 (iwmmxt_wmadds): Likewise.
19928 (iwmmxt_wmaddu): Likewise.
19929 (iwmmxt_tmia): Likewise.
19930 (iwmmxt_tmiaph): Likewise.
19931 (iwmmxt_tmiabb): Likewise.
19932 (iwmmxt_tmiatb): Likewise.
19933 (iwmmxt_tmiabt): Likewise.
19934 (iwmmxt_tmiatt): Likewise.
19935 (iwmmxt_tmovmskb): Likewise.
19936 (iwmmxt_tmovmskh): Likewise.
19937 (iwmmxt_tmovmskw): Likewise.
19938 (iwmmxt_waccb): Likewise.
19939 (iwmmxt_wacch): Likewise.
19940 (iwmmxt_waccw): Likewise.
19941 (iwmmxt_waligni): Likewise.
19942 (iwmmxt_walignr): Likewise.
19943 (iwmmxt_walignr0): Likewise.
19944 (iwmmxt_walignr1): Likewise.
19945 (iwmmxt_walignr2): Likewise.
19946 (iwmmxt_walignr3): Likewise.
19947 (iwmmxt_wsadb): Likewise.
19948 (iwmmxt_wsadh): Likewise.
19949 (iwmmxt_wsadbz): Likewise.
19950 (iwmmxt_wsadhz): Likewise.
19951 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
19952 (iwmmxt_wabsdiffb): Likewise.
19953 (iwmmxt_wabsdiffh): Likewise.
19954 (iwmmxt_wabsdiffw): Likewise.
19955 (iwmmxt_waddsubhx): Likewise
19956 (iwmmxt_wsubaddhx): Likewise.
19957 (addc<mode>3): Likewise.
19958 (iwmmxt_avg4): Likewise.
19959 (iwmmxt_avg4r): Likewise.
19960 (iwmmxt_wmaddsx): Likewise.
19961 (iwmmxt_wmaddux): Likewise.
19962 (iwmmxt_wmaddsn): Likewise.
19963 (iwmmxt_wmaddun): Likewise.
19964 (iwmmxt_wmulwsm): Likewise.
19965 (iwmmxt_wmulwum): Likewise.
19966 (iwmmxt_wmulsmr): Likewise.
19967 (iwmmxt_wmulumr): Likewise.
19968 (iwmmxt_wmulwsmr): Likewise.
19969 (iwmmxt_wmulwumr): Likewise.
19970 (iwmmxt_wmulwl): Likewise.
19971 (iwmmxt_wqmulm): Likewise.
19972 (iwmmxt_wqmulwm): Likewise.
19973 (iwmmxt_wqmulmr): Likewise.
19974 (iwmmxt_wqmulwmr): Likewise.
19975 (iwmmxt_waddbhusm): Likewise.
19976 (iwmmxt_waddbhusl): Likewise.
19977 (iwmmxt_wqmiabb): Likewise.
19978 (iwmmxt_wqmiabt): Likewise.
19979 (iwmmxt_wqmiatb): Likewise.
19980 (iwmmxt_wqmiatt): Likewise.
19981 (iwmmxt_wqmiabbn): Likewise.
19982 (iwmmxt_wqmiabtn): Likewise.
19983 (iwmmxt_wqmiatbn): Likewise.
19984 (iwmmxt_wqmiattn): Likewise.
19985 (iwmmxt_wmiabb): Likewise.
19986 (iwmmxt_wmiabt): Likewise.
19987 (iwmmxt_wmiatb): Likewise.
19988 (iwmmxt_wmiatt): Likewise.
19989 (iwmmxt_wmiabbn): Likewise.
19990 (iwmmxt_wmiabtn): Likewise.
19991 (iwmmxt_wmiatbn): Likewise.
19992 (iwmmxt_wmiattn): Likewise.
19993 (iwmmxt_wmiawbb): Likewise.
19994 (iwmmxt_wmiawbt): Likewise.
19995 (iwmmxt_wmiawtb): Likewise.
19996 (iwmmxt_wmiawtt): Likewise.
19997 (iwmmxt_wmiawbbn): Likewise.
19998 (iwmmxt_wmiawbtn): Likewise.
19999 (iwmmxt_wmiawtbn): Likewise.
20000 (iwmmxt_wmiawttn): Likewise.
20001 (iwmmxt_wmerge): Likewise.
20002 (iwmmxt_tandc<mode>3): Likewise.
20003 (iwmmxt_torc<mode>3): Likewise.
20004 (iwmmxt_torvsc<mode>3): Likewise.
20005 (iwmmxt_textrc<mode>3): Likewise.
20006 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
20007 (wmmxt_pack): Likewise.
20008 (wmmxt_mult_c1): Likewise.
20009 (wmmxt_mult_c2): Likewise.
20010 (wmmxt_alu_c1): Likewise.
20011 (wmmxt_alu_c2): Likewise.
20012 (wmmxt_alu_c3): Likewise.
20013 (wmmxt_transfer_c1): Likewise.
20014 (wmmxt_transfer_c2): Likewise.
20015 (wmmxt_transfer_c3): Likewise.
20016 (marvell_f_iwmmxt_wstr): Likewise.
20017 (marvell_f_iwmmxt_wldr): Likewise.
20019 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
20021 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
20023 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
20026 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
20027 * lra-constraints.c (need_for_split_p): Check call used hard regs
20028 living through calls.
20030 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
20031 call used regs for call insn.
20033 2013-06-28 Jakub Jelinek <jakub@redhat.com>
20036 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
20037 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
20039 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
20041 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
20043 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
20045 * lra-constraints.c (need_for_split_p): Check call used hard regs
20046 living through calls.
20048 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
20051 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
20052 to tie with any other modes. Eliminate Altivec vector mode tests,
20053 since these are a subset of ALTIVEC or VSX vector modes. Simplify
20054 code, to return 0 if testing MODE2 for a condition, if we've
20055 already tested MODE1 for the same condition.
20057 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
20059 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
20062 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
20064 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
20065 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
20067 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
20069 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
20071 (aarch64_symbolic_constant_p): Remove.
20072 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
20073 static. Fix line length and white space.
20074 (aarch64_symbolic_constant_p): Remove.
20075 * config/aarch64/predicates.md (aarch64_valid_symref):
20076 Use aarch64_classify_symbol_expression.
20078 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20080 * config/arm/constraints.md (Ts): New constraint.
20081 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
20083 (compare_scc): Use "Ts" constraint for operand 0.
20084 (ior_scc_scc): Likewise.
20085 (and_scc_scc): Likewise.
20086 (and_scc_scc_nodom): Likewise.
20087 (ior_scc_scc_cmp): Likewise for operand 7.
20088 (and_scc_scc_cmp): Likewise.
20089 * config/arm/thumb2.md (thumb2_movsi_insn):
20090 Add alternatives for 16-bit encodings.
20091 (thumb2_movhi_insn): Likewise.
20092 (thumb2_movsicc_insn): Likewise.
20093 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
20094 (thumb2_negscc): Use "Ts" constraint.
20095 Move mvn instruction outside cond_exec block.
20096 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
20097 for 16-bit encodings.
20099 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20101 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
20103 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
20104 (mulsi3subsi): Likewise.
20105 (mulsidi3adddi): Likewise.
20106 (mulsidi3_v6): Likewise.
20107 (umulsidi3_v6): Likewise.
20108 (umulsidi3adddi_v6): Likewise.
20109 (smulsi3_highpart_v6): Likewise.
20110 (umulsi3_highpart_v6): Likewise.
20111 (mulhisi3tb): Likewise.
20112 (mulhisi3bt): Likewise.
20113 (mulhisi3tt): Likewise.
20114 (maddhisi4): Likewise.
20115 (maddhisi4tb): Likewise.
20116 (maddhisi4tt): Likewise.
20117 (maddhidi4): Likewise.
20118 (maddhidi4tb): Likewise.
20119 (maddhidi4tt): Likewise.
20120 (zeroextractsi_compare0_scratch): Likewise.
20121 (insv_zero): Likewise.
20122 (insv_t2): Likewise.
20123 (anddi_notzesidi_di): Likewise.
20124 (anddi_notsesidi_di): Likewise.
20125 (andsi_notsi_si): Likewise.
20126 (iordi_zesidi_di): Likewise.
20127 (xordi_zesidi_di): Likewise.
20128 (andsi_iorsi3_notsi): Likewise.
20129 (smax_0): Likewise.
20130 (smax_m1): Likewise.
20131 (smin_0): Likewise.
20132 (not_shiftsi): Likewise.
20133 (unaligned_loadsi): Likewise.
20134 (unaligned_loadhis): Likewise.
20135 (unaligned_loadhiu): Likewise.
20136 (unaligned_storesi): Likewise.
20137 (unaligned_storehi): Likewise.
20138 (extv_reg): Likewise.
20139 (extzv_t2): Likewise.
20140 (divsi3): Likewise.
20141 (udivsi3): Likewise.
20142 (arm_zero_extendhisi2addsi): Likewise.
20143 (arm_zero_extendqisi2addsi): Likewise.
20144 (compareqi_eq0): Likewise.
20145 (arm_extendhisi2_v6): Likewise.
20146 (arm_extendqisi2addsi): Likewise.
20147 (arm_movt): Likewise.
20148 (thumb2_ldrd): Likewise.
20149 (thumb2_ldrd_base): Likewise.
20150 (thumb2_ldrd_base_neg): Likewise.
20151 (thumb2_strd): Likewise.
20152 (thumb2_strd_base): Likewise.
20153 (thumb2_strd_base_neg): Likewise.
20154 (arm_negsi2): Add alternative for 16-bit encoding.
20155 (arm_one_cmplsi2): Likewise.
20157 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20159 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
20160 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
20161 (movdfcc): Likewise.
20162 * config/arm/vfp.md (*thumb2_movsf_vfp):
20163 Disable predication for arm_restrict_it.
20164 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
20165 (*thumb2_movdfcc_vfp): Likewise.
20166 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
20167 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
20168 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
20169 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
20170 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
20171 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
20172 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
20173 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
20174 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
20175 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
20176 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
20177 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
20178 Disable predication for arm_restrict_it.
20180 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
20182 * config/i386/bmiintrin.h (_bextr_u32): New.
20183 (_bextr_u64): Ditto.
20185 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
20187 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
20188 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
20189 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
20190 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
20191 * config/mips/n32-elf.h: ...this new file.
20193 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
20196 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
20197 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
20199 2013-06-27 Catherine Moore <clm@codesourcery.com>
20201 * config/mips/mips-tables.opt: Regenerate.
20202 * config/mips/mips-cpus.def: Add m14ke and m14kec.
20203 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
20204 * doc/invoke.texi: Add -m14kc.
20206 2013-06-27 Jakub Jelinek <jakub@redhat.com>
20209 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
20210 constraints of operand 1 and 2.
20213 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
20214 to match RTL canonicalization. Swap predicates and
20215 constraints of operand 1 and 2.
20217 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
20219 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
20220 Process OP_INOUT regs for splitting too.
20222 2013-06-27 Jakub Jelinek <jakub@redhat.com>
20224 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
20225 decl before the loop, initialize to NULL.
20226 (vectorizable_load): Initialize ptr_incr to NULL.
20228 2013-06-27 Martin Jambor <mjambor@suse.cz>
20231 * ipa-ref.h (ipa_maybe_record_reference): Declare.
20232 * ipa-ref.c (ipa_maybe_record_reference): New function.
20233 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
20234 * ipa-cp.c (create_specialized_node): Record potential references from
20236 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
20238 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
20240 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
20241 parameter 'mode' of type 'enum machine_mode mode'; change to pass
20242 'mode' to force_reg.
20243 (aarch64_add_offset): Update calls to aarch64_force_temporary.
20244 (aarch64_expand_mov_immediate): Likewise.
20246 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
20248 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
20249 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
20251 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20253 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
20254 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
20255 (struct machine_function): Add tbegin_p.
20256 (s390_canonicalize_comparison): Fold CC mode compares to
20257 conditional jump if possible.
20258 (s390_emit_jump): Return the emitted jump.
20259 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
20260 Handle CCRAWmode compares.
20261 (s390_option_override): Default to -mhtm if available.
20262 (s390_reg_clobbered_rtx): Handle floating point regs as well.
20263 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
20264 FPRs instead of df_regs_ever_live_p.
20265 (s390_optimize_nonescaping_tx): New function.
20266 (s390_init_frame_layout): Extend clobbered_regs array to cover
20268 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
20269 (s390_expand_tbegin): New function.
20270 (enum s390_builtin): New enum definition.
20271 (code_for_builtin): New array definition.
20272 (s390_init_builtins): New function.
20273 (s390_expand_builtin): New function.
20274 (TARGET_INIT_BUILTINS): Define.
20275 (TARGET_EXPAND_BUILTIN): Define.
20276 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
20277 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
20278 (s390_alc_comparison): Likewise.
20279 * config/s390/s390-modes.def: Add CCRAWmode.
20280 * config/s390/s390.h (processor_flags): Add PF_TX.
20281 (TARGET_CPU_HTM): Define macro.
20282 (TARGET_HTM): Define macro.
20283 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
20284 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
20285 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
20286 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
20287 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
20288 (TBEGIN_MASK, TBEGINC_MASK): New constants.
20289 ("*cc_to_int"): Move up.
20290 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
20291 constants other than 0.
20292 ("*ccraw_to_int"): New insn and splitter definition.
20293 ("tbegin", "tbegin_nofloat", "tbegin_retry")
20294 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
20295 ("tx_assist"): New expander.
20296 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
20297 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
20298 * config/s390/s390.opt: Add -mhtm option.
20299 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
20300 * config/s390/htmxlintrin.h: New file.
20301 * config/s390/htmintrin.h: New file.
20302 * config/s390/s390intrin.h: New file.
20303 * doc/extend.texi: Document htm builtins.
20304 * config.gcc: Add the new header files to extra_headers.
20306 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
20308 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
20309 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
20311 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
20312 Pat Haugen <pthaugen@us.ibm.com>
20313 Peter Bergner <bergner@vnet.ibm.com>
20315 * config/rs6000/power8.md: New.
20316 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
20317 setting for power8 entry.
20318 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
20319 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
20320 test for Power4/Power5 only.
20321 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
20323 (force_new_group): Adjust comment.
20324 * config/rs6000/rs6000.md: Include power8.md.
20326 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
20328 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
20329 * config/arm/arm-protos.h (arm_max_conditional_execute): New
20331 (tune_params): Update comment.
20332 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
20333 (arm_max_conditional_execute): New function.
20334 (thumb2_final_prescan_insn): Use max_insn_skipped and
20335 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
20337 2013-06-25 Jakub Jelinek <jakub@redhat.com>
20339 PR tree-optimization/57705
20340 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
20341 SSA_NAME step, provided that it is not defined inside the loop.
20342 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
20343 (get_initial_def_for_induction): Handle SSA_NAME IV step.
20345 2013-06-25 Martin Jambor <mjambor@suse.cz>
20347 PR middle-end/57670
20348 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
20349 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
20351 (ipa_note_param_call): Initialize member_ptr flag.
20352 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
20353 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
20354 (ipa_write_indirect_edge_info): Stream member_ptr flag.
20355 (ipa_read_indirect_edge_info): Likewise.
20357 2013-06-25 Richard Biener <rguenther@suse.de>
20359 PR middle-end/56977
20360 * passes.c (init_optimization_passes): Move pass_fold_builtins
20361 and pass_dce earlier with -Og.
20363 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
20365 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
20366 <BIT_FIELD_REF>: Remove trailing TAB.
20367 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
20370 2013-06-24 Martin Jambor <mjambor@suse.cz>
20372 PR tree-optimization/57358
20373 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
20374 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
20375 (ipa_analyze_params_uses): Generate pessimistic info when true.
20377 2013-06-24 Martin Jambor <mjambor@suse.cz>
20379 PR tree-optimization/57539
20380 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
20381 global.inlined_to of the new node to it. All callers changed.
20382 * ipa-inline-transform.c (clone_inlined_nodes): New variable
20383 inlining_into, pass it to cgraph_clone_node.
20384 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
20385 ipa_free_edge_args_substructures.
20386 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
20387 rdesc linked list. Do not assert rdesc edges have inlined caller.
20388 Assert we have found an rdesc in the rdesc list.
20390 2013-06-24 Richard Biener <rguenther@suse.de>
20392 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
20393 (pointer_set_lookup): Declare.
20394 (class pointer_map): New template class implementing a
20395 generic pointer to T map.
20396 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
20397 pointer_map<T>::contains, pointer_map<T>::insert,
20398 pointer_map<T>::traverse): New functions.
20399 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
20400 (pointer_set_lookup): New function.
20401 (pointer_set_contains): Use pointer_set_lookup.
20402 (pointer_set_insert): Likewise.
20403 (insert_aux): Remove.
20404 (struct pointer_map_t): Embed a pointer_set_t.
20405 (pointer_map_create): Adjust.
20406 (pointer_map_destroy): Likewise.
20407 (pointer_map_contains): Likewise.
20408 (pointer_map_insert): Likewise.
20409 (pointer_map_traverse): Likewise.
20410 * tree-streamer.h (struct streamer_tree_cache_d): Use a
20411 pointer_map<unsigned> instead of a pointer_map_t.
20412 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
20413 (streamer_tree_cache_lookup): Likewise.
20414 (streamer_tree_cache_create): Likewise.
20415 (streamer_tree_cache_delete): Likewise.
20416 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
20417 pointer_map<unsigned> instead of a pointer_map_t.
20418 (lto_init_tree_ref_encoder): Adjust.
20419 (lto_destroy_tree_ref_encoder): Likewise.
20420 * lto-section-out.c (lto_output_decl_index): Likewise.
20421 (lto_record_function_out_decl_state): Likewise.
20422 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
20424 2013-06-24 Richard Biener <rguenther@suse.de>
20426 PR tree-optimization/57488
20427 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
20429 2013-06-24 Alan Modra <amodra@gmail.com>
20431 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
20432 (gen_easy_altivec_constant): Likewise.
20433 * config/rs6000/predicates.md (easy_vector_constant_add_self,
20434 easy_vector_constant_msb): Likewise.
20436 2013-06-23 Jakub Jelinek <jakub@redhat.com>
20439 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
20440 add missing return true.
20442 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
20445 * config/sh/predicates.md (general_extend_operand): Invoke
20446 general_movsrc_operand for memory operands.
20447 (general_movsrc_operand): Allow reg+reg addressing, do not use
20448 general_operand for memory operands.
20450 2013-06-23 Sriraman Tallam <tmsriram@google.com>
20452 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
20453 when current target options does not apply.
20454 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
20455 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
20456 * config/i386/bmiintrin.h: Pass appropriate target
20457 attributes to header.
20458 * config/i386/mmintrin.h: Ditto.
20459 * config/i386/nmmintrin.h: Ditto.
20460 * config/i386/avx2intrin.h: Ditto.
20461 * config/i386/fxsrintrin.h: Ditto.
20462 * config/i386/tbmintrin.h: Ditto.
20463 * config/i386/xsaveintrin.h: Ditto.
20464 * config/i386/f16cintrin.h: Ditto.
20465 * config/i386/xtestintrin.h: Ditto.
20466 * config/i386/xsaveoptintrin.h: Ditto.
20467 * config/i386/bmi2intrin.h: Ditto.
20468 * config/i386/lzcntintrin.h: Ditto.
20469 * config/i386/smmintrin.h: Ditto.
20470 * config/i386/wmmintrin.h: Ditto.
20471 * config/i386/x86intrin.h: Remove all header include guards.
20472 * config/i386/prfchwintrin.h: Ditto.
20473 * config/i386/pmmintrin.h: Ditto.
20474 * config/i386/tmmintrin.h: Ditto.
20475 * config/i386/xmmintrin.h: Ditto.
20476 * config/i386/popcntintrin.h: Ditto.
20477 * config/i386/rdseedintrin.h: Ditto.
20478 * config/i386/ammintrin.h: Ditto.
20479 * config/i386/emmintrin.h: Ditto.
20480 * config/i386/immintrin.h: Remove all header include guards.
20481 * config/i386/fma4intrin.h: Ditto.
20482 * config/i386/lwpintrin.h: Ditto.
20483 * config/i386/xopintrin.h: Ditto.
20484 * config/i386/ia32intrin.h: Ditto.
20485 * config/i386/avxintrin.h: Ditto.
20486 * config/i386/rtmintrin.h: Ditto.
20487 * config/i386/fmaintrin.h: Ditto.
20488 * config/i386/mm3dnow.h: Ditto.
20490 2013-06-22 Sriraman Tallam <tmsriram@google.com>
20492 * common/config/i386/i386-common.c: Handle LZCNT.
20494 2013-06-22 Andi Kleen <ak@linux.intel.com>
20496 * doc/extend.texi: Use __atomic_store_n instead of
20497 __atomic_store in HLE example.
20499 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
20501 * config/sh/sh.c: Remove <cstdlib> workaround.
20503 2013-06-21 Andi Kleen <ak@linux.intel.com>
20505 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
20507 2013-06-21 Andi Kleen <ak@linux.intel.com>
20509 * doc/extend.texi: Document that __atomic_clear and
20510 __atomic_test_and_set should only be used with bool.
20512 2013-06-20 Jan Hubicka <jh@suse.cz>
20514 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
20515 types_same_for_odr.
20516 * tree.c (decls_same_for_odr): New function.
20517 (same_for_edr): New function.
20518 (types_same_for_odr): New function.
20519 (get_binfo_at_offset): Use it.
20520 * tree.h (types_same_for_odr): Declare.
20522 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
20523 Jason Merrill <jason@redhat.com>
20525 * system.h: Include <cstdlib> as well as <stdlib.h>.
20527 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
20530 * config/i386/i386.c (construct_container): Report error if
20531 long double is used with disabled x87 float returns.
20533 2013-06-20 Jan Hubicka <jh@suse.cz>
20535 * lto-cgraph.c (input_symtab): Do not set cgraph state.
20537 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
20539 PR rtl-optimization/57425
20540 PR rtl-optimization/57569
20541 * alias.c (write_dependence_p): Remove parameters mem_mode and
20542 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
20543 Changed all callers.
20544 (canon_anti_dependence): Get comments and semantics in sync.
20545 Add parameter mem_canonicalized. Changed all callers.
20546 * rtl.h (canon_anti_dependence): Update prototype.
20548 2013-06-20 Richard Biener <rguenther@suse.de>
20550 * data-streamer-in.c (streamer_read_uhwi): Optimize single
20551 byte case, inline streamer_read_uchar and defer section
20554 2013-06-20 Richard Biener <rguenther@suse.de>
20556 PR tree-optimization/57584
20557 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
20558 SSA names into the expanded expression that take part in
20559 abnormal coalescing.
20561 2013-06-19 Sharad Singhai <singhai@google.com>
20563 * gcov.c (print_usage): Handle new option.
20564 (process_args): Ditto.
20565 (get_gcov_intermediate_filename): New function.
20566 (output_intermediate_file): New function.
20567 (output_gcov_file): New function
20568 (generate_results): Handle new option.
20569 (release_function): Relase demangled name.
20570 (read_graph_file): Handle demangled name.
20571 (output_lines): Ditto.
20572 * doc/gcov.texi: Document gcov intermediate format.
20574 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
20577 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
20578 (lra_emit_add): Use the functions. Add comment about Y as an
20581 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
20584 * collect2.c (collect_atexit): New.
20585 (collect_exit): Delete.
20586 (main): Register collect_atexit with atexit.
20587 (collect_wait): Change collect_exit to exit.
20589 * collect2.h (collect_exit): Delete.
20590 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
20592 2013-06-19 Wei Mi <wmi@google.com>
20594 PR rtl-optimization/57518
20595 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
20596 if regno is used in paradoxical subreg.
20597 (update_equiv_regs): Check pdx_subregs[regno] before
20598 set a reg to be equivalent with a mem.
20600 2013-06-19 Matthias Klose <doko@ubuntu.com>
20603 * file-find.h (find_a_file): Add a mode parameter.
20604 * file-find.c (find_a_file): Likewise.
20605 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
20606 with X_OK for the executables.
20607 * collect2.c (main): Call find_a_file with X_OK.
20609 2013-06-19 Steve Ellcey <sellcey@mips.com>
20612 * config/mips/mips.md (casesi_internal_mips16_<mode>):
20613 Use NEXT_INSN instead of next_real_insn.
20615 2013-06-19 Jan Hubicka <jh@suse.cz>
20617 * cgraph.h (const_value_known_p): Replace by ...
20618 (ctor_for_folding): .. this one.
20619 * cgraphunit.c (process_function_and_variable_attributes): Use it.
20620 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
20621 * expr.c (expand_expr_real_1): Likewise.
20622 (string_constant): Likewise.
20623 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
20624 * ipa.c (process_references): Likewise.
20625 (symtab_remove_unreachable_nodes): Likewise.
20626 * ipa-inline-analysis.c (param_change_prob): Likewise.
20627 * gimple-fold.c (canonicalize_constructor_val): Likewise.
20628 (get_base_constructor): Likwise.
20629 * varpool.c (varpool_remove_node): Likewise.
20630 (varpool_remove_initializer): LIkewise.
20631 (dump_varpool_node): LIkwise.
20632 (const_value_known_p): Rewrite to ...
20633 (ctor_for_folding): ... this one.
20635 2013-06-19 Jakub Jelinek <jakub@redhat.com>
20638 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
20639 PERSONALITY in $PATH derived prefixes.
20641 2013-06-19 Jeff Law <law@redhat.com>
20643 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
20646 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
20647 (simplify_bitwise_binary): Use it to simpify certain binary ops on
20650 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
20652 * config/arm/vfp.md: Move VFP instruction classification documentation
20654 * config/arm/arm.md: ... here. Update instruction classification
20657 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
20659 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
20660 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
20661 pattern. Use more efficient sequences on ARMv5 and Thumb2.
20663 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
20666 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
20667 with NEXT_INSN. Use tablejump_p to check for jump table data
20670 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
20673 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
20674 that now in C++ the value is correct per the C++ standards.
20676 2013-06-19 Richard Biener <rguenther@suse.de>
20678 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
20679 for global context.
20681 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20684 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20687 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
20688 with next_active_insn.
20690 2013-06-18 Sriraman Tallam <tmsriram@google.com>
20692 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
20693 functions are inlined during failures to flag an error.
20694 * tree-inline.c (expand_call_inline): Allow the error to be flagged
20695 in early inline pass.
20697 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
20699 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
20702 2013-06-18 Julian Brown <julian@codesourcery.com>
20704 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
20705 Permit virtual register pre-reload if !strict.
20706 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
20708 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
20710 * config/arm/neon.md (movmisalign<mode>): Use
20711 neon_perm_struct_or_reg_operand instead of
20712 neon_struct_or_register_operand.
20713 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
20714 neon_permissive_struct_operand instead of neon_struct_operand.
20715 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
20716 neon_vector_mem_operand.
20717 * config/arm/predicates.md (neon_struct_operand): Adjust call to
20718 neon_vector_mem_operand.
20719 (neon_permissive_struct_operand): New.
20720 (neon_struct_or_register_operand): Rename to...
20721 (neon_perm_struct_or_reg_operand): This. Adjust call to
20722 neon_vector_mem_operand.
20724 2013-06-18 Richard Biener <rguenther@suse.de>
20726 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
20727 * lto-streamer.h: Include pointer-set.h.
20728 (struct lto_decl_slot): Remove.
20729 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
20730 Remove next_index entry.
20731 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
20732 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
20733 (lto_init_tree_ref_encoder): Adjust.
20734 (lto_destroy_tree_ref_encoder): Likewise.
20735 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
20736 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
20737 (lto_output_decl_index): Adjust.
20738 (lto_new_out_decl_state): Likewise.
20739 (lto_record_function_out_decl_state): Likewise.
20740 * lto-streamer-out.c (copy_function): Likewise.
20742 2013-06-18 Richard Biener <rguenther@suse.de>
20744 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
20745 * cgraphunit.c: Include cfgloop.h.
20746 (init_lowered_empty_function): Initialize the loop tree.
20747 (assemble_thunk): Insert new BBs into loops.
20749 2013-06-18 Richard Biener <rguenther@suse.de>
20751 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
20752 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
20753 the map from cache entry to cache index optional.
20754 (streamer_tree_cache_replace_tree): Adjust accordingly.
20755 (streamer_tree_cache_append): Likewise.
20756 (streamer_tree_cache_delete): Likewise.
20757 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
20758 streamer cache map from cache entry to cache index.
20759 * lto-streamer-out.c (create_output_block): Adjust.
20761 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
20763 * config/arm/arm.md (attribute "insn"): Move multiplication and
20764 division attributes to...
20765 (attribute "type"): ... here. Remove mult.
20766 (attribute "mul32"): New attribute.
20767 (attribute "mul64"): Add umaal.
20768 (*arm_mulsi3): Update attributes.
20769 (*arm_mulsi3_v6): Likewise.
20770 (*thumb_mulsi3): Likewise.
20771 (*thumb_mulsi3_v6): Likewise.
20772 (*mulsi3_compare0): Likewise.
20773 (*mulsi3_compare0_v6): Likewise.
20774 (*mulsi_compare0_scratch): Likewise.
20775 (*mulsi_compare0_scratch_v6): Likewise.
20776 (*mulsi3addsi): Likewise.
20777 (*mulsi3addsi_v6): Likewise.
20778 (*mulsi3addsi_compare0): Likewise.
20779 (*mulsi3addsi_compare0_v6): Likewise.
20780 (*mulsi3addsi_compare0_scratch): Likewise.
20781 (*mulsi3addsi_compare0_scratch_v6): Likewise.
20782 (*mulsi3subsi): Likewise.
20783 (*mulsidi3adddi): Likewise.
20784 (*mulsi3addsi_v6): Likewise.
20785 (*mulsidi3adddi_v6): Likewise.
20786 (*mulsidi3_nov6): Likewise.
20787 (*mulsidi3_v6): Likewise.
20788 (*umulsidi3_nov6): Likewise.
20789 (*umulsidi3_v6): Likewise.
20790 (*umulsidi3adddi): Likewise.
20791 (*umulsidi3adddi_v6): Likewise.
20792 (*smulsi3_highpart_nov6): Likewise.
20793 (*smulsi3_highpart_v6): Likewise.
20794 (*umulsi3_highpart_nov6): Likewise.
20795 (*umulsi3_highpart_v6): Likewise.
20796 (mulhisi3): Likewise.
20797 (*mulhisi3tb): Likewise.
20798 (*mulhisi3bt): Likewise.
20799 (*mulhisi3tt): Likewise.
20800 (maddhisi4): Likewise.
20801 (*maddhisi4tb): Likewise.
20802 (*maddhisi4tt): Likewise.
20803 (maddhidi4): Likewise.
20804 (*maddhidi4tb): Likewise.
20805 (*maddhidi4tt): Likewise.
20806 (divsi3): Likewise.
20807 (udivsi3): Likewise.
20808 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
20809 (thumb2_mulsi_short_compare0): Likewise.
20810 (thumb2_mulsi_short_compare0_scratch): Likewise.
20811 * config/arm/arm1020e.md (1020mult1): Update attribute change.
20812 (1020mult2): Likewise.
20813 (1020mult3): Likewise.
20814 (1020mult4): Likewise.
20815 (1020mult5): Likewise.
20816 (1020mult6): Likewise.
20817 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
20819 (cortex_a15_mult64): Likewise.
20820 (cortex_a15_sdiv): Likewise.
20821 (cortex_a15_udiv): Likewise.
20822 * config/arm/arm1026ejs.md (mult1): Update attribute change.
20828 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
20829 (pj4_ir_div): Likewise.
20830 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
20831 (11_mult2): Likewise.
20832 (11_mult3): Likewise.
20833 (11_mult4): Likewise.
20834 (11_mult5): Likewise.
20835 (11_mult6): Likewise.
20836 (11_mult7): Likewise.
20837 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
20838 (cortex_a8_mla): Likewise.
20839 (cortex_a8_mull): Likewise.
20840 (cortex_a8_smulwy): Likewise.
20841 (cortex_a8_smlald): Likewise.
20842 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
20843 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
20844 (cortex_r4_mul_3): Likewise.
20845 (cortex_r4_mla_4): Likewise.
20846 (cortex_r4_mla_3): Likewise.
20847 (cortex_r4_smlald): Likewise.
20848 (cortex_r4_mull): Likewise.
20849 (cortex_r4_sdiv): Likewise.
20850 (cortex_r4_udiv): Likewise.
20851 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
20852 (cortex_a7_idiv): Likewise.
20853 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
20854 (9_mult2): Likewise.
20855 (9_mult3): Likewise.
20856 (9_mult4): Likewise.
20857 (9_mult5): Likewise.
20858 (9_mult6): Likewise.
20859 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
20860 (cortex_a53_sdiv): Likewise.
20861 (cortex_a53_udiv): Likewise.
20862 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
20863 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
20864 (mp626_mult2): Likewise.
20865 (mp626_mult3): Likewise.
20866 (mp626_mult4): Likewise.
20867 * config/arm/fa526.md (526_mult1): Update attribute change.
20868 (526_mult2): Likewise.
20869 * config/arm/arm-generic.md (mult): Update attribute change.
20870 (mult_ldsched_strongarm): Likewise.
20871 (mult_ldsched): Likewise.
20872 (multi_cycle): Likewise.
20873 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
20874 * config/arm/fa606te.md (606te_mult1): Update attribute change.
20875 (606te_mult2): Likewise.
20876 (606te_mult3): Likewise.
20877 (606te_mult4): Likewise.
20878 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
20879 (cortex_a9_mac16): Likewise.
20880 (cortex_a9_multiply): Likewise.
20881 (cortex_a9_mac): Likewise.
20882 (cortex_a9_multiply_long): Likewise.
20883 * config/arm/fa626te.md (626te_mult1): Update attribute change.
20884 (626te_mult2): Likewise.
20885 (626te_mult3): Likewise.
20886 (626te_mult4): Likewise.
20888 2013-06-18 Richard Biener <rguenther@suse.de>
20891 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
20893 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20896 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
20897 with next_active_insn.
20899 2013-06-18 Alan Modra <amodra@gmail.com>
20901 * config/rs6000/rs6000.h (enum data_align): New.
20902 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
20903 (DATA_ABI_ALIGNMENT): Define.
20904 (CONSTANT_ALIGNMENT): Correct comment.
20905 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
20906 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
20908 2013-06-17 David Malcolm <dmalcolm@redhat.com>
20910 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
20911 ATTRIBUTE_UNUSED marking.
20913 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
20915 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
20916 alternative and update.
20917 (aarch64_dup_lanedi): Delete.
20918 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
20919 * config/aarch64/aarch64-simd-builtins.def: Update.
20921 2013-06-17 Richard Biener <rguenther@suse.de>
20923 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
20924 (lto_input_scc): Declare.
20925 (lto_input_tree_1): Likewise.
20926 (struct lto_stats_d): Add num_tree_bodies_output and
20927 num_pickle_refs_output.
20928 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
20929 (lto_read_tree_1): Split out from ...
20930 (lto_read_tree): ... this.
20931 (lto_input_scc): New function.
20932 (lto_input_tree_1): Split out from ...
20933 (lto_input_tree): ... this. Handle LTO_tree_scc.
20934 (lto_data_in_create): Create the streamer cache without hashes.
20935 * lto-streamer-out.c (create_output_block): Create the streamer
20936 cache with hashes when not doing WPA.
20937 (lto_write_tree_1): Split out from ...
20938 (lto_write_tree): ... this.
20939 (get_symbol_initial_value): New function.
20940 (lto_output_tree_1): Split out from ...
20941 (lto_output_tree): ... this. Write trees as series of SCCs
20942 using a DFS walk via DFS_write_tree.
20943 (struct sccs, struct scc_entry): New types.
20944 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
20945 (DFS_write_tree_body): New function.
20946 (DFS_write_tree): Likewise.
20947 (hash_tree): Likewise.
20948 (scc_entry_compare): Likewise.
20949 (hash_scc): Likewise.
20950 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
20951 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
20952 TREE_CHAIN as regular reference.
20953 (streamer_read_integer_cst): Remove.
20954 (streamer_get_pickled_tree): Adjust.
20955 * tree-streamer-out.c (streamer_write_chain): Disable streaming
20956 of DECL_EXTERNALs in BLOCK_VARS for now.
20957 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
20959 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
20960 Add hash value argument and record that if hashes are recorded
20962 (streamer_tree_cache_insert_1): Adjust.
20963 (streamer_tree_cache_insert): Likewise.
20964 (streamer_tree_cache_insert_at): Rename to ...
20965 (streamer_tree_cache_replace_tree): ... this and adjust.
20966 (streamer_tree_cache_append): Adjust.
20967 (record_common_node): Likewise.
20968 (streamer_tree_cache_create): Add argument whether to
20969 record hash values together with trees.
20970 (streamer_tree_cache_delete): Adjust.
20971 * tree-streamer.h (struct streamer_tree_cache_d): Add
20973 (streamer_read_integer_cst): Remove.
20974 (streamer_tree_cache_insert): Adjust.
20975 (streamer_tree_cache_append): Likewise.
20976 (streamer_tree_cache_insert_at): Rename to ...
20977 (streamer_tree_cache_replace_tree): ... this and adjust.
20978 (streamer_tree_cache_create): Add argument whether to record hashes.
20979 (streamer_tree_cache_get): Rename to ...
20980 (streamer_tree_cache_get_tree): ... this.
20981 (streamer_tree_cache_get_hash): New function.
20982 * tree.c (cache_integer_cst): New function.
20983 * tree.h (cache_integer_cst): Declare.
20984 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
20985 * lto-symtab.c (lto_varpool_replace_node): Only release
20986 DECL_INITIAL of non-prevailing decls.
20987 * varpool.c (varpool_remove_initializer): Do not release
20988 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
20990 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
20992 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
20993 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
20994 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
20995 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
20996 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
20997 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
20998 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
20999 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
21000 instead of TARGET_64BIT.
21001 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
21002 Require ISA_HAS_<D>DIV.
21004 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
21006 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
21007 (mips*-*-linux*): Move default with_llsc setting to where other
21009 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
21011 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
21012 Likewise. Remove default with_tune setting. Move default float
21013 setting to its own block. Handle with_llsc in the same block as above.
21015 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
21017 PR rtl-optimization/57425
21018 PR rtl-optimization/57569
21019 * alias.c (write_dependence_p): Add new parameters mem_mode,
21020 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
21021 Changed all callers.
21022 (canon_anti_dependence): New function.
21023 * cse.c (check_dependence): Use canon_anti_dependence.
21024 * cselib.c (cselib_invalidate_mem): Likewise.
21025 * rtl.h (canon_anti_dependence): Declare.
21027 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
21029 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
21030 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
21031 ".set mips3" for 64-bit targets.
21033 2013-06-15 Dehao Chen <dehao@google.com>
21035 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
21036 * gimple-low.c (gimple_check_call_matching_types): Likewise.
21037 (gimple_check_call_args): Likewise.
21038 * value-prof.c (check_ic_target): Likewise.
21039 * ipa-inline.c (early_inliner): Likewise.
21040 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
21041 * cgraph.c (cgraph_create_edge_1): Likewise.
21042 (cgraph_make_edge_direct): Likewise.
21044 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
21047 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
21048 rs6000_output_move_128bit to handle emitting quad memory
21049 operations. Set attribute length to 8 bytes.
21051 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
21053 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
21055 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
21056 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
21057 (aarch64_<su>mlsl<mode>): Likewise.
21059 2013-06-14 Mike Stump <mikestump@comcast.net>
21061 * Makefile.in (TARGET_H): Add insn-codes.h.
21063 2013-06-14 Alan Modra <amodra@gmail.com>
21065 PR middle-end/57134
21066 PR middle-end/57586
21067 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
21068 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
21069 bitfield expansion when EXPAND_MEMORY.
21070 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
21072 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
21074 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
21075 test for clearing quad memory on 32-bit later.
21077 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
21079 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
21080 (fold_negate_expr): Likewise.
21081 (fold_real_zero_addition_p): Handle vectors.
21082 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
21084 2013-06-14 Alan Modra <amodra@gmail.com>
21086 * varasm.c (force_const_mem): Revert 2013-06-07 change.
21088 2013-06-13 Jan Hubicka <jh@suse.cz>
21090 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
21091 Local comdats are not externally visible.
21092 * symtab.c (dump_symtab_base): Dump externally visible.
21093 (verify_symtab_base): Verify back links in the symtab hash.
21095 2013-06-13 Bin Cheng <bin.cheng@arm.com>
21097 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
21098 CONVERT_EXPR as equal nodes.
21100 2013-06-13 Bin Cheng <bin.cheng@arm.com>
21102 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
21104 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
21106 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
21107 Generalize to complex and vector.
21108 * tree.c (build_all_ones_cst): New function.
21109 * tree.h (build_all_ones_cst): Declare it.
21111 2013-06-13 Alan Modra <amodra@gmail.com>
21113 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
21114 * config/rs6000/rs6000.md (signbittf2): New insn.
21115 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
21116 (abstf2_internal, cmptf_internal2): Likewise.
21117 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
21119 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
21120 Pat Haugen <pthaugen@us.ibm.com>
21121 Peter Bergner <bergner@vnet.ibm.com>
21123 * config/rs6000/rs6000.c (emit_load_locked): Add support for
21124 power8 byte, half-word, and quad-word atomic instructions.
21125 (emit_store_conditional): Likewise.
21126 (rs6000_expand_atomic_compare_and_swap): Likewise.
21127 (rs6000_expand_atomic_op): Likewise.
21129 * config/rs6000/sync.md (larx): Add new modes for power8.
21131 (AINT): New mode iterator to include TImode as well as normal
21132 integer modes on power8.
21133 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
21134 that VSX registers are not considered. Use AINT mode iterator
21135 instead of INT1 to allow inclusion of quad word atomic operations
21137 (load_locked<mode>): Likewise.
21138 (store_conditional<mode>): Likewise.
21139 (atomic_compare_and_swap<mode>): Likewise.
21140 (atomic_exchange<mode>): Likewise.
21141 (atomic_nand<mode>): Likewise.
21142 (atomic_fetch_<fetchop_name><mode>): Likewise.
21143 (atomic_nand_fetch<mode>): Likewise.
21144 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
21146 (ATOMIC): On power8, add QImode, HImode modes.
21147 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
21148 modes that promote to SImode.
21149 (load_lockedti): Convert TImode arguments to PTImode, so that we
21150 get a guaranteed even/odd register pair.
21151 (load_lockedpti): Likewise.
21152 (store_conditionalti): Likewise.
21153 (store_conditionalpti): Likewise.
21155 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
21156 atomic load/store instructions.
21159 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
21161 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
21163 (insn_count): New attribute, with most cases extracted from...
21164 (length): ...here. Redefine most cases in terms of insn_count.
21165 (single_insn): Delete.
21166 (can_delay): Use insn_count to check for single instructions.
21167 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
21168 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
21169 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
21170 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
21171 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
21172 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
21173 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
21174 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
21175 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
21176 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
21177 rather than "length".
21178 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
21179 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
21180 Use "insn_count" rather than "length".
21181 * config/mips/mips-dsp.md
21182 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
21183 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
21186 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
21188 PR tree-optimization/57361
21189 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
21191 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
21193 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
21195 (aarch64_simd_combine<mode>): New instruction expansion.
21196 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
21197 function prototype.
21198 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
21199 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
21201 2013-06-12 Jan Hubicka <jh@suse.cz>
21203 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
21204 decl has when in streaming stage.
21205 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
21206 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
21208 2013-06-12 Roland Stigge <stigge@antcom.de>
21211 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
21213 2013-06-12 Jakub Jelinek <jakub@redhat.com>
21215 PR tree-optimization/57537
21216 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
21217 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
21219 2013-06-12 Richard Biener <rguenther@suse.de>
21221 * data-streamer.h (streamer_write_char_stream): CSE
21222 obs->current_pointer.
21223 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
21224 streamer_write_char_stream manually and optimize the resulting loop.
21225 (streamer_write_hwi_stream): Likewise.
21227 2013-06-12 Jan Hubicka <jh@suse.cz>
21229 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
21230 * cgraph.h (varpool_create_empty_node): Declare.
21231 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
21233 * symtab.c (symtab_unregister_node): Be lax about missin entries
21235 (symtab_get_node): Update comment.
21236 * varpool.c (varpool_create_empty_node): Break out from ...
21237 (varpool_node_for_decl): ... here.
21238 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
21240 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
21242 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
21243 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
21244 part. Use straight-line flow at the end.
21245 <COMPONENT_REF>: Remove superfluous else.
21246 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
21248 2013-06-12 Jakub Jelinek <jakub@redhat.com>
21251 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
21252 target hook even for !TREE_PUBLIC decls. If no resolution info
21253 is available, return false for common and external decls.
21255 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
21257 * config/rl78/constraints.md (U): New constraint.
21258 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
21261 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
21264 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
21265 to allow returning address to AT_PLATFORM name.
21267 2013-06-11 Jan Hubicka <jh@suse.cz>
21269 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
21270 * cgraph.h (symtab_node_base): Add weakref flag.
21271 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
21272 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
21273 (output_weakrefs): Use weakref flag.
21274 * fold-const.c (simple_operand_p): Handle WEAK.
21275 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
21276 * ipa.c (varpool_externally_visible_p): Drop weakref.
21277 (function_and_variable_visibility): Update comment; fix weakref
21278 sanity checks; do not clear DECL_WEAK on them.
21279 * lto-cgraph.c (lto_output_node): update.
21280 (lto_output_varpool_node): Update.
21281 (input_overwrite_node): Update.
21282 (input_node): Update.
21283 (input_varpool_node): Update.
21284 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
21285 (lto_symtab_merge_symbols): Add sanity check.
21286 (lto_symtab_prevailing_decl): Do not special case weakrefs.
21287 * passes.c (rest_of_decl_compilation): Set static flag, too.
21288 * symtab.c (dump_symtab_base): Dump weakref.
21289 (verify_symtab_base): Sanity check weakrefs.
21290 (symtab_make_decl_local): Remove duplicated code.
21291 (symtab_alias_ultimate_target): Simplify.
21292 * varpool.c (varpool_create_variable_alias): Set weakref flag.
21294 2013-06-11 Tom de Vries <tom@codesourcery.com>
21296 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
21297 sequence_vect == NULL.
21299 2013-06-11 DJ Delorie <dj@redhat.com>
21301 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
21302 (rl78_unwind_word_mode): New.
21304 2013-06-11 David Malcolm <dmalcolm@redhat.com>
21306 * final.c (debug_prefix_maps): Make static.
21308 2013-06-11 David Malcolm <dmalcolm@redhat.com>
21310 * function.c (initial_trampoline): Remove stray copy.
21312 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
21314 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
21316 2013-06-11 Martin Jambor <mjambor@suse.cz>
21318 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
21319 within bounds at the beginning of the function.
21321 2013-06-11 Alan Modra <amodra@gmail.com>
21323 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
21325 (get_named_section): Don't NULL !DECL_P decl.
21327 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
21329 * doc/invoke.texi (core-avx2): Document.
21331 (atom): Updated with MOVBE.
21333 2013-06-11 Richard Biener <rguenther@suse.de>
21335 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
21337 2013-06-11 Anton Blanchard <anton@samba.org>
21339 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
21340 correct shift value in little-endian mode.
21342 2013-06-11 Jakub Jelinek <jakub@redhat.com>
21345 * varasm.c (get_variable_align): Move #endif to the right place.
21347 2013-06-10 Cary Coutant <ccoutant@google.com>
21349 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
21350 for hash so that hash table traversal order is deterministic.
21352 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
21353 Pat Haugen <pthaugen@us.ibm.com>
21354 Peter Bergner <bergner@vnet.ibm.com>
21356 * config/rs6000/vector.md (GPR move splitter): Do not split moves
21357 of vectors in GPRS if they are direct moves or quad word load or
21360 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
21362 (direct_move_p): Likewise.
21363 (quad_load_store_p): Likewise.
21365 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
21366 classes into bins based on the physical register type.
21367 (reg_class_to_reg_type): Likewise.
21368 (IS_STD_REG_TYPE): Likewise.
21369 (IS_FP_VECT_REG_TYPE): Likewise.
21370 (reload_fpr_gpr): Arrays to determine what insn to use if we can
21371 use direct move instructions.
21372 (reload_gpr_vsx): Likewise.
21373 (reload_vsx_gpr): Likewise.
21374 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
21375 information that is a simplification of register classes. Also
21376 precalculate direct move reload helpers.
21377 (direct_move_p): New function to return true if the operation can
21378 be done as a direct move instruciton.
21379 (quad_load_store_p): New function to return true if the operation
21380 is a quad memory operation.
21381 (rs6000_legitimize_address): If quad memory, only allow register
21382 indirect for TImode addresses.
21383 (rs6000_legitimate_address_p): Likewise.
21384 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
21385 (rs6000_reload_register_type): Likewise.
21386 (register_to_reg_type): Return register type.
21387 (rs6000_secondary_reload_simple_move): New helper function for
21388 secondary reload and secondary memory needed to identify anything
21389 that is a simple move, and does not need reloading.
21390 (rs6000_secondary_reload_direct_move): New helper function for
21391 secondary reload to identify cases that can be done with several
21392 instructions via the direct move instructions.
21393 (rs6000_secondary_reload_move): New helper function for secondary
21394 reload to identify moves between register types that can be done.
21395 (rs6000_secondary_reload): Add support for quad memory operations
21396 and for direct move.
21397 (rs6000_secondary_memory_needed): Likewise.
21398 (rs6000_debug_secondary_memory_needed): Change argument names.
21399 (rs6000_output_move_128bit): New function to return the move to
21400 use for 128-bit moves, including knowing about the various
21401 limitations of quad memory operations.
21403 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
21404 memory operations. call rs6000_output_move_128bit for the actual
21405 instruciton(s) to generate.
21406 (vsx_movti_64bit): Likewise.
21408 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
21409 (UNSPEC_P8V_MTVSRWZ): Likewise.
21410 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
21411 (UNSPEC_P8V_MTVSRD): Likewise.
21412 (UNSPEC_P8V_XXPERMDI): Likewise.
21413 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
21414 (UNSPEC_FUSION_GPR): Likewise.
21415 (FMOVE128_GPR): New iterator for direct move.
21416 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
21417 (f32_sv): Likewise.
21418 (f32_dm): Likewise.
21419 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
21420 loads and direct move instructions.
21421 (zero_extendsidi2_lfiwzx): Likewise.
21422 (extendsidi2_lfiwax): Likewise.
21423 (extendsidi2_nocell): Likewise.
21424 (floatsi<mode>2_lfiwax): Likewise.
21425 (lfiwax): Likewise.
21426 (floatunssi<mode>2_lfiwzx): Likewise.
21427 (lfiwzx): Likewise.
21428 (fix_trunc<mode>_stfiwx): Likewise.
21429 (fixuns_trunc<mode>_stfiwx): Likewise.
21430 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
21431 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
21432 (parity<mode>2_cmpb): Set length/type attr.
21433 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
21434 for 'mr.' to fast_compare.
21435 (bpermd_<mode>): Change type attr to popcnt.
21436 (p8_fmrgow_<mode>): New insns for power8 direct move support.
21437 (p8_mtvsrwz_1): Likewise.
21438 (p8_mtvsrwz_2): Likewise.
21439 (reload_fpr_from_gpr<mode>): Likewise.
21440 (p8_mtvsrd_1): Likewise.
21441 (p8_mtvsrd_2): Likewise.
21442 (p8_xxpermdi_<mode>): Likewise.
21443 (reload_vsx_from_gpr<mode>): Likewise.
21444 (reload_vsx_from_gprsf): Likewise.
21445 (p8_mfvsrd_3_<mode>): LIkewise.
21446 (reload_gpr_from_vsx<mode>): Likewise.
21447 (reload_gpr_from_vsxsf): Likewise.
21448 (p8_mfvsrd_4_disf): Likewise.
21449 (multi-word GPR splits): Do not split direct moves or quad memory
21452 2013-06-10 David Malcolm <dmalcolm@redhat.com>
21454 * tree-into-ssa.c (interesting_blocks): Make static.
21456 2013-06-10 Jakub Jelinek <jakub@redhat.com>
21459 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
21460 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
21461 Use DATA_ABI_ALIGNMENT for that case instead if defined.
21462 (get_variable_align): New function.
21463 (get_variable_section, emit_bss, emit_common,
21464 assemble_variable_contents, place_block_symbol): Use
21465 get_variable_align instead of DECL_ALIGN.
21466 (assemble_noswitch_variable): Add align argument, use it
21467 instead of DECL_ALIGN.
21468 (assemble_variable): Adjust caller. Use get_variable_align
21469 instead of DECL_ALIGN.
21470 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
21472 (DATA_ABI_ALIGNMENT): Define.
21473 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
21474 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
21475 opt is false, only return the psABI mandated alignment increase.
21476 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
21477 (DATA_ABI_ALIGNMENT): ... this.
21478 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
21479 (DATA_ABI_ALIGNMENT): ... this.
21480 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
21481 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
21482 (DATA_ABI_ALIGNMENT): ... this.
21483 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
21484 * doc/tm.texi: Regenerated.
21486 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
21488 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
21489 cmp_code to construct REG_EQUAL note.
21491 2013-06-09 Jakub Jelinek <jakub@redhat.com>
21494 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
21495 that operands[2] doesn't overlap with operands[0].
21497 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
21498 Jan Hubicka <jh@suse.cz>
21500 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
21501 hack to mark symbols as used.
21503 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
21505 PR rtl-optimization/57559
21506 * lra-constraints.c (process_alt_operands): Don't discourage
21507 memory with known offset for offsetable memory constraint.
21508 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
21510 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
21512 * varasm.c (struct oc_local_state): Reorder fields.
21513 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
21514 and adjust accordingly.
21515 (output_constructor): Reorder initialization code and adjust call to
21516 output_constructor_bitfield.
21518 2013-06-07 Jan Hubicka <jh@suse.cz>
21520 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
21522 2013-06-07 David Malcolm <dmalcolm@redhat.com>
21524 * tree-object-size.c (unknown): Make const.
21526 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21528 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
21529 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
21530 for last alternative in the cpu_facility attribute.
21532 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21535 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
21536 (xordi3): Change operand 2 constraint to arm_xordi_operand.
21537 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
21538 * config/arm/constraints.md (Dg): New constraint.
21539 * config/arm/neon.md (xordi3_neon): Remove.
21540 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
21541 * config/arm/predicates.md (arm_xordi_operand): New predicate.
21543 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21545 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
21546 Clean up alternatives.
21548 2013-06-07 Alan Modra <amodra@gmail.com>
21550 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
21553 2013-06-07 Alan Modra <amodra@gmail.com>
21555 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
21557 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21559 * config/arm/constraints.md (Df): New constraint.
21560 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
21561 Correct length attribute for last two alternatives.
21563 2013-06-07 Alan Modra <amodra@gmail.com>
21565 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
21566 override user -mfp-in-toc.
