1 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
4 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
7 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
9 * flag-types.h (enum ivar_visibility): Add.
11 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
13 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
16 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
18 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
20 2014-04-24 Radovan Obradovic <robradovic@mips.com>
21 Tom de Vries <tom@codesourcery.com>
23 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
24 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
26 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
27 * emit-rtl.c (try_split): Same.
29 2014-04-24 Radovan Obradovic <robradovic@mips.com>
30 Tom de Vries <tom@codesourcery.com>
32 * common.opt (fuse-caller-save): New option.
34 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
36 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
39 2014-04-24 Richard Biener <rguenther@suse.de>
41 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
42 during TER and instead use the sepops interface for expanding
43 non-GIMPLE_SINGLE_RHS.
45 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
47 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
48 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
50 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
52 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
53 assembler 64-bit option.
54 * configure: Regenerate.
56 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
58 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
59 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
60 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
61 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
62 (TARGET_CRYPTO): Take TARGET_SIMD into account.
64 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
66 * config/aarch64/aarch64-builtins.c
67 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
68 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
69 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
70 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
72 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
73 (Vrevsuff): New mode attribute.
75 2014-04-24 Terry Guo <terry.guo@arm.com>
77 * config/arm/arm.h (machine_function): Define variable
79 * config/arm/arm.c (after_arm_reorg): Remove the definition.
80 (arm_split_constant): Update the way to access variable
83 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
85 2014-04-23 Tom de Vries <tom@codesourcery.com>
87 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
89 2014-04-23 David Malcolm <dmalcolm@redhat.com>
91 * is-a.h: Update comments to reflect the following changes to the
92 "pointerness" of the API, making the template parameter match the
93 return type, allowing use of is-a.h with typedefs of pointers.
94 (is_a_helper::cast): Return a T rather then a pointer to a T, so
95 that the return type matches the parameter to the is_a_helper.
99 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
100 pointer from the is-a.h API.
102 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
103 (is_a_helper <cgraph_node *>::test): ...this, matching change to
105 (is_a_helper <varpool_node>::test): Likewise, convert to...
106 (is_a_helper <varpool_node *>::test): ...this.
108 (varpool_first_variable): Update for removal of implicit pointer
110 (varpool_next_variable): Likewise.
111 (varpool_first_static_initializer): Likewise.
112 (varpool_next_static_initializer): Likewise.
113 (varpool_first_defined_variable): Likewise.
114 (varpool_next_defined_variable): Likewise.
115 (cgraph_first_defined_function): Likewise.
116 (cgraph_next_defined_function): Likewise.
117 (cgraph_first_function): Likewise.
118 (cgraph_next_function): Likewise.
119 (cgraph_first_function_with_gimple_body): Likewise.
120 (cgraph_next_function_with_gimple_body): Likewise.
121 (cgraph_alias_target): Likewise.
122 (varpool_alias_target): Likewise.
123 (cgraph_function_or_thunk_node): Likewise.
124 (varpool_variable_node): Likewise.
125 (symtab_real_symbol_p): Likewise.
126 * cgraphunit.c (referred_to_p): Likewise.
127 (analyze_functions): Likewise.
128 (handle_alias_pairs): Likewise.
129 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
130 * gimple-ssa.h (gimple_vuse_op): Likewise.
131 (gimple_vdef_op): Likewise.
132 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
133 * gimple.c (gimple_build_asm_1): Likewise.
134 (gimple_build_try): Likewise.
135 (gimple_build_resx): Likewise.
136 (gimple_build_eh_dispatch): Likewise.
137 (gimple_build_omp_for): Likewise.
138 (gimple_omp_for_set_clauses): Likewise.
140 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
141 