21567 (offsettable_ok_by_alignment): Consider just the current access
21568 rather than the whole object, unless BLKmode. Handle
21569 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
21570 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
21571 for -mcmodel=medium.
21572 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
21573 override user -mfp-in-toc or -msum-in-toc. Default to
21574 -mno-fp-in-toc for -mcmodel=medium.
21576 2013-06-06 DJ Delorie <dj@redhat.com>
21578 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
21579 TARGET_VALID_POINTER_MODE.
21581 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
21582 Pat Haugen <pthaugen@us.ibm.com>
21583 Peter Bergner <bergner@vnet.ibm.com>
21585 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
21586 Document new power8 builtins.
21588 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
21589 condition code register, to allow 128-bit logical operations to be
21590 done in the VSX or GPR registers.
21591 (nor<mode>3): Use the canonical form for nor.
21592 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
21593 vclz*, and vpopcnt* vector instructions.
21594 (nand<mode>3): Likewise.
21595 (orc<mode>3): Likewise.
21596 (clz<mode>2): LIkewise.
21597 (popcount<mode>2): Likewise.
21599 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
21600 that only the GPRs are recognized.
21602 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
21603 support for new power8 builtins.
21605 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
21607 (xscvdpspn): Likewise.
21613 (vpopcnt): Likewise.
21614 (vpopcntb): Likewise.
21615 (vpopcnth): Likewise.
21616 (vpopcntw): Likewise.
21617 (vpopcntd): Likewise.
21619 (vmrgew): Likewise.
21620 (vmrgow): Likewise.
21622 (eqv_v16qi3): Likewise.
21623 (eqv_v8hi3): Likewise.
21624 (eqv_v4si3): Likewise.
21625 (eqv_v2di3): Likewise.
21626 (eqv_v4sf3): Likewise.
21627 (eqv_v2df3): Likewise.
21629 (nand_v16qi3): Likewise.
21630 (nand_v8hi3): Likewise.
21631 (nand_v4si3): Likewise.
21632 (nand_v2di3): Likewise.
21633 (nand_v4sf3): Likewise.
21634 (nand_v2df3): Likewise.
21636 (orc_v16qi3): Likewise.
21637 (orc_v8hi3): Likewise.
21638 (orc_v4si3): Likewise.
21639 (orc_v2di3): Likewise.
21640 (orc_v4sf3): Likewise.
21641 (orc_v2df3): Likewise.
21643 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
21644 allow power8 quad mode in 64-bit.
21645 (rs6000_builtin_vectorized_function): Add support to vectorize
21646 ISA 2.07 count leading zeros, population count builtins.
21647 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
21648 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
21649 (builtin_function_type): Add vgbbd builtin function which takes an
21651 (altivec_expand_vec_perm_const): Add support for new power8 merge
21654 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
21655 that does not include TImdoe for use with 32-bit.
21656 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
21658 (UNSPEC_VSX_CVDPSPN): Likewise.
21659 (vsx_xscvdpspn): Likewise.
21660 (vsx_xscvspdpn): Likewise.
21661 (vsx_xscvdpspn_scalar): Likewise.
21662 (vsx_xscvspdpn_directmove): Likewise.
21663 (vsx_and<mode>3): Split logical operations into 32-bit and
21664 64-bit. Add support to do logical operations on TImode as well as
21665 VSX vector types. Allow logical operations to be done in either
21666 VSX registers or in general purpose registers in 64-bit mode. Add
21667 splitters if GPRs were used. For AND, add clobber of CCmode to
21668 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
21670 (vsx_and<mode>3_32bit): Likewise.
21671 (vsx_and<mode>3_64bit): Likewise.
21672 (vsx_ior<mode>3): Likewise.
21673 (vsx_ior<mode>3_32bit): Likewise.
21674 (vsx_ior<mode>3_64bit): Likewise.
21675 (vsx_xor<mode>3): Likewise.
21676 (vsx_xor<mode>3_32bit): Likewise.
21677 (vsx_xor<mode>3_64bit): Likewise.
21678 (vsx_one_cmpl<mode>2): Likewise.
21679 (vsx_one_cmpl<mode>2_32bit): Likewise.
21680 (vsx_one_cmpl<mode>2_64bit): Likewise.
21681 (vsx_nor<mode>3): Likewise.
21682 (vsx_nor<mode>3_32bit): Likewise.
21683 (vsx_nor<mode>3_64bit): Likewise.
21684 (vsx_andc<mode>3): Likewise.
21685 (vsx_andc<mode>3_32bit): Likewise.
21686 (vsx_andc<mode>3_64bit): Likewise.
21687 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
21688 and xxlorc instructions.
21689 (vsx_eqv<mode>3_64bit): Likewise.
21690 (vsx_nand<mode>3_32bit): Likewise.
21691 (vsx_nand<mode>3_64bit): Likewise.
21692 (vsx_orc<mode>3_32bit): Likewise.
21693 (vsx_orc<mode>3_64bit): Likewise.
21695 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
21697 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
21699 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
21700 (p8_vmrgow): Likewise.
21701 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
21702 GPRs to be split under VSX.
21703 (p8v_clz<mode>2): Add power8 count leading zero support.
21704 (p8v_popcount<mode>2): Add power8 population count support.
21705 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
21708 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
21711 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
21713 (vec_nand): Likewise.
21714 (vec_vclz): Likewise.
21715 (vec_vclzb): Likewise.
21716 (vec_vclzd): Likewise.
21717 (vec_vclzh): Likewise.
21718 (vec_vclzw): Likewise.
21719 (vec_vgbbd): Likewise.
21720 (vec_vmrgew): Likewise.
21721 (vec_vmrgow): Likewise.
21722 (vec_vpopcnt): Likewise.
21723 (vec_vpopcntb): Likewise.
21724 (vec_vpopcntd): Likewise.
21725 (vec_vpopcnth): Likewise.
21726 (vec_vpopcntw): Likewise.
21728 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
21730 PR rtl-optimization/57468
21731 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
21734 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
21736 PR rtl-optimization/57459
21737 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
21738 type when setting live regs.
21740 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
21742 * config/s390/s390.opt (mlra): New option.
21743 * config/s390/s390.c (s390_decompose_address): Check displacement
21744 for all registers for LRA.
21745 (s390_secondary_reload): Don't used secondary reloads for LRA.
21746 (s390_lra_p): New function.
21747 (TARGET_LRA_P): Define.
21748 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
21749 of attribute cpu_facility to zarch for the last alternative.
21750 (*cmpmem_short): Ditto.
21752 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
21754 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
21755 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
21756 (arm_expand_prologue): Likewise.
21758 2013-06-06 Teresa Johnson <tejohnson@google.com>
21761 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
21762 as this is now done by redirect_edge_and_branch_force.
21763 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
21764 barriers, and fix interaction with splitting.
21765 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
21766 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
21767 reflect changes made in the routine.
21768 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
21769 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
21770 since this is called in cfglayout mode, and replace partition fixup
21771 with assert as that is now done by force_nonfallthru_and_redirect.
21772 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
21773 already be marked with region crossing note.
21774 (insert_section_boundary_note): Make non-static, gate on flag
21775 has_bb_partition, rewrite to also check for multiple partitions.
21776 (rest_of_handle_reorder_blocks): Remove call to
21777 insert_section_boundary_note, now done later during free_cfg.
21778 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
21779 * bb-reorder.h (insert_section_boundary_note): Declare.
21780 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
21781 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
21782 invoke insert_section_boundary_note.
21783 (try_redirect_by_replacing_jump): Remove unnecessary
21784 check for region crossing note.
21785 (fixup_partition_crossing): New function.
21786 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
21787 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
21788 in non-cfglayout mode.
21789 (force_nonfallthru_and_redirect): Fixup partition boundaries,
21790 remove old code that tried to do this. Emit barrier correctly
21791 when we are in cfglayout mode.
21792 (last_bb_in_partition): New function.
21793 (rtl_split_edge): Correctly fixup partition boundaries.
21794 (commit_one_edge_insertion): Remove old code that tried to
21795 fixup region crossing edge since this is now handled in
21796 split_block, and set up insertion point correctly since
21797 block may now end in a jump.
21798 (verify_hot_cold_block_grouping): Guard against checking when not in
21799 linearized RTL mode.
21800 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
21802 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
21803 rtl_verify_flow_info, so not called in cfglayout mode.
21804 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
21805 (fixup_reorder_chain): Remove old code that attempted to fixup region
21806 crossing note as this is now handled in force_nonfallthru_and_redirect.
21807 (duplicate_insn_chain): Don't duplicate switch section notes.
21808 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
21810 * basic-block.h (emit_barrier_after_bb): Declare.
21812 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21814 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
21815 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
21816 arm_usatsihi): Adjust alternatives for arm_restrict_it.
21818 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21820 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
21822 * config/arm/ldmstm.md: Regenerate.
21824 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21826 * config/arm/sync.md (atomic_loaddi_1):
21827 Disable predication for arm_restrict_it.
21828 (arm_load_exclusive<mode>): Likewise.
21829 (arm_load_exclusivesi): Likewise.
21830 (arm_load_exclusivedi): Likewise.
21831 (arm_load_acquire_exclusive<mode>): Likewise.
21832 (arm_load_acquire_exclusivesi): Likewise.
21833 (arm_load_acquire_exclusivedi): Likewise.
21834 (arm_store_exclusive<mode>): Likewise.
21835 (arm_store_exclusive<mode>): Likewise.
21836 (arm_store_release_exclusivedi): Likewise.
21837 (arm_store_release_exclusive<mode>): Likewise.
21839 2013-06-06 Richard Biener <rguenther@suse.de>
21841 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
21843 (lto_tag_is_tree_code_p): Adjust.
21844 (lto_tag_is_gimple_code_p): Likewise.
21845 (lto_gimple_code_to_tag): Likewise.
21846 (lto_tag_to_gimple_code): Likewise.
21847 (lto_tree_code_to_tag): Likewise.
21848 (lto_tag_to_tree_code): Likewise.
21849 * data-streamer.h (streamer_write_hwi_in_range): Use
21850 uhwi streaming to stream the normalized range.
21851 (streamer_read_hwi_in_range): Likewise.
21853 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21855 * config/arm/arm.md (enabled_for_depr_it): New attribute.
21856 (predicable_short_it): Likewise.
21857 (predicated): Likewise.
21858 (enabled): Handle above.
21859 (define_cond_exec): Set predicated attribute to yes.
21861 2013-06-05 Mike Stump <mikestump@comcast.net>
21863 * gdbinit.in (__FUNCTION__): Add.
21865 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
21867 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
21868 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
21870 2013-06-05 Jan Hubicka <jh@suse.cz>
21872 * varasm.c (mark_decl_referenced): Revert the removal until targets
21875 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
21877 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
21878 instead of mark_decl_referenced.
21880 2013-06-05 Jan Hubicka <jh@suse.cz>
21882 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
21883 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
21884 and symtab_used_from_object_file_p.
21885 (cgraph_make_node_local_1): Clear forced_by_abi.
21886 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
21887 * cgraph.h (symtab_node_base): Add forced_by_abi.
21888 (decide_is_variable_needed): Remove.
21889 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
21890 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
21891 (decide_is_symbol_needed): ... this one; handle symbols in general;
21892 always analyze virtuals; honnor forced_by_abi.
21893 (cgraph_finalize_function): Update.
21894 (varpool_finalize_decl): Update.
21895 (symbol_defined_and_needed): Remove.
21896 (analyze_functions): Update.
21897 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
21898 output_refs, input_overwrite_node): Handle forced_by_abi.
21899 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
21900 (address_taken_from_non_vtable_p): ... this one.
21901 (comdat_can_be_unshared_p_1): New function.
21902 (cgraph_comdat_can_be_unshared_p): Rename to ...
21903 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
21904 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
21905 (function_and_variable_visibility): Clear forced_by_abi as needed.
21906 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
21907 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
21908 * symtab.c (dump_symtab_base): Dump forced_by_abi.
21909 * varpool.c (decide_is_variable_needed): Remove.
21911 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21913 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
21914 (arm_option_override): Override arm_restrict_it where appropriate.
21915 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
21916 * config/arm/arm.opt (mrestrict-it): New command-line option.
21917 * doc/invoke.texi: Document -mrestrict-it.
21919 2013-06-05 David Malcolm <dmalcolm@redhat.com>
21921 * tsan.c (tsan_atomic_table): Make const.
21923 2013-06-05 Richard Biener <rguenther@suse.de>
21925 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
21926 index associated with the tree we are supposed to replace.
21927 * tree-streamer-out.c (pack_ts_base_value_fields): Output
21928 TREE_ASM_WRITTEN as zero for everything but SSA names.
21930 2013-06-05 David Malcolm <dmalcolm@redhat.com>
21932 * tree-ssa-structalias.c (call_stmt_vars): Make static.
21934 2013-06-04 Jan Hubicka <jh@suse.cz>
21936 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
21937 (input_node, input_varpool_node): Handle correctly external same
21939 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
21940 nodes at ltrans stage.
21942 2013-06-04 Jan Hubicka <jh@suse.cz>
21944 * ipa-inline.c (update_caller_keys): Fix availability test.
21945 (update_callee_keys): Likewise.
21946 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
21947 to follow ELF standard.
21949 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
21951 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
21952 (mips64r5900el-*-elf*): New configurations.
21953 * config/mips/mips-cpus.def (r5900): New processor.
21954 * config/mips/mips-tables.opt: Regenerate.
21955 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
21956 (mips_issue_rate): Handle PROCESSOR_R5900.
21957 (mips_reorg_process_insns): Force reorder mode for the R5900.
21958 * config/mips/mips.h (TARGET_MIPS5900): Define.
21959 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
21961 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
21963 * config/mips/mips.md (processor): Add r5900.
21964 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
21966 2013-06-04 Ian Bolton <ian.bolton@arm.com>
21968 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
21969 into function to generate MOVI instruction.
21970 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
21971 (aarch64_preferred_simd_mode): Turn into wrapper.
21972 (aarch64_output_scalar_simd_mov_immediate): New function.
21973 * config/aarch64/aarch64-protos.h: Add prototype for above.
21975 2013-06-04 Ian Bolton <ian.bolton@arm.com>
21977 * config/aarch64/aarch64.c (simd_immediate_info): Remove
21978 element_char member.
21979 (sizetochar): Return signed char.
21980 (aarch64_simd_valid_immediate): Remove elchar and other
21981 unnecessary variables.
21982 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
21983 Calculate element_char as required.
21984 * config/aarch64/aarch64-protos.h: Update and move prototype
21985 for aarch64_output_simd_mov_immediate.
21986 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
21989 2013-06-04 Ian Bolton <ian.bolton@arm.com>
21991 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
21992 information completed by aarch64_simd_valid_immediate.
21993 (aarch64_legitimate_constant_p): Update arguments.
21994 (aarch64_simd_valid_immediate): Work with struct rather than many
21996 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
21997 (aarch64_simd_make_constant): Update arguments.
21998 (aarch64_output_simd_mov_immediate): Work with struct rather than
21999 many pointers. Output immediate directly rather than as operand.
22000 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
22002 * config/aarch64/constraints.md (Dn): Update arguments.
22004 2013-06-04 Ian Bolton <ian.bolton@arm.com>
22006 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
22008 (aarch64_simd_immediate_valid_for_move): Remove.
22009 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
22010 (aarch64_simd_make_constant): Update call.
22011 (aarch64_output_simd_mov_immediate): Update call.
22012 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
22014 * config/aarch64/constraints.md (Dn): Update call.
22016 2013-06-04 Ian Bolton <ian.bolton@arm.com>
22018 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
22019 return type to bool for prototype.
22020 (aarch64_legitimate_constant_p): Check for true instead of not -1.
22021 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
22022 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
22024 2013-06-04 Catherine Moore <clm@codesourcery.com>
22026 * config/mips/mips.opt (meva): New.
22027 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
22028 (ASM_SPEC): Handle -meva.
22029 * doc/invoke.texi (meva): Document.
22031 2013-06-04 Alan Modra <amodra@gmail.com>
22033 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
22036 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22038 * rtl.def: Add extra fourth optional field to define_cond_exec.
22039 * gensupport.c (process_one_cond_exec): Process attributes from
22041 * doc/md.texi: Document fourth field in define_cond_exec.
22043 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
22045 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
22046 out the processing order as in store_bit_field_1.
22048 2013-06-04 Jan Hubicka <jh@suse.cz>
22050 PR middle-end/57500
22051 * cgraphunit.c (cgraph_process_same_body_aliases): Create
22052 non-VAR_DECL node if it does not exist yet.
22054 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
22056 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
22057 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
22058 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
22059 target_cpu_default setting.
22061 2013-06-03 Teresa Johnson <tejohnson@google.com>
22063 * dumpfile.c (opt_info_switch_p): Change -fopt-info
22064 default to -fopt-info=optimized instead of all.
22065 * doc/invoke.texi: Ditto.
22066 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
22067 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
22068 (execute_vect_slp): Emit BB vectorization success under
22069 MSG_OPTIMIZED_LOCATIONS.
22070 * tree-vect-slp.c (vect_slp_transform_bb): Change
22071 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
22072 * tree-vect-loop.c (vect_transform_loop): Ditto.
22074 2013-06-03 Jason Merrill <jason@redhat.com>
22077 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
22078 Use TARGET_EXPR for C++.
22080 2013-06-03 Jakub Jelinek <jakub@redhat.com>
22082 PR rtl-optimization/57268
22083 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
22084 if DEBUG_INSN_P (insn).
22087 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
22089 PR rtl-optimization/57268
22090 * sched-deps.c (sched_analyze_2): Flush dependence lists if
22091 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
22093 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
22095 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
22096 (ix86_avoid_lea_for_addr): Likewise.
22097 (exact_dependency_1): Likewise.
22098 (ix86_adjust_cost): Likewise.
22099 (swap_top_of_ready_list): Fix formatting and !reload_completed check
22101 (do_reorder_for_imul): Fix typo, formatting and
22102 !reload_completed check removed.
22103 (ix86_sched_reorder): Fix typo and formatting.
22104 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
22107 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
22109 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
22111 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
22113 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
22114 <STRING_CST>: Likewise.
22115 <VECTOR_CST>: Likewise.
22117 2013-06-01 Janus Weil <janus@gcc.gnu.org>
22118 Mikael Morin <mikael@gcc.gnu.org>
22120 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
22121 * config.in: Regenerated.
22122 * configure: Regenerated.
22124 2013-06-01 Jan Hubicka <jh@suse.cz>
22126 PR middle-end/57366
22127 * cgraphunit.c (compile): When weakref is not supported,
22128 set up transparent aliases before final output pass.
22129 * varasm.c (assemble_alias): Do not try to do it here.
22131 2013-06-01 Jan Hubicka <jh@suse.cz>
22133 PR middle-end/57467
22134 * passes.c (for_per_function): Skip unanalyzed functions.
22136 2013-06-01 Jan Hubicka <jh@suse.cz>
22138 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
22139 (lto_symtab_merge_symbols_1): ... this one.
22140 (lto_symtab_merge_cgraph_nodes): Rename to ...
22141 (lto_symtab_merge_symbols): ... this one; simplify.
22142 * cgraph.c (same_body_aliases_done): Rename to ...
22143 (cpp_implicit_aliases_done): ... this one.
22144 (cgraph_create_function_alias): Update.
22145 (cgraph_same_body_alias): Update.
22146 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
22147 (verify_edge_corresponds_to_fndecl): Simplify.
22148 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
22149 (cgraph_node): Remove same_body_alias.
22150 (varpool_node): Remove alias_of and extra_name_alias.
22151 (same_body_aliases_done): Rename to ..
22152 (cpp_implicit_aliases_done): ... this one.
22153 (symtab_alias_ultimate_target): Add default parameter.
22154 (symtab_resolve_alias): New function.
22155 (fixup_same_cpp_alias_visibility): Declare.
22156 (cgraph_function_node): Add default parameter.
22157 (cgraph_node_asm_name): Likewise.
22158 (cgraph_function_or_thunk_node): Add default parameter; do
22159 not ICE when it is NULL.
22160 (varpool_variable_node): Likewise.
22161 * tree-emutls.c (create_emultls_var): Update.
22162 (ipa_lower_emutls): Update.
22163 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
22164 (cgraph_reset_node): Reset alias info.
22165 (cgraph_finalize_function): Update.
22166 (fixup_same_cpp_alias_visibility): Move to symtab.c.
22167 (analyze_function): Simplify.
22168 (cgraph_process_same_body_aliases): Simplify.
22169 (analyze_functions): Fixup same body aliases.
22170 (handle_alias_pairs): Simplify.
22171 (assemble_thunk): Update.
22172 (assemble_thunks_and_aliases): Update.
22173 (output_weakrefs): Rewrite.
22174 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
22175 (lto_output_varpool_node): Likewise.
22176 (compute_ltrans_boundary): Remve assert.
22177 (get_alias_symbol): New functoin.
22178 (input_node): Rewrite alias handling.
22179 (input_varpool_node): Likewise.
22180 * ipa-pure-const.c (propagate_pure_const): Fix formating.
22181 * ipa.c (process_references): Handle weakrefs correctly.
22182 (symtab_remove_unreachable_nodes): Likewise.
22183 * trans-mem.c (get_cg_data): Update.
22184 (ipa_tm_create_version_alias): Update.
22185 (ipa_tm_execute): Update.
22186 * symtab.c (dump_symtab_base): Dump aliases.
22187 (verify_symtab_base): Verify aliases.
22188 (symtab_node_availability): New function.
22189 (symtab_alias_ultimate_target): Simplify.
22190 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
22191 handle all the fixup cases.
22192 (symtab_resolve_alias): New function.
22193 * passes.c (ipa_write_summaries): Handle weakrefs.
22194 * varpool.c (varpool_analyze_node): Simplify.
22195 (assemble_aliases): Update.
22196 (varpool_create_variable_alias): Simplify.
22197 (varpool_extra_name_alias): Simplify.
22198 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
22199 (lto_symtab_merge_symbols): ... this one.
22201 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
22204 PR rtl-optimization/57268
22205 * sched-deps.c (sched_analyze_2): Flush dependence lists if
22206 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
22208 2013-06-01 Tobias Burnus <burnus@net-b.de>
22210 Partially reverted:
22211 2013-05-31 Tobias Burnus <burnus@net-b.de>
22213 PR middle-end/57073
22214 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
22217 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
22219 PR rtl-optimization/57268
22220 * sched-deps.c (sched_analyze_2): Flush dependence lists if
22221 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
22223 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
22225 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
22226 unordered comparison operators when -fno-trapping-math is in effect
22228 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
22229 and implement unordered comparison operators properly on the e500.
22231 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
22233 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
22234 for constant scalar integers.
22235 (simplify_relational_operation_1): Likewise.
22237 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
22239 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
22240 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
22243 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
22244 Igor Zamyatin <igor.zamyatin@intel.com>
22246 Silvermont (SLM) architecture performance tuning.
22247 * config/i386/i386.h (enum ix86_tune_indices): Add
22248 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
22249 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
22251 * config/i386/i386.c (initial_ix86_tune_features)
22252 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
22253 (ix86_lea_outperforms): Handle Silvermont tuning.
22254 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
22256 (ix86_use_lea_for_mov): Likewise.
22257 (ix86_avoid_lea_for_addr): Likewise.
22258 (ix86_lea_for_add_ok): Likewise.
22259 (exact_dependency_1): New function.
22260 (exact_store_load_dependency): Likewise.
22261 (ix86_adjust_cost): Handle Silvermont tuning.
22262 (do_reoder_for_imul): Likewise.
22263 (swap_top_of_ready_list): New function.
22264 (ix86_sched_reorder): Changed to handle Silvermont tuning.
22266 * config/i386/i386.md (peepholes that split memory operand in fp
22269 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
22271 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
22272 Remove un-necessary braces.
22274 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
22276 * config/aarch64/aarch64.c (aarch64_classify_symbol):
22277 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
22279 2013-05-31 Tobias Burnus <burnus@net-b.de>
22281 PR middle-end/57073
22282 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
22284 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22287 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
22288 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
22289 * config/arm/neon.md (iordi3_neon): Remove.
22290 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
22291 * config/arm/predicates.md (imm_for_neon_logic_operand):
22292 Move to earlier in the file.
22293 (neon_logic_op2): Likewise.
22294 (arm_iordi_operand_neon): New predicate.
22296 2013-05-31 Richard Biener <rguenther@suse.de>
22298 PR tree-optimization/57478
22299 PR tree-optimization/57453
22300 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
22303 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
22305 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
22306 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
22308 2013-05-30 Tobias Burnus <burnus@net-b.de>
22309 Thomas Koenig <tkoenig@gcc.gnu.org>
22311 PR middle-end/57073
22312 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
22313 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
22315 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
22317 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
22319 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
22321 * target.def (register_usage_leveling_p): New hook.
22322 * targhooks.c (default_register_usage_leveling_p): New.
22323 * targhooks.h (default_register_usage_leveling_p): New prototype.
22324 * lra-assigns.c (register_usage_leveling_p): Use the hook.
22325 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
22326 * doc/tm.texi: Update.
22327 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
22329 2013-05-30 Ian Bolton <ian.bolton@arm.com>
22331 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
22332 (*insv_reg<mode>): New define_insn.
22334 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
22336 PR rtl-optimization/57439
22337 * postreload.c (move2add_valid_value_p): Check that we have
22338 a zero subreg_regno_offset when accessing the register in
22339 the requested mode.
22341 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
22342 Igor Zamyatin <igor.zamyatin@intel.com>
22344 Silvermont (SLM) architecture pipeline model, tuning and
22346 * config.gcc: Add slm config options and target.
22348 * config/i386/slm.md: New.
22350 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
22352 * config/i386/i386-c.c (ix86_target_macros_internal): New case
22354 (ix86_target_macros_internal): Likewise.
22356 * config/i386/i386.c (slm_cost): New cost.
22357 (m_SLM): New macro flag.
22358 (initial_ix86_tune_features): Set m_SLM.
22359 (x86_accumulate_outgoing_args): Likewise.
22360 (x86_arch_always_fancy_math_387): Likewise.
22361 (processor_target_table): Add slm cost.
22362 (cpu_names): Add slm cpu name.
22363 (x86_option_override_internal): Set SLM ISA.
22364 (ix86_issue_rate): New case PROCESSOR_SLM.
22365 (ia32_multipass_dfa_lookahead): Likewise.
22366 (fold_builtin_cpu): Add slm.
22368 * config/i386/i386.h (TARGET_SLM): New target macro.
22369 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
22370 (processor_type): Add PROCESSOR_SLM.
22372 * config/i386/i386.md (cpu): Add new value "slm".
22373 (slm.md): Include slm.md.
22375 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
22376 Zhenqiang Chen <zhenqiang.chen@linaro.org>
22378 * config/arm/arm-protos.h: Add and update function protos.
22379 * config/arm/arm.c (use_simple_return_p): New added.
22380 (thumb2_expand_return): Check simple_return flag.
22381 * config/arm/arm.md: Add simple_return and conditional simple_return.
22382 * config/arm/iterators.md: Add iterator for return and simple_return.
22384 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
22386 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
22387 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
22388 (arm_emit_vfp_multi_reg_pop): Likewise.
22389 (thumb2_emit_ldrd_pop): Likewise.
22390 (arm_expand_epilogue): Add misc REG_CFA notes.
22391 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
22393 2013-05-29 Lawrence Crowl <crowl@google.com>
22395 * config/arm/t-arm: Update for below.
22397 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
22398 Change type to hash_table. Update dependent calls and types.
22400 * config/i386/t-cygming: Update for below.
22402 * config/i386/t-interix: Update for below.
22404 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
22405 Change type to hash_table. Update dependent calls and types.
22406 (i386_find_on_wrapper_list::wrappers): Likewise.
22408 * config/ia64/t-ia64: Update for below.
22410 * config/ia64/ia64.c (bundle_state_table):
22411 Change type to hash_table. Update dependent calls and types.
22413 * config/mips/mips.c (mips_reorg_process_insns::htab):
22414 Change type to hash_table. Update dependent calls and types.
22416 * config/sol2.c (solaris_comdat_htab):
22417 Change type to hash_table. Update dependent calls and types.
22419 * config/t-sol2: Update for above.
22421 2013-05-29 Teresa Johnson <tejohnson@google.com>
22423 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
22424 functions are not yet marked as defined.
22426 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
22427 Pat Haugen <pthaugen@us.ibm.com>
22428 Peter Bergner <bergner@vnet.ibm.com>
22430 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
22434 (vrotl<mode>3): Likewise.
22435 (vashl<mode>3): Likewise.
22436 (vlshr<mode>3): Likewise.
22437 (vashr<mode>3): Likewise.
22439 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
22440 support for power8 V2DI builtins.
22442 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
22443 power8 V2DI builtins.
22444 (vupkhsw): Likewise.
22445 (vupklsw): Likewise.
22446 (vaddudm): Likewise.
22447 (vminsd): Likewise.
22448 (vmaxsd): Likewise.
22449 (vminud): Likewise.
22450 (vmaxud): Likewise.
22451 (vpkudum): Likewise.
22452 (vpksdss): Likewise.
22453 (vpkudus): Likewise.
22454 (vpksdus): Likewise.
22459 (vsubudm): Likewise.
22460 (vcmpequd): Likewise.
22461 (vcmpgtsd): Likewise.
22462 (vcmpgtud): Likewise.
22463 (vcmpequd_p): Likewise.
22464 (vcmpgtsd_p): Likewise.
22465 (vcmpgtud_p): Likewise.
22466 (vupkhsw): Likewise.
22467 (vupklsw): Likewise.
22468 (vaddudm): Likewise.
22469 (vmaxsd): Likewise.
22470 (vmaxud): Likewise.
22471 (vminsd): Likewise.
22472 (vminud): Likewise.
22473 (vpksdss): Likewise.
22474 (vpksdus): Likewise.
22475 (vpkudum): Likewise.
22476 (vpkudus): Likewise.
22481 (vsubudm): Likewise.
22483 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
22484 support for power8 V2DI instructions.
22486 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
22487 power8 V2DI instructions. Combine pack and unpack insns to use an
22488 iterator for each mode. Check whether a particular mode supports
22489 Altivec instructions instead of just checking TARGET_ALTIVEC.
22490 (UNSPEC_VPKUWUM): Likewise.
22491 (UNSPEC_VPKSHSS): Likewise.
22492 (UNSPEC_VPKSWSS): Likewise.
22493 (UNSPEC_VPKUHUS): Likewise.
22494 (UNSPEC_VPKSHUS): Likewise.
22495 (UNSPEC_VPKUWUS): Likewise.
22496 (UNSPEC_VPKSWUS): Likewise.
22497 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
22498 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
22499 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
22500 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
22501 (UNSPEC_VUPKHSB): Likewise.
22502 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
22503 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
22504 (UNSPEC_VUPKHSH): Likewise.
22505 (UNSPEC_VUPKLSB): Likewise.
22506 (UNSPEC_VUPKLSH): Likewise.
22508 (VI_char): Likewise.
22509 (VI_scalar): Likewise.
22510 (VI_unit): Likewise.
22512 (VP_small): Likewise.
22513 (VP_small_lc): Likewise.
22514 (VU_char): Likewise.
22515 (add<mode>3): Likewise.
22516 (altivec_vaddcuw): Likewise.
22517 (altivec_vaddu<VI_char>s): Likewise.
22518 (altivec_vadds<VI_char>s): Likewise.
22519 (sub<mode>3): Likewise.
22520 (altivec_vsubcuw): Likewise.
22521 (altivec_vsubu<VI_char>s): Likewise.
22522 (altivec_vsubs<VI_char>s): Likewise.
22523 (altivec_vavgs<VI_char>): Likewise.
22524 (altivec_vcmpbfp): Likewise.
22525 (altivec_eq<mode>): Likewise.
22526 (altivec_gt<mode>): Likewise.
22527 (altivec_gtu<mode>): Likewise.
22528 (umax<mode>3): Likewise.
22529 (smax<mode>3): Likewise.
22530 (umin<mode>3): Likewise.
22531 (smin<mode>3): Likewise.
22532 (altivec_vpkuhum): Likewise.
22533 (altivec_vpkuwum): Likewise.
22534 (altivec_vpkshss): Likewise.
22535 (altivec_vpkswss): Likewise.
22536 (altivec_vpkuhus): Likewise.
22537 (altivec_vpkshus): Likewise.
22538 (altivec_vpkuwus): Likewise.
22539 (altivec_vpkswus): Likewise.
22540 (altivec_vpks<VI_char>ss): Likewise.
22541 (altivec_vpks<VI_char>us): Likewise.
22542 (altivec_vpku<VI_char>us): Likewise.
22543 (altivec_vpku<VI_char>um): Likewise.
22544 (altivec_vrl<VI_char>): Likewise.
22545 (altivec_vsl<VI_char>): Likewise.
22546 (altivec_vsr<VI_char>): Likewise.
22547 (altivec_vsra<VI_char>): Likewise.
22548 (altivec_vsldoi_<mode>): Likewise.
22549 (altivec_vupkhsb): Likewise.
22550 (altivec_vupkhs<VU_char>): Likewise.
22551 (altivec_vupkls<VU_char>): Likewise.
22552 (altivec_vupkhsh): Likewise.
22553 (altivec_vupklsb): Likewise.
22554 (altivec_vupklsh): Likewise.
22555 (altivec_vcmpequ<VI_char>_p): Likewise.
22556 (altivec_vcmpgts<VI_char>_p): Likewise.
22557 (altivec_vcmpgtu<VI_char>_p): Likewise.
22558 (abs<mode>2): Likewise.
22559 (vec_unpacks_hi_v16qi): Likewise.
22560 (vec_unpacks_hi_v8hi): Likewise.
22561 (vec_unpacks_lo_v16qi): Likewise.
22562 (vec_unpacks_hi_<VP_small_lc>): Likewise.
22563 (vec_unpacks_lo_v8hi): Likewise.
22564 (vec_unpacks_lo_<VP_small_lc>): Likewise.
22565 (vec_pack_trunc_v8h): Likewise.
22566 (vec_pack_trunc_v4si): Likewise.
22567 (vec_pack_trunc_<mode>): Likewise.
22569 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
22571 (vec_vmaxsd): Likewise.
22572 (vec_vmaxud): Likewise.
22573 (vec_vminsd): Likewise.
22574 (vec_vminud): Likewise.
22575 (vec_vpksdss): Likewise.
22576 (vec_vpksdus): Likewise.
22577 (vec_vpkudum): Likewise.
22578 (vec_vpkudus): Likewise.
22579 (vec_vrld): Likewise.
22580 (vec_vsld): Likewise.
22581 (vec_vsrad): Likewise.
22582 (vec_vsrd): Likewise.
22583 (vec_vsubudm): Likewise.
22584 (vec_vupkhsw): Likewise.
22585 (vec_vupklsw): Likewise.
22587 2013-05-29 Jan Hubicka <jh@suse.cz>
22589 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
22590 flags; reorder rest of fields in more consistent way.
22591 (varpool_node): Remove analyzed, finalized and alias.
22592 (cgraph_ndoe): Likewise.
22593 (symtab_alias_ultimate_target): New function.
22594 (cgraph_function_node): Move offline.
22595 (cgraph_reset_node): Declare.
22596 (cgraph_comdat_can_be_unshared_p): Remove.
22597 (varpool_remove_initializer): Declare.
22598 (varpool_first_defined_variable, varpool_next_defined_variable
22599 cgraph_first_defined_function, cgraph_next_defined_function): Update.
22600 (cgraph_function_with_gimple_body_p): Update.
22601 (varpool_all_refs_explicit_p): Update.
22602 (symtab_alias_target): New function.
22603 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
22604 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
22605 (cgraph_function_or_thunk_node): Simplify using
22606 symtab_alias_ultimate_target.
22607 (varpool_variable_node): Likewise.
22608 * cgraph.c (cgraph_create_function_alias): Update.
22609 (cgraph_add_thunk): Update.
22610 (cgraph_remove_node): Update.
22611 (dump_cgraph_node): Do not dump removed flags.
22612 (cgraph_function_body_availability): Update.
22613 (cgraph_propagate_frequency): Update.
22614 (verify_cgraph_node): Check sanity of local flag.
22615 (cgraph_function_node): Move here from cgraph.h; revamp for
22616 cgraph_function_or_thunk_node.
22617 * lto-symtab.c (lto_varpool_replace_node): Update.
22618 (lto_symtab_resolve_can_prevail_p): Update.
22619 (lto_symtab_merge_cgraph_nodes): Update.
22620 * ipa-cp.c (determine_versionability, initialize_node_lattices,
22621 propagate_constants_accross_call, devirtualization_time_bonus,
22622 ipcp_propagate_stage): Update.
22623 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
22624 * ipa-inline-transform.c (clone_inlined_nodes,
22625 preserve_function_body_p): Update.
22626 * ipa-reference.c (propagate): Update.
22627 (write_node_summary_p): Update.
22628 * toplev.c (wrapup_global_declaration_2): Update.
22629 * cgraphunit.c (cgraph_analyze_function): Rename to ...
22630 (analyze_function) ... this one.
22631 (cgraph_process_new_functions): Update.
22632 (cgraph_reset_node): Export.
22633 (cgraph_finalize_function): Update.
22634 (cgraph_add_new_function): Update.
22635 (process_function_and_variable_attributes): Update.
22636 (varpool_finalize_decl): Update.
22637 (symbol_finalized): Remove.
22638 (symbol_finalized_and_needed): Rename to ...
22639 (symbol_defined_and_needed): ... update.
22640 (cgraph_analyze_functions): Update.
22641 (handle_alias_pairs): Update.
22642 (mark_functions_to_output): Update.
22643 (assemble_thunk): Update.
22644 (output_in_order): Update.
22645 (output_weakrefs): Update.
22646 (finalize_compilation_unit): Update.
22647 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
22648 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
22649 input_node, input_varpool_node): Update.
22650 * dbxout.c (dbxout_expand_expr): Update.
22651 * cgraphclones.c (cgraph_clone_node): Update.
22652 (cgraph_copy_node_for_versioning): Update.
22653 (cgraph_materialize_clone): Update.
22654 (cgraph_materialize_all_clones): Update.
22655 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
22656 propagate_pure_const, propagate_nothrow): Update.
22657 * lto-streamer-out.c (lto_output, write_symbol): Update.
22658 * ipa-utils.c (ipa_reverse_postorder): Update.
22659 * ipa-inline.c (can_inline_edge_p): Update.
22660 (update_caller_keys, ipa_inline): Update.
22661 * dwarf2out.c (reference_to_unused,
22662 premark_types_used_by_global_vars_helper): Update.
22663 * tree-eh.c (tree_could_trap_p): Update.
22664 * ipa-split.c (consider_split, execute_split_functions): Update.
22665 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
22666 has_addr_references_p): Update; move ahead in file for better
22668 (process_references): Simplify.
22669 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
22670 bodies are removed.
22671 (cgraph_comdat_can_be_unshared_p): Make static.
22672 (cgraph_externally_visible_p): Update.
22673 (varpool_externally_visible_p): Update.
22674 (function_and_variable_visibility): Update.
22675 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
22676 ipa_tm_mark_force_output_node): Update.
22677 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
22678 estimate_edge_devirt_benefit, inline_generate_summary,
22679 inline_write_summary): Update.
22680 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
22681 * ipa-prop.c (ipa_compute_jump_functions): Update.
22682 (ipa_print_node_params, ipa_prop_read_section,
22683 ipa_update_after_lto_read, read_replacements_section): Update.
22684 * varasm.c (mark_decl_referenced): Update.
22685 (assemble_alias, dump_tm_clone_pairs): Update.
22686 * tree-inline.c (copy_bb): Update.
22687 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
22689 * symtab.c (dump_symtab_base): Print new flags.
22690 (verify_symtab_base): Verify new flags.
22691 (symtab_alias_ultimate_target): New function.
22692 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
22693 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
22695 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
22697 * i386.c (ix86_get_function_versions_dispatcher,
22698 ix86_generate_version_dispatcher_body): Update.
22699 (fold_builtin_cpu): Use varpool_add_new_variable.
22700 * varpool.c (varpool_remove_initializer): Break out from ...
22701 (varpool_remove_node): ... this one.
22702 (dump_varpool_node, varpool_node_for_asm,
22703 cgraph_variable_initializer_availability, varpool_analyze_node,
22704 varpool_assemble_decl, varpool_remove_unreferenced_decls,
22705 varpool_finalize_named_section_flags, varpool_create_variable_alias):
22708 2013-05-29 Jan Hubicka <jh@suse.cz>
22710 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
22712 2013-05-29 Easwaran Raman <eraman@google.com>
22714 PR tree-optimization/57442
22715 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
22716 when control exits the main loop.
22718 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
22720 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
22722 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
22723 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
22724 * rx/t-rx: Add rx100 under multi library matches option for nofpu
22727 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22729 PR tree-optimization/57441
22730 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
22731 Don't limit size of incr_vec to number of candidates.
22733 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
22735 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
22736 and mips16 directories.
22737 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
22738 (MULTILIB_DIRNAMES): Ditto.
22739 (MULTILIB_EXCEPTIONS): Add new exceptions.
22740 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
22741 (MULTILIB_DIRNAMES): Ditto.
22742 (MULTILIB_EXCEPTIONS): Add new exceptions.
22744 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
22745 Marcus Shawcroft <marcus.shawcroft@arm.com>
22747 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
22748 SYMBOL_TINY_ABSOLUTE.
22749 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
22750 SYMBOL_TINY_ABSOLUTE.
22751 (aarch64_expand_mov_immediate): Likewise.
22752 (aarch64_classify_symbol): Likewise.
22753 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
22754 Permit SYMBOL_TINY_ABSOLUTE.
22755 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
22757 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
22758 Marcus Shawcroft <marcus.shawcroft@arm.com>
22760 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
22761 Refactor if/switch. Replace gcc_assert with if.
22763 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
22765 * config/i386/i386.c (initial_ix86_tune_features): Enable
22766 FP Reassociation for AMD bdver1 and bdver2.
22768 2013-05-29 Martin Jambor <mjambor@suse.cz>
22770 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
22771 and IMAGPART_EXPR do not occur within other handled_components.
22773 2013-05-29 Richard Biener <rguenther@suse.de>
22775 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
22776 access on whether the use is in the BB we currently try to
22778 (vect_bb_vectorization_profitable_p): Pass the BB we currently
22779 vectorize to vect_bb_slp_scalar_cost.
22781 2013-05-29 Richard Biener <rguenther@suse.de>
22783 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
22784 computing scalar cost offsetted by stmts that are kept live
22786 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
22787 for computation of scalar cost.
22789 2013-05-28 Steve Ellcey <sellcey@mips.com>
22791 * config/mips/mips-cpus.def (mips32r2): Change processor type.
22793 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
22795 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
22796 array notation built-in reduction functions.
22797 * doc/passes.texi (Passes): Added documentation about changes done
22799 * doc/invoke.texi (C Dialect Options): Added documentation about
22800 the -fcilkplus flag.
22801 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
22802 (BUILTINS_DEF): Depend on cilkplus.def.
22803 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
22804 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
22805 * cilkplus.def: New file.
22807 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
22809 PR rtl-optimization/57439
22810 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
22812 2013-05-28 Easwaran Raman <eraman@google.com>
22814 PR tree-optimization/57337
22815 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
22816 (find_insert_point): Correctly identify the insertion point
22817 when two statements with the same UID is compared.
22819 2013-05-28 Richard Biener <rguenther@suse.de>
22821 PR tree-optimization/56787
22822 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
22823 from the list of data references.
22824 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
22826 (vect_analyze_loop_operations): Likewise.
22827 (vect_transform_loop): Remove clobbers.
22829 2013-05-28 Martin Jambor <mjambor@suse.cz>
22831 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
22832 and REALPART_EXPRs have scalar type.
22834 2013-05-28 Richard Biener <rguenther@suse.de>
22836 PR tree-optimization/57411
22837 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
22839 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
22840 virtual operand propagation.
22842 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
22844 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
22845 destination register for bmasksi_vis.
22846 (vector_init_bshuffle): Likewise.
22847 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
22849 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
22851 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
22852 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
22853 mode if the instruction isn't available in the original mode.
22854 * config/sparc/sparc.opt (mfix-ut699): New option.
22855 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
22856 (divdf3): Turn into expander.
22857 (divdf3_nofix): New insn.
22858 (divdf3_fix): Likewise.
22859 (divsf3): Disable if -mfix-ut699.
22860 (sqrtdf2): Turn into expander.
22861 (sqrtdf2_nofix): New insn.
22862 (sqrtdf2_fix): Likewise.
22863 (sqrtsf2): Disable if -mfix-ut699.
22865 2013-05-27 Richard Biener <rguenther@suse.de>
22867 PR middle-end/57412
22868 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
22869 block for the new loop.
22871 2013-05-27 Richard Biener <rguenther@suse.de>
22873 PR tree-optimization/57343
22874 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
22875 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
22876 (number_of_iterations_cond): Do not build the folded tree.
22878 2013-05-27 Richard Biener <rguenther@suse.de>
22881 PR middle-end/57381
22882 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
22883 OEP_CONSTANT_ADDRESS_OF retained.
22885 PR tree-optimization/57417
22886 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
22887 for unchanged base.
22888 (set_ssa_val_to): Compare addresses using
22889 get_addr_base_and_unit_offset.
22891 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
22893 PR rtl-optimization/56833
22894 * postreload.c (move2add_record_mode): New function.
22895 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
22896 (move2add_use_add2_insn): Use move2add_record_sym_value.
22897 (move2add_use_add3_insn): Likewise.
22898 (reload_cse_move2add): Use move2add_valid_value_p and
22899 move2add_record_mode. Invalidate call-clobbered and REG_INC
22900 affected regs by setting reg_mode to VOIDmode.
22901 (move2add_note_store): Don't pretend the inside of a SUBREG is
22902 the actual destination. Invalidate single/leading registers by
22903 setting reg_mode to VOIDmode.
22904 Use move2add_record_sym_value, move2add_valid_value_p and
22905 move2add_record_mode.
22907 2013-05-27 Richard Biener <rguenther@suse.de>
22909 PR tree-optimization/57396
22910 * tree-affine.c (double_int_constant_multiple_p): Properly
22911 return false for val == 0 and div != 0.
22913 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
22915 * config/mips/mips.h: Use #elif in preprocessor conditions.
22917 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
22920 * config/mips/constraints.md (kl): New constraint.
22921 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
22922 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
22923 constraint for operand 0. Split after CSE for MIPS16. Emit a move
22924 from LO for MIPS16.
22925 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
22927 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
22930 * config/mips/mips.c (mips_can_inline_p): New function.
22931 (TARGET_CAN_INLINE_P): Define.
22933 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
22935 * sched-int.h (ds_t, dw_t): Make unsigned int.
22936 Fix documentation that describes how all the ds_t bits are used.
22937 Reserve the last bit for delayed-branch scheduling.
22938 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
22939 (BITS_PER_DEP_WEAK): Fix definition and documentation.
22940 (gen_dep_weak_1): Remove prototype.
22941 * sched-deps.c (get_dep_weak_1): Make static.
22942 * target.def (speculate_insn, needs_block_p, gen_spec_check,
22943 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
22944 * doc/tm.texi: Regenerate.
22945 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
22947 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
22950 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
22952 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
22953 Sandra Loosemore <sandra@codesourcery.com>
22955 * config.gcc (powerpc-*): Allow native for with-cpu.
22957 2013-05-24 Jeff Law <law@redhat.com>
22959 PR tree-optimization/57124
22960 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
22961 conversion feeding a condition if the range has an overflow
22962 if -fstrict-overflow. Add warnings for when we do make the
22965 2013-05-24 Dehao Chen <dehao@google.com>
22967 * tree-cfg.c (locus_discrim_map): Fix the typo.
22968 (locus_discrim_hasher): Likewise.
22969 (locus_discrim_hasher::hash): Likewise.
22970 (locus_discrim_hasher::equal): Likewise.
22972 2013-05-24 Martin Jambor <mjambor@suse.cz>
22974 PR tree-optimization/57294
22975 * cgraph.h (ipa_record_stmt_references): Declare.
22976 * cgraphbuild.c (ipa_record_stmt_references): New function.
22977 (build_cgraph_edges): Use ipa_record_stmt_references.
22978 (rebuild_cgraph_edges): Likewise.
22979 (cgraph_rebuild_references): Likewise.
22980 * ipa-prop.c (ipa_modify_call_arguments): Discard references
22981 associated with the old statement and build references from the
22982 newly built statements.
22983 * ipa-ref.c (ipa_remove_stmt_references): New function.
22984 * ipa-ref.h (ipa_remove_stmt_references): Declare.
22986 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
22988 * lra-constraints.c (emit_spill_move): Use smaller mode for
22990 (check_and_process_move): Consider mem-reg moves for secondary
22992 (curr_insn_transform): Don't lose insns emitted before for
22993 secondary memory moves.
22994 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
22995 reg set up in the current insn.
22997 2013-05-24 Dehao Chen <dehao@google.com>
22999 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
23001 (locus_descrim_hasher::equal): Likewise.
23002 (build_gimple_cfg): New discriminator assignment algorithm.
23003 (make_edges): Likewise.
23004 (next_discriminator_for_locus): Likewise.
23005 (same_line_p): Likewise.
23006 (assign_discriminators): Likewise.
23007 (make_cond_expr_edges): Likewise.
23008 (make_gimple_switch_edges): Likewise.
23009 (make_goto_expr_edges): Likewise.
23010 (make_gimple_asm_edges): Likewise.
23012 2013-05-24 Ian Bolton <ian.bolton@arm.com>
23014 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
23015 X format specifier to only display bottom 16 bits.
23016 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
23017 immediate to match for operand 2, since it will be masked.
23019 2013-05-24 Richard Biener <rguenther@suse.de>
23021 PR tree-optimization/57287
23022 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
23023 all SSA names that occur in abnormal PHIs.
23025 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
23028 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
23029 that index is not negative.
23031 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
23033 PR rtl-optimization/55177
23034 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
23035 (simplify_byte_swapping_operation): New.
23036 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
23037 (simplify_relational_operation_1): Deal with BSWAP.
23039 2013-05-23 Richard Henderson <rth@redhat.com>
23042 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
23043 (ix86_reorg): Call it.
23045 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
23048 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
23049 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
23050 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
23052 2013-05-23 Christian Bruel <christian.bruel@st.com>
23055 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
23057 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
23058 Marcus Shawcroft <marcus.shawcroft@arm.com>
23060 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
23061 * config/aarch64/constraints.md (Usa): Remove.
23062 * doc/md.texi (AArch64 Usa): Remove.
23064 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
23065 Marcus Shawcroft <marcus.shawcroft@arm.com>
23067 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
23068 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
23069 * config/aarch64/predicates.md (aarch64_const_address): Remove.
23070 (aarch64_mov_operand): Use aarch64_mov_operand_p.
23072 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
23074 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
23075 instruction (AdvSIMD).
23076 * config/aarch64/aarch64-builtins.c
23077 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
23078 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
23080 2013-05-23 Martin Jambor <mjambor@suse.cz>
23082 PR middle-end/57347
23083 * tree.h (contains_bitfld_component_ref_p): Declare.
23084 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
23085 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
23087 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
23088 not access a bit-field. Assert all final offsets are byte-aligned.
23090 2013-05-23 Richard Biener <rguenther@suse.de>
23092 PR tree-optimization/57380
23093 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
23094 least one invariant or re-used load.
23095 * passes.c (init_optimization_passes): Move pass_phiprop before
23098 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
23100 * config/aarch64/aarch64-simd.md
23101 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
23103 2013-05-23 Richard Biener <rguenther@suse.de>
23105 PR middle-end/57381
23106 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
23107 OEP_CONSTANT_ADDRESS_OF retained.
23109 2013-05-23 Jakub Jelinek <jakub@redhat.com>
23111 PR middle-end/57344
23112 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
23113 don't lower unit. Handle unit not being always BITS_PER_WORD.
23115 2013-05-23 Richard Biener <rguenther@suse.de>
23117 PR rtl-optimization/57341
23118 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
23119 instead of true_dependence.
23121 2013-05-22 David Malcolm <dmalcolm@redhat.com>
23123 * bb-reorder.c (branch_threshold): Make const.
23124 (exec_threshold): Ditto.
23126 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
23127 Pat Haugen <pthaugen@us.ibm.com>
23128 Peter Bergner <bergner@vnet.ibm.com>
23130 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
23131 documentation for the power8 crypto builtins.
23133 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
23135 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
23136 macros for defining power8 builtin functions.
23137 (BU_P8V_AV_2): Likewise.
23138 (BU_P8V_AV_P): Likewise.
23139 (BU_P8V_VSX_1): Likewise.
23140 (BU_P8V_OVERLOAD_1): Likewise.
23141 (BU_P8V_OVERLOAD_2): Likewise.
23142 (BU_CRYPTO_1): Likewise.
23143 (BU_CRYPTO_2): Likewise.
23144 (BU_CRYPTO_3): Likewise.
23145 (BU_CRYPTO_OVERLOAD_1): Likewise.
23146 (BU_CRYPTO_OVERLOAD_2): Likewise.
23147 (XSCVSPDP): Fix typo, point to the correct instruction.
23148 (VCIPHER): Add power8 crypto builtins.
23149 (VCIPHERLAST): Likewise.
23150 (VNCIPHER): Likewise.
23151 (VNCIPHERLAST): Likewise.
23152 (VPMSUMB): Likewise.
23153 (VPMSUMH): Likewise.
23154 (VPMSUMW): Likewise.
23155 (VPERMXOR_V2DI): Likewise.
23156 (VPERMXOR_V4SI: Likewise.
23157 (VPERMXOR_V8HI: Likewise.
23158 (VPERMXOR_V16QI: Likewise.
23159 (VSHASIGMAW): Likewise.
23160 (VSHASIGMAD): Likewise.
23161 (VPMSUM): Likewise.
23162 (VPERMXOR): Likewise.
23163 (VSHASIGMA): Likewise.
23165 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
23166 __CRYPTO__ if the crypto instructions are available.
23167 (altivec_overloaded_builtins): Add support for overloaded power8
23170 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
23171 support for power8 crypto builtins.
23172 (builtin_function_type): Likewise.
23173 (altivec_init_builtins): Add support for builtins that take vector
23174 long long (V2DI) arguments.
23176 * config/rs6000/crypto.md: New file, define power8 crypto
23179 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
23180 Pat Haugen <pthaugen@us.ibm.com>
23181 Peter Bergner <bergner@vnet.ibm.com>
23183 * doc/invoke.texi (Option Summary): Add power8 options.
23184 (RS/6000 and PowerPC Options): Likewise.
23186 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
23187 constraints.md instead of rs6000.h. Reorder w* constraints. Add
23188 wm, wn, wr documentation.
23190 * config/rs6000/constraints.md (wm): New constraint for VSX
23191 registers if direct move instructions are enabled.
23192 (wn): New constraint for no registers.
23193 (wq): New constraint for quad word even GPR registers.
23194 (wr): New constraint if 64-bit instructions are enabled.
23195 (wv): New constraint if power8 vector instructions are enabled.
23196 (wQ): New constraint for quad word memory locations.
23198 * config/rs6000/predicates.md (const_0_to_15_operand): New
23199 constraint for 0..15 for crypto instructions.
23200 (gpc_reg_operand): If VSX allow registers in VSX registers as well
23201 as GPR and floating point registers.
23202 (int_reg_operand): New predicate to match only GPR registers.
23203 (base_reg_operand): New predicate to match base registers.
23204 (quad_int_reg_operand): New predicate to match even GPR registers
23205 for quad memory operations.
23206 (vsx_reg_or_cint_operand): New predicate to allow vector logical
23207 operations in both GPR and VSX registers.
23208 (quad_memory_operand): New predicate for quad memory operations.
23209 (reg_or_indexed_operand): New predicate for direct move support.
23211 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
23212 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
23213 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
23214 (POWERPC_MASKS): Add power8 options.
23215 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
23218 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
23219 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
23221 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
23222 (-mpower8-fusion): New power8 options.
23223 (-mpower8-fusion-sign): Likewise.
23224 (-mpower8-vector): Likewise.
23225 (-mcrypto): Likewise.
23226 (-mdirect-move): Likewise.
23227 (-mquad-memory): Likewise.
23229 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
23230 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
23232 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
23233 (rs6000_debug_vector_unit): Add p8_vector.
23234 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
23235 definitions. Also print fusion state.
23236 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
23237 (rs6000_builtin_mask_calculate): Add power8 builtin support.
23238 (rs6000_option_override_internal): Add support for power8.
23239 (rs6000_common_init_builtins): Add debugging for skipped builtins
23240 if -mdebug=builtin.
23241 (rs6000_adjust_cost): Add power8 support.
23242 (rs6000_issue_rate): Likewise.
23243 (insn_must_be_first_in_group): Likewise.
23244 (insn_must_be_last_in_group): Likewise.
23245 (force_new_group): Likewise.
23246 (rs6000_register_move_cost): Likewise.
23247 (rs6000_opt_masks): Likewise.
23249 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
23250 power8 capable assembler, default to power7 options.
23251 (TARGET_DIRECT_MOVE): Likewise.
23252 (TARGET_CRYPTO): Likewise.
23253 (TARGET_P8_VECTOR): Likewise.
23254 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
23255 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
23256 (VECTOR_MEM_P8_VECTOR_P): Likewise.
23257 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
23258 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
23259 (TARGET_XSCVDPSPN): Likewise.
23260 (TARGET_XSCVSPDPN): Likewsie.
23261 (TARGET_SYNC_HI_QI): Likewise.
23262 (TARGET_SYNC_TI): Likewise.
23263 (MASK_CRYPTO): Likewise.
23264 (MASK_DIRECT_MOVE): Likewise.
23265 (MASK_P8_FUSION): Likewise.
23266 (MASK_P8_VECTOR): Likewise.
23267 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
23268 temporary used by some of the direct move instructions to get two FP
23269 temporary registers does not force creation of a stack frame.
23270 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
23271 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
23272 that any VSX registers are tieable, even if they are also an
23273 Altivec vector mode.
23274 (r6000_reg_class_enum): Add wm, wr, wv constraints.
23275 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
23276 (RS6000_BTM_CRYPTO): Likewise.
23277 (RS6000_BTM_COMMON): Likewise.
23279 * config/rs6000/rs6000.md (cpu attribute): Add power8.
23280 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
23281 (enum rs6000_vector): Add power8 vector support.
23283 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23287 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
23288 (any_sibcall_could_use_r3): this and handle indirect calls.
23289 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
23291 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23293 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
23295 2013-05-22 Richard Biener <rguenther@suse.de>
23297 PR middle-end/57349
23298 * profile.c (branch_prob): Do not split blocks that are
23299 abnormally receiving from ECF_RETURNS_TWICE functions.
23301 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
23303 * recog.c (offsettable_address_addr_space_p): Fix calculation of
23304 address mode. Move pointer mode initialization to the same place.
23306 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
23308 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
23309 while it has any effect.
23311 2013-05-21 Easwaran Raman <eraman@google.com>
23313 PR tree-optimization/57322
23314 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
23315 UID of the statement added to the BB to be 1.
23317 2013-05-21 Jakub Jelinek <jakub@redhat.com>
23319 PR tree-optimization/57331
23320 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
23321 of conversion from pointer type to integral type with integer.
23323 2013-05-21 Martin Jambor <mjambor@suse.cz>
23326 * ipa-prop.c (ipa_read_node_info): Process param_used and
23327 controlled_uses in the same order as when writing.
23329 2013-05-21 Magnus Granberg <baldrick@free.fr>
23332 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
23334 2013-05-21 Richard Biener <rguenther@suse.de>
23336 PR tree-optimization/57318
23337 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
23338 estimate stmts with side-effects as likely eliminated.
23340 2013-05-21 Richard Biener <rguenther@suse.de>
23342 PR tree-optimization/57330
23343 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
23344 preserve the call stmts fntype.
23346 2013-05-21 Richard Biener <rguenther@suse.de>
23348 PR tree-optimization/57303
23349 * tree-ssa-sink.c (statement_sink_location): Improve killing
23350 stmt detection and properly handle self-assignments.
23352 2013-05-21 Christian Bruel <christian.bruel@st.com>
23354 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
23355 spanning registers. LEAF_REG_REMAP is supported only for contiguous
23356 registers. Set register size out of the PARALLEL loop.
23358 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
23361 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
23362 (*fmasf4, *fmasf4_media): New insns.
23364 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
23366 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
23367 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
23368 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
23369 (mips_idiv_insns): Update the comments to say that the returned
23370 instruction counts are in units of BASE_INSN_LENGTH.
23371 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
23372 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
23373 using 2 rather than 4 as the length of indirect MIPS16 and
23374 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
23375 length of a NOP. Don't divide MIPS16 lengths by 2.
23376 (mips16_split_long_branches): Assume a branch is long if the
23377 length is greater than 4 rather than 8.
23378 * config/mips/mips.md (length): Give MIPS16 lengths directly,
23379 rather than multiplying them by 2. Multiply instruction counts
23380 by BASE_INSN_LENGTH rather than 4.
23381 (*jump_mips16, tls_get_tp_mips16_<mode>)
23382 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
23384 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
23386 * config/mips/mips.md (extended_mips16): Remove branch case.
23387 (length): Remove duplicated extended_mips16 test.
23389 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
23391 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
23393 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
23395 * recog.h (Recog_data): Rename to...
23396 (recog_data_d): ...this.
23397 (recog_data): Update accordingly.
23398 * recog.c (recog_data): Likewise.
23399 * reload.c (save_recog_data): Likewise.
23400 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
23401 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
23403 2013-05-17 Julian Brown <julian@codesourcery.com>
23405 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
23406 found in a REG_EQUAL note, invalidate it.
23408 2013-05-17 Easwaran Raman <eraman@google.com>
23410 * tree-ssa-reassoc.c (find_insert_point): New function.
23411 (insert_stmt_after): Likewise.
23412 (get_def_stmt): Likewise.
23413 (ensure_ops_are_available): Likewise.
23414 (not_dominated_by): Likewise.
23415 (rewrite_expr_tree): Do not move statements beyond what is
23416 necessary. Remove call to swap_ops_for_binary_stmt...
23417 (reassociate_bb): ... and move it here.
23418 (build_and_add_sum): Assign UIDs for new statements.
23419 (linearize_expr): Likewise.
23420 (do_reassoc): Renumber gimple statement UIDs.
23422 2013-05-17 Jan Hubicka <jh@suse.cz>
23424 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
23426 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
23427 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
23429 (output_weakrefs): Update.
23431 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
23432 Martin Jambor <mjambor@suse.cz>
23434 PR middle-end/57276
23435 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
23436 value that corresponds to the given aggval is found in values vector.
23438 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
23440 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
23441 sse, sse2, sse3, ssse3 and sse4a flags to options.
23443 2013-05-17 David Malcolm <dmalcolm@redhat.com>
23445 * gengtype-state.c: (s_expr_writer): New class, to handle
23446 prettifying of output layout of s-expressions.
23447 (state_writer): New class, to write out gtype.state.
23448 (state_written_type_count): Move this variable into member data of
23450 (s_expr_writer::s_expr_writer): New code: constructor for new class
23451 (state_writer::state_writer(): ditto
23452 (s_expr_writer::write_new_line): New function
23453 (s_expr_writer::write_any_indent): ditto
23454 (s_expr_writer::begin_s_expr): ditto
23455 (s_expr_writer::end_s_expr): ditto
23456 (write_state_fileloc): convert to method of state_writer...
23457 (state_writer:: write_state_fileloc): ...and use methods of
23458 s_expr_writer to write indentation into the gtype.state output file
23459 to visually represent the hierarchical structure of the list
23461 (write_state_fields): ditto, renaming to...
23462 (state_writer::write_state_fields)
23463 (write_state_a_string): ditto, renaming to...
23464 (state_writer::write_state_a_string)
23465 (write_state_string_option): ditto, renaming to...
23466 (state_writer::write_state_string_option)
23467 (write_state_type_option): ditto, renaming to...
23468 (state_writer::write_state_type_option)
23469 (write_state_nested_option): ditto, renaming to...
23470 (state_writer::write_state_nested_option)
23471 (write_state_option): ditto, renaming to...
23472 (state_writer::write_state_option)
23473 (write_state_options): ditto, renaming to...
23474 (state_writer::write_state_options)
23475 (write_state_lang_bitmap): ditto, renaming to...
23476 (state_writer::write_state_lang_bitmap)
23477 (write_state_version): ditto, renaming to...
23478 (state_writer::write_state_version)
23479 (write_state_scalar_type): ditto, renaming to...
23480 (state_writer::write_state_scalar_type)
23481 (write_state_string_type): ditto, renaming to...
23482 (state_writer::write_state_string_type)
23483 (write_state_undefined_type): ditto, renaming to...
23484 (state_writer::write_state_undefined_type)
23485 (write_state_struct_union_type): ditto, renaming to...
23486 (state_writer::write_state_struct_union_type)
23487 (write_state_struct_type): ditto, renaming to...
23488 (state_writer::write_state_struct_type)
23489 (write_state_user_struct_type): ditto, renaming to...
23490 (state_writer::write_state_user_struct_type)
23491 (write_state_lang_struct_type): ditto, renaming to...
23492 (state_writer::write_state_lang_struct_type)
23493 (write_state_param_struct_type): ditto, renaming to...
23494 (state_writer::write_state_param_struct_type)
23495 (write_state_pointer_type): ditto, renaming to...
23496 (state_writer::write_state_pointer_type)
23497 (write_state_array_type): ditto, renaming to...
23498 (state_writer::write_state_array_type)
23499 (write_state_gc_used): ditto, renaming to...
23500 (state_writer::write_state_gc_used)
23501 (write_state_common_type_content): ditto, renaming to...
23502 (state_writer::write_state_common_type_content)
23503 (write_state_type): ditto, renaming to...
23504 (state_writer::write_state_type)
23505 (write_state_pair_list): ditto, renaming to...
23506 (state_writer::write_state_pair_list)
23507 (write_state_pair): ditto, renaming to...
23508 (state_writer::write_state_pair)
23509 (write_state_typedefs): ditto, renaming to...
23510 (state_writer::write_state_typedefs)
23511 (write_state_structures): ditto, renaming to...
23512 (state_writer::write_state_structures)
23513 (write_state_param_structs): ditto, renaming to...
23514 (state_writer::write_state_param_structs)
23515 (write_state_variables): ditto, renaming to...
23516 (state_writer::write_state_variables)
23517 (write_state_srcdir): ditto, renaming to...
23518 (state_writer::write_state_srcdir)
23519 (write_state_files_list): ditto, renaming to...
23520 (state_writer::write_state_files_list)
23521 (write_state_languages): ditto, renaming to...
23522 (state_writer::write_state_languages)
23523 (write_state): create a state_writer instance and use it when
23524 writing out the state file
23526 2013-05-17 Mike Stump <mikestump@comcast.net>
23528 PR rtl-optimization/57304
23529 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
23530 accessing DF_REF_REAL_LOC.
23532 2013-05-17 Jakub Jelinek <jakub@redhat.com>
23534 PR rtl-optimization/57281
23535 PR rtl-optimization/57300
23536 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
23537 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
23538 what the other splitter did if the registers are dead.
23540 2013-05-17 Richard Biener <rguenther@suse.de>
23542 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
23545 2013-05-17 Jakub Jelinek <jakub@redhat.com>
23547 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
23550 2013-05-17 Marek Polacek <polacek@redhat.com>
23552 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
23553 length when doing non-zero store of storing '\0' to '\0'.
23555 2013-05-17 Jakub Jelinek <jakub@redhat.com>
23557 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
23558 vect_external_def oprnd1 with loop_vinfo, try to emit
23559 optional cast, negation and and stmts on the loop preheader
23560 edge instead of into the pattern def seq.
23562 PR tree-optimization/57051
23563 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
23564 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
23566 2013-05-16 Nick Clifton <nickc@redhat.com>
23568 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
23569 (rl78_is_naked_func): New function.
23570 (rl78_expand_prologue): Skip prologue generation for naked functions.
23571 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
23572 * doc/extend.texi (naked): Add RL78 to the list of processors
23573 that supports this attribute.
23575 2013-05-16 Jeff Law <law@redhat.com>
23577 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
23579 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
23581 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
23582 cache parameters using detect_caches_amd also for CYRIX,
23583 NSC and TM2 signatures.
23585 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
23586 Dzianis Kahanovich <mahatma@eu.by>
23590 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
23591 VIA/Centaur processors and determine their cache parameters
23592 using detect_caches_amd.
23594 2013-05-16 Teresa Johnson <tejohnson@google.com>
23596 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
23597 (rtl_verify_edges): New function.
23598 (rtl_verify_bb_insns): Ditto.
23599 (rtl_verify_bb_pointers): Ditto.
23600 (rtl_verify_bb_insn_chain): Ditto.
23601 (rtl_verify_fallthru): Ditto.
23602 (rtl_verify_bb_layout): Ditto.
23603 (rtl_verify_flow_info_1): Outline checks into new functions.
23604 (rtl_verify_flow_info): Ditto.
23606 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
23608 * cfghooks.c (copy_bbs): Add update_dominance argument.
23609 * cfghooks.h (copy_bbs): Update prototype.
23610 * tree-cfg.c (gimple_duplicate_sese_region):
23611 Add update_dominance argument.
23612 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
23613 * tree-ssa-loop-ch.c (copy_loop_headers): Update
23614 gimple_duplicate_sese_region call.
23615 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
23616 Update copy_bbs call.
23617 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
23618 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
23620 2013-05-16 Jakub Jelinek <jakub@redhat.com>
23622 * tree-vectorizer.h (NUM_PATTERNS): Increment.
23623 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
23624 vect_recog_rotate_pattern.
23625 (vect_recog_rotate_pattern): New function.
23627 2013-05-16 Jason Merrill <jason@redhat.com>
23629 * Makefile.in (LLINKER): New variable.
23630 (mostlyclean): Remove link mutex.
23631 * configure.ac: Handle --enable-link-mutex.
23632 * lock-and-run.sh: New script.
23634 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23637 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
23640 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23642 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
23644 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
23646 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
23647 * config/arm/arm.c (next_consecutive_mem): New function.
23648 (gen_movmem_ldrd_strd): Likewise.
23649 * config/arm/arm.md (movmemqi): Update condition and code.
23650 (unaligned_loaddi, unaligned_storedi): New patterns.
23652 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23654 * config.gcc: Obsolete *-*-solaris2.9*.
23655 * doc/install.texi (Specific, *-*-solaris2*): Document it.
23657 2013-05-16 Richard Biener <rguenther@suse.de>
23659 * passes.c (init_optimization_passes): Move pass_parallelize_loops
23660 earlier, after GRAPHITE transforms and IV canonicalization.
23662 2013-05-16 Jakub Jelinek <jakub@redhat.com>
23664 * omp-low.c (extract_omp_for_data): For collapsed loops,
23665 if at least one of the loops is known at compile time to
23666 iterate zero times, set count to 0.
23667 (expand_omp_regimplify_p): New function.
23668 (expand_omp_for_generic): For collapsed loops, if at least
23669 one of the loops isn't known to iterate at least once,
23670 add runtime check with setting count to 0.
23671 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
23672 For unsigned types if it isn't known at compile time that
23673 the loop will iterate at least once, add runtime check to bypass
23674 the whole loop if initial condition isn't true.
23676 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
23678 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
23680 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
23682 PR middle-end/57286
23683 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
23684 transformations to avoid an infinite loop.
23686 2013-05-16 Marek Polacek <polacek@redhat.com>
23688 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
23690 2013-05-15 Leif Ekblad <leif@rdos.net>
23692 * config/i386/i386.c (ix86_decompose_address): Use
23693 DEFAULT_TLS_SEG_REG to access TLS segment register.
23694 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
23695 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
23696 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
23698 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
23701 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
23702 sibling calls to functions that would normally be lazily bound,
23703 unless $gp is call-clobbered.
23705 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
23707 * config/i386/i386.c (ix86_option_override_internal): Update
23708 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
23709 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
23710 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
23711 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
23713 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
23715 2013-05-15 Andreas Schwab <schwab@suse.de>
23717 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
23718 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
23721 2013-05-15 Teresa Johnson <tejohnson@google.com>
23723 * loop-unroll.c (report_unroll_peel): Check decision before
23724 emitting unroll/peel message.
23726 2013-05-15 Teresa Johnson <tejohnson@google.com>
23728 * function.h (has_bb_partition): New rtl_data flag.
23729 (bb_reorder_complete): Ditto.
23730 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
23731 instead of flag_reorder_blocks_and_partition.
23732 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
23733 with some enhancements.
23734 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
23735 * bb-reorder.c (connect_traces): Check for has_bb_partition
23736 instead of flag_reorder_blocks_and_partition.
23737 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
23738 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
23739 verify_hot_cold_block_grouping.
23740 (partition_hot_cold_basic_blocks): Set has_bb_partition.
23742 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23745 * config/arm/predicates.md (call_insn_operand): New predicate.
23746 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
23747 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
23748 if insn is not a tail call.
23749 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
23751 * config/arm/arm.h (enum reg_class): New caller save register class.
23752 (REG_CLASS_NAMES): Likewise.
23753 (REG_CLASS_CONTENTS): Likewise.
23754 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
23757 2013-05-15 Richard Biener <rguenther@suse.de>
23759 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
23760 of MSG_OPTIMIZED_LOCATIONS.
23761 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
23762 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
23764 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
23765 of MSG_OPTIMIZED_LOCATIONS.
23766 (execute_vect_slp): Likewise.
23767 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
23768 (vect_create_cond_for_alias_checks): Likewise.
23769 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
23770 (vect_recog_widen_mult_pattern): Likewise.
23771 (vect_recog_widen_sum_pattern): Likewise.
23772 (vect_recog_over_widening_pattern): Likewise.
23773 (vect_recog_widen_shift_pattern): Likewise.
23774 (vect_recog_vector_vector_shift_pattern): Likewise.
23775 (vect_recog_divmod_pattern): Likewise.
23776 (vect_recog_mixed_size_cond_pattern): Likewise.
23777 (vect_recog_bool_pattern): Likewise.
23778 (vect_pattern_recog_1): Likewise.
23780 2013-05-15 Martin Jambor <mjambor@suse.cz>
23782 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
23783 non-functions to builtin_unreachable.
23784 * ipa-inline-transform.c (inline_call): Do not assert estimates were
23785 correct when new direct edges were discovered.
23787 2013-05-15 Martin Jambor <mjambor@suse.cz>
23789 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
23790 header, print symbol order instead of node uid, print more information
23791 about indirect edge targets.
23792 (ipa_make_edge_direct_to_target): Print symbol order instead of node
23794 (ipa_make_edge_direct_to_target): Likewise.
23795 (remove_described_reference): Likewise.
23796 (propagate_controlled_uses): Likewise.
23797 (ipa_print_node_params): Also print symbol order.
23798 (ipcp_transform_function): Print symbol order instead of node uids.
23799 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
23800 (cgraph_get_create_real_symbol_node): Likewise.
23801 * ipa-cp.c (print_lattice): Likewise.
23802 (print_all_lattices): Likewise.
23803 (determine_versionability): Likewise.
23804 (initialize_node_lattices): Likewise.
23805 (estimate_local_effects): Likewise.
23806 (update_profiling_info): Likewise.
23807 (create_specialized_node): Likewise.
23808 (perhaps_add_new_callers): Likewise.
23809 (decide_about_value): Likewise.
23810 (decide_whether_version_node): Likewise.
23811 (identify_dead_nodes): Likewise.
23812 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
23813 (dump_inline_summary): Likewise.
23814 (estimate_node_size_and_time): Likewise.
23815 (inline_analyze_function): Likewise.
23816 * ipa-inline.c (report_inline_failed_reason): Likewise.
23817 (want_early_inline_function_p): Likewise.
23818 (edge_badness): Likewise.
23819 (update_edge_key): Likewise.
23820 (inline_small_functions): Likewise. Add dumping of order to two other
23822 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
23823 instead of node uids.
23824 (propagate_pure_const): Likewise.
23825 (propagate_pure_const): Likewise.
23826 * ipa-utils.c (dump_cgraph_node_set): Likewise.
23827 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
23828 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
23830 * tree-pretty-print.c (dump_function_header): Likewise.
23831 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
23832 Print symbol order instead of node uids.
23834 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23836 * config/s390/s390.c (s390_register_move_cost): Don't impose the
23837 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
23839 2013-05-15 Richard Biener <rguenther@suse.de>
23841 PR tree-optimization/57275
23842 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
23843 return value for fail to do runtime alias checks for gather loads.
23845 2013-05-15 Jan Hubicka <jh@suse.cz>
23849 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
23850 weakrefs are not external.
23851 (lto_symtab_merge_decls): Fix thinko when dealing with
23852 non-lto_symtab decls.
23853 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
23854 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
23855 * varpool.c (dump_varpool_node): Dump more flags.
23857 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
23859 * config/i386/i386.c (processor_alias_table): Add instruction
23860 FSGSBASE for AMD bdver3 architecture.
23862 2013-05-14 Jakub Jelinek <jakub@redhat.com>
23864 * tree.c (warn_deprecated_use): Print file:line using locus color.
23865 * diagnostic.c (diagnostic_report_current_module): Print file:line
23866 and file:line:column using locus color.
23868 2013-05-14 Mike Stump <mikestump@comcast.net>
23870 * gdbinit.in: Add __null.
23872 2013-05-14 Mike Stump <mikestump@comcast.net>
23874 * recog.h: Rename struct recog_data to Recog_data.
23875 * recog.c: Likewise.
23876 * reload.c (can_reload_into): Likewise.
23877 * config/picochip/picochip.c: Likewise.
23879 2013-05-14 Mike Stump <mikestump@comcast.net>
23881 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
23883 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
23885 * resource.h (struct resources): Remove unch_memory member.
23886 (CLEAR_RESOURCE): Don't clear unch_memory.
23887 * resource.c (mark_referenced_resources): Don't set it.
23888 (mark_set_resources): Likewise.
23889 (mark_target_live_regs): Don't clear it.
23890 (init_resource_info): Likewise.
23891 * reorg.c (resource_conflicts_p): Don't compare it.
23892 (redundant_insn): Don't set it.
23894 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
23896 * emit-rtl.c (next_label): Remove unused function.
23897 (skip_consecutive_labels, link_cc0_insns): Move to ...
23898 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
23899 only place where these functions are used, and make them static.
23901 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
23903 * fold-const.c (fold_negate_expr): Handle vectors.
23904 (fold_truth_not_expr): Make it static.
23905 (fold_invert_truthvalue): New static function.
23906 (invert_truthvalue_loc): Handle vectors. Do not call
23907 fold_truth_not_expr directly.
23908 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
23909 <TRUTH_NOT_EXPR>: Do not cast to boolean.
23910 (fold_comparison): Handle vector constants.
23911 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
23912 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
23913 * tree.h (fold_truth_not_expr): Remove declaration.
23915 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
23917 * config/aarch64/aarch64-simd.md
23918 (aarch64_vcond_internal<mode>): Rename to...
23919 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
23920 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
23921 float modes. Clarify all iterator modes.
23922 (vcond<mode><mode>): Use new name for vcond expanders.
23923 (vcond<v_cmp_result><mode>): Likewise.
23924 (vcondu<mode><mode>: Likewise.
23925 * config/aarch64/iterators.md (VDQF_COND): New.
23927 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
23930 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
23931 variable for the shift amount. Check that we shift by non-negative
23934 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
23937 * config/arm/arm.h (EPILOGUE_USES): Only return true
23938 for LR_REGNUM after epilogue_completed.
23940 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
23942 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
23943 is error_mark_node.
23945 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23948 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
23949 and Solaris 11+/x86 with gld.
23950 * configure: Regenerate.
23952 2013-05-14 Jakub Jelinek <jakub@redhat.com>
23954 * expmed.c (expand_shift_1): Canonicalize rotates by
23955 constant bitsize / 2 to bitsize - 1.
23956 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
23957 case ROTATERT>: Likewise.
23960 2013-05-10 Jakub Jelinek <jakub@redhat.com>
23962 * config/i386/i386.md (rotateinv): New code attr.
23963 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
23964 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
23965 roll $31, %eax, etc.
23967 2013-05-14 Richard Biener <rguenther@suse.de>
23969 PR middle-end/57235
23970 * tree-eh.c (sink_clobbers): Give up for successors with
23971 multiple predecessors and no virtual uses.
23973 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
23975 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
23976 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
23978 2013-05-14 Jakub Jelinek <jakub@redhat.com>
23980 PR middle-end/57251
23981 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
23982 the case when both op0 and op1 have VOIDmode.
23984 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
23986 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
23987 in multiply-accumulate mode.
23989 2013-05-13 Guozhi Wei <carrot@google.com>
23991 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
23993 2013-05-13 Kai Tietz <ktietz@redhat.com>
23996 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
23997 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
23998 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
23999 * config/i386/i386.c (ix86_option_override_internal): Likewise.
24000 (ix86_expand_prologue): Likewise.
24001 (ix86_expand_split_stack_prologue): Likewise.
24002 (legitimate_pic_address_disp_p): Likewise.
24003 (legitimize_pic_address): Likewise.
24004 (legitimize_tls_address): Likewise.
24005 (legitimize_pe_coff_symbol): Likewise.
24006 (output_pic_addr_const): Likewise.
24007 (construct_plt_address): Likewise.
24008 (ix86_expand_call): Likewise.
24009 (x86_output_mi_thunk): Likewise.
24010 (x86_function_profiler): Likewise.
24012 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
24014 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
24015 similar switch cases.
24016 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
24017 (aarch64_simd_mov_to_<mode>low): Delete.
24018 (aarch64_simd_mov_to_<mode>high): Delete.
24019 (move_lo_quad_<mode>): Add w<-r alternative.
24020 (aarch64_simd_move_hi_quad_<mode>): Likewise.
24021 (aarch64_simd_mov_from_*): Update type attribute.
24022 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
24025 2013-05-13 Jan Hubicka <jh@suse.cz>
24027 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
24028 * config/i386/i386.c (ix86_compute_frame_layout,
24029 ix86_expand_epilogue, emit_i387_cw_initialization,
24030 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
24031 ix86_local_alignment): Fix use of size/speed predicates.
24033 2013-05-13 Jakub Jelinek <jakub@redhat.com>
24035 PR tree-optimization/45216
24036 PR tree-optimization/57157
24037 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
24038 the (-Y) & (B - 1) variant if OP is |.
24039 * expmed.c (expand_shift_1): For rotations by const0_rtx just
24040 return shifted. Use (-op1) & (prec - 1) as other_amount
24041 instead of prec - op1.
24043 2013-05-13 Martin Jambor <mjambor@suse.cz>
24045 PR middle-end/42371
24046 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
24047 (ipa_constant_data): New type.
24048 (ipa_jump_func): Use ipa_constant_data to hold information about
24049 constant jump functions.
24050 (ipa_get_jf_constant): Adjust to jump function type changes.
24051 (ipa_get_jf_constant_rdesc): New function.
24052 (ipa_param_descriptor): New field controlled_uses.
24053 (ipa_get_controlled_uses): New function.
24054 (ipa_set_controlled_uses): Likewise.
24055 * ipa-ref.h (ipa_find_reference): Declare.
24056 * ipa-prop.c (ipa_cst_ref_desc): New type.
24057 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
24059 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
24060 New parameter cs. Adjust all callers.
24061 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
24062 (remove_described_reference): New function.
24063 (jfunc_rdesc_usable): Likewise.
24064 (try_make_edge_direct_simple_call): Decrement controlled use count,
24065 attempt to remove reference if it hits zero.
24066 (combine_controlled_uses_counters): New function.
24067 (propagate_controlled_uses): Likewise.
24068 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
24069 (ipa_edge_duplication_hook): Duplicate reference descriptions.
24070 (ipa_print_node_params): Print described use counter.
24071 (ipa_write_jump_function): Adjust to jump function type changes.
24072 (ipa_read_jump_function): New parameter CS, pass it to
24073 ipa_set_jf_constant. Adjust caller.
24074 (ipa_write_node_info): Stream controlled use count
24075 (ipa_read_node_info): Likewise.
24076 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
24078 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
24079 count. Remove cloning-added reference if it reaches zero.
24080 * ipa-ref.c (ipa_find_reference): New function.
24082 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
24084 * config/i386/i386.c (processor_target_table): Modified default
24085 alignment values for AMD BD and BT architectures.
24087 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
24089 * tree-vect-generic.c (uniform_vector_p): Move ...
24090 * tree.c (uniform_vector_p): ... here.
24091 * tree.h (uniform_vector_p): Declare it.
24092 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
24095 2013-05-13 Jakub Jelinek <jakub@redhat.com>
24097 PR tree-optimization/57230
24098 * tree-ssa-strlen.c (handle_char_store): Record length for
24099 array store from STRING_CST.
24101 PR tree-optimization/57230
24102 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
24105 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
24107 * config/epiphany/epiphany.c (epiphany_init): Check size of
24108 NUM_MODES_FOR_MODE_SWITCHING.
24109 (epiphany_expand_prologue):
24110 Remove CONFIG_REGNUM initial value handling code.
24111 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
24112 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
24113 (emit_set_fp_mode, epiphany_mode_after): Likewise.
24114 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
24115 Don't return 1 for FP_MODE_NONE.
24116 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
24117 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
24118 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
24119 * config/epiphany/epiphany.md (save_config): New pattern.
24121 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
24123 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
24125 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
24127 * config/i386/i386.md (memory): Handle sseishft1.
24128 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
24129 (*vec_extractv2di_1): Ditto.
24131 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
24133 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
24136 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
24138 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
24139 Add mthumb/march=armv7-a multilib.
24140 Add mthumb/march=armv7-r multilib.
24141 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
24143 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
24145 * config/v850/t-rtems: Add more multilibs.
24147 2013-05-10 Richard Biener <rguenther@suse.de>
24149 PR tree-optimization/57214
24150 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
24151 not propagate from SSA names that occur in abnormal PHI nodes.
24153 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
24155 * stor-layout.c (element_precision): New function.
24156 * machmode.h (element_precision): Declare it.
24157 * tree.c (build_minus_one_cst): New function.
24158 (element_precision): Likewise.
24159 * tree.h (build_minus_one_cst): Declare new function.
24160 (element_precision): Likewise.
24161 * fold-const.c (operand_equal_p): Use element_precision.
24162 (fold_binary_loc): Handle vector types.
24163 * convert.c (convert_to_integer): Use element_precision.
24164 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
24167 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
24169 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
24170 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
24171 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
24172 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
24173 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
24174 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
24175 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
24176 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
24177 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
24178 (Uuw8): New constraints.
24179 (Usb4): Move into alphabetical order.
24180 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
24181 (sd8_operand, ub8_operand, uw8_operand): New predicates.
24182 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
24183 previously unnamed patterns.
24184 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
24185 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
24186 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
24187 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
24188 of set_attr_alternative/if_then_else. Use extended_mips16 instead
24189 of specific lengths.
24191 2013-05-10 Jakub Jelinek <jakub@redhat.com>
24193 * config/i386/i386.md (rotateinv): New code attr.
24194 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
24195 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
24196 roll $31, %eax, etc.
24198 PR tree-optimization/45216
24199 PR tree-optimization/57157
24200 * tree-ssa-forwprop.c (simplify_rotate): New function.
24201 (ssa_forward_propagate_and_combine): Call it.
24203 2013-05-10 Richard Biener <rguenther@suse.de>
24205 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
24206 disable peeling when we version for aliasing.
24207 (vector_alignment_reachable_p): Honor explicit user alignment.
24208 (vect_supportable_dr_alignment): Likewise.
24209 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
24210 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
24211 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
24212 then peeling to arrange for the cost-model check to come first.
24214 2013-05-10 Alan Modra <amodra@gmail.com>
24216 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
24217 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
24218 * configure: Regenerate.
24220 2013-05-10 Alan Modra <amodra@gmail.com>
24223 * varasm.c (default_elf_select_section): Move !DECL_P check..
24224 (get_named_section): ..to here before calling get_section_name.
24226 (default_section_type_flags): Add DECL_P check.
24227 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
24228 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
24230 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
24232 * config/epiphany/epiphany.c (epiphany_expand_prologue):
24233 When using gen_stack_adjust_str with a register offset, add a
24234 REG_FRAME_RELATED_EXPR note.
24236 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
24238 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
24239 (*vec_extractv4si_zext_mem): Ditto.
24240 (*vec_extractv2di): Add 0->x and x->x alternatives.
24241 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
24242 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
24244 2013-05-09 Jason Merrill <jason@redhat.com>
24246 N3639 C++1y VLA support
24247 * gimplify.c (gimplify_vla_decl): Don't touch an existing
24250 * tree.c (build_constructor_va): New.
24251 * tree.h: Declare it.
24253 2013-05-09 Martin Jambor <mjambor@suse.cz>
24256 * gimple-fold.c (canonicalize_constructor_val): Call
24257 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
24259 2013-05-09 Jan Hubicka <jh@suse.cz>
24260 Richard Biener <rguenther@suse.de>
24263 * symtab.c (symtab_make_decl_local): Do not add private names.
24265 2013-05-09 Jan Hubicka <jh@suse.cz>
24268 * symtab.c (insert_to_assembler_name_hash): Handle clones.
24269 (unlink_from_assembler_name_hash): Likewise.
24270 (symtab_prevail_in_asm_name_hash, symtab_register_node,
24271 symtab_unregister_node, symtab_initialize_asm_name_hash,
24272 change_decl_assembler_name): Update.
24274 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
24276 * config/aarch64/aarch64.md: New movtf split.
24277 (*movtf_aarch64): Update.
24278 (aarch64_movdi_tilow): Handle TF modes and rename to
24279 aarch64_movdi_<mode>low.
24280 (aarch64_movdi_tihigh): Handle TF modes and rename to
24281 aarch64_movdi_<mode>high
24282 (aarch64_movtihigh_di): Handle TF modes and rename to
24283 aarch64_mov<mode>high_di
24284 (aarch64_movtilow_di): Handle TF modes and rename to
24285 aarch64_mov<mode>low_di
24286 (aarch64_movtilow_tilow): Remove spurious whitespace.
24287 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
24289 (aarch64_print_operand): Update.
24291 2013-05-09 Alan Modra <amodra@gmail.com>
24293 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
24295 * configure: Regenerate.
24297 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
24299 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
24300 splitter preparation statements.
24301 * config/i386/sse.md (*vec_extract* splitters): Ditto.
24302 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
24305 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24307 * gimple-ssa-strength-reduction.c (count_candidates): Change
24308 return value to int.
24309 (analyze_candidates_and_replace): Change type of length to int.
24311 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
24313 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
24314 (*vec_extract<mode>): Use VI12_128 mode iterator.
24315 (*vec_extract<mode>_mem): Ditto.
24316 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
24319 2013-05-08 Diego Novillo <dnovillo@google.com>
24324 2012-08-17 Diego Novillo <dnovillo@google.com>
24327 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
24328 * config.in: Regenerate.
24329 * configure: Regenerate.
24330 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
24332 2013-05-08 Jan Hubicka <jh@suse.cz>
24335 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
24336 * cgraph.h (symtab_node_base): Add unique_name.
24337 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
24338 input_overwrite_node, input_varpool_node): Stream unique_name.
24339 * cgraphclones.c (cgraph_create_virtual_clone,
24340 cgraph_function_versioning): Set unique_name.
24341 * ipa.c (function_and_variable_visibility): Set unique_name.
24343 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24345 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
24346 (alloc_cand_and_find_basis): Restrict conditional candidate
24347 processing to CAND_MULTs.
24349 2013-05-08 Jan Hubicka <jh@suse.cz>
24352 lto-symtab.c (lto_symtab_symbol_p): New function.
24353 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
24354 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
24355 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
24356 Skip static symbols.
24358 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
24360 PR tree-optimization/57200
24361 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
24362 Only call inform if the preceding warning_at returns true.
24364 2013-05-07 Han Shen <shenhan@google.com>
24366 * cfgexpand.c (record_or_union_type_has_array_p): New function.
24367 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
24368 * common.opt (fstack-protector-strong): New option.
24369 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
24370 * doc/invoke.texi (Optimization Options): Document
24371 "-fstack-protector-strong".
24372 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
24374 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
24376 * config/mips/mips.c (mips_machine_reorg2): Return 0.
24378 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
24380 * ira.c (update_equiv_regs): Add insn having equiv memory even if
24381 it is not lhs of the insn.
24382 (setup_reg_equiv): Remove insn having equiv memory which it is not
24384 * lra-constraints.c (process_address): Try to improve generation
24385 code for address base + disp.
24386 (lra_constraints): Make correct the code for checking insn setting
24387 up backward equivalence. Remove insn only if it is in the init
24389 * lra-eliminations.c (update_reg_eliminate): Change return value.
24390 (lra_eliminate): Use the result.
24392 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
24394 * config/i386/sse.md (ssescalarnummask): New mode attribute.
24395 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
24396 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
24397 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
24398 register target operands.
24399 (*vec_extractv8hi_sse2): New pattern.
24400 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
24401 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
24402 (*vec_extract<mode>_mem): New insn and split pattern.
24404 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
24406 * config/arm/arm.c (arm_asan_shadow_offset): New function.
24407 (TARGET_ASAN_SHADOW_OFFSET): Define.
24408 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
24409 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
24411 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24413 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
24414 (incr_vec_index): Return -1 if increment not found.
24415 (create_add_on_incoming_edge): Assert if increment not found.
24416 (record_increment): Limit number of increments recorded.
24417 (all_phi_incrs_profitable): Return false if an increment not found.
24418 (replace_profitable_candidates): Don't process increments that were
24420 (analyze_candidates_and_replace): Limit size of incr_vec.
24422 2013-05-07 Richard Biener <rguenther@suse.de>
24424 * calls.c (special_function_p): setjmp-like functions are leaf.
24425 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
24426 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
24428 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
24430 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
24431 (aarch64_simd_mov<mode>): New expander.
24432 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
24433 (aarch64_simd_mov_to_<mode>high): Likewise.
24434 (aarch64_simd_mov_from_<mode>low): Likewise.
24435 (aarch64_simd_mov_from_<mode>high): Likewise.
24436 (aarch64_dup_lane<mode>): Update.
24437 (aarch64_dup_lanedi): New instruction pattern.
24438 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
24439 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
24441 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24443 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
24444 (replace_mult_candidate): Remove unnecessary argument; remove
24445 unnecessary parameter from call to introduce_cast_before_cand.
24446 (replace_unconditional_candidate): Remove unnecessary parameter
24447 from call to replace_mult_candidate.
24448 (replace_conditional_candidate): Likewise.
24449 (insert_initializers): Use make_temp_ssa_name.
24450 (introduce_cast_before_cand): Remove unnecessary argument; use
24451 make_temp_ssa_name.
24452 (replace_one_candidate): Remove unnecessary argument; remove
24453 unnecessary parameter from calls to introduce_cast_before_cand.
24454 (replace_profitable_candidates): Remove unnecessary parameters
24455 from calls to replace_one_candidate.
24457 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24459 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
24460 phi def as possibly hiding a basis for a CAND_ADD whose operands
24461 have been commuted in the analysis.
24462 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
24464 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24466 * config/aarch64/aarch64.md
24467 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
24468 shift value between 0-4.
24470 2013-05-07 Richard Biener <rguenther@suse.de>
24472 * double-int.h (rshift): New overload.
24473 * double-int.c (rshift): New function.
24474 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
24475 (create_reference_ops_from_ref): Remove.
24476 (vn_reference_insert): Use shared ops for constructing the
24477 reference and copy it.
24479 2013-05-07 Richard Biener <rguenther@suse.de>
24481 PR middle-end/57190
24482 * tree-eh.c (sink_clobbers): Properly propagate
24483 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
24485 2013-05-07 Jakub Jelinek <jakub@redhat.com>
24487 PR tree-optimization/57149
24488 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
24489 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
24490 collect_phi_def_edges, execute_late_warn_uninitialized): Use
24491 uninit_undefined_value_p instead of ssa_undefined_value_p.
24494 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
24495 for modifier == EXPAND_INITIALIZER.
24497 2013-05-07 Anton Blanchard <anton@samba.org>
24499 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
24500 for powerpc64 little endian.
24501 * configure: Regenerate.
24503 2013-05-06 Graham Stott <grahams@btinternet.com>
24505 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
24506 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
24507 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
24508 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
24510 2013-05-06 Graham Stott <grahams@btinternet.com>
24512 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
24513 codes which allow non-lvalues.
24515 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
24517 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
24518 components are all 1s.
24519 (integer_minus_onep): New function.
24520 * tree.h (integer_minus_onep): Declare it.
24521 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
24522 integer_minus_onep instead of integer_all_onesp.
24524 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
24527 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
24528 variations of these patterns.
24530 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
24532 * config/i386/i386.md (isa): Add x64_sse4 member.
24533 (enabled): Handle x64_sse4.
24534 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
24535 instruction for 64bit SSE4_1 targets. Update insn attributes.
24536 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
24537 instruction for SSE4_1 targets. Update insn attributes.
24538 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
24539 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
24540 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
24542 (*vec_extractv4si): Rename from *sse4_1_pextrd.
24543 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
24544 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
24546 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
24549 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
24551 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
24553 * final.c (do_assembler_dialects): Don't handle curly braces and
24554 vertical bar escaped by % as dialect delimiters.
24555 (output_asm_insn): Print curly braces and vertical bar if escaped
24556 by % and ASSEMBLER_DIALECT defined.
24557 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
24558 * doc/tm.texi: Regenerated.
24560 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
24562 * config/mips/mips.c: Include tree-pass.h.
24563 (mips_reorg): Split in pre- and post-dbr_schedule parts.
24564 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
24565 (pass_mips_machine_reorg2): New machine specific pass.
24566 (insert_pass_mips_machine_reorg2): New pass plugin definition.
24567 (mips_option_override): Register the new pass.
24568 * rtl.h (cleanup_barriers): Remove prototype.
24569 (dbr_schedule): Likewise.
24570 * jump.c (cleanup_barriers): Make static.
24571 * reorg.c (dbr_schedule): Likewise.
24573 2013-05-06 Richard Biener <rguenther@suse.de>
24575 PR tree-optimization/57185
24576 * tree-parloops.c (add_field_for_reduction): Handle anonymous
24577 SSA names properly.
24579 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
24582 * config/i386/i386.c (add_parameter_dependencies): Add dependence
24583 between "first_arg" and "insn", not "last" and "insn".
24585 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24587 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
24588 (find_candidates_in_block): Re-enable slsr_process_phi.
24589 (create_phi_basis): Fix double counting of candidate adjustment.
24591 2013-05-06 Richard Biener <rguenther@suse.de>
24593 PR middle-end/57147
24594 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
24595 the edge is also fallthru, preserve it and just clear the
24597 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
24598 also complex, preserve that and just clear the fallthru flag.
24599 * tree-inline.c (update_ssa_across_abnormal_edges): Also
24600 update virtual operands.
24602 2013-05-06 Alan Modra <amodra@gmail.com>
24604 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
24605 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
24606 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
24607 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
24608 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
24609 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
24611 2013-05-06 Alan Modra <amodra@gmail.com>
24613 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
24614 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
24615 (DEFAULT_ASM_ENDIAN): Define.
24616 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
24617 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
24618 Update -K PIC clause from sysv4.h.
24619 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
24620 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
24622 2013-05-06 Alan Modra <amodra@gmail.com>
24624 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
24625 twice for little-endian.
24626 (ashrdi3_no_power, ashrdi3): Support little-endian.
24628 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
24631 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
24632 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
24634 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
24635 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
24636 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
24639 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
24640 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24642 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
24643 * config/pa/pa.opt: Make mbig-switch a no-op.
24644 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
24645 (CASE_VECTOR_MODE): Always return SImode.
24646 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
24647 for the !TARGET_BIG_SWITCH case.
24648 * config/pa/pa-linux.h: Likewise.
24649 * config/pa/pa-openbsd.h: Likewise.
24650 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
24651 * config/pa/pa.md (short_jump): Remove define_insn.
24652 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
24653 (casesi0): Remove define_insn.
24654 (type): Remove btable_branch.
24655 (pa_combine_type): Likewise.
24656 (in_nullified_branch_delay): Likewise.
24657 (in_call_delay): Likewise.
24658 (define_delay): Likewise.
24659 (define_insn_reservation "Z3"): Likewise.
24660 (define_insn_reservation "Z4"): Likewise.
24661 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
24662 (pa_adjust_insn_length): Remove adjustment for btable branches.
24663 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
24666 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
24668 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
24669 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
24670 Add m->r,x alternatives.
24671 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
24672 splitters using SWI48x mode iterator.
24673 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
24674 TARGET_64BIT. Add m->x alternative.
24675 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
24676 Add o->x alternative. Enable for TARGET_SSE.
24677 (sse_storeq): Remove expander.
24678 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
24679 with memory input operand.
24680 (*vec_extractv2di_1 splitter): New.
24681 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
24682 * config/i386/i386.md (ssevecmodelower): New mode attribute.
24684 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
24686 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
24687 (INT_LOWPART): Delete.
24688 (extract_MB): Adjust.
24689 (extract_ME): Adjust.
24690 (print_operand): Adjust.
24692 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
24694 * config/rs6000/predicates.md (reg_or_add_cint_operand,
24695 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
24696 (reg_or_logical_cint_operand, easy_fp_constant,
24697 logical_const_operand): Delete "CONST_DOUBLE" case.
24698 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
24699 "HOST_BITS_PER_WIDE_INT == 64" test.
24700 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
24701 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
24702 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
24703 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
24705 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
24706 CONST_DOUBLE DImode/VOIDmode case.
24707 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
24708 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
24709 CONST_DOUBLE VOIDmode case.
24710 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
24711 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
24712 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
24713 Delete CONST_DOUBLE case.
24714 (splitters for mov FMOVE64 const_double): Delete
24715 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
24716 "HOST_BITS_PER_WIDE_INT >= 64" test.
24717 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
24719 (mov DI const_double): Delete.
24721 2013-05-04 Jakub Jelinek <jakub@redhat.com>
24723 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
24724 on op shows all bits zero in mode of a lowpart subreg, return zero.
24726 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
24729 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
24730 to save TFmode registers and DImode to save TImode registers for
24731 caller save operations.
24732 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
24733 mark being partially clobbered since they only use the first
24736 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
24737 and TDmode only use the upper 64-bits of each VSX register.
24739 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24741 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
24742 (find_candidates_in_block): Disable slsr_process_phi.
24744 2013-05-03 Guozhi Wei <carrot@google.com>
24746 * coverage.c (coverage_obj_init): Move the construction of gcov
24748 (build_init_ctor): ... here.
24750 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24752 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
24753 (slsr_cand_d): Redefine def_phi.
24754 (stride_status, phi_adjust_status, count_phis_status): New enums.
24755 (find_phi_def): New.
24756 (find_basis_for_base_expr): New.
24757 (find_basis_for_candidate): Handle hidden bases.
24758 (alloc_cand_and_find_basis): Handle phi candidates.
24759 (slsr_process_phi): New.
24760 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
24761 (create_mul_imm_cand): Likewise.
24762 (create_add_ssa_cand): Exclude phi base candidates.
24763 (create_add_imm_cand): Likewise.
24764 (slsr_process_cast): Likewise.
24765 (slsr_process_copy): Likewise.
24766 (find_candidates_in_block): Handle phi candidates.
24767 (dump_candidate): Likewise.
24768 (unconditional_cands): Delete.
24769 (unconditional_cands_with_known_stride_p): Delete.
24770 (phi_dependent_cand_p): New.
24771 (cand_increment): Handle phi-dependent candidates.
24772 (replace_dependent): Delete.
24773 (replace_mult_candidate): New.
24774 (replace_unconditional_candidate): New.
24775 (incr_vec_index): Move to avoid forward reference.
24776 (create_add_on_incoming_edge): New.
24777 (create_phi_basis): New.
24778 (replace_dependents): Delete.
24779 (replace_conditional_candidate): New.
24780 (phi_add_costs): New.
24781 (replace_uncond_cands_and_profitable_phis): New.
24782 (record_increment): Handle phi adjustments.
24783 (record_phi_increments): New.
24784 (record_increments): Handle phi adjustments.
24785 (phi_incr_cost): New.
24786 (lowest_cost_path): Handle phis.
24787 (total_savings): Likewise.
24788 (analyze_increments): Likewise.
24789 (ncd_with_phi): New.
24790 (ncd_of_cand_and_phis): New.
24791 (nearest_common_dominator_for_cands): Handle phi increments.
24792 (all_phi_incrs_profitable): New.
24793 (replace_profitable_candidates): Handle phi-dependent candidates.
24794 (analyze_candidates_and_replace): Likewise.
24796 2013-05-03 Teresa Johnson <tejohnson@google.com>
24799 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
24800 do not exceed REG_BR_PROB_BASE.
24802 2013-05-03 Jeff Law <law@redhat.com>
24804 PR tree-optimization/57144
24805 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
24806 operand of the condition will bit into the new type when eliminating
24807 a cast feeding a condition.
24809 2013-05-03 Jakub Jelinek <jakub@redhat.com>
24811 PR rtl-optimization/57130
24812 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
24813 of COMPARE as in_code to the recursive call if needed.
24815 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
24817 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
24818 (enabled): Handle new members.
24819 * config/i386/sse.md (*vec_concatv2si): Merge from
24820 *vec_concatv2si_sse2 and vec_concatv2si_sse.
24821 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
24823 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
24825 PR tree-optimization/57027
24826 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
24827 for fnms opportunity, check we got the prerequisite kind
24828 of tree / gimple before using accessor functions.
24830 2013-05-03 Richard Biener <rguenther@suse.de>
24832 * double-int.h (lshift): New overload without precision
24833 and arith argument.
24834 (operator *=, operator +=, operator -=): Move ...
24835 * double-int.c (operator *=, operator +=, operator -=): ... here
24836 and implement more efficiently.
24837 (mul_double_with_sign): Remove.
24838 (lshift_double): Adjust to take unsinged shift argument, push
24839 dispatching code to callers.
24840 (mul_double_wide_with_sign): Add early out for callers that
24841 are not interested in high parts or overflow.
24842 (lshift): New function.
24843 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
24844 dispatch code here.
24845 (lrotate, rrotate): Use logical shifts.
24846 * expr.c (get_inner_reference): Use lshift.
24847 * fixed-value.c (do_fixed_divide): Likewise.
24848 * tree-dfa.c (get_ref_base_and_extent): Likewise.
24849 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
24850 (indirect_refs_may_alias_p): Likewise.
24851 (stmt_kills_ref_p_1): Likewise.
24853 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
24855 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
24857 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
24859 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
24860 scalar form of FABD instruction.
24862 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
24864 * lra-constraints.c (process_alt_operands): Add checking alt
24865 number to choose the best alternative.
24867 2013-05-02 Richard Biener <rguenther@suse.de>
24869 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
24870 bitmap and its handling.
24871 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
24873 2013-05-02 Richard Biener <rguenther@suse.de>
24875 PR middle-end/57140
24876 * tree-inline.c (copy_loops): Properly handle removed loops.
24877 (copy_cfg_body): Mark destination loops for fixup if source
24878 loops needed fixup.
24880 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
24883 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
24884 generating simple_return for naked functions.
24886 2013-05-02 Martin Jambor <mjambor@suse.cz>
24888 PR middle-end/56988
24889 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
24890 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
24892 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
24893 ipa_agg_replacement_value structures.
24894 (known_aggs_to_agg_replacement_list): Likewise.
24895 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
24896 (read_agg_replacement_chain): Likewise.
24897 (ipcp_transform_function): Also check that by_ref flags match.
24899 2013-05-02 Richard Biener <rguenther@suse.de>
24901 * graphds.h (struct graph): Add obstack member.
24902 * graphds.c (new_graph): Initialize obstack and allocate
24904 (add_edge): Allocate edge from the obstack.
24905 (free_graph): Free the obstack instead of all edges and vertices.
24907 2013-05-02 Teresa Johnson <tejohnson@google.com>
24909 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
24911 * cfg.c (update_bb_profile_for_threading): Ditto.
24912 * tree-inline.c (copy_bb): Ditto.
24913 (copy_edges_for_bb): Ditto.
24914 (initialize_cfun): Ditto.
24915 (copy_cfg_body): Ditto.
24916 (expand_call_inline): Ditto.
24917 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
24918 (estimate_node_size_and_time): Ditto.
24919 (inline_merge_summary): Ditto.
24920 * cgraphclones.c (cgraph_clone_edge): Ditto.
24921 (cgraph_clone_node): Ditto.
24922 * sched-rgn.c (compute_dom_prob_ps): Ditto.
24923 (compute_trg_info): Ditto.
24925 2013-05-02 Ian Bolton <ian.bolton@arm.com>
24927 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
24928 S reg when fp attribute set.
24929 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
24931 2013-05-02 Ian Bolton <ian.bolton@arm.com>
24933 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
24935 (*and_one_cmplsi3_compare0_uxtw): Likewise.
24936 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
24937 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
24939 2013-05-02 Richard Biener <rguenther@suse.de>
24941 * tree-scalar-evolution.c (scev_info_hasher): Remove.
24942 (struct instantiate_cache_entry): New type.
24943 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
24944 (struct instantiate_cache_type): New type.
24945 (set_instantiated_value, get_instantiated_value): Remove.
24946 (get_instantiated_value_entry): New function.
24947 (instantiate_scev_name): Use the new cache and adjust.
24948 (instantiate_scev_poly): Adjust.
24949 (instantiate_scev_binary): Likewise.
24950 (instantiate_array_ref): Likewise.
24951 (instantiate_scev_convert): Likewise.
24952 (instantiate_scev_not): Likewise.
24953 (instantiate_scev_3): Likewise.
24954 (instantiate_scev_2): Likewise.
24955 (instantiate_scev_r): Likewise.
24956 (instantiate_scev): Likewise.
24957 (resolve_mixers): Likewise.
24959 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
24962 * lra-constraints.c (best_small_class_operands_num): Remove.
24963 (process_alt_operands): Remove small_class_operands_num. Take
24964 small classes operands into losers and only if the operand is not
24965 matched. Modify debugging output.
24966 (curr_insn_transform): Remove best_small_class_operands_num.
24969 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
24971 * config/aarch64/aarch64-builtins.c
24972 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
24973 * config/aarch64/aarch64-simd-builtins.def
24974 (reduc_splus_): Add new modes.
24975 (reduc_uplus_): New.
24976 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
24977 (reduc_uplus_v4sf): Likewise.
24978 (reduc_splus_v4sf): Likewise.
24979 (aarch64_addv<mode>): Likewise.
24980 (reduc_uplus_<mode>): Likewise.
24981 (reduc_splus_<mode>): Likewise.
24982 (aarch64_addvv2di): Likewise.
24983 (reduc_uplus_v2di): Likewise.
24984 (reduc_splus_v2di): Likewise.
24985 (aarch64_addvv2si): Likewise.
24986 (reduc_uplus_v2si): Likewise.
24987 (reduc_splus_v2si): Likewise.
24988 (reduc_<sur>plus_<mode>): New.
24989 (reduc_<sur>plus_v2di): Likewise.
24990 (reduc_<sur>plus_v2si): Likewise.
24991 (reduc_<sur>plus_v4sf): Likewise.
24992 (aarch64_addpv4sf): Likewise.
24993 * config/aarch64/arm_neon.h
24994 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
24995 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
24996 add UNSPEC_SADDV, UNSPEC_UADDV.
24998 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
25000 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
25002 * config/aarch64/arm_neon.h
25003 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
25005 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
25007 * config/aarch64/aarch64-builtins
25008 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
25010 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
25012 * config/aarch64/aarch64-simd-builtins.def
25013 (reduc_smax_): New.
25014 (reduc_smin_): Likewise.
25015 (reduc_umax_): Likewise.
25016 (reduc_umin_): Likewise.
25017 (reduc_smax_nan_): Likewise.
25018 (reduc_smin_nan_): Likewise.
25021 (smax): Update for V2SF, V4SF and V2DF modes.
25024 (smin_nan): Likewise.
25025 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
25026 (<su><maxmin><mode>3): ...This, refactor.
25027 (s<maxmin><mode>3): New.
25028 (<maxmin_uns><mode>3): Likewise.
25029 (reduc_<maxmin_uns>_<mode>): Refactor.
25030 (reduc_<maxmin_uns>_v4sf): Likewise.
25031 (reduc_<maxmin_uns>_v2si): Likewise.
25032 (aarch64_<fmaxmin><mode>: Remove.
25033 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
25035 (vmin<q>_f<32,64>): Likewise.
25036 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
25038 (su): Add mappings for smax, smin, umax, umin.
25040 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
25041 (FMAXMIN): Rename as...
25042 (FMAXMIN_UNS): ...This.
25044 (fmaxminv): Likewise.
25045 (fmaxmin): Likewise.
25047 (maxmin_uns_op): Likewise.
25049 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
25051 * config/aarch64/arm_neon.h
25052 (vac<ge, gt><sd>_f<32, 64>): Rename to...
25053 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
25054 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
25056 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
25058 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
25059 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
25061 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
25063 * config/aarch64/aarch64-simd.md
25064 (vcond<mode>_internal): Handle special cases for constant masks.
25065 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
25066 (vcondu<mode><mode>): Likewise.
25067 (vcond<v_cmp_result><mode>): New.
25069 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
25071 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
25072 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
25073 * config/aarch64/aarch64-simd-builtins.def
25074 (cmeq): Update to BUILTIN_VALLDI.
25079 * config/aarch64/arm_neon.h
25080 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
25081 to builtins or C as appropriate.
25083 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
25085 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
25087 (cmhi): Rename to...
25089 * config/aarch64/aarch64-simd.md
25090 (simd_mode): Add SF.
25091 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
25092 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
25093 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
25094 (cstore<mode>_neg): ...This.
25095 * config/aarch64/iterators.md
25097 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
25098 (COMPARISONS): New.
25099 (UCOMPARISONS): Likewise.
25100 (optab): Add missing comparisons.
25106 (VCMP_S): Likewise.
25107 (VCMP_U): Likewise.
25108 (V_cmp_result): Add DF, SF modes.
25109 (v_cmp_result): Likewise.
25111 (vmtype): Likewise.
25112 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
25114 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
25116 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
25117 define_insn to define_insn_and_split.
25118 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
25119 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
25120 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
25121 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
25122 (thumb2_negscc): Likewise.
25124 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
25126 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
25128 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
25130 * config/arm/thumb2.md: Remove trailing whitespaces.
25132 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
25134 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
25135 Use gen_int_mode rather than GEN_INT.
25137 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
25139 * value-prof.c (stream_in_histogram_value): Remove the strayed
25142 2013-04-30 Richard Biener <rguenther@suse.de>
25144 PR middle-end/57122
25145 * cfghooks.c (split_edge): Properly check for the loop latch edge.
25147 2013-04-30 Richard Biener <rguenther@suse.de>
25149 PR middle-end/57107
25150 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
25152 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
25154 PR rtl-optimization/56957
25155 PR rtl-optimization/57105
25156 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
25157 variable. Use just INSN_UID for determining whether an insn
25158 should be only disconnected from the insn stream.
25159 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
25161 2013-04-30 Jakub Jelinek <jakub@redhat.com>
25163 PR tree-optimization/57104
25164 * tsan.c (instrument_expr): Don't instrument accesses to
25165 DECL_HARD_REGISTER VAR_DECLs.
25167 2013-04-30 Richard Biener <rguenther@suse.de>
25169 * function.h (loops_for_fn): New inline function.
25170 (set_loops_for_fn): Likewise.
25171 * cfgloop.h (place_new_loop): Add struct function parameter.
25172 (get_loop): Likewise.
25173 (get_loops): Likewise.
25174 (number_of_loops): Likewise.
25175 (fel_next): Adjust.
25176 (fel_init): Likewise.
25177 * cfg.c (get_loop_copy): Adjust.
25178 * cfgloop.c (flow_loops_dump): Likewise.
25179 (record_loop_exits): Likewise.
25180 (verify_loop_structure): Likewise.
25181 * cfgloopanal.c (mark_irreducible_loops): Likewise.
25182 (estimate_reg_pressure_cost): Likewise.
25183 (mark_loop_exit_edges): Likewise.
25184 * cfgloopmanip.c (place_new_loop): Likewise.
25185 (add_loop): Likewise.
25186 (duplicate_loop): Likewise.
25187 * graph.c (draw_cfg_nodes): Likewise.
25188 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
25189 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
25190 (extract_affine_chrec): Likewise.
25191 (build_scop_iteration_domain): Likewise.
25192 * graphite.c (graphite_initialize): Likewise.
25193 * ira-build.c (create_loop_tree_nodes): Likewise.
25194 (more_one_region_p): Likewise.
25195 (rebuild_regno_allocno_maps): Likewise.
25196 (mark_loops_for_removal): Likewise.
25197 (mark_all_loops_for_removal): Likewise.
25198 (remove_unnecessary_regions): Likewise.
25199 (ira_build): Likewise.
25200 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
25201 * loop-init.c (fix_loop_structure): Likewise.
25202 (gate_rtl_move_loop_invariants): Likewise.
25203 (gate_rtl_unswitch): Likewise.
25204 (gate_rtl_unroll_and_peel_loops): Likewise.
25205 (rtl_doloop): Likewise.
25206 * lto-streamer-in.c (input_cfg): Likewise.
25207 * lto-streamer-out.c (output_cfg): Likewise.
25208 * modulo-sched.c (sms_schedule): Likewise.
25209 * predict.c (tree_estimate_probability): Likewise.
25210 (tree_estimate_probability_driver): Likewise.
25211 (estimate_loops): Likewise.
25212 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
25213 (move_sese_region_to_fn): Likewise.
25214 (debug_loop_num): Likewise.
25215 * tree-chrec.c (chrec_evaluate): Likewise.
25216 (hide_evolution_in_other_loops_than_loop): Likewise.
25217 (chrec_component_in_loop_num): Likewise.
25218 (reset_evolution_in_loop): Likewise.
25219 (evolution_function_is_invariant_rec_p): Likewise.
25220 * tree-if-conv.c (main_tree_if_conversion): Likewise.
25221 * tree-inline.c (copy_loops): Likewise.
25222 (copy_cfg_body): Likewise.
25223 (tree_function_versioning): Likewise.
25224 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
25225 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
25227 (add_to_evolution_1): Likewise.
25228 (scev_const_prop): Likewise.
25229 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
25230 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
25231 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
25232 (tree_ssa_lim_initialize): Likewise.
25233 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
25234 (verify_loop_closed_ssa): Likewise.
25235 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
25236 (tree_ssa_loop_im): Likewise.
25237 (tree_ssa_loop_unswitch): Likewise.
25238 (tree_vectorize): Likewise.
25239 (check_data_deps): Likewise.
25240 (tree_ssa_loop_ivcanon): Likewise.
25241 (tree_ssa_loop_bounds): Likewise.
25242 (tree_complete_unroll): Likewise.
25243 (tree_complete_unroll_inner): Likewise.
25244 (tree_parallelize_loops): Likewise.
25245 (tree_ssa_loop_prefetch): Likewise.
25246 (tree_ssa_loop_ivopts): Likewise.
25247 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
25248 * tree-vectorizer.c (vectorize_loops): Likewise.
25250 2013-04-29 Mike Frysinger <vapier@gentoo.org>
25252 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
25253 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
25254 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
25255 with EABI_LINK_SPEC.
25257 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
25260 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
25263 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
25266 * lra-constraints.c (process_alt_operands): Discourage a bit more
25267 using memory for pseudos. Print cost dump for alternatives.
25268 Modify cost values for conflicts with early clobbers.
25269 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
25271 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
25274 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
25276 2013-04-29 Ian Bolton <ian.bolton@arm.com>
25278 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
25279 from/to S register.
25280 (movdi_aarch64): Support LDR/STR from/to D register.
25282 2013-04-29 Ian Bolton <ian.bolton@arm.com>
25284 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
25285 or higher by default.
25287 2013-04-29 Richard Biener <rguenther@suse.de>
25289 PR middle-end/57075
25290 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
25291 even if not adding abnormal edges for calls that can make
25294 2013-04-29 Richard Biener <rguenther@suse.de>
25296 PR middle-end/57103
25297 * tree-cfg.c (move_stmt_op): Fix condition under which to update
25299 (move_stmt_r): Remove redundant checking.
25301 2013-04-29 Teresa Johnson <tejohnson@google.com>
25304 * basic-block.h (apply_scale): New function.
25305 (apply_probability): Use apply_scale.
25306 * gimple-streamer-in.c (input_bb): Ditto.
25307 * lto-streamer-in.c (input_cfg): Ditto.
25308 * lto-cgraph.c (merge_profile_summaries): Ditto.
25309 * tree-optimize.c (execute_fixup_cfg): Ditto.
25310 * tree-inline.c (copy_bb): Update comment to use apply_scale.
25311 (copy_edges_for_bb): Ditto.
25312 (copy_cfg_body): Ditto.
25314 2013-04-29 Tom de Vries <tom@codesourcery.com>
25316 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
25317 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
25318 (tail_merge_optimize): Handle current_loops == NULL.
25320 2013-04-26 Jeff Law <law@redhat.com>
25322 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
25323 (simplify_cond_using_ranges): Generalize code to simplify
25324 COND_EXPRs where one argument is a constant and the other
25325 is an SSA_NAME created by an integral type conversion.
25327 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25329 * config/arm/arm.md (store_minmaxsi): Use only when
25330 optimize_insn_for_size_p.
25332 2013-04-29 Christian Bruel <christian.bruel@st.com>
25335 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
25337 2013-04-29 Richard Biener <rguenther@suse.de>
25339 PR middle-end/57089
25340 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
25341 loop tree make sure to schedule a fixup for the child as well.
25342 (expand_omp_for_generic): Properly add loops.
25343 (expand_omp_for_static_nochunk): Likewise.
25344 (expand_omp_for_static_chunk): Likewise.
25345 (expand_omp_for): For the degenerate case fixup loops.
25346 (expand_omp_sections): Fix default bb placement in loops.
25347 (expand_omp_atomic_pipeline): Properly add loops.
25349 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25351 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
25353 2013-04-29 Tom de Vries <tom@codesourcery.com>
25355 * tree-ssa-tail-merge.c: Update header comment.
25357 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
25359 * config/aarch64/arm_neon.h
25360 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
25361 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
25362 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
25363 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
25364 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
25365 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
25366 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
25367 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
25369 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
25371 * config/aarch64/aarch64-simd.md
25372 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
25373 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
25374 fix_trunc, fixuns_trunc.
25375 (ftrunc<VDQF:mode>2): New.
25376 * config/aarch64/iterators.md (optab): Add fix, fixuns.
25377 (fix_trunc_optab): New.
25379 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
25381 * config/aarch64/aarch64-builtins.c
25382 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
25383 iceilf, lround, iroundf.
25385 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
25388 * config/i386/i386.h (enum ix86_tune_indices)
25389 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
25390 New, split from X86_TUNE_INTER_UNIT_MOVES.
25391 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
25392 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
25393 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
25394 (TARGET_INTER_UNIT_MOVES): Remove.
25395 * config/i386/i386.c (initial_ix86_tune_features): Update.
25396 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
25397 (ix86_expand_convert_uns_didf_sse): Use
25398 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
25399 (ix86_expand_vector_init_one_nonzero): Ditto.
25400 (ix86_expand_vector_init_interleave): Ditto.
25401 (inline_secondary_memory_needed): Return true for moves from SSE class
25402 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
25403 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
25404 * config/i386/constraints.md (Yi, Ym): Depend on
25405 TARGET_INTER_UNIT_MOVES_TO_VEC.
25406 (Yj, Yn): New constraints.
25407 * config/i386/i386.md (*movdi_internal): Change constraints of
25408 operand 1 from Yi to Yj and from Ym to Yn.
25409 (*movsi_internal): Ditto.
25410 (*movdf_internal): Ditto.
25411 (*movsf_internal): Ditto.
25412 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
25413 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
25414 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
25415 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
25416 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
25417 * config/i386/sse.md (movdi_to_sse): Ditto.
25418 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
25419 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
25420 TARGET_INTER_UNIT_MOVES.
25421 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
25422 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
25423 instead of TARGET_INTER_UNIT_MOVES.
25424 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
25425 operand 1 from Yi to Yj and from Ym to Yn.
25427 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
25429 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
25430 (float_truncate_hi_): Likewise.
25431 (float_extend_lo_): Likewise.
25432 (float_truncate_lo_): Likewise.
25433 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
25434 (aarch64_float_extend_lo_v2df): Likewise.
25435 (vec_unpacks_hi_v4sf): Likewise.
25436 (aarch64_float_truncate_lo_v2sf): Likewise.
25437 (aarch64_float_truncate_hi_v4sf): Likewise.
25438 (vec_pack_trunc_v2df): Likewise.
25439 (vec_pack_trunc_df): Likewise.
25441 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
25443 * config/aarch64/aarch64-builtins.c
25444 (aarch64_fold_builtin): Fold float conversions.
25445 * config/aarch64/aarch64-simd-builtins.def
25446 (floatv2si, floatv4si, floatv2di): New.
25447 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
25448 * config/aarch64/aarch64-simd.md
25449 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
25450 * config/aarch64/iterators.md (FLOATUORS): New.
25451 (optab): Add float, floatuns.
25452 (su_optab): Likewise.
25454 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
25456 * config/aarch64/aarch64-builtins.c
25457 (aarch64_builtin_vectorized_function): Use new names for
25459 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
25460 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
25461 (fcvtzu): Split as...
25462 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
25463 (fcvtas): Split as...
25464 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
25465 (fcvtau): Split as...
25466 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
25467 (fcvtps): Split as...
25468 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
25469 (fcvtpu): Split as...
25470 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
25471 (fcvtms): Split as...
25472 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
25473 (fcvtmu): Split as...
25474 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
25475 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
25476 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
25477 (lfrintnusf, lfrintnudf): Likewise.
25478 * config/aarch64/aarch64-simd.md
25479 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
25481 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
25482 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
25483 (fcvt_pattern): Likewise.
25485 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
25487 * config/aarch64/aarch64-simd.md
25488 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
25489 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
25491 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
25493 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
25494 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
25495 (vrnd<a,m,n,p>_f32): Implement using builtins.
25496 (vrnd<i,x><q>_f<32, 64>): New.
25498 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
25500 * config/aarch64/aarch64-builtins.c
25501 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
25502 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
25503 (frintz): Rename to...
25505 (frintp): Rename to...
25507 (frintm): Rename to...
25509 (frinti): Rename to...
25510 (nearbyint): ...this.
25511 (frintx): Rename to...
25513 (frinta): Rename to...
25515 * config/aarch64/aarch64-simd.md
25516 (aarch64_frint<frint_suffix><mode>): Delete.
25517 (<frint_pattern><mode>2): Convert to insn.
25518 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
25519 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
25520 (frint_pattern): Likewise.
25521 (frint_suffix): Likewise.
25523 2013-04-29 Richard Biener <rguenther@suse.de>
25525 PR tree-optimization/57081
25526 * loop-init.c: Include tree-flow.h.
25527 (loop_optimizer_finalize): Free number of iteration estimates.
25528 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
25530 2013-04-29 Jakub Jelinek <jakub@redhat.com>
25532 PR tree-optimization/57083
25533 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
25534 non-singleton shift count range, zero extend low_bound for uns case.
25536 * config/i386/predicates.md (general_vector_operand): New predicate.
25537 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
25538 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
25539 if they aren't nonimmediate operands. If their original values
25540 satisfy const_vector_equal_evenodd_p, don't shift them.
25541 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
25542 predicates. For the SSE4.1 case force operands[{1,2}] into registers
25543 if not nonimmediate_operand.
25544 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
25545 instead of register_operand.
25546 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
25548 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
25550 * stor-layout.c (finalize_size_functions): Allocate a structure and
25551 reset cfun before dumping the functions.
25553 2013-04-27 Jakub Jelinek <jakub@redhat.com>
25555 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
25558 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
25559 use xop_pmacsdqh if uns_p.
25560 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
25561 the immediate rotate count.
25563 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
25565 * rtl.h (struct rtx_def): Add comment for field jump.
25566 (LRA_SUBREG_P): New macro.
25567 * recog.c (register_operand): Check LRA_SUBREG_P.
25568 * lra.c (lra): Add note at the end of RTL code. Align non-empty
25570 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
25571 (lra_final_code_change): Skip subreg change for operators.
25572 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
25573 if there are no operand changes.
25574 * lra-constraints.c (curr_insn_set): New.
25575 (match_reload): Set LRA_SUBREG_P.
25576 (emit_spill_move): Ditto.
25577 (check_and_process_move): Use curr_insn_set. Process only single
25578 set insns. Don't initialize sec_mem_p and change_p.
25579 (simplify_operand_subreg): Use LRA_SUBREG_P.
25580 (reg_in_class_p): New function.
25581 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
25582 of #ifdef. Add code to remove cycling.
25583 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
25584 non-null disp. Reload inner instead of disp when base and index
25585 are null. Try to put lo_sum into register.
25586 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
25587 (check_and_process_move): Move code for move cost check to
25588 simple_move_p. Remove equiv_substitution.
25589 (simple_move_p): New function.
25590 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
25591 curr_insn_set. Call check_and_process_move only for single set
25592 insns. Use the new function. Move call of check_and_process_move
25593 after operand equiv substitution and address process.
25595 2013-04-26 Jakub Jelinek <jakub@redhat.com>
25598 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
25599 with nonlocal goto receivers or returns twice calls, ignore
25600 unininitialized values from abnormal edges to nl goto receiver
25601 or returns twice call.
25603 2013-04-26 Jakub Jelinek <jakub@redhat.com>
25605 PR tree-optimization/57051
25606 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
25607 and VEC_RSHIFT_EXPR if shift count is a multiple of element
25610 2013-04-26 Richard Biener <rguenther@suse.de>
25612 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
25613 (expand_omp_taskreg): Likewise. Mark loops for fixup.
25614 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
25615 (fixup_loop_arrays_after_move): New function.
25616 (move_sese_region_to_fn): Properly outline the loop tree parts
25617 of the SESE region.
25619 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
25621 * config/i386/i386.md (type, unit): Fix long lines.
25623 2013-04-26 Richard Biener <rguenther@suse.de>
25625 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
25626 (lto-streamer-out.o): Likewise.
25627 * cfgloop.c (init_loops_structure): Export, add struct function
25628 argument and adjust.
25629 (flow_loops_find): Adjust.
25630 * cfgloop.h (enum loop_estimation): Add EST_LAST.
25631 (init_loops_structure): Declare.
25632 * lto-streamer-in.c: Include cfgloop.h.
25633 (input_cfg): Input the loop tree.
25634 * lto-streamer-out.c: Include cfgloop.h.
25635 (output_cfg): Output the loop tree.
25636 (output_struct_function_base): Do not drop PROP_loops.
25638 2013-03-26 Richard Biener <rguenther@suse.de>
25640 * tree-cfg.c (execute_build_cfg): Build the loop tree.
25641 (pass_build_cfg): Provide PROP_loops.
25642 (move_sese_region_to_fn): Remove loops that are outlined into fn
25644 * tree-inline.c: Include cfgloop.h.
25645 (initialize_cfun): Do not drop PROP_loops.
25646 (copy_loops): New function.
25647 (copy_cfg_body): Copy loop structure.
25648 (tree_function_versioning): Initialize destination loop tree.
25649 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
25650 (pass_parallelize_loops): Do IL verification.
25651 * loop-init.c (loop_optimizer_init): Fixup loops if required.
25652 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
25653 the CFG make sure we fixup loops as well.
25654 * tree-ssa-tail-merge.c: Include cfgloop.h.
25655 (replace_block_by): When merging loop latches mark loops for fixup.
25656 * lto-streamer-out.c (output_struct_function_base): Drop
25657 PROP_loops for now.
25658 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
25659 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
25660 * ipa-split.c: Include cfgloop.h.
25661 (split_function): Add the new return block to the loop tree root.
25662 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
25663 whether we have removed the forwarder block.
25664 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
25665 * cfgloop.h (place_new_loop): Declare.
25666 * cfgloopmanip.c (place_new_loop): Export.
25667 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
25668 (tree-switch-conversion.o): Likewise.
25669 (tree-complex.o): Likewise.
25670 (tree-inline.o): Likewise.
25671 (tree-ssa-tailmerge.o): Likewise.
25672 (ipa-split.o): Likewise.
25673 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
25674 (tree-ssa-copy.o): Likewise.
25675 * tree-switch-conversion.c: Include cfgloop.h
25676 (process_switch): If we emit a bit-test cascade, schedule loops
25678 * tree-complex.c: Include cfgloop.h.
25679 (expand_complex_div_wide): Properly add new basic-blocks to loops.
25680 * asan.c: Include cfgloop.h.
25681 (create_cond_insert_point): Properly add new basic-blocks to
25682 loops, schedule loop fixup.
25683 * cfgloop.c (verify_loop_structure): Check that looks are not
25685 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
25687 (expand_omp_for_generic): Likewise.
25688 (expand_omp_sections): Likewise.
25689 (expand_omp_atomic_pipeline): Schedule loops for fixup.
25690 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
25691 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
25692 is initialized, not when loops are present.
25693 * tree-parloops.c (parallelize_loops): Remove checking here.
25694 * passes.c (init_optimization_passes): Schedule a copy-propagation
25695 pass before complete unrolling of inner loops.
25697 2013-04-26 Jakub Jelinek <jakub@redhat.com>
25699 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
25700 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
25701 (colorize_init): Add argument to _WIN32 version.
25702 * toplev.c: Include diagnostic-color.h.
25703 (process_options): Default to -fdiagnostics-color=auto if
25704 GCC_COLORS env var is in the environment.
25705 * common.opt (fdiagnostics-color=): Add Var and Init.
25706 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
25707 env var is in the environment, the default is auto rather than never.
25709 * diagnostic.h (file_name_as_prefix): Add context argument.
25710 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
25711 the string as locus.
25712 * langhooks.c (lhd_print_error_function): Adjust caller.
25714 2013-04-25 Lawrence Crowl <crowl@google.com>
25716 * var-tracking.c (shared_hash_def::htab):
25717 Change type to hash_table. Update dependent calls and types.
25719 2013-04-25 Lawrence Crowl <crowl@google.com>
25721 * Makefile.in: Update as needed below.
25723 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
25724 Move declaration to after the type's method definitons.
25726 * attribs.c (htab_t scoped_attributes::attribute_hash):
25727 Change type to hash_table. Update dependent calls and types.
25729 * bitmap.c (htab_t bitmap_desc_hash):
25730 Change type to hash_table. Update dependent calls and types.
25732 * cselib.c (htab_t cselib_hash_table):
25733 Change type to hash_table. Update dependent calls and types.
25735 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
25736 (hash_string_slot_node): Move implementation into lto-streamer.h
25737 struct string_slot_hasher.
25738 (eq_string_slot_node): Likewise.
25740 * data-streamer-out.c: Update output_block::string_hash_table
25741 dependent calls and types.
25743 * dwarf2cfi.c (htab_t trace_index):
25744 Change type to hash_table. Update dependent calls and types.
25746 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
25747 Change type to hash_table. Update dependent calls and types.
25748 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
25749 (htab_t optimize_external_refs::map): Likewise.
25750 (htab_t output_comp_unit::extern_map): Likewise.
25751 (htab_t output_comdat_type_unit::extern_map): Likewise.
25752 (htab_t output_macinfo::macinfo_htab): Likewise.
25753 (htab_t optimize_location_lists::htab): Likewise.
25754 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
25756 * except.c (htab_t ehspec_hash_type):
25757 Change type to hash_table. Update dependent calls and types.
25758 (assign_filter_values::ttypes): Likewise.
25759 (assign_filter_values::ehspec): Likewise.
25760 (sjlj_assign_call_site_values::ar_hash): Likewise.
25761 (convert_to_eh_region_ranges::ar_hash): Likewise.
25763 * gcse.c (htab_t pre_ldst_table):
25764 Change type to hash_table. Update dependent calls and types.
25766 * ggc-common.c (htab_t saving_htab):
25767 Change type to hash_table. Update dependent calls and types.
25768 (htab_t loc_hash): Likewise.
25769 (htab_t ptr_hash): Likewise.
25770 (call_count): Rename ggc_call_count.
25771 (call_alloc): Rename ggc_call_alloc.
25772 (loc_descriptor): Rename make_loc_descriptor.
25773 (add_statistics): Rename ggc_add_statistics.
25775 * ggc-common.c (saving_htab):
25776 Change type to hash_table. Update dependent calls and types.
25778 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
25779 (push_gimplify_context): Likewise.
25780 (pop_gimplify_context): Likewise.
25781 (struct gimple_temp_hash_elt): Added.
25782 (struct gimplify_hasher): Likewise.
25783 (struct gimplify_ctx.temp_htab):
25784 Change type to hash_table. Update dependent calls and types.
25786 * gimple-fold.c: Include gimplify-ctx.h.
25788 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
25789 Change type to hash_table. Update dependent calls and types.
25790 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
25791 avoid potential global name collision.
25793 * gimplify.c: Include gimplify-ctx.h.
25794 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
25795 (htab_t gimplify_ctx::temp_htab):
25796 Update dependent calls and types for new type hash_table.
25797 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
25798 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
25800 * gimplify-ctx.h: New.
25801 (struct gimple_temp_hash_elt): Move from gimplify.c.
25802 (class gimplify_hasher): New.
25803 (struct gimplify_ctx): Move from gimple.h.
25804 (htab_t gimplify_ctx::temp_htab):
25805 Change type to hash_table. Update dependent calls and types.
25807 * graphite-clast-to-gimple.c: Include graphite-htab.h.
25808 (htab_t ivs_params::newivs_index):
25809 Change type to hash_table. Update dependent calls and types.
25810 (htab_t ivs_params::params_index): Likewise.
25811 (htab_t print_generated_program::params_index): Likewise.
25812 (htab_t gloog::newivs_index): Likewise.
25813 (htab_t gloog::params_index): Likewise.
25815 * graphite.c: Include graphite-htab.h.
25816 4htab_t graphite_transform_loops::bb_pbb_mapping):
25817 Change type to hash_table. Update dependent calls and types.
25819 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
25820 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
25821 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
25823 * graphite-dependences.c: Include graphite-htab.h.
25824 (loop_is_parallel_p): Change hash table type of parameter.
25826 * graphite-htab.h: New.
25827 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
25828 (extern find_pbb_via_hash): Move from graphite-poly.h.
25829 (extern loop_is_parallel_p): Move from graphite-poly.h.
25830 (extern get_loop_body_pbbs): Move from graphite-poly.h.
25832 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
25833 (extern loop_is_parallel_p): Move to graphite-htab.h.
25834 (extern get_loop_body_pbbs): Move to graphite-htab.h.
25836 * haifa-sched.c (htab_t delay_htab):
25837 Change type to hash_table. Update dependent calls and types.
25838 (htab_t delay_htab_i2): Likewise.
25840 * ira-color.c (htab_t allocno_hard_regs_htab):
25841 Change type to hash_table. Update dependent calls and types.
25843 * ira-costs.c (htab_t cost_classes_htab):
25844 Change type to hash_table. Update dependent calls and types.
25846 * loop-invariant.c (htab_t merge_identical_invariants::eq):
25847 Change type to hash_table. Update dependent calls and types.
25849 * loop-iv.c (htab_t bivs):
25850 Change type to hash_table. Update dependent calls and types.
25852 * loop-unroll.c (htab_t opt_info::insns_to_split):
25853 Change type to hash_table. Update dependent calls and types.
25854 (htab_t opt_info::insns_with_var_to_expand): Likewise.
25856 * lto-streamer.h (struct string_slot): Move from data-streamer.h
25857 (struct string_slot_hasher): New.
25858 (htab_t output_block::string_hash_table):
25859 Change type to hash_table. Update dependent calls and types.
25861 * lto-streamer-in.c (freeing_string_slot_hasher): New.
25862 (htab_t file_name_hash_table):
25863 Change type to hash_table. Update dependent calls and types.
25865 * lto-streamer-out.c: Update output_block::string_hash_table dependent
25868 * lto-streamer.c (htab_t tree_htab):
25869 Change type to hash_table. Update dependent calls and types.
25871 * omp-low.c: Include gimplify-ctx.h.
25873 * passes.c (htab_t name_to_pass_map):
25874 Change type to hash_table. Update dependent calls and types.
25875 (pass_traverse): Rename to passes_pass_traverse.
25877 * plugin.c (htab_t event_tab):
25878 Change type to hash_table. Update dependent calls and types.
25880 * postreload-gcse.c (htab_t expr_table):
25881 Change type to hash_table. Update dependent calls and types.
25882 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
25884 * sese.c (debug_rename_map_1): Make extern.
25885 (htab_t copy_bb_and_scalar_dependences::rename_map):
25886 Change type to hash_table. Update dependent calls and types.
25888 * sese.h (extern debug_rename_map): Move to .c file.
25890 * store-motion.c (htab_t store_motion_mems_table):
25891 Change type to hash_table. Update dependent calls and types.
25893 * trans-mem.c (htab_t tm_new_mem_hash):
25894 Change type to hash_table. Update dependent calls and types.
25896 * tree-browser.c (htab_t TB_up_ht):
25897 Change type to hash_table. Update dependent calls and types.
25899 * tree-cfg.c (htab_t discriminator_per_locus):
25900 Change type to hash_table. Update dependent calls and types.
25902 * tree-complex.c: Include tree-hasher.h
25903 (htab_t complex_variable_components):
25904 Change type to hash_table. Update dependent calls and types.
25906 * tree-eh.c (htab_t finally_tree):
25907 Change type to hash_table. Update dependent calls and types.
25909 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
25910 struct int_tree_hasher.
25911 (extern int_tree_map_eq): Likewise.
25912 (uid_decl_map_hash): Removed.
25913 (extern decl_tree_map_eq): Likewise.
25915 * tree-hasher.h: New.
25916 (struct int_tree_hasher): New.
25917 (typedef int_tree_htab_type): New.
25919 * tree-inline.c: Include gimplify-ctx.h.
25921 * tree-mudflap.c: Include gimplify-ctx.h.
25923 * tree-parloops.c: Include tree-hasher.h.
25924 (htab_t eliminate_local_variables_stmt::decl_address):
25925 Change type to hash_table. Update dependent calls and types.
25926 (htab_t separate_decls_in_region::decl_copies): Likewise.
25928 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
25929 Change type to hash_table. Update dependent calls and types.
25931 * tree-sra.c (candidates):
25932 Change type to hash_table. Update dependent calls and types.
25934 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
25936 (int_tree_map_hash): Likewise.
25938 * tree-ssa-dom.c (htab_t avail_exprs):
25939 Change type to hash_table. Update dependent calls and types.
25941 * tree-ssa-live.c (var_map_base_init::tree_to_index):
25942 Change type to hash_table. Update dependent calls and types.
25944 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
25945 Change type to hash_table. Update dependent calls and types.
25947 * tree-ssa-phiopt.c (seen_ssa_names):
25948 Change type to hash_table. Update dependent calls and types.
25950 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
25951 Change type to hash_table. Update dependent calls and types.
25953 * tree-ssa-uncprop.c (equiv):
25954 Change type to hash_table. Update dependent calls and types.
25956 2013-04-25 Jakub Jelinek <jakub@redhat.com>
25958 PR rtl-optimization/57003
25959 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
25960 call note_stores with kill_clobbered_value callback again after
25961 killing regs_invalidated_by_call.
25963 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
25965 * config/aarch64/aarch64-simd.md
25966 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
25967 (aarch64_simd_bsl<mode>): Likewise.
25968 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
25970 2013-04-25 Marek Polacek <polacek@redhat.com>
25972 PR tree-optimization/57066
25973 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
25975 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
25977 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
25979 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
25981 * config/aarch64/aarch64-builtins.c
25982 (aarch64_fold_builtin): New.
25983 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
25984 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
25985 * config/aarch64/aarch64-simd-builtins.def (abs): New.
25986 * config/aarch64/arm_neon.h
25987 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
25989 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
25990 Tejas Belagod <tejas.belagod@arm.com>
25992 * config/aarch64/aarch64-builtins.c
25993 (aarch64_gimple_fold_builtin): New.
25994 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
25995 * config/aarch64/aarch64-simd-builtins.def (addv): New.
25996 * config/aarch64/aarch64-simd.md (addpv4sf): New.
25997 (addvv4sf): Update.
25998 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
26000 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26002 * config/aarch64/aarch64.md
26003 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
26005 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26007 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
26008 (*ngcsi_uxtw): New pattern.
26010 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26011 Julian Brown <julian@codesourcery.com>
26013 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
26014 (TB_DREG): Add T_V4HF.
26015 (v4hf_UP): New macro.
26016 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
26017 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
26018 Handle initialisation of V4HF. Adjust initialisation of reinterpret
26020 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
26021 (arm_vector_mode_supported_p): Handle V4HF.
26022 (arm_mangle_map): Handle V4HFmode.
26023 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
26024 * config/arm/arm_neon_builtins.def: Add entries for
26025 vcvtv4hfv4sf, vcvtv4sfv4hf.
26026 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
26027 (neon_vcvtv4hfv4sf): Likewise.
26028 * config/arm/neon-gen.ml: Handle half-precision floating point
26030 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
26031 * config/arm/arm_neon.h: Regenerate.
26032 * config/arm/neon.ml (type elts): Add F16.
26033 (type vectype): Add T_float16x4, T_floatHF.
26034 (type vecmode): Add V4HF.
26035 (type features): Add Requires_FP_bit feature.
26036 (elt_width): Handle F16.
26037 (elt_class): Likewise.
26038 (elt_of_class_width): Likewise.
26039 (mode_of_elt): Refactor.
26040 (type_for_elt): Handle F16, fix error messages.
26041 (vectype_size): Handle T_float16x4.
26042 (vcvt_sh): New function.
26043 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
26044 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
26045 (string_of_mode): Handle V4HF.
26046 * doc/arm-neon-intrinsics.texi: Regenerate.
26048 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
26050 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
26051 format specifier in 'X' case.
26053 2013-04-25 Alan Modra <amodra@gmail.com>
26056 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
26057 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
26058 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
26059 Repeat for many other rotate/shift and mask patterns using subregs.
26060 Name lshiftrt insns.
26061 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
26062 on WORDS_BIG_ENDIAN.
26064 2013-04-25 Alan Modra <amodra@gmail.com>
26066 * config.gcc: Support little-endian powerpc-linux targets.
26067 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
26068 (LINK_OS_LINUX_SPEC): Define.
26069 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
26070 Preserve MASK_LITTLE_ENDIAN.
26071 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
26072 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
26073 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
26074 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
26075 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
26076 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
26077 Correct fp word order for little-endian. Don't shift toc entries
26078 smaller than a word for little-endian.
26079 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
26080 (bswapdi2 splits): Correct low-part subreg for little-endian.
26081 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
26082 low/high where such is correct only for be.
26083 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
26084 little-endian for -mcall-aixdesc.
26086 2013-04-25 Alan Modra <amodra@gmail.com>
26088 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
26089 replace_equiv_address_nv.
26091 2013-04-25 Alan Modra <amodra@gmail.com>
26093 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
26095 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
26098 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
26099 * rtl.h (struct rtx_def): ...
26101 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
26103 PR rtl-optimizations/57046
26104 * lra-constraints (split_reg): Set up lra_risky_transformations_p
26105 for multi-reg splits.
26107 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
26109 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
26111 2013-04-24 Sterling Augustine <saugustine@google.com>
26113 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
26114 (comp_dir_string, debug_str_dwo_section): New.
26115 (DEBUG_STR_DWO_SECTION): Rename to ...
26116 (DEBUG_DWO_STR_SECTION): ... this.
26117 (DEBUG_NORM_STR_SECTION): Delete.
26118 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
26119 (DEBUG_STR_DWO_SECTION_FLAGS): New.
26120 (find_AT_string): Move most logic to ...
26121 (find_AT_string_in_table): ... here. New.
26122 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
26123 add_skeleton_AT_string. Delete logic.
26124 (output_skeleton_debug_sections): Remove call to
26125 add_top_level_skeleton_die_attrs.
26126 (add_comp_dir_attribute): Move logic to comp_dir_string.
26127 (dwarf2out_init): Initialize debug_str_dwo_section.
26128 (output_indirect_string): Call find_string_form.
26129 (output_indirect_strings): Rewrite.
26130 (prune_unused_types): Empty skeleton_debug_str_hash.
26131 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
26132 (dwarf2out_finish): Call output_indirect_strings.
26134 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
26136 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
26138 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
26140 * rtl.h (struct rtx_def): Add comment for field jump.
26141 (LRA_SUBREG_P): New macro.
26142 * recog.c (register_operand): Check LRA_SUBREG_P.
26143 * lra.c (lra): Add note at the end of RTL code. Align non-empty
26145 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
26146 (lra_final_code_change): Skip subreg change for operators.
26147 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
26148 if there are no operand changes.
26149 * lra-constraints.c (curr_insn_set): New.
26150 (match_reload): Set LRA_SUBREG_P.
26151 (emit_spill_move): Ditto.
26152 (check_and_process_move): Use curr_insn_set. Process only single
26153 set insns. Don't initialize sec_mem_p and change_p.
26154 (simplify_operand_subreg): Use LRA_SUBREG_P.
26155 (reg_in_class_p): New function.
26156 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
26157 of #ifdef. Add code to remove cycling.
26158 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
26159 non-null disp. Reload inner instead of disp when base and index
26160 are null. Try to put lo_sum into register.
26161 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
26162 (check_and_process_move): Move code for move cost check to
26163 simple_move_p. Remove equiv_substitution.
26164 (simple_move_p): New function.
26165 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
26166 curr_insn_set. Call check_and_process_move only for single set
26167 insns. Use the new function. Move call of check_and_process_move
26168 after operand equiv substitution and address process.
26170 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
26172 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
26173 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
26174 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
26176 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
26178 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
26180 2013-04-24 Marek Polacek <polacek@redhat.com>
26182 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
26183 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
26184 (select_loops_exit_conditions): Likewise.
26185 (number_of_iterations_for_all_loops): Likewise.
26186 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
26187 (scev_analysis): Likewise.
26189 2013-04-02 Catherine Moore <clm@codesourcery.com>
26190 Chao-ying Fu <fu@mips.com>
26192 * config/mips/micromips.md (jraddiusp): New pattern.
26193 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
26194 instruction if possible.
26196 2013-04-24 Alan Modra <amodra@gmail.com>
26198 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
26200 2013-04-24 Julian Brown <julian@codesourcery.com>
26201 Chung-Lin Tang <cltang@codesourcery.com>
26203 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
26204 dependency behavior in enumeration type DIE generation. Add TODO note
26205 to comments about future DW_FORM_sdata/udata re-work of related code.
26207 2013-04-23 Lawrence Crowl <crowl@google.com>
26209 * Makefile.in: Update as needed below.
26211 * hash-table.h (class hash_table):
26212 Correct many methods with parameter types compare_type to the correct
26213 value_type. (Correct code was unlikely to notice the change.)
26214 (hash_table::elements_with_deleted) New.
26215 (class hashtable::iterator): New.
26216 (hashtable::begin()): New.
26217 (hashtable::end()): New.
26218 (FOR_EACH_HASH_TABLE_ELEMENT): New.
26220 * statistics.c (statistics_hashes):
26221 Change type to hash_table. Update dependent calls and types.
26223 * tree-into-ssa.c (var_infos):
26224 Change type to hash_table. Update dependent calls and types.
26226 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
26227 Change type to hash_table. Update dependent calls and types.
26229 * tree-ssa-loop-im.c (struct mem_ref.refs):
26230 Change type to hash_table. Update dependent calls and types.
26232 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
26233 Change type to hash_table. Update dependent calls and types.
26235 * tree-ssa-sccvn.c (vn_tables_s::nary):
26236 Change type to hash_table. Update dependent calls and types.
26237 (vn_tables_s::phis): Likewise.
26238 (vn_tables_s::references): Likewise.
26240 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
26241 (vn_reference_eq): Update parameter and return types.
26243 * tree-ssa-structalias.c (pointer_equiv_class_table):
26244 Change type to hash_table. Update dependent calls and types.
26245 (location_equiv_class_table): Likewise.
26247 * tree-vect-data-refs.c: Consequential changes for making
26248 peeling a hash_table.
26250 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
26251 (destroy_loop_vec_info): Dependent hash_table update.
26253 * tree-vectorizer.h (peeling_htab):
26254 Change type to hash_table. Update dependent calls and types.
26256 2013-04-23 Shiva Chen <shiva0217@gmail.com>
26258 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
26259 to check the register content is equal or not.
26260 * lra-constraints.c (match_reload): Use lra_assign_reg_val
26261 to assign register content record.
26262 * lra-eliminations.c (update_reg_eliminate): Use
26263 lra_update_reg_val_offset to update register content offset.
26264 * lra-int.h (struct lra_reg): Add offset member.
26265 (lra_reg_val_equal_p): New static inline function.
26266 (lra_update_reg_val_offset): New static inline function.
26267 (lra_assign_reg_val): New static inline function.
26268 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
26269 to assign register content record.
26270 (initialize_lra_reg_info_element): Initial offset to zero.
26272 2013-04-23 Catherine Moore <clm@codesourcery.com>
26274 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
26275 operands. Record compression.
26277 2013-04-23 Xinliang David Li <davidxl@google.com>
26279 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
26281 2013-04-23 Richard Biener <rguenther@suse.de>
26283 PR middle-end/57036
26284 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
26285 parameter, only add abnormal goto edges from the copied body
26286 if the call could perform abnormal gotos.
26287 (copy_cfg_body): Adjust.
26289 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
26291 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
26293 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
26295 * coretypes.h (gimple_stmt_iterator): Add struct to make
26298 2013-04-23 Richard Biener <rguenther@suse.de>
26300 PR tree-optimization/57026
26301 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
26302 from SSA names occuring in abnormal PHI nodes.
26304 2013-04-22 Andi Kleen <ak@linux.intel.com>
26306 * lto/lto.c (print_lto_report_1): Fix LTO report names.
26308 2013-04-22 Andi Kleen <ak@linux.intel.com>
26310 * lto/lto.c (print_lto_report_1): Declare early.
26311 (read_cgraph_and_symbols): Call print_lto_report_1 early.
26313 2013-04-22 Andi Kleen <ak@linux.intel.com>
26315 * common.opt (-flto-report-wpa): Add.
26316 * doc/invoke.texi (-flto-report-wpa): Add.
26317 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
26320 2013-04-22 Xinliang David Li <davidxl@google.com>
26322 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
26323 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
26324 * Makefile.in: New dependency
26326 David Daney <ddaney.cavm@gmail.com>
26328 * configure.ac (gcc_cv_as_micromips_support): Use the
26329 --fatal-warnings option.
26330 * configure: Regenerate.
26332 2013-04-22 Marek Polacek <polacek@redhat.com>
26335 * tsan.c (instrument_expr): Don't instrument expression
26336 in case its size is zero.
26338 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
26342 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26344 * config/alpha/alpha.c (TARGET_LRA_P): New define.
26346 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
26348 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
26349 (gimple_stmt_iterator): New typedef.
26350 * gimple.h (gimple_stmt_iterator): Rename to...
26351 (gimple_stmt_iterator_d): ... This.
26352 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
26353 trees be valid for GIMPLE and GENERIC.
26354 (TARGET_GIMPLE_FOLD_BUILTIN): New.
26355 * gimple-fold.c (gimple_fold_call): Call target hook
26356 gimple_fold_builtin.
26357 * hooks.c (hook_bool_gsiptr_false): New.
26358 * hooks.h (hook_bool_gsiptr_false): New.
26359 * target.def (fold_stmt): New.
26360 * doc/tm.texi: Regenerate.
26362 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
26365 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
26366 a set sp if no stack realignment.
26368 2013-04-22 Nick Clifton <nickc@redhat.com>
26370 * config.gcc (tilegx-linux): Extend extra_objs rather than
26372 (tilepro-linux): Likewise.
26374 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
26376 * config/aarch64/aarch64-builtins.c
26378 (CF0, CF1, CF2, CF3, CF4, CF10): New.
26379 (VAR<1-12>): Add MAP parameter.
26380 (BUILTIN_*): Likewise.
26381 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
26382 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
26383 (aarch64_ushl_n<mode>): Likewise.
26384 (aarch64_sshr_n<mode>): Likewise.
26385 (aarch64_ushr_n<mode>): Likewise.
26386 (aarch64_<maxmin><mode>): Likewise.
26387 (aarch64_sqrt<mode>): Likewise.
26388 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
26389 (vshr<q>_n_*): Likewise.
26391 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
26393 * config/aarch64/aarch64-builtins.c
26394 (aarch64_simd_builtin_type_mode): Handle SF types.
26396 (BUILTIN_GPF): Define.
26397 (aarch64_init_simd_builtins): Handle SF types.
26398 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
26399 (frecps): Likewise.
26400 (frecpx): Likewise.
26401 * config/aarch64/aarch64-simd.md
26402 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
26403 (aarch64_frecpe<mode>): New.
26404 (aarch64_frecps<mode>): Likewise.
26405 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
26406 (v8type): Add frecp<esx>.
26407 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
26408 (aarch64_frecps<mode>): Likewise.
26409 * config/aarch64/iterators.md (FRECP): New.
26410 (frecp_suffix): Likewise.
26411 * config/aarch64/arm_neon.h
26412 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
26414 2013-04-22 Christian Bruel <christian.bruel@st.com>
26417 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
26418 (REG_CLASS_NAMES): Idem.
26419 (REG_CLASS_CONTENTS): Idem.
26420 (REGCLASS_HAS_FP_REG): Idem.
26421 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
26422 (sh_conditional_register_usage): Idem.
26424 2013-04-21 Jeff Law <law@redhat.com>
26426 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
26427 (ssa_forward_propagate_and_combine): Use it.
26429 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
26431 * lra.c: Update the flow chart diagram.
26433 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
26435 PR rtl-optimization/56847
26436 * lra-constraints.c (process_alt_operands): Discourage alternative
26437 with non-matche doffsettable memory constraint fro memory with
26440 2013-04-19 Richard Biener <rguenther@suse.de>
26442 PR tree-optimization/56982
26443 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
26445 * gimplify.c (gimplify_call_expr): Notice special calls.
26446 (gimplify_modify_expr): Likewise.
26447 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
26448 abnormal control flow receivers.
26449 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
26450 in the same way as cfun->has_nonlocal_labels.
26451 (gimple_purge_dead_abnormal_call_edges): Likewise.
26452 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
26453 receivers start a basic-block.
26455 2013-04-19 Richard Biener <rguenther@suse.de>
26457 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
26459 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
26460 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
26461 (SLP_TREE_LOAD_PERMUTATION): Add.
26462 (vect_transform_slp_perm_load): Adjust prototype.
26463 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
26464 (vect_free_slp_instance): Likewise.
26465 (vect_create_new_slp_node): Likewise.
26466 (vect_supported_slp_permutation_p): Remove.
26467 (vect_slp_rearrange_stmts): Adjust.
26468 (vect_supported_load_permutation_p): Likewise. Inline
26469 vect_supported_slp_permutation_p here.
26470 (vect_analyze_slp_instance): Compute load permutations per
26471 slp node instead of per instance.
26472 (vect_get_slp_defs): Adjust.
26473 (vect_transform_slp_perm_load): Likewise.
26474 (vect_schedule_slp_instance): Remove redundant code.
26475 (vect_schedule_slp): Remove hack for PR56270, add it ...
26476 * tree-vect-stmts.c (vectorizable_load): ... here, do not
26477 CSE loads for SLP. Adjust.
26479 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
26481 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
26482 spelling in two comments.
26484 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
26487 * config/arm/arm.c (load_multiple_sequence): Require SP
26488 as base register for loads if SP is in the register list.
26490 2013-04-19 Martin Jambor <mjambor@suse.cz>
26492 PR tree-optimization/56718
26493 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
26494 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
26495 and made public. Adjusted all callers.
26496 (ipa_intraprocedural_devirtualization): New function.
26497 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
26498 (ipa_intraprocedural_devirtualization): Likewise.
26499 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
26501 2013-04-19 Richard Biener <rguenther@suse.de>
26503 PR tree-optimization/57000
26504 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
26506 2013-04-19 Terry Guo <terry.guo@arm.com>
26508 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
26510 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
26511 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
26512 (cortex_m4_fmacs): Use new reservations.
26513 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
26515 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
26517 PR rtl-optimization/56999
26518 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
26520 (lra_coalesce): Remove split_origin_bitmap and related code.
26521 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
26522 ranges if necessary.
26524 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
26526 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
26528 (ix86_expand_call): Remove clobbered_registers array and use
26529 x86_64_ms_sysv_extra_clobbered_registers instead.
26530 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
26532 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
26534 * config/i386/i386.md (*call_rex64_ms_sysv): Use
26535 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
26536 (*call_value_rex64_ms_sysv): Ditto.
26538 2013-04-18 Cary Coutant <ccoutant@google.com>
26540 * dwarf2out.c (output_pubnames): Check die_perennial_p of
26541 parent instead of die_mark.
26543 2013-04-18 Diego Novillo <dnovillo@google.com>
26545 * gimple.c (create_gimple_tmp): New.
26546 (get_expr_type): New.
26547 (build_assign): New.
26548 (build_type_cast): New.
26549 * gimple.h (enum ssa_mode): Define.
26550 (gimple_seq_set_location): New.
26551 * asan.c (build_check_stmt): Change some gimple_build_* calls
26552 to use build_assign and build_type_cast.
26554 2013-04-18 Richard Biener <rguenther@suse.de>
26556 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
26557 handle negative step. Remove redundant checks.
26558 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
26559 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
26560 for negative step and grouped loads fail to vectorize.
26562 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
26564 * emit-rtl.c (reset_insn_used_flags): New function.
26565 (reset_all_used_flags): Use it.
26566 (verify_insn_sharing): New function.
26567 (verify_rtl_sharing): Fix verification for SEQUENCEs.
26569 2013-04-18 Jakub Jelinek <jakub@redhat.com>
26571 PR tree-optimization/56984
26572 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
26573 and (x >> M) >= N don't register any assertion if N << M is the
26576 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
26578 * lower-subreg.c (resolve_simple_move): If called self-recursive,
26579 do not delete_insn insns that have not yet been emitted, only
26580 unlink them with remove_insn.
26581 * df-scan.c (df_insn_delete): Revert r197492.
26583 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
26585 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
26586 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
26588 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
26590 * config/arm/arm.md (movsicc_insn): Convert define_insn into
26591 define_insn_and_split.
26592 (and_scc,ior_scc,negscc): Likewise.
26593 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
26595 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
26597 * config/arm/arm.c (use_return_insn): Return 0 for targets that
26598 can benefit from using a sequence of LDRD instructions in epilogue
26599 instead of a single LDM instruction.
26601 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
26604 * doc/extend.texi: Fix typo.
26606 2013-04-17 Richard Biener <rguenther@suse.de>
26608 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
26609 (vect_build_slp_tree): ... here.
26610 (vect_build_slp_tree_1): Compute which stmts of the SLP group
26611 match. Remove special-casing of mismatched complex loads.
26612 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
26613 re-try the match with swapped commutative operands.
26614 (vect_supported_load_permutation_p): Remove special-casing of
26615 mismatched complex loads.
26616 (vect_analyze_slp_instance): Adjust.
26618 2013-04-17 Richard Biener <rguenther@suse.de>
26620 PR rtl-optimization/56921
26621 * cfgloop.h (struct loop): Add simple_loop_desc member.
26622 (struct niter_desc): Mark with GTY(()).
26623 (simple_loop_desc): Do not use aux field but simple_loop_desc.
26624 * loop-iv.c (get_simple_loop_desc): Likewise.
26625 (free_simple_loop_desc): Likewise.
26628 2013-04-16 Richard Biener <rguenther@suse.de>
26630 PR rtl-optimization/56921
26631 * loop-init.c (pass_rtl_move_loop_invariants): Add
26632 TODO_do_not_ggc_collect to todo_flags_finish.
26633 (pass_rtl_unswitch): Same.
26634 (pass_rtl_unroll_and_peel_loops): Same.
26635 (pass_rtl_doloop): Same.
26637 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
26639 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
26640 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
26641 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
26643 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
26644 * tree-streamer.c (record_common_node): Adjust reference in comment.
26646 2013-04-17 Terry Guo <terry.guo@arm.com>
26648 * config/arm/cortex-m4.md: Add a new bypass.
26650 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26652 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
26654 (*subs_<optab><mode>_multp2): New pattern.
26655 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
26656 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
26658 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26660 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
26661 (*subs_mul_imm_<mode>): New pattern.
26663 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
26666 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
26667 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
26668 (vsx_movti_32bit): Change j->wa to O->wa.
26670 2013-04-16 Richard Biener <rguenther@suse.de>
26672 PR rtl-optimization/56921
26673 * loop-init.c (pass_rtl_move_loop_invariants): Add
26674 TODO_do_not_ggc_collect to todo_flags_finish.
26675 (pass_rtl_unswitch): Same.
26676 (pass_rtl_unroll_and_peel_loops): Same.
26677 (pass_rtl_doloop): Same.
26679 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
26681 * config/arm/arm.c (emit_multi_reg_push): New declaration
26682 for an existing function.
26683 (arm_emit_strd_push): New function.
26684 (arm_expand_prologue): Used here.
26685 (arm_emit_ldrd_pop): New function.
26686 (arm_expand_epilogue): Used here.
26687 (arm_get_frame_offsets): Update condition.
26688 (arm_emit_multi_reg_pop): Add a special case for load of a single
26689 register with writeback.
26691 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
26693 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
26696 2013-04-16 Richard Biener <rguenther@suse.de>
26698 PR tree-optimization/56756
26699 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
26700 (first_mem_ref_loc): New.
26701 (execute_sm): Place the load temporarily before a previous
26702 access instead of in the latch edge to ensure its SSA dependencies
26703 are defined at points dominating the load.
26705 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
26707 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
26708 correct fix by moving header and footer insn to the footer of
26709 the merged basic block. Clear BB_END of the merged-away block.
26711 PR middle-end/43631
26712 * emit-rtl.c (make_note_raw): New function.
26713 (link_insn_into_chain): New static inline function.
26714 (add_insn): Use it.
26715 (add_insn_before, add_insn_after): Factor insn chain linking code...
26716 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
26717 using link_insn_into_chain.
26718 (note_outside_basic_block_p): New helper function for emit_note_after
26719 and emit_note_before.
26720 (emit_note_after): Use nobb variant of add_insn_after if the note
26721 should not be contained in a basic block.
26722 (emit_note_before): Use nobb variant of add_insn_before if the note
26723 should not be contained in a basic block.
26724 (emit_note_copy): Use make_note_raw.
26725 (emit_note): Likewise.
26726 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
26727 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
26728 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
26729 the moved barrier the tail of the basic block it follows.
26730 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
26732 2013-04-15 Jakub Jelinek <jakub@redhat.com>
26734 PR tree-optimization/56962
26735 * gimple-ssa-strength-reduction.c (record_increment): Only set
26736 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
26737 either rhs1 or rhs2 is equal to c->base_expr.
26739 2013-04-15 Richard Biener <rguenther@suse.de>
26741 PR tree-optimization/56933
26742 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
26744 (GROUP_READ_WRITE_DEPENDENCE): Remove.
26745 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
26746 * tree-vect-data-refs.c (vect_analyze_group_access): Move
26747 dependence check ...
26748 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
26750 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
26751 GROUP_READ_WRITE_DEPENDENCE.
26753 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26755 * emit-rtl.c (reset_all_used_flags): New function.
26756 (verify_rtl_sharing): Call reset_all_used_flags before and after
26757 performing the checks.
26759 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
26761 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
26762 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
26763 * config/arm/constraints.md (De): New constraint.
26764 * config/arm/neon.md (anddi3_neon): Delete.
26765 (neon_vand<mode>): Expand to standard anddi3 pattern.
26766 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
26767 Move earlier in the file.
26768 (neon_inv_logic_op2): Likewise.
26769 (arm_anddi_operand_neon): New predicate.
26771 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
26773 * configure.ac (gcc_cv_ld_as_needed): Set
26774 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
26775 Use -z ignore, -z record on *-*-solaris2*.
26776 (HAVE_LD_AS_NEEDED): Update comment.
26777 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
26778 * configure: Regenerate.
26779 * config.in: Regenerate.
26780 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
26781 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
26782 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
26783 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
26784 equivalents. Fix markup.
26785 * doc/tm.texi: Regenerate.
26787 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
26789 * config/i386/i386.opt: New option mstack-protector-guard=.
26790 * config/i386/i386-opts.h: Add enum stack_protector_guard.
26791 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
26792 TARGET_SSP_TLS_GUARD.
26793 * config/i386/i386.c (ix86_option_override_internal): Set
26794 ix86_stack_protector_guard.
26795 * config/i386/i386.md (stack_protect_set): Enable for
26796 TARGET_SSP_TLS_GUARD only.
26797 (stack_protect_set_<mode>): Ditto.
26798 (stack_protect_test): Ditto.
26799 (stack_protect_test_<mode>): Ditto.
26800 * doc/invoke.texi (i386 Option): Document.
26802 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
26805 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
26806 (S_MODES): Set H_MODE bit.
26807 (SF_MODES): Set only S_MODE and SF_MODE bits.
26808 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
26809 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
26810 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
26811 <MODE_FLOAT>: Likewise.
26813 2013-04-15 Joey Ye <joey.ye@arm.com>
26815 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
26817 2013-04-15 Joey Ye <joey.ye@arm.com>
26819 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
26821 (thumb_far_jump_used_p): Count instruction size and set
26824 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
26826 * reorg.c (fill_simple_delay_slots): Reindent block of code.
26827 * resource.c (mark_target_live_regs): Reformat conditional block.
26829 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
26831 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
26832 notes, they are emitted only just before final.
26833 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
26835 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
26837 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
26838 * cfgrtl.c (delete_insn): Call it here instead.
26839 * lra-spills.c (lra_final_code_change): Use delete_insn.
26840 * haifa-sched.c (sched_remove_insn): Likewise.
26841 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
26842 returning to the nop pool.
26843 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
26844 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
26846 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
26848 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
26849 * doc/tm.texi: Regenerated.
26851 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
26853 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
26856 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
26858 * df-core.c (df_find_def): Compare register numbers.
26859 (df_find_use): Likewise.
26861 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
26864 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
26865 lra_in_progress for return.
26867 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
26869 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
26870 define_insn into define_insn_and_split and emit movsicc patterns.
26872 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
26874 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
26876 2013-04-12 Richard Biener <rguenther@suse.de>
26878 * tree-pass.h (TODO_do_not_ggc_collect): New.
26879 * passes.c (execute_one_ipa_transform_pass): Honor
26880 TODO_do_not_ggc_collect.
26881 (execute_one_pass): Likewise.
26884 2013-04-10 Richard Biener <rguenther@suse.de>
26886 * passes.c (init_optimization_passes): Remove reload pass.
26887 * ira.c (do_reload): Merge into ...
26889 (rest_of_handle_reload): Remove.
26890 (pass_reload): Likewise.
26891 * config/i386/i386.c (ix86_option_override): Refer to ira instead
26892 of reload for vzeroupper pass placement.
26894 2013-04-12 Jakub Jelinek <jakub@redhat.com>
26896 PR tree-optimization/56918
26897 PR tree-optimization/56920
26898 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
26899 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
26900 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
26901 use wide_mul_with_sign method.
26903 2013-04-12 Richard Biener <rguenther@suse.de>
26905 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
26906 not be considered a gimple constant.
26908 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
26910 * fold-const.c (const_binop): Handle vector shifts by a scalar.
26911 (fold_binary_loc): Call const_binop also for mixed vector-scalar
26914 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
26915 Jakub Jelinek <jakub@redhat.com>
26917 * opts.c: Include diagnostic-color.h.
26918 (common_handle_option): Handle OPT_fdiagnostics_color_.
26919 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
26920 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
26921 (diagnostic-color.o): New.
26922 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
26923 (diagnostic_color_rule): New enum.
26924 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
26925 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
26926 the location string.
26927 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
26928 either NULL, or color kind.
26929 * diagnostic-color.c: New file.
26930 * diagnostic-color.h: New file.
26931 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
26933 * doc/invoke.texi (-fdiagnostics-color): Document.
26934 * pretty-print.h (pp_show_color): Define.
26935 (struct pretty_print_info): Add show_color field.
26936 * diagnostic.c: Include diagnostic-color.h.
26937 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
26938 macros. Colorize error:, warning: etc. strings and also the location
26940 (diagnostic_show_locus): Colorize the caret line.
26941 * pretty-print.c: Include diagnostic-color.h.
26942 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
26943 inside of %< %> quotes or quoted through q format modifier.
26945 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26947 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
26949 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26951 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
26952 code in CC_NZ mode.
26953 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
26956 2013-04-11 Marek Polacek <polacek@redhat.com>
26958 PR tree-optimization/48184
26959 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
26961 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
26963 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
26964 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
26965 (skip_simple_arithmetic): Tidy up.
26966 * tree.h (skip_simple_constant_arithmetic): Declare.
26968 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26970 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
26972 2013-04-11 Richard Biener <rguenther@suse.de>
26974 * tree-vect-loop.c (get_initial_def_for_induction): Properly
26975 generate vector constants.
26977 2013-04-11 Richard Biener <rguenther@suse.de>
26979 PR tree-optimization/56878
26980 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
26981 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
26983 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
26984 Prefer to align the DR with the most invariant base address.
26986 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
26988 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
26991 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
26993 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
26994 floating-point vector comparisons against 0.
26996 2013-04-11 Jakub Jelinek <jakub@redhat.com>
26998 PR tree-optimization/56899
26999 * fold-const.c (extract_muldiv_1): Apply distributive law
27000 only if TYPE_OVERFLOW_WRAPS (ctype).
27002 2013-04-11 Bin Cheng <bin.cheng@arm.com>
27005 * ira-costs.c (scan_one_insn): Check whether the source rtx of
27006 loading has side effect.
27008 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
27010 * config/sparc/sparc.c: Include tree-pass.h.
27011 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
27012 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
27013 head of file. Change return type. Split off gate function.
27014 (sparc_gate_work_around_errata): New function.
27015 (pass_work_around_errata): New pass definition.
27016 (insert_pass_work_around_errata) New pass insert definition to
27017 insert pass_work_around_errata just after delayed-branch scheduling.
27018 (sparc_option_override): Insert the pass.
27019 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
27021 2013-04-10 David S. Miller <davem@davemloft.net>
27023 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
27024 or -mcpu=hypersparc.
27026 * target.def (cstore_mode): New hook.
27027 * target.h: Include insn-codes.h
27028 * targhooks.c: Likewise.
27029 (default_cstore_mode): New function.
27030 * targhooks.h: Declare it.
27031 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
27032 * doc/tm.texi: Rebuild.
27033 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
27034 target hook, rather than inspecting the insn_data.
27035 * config/sparc/sparc.c (sparc_cstore_mode): New function.
27036 (TARGET_CSTORE_MODE): Redefine.
27037 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
27039 * config/sparc/predicates.md (cstore_result_operand): New special
27041 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
27042 Use it for operand 0.
27043 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
27044 (*snesi_special): Likewise.
27045 (*snesi_zero): Likewise.
27046 (*seqsi_zero): Likewise.
27047 (*sltu_insn): Likewise.
27048 (*sgeu_insn): Likewise.
27049 (*seqdi_special): Make operand 0 and comparison operation be of
27051 (*snedi_special): Likewise.
27052 (*snedi_special_vis3): Likewise.
27053 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
27054 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
27055 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
27056 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
27057 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
27058 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
27059 (*sltu_extend_sp64): Likewise.
27060 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
27061 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
27062 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
27063 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
27064 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
27066 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
27068 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
27069 (aarch64_start_file): Use the new function.
27071 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
27072 Jason Merrill <jason@redhat.com>
27074 * common.opt: Add -gdwarf.
27075 * opts.c (common_handle_option): Handle it.
27076 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
27078 2013-04-10 Richard Biener <rguenther@suse.de>
27080 * passes.c (execute_todo): Do not call ggc_collect conditional here.
27081 (execute_one_ipa_transform_pass): But unconditionally here.
27082 (execute_one_pass): And here.
27083 (init_optimization_passes): Remove reload pass.
27084 * tree-pass.h (TODO_ggc_collect): Remove.
27085 (pass_reload): Likewise.
27086 * ira.c (do_reload): Merge into ...
27088 (rest_of_handle_reload): Remove.
27089 (pass_reload): Likewise.
27090 * config/i386/i386.c (ix86_option_override): Refer to ira instead
27091 of reload for vzeroupper pass placement.
27092 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
27093 and todo_flags_finish of all passes.
27095 2013-04-10 Richard Biener <rguenther@suse.de>
27097 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
27098 first_const_oprnd field, rename first_def_type to first_op_type.
27099 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
27100 (vect_get_and_check_slp_defs): Always use the type of the
27101 operand. Allow mixed vect_external_def, vect_constant_def types.
27102 (vect_get_constant_vectors): Handle mixed vect_external_def,
27103 vect_constant_def types.
27105 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
27107 PR tree-optimization/55524
27108 * tree-ssa-math-opts.c
27109 (convert_mult_to_fma): Don't use an fms construct
27110 when we don't have an fms operation, but fnma, and it looks
27111 likely that we'll be able to use the latter.
27113 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
27115 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
27117 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
27118 inline fail caused by overwritable functions.
27120 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
27122 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
27123 unnecessary bits in the constant power of two case.
27125 2013-04-10 Richard Biener <rguenther@suse.de>
27127 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
27128 broken code swapping operands.
27129 (vect_build_slp_tree): Do not compute load permutations here.
27130 (vect_analyze_slp_instance): Compute load permutations here,
27131 after building the SLP tree.
27133 2013-04-09 Christian Bruel <christian.bruel@st.com>
27135 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
27136 of next/prev_real_insn.
27138 2013-04-09 Jan Hubicka <jh@suse.cz>
27140 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
27141 Drop aliased parameter.
27142 (function_and_variable_visibility): Do not handle alias pairs.
27143 * cgraph.c (varpool_externally_visible_p): Update prototype.
27144 * varpool.c (varpool_add_new_variable): Update.
27146 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
27148 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
27150 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
27152 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
27154 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
27156 2013-04-09 Marek Polacek <polacek@redhat.com>
27158 PR tree-optimization/48762
27159 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
27161 2013-04-09 Richard Biener <rguenther@suse.de>
27163 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
27165 (vect_build_slp_tree): Likewise.
27166 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
27167 calculating the cost of a SLP instance.
27168 (vect_analyze_slp_instance): Use it from here, after building
27171 2013-04-09 Jakub Jelinek <jakub@redhat.com>
27173 PR middle-end/56883
27174 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
27175 expand_omp_for_static_chunk): Use simple_p = true in
27176 force_gimple_operand_gsi calls when assigning to addressable decls.
27178 2013-04-09 Jeff Law <law@redhat.com>
27180 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
27181 when the boolean was created by converting a wider object which
27182 had a boolean range.
27184 2013-04-09 Richard Biener <rguenther@suse.de>
27186 * tree-vectorizer.h (slp_void_p): Remove.
27187 (slp_tree): Typedef before _slp_tree declaration.
27188 (struct _slp_tree): Use a vector of slp_tree as children.
27189 (vect_get_place_in_interleaving_chain): Remove.
27190 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
27192 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
27194 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
27195 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
27196 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
27197 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
27198 Use slp_node instead of slp_void_p and adjust.
27200 2013-04-09 Richard Biener <rguenther@suse.de>
27202 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
27203 work that is not necessary.
27205 2013-04-09 Jakub Jelinek <jakub@redhat.com>
27207 PR tree-optimization/56854
27208 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
27209 forward into clobber stmts if it would change MEM_REF lhs into
27212 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
27214 * tree.c (type_hash_lookup, type_hash_add): Make static.
27215 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
27217 2013-04-09 Richard Biener <rguenther@suse.de>
27219 * tree.h (unsave_expr_now): Remove.
27220 * tree-inline.c (mark_local_for_remap_r): Remove.
27221 (unsave_expr_1): Likewise.
27222 (unsave_r): Likewise.
27223 (unsave_expr_now): Likewise.
27224 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
27225 (propagate_tree_value): Likewise.
27227 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
27229 * doc/rtl.texi (sequence): Rewrite documentation to match the
27230 current use of SEQUENCE rtl objects.
27231 * rtl.def (SEQUENCE): Likewise.
27233 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
27234 Update documentation.
27235 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
27236 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
27238 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
27240 2013-04-08 Teresa Johnson <tejohnson@google.com>
27242 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
27243 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
27245 (estimate_edge_size_and_time): Add comment to suggest using rounding
27247 (estimate_node_size_and_time): Ditto.
27248 (remap_edge_change_prob): Use helper rounding divide methods.
27249 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
27250 (gimple_mod_pow2_value_transform): Ditto.
27251 (gimple_mod_subtract_transform): Ditto.
27252 (gimple_ic_transform): Ditto.
27253 (gimple_stringops_transform): Ditto.
27254 * stmt.c (conditional_probability): Ditto.
27255 (emit_case_dispatch_table): Ditto.
27256 * lto-cgraph.c (merge_profile_summaries): Ditto.
27257 * tree-optimize.c (execute_fixup_cfg): Ditto.
27258 * cfgcleanup.c (try_forward_edges): Ditto.
27259 * cfgloopmanip.c (scale_loop_profile): Ditto.
27261 (duplicate_loop_to_header_edge): Ditto.
27262 (lv_adjust_loop_entry_edge): Ditto.
27263 * tree-vect-loop.c (vect_transform_loop): Ditto.
27264 * profile.c (compute_branch_probabilities): Ditto.
27265 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
27266 * lto-streamer-in.c (input_cfg): Ditto.
27267 * gimple-streamer-in.c (input_bb): Ditto.
27268 * ipa-cp.c (update_profiling_info): Ditto.
27269 (update_specialized_profile): Ditto.
27270 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
27271 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
27273 * sched-rgn.c (compute_dom_prob_ps): Ditto.
27274 (compute_trg_info): Ditto.
27275 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
27276 (purge_dead_edges): Ditto.
27277 * loop-unswitch.c (unswitch_loop): Ditto.
27278 * cgraphclones.c (cgraph_clone_edge): Ditto.
27279 (cgraph_clone_node): Ditto.
27280 * tree-inline.c (copy_bb): Ditto.
27281 (copy_edges_for_bb): Ditto.
27282 (initialize_cfun): Ditto.
27283 (copy_cfg_body): Ditto.
27284 (expand_call_inline): Ditto.
27286 2013-04-08 Kai Tietz <ktietz@redhat.com>
27288 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
27289 TARGET_CYGWIN64 by TARGET_64BIT.
27291 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
27293 * config/epiphany/epiphany.md (GPR_1): New constant.
27294 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
27295 * config/epiphany/epiphany.c (gen_compare_reg):
27296 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
27297 is already in place.
27298 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
27299 Don't require being called during rtl expansion; If y operlaps r0,
27301 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
27302 (epiphany_expand_epilogue): Likewise.
27304 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
27305 Don't use CC_FPmode for ORDERED / UNORDERED.
27306 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
27308 * config/epiphany/constraints.md (CnL): New constraint.
27309 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
27310 * config/epiphany/predicates.md (add_operand): Allow 1024.
27312 * config/epiphany/epiphany.md (logical_op): New code iterator.
27313 (op_mnc): New code attribute.
27314 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
27315 (mov_f+1, mov_f+2): New peephole2 patterns.
27317 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
27318 (cstoresi4): Also allow re-use of zero result when doing a NE
27319 comparison to a non-zero operand.
27320 Use (clobber (scratch)) for first insn if the gpr output is not needed.
27322 * config/epiphany/epiphany.md (<insn_opname>v2si3):
27323 Use gen_addsi3_i / gen_subsi3_i.
27325 2013-04-08 Jakub Jelinek <jakub@redhat.com>
27329 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
27330 contain anything but clobbers, at most one __builtin_stack_restore,
27331 optionally debug stmts and final resx, and if it has at least one
27332 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
27333 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
27334 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
27335 which isn't defaut definition, remove them.
27336 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
27337 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
27338 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
27339 with MEM_REF LHS with SSA_NAME address.
27341 2013-04-08 Jeff Law <law@redhat.com>
27343 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
27345 2013-04-08 Richard Biener <rguenther@suse.de>
27347 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
27349 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
27350 determined vector type.
27351 (vect_analyze_data_refs): Likewise.
27352 (vect_get_new_vect_var): Adjust.
27353 (vect_create_destination_var): Preserve SSA name versions.
27354 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
27355 not dump anything here.
27357 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
27359 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
27360 Add member lr_slot_known.
27361 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
27363 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
27364 Remove code that sets lr_slot_offset according to what a previous
27365 version of epiphany_emit_save_restore used to do.
27366 (epiphany_emit_save_restore): When doing an lr save or restore,
27367 set/verify lr_slot_known and lr_slot_offset.
27369 2013-04-08 Xinyu Qi <xyqi@marvell.com>
27372 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
27375 2013-04-08 Richard Biener <rguenther@suse.de>
27377 * alias.c (find_base_term): Fix thinko in previous change.
27379 2013-04-08 Jakub Jelinek <jakub@redhat.com>
27381 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
27382 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
27383 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
27384 if possible to compute val.
27385 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
27386 For QImode integers don't require anything about precision. Use
27387 const_with_all_bytes_same to find out if the constant doesn't have
27388 repeated bytes in it.
27390 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
27392 * config/s390/s390.c (s390_expand_insv): Only accept insertions
27395 2013-04-08 Marek Polacek <polacek@redhat.com>
27397 PR rtl-optimization/48182
27398 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
27401 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27404 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
27405 nuses, make sure we have a label.
27407 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
27410 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
27411 (rs6000_emit_swdiv_low_precision): Remove.
27412 (rs6000_emit_swdiv): Rewrite to handle between one and four
27413 iterations of Newton-Raphson generally; modify required number of
27414 iterations for some cases.
27415 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
27417 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
27419 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
27420 set-but-unused variable.
27422 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
27423 basic blocks of released function bodies garbage-collectable.
27425 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
27426 (struct rtl_opt_pass): Add TODO_df_finish.
27428 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
27430 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
27432 * config/arm/constraints.md (q): New constraint.
27433 * config/arm/ldrdstrd.md: New file.
27434 * config/arm/arm.md (ldrdstrd.md) New include.
27435 (arm_movdi): Use "q" instead of "r" constraint
27436 for double-word memory access.
27437 (movdf_soft_insn): Likewise.
27438 * config/arm/vfp.md (movdi_vfp): Likewise.
27439 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
27440 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
27441 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
27442 (mem_ok_for_ldrd_strd): Likewise.
27443 (output_move_double): Update assertion.
27445 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
27447 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
27449 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
27451 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
27452 define_insn_and_split.
27453 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
27455 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
27457 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
27458 define_insn_and_split.
27459 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
27460 (shiftsi3_compare): New pattern.
27461 (rrx): New pattern.
27462 * config/arm/unspecs.md (UNSPEC_RRX): New.
27464 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
27466 * config/arm/arm.md (negdi_extendsidi): New pattern.
27467 (negdi_zero_extendsidi): Likewise.
27469 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
27471 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
27472 define_insn_and_split.
27473 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
27474 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
27476 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
27478 * config/arm/arm.md (arm_subdi3): Convert define_insn into
27479 define_insn_and_split.
27480 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
27481 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
27483 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
27485 * config/arm/arm.md (subsi3_carryin): New pattern.
27486 (subsi3_carryin_const): Likewise.
27487 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
27488 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
27490 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
27492 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
27494 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
27496 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
27497 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
27499 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
27501 * config/arm/arm.c (arm_expand_builtin): Change fcode
27502 type to unsigned int.
27504 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
27506 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
27508 2013-04-04 Ian Lance Taylor <iant@google.com>
27510 * doc/standards.texi (Standards): The Go frontend supports the Go 1
27513 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
27515 PR middle-end/56729
27516 * df-scan.c (df_insn_delete): Disable failing assert.
27518 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
27520 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
27521 New function prototype.
27522 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
27523 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
27524 (arm_builtin_vectorized_function): New function.
27526 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
27528 * config/arm/arm_neon_builtins.def: New file.
27529 * config/arm/arm.c (neon_builtin_data): Move contents to
27530 arm_neon_builtins.def.
27531 (enum arm_builtins): Include neon builtin definitions.
27532 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
27533 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
27535 2013-04-04 Marek Polacek <polacek@redhat.com>
27537 PR tree-optimization/48186
27538 * predict.c (maybe_hot_frequency_p): Return false if
27539 HOT_BB_FREQUENCY_FRACTION is 0.
27540 (cgraph_maybe_hot_edge_p): Likewise.
27542 2013-04-04 Richard Biener <rguenther@suse.de>
27544 PR tree-optimization/56826
27545 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
27548 2013-04-04 Richard Biener <rguenther@suse.de>
27550 PR tree-optimization/56213
27551 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
27552 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
27554 2013-04-04 Richard Biener <rguenther@suse.de>
27556 PR tree-optimization/56837
27557 * tree-loop-distribution.c (classify_partition): For non-zero
27558 values require that the value has the same precision as its
27559 mode to be useful as memset value.
27561 2013-04-03 Nick Clifton <nickc@redhat.com>
27563 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
27564 (fmssf4): Use fmsf.s on E3V5 architectures.
27565 (fnmasf4): Use fnmaf.s on E3V5 architectures.
27566 (fnmssf4): Use fnmsf.s on E3V5 architectures.
27568 2013-04-03 Jeff Law <law@redhat.com>
27570 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
27571 (lra-eliminations.o): Likewise.
27573 2013-04-03 Teresa Johnson <tejohnson@google.com>
27575 * gcov-io.c (compute_working_sets): Moved most of body of old
27576 compute_working_sets here from profile.c.
27577 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
27578 (gcov_working_set_t): Moved typedef here from basic-block.h
27579 (compute_working_set): Declare.
27580 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
27581 (get_working_sets): Renamed from compute_working_set,
27582 replace most of body with call to new compute_working_sets.
27583 (get_exec_counts): Replace call to compute_working_sets
27584 to get_working_sets.
27585 * profile.h (get_working_sets): Renamed from compute_working_set.
27586 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
27587 to get_working_sets.
27588 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
27589 * gcov-dump.c (dump_working_sets): New function.
27591 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
27593 * hwint.c (sext_hwi, zext_hwi): New functions.
27594 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
27595 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
27596 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
27597 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
27598 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
27599 (sext_hwi, zext_hwi): New functions.
27601 2013-04-03 Jeff Law <law@redhat.com>
27603 PR tree-optimization/56799
27604 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
27605 back test for widening conversion erroneously dropped in prior change.
27607 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
27610 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
27611 instead of next_real_insn.
27613 2013-04-03 Marek Polacek <polacek@redhat.com>
27616 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
27618 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
27621 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
27623 (thumb1_output_casesi): Likewise.
27624 (thumb2_output_casesi): Likewise.
27626 2013-04-03 Richard Biener <rguenther@suse.de>
27628 PR tree-optimization/56817
27629 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
27631 (tree_unroll_loops_completely_1): ... new function to manually
27632 walk the loop tree, properly defering outer loops of unrolled
27633 loops to later iterations.
27635 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
27637 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
27638 (vectorizable_load): Likewise.
27639 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
27640 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
27642 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
27644 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
27647 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
27649 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
27651 2013-04-03 Bin Cheng <bin.cheng@arm.com>
27653 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
27655 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
27657 PR tree-optimization/56790
27658 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
27661 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
27663 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
27665 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
27666 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
27669 2013-04-03 Jakub Jelinek <jakub@redhat.com>
27672 * tree.h (force_folding_builtin_constant_p): New decl.
27673 * builtins.c (force_folding_builtin_constant_p): New variable.
27674 (fold_builtin_constant_p): Fold immediately also if
27675 force_folding_builtin_constant_p.
27677 2013-04-03 Richard Biener <rguenther@suse.de>
27679 PR tree-optimization/56812
27680 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
27681 DRs of the same interleaving chain are independent.
27683 2013-04-02 Jason Merrill <jason@redhat.com>
27685 * gdbinit.in (pbb): Use debug fn.
27687 2013-04-02 Lawrence Crowl <crowl@google.com>
27689 * sese.h (struct ivtype_map_elt_s): Remove unused.
27690 (extern debug_ivtype_map): Remove unused.
27691 (extern eq_ivtype_map_elts): Remove unused.
27692 * sese.c (debug_ivtype_map): Removed unused.
27693 (debug_ivtype_map_1): Removed unused.
27694 (debug_ivtype_elt): Remove unused.
27695 (eq_ivtype_map_elts): Remove unused.
27697 2013-04-02 Kai Tietz <ktietz@redhat.com>
27700 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
27701 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
27702 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
27704 (legitimize_pe_coff_symbol): Likewise.
27705 (is_imported_p): New helper-function.
27706 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
27707 for Windows x64 targets.
27708 (ix86_expand_prologue): Optimize for pe-coff targets.
27709 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
27710 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
27711 medium/large code-model.
27712 (legitimize_pic_address): Likewise.
27713 (legitimize_tls_address): Likewise.
27714 (ix86_expand_call): Likewise.
27715 (x86_output_mi_thunk): Likewise.
27716 (get_dllimport_decl): Add new beimport argument.
27717 (construct_plt_address): Don't assert for x64 pe-coff targets.
27718 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
27720 (SYMBOL_FLAG_STUBVAR): New macro.
27721 (SYMBOL_REF_STUBVAR_P): Likewise.
27722 * config/i386/winnt.c (stub_list): New structure.
27723 (stub_head): New local variable.
27724 (i386_pe_record_stub): New function.
27725 (i386_pe_file_end): Emit refptr-stubs.
27727 2013-04-02 Jakub Jelinek <jakub@redhat.com>
27729 PR rtl-optimization/56745
27730 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
27731 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
27734 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
27735 and both of them are MEM_REFs, just compare first argument for
27736 equality and attempt to deal even with differing offsets.
27739 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
27740 of gimple_clobber_p to be MEM_REF.
27741 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
27742 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
27743 after gimplification.
27744 * asan.c (get_mem_ref_of_assignment): Don't instrument
27745 gimple_clobber_p stmts.
27746 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
27747 gimple_clobber_p stmt if they have MEM_REF lhs and
27748 are dead because of another gimple_clobber_p stmt.
27749 * tree-ssa-live.c (clear_unused_block_pointer): Treat
27750 gimple_clobber_p stmts like debug stmts.
27751 (remove_unused_locals): Remove clobbers with MEM_REF lhs
27752 that refer to unused VAR_DECLs or uninitialized values.
27753 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
27754 gimple_clobber_p stmts if they refer to removed parameters.
27755 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
27758 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
27760 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
27761 using SWI48 mode attribute.
27763 2013-04-02 Wei Mi <wmi@google.com>
27765 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
27766 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
27767 *<rotate_insn><mode>3_mask in i386.md.
27769 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
27771 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
27773 2013-04-02 Richard Biener <rguenther@suse.de>
27775 PR tree-optimization/56778
27776 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
27777 Runtime alias tests are not supported for gather loads.
27778 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
27779 stmts referenced from SSA operands before updating SSA form.
27781 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
27782 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
27784 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
27785 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
27786 * config/arm/cortex-a53.md: New file.
27787 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
27788 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
27789 * config/arm/arm.c (arm_issue_rate): Likewise.
27790 * config/arm/arm-tune.md: Regenerate
27791 * config/arm/arm-tables.opt: Regenerate.
27792 * config/arm/arm-cores.def: Add cortex-a53.
27794 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
27796 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
27799 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
27801 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
27802 scalar load/store operations using B/H registers.
27803 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
27805 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
27807 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
27809 * config/aarch64/aarch64.c
27810 (aarch64_simd_scalar_immediate_valid_for_move): New.
27811 * config/aarch64/aarch64-protos.h
27812 (aarch64_simd_scalar_immediate_valid_for_move): New.
27813 * config/aarch64/constraints.md (Dh, Dq): New.
27814 * config/aarch64/iterators.md (hq): New.
27816 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
27818 * reorg.c (get_branch_condition): Deal with conditional returns.
27819 (fill_simple_delay_slots): Remove dead code dealing with jumps.
27821 2013-04-01 Wei Mi <wmi@google.com>
27823 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
27824 Truncate operand 2 using %b asm operand modifier.
27825 (*<shift_insn><mode>3_mask): Ditto.
27826 (*<rotate_insn><mode>3_mask): Ditto.
27828 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
27830 PR middle-end/56798
27831 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
27833 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
27835 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
27837 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
27839 2013-03-30 Lawrence Crowl <crowl@google.com>
27841 * dse.c (clear_alias_sets): Remove never set.
27842 (disqualified_clear_alias_sets): Remove never set.
27843 (clear_alias_mode_pool): Remove never set.
27844 (dse_step0): Remove condition that is never true.
27845 (canon_address): Remove condition that is never true.
27846 (dse_step7): Remove condition that is never true.
27847 (rest_of_handle_dse): Remove condition that is never true.
27848 (rest_of_handle_dse::did_global): Remove never read from above.
27849 (dse_step2_spill): Remove never called from above.
27850 (dse_step5_spill): Remove never called from above.
27852 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
27854 * doc/md.texi (Standard Names) <casesi>: Update documentation for
27855 JUMP_TABLE_DATA changes.
27856 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
27857 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
27858 (Insns) <jump_table_data>: New entry.
27859 * doc/tm.texi: Regenerate.
27861 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
27863 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
27864 for table jump at the end of a basic block using tablejump_p.
27865 * targhooks.c (default_invalid_within_doloop): Likewise.
27866 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
27867 target hook implementation that is identical to the default hook.
27868 (rs6000_invalid_within_doloop): Remove.
27870 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
27871 unused variable from tablejump_p call.
27873 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
27874 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
27875 (INSN_DELETED_P): Likewise.
27876 (emit_jump_table_data): New prototype.
27877 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
27878 after 4th as unused.
27879 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
27880 * sched-vis.c (print_insn): Likewise.
27881 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
27882 insn for compatibility with back ends that use next_active_insn to
27883 identify jump table data.
27884 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
27885 (remove_insn): Likewise.
27886 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
27888 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
27889 (emit_jump_table_data): New function.
27891 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
27892 basic block, a JUMP_TABLE_DATA never is.
27893 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
27894 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
27895 off from code handling real insns.
27896 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
27897 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
27898 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
27899 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
27900 is not a NONDEBUG_INSN_P.
27901 * ira-costs.c (scan_one_insn): Likewise.
27902 * jump.c (mark_all_labels): Likewise.
27903 (mark_jump_label_1): Likewise.
27904 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
27905 * lra.c (get_insn_freq): Expect all insns reaching here to be in
27907 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
27908 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
27909 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
27910 JUMP_TABLE_DATA_P insns.
27911 (calculate_elim_costs_all_insns): Likewise.
27912 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
27913 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
27914 (delete_output_reload): Code style fixups.
27915 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
27916 insn flags on this non-insn.
27917 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
27918 as scheduling barriers, for pre-change compatibility.
27919 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
27920 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
27922 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
27923 redundant JUMP_TABLE_DATA_P test.
27924 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
27925 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
27926 (frv_for_each_packet): Likewise.
27927 * config/i386/i386.c (min_insn_size): Likewise.
27928 (ix86_avoid_jump_mispredicts): Likewise.
27929 * config/m32r/m32r.c (m32r_is_insn): Likewise.
27930 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
27931 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
27932 (mips16_insn_length): Robustify.
27933 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
27934 (mips16_split_long_branches): Likewise.
27935 * config/pa/pa.c (pa_combine_instructions): Likewise.
27936 * config/rs6000/rs6000.c (get_next_active_insn): Treat
27937 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
27938 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
27939 as contributing to pool range lengths.
27940 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
27941 Remove redundant JUMP_TABLE_DATA_P test.
27942 (sh_loop_align): Likewise.
27943 (split_branches): Likewise.
27944 (sh_insn_length_adjustment): Likewise.
27945 * config/spu/spu.c (get_branch_target): Likewise.
27947 2013-03-29 Jan Hubicka <jh@suse.cz>
27949 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
27950 gcov streaming; stream hot bb threshold to ltrans.
27951 * predict.c (get_hot_bb_threshold): Break out from ....
27952 (maybe_hot_count_p): ... here.
27953 (set_hot_bb_threshold): New function.
27954 * lto-section-in.c (lto_section_name): Add profile.
27955 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
27956 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
27957 and data-streamer.h
27958 (histogram_entry): New structure.
27959 (histogram, histogram_pool): New global vars.
27960 (histogram_hash): New structure.
27961 (histogram_hash::hash): New method.
27962 (histogram_hash::equal): Likewise.
27963 (account_time_size): New function.
27964 (cmp_counts): New function.
27965 (dump_histogram): New function.
27966 (ipa_profile_generate_summary): New function.
27967 (ipa_profile_write_summary): New function.
27968 (ipa_profile_read_summary): New function.
27969 (ipa_profile): Decide on threshold.
27970 (pass_ipa_profile): Add ipa_profile_write_summary and
27971 ipa_profile_read_summary.
27972 * Makefile.in (ipa.o): Update dependencies.
27973 * lto-streamer.h (LTO_section_ipa_profile): New section.
27975 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
27977 * tree.h (VAR_P): New.
27979 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
27982 * doc/invoke.texi ([-fwhole-program]): Fix typo.
27984 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
27986 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
27987 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
27988 (control_flow_insn_p): Likewise.
27989 * cfgrtl.c (duplicate_insn_chain): Likewise.
27990 * final.c (get_attr_length_1): Likewise.
27991 (shorten_branches): Likewise.
27992 (final_scan_insn): Likewise.
27993 * function.c (instantiate_virtual_regs): Likewise.
27994 * gcse.c (insert_insn_end_basic_block): Likewise.
27995 * ira-costs.c (scan_one_insn): Likewise.
27996 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
27997 * lra.c (check_rtl): Likewise.
27998 * reload1.c (elimination_costs_in_insn): Likewise.
27999 * reorg.c (follow_jumps): Likewise.
28001 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
28002 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
28003 (thumb_far_jump_used_p): Likewise.
28004 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
28005 (workaround_speculation): Likewise.
28006 (add_sched_insns_for_speculation): Likewise.
28007 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
28008 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
28009 (frv_for_each_packet): Likewise.
28010 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
28011 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
28012 (final_emit_insn_group_barriers): Likewise.
28013 * config/m32r/m32r.c (m32r_is_insn): Likewise.
28014 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
28015 (mips16_insn_length): Likewise.
28016 * config/pa/pa.c (pa_reorg): Likewise.
28017 (pa_combine_instructions): Likewise.
28018 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
28019 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
28020 (sh_reorg): Likewise.
28021 (split_branches): Likewise.
28022 * config/spu/spu.c (get_branch_target): Likewise.
28024 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
28027 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
28029 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
28030 Fix declaration name.
28032 2013-03-28 Lawrence Crowl <crowl@google.com>
28034 * graphds.h (struct graph.indicies): Remove unused.
28035 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
28036 (SCOP_ORIGINAL_PDDRS): Remove unused.
28037 * sese.h (extern insert_loop_close_phis): Removed unused.
28038 (extern insert_guard_phis): Removed unused.
28039 (extern ivtype_map_elt_info): Removed unused.
28040 (new_ivtype_map_elt): Removed unused.
28041 * sese.c (ivtype_map_elt_info): Removed unused.
28043 2013-03-28 Lawrence Crowl <crowl@google.com>
28045 * Makefile.in: Add several missing include dependences.
28047 (test-dump.o): New. This object is not added to any executable,
28048 but is present for ad-hoc testing.
28050 (debug (const bitmap_head_def &)): New.
28051 (debug (const bitmap_head_def *)): New.
28053 (extern debug (const bitmap_head_def &)): New.
28054 (extern debug (const bitmap_head_def *)): New.
28056 (debug (edge_def &)): New.
28057 (debug (edge_def *)): New.
28059 (debug (basic_block_def &)): New.
28060 (debug (basic_block_def *)): New.
28062 (dump_node (const_tree, int, FILE *)): Correct source file.
28064 (debug (die_struct &)): New.
28065 (debug (die_struct *)): New.
28067 (extern debug (die_struct &)): New.
28068 (extern debug (die_struct *)): New.
28069 * gimple-pretty-print.c
28070 (debug (gimple_statement_d &)): New.
28071 (debug (gimple_statement_d *)): New.
28072 * gimple-pretty-print.h
28073 (extern debug (gimple_statement_d &)): New.
28074 (extern debug (gimple_statement_d *)): New.
28076 (debug (ira_allocno_copy &)): New.
28077 (debug (ira_allocno_copy *)): New.
28078 (debug (ira_allocno &)): New.
28079 (debug (ira_allocno *)): New.
28081 (extern debug (ira_allocno_copy &)): New.
28082 (extern debug (ira_allocno_copy *)): New.
28083 (extern debug (ira_allocno &)): New.
28084 (extern debug (ira_allocno *)): New.
28086 (debug (live_range &)): New.
28087 (debug (live_range *)): New.
28089 (debug (lra_live_range &)): New.
28090 (debug (lra_live_range *)): New.
28092 (debug (lra_live_range &)): New.
28093 (debug (lra_live_range *)): New.
28095 (debug (omega_pb_d &)): New.
28096 (debug (omega_pb_d *)): New.
28098 (extern debug (omega_pb_d &)): New.
28099 (extern debug (omega_pb_d *)): New.
28101 (debug (const rtx_def &)): New.
28102 (debug (const rtx_def *)): New.
28104 (debug_tree (tree): Move within file.
28105 (debug_raw (const tree_node &)): New.
28106 (debug_raw (const tree_node *)): New.
28107 (dump_tree_via_hooks (const tree_node *, int)): New.
28108 (debug (const tree_node &)): New.
28109 (debug (const tree_node *)): New.
28110 (debug_verbose (const tree_node &)): New.
28111 (debug_verbose (const tree_node *)): New.
28112 (debug_head (const tree_node &)): New.
28113 (debug_head (const tree_node *)): New.
28114 (debug_body (const tree_node &)): New.
28115 (debug_body (const tree_node *)): New.
28116 (debug_vec_tree (tree): Move and reimplement in terms of dump.
28117 (debug (vec<tree, va_gc> &)): New.
28118 (debug (vec<tree, va_gc> *)): New.
28120 (extern debug (const rtx_def &)): New.
28121 (extern debug (const rtx_def *)): New.
28123 (debug_raw (simple_bitmap_def &)): New.
28124 (debug_raw (simple_bitmap_def *)): New.
28125 (debug (simple_bitmap_def &)): New.
28126 (debug (simple_bitmap_def *)): New.
28128 (extern debug (simple_bitmap_def &)): New.
28129 (extern debug (simple_bitmap_def *)): New.
28130 (extern debug_raw (simple_bitmap_def &)): New.
28131 (extern debug_raw (simple_bitmap_def *)): New.
28133 (debug (vinsn_def &)): New.
28134 (debug (vinsn_def *)): New.
28135 (debug_verbose (vinsn_def &)): New.
28136 (debug_verbose (vinsn_def *)): New.
28137 (debug (expr_def &)): New.
28138 (debug (expr_def *)): New.
28139 (debug_verbose (expr_def &)): New.
28140 (debug_verbose (expr_def *)): New.
28141 (debug (vec<rtx> &)): New.
28142 (debug (vec<rtx> *)): New.
28144 (extern debug (vinsn_def &)): New.
28145 (extern debug (vinsn_def *)): New.
28146 (extern debug_verbose (vinsn_def &)): New.
28147 (extern debug_verbose (vinsn_def *)): New.
28148 (extern debug (expr_def &)): New.
28149 (extern debug (expr_def *)): New.
28150 (extern debug_verbose (expr_def &)): New.
28151 (extern debug_verbose (expr_def *)): New.
28152 (extern debug (vec<rtx> &)): New.
28153 (extern debug (vec<rtx> *)): New.
28155 (_list_iter_cond_expr): Make inline instead of static.
28157 (debug (sreal &)): New.
28158 (debug (sreal *)): New.
28160 (extern debug (sreal &)): New.
28161 (extern debug (sreal *)): New.
28163 (extern debug_raw (const tree_node &)): New.
28164 (extern debug_raw (const tree_node *)): New.
28165 (extern debug (const tree_node &)): New.
28166 (extern debug (const tree_node *)): New.
28167 (extern debug_verbose (const tree_node &)): New.
28168 (extern debug_verbose (const tree_node *)): New.
28169 (extern debug_head (const tree_node &)): New.
28170 (extern debug_head (const tree_node *)): New.
28171 (extern debug_body (const tree_node &)): New.
28172 (extern debug_body (const tree_node *)): New.
28173 (extern debug (vec<tree, va_gc> &)): New.
28174 (extern debug (vec<tree, va_gc> *)): New.
28176 (debug (struct loop &)): New.
28177 (debug (struct loop *)): New.
28178 (debug_verbose (struct loop &)): New.
28179 (debug_verbose (struct loop *)): New.
28180 * tree-dump.c: Add header dependence.
28182 (extern debug (struct loop &)): New.
28183 (extern debug (struct loop *)): New.
28184 (extern debug_verbose (struct loop &)): New.
28185 (extern debug_verbose (struct loop *)): New.
28187 (debug (data_reference &)): New.
28188 (debug (data_reference *)): New.
28189 (debug (vec<data_reference_p> &)): New.
28190 (debug (vec<data_reference_p> *)): New.
28191 (debug (vec<ddr_p> &)): New.
28192 (debug (vec<ddr_p> *)): New.
28194 (extern debug (data_reference &)): New.
28195 (extern debug (data_reference *)): New.
28196 (extern debug (vec<data_reference_p> &)): New.
28197 (extern debug (vec<data_reference_p> *)): New.
28198 (extern debug (vec<ddr_p> &)): New.
28199 (extern debug (vec<ddr_p> *)): New.
28201 (debug (pt_solution &)): New.
28202 (debug (pt_solution *)): New.
28204 (extern debug (pt_solution &)): New.
28205 (extern debug (pt_solution *)): New.
28207 (debug (_var_map &)): New.
28208 (debug (_var_map *)): New.
28209 (debug (tree_live_info_d &)): New.
28210 (debug (tree_live_info_d *)): New.
28212 (extern debug (_var_map &)): New.
28213 (extern debug (_var_map *)): New.
28214 (extern debug (tree_live_info_d &)): New.
28215 (extern debug (tree_live_info_d *)): New.
28217 2013-03-28 Jan Hubicka <jh@suse.cz>
28219 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
28221 2013-03-28 Ian Bolton <ian.bolton@arm.com>
28223 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
28224 record only when desired or required.
28226 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
28228 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
28229 *vec_extractv2di_1_rex64. Use x64 isa attribute.
28231 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
28233 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
28234 (*andsi3_compare0_uxtw): New pattern.
28235 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
28236 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
28238 2013-03-28 Jan Hubicka <jh@suse.cz>
28240 * data-streamer-in.c (streamer_read_gcov_count): New function.
28241 * gimple-streamer-out.c: Include value-prof.h.
28242 (output_gimple_stmt): Output histogram.
28243 (output_bb): Use streamer_write_gcov_count.
28244 * value-prof.c: Include data-streamer.h
28245 (dump_histogram_value): Add HIST_TYPE_MAX.
28246 (stream_out_histogram_value): New function.
28247 (stream_in_histogram_value): New function.
28248 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
28249 (stream_out_histogram_value, stream_in_histogram_value): Declare.
28250 * data-streamer-out.c (streamer_write_gcov_count): New function.
28251 (streamer_write_gcov_count_stream): New function.
28252 * lto-cgraph.c (lto_output_edge): Update counter streaming.
28253 (lto_output_node): Likewise.
28254 (input_node, input_edge): Likewise.
28255 * lto-streamer-out.c (output_cfg): Update streaming.
28256 * lto-streamer-in.c (input_cfg): Likewise.
28257 * data-streamer.h (streamer_write_gcov_count,
28258 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
28259 * gimple-streamer-in.c: Include value-prof.h
28260 (input_gimple_stmt): Input histograms.
28261 (input_bb): Update profile streaming.
28263 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
28265 * genmodes.c (emit_max_int): New function.
28266 (emit_insn_modes_h): Added call to emit_max_function.
28267 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
28269 * machmode.def: Fixed comment.
28271 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
28273 * combine.c (try_combine): Removed useless assert.
28274 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
28276 2013-03-28 Marek Polacek <polacek@redhat.com>
28277 Richard Biener <rguenther@suse.de>
28279 PR tree-optimization/56695
28280 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
28281 build signed result of a vector comparison.
28282 * tree-cfg.c (verify_gimple_comparison): Check that a result
28283 of a vector comparison has signed type.
28285 2013-03-28 Richard Biener <rguenther@suse.de>
28287 PR tree-optimization/37021
28288 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
28289 do not restrict gaps between groups.
28290 * tree-vect-stmts.c (vectorizable_load): Properly account for
28291 a gap between groups.
28293 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
28295 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
28296 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
28299 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
28301 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
28302 * doc/extend.texi (Named Address Spaces): Ditto.
28303 (Variable Attributes): Ditto.
28305 2013-03-27 Kai Tietz <ktietz@redhat.com>
28307 * config.build: Add support for cygwin x64 target.
28308 * config.gcc: Likewise.
28309 * config.host: Likewise.
28310 * configure.ac: Likewise
28311 * configure: Regenerated.
28313 2013-03-27 Kai Tietz <ktietz@redhat.com>
28315 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
28316 * config/i386/t-cygwin-w64: New file.
28317 * config/i386/cygwin-w64.h: New file.
28318 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
28319 and add support for x64-cygwin target.
28320 (CPP_SPEC): Likewise.
28321 (CXX_WRAP_SPEC_LIST): Undefine before define.
28322 (LIBGCJ_SONAME): Use 15 as version.
28324 2013-03-27 Richard Biener <rguenther@suse.de>
28326 PR tree-optimization/56716
28327 * tree-ssa-structalias.c (perform_var_substitution): Adjust
28328 dumping for ref nodes.
28330 2013-03-27 Martin Jambor <mjambor@suse.cz>
28332 PR tree-optimization/55334
28333 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
28334 restricted pointers to arrays.
28336 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
28338 * Makefile.in (.SUFFIXES): Add .cc.
28339 (.c.o): Apply same recipe for implicit rule .cc.o.
28341 2013-03-27 Richard Biener <rguenther@suse.de>
28343 PR tree-optimization/37021
28344 * tree-vect-data-refs.c (vect_check_strided_load): Allow
28345 REALPART/IMAGPART_EXPRs around the supported refs.
28346 * tree-ssa-structalias.c (find_func_aliases): Assume that
28347 floating-point values are not used to transfer pointers.
28349 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
28351 * target.def (TARGET_HAS_IFUNC_P): New target hook.
28352 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
28353 * doc/tm.texi: Regenerate.
28354 * targhooks.h (default_has_ifunc_p): New.
28355 * targhooks.c (default_has_ifunc_p): Ditto.
28356 * config/linux-protos.h: New file.
28357 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
28358 hook for linux which disables support of indirect functions in android.
28359 * config/linux-android.c: New file.
28360 * config/t-linux-android.c: Ditto.
28361 * config.gcc: Added new object file linux-android.o.
28362 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
28363 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
28364 * varasm.c (do_assemble_alias): Likewise.
28365 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
28366 doesn't support indirect functions.
28367 * configure: Regenerate.
28369 2013-03-27 Bin Cheng <bin.cheng@arm.com>
28372 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
28373 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
28376 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
28378 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
28380 2013-03-27 Terry Guo <terry.guo@arm.com>
28382 * config/arm/arm-cores.def: Added core cortex-r7.
28383 * config/arm/arm-tune.md: Regenerated.
28384 * config/arm/arm-tables.opt: Regenerated.
28385 * doc/invoke.texi: Added entry for core cortex-r7.
28387 2013-03-27 Walter Lee <walt@tilera.com>
28389 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
28390 double-decrement of next_scratch_regno.
28392 2013-03-27 Walter Lee <walt@tilera.com>
28394 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
28396 (insn_v1mulus): Ditto.
28397 (insn_v2muls): Ditto.
28399 2013-03-27 Walter Lee <walt@tilera.com>
28401 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
28402 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
28404 2013-03-27 Walter Lee <walt@tilera.com>
28406 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
28407 (*sibcall_value): Ditto.
28409 2013-03-27 Walter Lee <walt@tilera.com>
28411 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
28412 (insn_mnz_v8qi): ... this ...
28413 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
28415 (insn_v<n>mnz): Replaced by ...
28416 (insn_v1mnz): ... this ...
28417 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
28419 (insn_mz_<mode>): Replaced by ...
28420 (insn_mz_v8qi): ... this ...
28421 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
28423 (insn_v<n>mz): Replaced by ...
28424 (insn_v1mz): ... this ...
28425 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
28428 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
28430 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
28432 2013-03-26 Roland McGrath <mcgrathr@google.com>
28434 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
28435 than fprintf with a non-constant, non-format string.
28437 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
28439 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
28440 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
28441 operand 0 predicate.
28442 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
28443 attribute. Use general_x64nomem_operand as operand 1 predicate.
28444 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
28445 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
28446 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
28447 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
28448 (mov<mode>_insv_1): Remove expander. Merge insn with
28449 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
28450 Use general_x64nomem_operand as operand 1 predicate.
28451 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
28452 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
28453 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
28454 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
28455 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
28456 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
28457 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
28458 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
28459 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
28460 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
28461 (general_x64nomem_operand): Ditto.
28463 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
28465 * config/rtems.opt: Add -pthread option.
28467 2013-03-26 Richard Biener <rguenther@suse.de>
28469 * alias.c (find_base_term): Avoid redundant and not used recursion.
28470 (base_alias_check): Get the initial base term from the caller.
28471 (true_dependence_1): Compute and pass base terms to base_alias_check.
28472 (write_dependence_p): Likewise.
28473 (may_alias_p): Likewise.
28475 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
28477 * config/aarch64/aarch64.c (aarch64_classify_address): Support
28478 PC-relative load in SI modes and above only.
28480 2013-03-26 Xinyu Qi <xyqi@marvell.com>
28482 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
28483 * config/arm/iwmmxt.md (WCGR0): Update.
28484 (WCGR1, WCGR2, WCGR3): Likewise.
28486 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
28488 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
28489 Use x64 and nox64 isa attributes.
28491 2013-03-26 Richard Biener <rguenther@suse.de>
28493 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
28494 alignment computations and rely on get_object_alignment_1
28495 for the !TYPE_P case.
28496 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
28498 2013-03-26 Walter Lee <walt@tilera.com>
28500 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
28501 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
28503 2013-03-25 Jeff Law <law@redhat.com>
28505 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
28506 check for INTEGRAL_TYPE_P that was missing due to checking in
28507 wrong version of prior patch.
28509 2013-03-25 Walter Lee <walt@tilera.com>
28511 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
28512 TILEGX_INSN_SHUFFLEBYTES1.
28513 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
28515 (tilegx_builtins): Ditto.
28516 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
28518 2013-03-25 Walter Lee <walt@tilera.com>
28520 * config/tilegx/tilegx.md (floatsisf2): New pattern.
28521 (floatunssisf2): New pattern.
28522 (floatsidf2): New pattern.
28523 (floatunssidf2): New pattern.
28525 2013-03-25 Walter Lee <walt@tilera.com>
28527 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
28528 tests for constraint J, K, N, P.
28530 2013-03-25 Walter Lee <walt@tilera.com>
28532 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
28533 Use indirect/pcrel encoding.
28534 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
28537 2013-03-25 Steve Ellcey <sellcey@mips.com>
28539 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
28540 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
28541 * config/mips/mips.c (mips_option_override): Set IMADD default.
28542 * config/mips/mips.h (PTF_AVOID_IMADD): New.
28543 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
28544 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
28545 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
28546 * doc/invoke.texi (-mimadd/-mno-imadd): New.
28548 2013-03-25 Jeff Law <law@redhat.com>
28550 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
28551 slightly to avoid creating and folding useless trees. Simplify
28552 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
28554 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
28556 * config/i386/i386.md (*zero_extendsidi2): Merge with
28557 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
28558 * config/i386/predicates.md (x86_64_zext_operand): Rename from
28559 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
28560 targets. Clarify comment.
28562 2013-03-25 Martin Jambor <mjambor@suse.cz>
28564 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
28565 pass-through jump functions differently.
28566 (ipa_read_jump_function): Likewise. Also use setter functions to set
28569 2013-03-25 Martin Jambor <mjambor@suse.cz>
28571 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
28572 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
28574 (ipa_get_indirect_edge_target): New function.
28575 (devirtualization_time_bonus): New parameter known_aggs, pass it to
28576 ipa_get_indirect_edge_target. Update all callers.
28577 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
28578 ipa_get_indirect_edge_target_1 instead of calling
28579 ipa_get_indirect_edge_target.
28580 (create_specialized_node): Pass aggvlas to
28581 ipcp_discover_new_direct_edges.
28583 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
28585 * config/arm/arm.md (f_sels, f_seld): New types.
28586 (*cmov<mode>): New pattern.
28587 * config/arm/predicates.md (arm_vsel_comparison_operator): New
28590 2013-03-25 Kai Tietz <ktietz@redhat.com>
28592 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
28593 POSIX-printf for mingw-hosted builds.
28595 2013-03-25 Richard Biener <rguenther@suse.de>
28597 PR middle-end/56694
28598 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
28599 must-not-throw stmt location.
28601 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
28603 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
28604 Emit load-acquire versions when acq is true.
28605 (arm_emit_store_exclusive): Add rel parameter.
28606 Emit store-release versions when rel is true.
28607 (arm_split_compare_and_swap): Use acquire-release instructions
28609 of barriers when appropriate.
28610 (arm_split_atomic_op): Likewise.
28611 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
28612 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
28613 (VUNSPEC_SLX): Likewise.
28614 (VUNSPEC_LDA): Likewise.
28615 (VUNSPEC_STL): Likewise.
28616 * config/arm/sync.md (atomic_load<mode>): New pattern.
28617 (atomic_store<mode>): Likewise.
28618 (arm_load_acquire_exclusive<mode>): Likewise.
28619 (arm_load_acquire_exclusivesi): Likewise.
28620 (arm_load_acquire_exclusivedi): Likewise.
28621 (arm_store_release_exclusive<mode>): Likewise.
28623 2013-03-25 Catherine Moore <clm@codesourcery.com>
28625 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
28626 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
28627 * config/mip/predicates.md (lwsp_swsp_operand,
28628 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
28629 sb16_operand, db4_operand, db7_operand, ib3_operand,
28630 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
28631 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
28632 andi16_operand): New predicates.
28633 * config/mips/mips.md (compression): New attribute.
28634 (enabled): New attribute.
28635 (length): Consider compression in computing length.
28636 (shift_compression): New code attribute.
28637 (*add<mode>3): New operands. Record compression.
28638 (sub<mode>3): Likewise.
28639 (one_cmpl<mode>2): Likewise.
28640 (*and<mode>3): Likewise.
28641 (*ior<mode>3): Likewise.
28642 (unnamed pattern for xor): Likewise.
28643 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
28644 (*<optab><mode>3): Likewise.
28645 (*mov<mode>_internal: Likewise.
28646 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
28647 (mips_unsigned_immediate_p): New.
28648 (umips_lwsp_swsp_address_p): New.
28649 (m16_based_address_p): New.
28650 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
28651 (mips_unsigned_immediate_p): New prototype.
28652 (lwsp_swsp_address_p): New prototype.
28653 (m16_based_address_p): New prototype.
28654 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
28655 (mips_signed_immediate_p): New function.
28656 (m16_based_address_p): New function.
28657 (lwsp_swsp_address_p): New function.
28658 (mips_print_operand_punctuation): Recognize short delay slot insns
28659 for microMIPS.add<mode>3"
28661 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
28664 * config/arm/iterators.md (v_cmp_result): New mode attribute.
28665 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
28667 2013-03-25 Richard Biener <rguenther@suse.de>
28669 PR tree-optimization/56689
28670 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
28673 2013-03-25 Richard Biener <rguenther@suse.de>
28675 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
28677 (memory_references): Likewise.
28678 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
28679 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
28680 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
28681 (gather_mem_refs_in_loops): Fold into ...
28682 (analyze_memory_references): ... this. Move initialization
28683 to tree_ssa_lim_initialize.
28684 (fill_always_executed_in): Rename to ...
28685 (fill_always_executed_in_1): ... this.
28686 (fill_always_executed_in): Move contains_call computation to
28687 this new function from ...
28688 (tree_ssa_lim_initialize): ... here.
28689 (tree_ssa_lim): Call fill_always_executed_in.
28691 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
28693 * postreload.c (reload_combine): Fix code detecting returns.
28695 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
28697 * function.c (emit_use_return_register_into_block): On cc0 targets,
28698 do not emit the sequence between cc0 setter and user.
28700 2013-03-25 Kai Tietz <ktietz@redhat.com>
28702 * config/i386/predicates.md (local_symbolic_operand): Interpret
28703 dll-imported symbols as none-local.
28705 2013-03-25 Richard Biener <rguenther@suse.de>
28707 * tree-ssa-loop-im.c (struct depend): Remove.
28708 (struct lim_aux_data): Make depends a vec of gimples.
28709 (free_lim_aux_data): Adjust.
28710 (add_dependency): Likewise.
28711 (set_level): Likewise.
28713 2013-03-25 Richard Biener <rguenther@suse.de>
28715 PR middle-end/56434
28716 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
28717 the pointer returned by calls with ECF_MALLOC set.
28719 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
28721 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
28723 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
28725 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
28726 using MMXMODE mode iterator.
28727 (*move<mode>_internal): Merge with *movv2sf_internal and
28728 *movv2sf_internal_rex64 using MMXMODE mode iterator.
28730 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
28732 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
28733 (record_last_mem_set_info): Likewise.
28735 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
28736 of XNEWVEC followed by memset.
28737 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
28739 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
28741 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
28742 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
28743 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
28744 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
28745 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
28746 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
28747 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
28748 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
28749 BARRIER_P instead of GET_CODE.
28751 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
28753 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
28754 inaccuracy in the probing code.
28756 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
28757 (ctrapdi4): Likewise.
28759 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
28761 * calls.c (expand_call): Add missing guard to code handling return
28762 of non-BLKmode structures in MSB.
28763 * function.c (expand_function_end): Likewise.
28765 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
28767 * combine.c (try_combine): Adjust comment. Do not add the set of
28768 insn #0 if the destination indirectly is set or dies in insn #2.
28769 Tidy up code to distribute a new note.
28771 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
28773 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
28774 also for alternatives 16 and 17.
28776 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
28778 * config/i386/sse.md (*mov<mode>_internal): Merge with
28779 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
28780 Emit insn template depending on type attribute. Use
28781 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
28782 movd instead of movq mnemonic for interunit moves. Rewrite mode
28783 attribute calculation. Remove unit attribute calculation.
28784 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
28785 Set prefix_data16 attribute for DImode ssemov types.
28786 Use Ym instead of y for SSE-MMX conversion alternatives.
28787 Reorder operand constraints.
28789 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
28791 * df.h (df_insn_delete): Adjust prototype.
28792 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
28793 and let it decide whether mark the basic block dirty.
28794 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
28795 * df-scan.c (df_insn_info_delete): New helper function, split
28796 off from df_insn_delete.
28797 (df_scan_free_bb_info): Use it.
28798 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
28800 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
28801 that the insn is actually an insn and it has a non-NULL basic block.
28802 Do not mark basic block dirty if only deleting a DEBUG_INSN.
28804 2013-03-22 Richard Biener <rguenther@suse.de>
28806 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
28808 (mem_ref_alloc): Do not allocate them.
28809 (refs_independent_p): Do not query or maintain a cache.
28811 2013-03-22 Richard Biener <rguenther@suse.de>
28813 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
28814 (gather_mem_refs_in_loops): Do not compute it.
28815 (analyze_memory_references): Do not allocate it.
28816 (tree_ssa_lim_finalize): Do not free it.
28817 (for_all_locs_in_loop): Do not query all_refs_in_loop.
28819 2013-03-22 Richard Biener <rguenther@suse.de>
28821 * is-a.h (as_a): Use gcc_checking_assert.
28823 2013-03-22 Ian Bolton <ian.bolton@arm.com>
28825 * config/aarch64/aarch64.c (aarch64_print_operand): New
28826 format specifier for printing a constant in hex.
28827 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
28828 format specifier for printing second operand.
28830 2013-03-22 Richard Biener <rguenther@suse.de>
28832 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
28834 (gather_mem_refs_in_loops): Perform store accumulation here.
28835 (create_vop_ref_mapping_loop): Remove.
28836 (create_vop_ref_mapping): Likewise.
28837 (analyze_memory_references): Initialize refs_stored_in_loop.
28838 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
28839 (record_indep_loop): Remove.
28840 (record_dep_loop): New function.
28841 (ref_indep_loop_p_1): Adjust to only walk over references
28842 in the loop, not its subloops.
28843 (ref_indep_loop_p): Rename to ...
28844 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
28845 maintaining a more fine-grained cache.
28846 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
28847 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
28849 2013-03-22 Richard Biener <rguenther@suse.de>
28851 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
28852 (struct mem_ref): Make accesses_in_loop a vec of a vec of
28853 aggregate mem_ref_loc.
28854 (free_mem_ref_locs): Inline into ...
28855 (memref_free): ... this and adjust.
28856 (mem_ref_alloc): Adjust.
28857 (mem_ref_locs_alloc): Remove.
28858 (record_mem_ref_loc): Adjust.
28859 (get_all_locs_in_loop): Rewrite into ...
28860 (for_all_locs_in_loop): ... this iterator.
28861 (rewrite_mem_ref_loc): New functor.
28862 (rewrite_mem_refs): Use for_all_locs_in_loop.
28863 (sm_set_flag_if_changed): New functor.
28864 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
28865 (ref_always_accessed): New functor.
28866 (ref_always_accessed_p): Use for_all_locs_in_loop.
28868 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
28870 * tree-pass.h (PROP_gimple_lvec): New.
28871 * passes.c (dump_properties): Handle PROP_gimple_lvec.
28872 (init_optimization_passes): Move pass_lower_vector.
28873 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
28875 (pass_lower_vector): Provide PROP_gimple_lvec.
28876 (pass_lower_vector_ssa): Likewise.
28877 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
28879 2013-03-21 Mark Wielaard <mjw@redhat.com>
28881 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
28883 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
28885 * config/i386/i386.md (*movdi_internal): Disparage slightly
28886 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
28887 conversion alternatives.
28889 2013-03-21 Jakub Jelinek <jakub@redhat.com>
28891 PR middle-end/48087
28892 * diagnostic.def (DK_WERROR): New kind.
28893 * diagnostic.h (werrorcount): Define.
28894 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
28895 promoted to DK_ERROR, increment DK_WERROR counter instead of
28897 * toplev.c (toplev_main): Call print_ignored_options even if
28898 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
28899 even if just werrorcount is non-zero.
28902 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
28904 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
28905 (string_cst_pool_decl): New function.
28906 (optimize_one_addr_into_implicit_ptr): New function.
28907 (resolve_addr_in_expr): Optimize DWARF location expression
28908 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
28909 which doesn't live in memory, but has DW_AT_location or
28910 DW_AT_const_value, or refers to a string literal, into
28911 DW_OP_GNU_implicit_pointer.
28912 (optimize_location_into_implicit_ptr): New function.
28913 (resolve_addr): If removing DW_AT_location of a variable because
28914 it was DW_OP_addr of address of the variable, but the variable doesn't
28915 live in memory, try to emit const value attribute for the initializer.
28917 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
28919 * tree.h (VECTOR_TYPE_P): New macro.
28920 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
28921 TYPE_MODE): Use it.
28922 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
28923 VEC_COND_EXPR cannot be lvalues.
28924 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
28926 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
28928 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
28929 Restrict the transformation to equal modes.
28931 2013-03-21 Richard Biener <rguenther@suse.de>
28933 PR tree-optimization/39326
28934 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
28935 (MEM_ANALYZABLE): Adjust.
28936 (record_mem_ref_loc): Move bitmap ops ...
28937 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
28938 unanalyzable refs, do not record locations for it.
28939 (analyze_memory_references): Allocate ref zero as shared
28941 (refs_independent_p): Do not test for unanalyzed mems here.
28942 (ref_indep_loop_p_1): Special-case disambiguation against
28943 the unanalyzed ref.
28944 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
28946 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
28948 * config/arm/arm-protos.h (tune_params): Add
28949 prefer_neon_for_64bits field.
28950 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
28951 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
28952 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
28953 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
28954 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
28955 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
28956 (arm_option_override): Handle -mneon-for-64bits new option.
28957 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
28958 (prefer_neon_for_64bits): Declare new variable.
28959 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
28960 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
28961 (arch_enabled): Handle new arch types. Remove support for onlya8
28963 (one_cmpldi2): Use new arch names.
28964 (zero_extend<mode>di2, extend<mode>di2): Ditto.
28965 * config/arm/arm.opt (mneon-for-64bits): Add option.
28966 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
28967 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
28968 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
28970 * doc/invoke.texi (-mneon-for-64bits): Document.
28972 2013-03-21 Richard Biener <rguenther@suse.de>
28974 PR tree-optimization/39326
28975 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
28976 (sort_bbs_in_loop_postorder_cmp): New function.
28977 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
28979 2013-03-21 Richard Biener <rguenther@suse.de>
28981 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
28982 (vect_insert_into_interleaving_chain): Likewise.
28983 (vect_drs_dependent_in_basic_block): Inline ...
28984 (vect_slp_analyze_data_ref_dependence): ... here. New function,
28986 (vect_analyze_data_ref_dependence): ... here. Simplify.
28987 (vect_check_interleaving): Simplify.
28988 (vect_analyze_data_ref_dependences): Likewise. Split out ...
28989 (vect_slp_analyze_data_ref_dependences): ... this new function.
28990 (dr_group_sort_cmp): New function.
28991 (vect_analyze_data_ref_accesses): Compute data-reference groups
28992 here instead of in vect_analyze_data_ref_dependence. Use
28993 a more efficient algorithm.
28994 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
28995 vect_slp_analyze_data_ref_dependences. Call
28996 vect_analyze_data_ref_accesses earlier.
28997 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
28998 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
28999 (vect_slp_analyze_data_ref_dependences): New prototype.
29001 2013-03-21 Richard Biener <rguenther@suse.de>
29003 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
29004 ref is stored in the loop.
29005 (find_refs_for_sm): Walk only over all stores.
29006 (store_motion_loop): Allocate from lim_bitmap_obstack.
29007 (store_motion): Likewise.
29009 2013-03-21 Richard Biener <rguenther@suse.de>
29011 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
29012 Update virtual SSA form.
29014 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
29016 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
29017 * configure: Regenerate.
29018 * config.in: Regenerate.
29019 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
29020 if !HAVE_LD_EH_FRAME_CIEV3.
29022 2013-03-21 Richard Biener <rguenther@suse.de>
29024 * tree-cfg.c (verify_expr_no_block): New function.
29025 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
29026 nor DECL_VALUE_EXPR have locations with associated blocks.
29027 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
29028 (clear_unused_block_pointer): Remove code dealing with
29029 blocks in DECL_DEBUG_EXPR locations.
29031 2013-03-21 Richard Biener <rguenther@suse.de>
29033 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
29034 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
29035 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
29036 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
29037 instead of DECL_DEBUG_EXPR_IS_FROM.
29038 * gimplify.c (gimplify_modify_expr): Likewise.
29039 * tree-cfg.c (verify_expr_location_1): Likewise.
29040 * tree-complex.c (create_one_component_var): Likewise.
29041 * tree-sra.c (create_access_replacement): Likewise.
29042 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
29043 (clear_unused_block_pointer): Likewise.
29044 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
29045 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
29046 * var-tracking.c (var_debug_decl): Likewise.
29047 (track_expr_p): Likewise.
29048 * tree-inline.c (add_local_variables): Likewise. Set
29049 DECL_HAS_DEBUG_EXPR_P after copying it.
29050 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
29051 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
29053 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
29056 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
29057 * configure: Regenerate.
29058 * config.in: Regenerate.
29059 * config/i386/i386.md (*movdf_internal): Use
29060 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
29061 movd instead of movq mnemonic for interunit moves.
29062 (*movdi_internal): Ditto.
29064 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29066 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
29067 (abd<mode>_3): New pattern.
29068 (aba<mode>_3): New pattern.
29069 (fabd<mode>_3): New pattern.
29071 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29073 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
29074 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
29075 occurrence of REGISTER_PREFIX as its empty string.
29077 2013-03-20 Jeff Law <law@redhat.com>
29079 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
29080 addititional equivalences for equality comparisons between an SSA_NAME
29081 and a constant where the SSA_NAME was set from a widening conversion.
29083 2013-03-20 Walter Lee <walt@tilera.com>
29085 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
29087 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
29089 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
29090 depending on type attribute.
29091 (*movti_internal): Ditto.
29092 (*movtf_internal): Ditto.
29093 (*movxf_internal): Ditto.
29094 (*movdf_internal): Ditto.
29095 (*movsf_internal): Ditto.
29097 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
29099 * config/i386/i386.md (*movti_internal): Set prefix attribute to
29100 maybe_vex for sselog1 and ssemov types.
29101 (*movdi_internal): Reorder operand constraints.
29102 (*movsi_internal): Ditto. Set prefix attribute to
29103 maybe_vex for sselog1 and ssemov types.
29104 (*movtf_internal): Set prefix attribute to maybe_vex
29105 for sselog1 and ssemov types.
29106 (*movdf_internal): Ditto. Set prefix_data16 attribute for
29107 DImode ssemov types. Reorder operand constraints.
29108 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
29109 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
29110 attribute for SImode ssemov types. Reorder operand constraints.
29112 2013-03-20 Martin Jambor <mjambor@suse.cz>
29114 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
29115 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
29117 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
29119 * config/rs6000/predicates.md (indexed_address, update_address_mem
29120 update_indexed_address_mem): New predicates.
29121 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
29122 attribute for load/store instructions.
29123 * config/rs6000/dfp.md (movsd_store): Likewise.
29124 (movsd_load): Likewise.
29125 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
29126 (unnamed HI->DI extend define_insn): Likewise.
29127 (unnamed SI->DI extend define_insn): Likewise.
29128 (unnamed QI->SI extend define_insn): Likewise.
29129 (unnamed QI->HI extend define_insn): Likewise.
29130 (unnamed HI->SI extend define_insn): Likewise.
29131 (unnamed HI->SI extend define_insn): Likewise.
29132 (extendsfdf2_fpr): Likewise.
29133 (movsi_internal1): Likewise.
29134 (movsi_internal1_single): Likewise.
29135 (movhi_internal): Likewise.
29136 (movqi_internal): Likewise.
29137 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
29138 attribute for load/store instructions.
29139 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
29141 (mov<mode>_softfloat): Likewise.
29142 (mov<mode>_hardfloat32): Likewise.
29143 (mov<mode>_hardfloat64): Likewise.
29144 (mov<mode>_softfloat64): Likewise.
29145 (movdi_internal32): Likewise.
29146 (movdi_internal64): Likewise.
29147 (probe_stack_<mode>): Likewise.
29149 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
29151 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
29152 floating point, and decimal floating point to reload iterator.
29154 * config/rs6000/constraints.md (wl constraint): New constraints to
29155 return FLOAT_REGS if certain options are used to reduce the number
29156 of separate patterns that exist in the file.
29157 (wx constraint): Likewise.
29158 (wz constraint): Likewise.
29160 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
29161 -mdebug=reg, print wg, wl, wx, and wz constraints.
29162 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
29163 Initialize the reload functions for 64-bit binary/decimal floating
29165 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
29166 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
29167 create the buffer on the stack to overcome not having a 32-bit
29169 (rs6000_emit_move): Likewise.
29170 (rs6000_secondary_memory_needed_rtx): Likewise.
29171 (rs6000_alloc_sdmode_stack_slot): Likewise.
29172 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
29173 via xxlxor, just like DFmode 0.0.
29175 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
29176 define as 1 if we are running on a power7 or newer.
29177 (enum r6000_reg_class_enum): Add new constraints.
29179 * config/rs6000/dfp.md (movsd): Delete, combine with binary
29180 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
29181 with other moves by using conditional constraits (wg). Use LFIWZX
29182 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
29183 (movsd splitter): Likewise.
29184 (movsd_hardfloat): Likewise.
29185 (movsd_softfloat): Likewise.
29187 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
29188 binary and decimal floating point moves.
29189 (fmove_ok): New attributes to combine binary and decimal floating
29190 point moves, and to combine power6x (mfpgpr) moves along normal
29192 (real_value_to_target): Likewise.
29193 (f32_lr): Likewise.
29194 (f32_lm): Likewise.
29195 (f32_li): Likewise.
29196 (f32_sr): Likewise.
29197 (f32_sm): Likewise.
29198 (f32_si): Likewise.
29199 (movsf): Combine binary and decimal floating point moves. Combine
29200 power6x (mfpgpr) moves with other moves by using conditional
29201 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
29202 (mov<mode> for SFmode/SDmode); Likewise.
29203 (SFmode/SDmode splitters): Likewise.
29204 (movsf_hardfloat): Likewise.
29205 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
29206 (movsf_softfloat): Likewise.
29207 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
29209 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
29210 wx and wz constraints.
29212 * config/rs6000/constraints.md (wg constraint): New constraint to
29213 return FLOAT_REGS if -mmfpgpr (power6x) was used.
29215 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
29218 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
29219 -mdebug=reg, print wg, wl, wx, and wz constraints.
29220 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
29221 Initialize the reload functions for 64-bit binary/decimal floating
29223 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
29224 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
29225 create the buffer on the stack to overcome not having a 32-bit
29227 (rs6000_emit_move): Likewise.
29228 (rs6000_secondary_memory_needed_rtx): Likewise.
29229 (rs6000_alloc_sdmode_stack_slot): Likewise.
29230 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
29231 via xxlxor, just like DFmode 0.0.
29233 * config/rs6000/dfp.md (movdd): Delete, combine with binary
29234 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
29235 with other moves by using conditional constraits (wg). Use LFIWZX
29236 and STFIWX for loading SDmode on power7.
29237 (movdd splitters): Likewise.
29238 (movdd_hardfloat32): Likewise.
29239 (movdd_softfloat32): Likewise.
29240 (movdd_hardfloat64_mfpgpr): Likewise.
29241 (movdd_hardfloat64): Likewise.
29242 (movdd_softfloat64): Likewise.
29244 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
29245 64-bit binary and decimal floating point moves.
29246 (FMOVE64X): Likewise.
29247 (movdf): Combine 64-bit binary and decimal floating point moves.
29248 Combine power6x (mfpgpr) moves with other moves by using
29249 conditional constraits (wg).
29250 (mov<mode> for DFmode/DDmode): Likewise.
29251 (DFmode/DDmode splitters): Likewise.
29252 (movdf_hardfloat32): Likewise.
29253 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
29254 (movdf_softfloat32): Likewise.
29255 (movdf_hardfloat64_mfpgpr): Likewise.
29256 (movdf_hardfloat64): Likewise.
29257 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
29258 (movdf_softfloat64): Likewise.
29259 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
29260 (reload_<mode>_load): Move to later in the file so they aren't in
29261 the middle of the floating point move insns.
29262 (reload_<mode>_store): Likewise.
29264 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
29267 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
29268 constraint if -mdebug=reg.
29269 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
29270 Enable using dd reload support if needed.
29272 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
29273 binary and decimal floating point moves in rs6000.md.
29274 (movtd_internal): Likewise.
29276 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
29277 decimal floating point moves.
29279 (movtf_internal): Likewise.
29280 (mov<mode>_internal, TDmode/TFmode): Likewise.
29281 (movtf_softfloat): Likewise.
29282 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
29284 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
29285 movdi_internal64, using wg constraint for move direct operations.
29286 (movdi_internal64): Likewise.
29288 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
29289 MODES_TIEABLE_P for selected modes. Print the numerical value of
29290 the various virtual registers. Use GPR/FPR first/last values,
29291 instead of hard coding the register numbers. Print which modes
29292 have reload functions registered.
29293 (rs6000_option_override_internal): If -mdebug=reg, trace the options
29294 settings before/after setting cpu, target and subtarget settings.
29295 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
29296 and for secondary reload failures in rs6000_secondary_reload_inner.
29297 (rs6000_secondary_reload_fail): Likewise.
29298 (rs6000_secondary_reload_inner): Likewise.
29300 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
29301 macros for first/last GPR and FPR registers.
29302 (LAST_GPR_REGNO): Likewise.
29303 (FIRST_FPR_REGNO): Likewise.
29304 (LAST_FPR_REGNO): Likewise.
29306 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
29307 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
29308 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
29309 (vcond<mode><mode>): Likewise.
29310 (vcondu<mode><mode>): Likewise.
29311 (vector_gtu<mode>): Likewise.
29312 (vector_gte<mode>): Likewise.
29313 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
29314 to prevent the compiler from converting DImode operations to TImode.
29315 (ior<mode>3): Likewise.
29316 (and<mode>3): Likewise.
29317 (one_cmpl<mode>2): Likewise.
29318 (nor<mode>3): Likewise.
29319 (andc<mode>3): Likewise.
29321 * config/rs6000/constraints.md (wt constraint): New constraint
29322 that returns VSX_REGS if TImode is allowed in VSX registers.
29324 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
29325 constant under VSX.
29327 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
29328 similar to TImode, but it is restricted to being in the GPRs.
29330 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
29331 TImode to occupy a single VSX register.
29333 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
29334 -mvsx-timode for power7/power8.
29335 (power7 cpu): Likewise.
29336 (power8 cpu): Likewise.
29338 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
29339 sure that TFmode/TDmode take up two registers if they are ever
29340 allowed in the upper VSX registers.
29341 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
29343 (rs6000_init_hard_regno_mode_ok): Likewise.
29344 (rs6000_debug_reg_global): Add debugging for PTImode and wt
29345 constraint. Print if LRA is turned on.
29346 (rs6000_option_override_internal): Give an error if -mvsx-timode
29347 and VSX is not enabled.
29348 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
29349 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
29350 to reg+offset addressing. Use PTImode when checking offset
29351 addresses for validity.
29352 (reg_offset_addressing_ok_p): Likewise.
29353 (rs6000_legitimate_offset_address_p): Likewise.
29354 (rs6000_legitimize_address): Likewise.
29355 (rs6000_legitimize_reload_address): Likewise.
29356 (rs6000_legitimate_address_p): Likewise.
29357 (rs6000_eliminate_indexed_memrefs): Likewise.
29358 (rs6000_emit_move): Likewise.
29359 (rs6000_secondary_reload): Likewise.
29360 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
29361 reloads to fpr registers to continue to use reg+offset addressing,
29362 but 64-bit reloads to altivec registers need reg+reg addressing.
29363 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
29364 it. Treat LO_SUM like a PLUS operation.
29365 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
29366 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
29367 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
29368 registers to share a register with a smaller sized type, since VSX
29369 puts scalars in the upper 64-bits.
29370 (print_operand): Add support for PTImode.
29371 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
29372 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
29373 registers, but don't have arithmetic support.
29374 (rs6000_memory_move_cost): Add test for VSX.
29375 (rs6000_opt_masks): Add -mvsx-timode.
29377 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
29380 (VSr): Use wt constraint for TImode.
29381 (VSv): Drop TImode support.
29382 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
29383 (vsx_movti_64bit): Likewise.
29384 (vsx_movti_32bit): Likewise.
29385 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
29386 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
29387 one '?' on the appropriate output constraint. Do not allow TImode
29388 logical operations on 32-bit systems.
29389 (vsx_ior<mode>3): Likewise.
29390 (vsx_xor<mode>3): Likewise.
29391 (vsx_one_cmpl<mode>2): Likewise.
29392 (vsx_nor<mode>3): Likewise.
29393 (vsx_andc<mode>3): Likewise.
29394 (vsx_concat_<mode>): Likewise.
29395 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
29397 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
29398 OPTION_MASK_VSX_TIMODE.
29399 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
29400 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
29402 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
29403 (TI2 iterator): New iterator for TImode, PTImode.
29404 (wd mode attribute): Add values for vector types.
29405 (movti_string): Replace TI move operations with operations for TImode
29406 and PTImode. Add support for TImode being allowed in VSX registers.
29407 (mov<mode>_string, TImode/PTImode): Likewise.
29408 (movti_ppc64): Likewise.
29409 (mov<mode>_ppc64, TImode/PTImode): Likewise.
29410 (TI mode splitters): Likewise.
29412 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
29415 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
29417 PR tree-optimization/56355
29418 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
29419 Also handle integers with undefined overflow.
29421 2013-03-20 Catherine Moore <clm@codesourcery.com>
29422 Maciej W. Rozycki <macro@codesourcery.com>
29423 Tom de Vries <tom@codesourcery.com>
29424 Nathan Sidwell <nathan@codesourcery.com>
29425 Iain Sandoe <iain@codesourcery.com>
29426 Nathan Froyd <froydnj@codesourcery.com>
29427 Chao-ying Fu <fu@mips.com>
29429 * doc/extend.texi: (micromips, nomicromips, nocompression):
29430 Document new function attributes.
29431 * doc/invoke.texi (minterlink-compressed, mmicromips,
29432 m14k, m14ke, m14kec): Document new options.
29433 (minterlink-mips16): Update documentation.
29434 * doc/md.texi (ZC, ZD): Document new constraints.
29435 * configure.ac (gcc_cv_as_micromips): Check if linker
29436 supports the .set micromips directive.
29437 * configure: Regenerate.
29438 * config.in: Regenerate.
29439 * config/mips/mips-tables.opt: Regenerate.
29440 * config/mips/micromips.md: New file.
29441 * constraints.md (ZC, ZD): New constraints.
29442 * config/mips/predicates.md (movep_src_register): New predicate.
29443 (movep_src_operand): New predicate.
29444 (non_volatile_mem_operand): New predicate.
29445 * config/mips/mips.md (multimem): New type.
29446 (length): Differentiate between 17-bit and 18-bit branch offsets.
29447 (MOVEP1, MOVEP2): New mode iterator.
29448 (mov_<load>l): Use ZC constraint.
29449 (mov_<load>r): Likewise.
29450 (mov_<store>l): Likewise.
29451 (mov_<store>r): Likewise.
29452 (*branch_equality<mode>_inverted): Add microMIPS support.
29453 (*branch_equality<mode>): Likewise.
29454 (*jump_absolute): Likewise.
29455 (indirect_jump_<mode>): Likewise.
29456 (tablejump_<mode>): Likewise.
29457 (<optab>_internal): Likewise.
29458 (sibcall_internal): Likewise.
29459 (sibcall_value_internal): Likewise.
29460 (prefetch): Use constraint ZD.
29461 * config/mips/mips.opt (minterlink-compressed): New option.
29462 (minterlink-mips16): Now an alias for minterlink-compressed.
29463 (mmicromips): New option.
29464 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
29465 (compare_and_swap_12): Likewise.
29466 (sync_add<mode>): Likewise.
29467 (sync_<optab>_12): Likewise.
29468 (sync_old_<optab>_12): Likewise.
29469 (sync_new_<optab>_12): Likewise.
29470 (sync_nand_12): Likewise.
29471 (sync_old_nand_12): Likewise.
29472 (sync_new_nand_12): Likewise.
29473 (sync_sub<mode>): Likewise.
29474 (sync_old_add<mode>): Likewise.
29475 (sync_old_sub<mode>): Likewise.
29476 (sync_new_add<mode>): Likewise.
29477 (sync_new_sub<mode>): Likewise.
29478 (sync_<optab><mode>): Likewise.
29479 (sync_old_<optab><mode>): Likewise.
29480 (sync_new_<optab><mode>): Likewise.
29481 (sync_nand<mode>): Likewise.
29482 (sync_old_nand<mode>): Likewise.
29483 (sync_new_nand<mode>): Likewise.
29484 (sync_lock_test_and_set<mode>): Likewise.
29485 (test_and_set_12): Likewise.
29486 (atomic_compare_and_swap<mode>): Likewise.
29487 (atomic_exchange<mode>_llsc): Likewise.
29488 (atomic_fetch_add<mode>_llsc): Likewise.
29489 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
29490 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
29491 (umips_save_restore_pattern_p): Likewise.
29492 (umips_load_store_pair_p): Likewise.
29493 (umips_output_load_store_pair): Likewise.
29494 (umips_movep_target_p): Likewise.
29495 (umips_12bit_offset_address_p): Likewise.
29496 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
29497 (mips_base_mips16): Rename this...
29498 (mips_base_compression_flags): ...to this. Update all uses.
29499 (mips_attribute_table): Add micromips, nomicromips and nocompression.
29500 (mips_mips16_decl_p): Delete.
29501 (mips_nomips16_decl_p): Delete.
29502 (mips_get_compress_on_flags): New function.
29503 (mips_get_compress_off_flags): New function.
29504 (mips_get_compress_mode): New function.
29505 (mips_get_compress_on_name): New function.
29506 (mips_get_compress_off_name): New function.
29507 (mips_insert_attributes): Support multiple compression types.
29508 (mips_merge_decl_attributes): Likewise.
29509 (umips_12bit_offset_address_p): New function.
29510 (mips_start_function_definition): Emit .set micromips directive.
29511 (mips_call_may_need_jalx_p): New function.
29512 (mips_function_ok_for_sibcall): Add microMIPS support.
29513 (mips_print_operand_punctuation): Support short delay slots and
29515 (umips_swm_mask, umips_swm_encoding): New.
29516 (umips_build_save_restore): New function.
29517 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
29518 (was_mips16_p): Remove.
29519 (old_compression_mode): New.
29520 (mips_set_compression_mode): New function.
29521 (mips_set_current_function): Add microMIPS support.
29522 (mips_option_override): Likewise.
29523 (umips_save_restore_pattern_p): New function.
29524 (umips_output_save_restore): New function.
29525 (umips_load_store_pair_p_1): New function.
29526 (umips_load_store_pair_p): New function.
29527 (umips_output_load_store_pair_1): New function.
29528 (umips_output_load_store_pair): New function.
29529 (umips_movep_target_p) New function.
29530 (mips_prepare_pch_save): Add microMIPS support.
29531 * config/mips/mips.h (TARGET_COMPRESSION): New.
29532 (TARGET_CPU_CPP_BUILTINS): Update macro
29533 to use new compression flags and to support microMIPS.
29534 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
29535 (MIPS_ARCH_FLOAT_SPEC): Likewise.
29536 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
29537 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
29538 (ASM_SPEC): Support mmicromips and mno-micromips.
29539 (M16STORE_REG_P): New macro.
29540 (MIPS_CALL): Support TARGET_MICROMIPS.
29541 (MICROMIPS_J): New macro.
29542 (mips_base_mips16): Rename this...
29543 (mips_base_compression_flags): ...to this.
29544 (UMIPS_12BIT_OFFSET_P): New macro.
29545 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
29546 (MULTILIB_DIRNAMES): Likewise.
29547 2013-03-20 Richard Biener <rguenther@suse.de>
29549 PR tree-optimization/56661
29550 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
29551 the result does not have to be distinct.
29553 2013-03-20 Richard Biener <rguenther@suse.de>
29555 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
29558 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
29559 Steven Bosscher <steven@gcc.gnu.org>
29561 PR rtl-optimization/56605
29562 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
29564 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
29567 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
29568 that require movd instead of movq.
29570 2013-03-20 Richard Biener <rguenther@suse.de>
29572 * tree-ssa-structalias.c (struct variable_info): Add pointer
29573 to the first field of an aggregate with sub-vars. Make
29574 this and the pointer to the next subfield its ID.
29575 (vi_next): New function.
29576 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
29577 storedanything_id, integer_id): Increment by one.
29578 (new_var_info, get_call_vi, lookup_call_clobber_vi,
29579 get_call_clobber_vi): Adjust.
29580 (solution_set_expand): Simplify and speedup.
29581 (solution_set_add): Inline into ...
29582 (set_union_with_increment): ... this. Adjust accordingly.
29583 (do_sd_constraint): Likewise.
29584 (do_ds_constraint): Likewise.
29585 (do_complex_constraint): Simplify.
29586 (build_pred_graph): Adjust.
29587 (solve_graph): Likewise. Simplify and speedup.
29588 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
29589 get_constraint_for_component_ref, get_constraint_for_1,
29590 first_vi_for_offset, first_or_preceding_vi_for_offset,
29591 create_function_info_for, create_variable_info_for_1,
29592 create_variable_info_for, intra_create_variable_infos): Adjust.
29593 (init_base_vars): Push NULL for ID zero.
29594 (compute_points_to_sets): Adjust.
29596 2013-03-20 Richard Biener <rguenther@suse.de>
29598 * cfgloop.c (verify_loop_structure): Streamline and avoid
29599 ICEing on corrupt loop tree.
29600 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
29603 2013-03-20 Richard Biener <rguenther@suse.de>
29605 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
29606 check whether an SSA update is needed.
29608 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
29610 * config/mips/constraints.md (T): Rename to...
29614 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
29615 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
29617 2013-03-19 Ian Bolton <ian.bolton@arm.com>
29619 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
29620 (*subsi3_carryin_uxtw): Likewise.
29622 2013-03-19 Ian Bolton <ian.bolton@arm.com>
29624 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
29625 (*rorsi3_insn_uxtw): Likewise.
29627 2013-03-19 Ian Bolton <ian.bolton@arm.com>
29629 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
29630 (*extrsi5_insn_uxtw): Likewise.
29632 2013-03-19 Richard Biener <rguenther@suse.de>
29634 PR tree-optimization/56273
29635 * passes.c (init_optimization_passes): Move second VRP after DOM.
29637 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
29639 * config/i386/i386.md (*movti_internal): Merge from
29640 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
29641 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
29642 nox64 isa attributes.
29644 2013-03-18 Richard Biener <rguenther@suse.de>
29646 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
29648 (merge_node_constraints): Likewise.
29649 (build_succ_graph): Likewise.
29650 (valid_graph_edge): Inline into single caller.
29651 (unify_nodes): Likewise. Use bitmap_set_bit return value
29653 (scc_visit): Fix formatting and variable use.
29654 (do_sd_constraint): Use gcc_checking_assert.
29655 (do_ds_constraint): Likewise.
29656 (do_complex_constraint): Likewise.
29657 (condense_visit): Likewise. Cleanup.
29658 (dump_pred_graph): New function.
29659 (perform_var_substitution): Dump the pred-graph before
29660 variable substitution.
29661 (find_equivalent_node): Use gcc_checking_assert.
29662 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
29664 2013-03-18 Richard Biener <rguenther@suse.de>
29666 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
29667 Remove cond_expr_stmt_list argument and do not gimplify the
29669 (vect_loop_versioning): Adjust.
29670 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
29671 Cleanup to use less temporaries.
29672 (vect_create_data_ref_ptr): Cleanup.
29674 2013-03-18 Jakub Jelinek <jakub@redhat.com>
29676 PR tree-optimization/56635
29677 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
29678 require types_compatible_p types.
29680 2013-03-18 Nick Clifton <nickc@redhat.com>
29682 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
29683 spurious backslash.
29685 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
29686 Add missing line to comment describing function.
29688 2013-03-18 Richard Biener <rguenther@suse.de>
29690 PR tree-optimization/56210
29691 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
29692 Handle string / character search functions.
29693 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
29695 2013-03-18 Richard Biener <rguenther@suse.de>
29697 PR middle-end/56483
29698 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
29699 and implement properly.
29700 * gimple.h (gimple_cond_single_var_p): Remove.
29702 2013-03-18 Richard Biener <rguenther@suse.de>
29704 * tree-data-ref.h (find_data_references_in_loop): Declare.
29705 * tree-data-ref.c (get_references_in_stmt): Use a stack
29706 vector pre-allocated in the callers.
29707 (find_data_references_in_stmt): Adjust.
29708 (graphite_find_data_references_in_stmt): Likewise.
29709 (create_rdg_vertices): Likewise.
29710 (find_data_references_in_loop): Export.
29711 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
29712 Compute dependences here...
29713 (vect_analyze_data_refs): ...not here. When we encounter
29714 a non-vectorizable data reference in basic-block vectorization
29715 truncate the data reference vector. Do not bother to
29716 fixup data-dependence information for gather loads.
29717 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
29718 of data references, as reported.
29720 2013-03-18 Richard Biener <rguenther@suse.de>
29722 PR tree-optimization/3713
29723 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
29724 has_constants and expr.
29725 (stmt_has_constants): Properly valueize SSA names when deciding
29726 whether the stmt has constants.
29728 2013-03-18 Richard Biener <rguenther@suse.de>
29730 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
29731 whole function when there is nothing to do.
29732 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
29733 * tree-vectorizer.c (vectorize_loops): Update virtual and
29734 loop-closed SSA once.
29735 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
29737 2013-03-18 Richard Biener <rguenther@suse.de>
29739 PR middle-end/56113
29740 * domwalk.c (bb_postorder): New global static.
29741 (cmp_bb_postorder): New function.
29742 (walk_dominator_tree): Replace scheme imposing an order for
29743 visiting dominator sons by one sorting them at the time they
29744 are pushed on the stack.
29746 2013-03-18 Richard Biener <rguenther@suse.de>
29748 PR tree-optimization/39326
29749 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
29750 (struct mem_ref): Replace mem member with ao_ref typed member.
29751 (MEM_ANALYZABLE): Adjust.
29752 (memref_eq): Likewise.
29753 (mem_ref_alloc): Likewise.
29754 (gather_mem_refs_stmt): Likewise.
29755 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
29756 (execute_sm_if_changed_flag_set): Adjust.
29757 (execute_sm): Likewise.
29758 (ref_always_accessed_p): Likewise.
29759 (refs_independent_p): Likewise.
29760 (can_sm_ref_p): Likewise.
29762 2013-03-18 Jakub Jelinek <jakub@redhat.com>
29765 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
29766 return 1 even for !unsignedp.
29768 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
29770 * config/i386/i386.md (isa): Add x64 and nox64.
29771 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
29772 (*pushtf): Enable *roF alternative for x64 isa only.
29773 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
29774 mode attribute of integer alternatives to DImode for TARGET_64BIT.
29775 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
29776 (*movtf_internal): Merge from *movtf_internal_rex64 and
29777 *movtf_internal_sse. Use x64 and nox64 isa attributes.
29778 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
29779 nox64 isa attributes.
29780 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
29781 nox64 isa attributes.
29782 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
29784 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
29786 * config/alpha/alpha.c (TARGET_LRA_P): New define.
29788 2013-03-17 Jakub Jelinek <jakub@redhat.com>
29791 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
29792 class names. Remove trailing comma after "ALL_REGS".
29794 2013-03-16 Jan Hubicka <jh@suse.cz>
29796 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
29797 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
29798 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
29799 of cgraph_get_create_node.
29800 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
29802 2013-03-16 Jason Merrill <jason@redhat.com>
29805 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
29806 with DW_AT_default_value.
29808 2013-03-16 Jakub Jelinek <jakub@redhat.com>
29810 * BASE-VER: Set to 4.9.0.
29812 2013-03-14 Andi Kleen <ak@linux.intel.com>
29815 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
29816 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
29817 Document _x* TSX intrinsics.
29819 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
29820 David Holsgrove <david.holsgrove@xilinx.com>
29822 * configure.ac: Add MicroBlaze TLS support detection.
29823 * configure: Regenerate.
29824 * config/microblaze/microblaze-protos.h
29825 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
29826 symbol_mentioned_p, label_mentioned_p): Add prototypes.
29827 * config/microblaze/microblaze.c (microblaze_address_type): Add
29828 ADDRESS_TLS and tls_reloc address types.
29829 (microblaze_address_info): Add tls_reloc.
29830 (TARGET_HAVE_TLS): Define.
29831 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
29832 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
29833 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
29834 load_tls_operand, microblaze_call_tls_get_addr,
29835 microblaze_legitimize_tls_address): New functions.
29836 (microblaze_classify_unspec): Handle UNSPEC_TLS.
29837 (get_base_reg): Use microblaze_tls_symbol_p.
29838 (microblaze_classify_address): Handle TLS.
29839 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
29840 label_mentioned_p and microblaze_tls_referenced_p.
29841 (microblaze_legitimize_address): Handle TLS.
29842 (microblaze_address_insns): Handle ADDRESS_TLS.
29843 (pic_address_needs_scratch): Handle TLS.
29844 (print_operand_address): Handle TLS.
29845 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
29846 (microblaze_expand_move): Handle TLS.
29847 (microblaze_legitimate_constant_p): Check
29848 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
29849 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
29850 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
29851 (PIC_OFFSET_TABLE_REGNUM): Set.
29852 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
29853 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
29854 (addsi3, movsi_internal2, movdf_internal): Update constraints
29855 * config/microblaze/predicates.md (arith_plus_operand): Define
29856 (move_operand): Redefine as move_src_operand,
29857 check microblaze_tls_referenced_p.
29859 2013-03-14 Ian Bolton <ian.bolton@arm.com>
29861 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
29862 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
29864 2013-03-14 Ian Bolton <ian.bolton@arm.com>
29866 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
29869 2013-03-14 Jakub Jelinek <jakub@redhat.com>
29871 PR tree-optimization/53265
29872 * common.opt (Waggressive-loop-optimizations): New option.
29873 * tree-ssa-loop-niter.c: Include tree-pass.h.
29874 (do_warn_aggressive_loop_optimizations): New function.
29875 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
29876 if number_of_latch_executions returned constant.
29877 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
29878 early. If number_of_latch_executions returned constant, set
29879 nb_iterations_upper_bound back to it.
29880 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
29882 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
29883 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
29885 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
29886 (MULTILIB_OSDIRNAMES): Set.
29887 * genmultilib: If defaultosdirname doesn't start with :: , set
29888 defaultosdirname2 instead, clear it and emit two . multilib_raw
29889 entries instead of just one.
29891 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
29893 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
29894 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
29895 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
29896 (SUBTARGET_OVERRIDE_OPTIONS): New.
29898 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
29901 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
29902 (musermode): Convert to Var(TARGET_USERMODE).
29903 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
29904 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
29905 * config/sh/sh.c (sh_option_override): Use
29906 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
29907 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
29909 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
29911 (udivsi3_i4_single, divsi3_i4_single): Use
29912 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
29914 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
29916 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
29919 2013-03-13 Richard Biener <rguenther@suse.de>
29921 PR tree-optimization/56608
29922 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
29923 calls when vectorizing basic-blocks.
29925 2013-03-13 Jakub Jelinek <jakub@redhat.com>
29928 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
29931 2013-03-12 Jakub Jelinek <jakub@redhat.com>
29933 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
29935 2013-03-11 Jan Hubicka <jh@suse.cz>
29938 * lto-streamer-out.c (output_symbol_p): Skip references from
29939 constructors of external variables.
29941 2013-03-11 Jan Hubicka <jh@suse.cz>
29943 PR middle-end/56571
29944 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
29946 * emit-rtl.c (verify_rtx_sharing): Likewise.
29947 (copy_insn_1): Likewise.
29948 * rtl.c (copy_rtx): Likewise.
29950 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
29953 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
29954 output_operand_lossage message.
29956 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
29959 * arm.c (shift_op): Validate RTL pattern on the fly.
29960 (arm_print_operand, case 'S'): Don't use shift_operator to validate
29963 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29966 * config/pa/pa.md (call_value): Check for calls to powf and direct to
29967 new call patterns that clobber %fr12.
29968 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
29969 split and postreload patterns.
29970 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
29971 registers %fr12 and %fr12R as call used.
29973 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
29975 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
29976 (canon_address, record_store, replace_read, check_mem_read_rtx,
29977 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
29978 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
29979 rest_of_handle_dse): Likewise.
29981 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
29983 PR middle-end/56524
29984 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
29986 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
29987 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
29988 (save_optabs_if_changed): Replace with...
29989 (init_tree_optimization_optabs): ...this.
29990 * optabs.c (save_optabs_if_changed): Rename to...
29991 (init_tree_optimization_optabs): ...this. Take the optimization node
29992 as argument. Do nothing if the base optabs are already correct.
29993 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
29994 to recompute optabs.
29995 * function.h (function): Remove optabs field.
29996 * function.c (invoke_set_current_function_hook): Call
29997 init_tree_optimization_optabs. Use the result to initialize
30000 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
30002 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
30003 if GTMA_HAS_NO_INSTRUMENTATION.
30004 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
30005 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
30006 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
30007 * gimple-pretty-print.c (dump_gimple_transaction): Handle
30008 GTMA_HAS_NO_INSTRUMENTATION.
30010 2013-03-08 Jakub Jelinek <jakub@redhat.com>
30012 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
30015 2013-03-08 Marek Polacek <polacek@redhat.com>
30016 Jakub Jelinek <jakub@redhat.com>
30018 PR tree-optimization/56478
30019 * predict.c (is_comparison_with_loop_invariant_p): Change the
30020 type of loop_step to tree.
30021 (predict_loops): Adjust.
30022 (predict_iv_comparison): Perform the computations on double_ints.
30024 2013-03-08 Richard Biener <rguenther@suse.de>
30026 PR tree-optimization/56570
30027 * tree-cfg.c (verify_expr_location_1): Verify locations for
30029 * tree-sra.c (create_access_replacement): Strip locations
30030 from DECL_DEBUG_EXPRs.
30032 2013-03-08 Richard Biener <rguenther@suse.de>
30034 * tree-inline.c (expand_call_inline): Do not associate
30035 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
30036 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
30038 2013-03-08 Richard Biener <rguenther@suse.de>
30040 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
30041 or block changes with -Og. Fix for location / block encoding
30042 changes and PHI arguments with locations.
30044 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
30046 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
30048 (struct output_info): Likewise.
30049 (register_overhead): Remove bad gcc_assert.
30050 (bitmap_find_bit): If there is only a single bitmap element, do not
30051 count a miss as a search.
30052 (print_statistics): Update for counter type changes.
30053 (dump_bitmap_statistics): Likewise. Print headers such that they
30054 are properly lined up with the printed counters.
30056 2013-03-07 Jakub Jelinek <jakub@redhat.com>
30058 PR tree-optimization/56559
30059 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
30060 check that it has only a single use.
30062 2013-03-07 Richard Biener <rguenther@suse.de>
30064 * doc/invoke.texi (fwhole-program): Discourage use in combination
30067 2013-03-06 Jakub Jelinek <jakub@redhat.com>
30069 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
30071 PR tree-optimization/56539
30072 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
30073 instead of GSI_CONTINUE_LINKING as last argument to
30074 force_gimple_operand_gsi. Adjust function comment.
30076 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
30079 PR middle-end/56548
30080 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
30081 promoted mode, convert the result back to the original mode.
30083 2013-03-06 Richard Biener <rguenther@suse.de>
30085 PR middle-end/56294
30086 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
30087 (insert_updated_phi_nodes_compare_uids): New function.
30088 (update_ssa): Sort symbols_to_rename after UID before
30089 traversing it to insert PHI nodes.
30091 2013-03-06 Richard Biener <rguenther@suse.de>
30093 PR middle-end/50494
30094 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
30095 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
30098 2013-02-13 Richard Biener <rguenther@suse.de>
30101 * varasm.c (output_constant_def_1): Get the decl representing
30102 the constant as argument.
30103 (output_constant_def): Wrap output_constant_def_1.
30104 (make_decl_rtl): Use output_constant_def_1 with the decl
30105 representing the constant.
30106 (build_constant_desc): Optionally re-use a decl already
30107 representing the constant.
30108 (tree_output_constant_def): Adjust.
30110 2013-03-06 Joey Ye <joey.ye@arm.com>
30113 * gcc.c (convert_white_space): New function.
30114 (main): Handles white space in function name.
30116 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
30119 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
30120 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
30121 to SH_DIV_CALL_TABLE for TARGET_SH2.
30122 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
30124 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
30125 call-table options.
30127 2013-03-05 Sterling Augustine <saugustine@google.com>
30128 Cary Coutant <ccoutant@google.com>
30131 * dwarf2out.c (resolve_addr): Don't call
30132 remove_loc_list_addr_table_entries a second time for the same
30135 2013-03-05 Jakub Jelinek <jakub@redhat.com>
30138 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
30139 (avoid_complex_debug_insns): New function.
30140 (expand_debug_locations): Call it.
30142 PR rtl-optimization/56484
30143 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
30144 lifetimes of hard registers on small register class machines.
30146 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
30148 * config/microblaze/microblaze-protos.h: Rename
30149 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
30150 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
30152 (microblaze_fast_interrupt_function_p): New function.
30153 (microblaze_is_interrupt_handler): Rename to
30154 microblaze_is_interrupt_variant and add fast_interrupt check.
30155 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
30156 (save_restore_insns): Likewise.
30157 (compute_frame_size): Likewise.
30158 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
30159 (microblaze_globalize_label): Likewise.
30160 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
30161 * config/microblaze/microblaze.md: Use wrapper
30162 microblaze_is_interrupt_variant.
30164 2013-03-05 Kai Tietz <ktietz@redhat.com>
30166 * sdbout.c (sdbout_one_type): Switch to current function's section
30167 supporting cold/hot.
30169 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
30171 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
30174 2013-03-05 Jakub Jelinek <jakub@redhat.com>
30176 PR middle-end/56461
30177 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
30178 if VALGRIND_GET_VBITS is defined, temporarily make object
30179 memory all defined, and restore previous valgrind addressability
30180 and definability afterwards. Free this_object at the end.
30182 PR middle-end/56461
30183 * lra.c (lra): Call lra_clear_live_ranges if live_p,
30184 right before calling lra_create_live_ranges, also call it
30185 when clearing live_p. Only call lra_clear_live_ranges
30186 at the end if live_p.
30188 PR middle-end/56461
30189 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
30191 2013-03-05 Richard Biener <rguenther@suse.de>
30193 PR tree-optimization/56521
30194 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
30197 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
30200 * except.h (remove_unreachable_eh_regions): New prototype.
30201 * except.c (remove_eh_handler_splicer): New function, split out
30202 of remove_eh_handler.
30203 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
30204 warning about running it on many EH regions one at a time.
30205 (remove_unreachable_eh_regions_worker): New function, walk the
30206 EH tree in depth-first order and remove non-marked regions.
30207 (remove_unreachable_eh_regions): New function.
30208 * tree-eh.c (mark_reachable_handlers): New function, split out
30209 from remove_unreachable_handlers.
30210 (remove_unreachable_handlers): Use mark_reachable_handlers and
30211 remove_unreachable_eh_regions.
30212 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
30213 and remove_unreachable_eh_regions.
30215 2013-03-05 Richard Biener <rguenther@suse.de>
30217 PR middle-end/56525
30218 * loop-init.c (fix_loop_structure): Remove loops in two stages,
30219 not freeing them until the end.
30221 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
30223 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
30225 2013-03-05 Richard Biener <rguenther@suse.de>
30227 PR tree-optimization/56270
30228 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
30229 of loads after scheduling an SLP instance.
30231 2013-03-05 Jakub Jelinek <jakub@redhat.com>
30233 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
30235 (check_gcc_parallelize): Run guality.exp as a separate job from
30236 vect.exp with unsorted.exp and $(dg_target_exps) separately from
30237 struct-layout-1.exp with stackalign.exp.
30239 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
30241 PR middle-end/56461
30242 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
30243 load_index sbitmap even if some bit in it isn't set.
30245 PR middle-end/56461
30246 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
30247 (discover_iteration_bound_by_body_walk): Change queues to
30248 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
30249 spelling in comment. Call safe_push on queues[bound_index] directly.
30250 Release queues[queue_index] in every iteration unconditionally.
30251 Release bounds vector.
30253 PR middle-end/56461
30254 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
30255 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
30256 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
30259 2013-03-05 Richard Biener <rguenther@suse.de>
30262 * tree-inline.c (remap_blocks_to_null): New function.
30263 (expand_call_inline): When expanding a call stmt without
30264 an associated block inline remap all callee blocks to NULL.
30266 2013-03-05 Jakub Jelinek <jakub@redhat.com>
30268 PR rtl-optimization/56494
30269 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
30270 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
30271 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
30273 PR middle-end/56461
30274 * sel-sched-ir.c (free_sched_pools): Release
30275 succs_info_pool.stack[succs_info_pool.max_top] vectors too
30276 if succs_info_pool.max_top isn't -1.
30279 * opts.c (opts_obstack, opts_concat): Moved to...
30280 * opts-common.c (opts_obstack, opts_concat): ... here.
30282 2013-03-04 Jakub Jelinek <jakub@redhat.com>
30284 PR middle-end/56461
30285 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
30287 2013-03-04 Martin Jambor <mjambor@suse.cz>
30289 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
30290 all appropriate places.
30292 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
30294 PR tree-optimization/56424
30295 * ipa-split.c (split_function): Do not set the RSO flag if result is
30296 not by reference and its type is a register type.
30298 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
30300 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
30301 (microblaze_legitimate_pic_operand): Likewise
30302 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
30303 new function microblaze_legitimate_pic_operand
30304 * config/microblaze/microblaze-protos.h
30305 (microblaze_legitimate_pic_operand): Declare.
30307 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
30309 * config/microblaze/predicates.md (call_insn_simple_operand):
30310 New predicate for supported rtx code types.
30311 * config/microblaze/microblaze.md (call_internal1): Use
30312 call_insn_simple_operand predicate.
30314 2013-03-04 Jakub Jelinek <jakub@redhat.com>
30316 PR middle-end/56461
30317 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
30318 partitions.ordered_remove.
30320 PR middle-end/56461
30321 * tree-vect-stmts.c (vectorizable_conversion): Don't call
30322 vec_oprnds0.create (1) for modifier == NONE.
30324 PR middle-end/56461
30325 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
30326 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
30327 vec_oprnds1 right before pushing anything to it for
30330 PR middle-end/56461
30331 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
30332 set nbbs to 0 instead of having separate code path.
30333 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
30334 instead of false as last argument if returning NULL.
30336 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
30338 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
30339 the attribute is now called "target" instead of "option".
30340 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
30341 * doc/tm.texi.in (Target Attributes): Likewise document the correct
30342 attribute/pragma name for TARGET_OPTION_VALID_P and
30343 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
30344 * doc/tm.texi: Regenerated.
30346 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
30348 * config/microblaze/microblaze.c:
30349 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
30350 * config/microblaze/microblaze.h: Add -mxl-reorder to
30352 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
30353 instructions emitted if TARGET_REORDER.
30354 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
30355 or 0 for -m/-mno case, but initialises as 2 to detect default use case
30358 2013-03-01 Xinliang David Li <davidxl@google.com>
30360 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
30363 2013-03-01 Jakub Jelinek <jakub@redhat.com>
30365 PR middle-end/56461
30366 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
30367 vector even when returning true. Fix up function comment formatting.
30369 PR middle-end/56461
30370 * ira-build.c (ira_loop_nodes_count): New variable.
30371 (create_loop_tree_nodes): Initialize it.
30372 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
30374 PR middle-end/56461
30375 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
30376 method on dr_chain and result_chain.
30377 * tree-vect-stmts.c (vectorizable_store): Only call
30378 result_chain.create if j == 0.
30380 PR middle-end/56461
30381 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
30382 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
30383 before overwriting it.
30385 2013-03-01 Tobias Burnus <burnus@net-b.de>
30387 * doc/extended.texi (C Extensions): Change order in @menu
30389 (Other MIPS Built-in Functions): Move last MIPS entry before
30390 "picoChip Built-in Functions".
30391 (SH Built-in Functions): Move after RX Built-in Functions.
30392 * doc/gcc.texi (Introduction): Change order in @menu
30394 * doc/md.texi (Constraints): Ditto.
30395 * gty.texi (Type Information): Ditto.
30396 (User-provided marking routines for template types): Make
30398 * doc/invoke.texi (AArch64 Options): Move before
30399 "Adapteva Epiphany Options".
30401 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
30402 Jakub Jelinek <jakub@redhat.com>
30405 * asan.c (gate_asan): Lookup no_sanitize_address instead of
30406 no_address_safety_analysis attribute.
30407 * doc/extend.texi (no_address_safety_attribute): Rename to
30408 no_sanitize_address attribute, mention no_address_safety_analysis
30409 attribute as deprecated alias.
30411 2013-02-28 Jakub Jelinek <jakub@redhat.com>
30413 PR middle-end/56461
30414 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
30415 type to vec<vec<tree> > *.
30416 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
30417 to be vec<tree> instead of vec<tree> *, set vec_defs
30418 to vNULL and call vec_defs.create (number_of_vects), adjust other
30420 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
30421 vectorizable_condition): Adjust vect_get_slp_defs callers.
30423 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
30425 * config/aarch64/aarch64.c
30426 (aarch64_float_const_representable): Remove unused variable.
30428 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
30430 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
30432 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
30434 * config/aarch64/aarch64-builtins.c
30435 (aarch64_init_simd_builtins): Make static.
30437 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
30439 * config/aarch64/aarch64.c
30440 (aarch64_simd_make_constant): Make static.
30442 2013-02-28 Martin Jambor <mjambor@suse.cz>
30444 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
30445 with no initialization to the RHS of debug statements.
30447 2013-02-28 Martin Jambor <mjambor@suse.cz>
30449 PR tree-optimization/56294
30450 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
30452 (get_access_replacement): Do not call create_access_replacement.
30453 Assert a replacement exists.
30454 (get_repl_default_def_ssa_name): Create the replacement declaration
30457 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
30459 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
30460 final_end_function.
30462 2013-02-28 Marek Polacek <polacek@redhat.com>
30464 PR rtl-optimization/56466
30465 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
30466 if we're changing a loop.
30467 (peel_loops_completely): Likewise.
30469 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
30472 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
30474 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
30477 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
30478 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
30479 INTX_FTYPE_FX, FX_FTYPE_INTX.
30480 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
30482 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
30484 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
30485 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
30486 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
30487 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
30488 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
30489 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
30490 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
30491 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
30492 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
30493 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
30494 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
30495 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
30496 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
30497 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
30498 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
30499 (avrxmega6): Increase max flash segments from 5 to 6.
30500 * config/avr/t-multilib: Regenerate.
30501 * config/avr/avr-tables.opt: Regenerate.
30502 * doc/avr-mmcu.texi: Regenerate.
30504 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
30506 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
30507 (avr_device_to_arch): Rename to avr_device_to_ld.
30508 (avr_device_to_as): New prototype.
30509 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
30510 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
30511 * config/avr/driver-avr.c (avr_device_to_as): New.
30512 (avr_device_to_arch): Rename to avr_device_to_ld.
30514 2013-02-27 Jakub Jelinek <jakub@redhat.com>
30516 PR middle-end/56461
30517 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
30518 method on dr_chain and result_chain.
30520 PR middle-end/56461
30521 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
30522 pointer_set_destroy on not_executed_last_iteration.
30524 PR middle-end/56461
30525 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
30527 PR middle-end/56461
30528 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
30529 FOR_EACH_DEFINED_FUNCTION when freeing state.
30531 PR middle-end/56461
30532 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
30534 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
30537 PR middle-end/56461
30538 * ipa-cp.c (decide_whether_version_node): Call vec_free on
30539 known_aggs[i].items and release known_aggs vector.
30541 PR middle-end/56461
30542 * ipa-reference.c (propagate): Free node_info even for alias nodes.
30544 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
30546 * config/microblaze/microblaze.c (microblaze_emit_compare):
30547 Use xor for EQ/NE comparisions.
30548 * config/microblaze/microblaze.md (cstoresf4): Add constraints
30549 (cbranchsf4): Adjust operator to comparison_operator.
30551 2013-02-27 Jakub Jelinek <jakub@redhat.com>
30553 PR middle-end/56461
30554 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
30556 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
30557 vec_safe_push, always update *slot.
30558 (redirect_edge_var_map_clear): Use vec_free.
30559 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
30560 (free_var_map_entry): Use vec_free.
30561 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
30562 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
30564 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
30566 PR middle-end/45472
30567 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
30568 when the may_trap_p bit of the exprs being merged differs.
30569 Reorder tests for speculativeness in the logical and operator.
30571 2013-02-27 Jakub Jelinek <jakub@redhat.com>
30573 * incpath.c (add_standard_paths): Use reconcat instead of concat
30574 where appropriate and avoid leaking memory.
30576 * opts.h: Include obstack.h.
30577 (opts_concat): New prototype.
30578 (opts_obstack): New declaration.
30579 * opts.c (opts_concat): New function.
30580 (opts_obstack): New variable.
30581 (init_options_struct): Call gcc_init_obstack on opts_obstack.
30582 (finish_options): Use opts_concat instead of concat
30583 and XOBNEWVEC instead of XNEWVEC.
30584 * opts-common.c (generate_canonical_option, decode_cmdline_option,
30585 generate_option): Likewise.
30586 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
30587 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
30590 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
30591 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
30593 2013-02-26 Jakub Jelinek <jakub@redhat.com>
30595 PR middle-end/56461
30596 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
30598 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
30600 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
30601 (arm_block_move_unaligned_straight): Likewise.
30602 (arm_adjust_block_mem): Likewise.
30604 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
30607 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
30608 temp, cond and label.
30609 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
30612 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
30613 * config/c6x/c6x.h (dbx_register_map): Update declaration.
30616 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
30617 of prologue/epilogue functions.
30620 * config/tilegx/tilegx.c (tilegx_expand_prologue):
30621 Remove unused variable cfa_offset.
30622 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
30625 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
30626 type promotion to unsigned.
30629 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
30630 for HOST_WIDE_INT of 32 bit / same size as int.
30631 (arm_block_move_unaligned_straight): Likewise.
30632 (arm_adjust_block_mem): Likewise.
30635 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
30638 2013-02-26 Marek Polacek <polacek@redhat.com>
30640 PR tree-optimization/56426
30641 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
30643 2013-02-26 Richard Biener <rguenther@suse.de>
30646 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
30647 unused variable loops.
30649 2013-02-26 Jakub Jelinek <jakub@redhat.com>
30651 PR tree-optimization/56448
30652 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
30653 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
30654 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
30655 later operands of the references, or even first operand for
30656 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
30658 PR tree-optimization/56443
30659 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
30660 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
30661 to type_for_mode langhook.
30663 2013-02-25 Matt Turner <mattst88@gmail.com>
30665 * doc/invoke.texi: Document r4700.
30667 2013-02-25 Richard Biener <rguenther@suse.de>
30669 PR tree-optimization/56175
30670 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
30672 (simplify_bitwise_binary): ... here. Also guard the conversion
30673 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
30675 2013-02-25 Catherine Moore <clm@codesourcery.com>
30678 2013-02-24 Catherine Moore <clm@codesourcery.com>
30679 Maciej W. Rozycki <macro@codesourcery.com>
30680 Tom de Vries <tom@codesourcery.com>
30681 Nathan Sidwell <nathan@codesourcery.com>
30682 Iain Sandoe <iain@codesourcery.com>
30683 Nathan Froyd <froydnj@codesourcery.com>
30684 Chao-ying Fu <fu@mips.com>
30686 * doc/extend.texi: (micromips, nomicromips, nocompression):
30687 Document new function attributes.
30688 * doc/invoke.texi (minterlink-compressed, mmicromips,
30689 m14k, m14ke, m14kec): Document new options.
30690 (minterlink-mips16): Update documentation.
30691 * doc/md.texi (ZC, ZD): Document new constraints.
30692 * configure.ac (gcc_cv_as_micromips): Check if linker
30693 supports the .set micromips directive.
30694 * configure: Regenerate.
30695 * config.in: Regenerate.
30696 * config/mips/mips-tables.opt: Regenerate.
30697 * config/mips/micromips.md: New file.
30698 * constraints.md (ZC, AD): New constraints.
30699 * config/mips/predicates.md (movep_src_register): New predicate.
30700 (movep_src_operand): New predicate.
30701 (non_volatile_mem_operand): New predicate.
30702 * config/mips/mips.md (multimem): New type.
30703 (length): Differentiate between 17-bit and 18-bit branch offsets.
30704 (MOVEP1, MOVEP2): New mode iterator.
30705 (mov_<load>l): Use ZC constraint.
30706 (mov_<load>r): Likewise.
30707 (mov_<store>l): Likewise.
30708 (mov_<store>r): Likewise.
30709 (*branch_equality<mode>_inverted): Add microMIPS support.
30710 (*branch_equality<mode>): Likewise.
30711 (*jump_absolute): Likewise.
30712 (indirect_jump_<mode>): Likewise.
30713 (tablejump_<mode>): Likewise.
30714 (<optab>_internal): Likewise.
30715 (sibcall_internal): Likewise.
30716 (sibcall_value_internal): Likewise.
30717 (prefetch): Use constraint ZD.
30718 * config/mips/mips.opt (minterlink-compressed): New option.
30719 (minterlink-mips16): Now an alias for minterlink-compressed.
30720 (mmicromips): New option.
30721 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
30722 (compare_and_swap_12): Likewise.
30723 (sync_add<mode>): Likewise.
30724 (sync_<optab>_12): Likewise.
30725 (sync_old_<optab>_12): Likewise.
30726 (sync_new_<optab>_12): Likewise.
30727 (sync_nand_12): Likewise.
30728 (sync_old_nand_12): Likewise.
30729 (sync_new_nand_12): Likewise.
30730 (sync_sub<mode>): Likewise.
30731 (sync_old_add<mode>): Likewise.
30732 (sync_old_sub<mode>): Likewise.
30733 (sync_new_add<mode>): Likewise.
30734 (sync_new_sub<mode>): Likewise.
30735 (sync_<optab><mode>): Likewise.
30736 (sync_old_<optab><mode>): Likewise.
30737 (sync_new_<optab><mode>): Likewise.
30738 (sync_nand<mode>): Likewise.
30739 (sync_old_nand<mode>): Likewise.
30740 (sync_new_nand<mode>): Likewise.
30741 (sync_lock_test_and_set<mode>): Likewise.
30742 (test_and_set_12): Likewise.
30743 (atomic_compare_and_swap<mode>): Likewise.
30744 (atomic_exchange<mode>_llsc): Likewise.
30745 (atomic_fetch_add<mode>_llsc): Likewise.
30746 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
30747 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
30748 (umips_save_restore_pattern_p): Likewise.
30749 (umips_load_store_pair_p): Likewise.
30750 (umips_output_load_store_pair): Likewise.
30751 (umips_movep_target_p): Likewise.
30752 (umips_12bit_offset_address_p): Likewise.
30753 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
30754 (mips_base_mips16): Rename this...
30755 (mips_base_compression_flags): ...to this. Update all uses.
30756 (mips_attribute_table): Add micromips, nomicromips and nocompression.
30757 (mips_mips16_decl_p): Delete.
30758 (mips_nomips16_decl_p): Delete.
30759 (mips_get_compress_on_flags): New function.
30760 (mips_get_compress_off_flags): New function.
30761 (mips_get_compress_mode): New function.
30762 (mips_get_compress_on_name): New function.
30763 (mips_get_compress_off_name): New function.
30764 (mips_insert_attributes): Support multiple compression types.
30765 (mips_merge_decl_attributes): Likewise.
30766 (umips_12bit_offset_address_p): New function.
30767 (mips_start_function_definition): Emit .set micromips directive.
30768 (mips_call_may_need_jalx_p): New function.
30769 (mips_function_ok_for_sibcall): Add microMIPS support.
30770 (mips_print_operand_punctuation): Support short delay slots and
30772 (umips_swm_mask, umips_swm_encoding): New.
30773 (umips_build_save_restore): New function.
30774 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
30775 (was_mips16_p): Remove.
30776 (old_compression_mode): New.
30777 (mips_set_compression_mode): New function.
30778 (mips_set_current_function): Add microMIPS support.
30779 (mips_option_override): Likewise.
30780 (umips_save_restore_pattern_p): New function.
30781 (umips_output_save_restore): New function.
30782 (umips_load_store_pair_p_1): New function.
30783 (umips_load_store_pair_p): New function.
30784 (umips_output_load_store_pair_1): New function.
30785 (umips_output_load_store_pair): New function.
30786 (umips_movep_target_p) New function.
30787 (mips_prepare_pch_save): Add microMIPS support.
30788 * config/mips/mips.h (TARGET_COMPRESSION): New.
30789 (TARGET_CPU_CPP_BUILTINS): Update macro
30790 to use new compression flags and to support microMIPS.
30791 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
30792 (MIPS_ARCH_FLOAT_SPEC): Likewise.
30793 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
30794 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
30795 (ASM_SPEC): Support mmicromips and mno-micromips.
30796 (M16STORE_REG_P): New macro.
30797 (MIPS_CALL): Support TARGET_MICROMIPS.
30798 (MICROMIPS_J): New macro.
30799 (mips_base_mips16): Rename this...
30800 (mips_base_compression_flags): ...to this.
30801 (UMIPS_12BIT_OFFSET_P): New macro.
30802 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
30803 (MULTILIB_DIRNAMES): Likewise.
30805 2013-02-25 Tom de Vries <tom@codesourcery.com>
30807 PR rtl-optimization/56131
30808 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
30809 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
30810 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
30812 2013-02-25 Tobias Burnus <burnus@net-b.de>
30814 * doc/invoke.texi (-fsanitize=): Move from optimization
30815 to debugging options.
30817 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
30819 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
30821 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
30822 Alexander Monakov <amonakov@ispras.ru>
30824 PR middle-end/56077
30825 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
30826 flush pending lists also on non-jumps. Adjust comment.
30828 2013-02-24 Catherine Moore <clm@codesourcery.com>
30829 Maciej W. Rozycki <macro@codesourcery.com>
30830 Tom de Vries <tom@codesourcery.com>
30831 Nathan Sidwell <nathan@codesourcery.com>
30832 Iain Sandoe <iain@codesourcery.com>
30833 Nathan Froyd <froydnj@codesourcery.com>
30834 Chao-ying Fu <fu@mips.com>
30836 * doc/extend.texi: (micromips, nomicromips, nocompression):
30837 Document new function attributes.
30838 * doc/invoke.texi (minterlink-compressed, mmicromips,
30839 m14k, m14ke, m14kec): Document new options.
30840 (minterlink-mips16): Update documentation.
30841 * doc/md.texi (ZC, ZD): Document new constraints.
30842 * configure.ac (gcc_cv_as_micromips): Check if linker
30843 supports the .set micromips directive.
30844 * configure: Regenerate.
30845 * config.in: Regenerate.
30846 * config/mips/mips-tables.opt: Regenerate.
30847 * config/mips/micromips.md: New file.
30848 * constraints.md (ZC, AD): New constraints.
30849 * config/mips/predicates.md (movep_src_register): New predicate.
30850 (movep_src_operand): New predicate.
30851 (non_volatile_mem_operand): New predicate.
30852 * config/mips/mips.md (multimem): New type.
30853 (length): Differentiate between 17-bit and 18-bit branch offsets.
30854 (MOVEP1, MOVEP2): New mode iterator.
30855 (mov_<load>l): Use ZC constraint.
30856 (mov_<load>r): Likewise.
30857 (mov_<store>l): Likewise.
30858 (mov_<store>r): Likewise.
30859 (*branch_equality<mode>_inverted): Add microMIPS support.
30860 (*branch_equality<mode>): Likewise.
30861 (*jump_absolute): Likewise.
30862 (indirect_jump_<mode>): Likewise.
30863 (tablejump_<mode>): Likewise.
30864 (<optab>_internal): Likewise.
30865 (sibcall_internal): Likewise.
30866 (sibcall_value_internal): Likewise.
30867 (prefetch): Use constraint ZD.
30868 * config/mips/mips.opt (minterlink-compressed): New option.
30869 (minterlink-mips16): Now an alias for minterlink-compressed.
30870 (mmicromips): New option.
30871 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
30872 (compare_and_swap_12): Likewise.
30873 (sync_add<mode>): Likewise.
30874 (sync_<optab>_12): Likewise.
30875 (sync_old_<optab>_12): Likewise.
30876 (sync_new_<optab>_12): Likewise.
30877 (sync_nand_12): Likewise.
30878 (sync_old_nand_12): Likewise.
30879 (sync_new_nand_12): Likewise.
30880 (sync_sub<mode>): Likewise.
30881 (sync_old_add<mode>): Likewise.
30882 (sync_old_sub<mode>): Likewise.
30883 (sync_new_add<mode>): Likewise.
30884 (sync_new_sub<mode>): Likewise.
30885 (sync_<optab><mode>): Likewise.
30886 (sync_old_<optab><mode>): Likewise.
30887 (sync_new_<optab><mode>): Likewise.
30888 (sync_nand<mode>): Likewise.
30889 (sync_old_nand<mode>): Likewise.
30890 (sync_new_nand<mode>): Likewise.
30891 (sync_lock_test_and_set<mode>): Likewise.
30892 (test_and_set_12): Likewise.
30893 (atomic_compare_and_swap<mode>): Likewise.
30894 (atomic_exchange<mode>_llsc): Likewise.
30895 (atomic_fetch_add<mode>_llsc): Likewise.
30896 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
30897 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
30898 (umips_save_restore_pattern_p): Likewise.
30899 (umips_load_store_pair_p): Likewise.
30900 (umips_output_load_store_pair): Likewise.
30901 (umips_movep_target_p): Likewise.
30902 (umips_12bit_offset_address_p): Likewise.
30903 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
30904 (mips_base_mips16): Rename this...
30905 (mips_base_compression_flags): ...to this. Update all uses.
30906 (mips_attribute_table): Add micromips, nomicromips and nocompression.
30907 (mips_mips16_decl_p): Delete.
30908 (mips_nomips16_decl_p): Delete.
30909 (mips_get_compress_on_flags): New function.
30910 (mips_get_compress_off_flags): New function.
30911 (mips_get_compress_mode): New function.
30912 (mips_get_compress_on_name): New function.
30913 (mips_get_compress_off_name): New function.
30914 (mips_insert_attributes): Support multiple compression types.
30915 (mips_merge_decl_attributes): Likewise.
30916 (umips_12bit_offset_address_p): New function.
30917 (mips_start_function_definition): Emit .set micromips directive.
30918 (mips_call_may_need_jalx_p): New function.
30919 (mips_function_ok_for_sibcall): Add microMIPS support.
30920 (mips_print_operand_punctuation): Support short delay slots and
30922 (umips_swm_mask, umips_swm_encoding): New.
30923 (umips_build_save_restore): New function.
30924 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
30925 (was_mips16_p): Remove.
30926 (old_compression_mode): New.
30927 (mips_set_compression_mode): New function.
30928 (mips_set_current_function): Add microMIPS support.
30929 (mips_option_override): Likewise.
30930 (umips_save_restore_pattern_p): New function.
30931 (umips_output_save_restore): New function.
30932 (umips_load_store_pair_p_1): New function.
30933 (umips_load_store_pair_p): New function.
30934 (umips_output_load_store_pair_1): New function.
30935 (umips_output_load_store_pair): New function.
30936 (umips_movep_target_p) New function.
30937 (mips_prepare_pch_save): Add microMIPS support.
30938 * config/mips/mips.h (TARGET_COMPRESSION): New.
30939 (TARGET_CPU_CPP_BUILTINS): Update macro
30940 to use new compression flags and to support microMIPS.
30941 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
30942 (MIPS_ARCH_FLOAT_SPEC): Likewise.
30943 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
30944 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
30945 (ASM_SPEC): Support mmicromips and mno-micromips.
30946 (M16STORE_REG_P): New macro.
30947 (MIPS_CALL): Support TARGET_MICROMIPS.
30948 (MICROMIPS_J): New macro.
30949 (mips_base_mips16): Rename this...
30950 (mips_base_compression_flags): ...to this.
30951 (UMIPS_12BIT_OFFSET_P): New macro.
30952 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
30953 (MULTILIB_DIRNAMES): Likewise.
30955 2013-02-24 Jakub Jelinek <jakub@redhat.com>
30958 * target-globals.c (save_target_globals): For init_reg_sets and
30959 target_reinit remporarily set this_fn_optabs to this_target_optabs.
30961 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
30963 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
30964 * config/aarch64/t-aarch64
30965 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
30967 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
30969 PR inline-asm/56148
30970 * lra-constraints.c (process_alt_operands): Reload operand
30971 conflicting with earlier clobber only if no more other conflicting
30974 2013-02-22 Jakub Jelinek <jakub@redhat.com>
30977 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
30978 if not linking a shared library.
30980 2013-02-22 Seth LaForge <sethml@google.com>
30982 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
30984 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
30986 * config/arm/arm.md (split for extendsidi): Update condition.
30987 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
30988 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
30989 (qhs_zextenddi_cstr): Likewise.
30991 2013-02-21 Jakub Jelinek <jakub@redhat.com>
30993 PR middle-end/56420
30994 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
30995 avoid signed wrapping.
30996 (expand_mult): Handle properly multiplication by
30997 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
30998 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
30999 in the compiler if coeff is HOST_WIDE_INT_MIN.
31000 (expand_divmod): Don't make ext_op1 static, change it's type to
31001 uhwi. Avoid undefined behavior in -INTVAL (op1).
31003 PR rtl-optimization/50339
31004 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
31006 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
31007 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
31008 into splitting_ashiftrt field.
31009 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
31011 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
31014 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
31016 PR middle-end/56108
31017 * trans-mem.c (execute_tm_mark): Do not expand transactions that
31018 are sure to go irrevocable.
31020 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
31022 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
31023 scalars are valid operands.
31025 2013-02-21 Martin Jambor <mjambor@suse.cz>
31027 PR tree-optimization/56310
31028 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
31029 only matching indices and non-negative final offsets.
31030 (intersect_aggregates_with_edge): Pass src_idx to
31031 agg_replacements_to_vector. Pass src_idx insstead of index to
31032 intersect_with_agg_replacements.
31034 2013-02-21 Martin Jambor <mjambor@suse.cz>
31036 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
31037 instead of hard-wired defaults.
31039 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
31041 * doc/invoke.texi (MIPS Options): Update documentation of the
31042 floating-point multiply-accumulate instruction restrictions.
31044 2013-02-21 Kostya Serebryany <kcc@google.com>
31046 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
31047 asan_shadow_offset on x86_64 linux.
31049 2013-02-21 Richard Biener <rguenther@suse.de>
31051 PR tree-optimization/56415
31053 2013-02-11 Richard Biener <rguenther@suse.de>
31055 PR tree-optimization/56273
31056 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
31059 2013-02-21 Jakub Jelinek <jakub@redhat.com>
31062 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
31065 PR inline-asm/56405
31066 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
31067 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
31069 2013-02-20 Jan Hubicka <jh@suse.cz>
31071 PR tree-optimization/56265
31072 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
31073 when target is referenced for first time.
31075 2013-02-20 Richard Biener <rguenther@suse.de>
31077 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
31078 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
31079 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
31080 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
31081 not return anything.
31082 (rename_ssa_copies): Do not remove unused locals.
31083 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
31084 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
31085 * passes.c (execute_function_todo): Do not schedule unused locals
31086 removal if cleanup_tree_cfg did something.
31087 * tree-ssa-live.c (remove_unused_locals): Dump statistics
31088 about the number of removed locals.
31090 2013-02-20 Richard Biener <rguenther@suse.de>
31092 PR tree-optimization/56398
31093 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
31095 2013-02-20 Martin Jambor <mjambor@suse.cz>
31097 PR tree-optimization/55334
31098 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
31099 restricted pointers to arrays.
31101 2013-02-20 Richard Biener <rguenther@suse.de>
31102 Jakub Jelinek <jakub@redhat.com>
31104 PR tree-optimization/56396
31105 * tree-ssa-ccp.c (n_const_val): New static variable.
31106 (get_value): Return NULL for SSA names we don't have a lattice
31108 (ccp_initialize): Initialize n_const_val.
31109 * tree-ssa-copy.c (n_copy_of): New static variable.
31110 (init_copy_prop): Initialize n_copy_of.
31111 (get_value): Return NULL_TREE for SSA names we don't have a
31114 2013-02-20 Martin Jambor <mjambor@suse.cz>
31116 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
31118 2013-02-20 Richard Biener <rguenther@suse.de>
31120 * genpreds.c (write_lookup_constraint): Do not compare first
31121 letter of the constraint again.
31123 2013-02-20 Richard Biener <rguenther@suse.de>
31125 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
31127 (get_use_iv_cost): Terminate hashtable walk when coming across
31130 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
31132 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
31133 reassociation for avx2 targets.
31135 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
31137 * config/microblaze/microblaze.c: microblaze_has_clz = 0
31138 Add version check for v8.10.a to enable microblaze_has_clz
31139 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
31140 version and TARGET_PATTERN_COMPARE check
31141 * config/microblaze/microblaze.md: New clzsi2 instruction
31143 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
31145 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
31146 function before branching.
31148 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
31150 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
31152 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
31154 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
31156 PR middle-end/55889
31157 * sel-sched.c: Include ira.h.
31158 (implicit_clobber_conflict_p): New function.
31159 (moveup_expr): Use it.
31160 * Makefile.in (sel-sched.o): Depend on ira.h.
31162 2013-02-19 Richard Biener <rguenther@suse.de>
31164 PR tree-optimization/56384
31165 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
31166 (vn_hash_type): Split out from ...
31167 (vn_hash_constant_with_type): ... here.
31168 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
31169 (vn_phi_eq): Compare types from vn_phi_s structure.
31170 (vn_phi_lookup): Populate vn_phi_s type.
31171 (vn_phi_insert): Likewise.
31173 2013-02-19 Jakub Jelinek <jakub@redhat.com>
31175 PR tree-optimization/56350
31176 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
31177 if haven't found reduction or nested cycle operand, rather than
31178 asserting we must find it.
31180 PR tree-optimization/56381
31181 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
31184 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
31185 Jakub Jelinek <jakub@redhat.com>
31188 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
31189 (swap_optab_enable): Same.
31190 (init_all_optabs): Use argument instead of global.
31191 * tree.h (struct tree_optimization_option): New field target_optabs.
31192 * expr.h (init_all_optabs): Add argument to prototype.
31193 (TREE_OPTIMIZATION_OPTABS): New.
31194 (save_optabs_if_changed): Protoize.
31195 * optabs.h: Declare this_fn_optabs.
31196 * optabs.c (save_optabs_if_changed): New.
31197 Declare this_fn_optabs.
31198 (init_optabs): Add argument to init_all_optabs() call.
31199 * function.c (invoke_set_current_function_hook): Handle per
31201 * function.h (struct function): New field optabs.
31202 * config/mips/mips.c (mips_set_mips16_mode): Handle when
31203 optimization_current_node has changed.
31204 * target-globals.h (save_target_globals_default_opts): Protoize.
31205 * target-globals.c (save_target_globals_default_opts): New.
31207 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
31210 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
31211 registers %fr12 and %fr12R as call used.
31214 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
31215 and HImode, require all displacements to be an integer multiple of
31217 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
31218 only allow QImode and HImode when reload is in progress and strict is
31219 true. Likewise for symbolic addresses. Use base14_operand to check
31220 displacements in REG+BASE addresses.
31222 2013-02-18 Richard Biener <rguenther@suse.de>
31224 PR tree-optimization/56366
31225 * tree-vect-loop.c (get_initial_def_for_induction): Properly
31226 handle sign-conversion of outer-loop initial induction value.
31228 2013-02-18 Richard Biener <rguenther@suse.de>
31230 PR middle-end/56349
31231 * cfghooks.c (merge_blocks): If we merge a latch into another
31232 block adjust references to it.
31233 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
31234 (verify_loop_structure): Verify that a recorded latch is in fact
31237 2013-02-18 Richard Biener <rguenther@suse.de>
31239 PR tree-optimization/56321
31240 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
31241 order SSA name release and virtual operand unlinking.
31243 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
31245 * config/microblaze/microblaze.md (save_stack_block): Define.
31246 (restore_stack_block): Likewise.
31248 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
31250 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
31251 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
31252 * config/microblaze/microblaze.c (microblaze_option_override):
31253 Bail out early for PIC modes when target does not support PIC.
31255 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
31257 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
31258 Replace with a microblaze version.
31259 (microblaze_trampoline_init): Adapt for microblaze.
31260 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
31263 2013-02-16 Jakub Jelinek <jakub@redhat.com>
31264 Dodji Seketeli <dodji@redhat.com>
31267 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
31268 (instrument_mem_region_access): Do not forget to always put
31269 instrumentation of the of 'base' and 'base + len' in a "if (len !=
31270 0) statement, even for cases where either 'base' or 'base + len'
31271 are not instrumented -- because they have been previously
31272 instrumented. Simplify the logic by putting all the statements
31273 instrument 'base + len' inside a sequence, and then insert that
31274 sequence right before the current insertion point. Then, to
31275 instrument 'base + len', just get an iterator on that statement.
31276 And do not forget to update the pointer to iterator the function
31277 received as argument.
31279 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
31281 PR rtl-optimization/56348
31282 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
31284 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
31286 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
31287 (clean_graph_dump_file): Pass base to start_graph_dump.
31289 2013-02-14 Richard Henderson <rth@redhat.com>
31292 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
31294 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
31296 * collect2-aix.h: Define F_LOADONLY.
31298 2013-02-14 Richard Biener <rguenther@suse.de>
31301 * varasm.c (output_constant_def_1): Get the decl representing
31302 the constant as argument.
31303 (output_constant_def): Wrap output_constant_def_1.
31304 (make_decl_rtl): Use output_constant_def_1 with the decl
31305 representing the constant.
31306 (build_constant_desc): Optionally re-use a decl already
31307 representing the constant.
31308 (tree_output_constant_def): Adjust.
31310 2013-02-14 Dodji Seketeli <dodji@redhat.com>
31313 * asan.c (instrument_builtin_call): Really put the length of the
31314 second source argument into src1_len.
31316 2013-02-13 Jakub Jelinek <jakub@redhat.com>
31318 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
31319 argument. If it is false, don't create edge from then_bb to
31321 (insert_if_then_before_iter): Pass true to it.
31322 (build_check_stmt): Pass false to it.
31323 (transform_statements): Flush hash table only on extended basic
31324 block boundaries, rather than at the beginning of every bb.
31325 Don't flush hash table on nonfreeing_call_p calls.
31326 * tree-flow.h (nonfreeing_call_p): New prototype.
31327 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
31329 2013-02-13 David S. Miller <davem@davemloft.net>
31331 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
31333 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
31336 * ira.c (max_regno_before_ira): Move from ...
31338 (fix_reg_equiv_init): Use max_regno_before_ira instead of
31341 2013-02-13 Jakub Jelinek <jakub@redhat.com>
31343 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
31345 2013-02-13 Richard Biener <rguenther@suse.de>
31348 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
31349 globals in MEM_REFs.
31351 2013-02-13 Richard Biener <rguenther@suse.de>
31353 * loop-init.c (loop_optimizer_init): Clear loop state when
31354 re-initializing preserved loops.
31355 * loop-unswitch.c (unswitch_single_loop): Return whether
31356 we unswitched the loop. Do not verify loop state here.
31357 (unswitch_loops): When we unswitched a loop discover new loops.
31359 2013-02-13 Kostya Serebryany <kcc@google.com>
31361 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
31363 * sanitizer.def: Rename __asan_init to __asan_init_v1.
31365 2013-02-12 Dodji Seketeli <dodji@redhat.com>
31367 Avoid instrumenting duplicated memory access in the same basic block
31368 * Makefile.in (asan.o): Add new dependency on hash-table.h
31369 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
31370 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
31371 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
31372 (free_mem_ref_resources, has_mem_ref_been_instrumented)
31373 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
31374 (get_mem_ref_of_assignment): New functions.
31375 (get_mem_refs_of_builtin_call): Extract from
31376 instrument_builtin_call and tweak a little bit to make it fit with
31378 (instrument_builtin_call): Use the new
31379 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
31380 of is_gimple_builtin_call.
31381 (instrument_derefs, instrument_mem_region_access): Insert the
31382 instrumented memory reference into the hash table.
31383 (maybe_instrument_assignment): Renamed instrument_assignment into
31384 this, and change it to advance the iterator when instrumentation
31385 actually happened and return true in that case. This makes it
31386 homogeneous with maybe_instrument_assignment, and thus give a
31387 chance to callers to be more 'regular'.
31388 (transform_statements): Clear the memory reference hash table
31389 whenever we enter a new BB, when we cross a function call, or when
31390 we are done transforming statements. Use
31391 maybe_instrument_assignment instead of instrumentation. No more
31392 need to special case maybe_instrument_assignment and advance the
31393 iterator after calling it; it's now handled just like
31394 maybe_instrument_call. Update comment.
31396 2013-02-13 Richard Biener <rguenther@suse.de>
31398 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
31399 Fix loop discovery code.
31401 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
31403 PR inline-asm/56148
31404 * lra-constraints.c (process_alt_operands): Match early clobber
31405 operand with itself. Check conflicts with earlyclobber only if
31406 the operand is not reloaded. Prefer to reload conflicting operand
31407 if earlyclobber and matching operands are the same.
31409 2013-02-12 Richard Biener <rguenther@suse.de>
31412 * lto-streamer-out.c (write_symbol): Do not output symbols
31413 for hard register variables.
31415 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
31418 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
31419 (umulsidi3_insn, mulsidi3_insn): New insns.
31421 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
31423 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
31424 (struct tune_params): Add vec_costs field.
31425 * config/arm/arm.c (arm_builtin_vectorization_cost)
31426 (arm_add_stmt_cost): New functions.
31427 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
31428 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
31429 (arm_default_vec_cost): New struct of type cpu_vec_costs.
31430 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
31431 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
31432 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
31433 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
31435 2013-02-12 Richard Biener <rguenther@suse.de>
31438 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
31439 decls again if possible.
31441 2013-02-12 Richard Biener <rguenther@suse.de>
31443 PR middle-end/56288
31444 * tree-ssa.c (verify_ssa_name): Fix check, move
31445 SSA_NAME_IN_FREE_LIST check up.
31447 2013-02-12 Jakub Jelinek <jakub@redhat.com>
31448 Steven Bosscher <steven@gcc.gnu.org>
31450 PR rtl-optimization/56151
31451 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
31452 equal to op0 or op1, and last_insn pattern is CODE operation
31453 with MEM dest and one of the operands matches that MEM.
31455 2013-02-11 Sriraman Tallam <tmsriram@google.com>
31457 * doc/extend.texi: Document Function Multiversioning and "default"
31458 parameter string to target attribute.
31459 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
31460 target attribute parameter is "default".
31461 (ix86_compare_version_priority): Remove checks for target attribute.
31462 (ix86_mangle_function_version_assembler_name): Change error to sorry.
31463 Remove check for target attribute equal to NULL. Add assert.
31464 (ix86_generate_version_dispatcher_body): Change error to sorry.
31466 2013-02-11 Iain Sandoe <iain@codesourcery.com>
31467 Jack Howarth <howarth@bromo.med.uc.edu>
31468 Patrick Marlier <patrick.marlier@gmail.com>
31471 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
31472 define ENDFILE_SPEC as TM_DESTRUCTOR.
31473 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
31475 2013-02-11 Alexander Potapenko <glider@google.com>
31476 Jack Howarth <howarth@bromo.med.uc.edu>
31477 Jakub Jelinek <jakub@redhat.com>
31480 * config/darwin.c (cdtor_record): Rename ctor_record.
31481 (sort_cdtor_records): Rename sort_ctor_records.
31482 (finalize_dtors): New routine to sort destructors by
31483 priority before use in assemble_integer.
31484 (machopic_asm_out_destructor): Use finalize_dtors if needed.
31486 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
31488 PR rtl-optimization/56275
31489 * simplify-rtx.c (avoid_constant_pool_reference): Check that
31490 offset is non-negative and less than cmode size before
31491 calling simplify_subreg.
31493 2013-02-11 Richard Biener <rguenther@suse.de>
31495 PR tree-optimization/56264
31496 * cfgloop.h (fix_loop_structure): Adjust prototype.
31497 * loop-init.c (fix_loop_structure): Return the number of
31498 newly discovered loops.
31499 * tree-cfgcleanup.c (repair_loop_structures): When new loops
31500 are discovered, do a full loop-closed SSA rewrite.
31502 2013-02-11 Richard Biener <rguenther@suse.de>
31504 PR tree-optimization/56273
31505 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
31507 (check_array_ref): Fix missing newline in dumps.
31508 (search_for_addr_array): Likewise.
31510 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
31512 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
31514 2013-02-09 Jakub Jelinek <jakub@redhat.com>
31517 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
31519 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
31521 PR rtl-optimization/56246
31522 * lra-constraints.c (simplify_operand_subreg): Try to reuse
31524 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
31525 constraints are satisfied.
31527 2013-02-08 Jeff Law <law@redhat.com>
31530 * emit-rtl.c (reg_is_parm_p): New function.
31531 * regs.h (reg_is_parm_p): New prototype.
31532 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
31533 callee-clobbered registers.
31535 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
31538 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
31539 If there is no implicit builtin declaration, just return NULL.
31541 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
31543 * config/i386/sse.md (FMAMODEM): New mode iterator.
31544 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
31545 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
31547 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
31549 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
31550 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
31551 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
31553 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
31555 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
31556 (microblaze*-*-elf): Likewise.
31557 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
31559 * config/microblaze/microblaze-c.c: Add builtin defines for
31560 _LITTLE_ENDIAN and _BIG_ENDIAN.
31561 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
31562 add to TARGET_DEFAULT flags.
31563 Expand ASM_SPEC and LINK_SPEC.
31564 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
31565 * config/microblaze/microblaze.md: Update extendsidi2 and
31566 movdi_internal instructions to use low-order / high-order reg
31568 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
31569 options and inversemask / mask of LITTLE_ENDIAN.
31570 * config/microblaze/t-microblaze: Expand multilib options to
31571 include mlittle-endian (le) and update exceptions patterns.
31573 2013-02-08 Jakub Jelinek <jakub@redhat.com>
31575 PR rtl-optimization/56195
31576 * lra-constraints.c (get_reload_reg): Don't reuse regs
31577 if they have smaller mode than requested, if they have
31578 wider mode than requested, try to return a SUBREG.
31580 PR tree-optimization/56250
31581 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
31582 if type is unsigned and code isn't MULT_EXPR.
31584 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
31586 PR tree-optimization/56064
31587 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
31588 bits according to mode.
31589 * fixed-value.h (fixed_from_double_int)
31590 (const_fixed_from_double_int): Adjust comments.
31592 2013-02-08 Richard Biener <rguenther@suse.de>
31595 * lto-streamer.h (struct data_in): Remove current_file, current_line
31596 and current_col members.
31597 * lto-streamer-out.c (lto_output_location): Stream changed bits
31598 en-block for efficiency.
31599 * lto-streamer-in.c (clear_line_info): Remove.
31600 (lto_input_location): Cache current file, line and column
31601 globally via local statics. Read changed bits en-block.
31602 (input_function): Do not call clear_line_info.
31603 (lto_read_body): Likewise.
31604 (lto_input_toplevel_asms): Likewise.
31606 2013-02-08 Michael Matz <matz@suse.de>
31608 PR tree-optimization/52448
31609 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
31610 (nt_call_phase): New static.
31611 (add_or_mark_expr): Only mark accesses with newer phase than any
31613 (nonfreeing_call_p): New.
31614 (nt_init_block): Update nt_call_phase, mark blocks as visited.
31615 (nt_fini_block): Keep blocks marked as visited.
31616 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
31618 2013-02-08 Richard Biener <rguenther@suse.de>
31620 * ira.c (ira): Free broken dominator information.
31622 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
31624 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
31626 2013-02-08 Marek Polacek <polacek@redhat.com>
31628 * cfgloop.c (verify_loop_structure): Add more checking of headers.
31630 2013-02-08 Richard Biener <rguenther@suse.de>
31632 PR middle-end/56181
31633 * cfgloop.h (flow_loops_find): Adjust.
31634 (bb_loop_header_p): Declare.
31635 * cfgloop.c (bb_loop_header_p): New function split out from ...
31636 (flow_loops_find): ... here. Adjust function signature,
31637 support incremental loop structure update.
31638 (verify_loop_structure): Cleanup. Verify a loop is a loop.
31639 * cfgloopmanip.c (fix_loop_structure): Move ...
31640 * loop-init.c (fix_loop_structure): ... here.
31641 (apply_loop_flags): Split out from ...
31642 (loop_optimizer_init): ... here.
31643 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
31644 in incremental mode, only remove dead loops here.
31646 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
31649 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
31650 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
31651 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
31652 (*round<mode>3.libgcc): New insns for fixed-modes.
31653 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
31654 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
31655 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
31656 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
31657 implementations. Define to __builtin_avr_absFX,
31658 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
31659 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
31660 __builtin_avr_countlsFX, respectively.
31661 * config/avr/avr-c.c (target.h): Include it.
31662 (enum avr_builtin_id): New enum.
31663 (avr_resolve_overloaded_builtin): New static function.
31664 (avr_register_target_pragmas): Use it to set
31665 targetm.resolve_overloaded_builtin.
31666 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
31667 tree nodes used by DEF_BUILTIN.
31668 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
31669 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
31670 <AVR_BUILTIN_xxBITS>: Same.
31672 2013-02-08 Richard Biener <rguenther@suse.de>
31674 * cfgloop.c (verify_loop_structure): Properly handle
31675 a loop exiting to another loop header.
31676 * ira-int.h (ira_loops): Remove.
31677 * ira.c (ira_loops): Remove.
31678 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
31679 (do_reload): Use loop_optimizer_finalize.
31680 * ira-build.c (create_loop_tree_nodes): Use get_loops and
31681 number_of_loops to access the loop tree.
31682 (more_one_region_p): Likewise.
31683 (finish_loop_tree_nodes): Likewise.
31684 (rebuild_regno_allocno_maps): Likewise.
31685 (mark_loops_for_removal): Likewise.
31686 (mark_all_loops_for_removal): Likewise.
31687 (remove_unnecessary_regions): Likewise.
31688 (ira_build): Likewise.
31689 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
31691 2013-02-08 Richard Biener <rguenther@suse.de>
31693 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
31694 * ipa-pure-const.c (analyze_function): Avoid calling
31695 mark_irreducible_loops twice.
31696 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
31698 2013-02-07 David S. Miller <davem@davemloft.net>
31700 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
31702 * var-tracking.c (vt_add_function_parameter): Test the presence of
31703 HAVE_window_save properly and do not remap argument registers when
31704 we have a leaf function.
31706 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
31709 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
31711 * config/i386/i386.c (ix86_print_operand): Ditto.
31713 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
31715 * lra-constraints.c (process_alt_operands): Fix recently added comment.
31717 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
31719 PR rtl-optimization/56225
31720 * lra-constraints.c (process_alt_operands): Check that reload hard
31721 reg can hold value for strict_low_part.
31723 2013-02-07 Jakub Jelinek <jakub@redhat.com>
31726 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
31727 dwarf2out_end_function.
31728 (in_first_function_p, maybe_at_text_label_p,
31729 first_loclabel_num_not_at_text_label): New variables.
31730 (dwarf2out_var_location): In the first function find out
31731 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
31732 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
31735 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
31737 PR rtl-optimization/56178
31738 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
31739 SUBREG of a register. Tidy up related block of code.
31740 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
31741 note if the source is a register or a SUBREG of a register.
31743 2013-02-07 Jakub Jelinek <jakub@redhat.com>
31746 * config/rs6000/rs6000.md (ptrm): New mode attr.
31747 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
31748 call_value_indirect_aix<pttrsize>,
31749 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
31752 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
31754 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
31755 if -bnortl. Convert to strcmp and strncmp.
31757 2013-02-07 Alan Modra <amodra@gmail.com>
31760 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
31761 addresses won't wrap when offsetting.
31762 (rs6000_secondary_reload): Provide secondary reloads needed for
31763 wrapping LO_SUM addresses.
31765 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
31767 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
31768 MACH, just __MACH__.
31770 2013-02-06 Richard Biener <rguenther@suse.de>
31772 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
31773 instead of calling fix_loop_structure.
31775 2013-02-06 Jakub Jelinek <jakub@redhat.com>
31777 PR middle-end/56217
31778 * omp-low.c (use_pointer_for_field): Return false if
31779 lower_send_shared_vars doesn't generate any copy-out code.
31781 2013-02-06 Tom de Vries <tom@codesourcery.com>
31783 PR rtl-optimization/56131
31784 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
31785 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
31786 of the label is NULL. Add comment.
31788 2013-02-05 Jakub Jelinek <jakub@redhat.com>
31790 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
31793 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
31794 (STATIC_LIBTSAN_LIBS): Likewise.
31795 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
31796 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
31797 is defined, don't add anything else beyond that.
31798 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
31799 (LINK_COMMAND_SPEC): Use them.
31801 PR tree-optimization/56205
31802 * tree-stdarg.c (check_all_va_list_escapes): Return true if
31803 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
31804 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
31806 2013-02-05 Richard Biener <rguenther@suse.de>
31808 PR tree-optimization/53342
31809 PR tree-optimization/53185
31810 * tree-vectorizer.h (vect_check_strided_load): Remove.
31811 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
31812 not disallow peeling for vectorized strided loads.
31813 (vect_check_strided_load): Make static and simplify.
31814 (vect_analyze_data_refs): Adjust.
31815 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
31816 correctly when vectorizing strided loads.
31818 2013-02-05 Richard Biener <rguenther@suse.de>
31820 * doc/install.texi: Refer to ISL, not PPL.
31822 2013-02-05 Jan Hubicka <jh@suse.cz>
31824 PR tree-optimization/55789
31825 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
31827 2013-02-05 Jan Hubicka <jh@suse.cz>
31829 PR tree-optimization/55789
31830 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
31831 the dead call anyway.
31833 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
31836 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
31838 2013-02-04 Alexander Potapenko <glider@google.com>
31839 Jack Howarth <howarth@bromo.med.uc.edu>
31840 Jakub Jelinek <jakub@redhat.com>
31843 * config/darwin.c (sort_ctor_records): Stabilized qsort
31844 on constructor priority by using original position.
31845 (finalize_ctors): New routine to sort constructors by
31846 priority before use in assemble_integer.
31847 (machopic_asm_out_constructor): Use finalize_ctors if needed.
31849 2013-02-04 Jakub Jelinek <jakub@redhat.com>
31852 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
31853 about visibility on artificial decls.
31854 * config/sol2.c (solaris_assemble_visibility): Likewise.
31856 2013-02-04 Kai Tietz <ktietz@redhat.com>
31859 * config/i386/i386.c (function_value_ms_64): Add additional valtype
31860 argument and improve checking of return-argument types for 16-byte
31862 (ix86_function_value_1): Add additional valtype argument on call
31863 of function_value_64.
31864 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
31865 handling infunction_value_64 function.
31867 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
31869 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
31871 2013-02-04 Richard Biener <rguenther@suse.de>
31873 PR tree-optimization/56188
31874 * tree-ssa-structalias.c (label_visit): Consider case with
31875 initially non-empty points-to set.
31876 (perform_var_substitution): Dump node mapping and clean up.
31878 2013-02-04 Richard Guenther <rguenther@suse.de>
31881 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
31882 node prevail as last resort.
31883 (lto_symtab_merge_decls): Remove guard on LTRANS here.
31884 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
31886 2013-02-04 Richard Biener <rguenther@suse.de>
31888 PR tree-optimization/56113
31889 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
31891 (equiv_class_lookup_or_add): ... this.
31892 (label_visit): Adjust and fix error in previous patch.
31893 (perform_var_substitution): Adjust.
31895 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
31897 * config/sh/divtab.c: Fix formatting and comments throughout the file.
31898 * config/sh/sh4-300.md: Likewise.
31899 * config/sh/sh4a.md: Likewise.
31900 * config/sh/constraints.md: Likewise.
31901 * config/sh/sh.md: Likewise.
31902 * config/sh/netbsd-elf.h: Likewise.
31903 * config/sh/predicates.md: Likewise.
31904 * config/sh/sh-protos.h: Likewise.
31905 * config/sh/ushmedia.h: Likewise.
31906 * config/sh/linux.h: Likewise.
31907 * config/sh/sh.c: Likewise.
31908 * config/sh/superh.h: Likewise.
31909 * config/sh/elf.h: Likewise.
31910 * config/sh/sh4.md: Likewise.
31911 * config/sh/sh.h: Likewise.
31913 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
31915 * config/pa/constraints.md: Adjust unused letters. Change "T"
31916 constraint to match_test floating_point_store_memory_operand().
31917 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
31918 (base14_operand): New.
31919 (floating_point_store_memory_operand): New.
31920 (integer_store_memory_operand): Revise to use base14_operand and
31921 reg_plus_base_memory_operand.
31922 (move_dest_operand): Allow symbolic_memory_operands.
31923 (symbolic_memory_operand): Check for LO_SOM.
31924 (symbolic_operand): Change default case to break.
31925 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
31926 CONST_DOUBLE values to be reloaded by putting them into memory when
31927 the destination is a floating point register.
31928 (movdf): Remove code to handle CONST_DOUBLE.
31930 (reload_indf_r1): New.
31931 (reload_insf_r1): New.
31932 Consistently use "Q" and "T" constraints with integer and floating
31933 point move instructions, respectively.
31934 (movdi): Remove FAIL.
31935 Change predicate for source operand unamed DImode move from
31936 general_operand to move_src_operand.
31937 (umulsidi3): Change predicate for destination operand to
31939 Likewise for similar unamed patterns.
31940 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
31941 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
31942 (hppa_legitimize_address): Simplify mask calculation.
31943 (pa_emit_move_sequence): Revised handling of secondary reloads from
31944 REG+D addresses for floating point loads and stores. Directly handle
31945 loading CONST0_RTX (mode) to a floating point register.
31946 (pa_secondary_reload): Handle reloading DF and SFmode constant values
31947 to floating point registers. Don't restrict secondary reloads to
31948 floating point registers to integer modes. Revise some comments and
31950 (TARGET_LEGITIMATE_ADDRESS_P): Define.
31951 (pa_legitimate_address_p): New.
31952 (pa_legitimize_reload_address): New.
31953 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
31954 (STRICT_REG_OK_FOR_BASE_P): New.
31955 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
31956 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
31958 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
31959 Andrew Dixie <andrewd@gentrack.com>
31961 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
31964 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
31966 * expmed.c (extract_bit_field_1): Pass the full width of the
31967 structure to get_best_reg_extraction_insn.
31969 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
31972 * configure.ac (use_cxa_atexit): Add AIX.
31973 * configure: Regenerate.
31975 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
31977 2013-02-01 Jakub Jelinek <jakub@redhat.com>
31980 * final.c (need_profile_function): New variable.
31981 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
31982 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
31983 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
31984 notes, targetm.asm_out.function_prologue doesn't emit anything,
31985 HAVE_prologue and profiler should be emitted before prologue,
31986 set need_profile_function instead of emitting it.
31987 (final_scan_insn): If need_profile_function, emit
31988 profile_function on the first NOTE_INSN_BASIC_BLOCK or
31989 NOTE_INSN_FUNCTION_BEG note.
31991 2013-02-01 Richard Henderson <rth@redhat.com>
31993 * config/rs6000/rs6000.md (smulditi3): New.
31996 * config/alpha/alpha.md (umulditi3): New.
31998 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
32000 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
32001 (ASM_OUTPUT_ALIGNED_LOCAL): New.
32003 2013-02-01 Richard Biener <rguenther@suse.de>
32005 PR tree-optimization/56113
32006 * tree-ssa-structalias.c (label_visit): Reduce work for
32007 single-predecessor nodes.
32009 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
32011 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
32012 range isn't testing for zero.
32014 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
32016 PR middle-end/56113
32017 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
32019 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
32020 Nick Clifton <nickc@redhat.com>
32022 * config/v850/constraints.md (Q): Define as a memory constraint.
32023 * config/v850/predicates.md (label_ref_operand): New predicate.
32024 (e3v5_shift_operand): New predicate.
32025 (ior_operator): New predicate.
32026 * config/v850/t-v850: Add e3v5 multilib.
32027 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
32028 (v850_gen_movdi): Prototype.
32029 * config/v850/v850.c: Add support for e3v5 architecture.
32030 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
32032 (construct_save_jarl): Add e3v5 long JARL support.
32033 (v850_adjust_insn_length): New function. Adjust length of call
32034 insns when using e3v5 instructions.
32035 (v850_gen_movdi): New function: Generate instructions to move a
32037 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
32038 (CPP_SPEC): Define __v850e3v5__ as appropriate.
32039 (TARGET_USE_FPU): Enable for e3v5.
32040 (CONST_OK_FOR_W): New macro.
32041 (ADJUST_INSN_LENGTH): Define.
32042 * config/v850/v850.md (UNSPEC_LOOP): Define.
32043 (attr cpu): Add v850e3v5.
32044 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
32045 (movdi): New pattern.
32046 (movdi_internal): New pattern.
32047 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
32048 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
32049 (cstoresf4): Likewise.
32050 (cstoredf4): Likewise.
32051 (insv): New pattern.
32052 (rotlso3_a): New pattern.
32053 (rotlsi3_b): New pattern
32054 (rotlsi3_v850e3v5): New pattern.
32055 (doloop_begin): New pattern.
32056 (fix_loop_counter): New pattern.
32057 (doloop_end): New pattern.
32058 (branch_normal): Add e3v5 long branch support.
32059 (branch_invert): Likewise.
32060 (branch_z_normal): Likewise.
32061 (branch_z_invert): Likewise.
32062 (branch_nz_normal): Likewise.
32063 (branch_nz_invert): Likewise.
32064 (call_internal_short): Add e3v5 register-indirect JARL support.
32065 (call_internal_long): Likewise.
32066 (call_value_internal_short): Likewise.
32067 (call_value_internal_long): Likewise.
32068 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
32069 (mloop): New option.
32070 * config.gcc: Add support for configuring v840e3v5 target.
32071 * doc/invoke.texi: Document new v850 specific command line options.
32073 2013-01-31 Paul Koning <ni1d@arrl.net>
32077 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
32078 children if parent is a class.
32079 (prune_unused_types_prune): Don't add DW_AT_declaration.
32081 2013-01-31 Richard Biener <rguenther@suse.de>
32083 PR tree-optimization/56157
32084 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
32085 match up operand with SLP child.
32087 2013-01-31 Jason Merrill <jason@redhat.com>
32090 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
32091 parameters the first time.
32092 (gen_scheduled_generic_parms_dies): Check completeness here.
32094 2013-01-31 Richard Biener <rguenther@suse.de>
32096 PR middle-end/53073
32097 * common.opt (faggressive-loop-optimizations): New flag,
32098 enabled by default.
32099 * doc/invoke.texi (faggressive-loop-optimizations): Document.
32100 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
32101 infer_loop_bounds_from_undefined by it.
32103 2013-01-31 Richard Biener <rguenther@suse.de>
32105 PR tree-optimization/56150
32106 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
32107 visit virtual operands.
32108 (find_uses_to_rename_bb): Likewise.
32110 2013-01-31 Richard Biener <rguenther@suse.de>
32112 PR tree-optimization/56150
32113 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
32114 mixed store non-store stmts.
32116 2013-01-30 Jakub Jelinek <jakub@redhat.com>
32119 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
32120 LIBASAN_EARLY_SPEC is defined.
32121 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
32122 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
32124 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
32127 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
32128 invalid args instead of ICEing on it.
32129 (ix86_valid_target_attribute_tree): Return error_mark_node if
32130 ix86_valid_target_attribute_inner_p failed.
32131 (ix86_valid_target_attribute_p): Return false only if
32132 ix86_valid_target_attribute_tree returned error_mark_node. Allow
32133 target("default") attribute.
32134 (sorted_attr_string): Change argument from const char * to tree,
32135 merge in all target attribute arguments rather than just one.
32136 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
32137 instead of free. Avoid using strcat.
32138 (ix86_mangle_function_version_assembler_name): Mangle
32139 target("default") as if no target attribute is present. Adjust
32140 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
32141 instead of xmalloc and XDELETEVEC instead of free.
32142 (ix86_function_versions): Don't return true if one of the decls
32143 doesn't have target attribute. If they don't and one of the decls
32144 is DECL_FUNCTION_VERSIONED, report an error. Adjust
32145 sorted_attr_string caller. Use XDELETEVEC instead of free.
32146 (ix86_supports_function_versions): Remove.
32147 (make_name): Fix up formatting.
32148 (make_dispatcher_decl): Remove resolver_name and its initialization.
32149 Avoid leaking memory.
32150 (is_function_default_version): Return true if there is
32151 target("default") attribute rather than no target attribute at all.
32152 (make_resolver_func): Avoid leaking memory.
32153 (ix86_generate_version_dispatcher_body): Likewise.
32154 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
32155 * target.def (supports_function_versions): Remove.
32156 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
32157 * doc/tm.texi: Regenerated.
32159 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
32161 PR rtl-optimization/56144
32162 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
32163 for values with side effects.
32165 2013-01-30 Richard Biener <rguenther@suse.de>
32167 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
32168 (sparseset_pop): Likewise.
32169 * cfganal.c (compute_idf): Likewise. Increase work-stack size
32170 to be able to use quick_push in the worker loop.
32172 2013-01-30 Marek Polacek <polacek@redhat.com>
32174 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
32176 2013-01-30 Richard Biener <rguenther@suse.de>
32179 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
32181 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
32183 PR tree-optimization/56064
32184 * fixed-value.c (fixed_from_double_int): New function.
32185 * fixed-value.h (fixed_from_double_int): New prototype.
32186 (const_fixed_from_double_int): New static inline function.
32187 * fold-const.c (native_interpret_fixed): New static function.
32188 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
32189 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
32190 (native_encode_fixed): New static function.
32191 (native_encode_expr) <FIXED_CST>: Use it.
32192 (native_interpret_int): Move double_int worker code to...
32193 * double-int.c (double_int::from_buffer): ...this new static method.
32194 * double-int.h (double_int::from_buffer): Prototype it.
32196 2013-01-30 Richard Biener <rguenther@suse.de>
32198 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
32199 New pointer-map and obstack.
32200 (init_alias_vars): Allocate pointer-map and obstack.
32201 (delete_points_to_sets): Free them.
32202 (find_what_var_points_to): Cache result.
32203 (find_what_p_points_to): Adjust for changed interface of
32204 find_what_var_points_to.
32205 (compute_points_to_sets): Likewise.
32206 (ipa_pta_execute): Likewise.
32208 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
32210 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
32211 * configure: Regenerate.
32212 * config.in: Regenerate.
32213 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
32214 #nobits/#progbits if supported.
32216 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
32219 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
32220 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
32221 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
32223 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
32225 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
32226 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
32228 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
32230 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
32231 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
32233 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
32235 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
32237 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
32238 * config/arm/cortex-a7.md: New bypasses using
32239 arm_mac_accumulator_is_result.
32241 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
32243 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
32244 (cortex_a7_neon_mla): Likewise.
32245 (cortex_a7_fpfmad): New reservation.
32246 (cortex_a7_fpmacs): Use ffmas and update required units.
32247 (cortex_a7_fpmuld): Update required units and latency.
32248 (cortex_a7_fpmacd): Likewise.
32249 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
32250 (cortex_a7_neon). Likewise.
32251 (bypass) Update participating units.
32253 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
32255 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
32256 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
32258 * config/arm/vfp11.md (vfp_farith): Use ffmas.
32259 (vfp_fmul): Use ffmad.
32260 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
32261 (cortex_r4_fmacd): Use ffmad.
32262 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
32263 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
32264 (cortex_a9_fmacd): Use ffmad.
32265 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
32266 (cortex_a8_vfp_macd): Use ffmad.
32267 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
32268 (cortex_a5_fpmacd): Use ffmad.
32269 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
32270 (cortex_a15_vfp_macd): Use ffmad.
32271 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
32273 2013-01-29 Jason Merrill <jason@redhat.com>
32276 * varasm.c (default_assemble_visibility): Don't warn about
32277 visibility on artificial decls.
32279 2013-01-29 Richard Biener <rguenther@suse.de>
32281 PR tree-optimization/56113
32282 * tree-ssa-structalias.c (equiv_class_lookup): Also return
32284 (label_visit): Free duplicate bitmaps and record the leader instead.
32285 (perform_var_substitution): Adjust.
32287 2013-01-29 Richard Biener <rguenther@suse.de>
32289 PR tree-optimization/55270
32290 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
32291 the CFG, schedule loops for fixup.
32293 2013-01-29 Nick Clifton <nickc@redhat.com>
32295 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
32298 2013-01-28 Leif Ekblad <leif@rdos.net>
32300 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
32301 * config/i386/i386.h (TARGET_RDOS): New macro.
32302 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
32303 * config/i386/i386.c (ix86_option_override_internal): For 64bit
32304 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
32305 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
32306 DEFAULT_LARGE_SECTION_THRESHOLD.
32307 * config/i386/i386.md (R14_REG, R15_REG): New constants.
32308 * config/i386/rdos.h: New file.
32309 * config/i386/rdos64.h: New file.
32311 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
32314 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
32317 2013-01-28 Jakub Jelinek <jakub@redhat.com>
32319 PR rtl-optimization/56117
32320 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
32321 call cselib_lookup_from_insn on the MEM before calling
32322 add_insn_mem_dependence.
32324 2013-01-28 Richard Biener <rguenther@suse.de>
32326 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
32327 to a stmt that didn't have one.
32328 (copy_phis_for_bb): Likewise for PHI arguments.
32329 (copy_debug_stmt): Likewise for debug stmts.
32331 2013-01-28 Richard Biener <rguenther@suse.de>
32333 PR tree-optimization/56034
32334 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
32335 (partition_builtin_p): Adjust.
32336 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
32337 it is the last partition.
32338 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
32339 up the vertex for the definition.
32340 (classify_partition): Classify whether a partition is a
32341 PKIND_REDUCTION, thus has uses outside of the loop.
32342 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
32343 Merge all PKIND_REDUCTION partitions into the last partition.
32344 (tree_loop_distribution): Seed partitions from reductions as well.
32346 2013-01-28 Jakub Jelinek <jakub@redhat.com>
32348 PR tree-optimization/56125
32349 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
32350 pow(x,c) into sqrt(x) * powi(x, n/2) or
32351 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
32352 optimizing for size.
32353 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
32354 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
32357 PR tree-optimization/56094
32358 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
32359 to UNKNOWN_LOCATION while gimplifying expr.
32361 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
32364 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
32365 operand 0 in movabs insn template for -masm=intel asm alternative.
32366 (*movabs<mode>_2): Ditto for operand 1.
32368 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
32371 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
32374 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
32376 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
32379 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
32381 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
32382 Undef to avoid warning.
32384 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
32386 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
32387 * configure: Regenerate.
32389 2013-01-25 Jakub Jelinek <jakub@redhat.com>
32391 PR tree-optimization/56098
32392 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
32393 for stmts with volatile ops.
32394 (cond_store_replacement): Don't optimize if assign has volatile ops.
32395 (cond_if_else_store_replacement_1): Don't optimize if either
32396 then_assign or else_assign have volatile ops.
32397 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
32400 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
32402 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
32404 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
32406 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
32407 missing ':' in asm example.
32409 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
32411 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
32412 entries into lane and laneq entries.
32413 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
32414 Remove AdvSIMD scalar modes.
32415 (aarch64_sq<r>dmulh_laneq<mode>): New.
32416 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
32418 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
32419 builtin implementations to relfect changes in RTL in aarch64-simd.md.
32420 * config/aarch64/iterators.md (VCOND): New.
32423 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
32426 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
32427 Add NULL LIBNAME argument to existing definitions.
32428 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
32429 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
32430 * config/avr/avr.c (DEF_BUILTIN): Same.
32431 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
32432 (avr_expand_builtin): Expand to a vanilla call if a libgcc
32433 implementation is available (DECL_ASSEMBLER_NAME is set).
32434 (avr_fold_absfx): New static function.
32435 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
32436 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
32437 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
32438 AVR_BUILTIN_ABSLLK.
32439 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
32440 (abshk, absk, abslk, absllk): Provide as static inline functions.
32442 2013-01-25 Marek Polacek <polacek@redhat.com>
32444 PR tree-optimization/56035
32445 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
32447 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
32449 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
32450 (*movtf_internal_rex64): Add (!o,C) alternative
32451 (*movxf_internal_rex64): Ditto.
32452 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
32454 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
32456 * doc/invoke.texi: fix typo.
32457 * doc/objc.texi: fix typo.
32459 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
32461 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
32462 for the first two alternatives.
32464 2013-01-24 Diego Novillo <dnovillo@google.com>
32466 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
32467 (ggc-zone.o): Remove.
32468 * configure.ac: Remove option --with-gc.
32469 * configure: Re-generate.
32470 * doc/install.texi: Remove documentation for --with-gc.
32471 * gengtype.c (write_enum_defn): Remove. Update all users.
32472 (write_Types_process_field): Remove generation of gt_e_* argument.
32473 (output_type_enum): Remove. Update all users.
32474 (write_enum_defn): Remove. Update all users.
32475 (enum alloc_zone): Remove. Update all users.
32476 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
32477 * ggc-common.c (ggc_splay_alloc): Remove first argument.
32478 Update all callers.
32479 (struct ptr_data): Remove field TYPE. Update all users.
32480 (gt_pch_note_object): Remove argument TYPE. Update all users.
32481 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
32483 * ggc-none.c (ggc_alloc_typed_stat): Remove.
32484 (struct alloc_zone): Remove.
32485 (ggc_internal_alloc_zone_stat): Remove.
32486 (ggc_internal_cleared_alloc_zone_stat): Remove.
32487 * ggc-page.c (ggc_alloc_typed_stat): Remove.
32488 (ggc_pch_count_object): Remove last argument. Update all users.
32489 (ggc_pch_alloc_object): Remove last argument. Update all users.
32490 (struct alloc_zone): Remove.
32491 * ggc-zone.c: Remove.
32492 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
32493 (struct alloc_zone): Remove.
32494 (ggc_alloc_typed_stat): Remove.
32495 (ggc_alloc_typed): Remove.
32496 (ggc_splay_alloc): Remove first argument.
32497 (rtl_zone): Remove. Update all users.
32498 (tree_zone): Remove. Update all users.
32499 (tree_id_zone): Remove. Update all users.
32500 (ggc_internal_zone_alloc_stat): Remove. Update all users.
32501 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
32502 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
32503 * tree-ssanames.c: Remove references to zone allocator in comments.
32505 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
32507 * config/avr/avr.c (avr_out_fract): Make register numbers that
32508 might be outside of source operand signed.
32510 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
32512 * config/i386/constraints.md (Yf): New constraint.
32513 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
32514 of f constraint to conditionaly disable x87 register preferences.
32515 (*movdf_internal): Ditto.
32516 (*movsf_internal): Ditto.
32518 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
32520 PR inline-asm/55934
32521 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
32522 that have operands with impossible constraints.
32523 Add a FIXME for a speed-up opportunity.
32524 * lra-constraints.c (process_alt_operands): Verify that a class
32525 selected from constraints on asms is valid for the operand mode.
32526 (curr_insn_transform): Remove incorrect comment.
32528 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
32530 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
32531 TOC operand is a valid symbol ref in the constant pool.
32533 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
32535 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
32537 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
32540 * config/avr/stdfix.h: New file.
32541 * t-avr (stdfix-gcc.h): New rule to build it.
32542 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
32544 2013-01-23 Kostya Serebryany <kcc@google.com>
32546 * config/darwin.h: remove dependency on
32547 CoreFoundation (asan on Mac OS).
32549 2013-01-23 Jakub Jelinek <jakub@redhat.com>
32552 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
32553 instead of cmpdi_operand for first comparison operand.
32554 Don't assert that comparison operands aren't both constants.
32556 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
32558 * doc/install.texi (Downloading the Source): Update references to
32559 downloading separate components.
32561 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
32563 * doc/extend.texi (__int128): Improve grammar.
32565 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
32568 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
32569 alternative to (o,r).
32570 (*movdi_internal_rex64): Remove (!o,n) alternative.
32571 (DImode immediate->memory splitter): Remove.
32572 (DImode immediate->memory peephole2): Remove.
32573 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
32574 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
32575 alternative to (!o,*r).
32576 (*movtf_internal_sse): New pattern.
32577 (*movxf_internal_rex64): New pattern.
32578 (*movxf_internal): Disable for TARGET_64BIT.
32579 (*movdf_internal_rex64): Remove (!o,F) alternative.
32581 2013-01-22 Jakub Jelinek <jakub@redhat.com>
32583 PR middle-end/56074
32584 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
32585 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
32586 * tree-vect-loop-manip.c (find_loop_location): Also ignore
32587 stmt locations where LOCATION_LOCUS of the stmt location is
32588 UNKNOWN_LOCATION or BUILTINS_LOCATION.
32591 * config/i386/i386.md (UNSPEC_STOS): New.
32592 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
32593 *strsetqi_1): Add UNSPEC_STOS.
32595 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
32598 * doc/invoke.texi: Remove left over -Wsynth example.
32600 2013-01-21 Jakub Jelinek <jakub@redhat.com>
32602 PR tree-optimization/56051
32603 * fold-const.c (fold_binary_loc): Don't fold
32604 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
32605 a narrowing conversion, or widening conversion from signed
32608 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
32610 PR rtl-optimization/56023
32611 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
32612 dependent on debug instruction.
32614 2013-01-21 Martin Jambor <mjambor@suse.cz>
32616 PR middle-end/56022
32617 * function.c (allocate_struct_function): Call
32618 invoke_set_current_function_hook earlier.
32620 2013-01-21 Jakub Jelinek <jakub@redhat.com>
32622 * reload1.c (init_reload): Only initialize reload_obstack
32623 during the first call.
32625 2013-01-21 Marek Polacek <polacek@redhat.com>
32627 * cfgloop.c (verify_loop_structure): Fix up grammar.
32629 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
32631 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
32632 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
32634 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
32637 * config/arm/marvell-pj4.md: Update copyright year.
32638 Fix up use of alu to alu_reg and simple_alu_imm.
32640 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
32642 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
32644 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
32647 * lra-constraints.c (curr_insn_transform): Don't reuse original
32648 insn for secondary memory move when memory mode should be different.
32650 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
32652 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
32653 atomic_storedi_1): New patterns.
32655 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
32657 btver2 pipeline descriptions.
32658 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
32660 * config/i386/i386.md (btver2_decode): New type attributes.
32661 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
32663 * config/i386/btver2.md: New file describing btver2 pipelines.
32665 2013-01-19 Andrew Pinski <apinski@cavium.com>
32667 PR tree-optimization/52631
32668 * tree-ssa-sccvn (visit_use): Before looking up the original
32669 statement, try looking up the simplified expression.
32671 2013-01-19 Anthony Green <green@moxielogic.com>
32673 * config/moxie/moxie.c (moxie_expand_prologue): Set
32674 current_function_static_stack_size.
32676 2013-01-18 Jakub Jelinek <jakub@redhat.com>
32678 PR tree-optimization/56029
32679 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
32680 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
32682 2013-01-18 Sharad Singhai <singhai@google.com>
32684 PR tree-optimization/55995
32685 * dumpfile.c (dump_loc): Print location only if available.
32686 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
32688 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
32691 * lra-constraints.c (curr_insn_transform): Reuse original insn for
32692 secondary memory move.
32693 (inherit_reload_reg): Use rclass instead of cl for
32694 check_secondary_memory_needed_p.
32696 2013-01-18 Jakub Jelinek <jakub@redhat.com>
32698 PR middle-end/56015
32699 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
32700 the case where writing real complex part of target modifies op1.
32702 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
32704 * config/aarch64/aarch64-simd.md
32705 (aarch64_vcond_internal<mode>): Handle unordered cases.
32706 * config/aarch64/iterators.md (v_cmp_result): New.
32708 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
32709 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
32711 * config/arm/marvell-pj4.md: New file.
32712 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
32713 * config/arm/arm.md (generic_sched): Add marvell_pj4.
32714 (generic_vfp): Likewise.
32715 * config/arm/arm-cores.def: Add marvell-pj4.
32716 * config/arm/arm-tune.md: Regenerate.
32717 * config/arm/arm-tables.opt: Regenerate.
32718 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
32719 * doc/invoke.texi: Document marvell-pj4.
32721 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
32723 * config/aarch64/arm_neon.h: Map scalar types to standard types.
32725 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
32730 * var-tracking.c (negative_power_of_two_p): New.
32731 (global_get_addr_cache, local_get_addr_cache): New.
32732 (get_addr_from_global_cache, get_addr_from_local_cache): New.
32733 (vt_canonicalize_addr): Rewrite using the above. Adjust the
32735 (vt_stack_offset_p): Remove.
32736 (vt_canon_true_dep): Always canonicalize loc's address.
32737 (clobber_overlapping_mems): Make sure we have a MEM.
32738 (local_get_addr_clear_given_value): New.
32739 (val_reset): Clear local cached entries.
32740 (compute_bb_dataflow): Create and release the local cache.
32741 Disable duplicate MEMs clobbering.
32742 (emit_notes_in_bb): Clobber MEMs likewise.
32743 (vt_emit_notes): Create and release the local cache.
32744 (vt_initialize, vt_finalize): Create and release the global
32745 cache, respectively.
32746 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
32748 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
32750 PR libmudflap/53359
32751 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
32752 not found in the symtab.
32754 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
32757 PR rtl-optimization/55547
32758 PR rtl-optimization/53827
32761 * alias.c (offset_overlap_p): New, factored out of...
32762 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
32763 the conservative special case for symbolic constants. Don't
32764 adjust zero sizes on alignment.
32766 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
32768 PR rtl-optimization/52573
32769 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
32770 REG_UNUSED for the same register.
32772 2013-01-17 Richard Biener <rguenther@suse.de>
32773 Marek Polacek <polacek@redhat.com>
32775 PR rtl-optimization/55833
32776 * loop-unswitch.c (unswitch_loops): Move loop verification...
32777 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
32778 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
32779 Set it to true when we're removing a loop from hierarchy tree in
32780 an irreducible region.
32781 (fix_bb_placements): Adjust caller.
32782 (fix_loop_placements): Likewise.
32784 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
32786 * config/avr/builtins.def (DEF_BUILTIN): Factor out
32787 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
32788 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
32789 Remove ID. Adjust comments.
32790 * config/avr/avr-c.c (avr_builtin_name): Remove.
32791 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
32792 * config/avr/avr.c (avr_tolower): New static function.
32793 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
32794 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
32795 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
32798 2013-01-17 Jan Hubicka <jh@suse.cz>
32800 PR tree-optimization/55273
32801 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
32803 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
32806 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
32807 store through atomic_store<mode>_1.
32808 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
32810 2013-01-17 Martin Jambor <mjambor@suse.cz>
32812 PR tree-optimizations/55264
32813 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
32814 for virtual methods.
32815 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
32816 virtual methods before inlining is over.
32817 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
32819 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
32822 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
32824 PR rtl-optimization/56005
32825 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
32826 pending reads for prefetch.
32828 2013-01-16 Ian Bolton <ian.bolton@arm.com>
32830 * config/aarch64/aarch64.md
32831 (*cstoresi_neg_uxtw): New pattern.
32832 (*cmovsi_insn_uxtw): New pattern.
32833 (*<optab>si3_uxtw): New pattern.
32834 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
32835 (*<optab>si3_insn_uxtw): New pattern.
32836 (*bswapsi2_uxtw): New pattern.
32838 2013-01-16 Richard Biener <rguenther@suse.de>
32840 * tree-inline.c (tree_function_versioning): Remove set but
32841 never used variable.
32843 2013-01-16 Richard Biener <rguenther@suse.de>
32845 PR tree-optimization/55964
32846 * tree-flow.h (rename_variables_in_loop): Remove.
32847 (rename_variables_in_bb): Likewise.
32848 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
32849 (copy_loop_before): Adjust and delete update-ssa status.
32850 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
32851 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
32852 (rename_variables_in_loop): Remove.
32853 (slpeel_update_phis_for_duplicate_loop): Likewise.
32854 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
32855 use available cfg machinery instead of duplicating it.
32856 Update PHI nodes and perform poor-mans SSA update here.
32857 (slpeel_tree_peel_loop_to_edge): Adjust.
32859 2013-01-16 Richard Biener <rguenther@suse.de>
32861 PR tree-optimization/54767
32862 PR tree-optimization/53465
32863 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
32864 (vrp_visit_phi_node): For PHI arguments coming via backedges
32865 drop all symbolical range information.
32866 (execute_vrp): Compute backedges.
32868 2013-01-16 Richard Biener <rguenther@suse.de>
32870 * doc/install.texi: Update CLooG and ISL requirements to
32873 2013-01-16 Christian Bruel <christian.bruel@st.com>
32876 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
32877 (broken_move): Handle UNSPECV_SP_SWITCH_B.
32878 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
32880 2013-01-16 DJ Delorie <dj@redhat.com>
32882 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
32883 (UNSPECV_SP_SWITCH_E): New.
32884 (sp_switch_1): Change to an unspec.
32885 (sp_switch_2): Change to an unspec. Don't use post-inc when we
32888 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
32890 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
32891 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
32892 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
32893 (expand_mem_thread_fence): Ditto.
32894 (expand_mem_signal_fence): Ditto.
32895 (expand_atomic_load): Ditto.
32896 (expand_atomic_store): Ditto.
32898 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
32900 PR rtl-optimization/55547
32901 PR rtl-optimization/53827
32904 * alias.c (memrefs_conflict_p): Set sizes to negative after
32907 2013-01-15 Jakub Jelinek <jakub@redhat.com>
32910 * function.c (thread_prologue_and_epilogue_insns): Always
32911 add crtl->drap_reg to set_up_by_prologue.set, even if
32912 stack_realign_drap is false.
32914 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
32916 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
32917 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
32918 *call): Fix indention.
32920 2013-01-15 Tom de Vries <tom@codesourcery.com>
32923 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
32926 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
32928 PR rtl-optimization/55153
32929 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
32931 2013-01-15 Martin Jambor <mjambor@suse.cz>
32933 PR tree-optimization/55920
32934 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
32935 accesses as grp_to_be_debug_replaced.
32937 2013-01-15 Jakub Jelinek <jakub@redhat.com>
32939 PR tree-optimization/55920
32940 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
32941 there is non-useless type conversion needed from debug rhs to lhs,
32942 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
32944 2013-01-15 Joseph Myers <joseph@codesourcery.com>
32945 Mikael Pettersson <mikpe@it.uu.se>
32948 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
32950 (ASM_OUTPUT_CASE_LABEL): Remove.
32951 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
32952 * final.c (shorten_branches): Update alignment of labels before
32953 jump tables if CASE_VECTOR_SHORTEN_MODE.
32955 2013-01-15 Richard Biener <rguenther@suse.de>
32958 * system.h: Do not include gmp.h for building host tools.
32960 2013-01-15 Richard Biener <rguenther@suse.de>
32962 PR middle-end/55882
32963 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
32964 account for bitpos when computing alignment.
32966 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
32968 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
32969 (ix86_target_macros_internal): Likewise.
32971 * config/i386/i386.c (m_CORE2I7): Removed.
32972 (m_CORE_HASWELL): New macro.
32973 (m_CORE_ALL): Likewise.
32974 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
32975 (initial_ix86_arch_features): Likewise.
32976 (processor_target_table): Initializations for Core avx2.
32977 (cpu_names): New names "core-avx2".
32978 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
32979 PROCESSOR_CORE_HASWELL.
32980 (ix86_issue_rate): New case.
32981 (ia32_multipass_dfa_lookahead): Likewise.
32982 (ix86_sched_init_global): Likewise.
32984 * config/i386/i386.h (TARGET_HASWELL): New macro.
32985 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
32986 (processor_type): New PROCESSOR_HASWELL.
32988 2013-01-15 Jakub Jelinek <jakub@redhat.com>
32990 PR tree-optimization/55955
32991 * tree-vect-loop.c (vectorizable_reduction): Give up early on
32992 *SHIFT_EXPR and *ROTATE_EXPR codes.
32994 PR tree-optimization/48766
32995 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
32996 -ftrapv disable -fwrapv.
32998 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
33001 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
33002 etc. to 1 and not to __flash.
33003 Use LL suffix for __INT24_MAX__ with -mint8.
33004 Use ULL suffix for __UINT24_MAX__ with -mint8.
33006 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
33008 * config/avr/avr-arch.h
33009 (struct base_arch_s): Use typedef avr_arch_t instead.
33010 (struct arch_info_s): Use typedef avr_arch_info_t instead.
33011 (struct mcu_type_s): Use typedef avr_mcu_t instead.
33012 * config/avr/avr.c: Same.
33013 * config/avr/avr-devices.c: Same.
33014 * config/avr/driver-avr.c: Same.
33015 * config/avr/gen-avr-mmcu-texi.c: Same.
33016 * config/avr/avr-mcus.def: Adjust comment.
33018 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
33020 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
33021 * config/aarch64/iterators.md (VALLDI): New.
33023 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
33024 Andi Kleen <ak@linux.intel.com>
33027 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
33028 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
33031 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
33033 * config/avr/avr-stdint.h: Remove trailing blanks.
33034 * config/avr/avr-log.h: Same.
33035 * config/avr/avr-arch.h: Same.
33036 * config/avr/avr-devices.c: Same.
33037 * config/avr/avr-dimode.md: Same.
33038 * config/avr/predicates.md: Same.
33039 * config/avr/avr-c.c: Same. And fix typo.
33041 * config/avr/avr-protos.h: Same. And:
33042 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
33043 (init_cumulative_args): Rename to avr_init_cumulative_args.
33044 (expand_prologue): Rename to avr_expand_prologue.
33045 (expand_epilogue): Rename to avr_expand_epilogue.
33046 (adjust_insn_length): Rename to avr_adjust_insn_length.
33047 (notice_update_cc): Rename to avr_notice_update_cc.
33048 (final_prescan_insn): Rename to avr_final_prescan_insn.
33049 * config/avr/avr.c: Same.
33050 * config/avr/avr.h: Same.
33051 * config/avr/avr.md: Remove trailing blanks.
33052 (prologue): Use avr_expand_prologue.
33053 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
33055 2013-01-14 Richard Biener <rguenther@suse.de>
33057 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
33058 verify_location, collect_subblocks): New functions.
33059 (verify_gimple_in_cfg): Verify that locations only reference
33060 BLOCKs in the functions BLOCK tree.
33062 2013-01-14 Richard Biener <rguenther@suse.de>
33064 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
33066 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
33068 (insert_out_of_ssa_copy_on_edge): Likewise.
33069 (rewrite_close_phi_out_of_ssa): Likewise.
33070 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
33072 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
33073 propagated constants.
33074 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
33077 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
33079 * config/avr/avr-modes.def: Add GPL copyright notice.
33081 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
33083 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
33084 MEMMODEL_MASK to determine memory model.
33085 (atomic_store<mode>): Ditto from operands[2].
33086 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
33088 2013-01-13 Jakub Jelinek <jakub@redhat.com>
33091 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
33092 (fold_gimple_assign): Don't call unshare_expr here.
33093 (fold_ctor_reference): Call unshare_expr.
33095 2013-01-13 Terry Guo <terry.guo@arm.com>
33097 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
33098 * doc/fragments.texi: Document MULTILIB_REUSE.
33099 * gcc.c (multilib_reuse): New internal spec.
33100 (set_multilib_dir): Also search multilib from multilib_reuse.
33101 * genmultilib (tmpmultilib3): Refactor code.
33102 (tmpmultilib4): Ditto.
33103 (multilib_reuse): New multilib argument.
33105 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
33107 * Makefile.in: Update copyright.
33109 2013-01-12 Tom de Vries <tom@codesourcery.com>
33111 PR middle-end/55890
33112 * calls.c (expand_call): Check if arg_nr is valid.
33114 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
33116 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
33117 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
33118 documentation. Add missing '__' in front of
33119 __builtin_ia32_packssdw256.
33121 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
33124 * config/s390/s390.c (s390_preferred_reload_class): Do not return
33125 NO_REGS for larl operands.
33126 (s390_reload_larl_operand): Use s390_load_address instead of
33129 2013-01-11 Richard Biener <rguenther@suse.de>
33131 * tree-cfg.c (verify_node_sharing_1): Split out from ...
33132 (verify_node_sharing): ... here.
33133 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
33135 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
33137 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
33138 Substitute TREECHECKING.
33139 * configure: Regenerate.
33140 * Makefile.in (TREECHECKING): New.
33142 2013-01-11 Richard Guenther <rguenther@suse.de>
33144 PR tree-optimization/44061
33145 * tree-vrp.c (extract_range_basic): Compute zero as
33146 value-range for __builtin_constant_p of function parameters.
33148 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
33150 Update copyright years.
33152 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
33154 PR rtl-optimization/55672
33155 * lra-eliminations.c (mark_not_eliminable): Permit addition with
33156 const to be eliminable.
33158 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
33160 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
33161 * configure: Regenerate.
33163 2013-01-10 Richard Biener <rguenther@suse.de>
33165 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
33167 2013-01-10 Richard Biener <rguenther@suse.de>
33170 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
33171 locations for virtual PHI arguments.
33172 (rewrite_update_phi_arguments): Likewise.
33174 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
33176 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
33179 2013-01-10 Jakub Jelinek <jakub@redhat.com>
33181 PR tree-optimization/55921
33182 * tree-complex.c (expand_complex_asm): New function.
33183 (expand_complex_operations_1): Call it for GIMPLE_ASM.
33185 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
33188 * config/s390/s390.c (s390_symref_operand_p)
33189 (s390_loadrelative_operand_p): Merge the two functions.
33190 (s390_check_qrst_address, print_operand_address): Add parameters
33191 to s390_loadrelative_operand_p invokation.
33192 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
33193 (s390_reload_larl_operand, s390_secondary_reload): Use
33194 s390_loadrelative_operand_p instead of s390_symref_operand_p.
33195 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
33197 2013-01-09 Mike Stump <mikestump@comcast.net>
33199 * dse.c (record_store): Remove unnecessary assert.
33201 2013-01-09 Jan Hubicka <jh@suse.cz>
33203 PR tree-optimization/55569
33204 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
33205 * cfgloop.h (scale_loop_profile): Likewise.
33207 2013-01-09 Jan Hubicka <jh@suse.cz>
33210 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
33212 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
33214 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
33216 PR middle-end/55114
33217 * expr.h (maybe_emit_group_store): Declare.
33218 * expr.c (maybe_emit_group_store): New function.
33219 * builtins.c (expand_builtin_int_roundingfn): Call it.
33220 (expand_builtin_int_roundingfn_2): Likewise.
33222 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
33224 PR rtl-optimization/55829
33225 * lra-constraints.c (match_reload): Add code for absent output.
33226 (curr_insn_transform): Add code for reloads of matched inputs
33229 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
33231 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
33232 attribute of movddup insn to DF.
33233 (*vec_interleave_lowv2df): Ditto.
33234 (vec_dupv2df): Ditto.
33236 2013-01-09 Jan Hubicka <jh@suse.cz>
33238 PR tree-optimiation/55875
33239 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
33240 EVERY_ITERATION parameter.
33241 (number_of_iterations_exit): Check if exit is executed every iteration.
33242 (idx_infer_loop_bounds): Similarly here.
33243 (n_of_executions_at_most): Simplify
33244 to only test for cases where statement is dominated by the
33245 particular bound; handle correctly the "postdominance" test.
33246 (scev_probably_wraps_p): Use max loop iterations info
33247 as a global bound first.
33249 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
33250 Nick Clifton <nickc@redhat.com>
33252 * config/v850/v850.md (cbranchsf4): New pattern.
33253 (cstoresf4): New pattern.
33254 (cbranchdf4): New pattern.
33255 (cstoredf4): New pattern.
33256 (movsicc): Disallow floating point comparisons.
33257 (cmpsf_le_insn): Fix order of operators.
33258 (cmpsf_lt_insn): Likewise.
33259 (cmpsf_eq_insn): Likewise.
33260 (cmpdf_le_insn): Likewise.
33261 (cmpdf_lt_insn): Likewise.
33262 (cmpdf_eq_insn): Likewise.
33263 (cmpsf_ge_insn): Use LE comparison.
33264 (cmpdf_ge_insn): Likewise.
33265 (cmpsf_gt_insn): Use LT comparison.
33266 (cmpdf_gt_insn): Likewise.
33267 (cmpsf_ne_insn): Delete pattern.
33268 (cmpdf_ne_insn): Delete pattern.
33269 * config/v850/v850.c (v850_gen_float_compare): Use
33270 gen_cmpdf_eq_insn for NE comparison.
33271 (v850_float_z_comparison_operator)
33272 (v850_float_nz_comparison_operator): Move from here ...
33273 * config/v850/predicates.md: ... to here. Move GT and GE
33274 comparisons into v850_float_z_comparison_operator.
33275 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
33277 (v850_float_nz_comparison_operator): Likewise.
33279 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
33281 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
33282 with calls to gen_insvsi/gen_insvdi.
33284 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
33286 * config/i386/i386.c (initial_ix86_tune_features): Set up
33287 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
33289 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
33290 Jakub Jelinek <jakub@redhat.com>
33292 PR tree-optimization/48189
33293 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
33294 If nitercst is 0, don't predict the exit edge.
33296 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
33298 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
33299 in asm_fprintf with reg_names.
33300 (aarch64_print_operand_address): Likewise.
33301 (aarch64_return_addr): Likewise.
33302 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
33304 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
33306 * config/pa/pa.h (VAL_U6_BITS_P): Define.
33307 (INT_U6_BITS): Likewise.
33308 * config/pa/predicates.md (uint6_operand): New predicate.
33309 (shift5_operand, shift6_operand): Likewise.
33310 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
33312 (lshrdi3): Use shift6_operand.
33313 (shrpsi4, shrpdi4): New insn patterns.
33314 (extzv): Delete expander.
33315 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
33316 predicates in unamed zero extract patterns. Tighten common constraint.
33317 (extv): Delete expander.
33318 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
33319 predicates in unamed sign extract patterns. Tighten common constraint.
33320 (insv): Delete expander.
33321 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
33322 predicates in unamed insert patterns. Tighten common constraint.
33323 Change uint32_operand predicate to uint6_operand predicate in unamed
33324 DImode pattern to insert constant values of type 1...1xxxx.
33326 2013-01-04 Jan Hubicka <jh@suse.cz>
33328 PR tree-optimization/55823
33329 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
33332 2013-01-08 Jakub Jelinek <jakub@redhat.com>
33333 Uros Bizjak <ubizjak@gmail.com>
33335 PR rtl-optimization/55845
33336 * df-problems.c (can_move_insns_across): Stop scanning at
33337 volatile_insn_p source instruction or give up if
33338 across_from .. across_to range contains any volatile_insn_p
33341 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
33343 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
33344 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
33346 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
33347 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
33349 2013-01-08 Jakub Jelinek <jakub@redhat.com>
33352 * asan.c (asan_clear_shadow): New function.
33353 (asan_emit_stack_protection): Use it.
33355 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
33357 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
33358 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
33359 with tab instead of space.
33361 2013-01-08 Nick Clifton <nickc@redhat.com>
33363 * config/rl78/rl78.c (rl78_expand_prologue): Always select
33364 register bank 0 at the start of an interrupt handler.
33365 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
33368 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
33370 * config/aarch64/aarch64-simd.md
33371 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
33372 (aarch64_simd_bsl): Likewise.
33373 (aarch64_vcond_internal<mode>): Likewise.
33374 (vcond<mode><mode>): Likewise.
33375 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
33376 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
33378 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
33380 * config/aarch64/aarch64-builtins.c
33381 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
33383 2013-01-08 Martin Jambor <mjambor@suse.cz>
33386 * tree-sra.c (analyze_access_subtree): Return true also after
33387 potentially creating a debug-only replacement.
33389 2013-01-08 Jakub Jelinek <jakub@redhat.com>
33391 PR middle-end/55890
33392 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
33394 PR tree-optimization/54120
33395 * tree-vrp.c (range_fits_type_p): Don't allow
33396 src_precision < precision from signed vr to unsigned_p
33397 if vr->min or vr->max is negative.
33398 (simplify_float_conversion_using_ranges): Test can_float_p
33399 against CODE_FOR_nothing.
33401 2013-01-08 Jakub Jelinek <jakub@redhat.com>
33402 Richard Biener <rguenther@suse.de>
33404 PR middle-end/55851
33405 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
33406 types instead of just INTEGER_TYPE types.
33408 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
33410 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
33413 2013-01-07 Steve Ellcey <sellcey@mips.com>
33416 * config/mips/mips.opt: Change mad to mmad to match documentation.
33418 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
33421 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
33422 .progmemx.data now.
33424 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
33427 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
33428 (avr_addrspace_t): Add .section_name field.
33429 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
33431 (avr_addrspace): Same. Initialize .section_name. Remove last
33432 NULL entry. Put __memx into .progmemx.data.
33433 (progmem_section_prefix): Remove.
33434 (avr_asm_init_sections): No need to initialize progmem_section.
33435 (avr_asm_named_section): Use avr_addrspace[].section_name to get
33436 section name prefix.
33437 (avr_asm_select_section): Ditto. And use get_unnamed_section to
33438 retrieve the progmem section.
33439 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
33440 boundary to run over avr_addrspace[].
33441 (avr_register_target_pragmas): Ditto.
33443 2013-01-06 Jakub Jelinek <jakub@redhat.com>
33445 * varasm.c (output_constant_def_contents): For asan_protect_global
33446 protected strings, adjust DECL_ALIGN if needed, before testing for
33448 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
33449 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
33451 (output_object_block): For asan protected decls, emit asan padding
33452 after their contents.
33453 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
33454 (asan_finish_file): Test it here instead.
33456 2013-01-07 Nick Clifton <nickc@redhat.com>
33457 Matthias Klose <doko@debian.org>
33458 Doug Kwan <dougkwan@google.com>
33459 H.J. Lu <hongjiu.lu@intel.com>
33462 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
33464 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
33466 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
33468 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
33470 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
33472 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
33475 * doc/install.texi (Cross-Compiler-Specific Options): Document
33478 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
33480 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
33481 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
33482 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
33483 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
33484 vqmovun_high_s64): Fix source operand number and update copyright.
33486 2013-01-07 Richard Biener <rguenther@suse.de>
33488 PR middle-end/55890
33489 * gimple.h (gimple_call_builtin_p): New overload.
33490 * gimple.c (validate_call): New function.
33491 (gimple_call_builtin_p): Likewise.
33492 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
33493 Use gimple_call_builtin_p.
33494 (find_func_clobbers): Likewise.
33495 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
33496 (strlen_optimize_stmt): Likewise.
33498 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
33500 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
33501 (vld1q_dup_*): Likewise.
33502 (vld1_*): Likewise.
33503 (vld1q_*): Likewise.
33504 (vld1_lane_*): Likewise.
33505 (vld1q_lane_*): Likewise.
33507 2013-01-07 Richard Biener <rguenther@suse.de>
33509 * lto-streamer.h (LTO_minor_version): Bump to 2.
33511 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
33513 * config/aarch64/aarch64-protos.h
33514 (aarch64_const_double_zero_rtx_p): Rename to...
33515 (aarch64_float_const_zero_rtx_p): ...this.
33516 (aarch64_float_const_representable_p): New.
33517 (aarch64_output_simd_mov_immediate): Likewise.
33518 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
33519 move immediate case.
33520 * config/aarch64/aarch64.c
33521 (aarch64_const_double_zero_rtx_p): Rename to...
33522 (aarch64_float_const_zero_rtx_p): ...this.
33523 (aarch64_print_operand): Allow printing of new constants.
33524 (aarch64_valid_floating_const): New.
33525 (aarch64_legitimate_constant_p): Check for valid floating-point
33527 (aarch64_simd_valid_immediate): Likewise.
33528 (aarch64_vect_float_const_representable_p): New.
33529 (aarch64_float_const_representable_p): Likewise.
33530 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
33531 (aarch64_output_simd_mov_immediate): New.
33532 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
33533 (*movdf_aarch64): Likewise.
33534 * config/aarch64/constraints.md (Ufc): New.
33535 (Y): call aarch64_float_const_zero_rtx.
33536 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
33538 2013-01-07 Richard Biener <rguenther@suse.de>
33540 PR tree-optimization/55888
33541 PR tree-optimization/55862
33542 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
33543 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
33544 not if it is contained therein.
33546 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
33548 * config/avr/t-avr: Typo.
33550 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
33553 * config/avr/t-avr: Don't automatically rebuild
33554 $(srcdir)/config/avr/t-multilib
33555 $(srcdir)/config/avr/avr-tables.opt
33556 $(srcdir)/doc/avr-mmcu.texi
33557 (avr-mcus): New phony target to build them on request.
33558 (s-avr-mlib, s-avr-mmcu-texi): Remove.
33559 * avr/avr-mcus.def: Adjust comments.
33561 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
33563 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
33565 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
33567 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
33569 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
33571 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
33573 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
33575 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
33576 to generate profiling.
33577 * config/rs6000/aix64.h (LIB_SPEC): Same.
33579 2013-01-04 Andrew Pinski <apinski@cavium.com>
33581 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
33583 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
33585 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
33587 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
33589 (ix86_expand_move): Ditto.
33590 (ix86_zero_extend_to_Pmode): Ditto.
33591 (ix86_expand_call): Ditto.
33592 (ix86_expand_special_args_builtin): Ditto.
33593 (ix86_expand_builtin): Ditto.
33595 2013-01-04 Richard Biener <rguenther@suse.de>
33597 PR tree-optimization/55862
33598 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
33599 translating them through PHI nodes.
33601 2013-01-04 Martin Jambor <mjambor@suse.cz>
33603 PR tree-optimization/55755
33604 * tree-sra.c (sra_modify_assign): Do not check that an access has no
33605 children when trying to avoid producing a VIEW_CONVERT_EXPR.
33607 2013-01-04 Marek Polacek <polacek@redhat.com>
33609 PR middle-end/55859
33610 * opts.c (default_options_optimization): Clarify error message.
33612 2013-01-04 Richard Biener <rguenther@suse.de>
33614 PR middle-end/55863
33615 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
33618 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
33621 * config/pa/pa.md (movsi): Revert previous change.
33622 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
33625 2013-01-03 Richard Henderson <rth@redhat.com>
33627 * config/i386/i386.c (ix86_expand_move): Always assign to op1
33628 after eliminating TLS symbols.
33630 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
33633 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
33634 * graphite-poly.c (debug_gmp_value): Likewise.
33636 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
33639 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
33640 selected code model, define __code_mode_small__, __code_model_medium__,
33641 __code_model_large__, __code_model_32__ or __code_model_kernel__.
33642 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
33643 xchg temporary register with %k. Declare temporary register as
33645 [__x86_64__]: For medium and large code models, preserve %rbx register.
33647 2013-01-03 Richard Biener <rguenther@suse.de>
33649 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
33650 (dump_subscript): Adjust.
33651 (finalize_ddr_dependent): Do not dump redundant info.
33652 (analyze_siv_subscript): Adjust.
33653 (subscript_dependence_tester): Likewise.
33654 (compute_affine_dependence): Likewise.
33656 2013-01-03 Richard Biener <rguenther@suse.de>
33659 2013-01-03 Richard Biener <rguenther@suse.de>
33661 PR tree-optimization/55857
33662 * tree-vect-stmts.c (vectorizable_load): Do not setup
33663 re-alignment for invariant loads.
33665 2013-01-02 Richard Biener <rguenther@suse.de>
33667 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
33668 invariant load do not generate a vector load from the scalar location.
33670 2013-01-03 Richard Biener <rguenther@suse.de>
33672 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
33673 for not vectorizing.
33674 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
33675 not build INDIRECT_REFs, call get_name once only.
33676 (vect_create_data_ref_ptr): Likewise. Dump base object kind
33677 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
33679 2013-01-03 Richard Biener <rguenther@suse.de>
33681 PR tree-optimization/55857
33682 * tree-vect-stmts.c (vectorizable_load): Do not setup
33683 re-alignment for invariant loads.
33685 2013-01-03 Richard Biener <rguenther@suse.de>
33688 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
33689 prefer a built-in decl.
33691 2013-01-03 Jakub Jelinek <jakub@redhat.com>
33693 * gcc.c (process_command): Update copyright notice dates.
33694 * gcov.c (print_version): Likewise.
33695 * gcov-dump.c (print_version): Likewise.
33697 PR rtl-optimization/55838
33698 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
33699 iv0.step, iv1.step and step.
33701 2013-01-03 Jakub Jelinek <jakub@redhat.com>
33702 Marc Glisse <marc.glisse@inria.fr>
33704 PR tree-optimization/55832
33705 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
33706 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
33707 integer_{one,zero}_node.
33709 2013-01-03 Jakub Jelinek <jakub@redhat.com>
33712 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
33713 * var-tracking.c (reverse_op): Don't add reverse ops to
33714 VALUEs that have already
33715 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
33717 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
33719 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
33721 2013-01-02 Teresa Johnson <tejohnson@google.com>
33723 * dumpfile.c (dump_loc): Print filename with location.
33724 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
33725 new location_t parameter to emit complete unroll message with
33726 new dump framework.
33727 (canonicalize_loop_induction_variables): Compute loops location
33728 and pass to try_unroll_loop_completely.
33729 * loop-unroll.c (report_unroll_peel): New function.
33730 (peel_loops_completely): Use new dump format with location
33731 for main dumpfile message, and invoke report_unroll_peel on success.
33732 (decide_unrolling_and_peeling): Ditto.
33733 (decide_peel_once_rolling): Remove old dumpfile message subsumed
33734 by report_unroll_peel.
33735 (decide_peel_completely): Ditto.
33736 (decide_unroll_constant_iterations): Ditto.
33737 (decide_unroll_runtime_iterations): Ditto.
33738 (decide_peel_simple): Ditto.
33739 (decide_unroll_stupid): Ditto.
33740 * cfgloop.c (get_loop_location): New function.
33741 * cfgloop.h (get_loop_location): Declare.
33743 2013-01-02 Sriraman Tallam <tmsriram@google.com>
33745 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
33748 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
33750 PR middle-end/55198
33751 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
33752 BLKmode objects when EXPAND_MEMORY is specified.
33754 2013-01-02 Sriraman Tallam <tmsriram@google.com>
33756 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
33758 (fold_builtin_cpu): Do not share cpu model decls across statements.
33760 2013-01-02 Jason Merrill <jason@redhat.com>
33763 * tree.c (build_array_type_1): Revert earlier change.
33765 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
33767 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
33769 * config/aarch64/aarch64-tune.md: Re-generate.
33771 2013-01-02 Richard Biener <rguenther@suse.de>
33773 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
33774 invariant load do not generate a vector load from the scalar location.
33776 2013-01-02 Richard Biener <rguenther@suse.de>
33779 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
33780 * configure: Regenerate.
33782 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
33784 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
33785 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
33786 (expand_builtin_int_roundingfn_2): Keep the original target around
33787 for the fallback case.
33789 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
33791 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
33792 to be clear for sign changes.
33794 2013-01-01 Jan Hubicka <jh@suse.cz>
33796 * ipa-inline-analysis.c: Fix formatting.
33798 2013-01-01 Jakub Jelinek <jakub@redhat.com>
33800 PR tree-optimization/55831
33801 * tree-vect-loop.c (get_initial_def_for_induction): Use
33802 gsi_after_labels instead of gsi_start_bb.
33804 Copyright (C) 2013 Free Software Foundation, Inc.
33806 Copying and distribution of this file, with or without modification,
33807 are permitted in any medium without royalty provided the copyright
33808 notice and this notice are preserved.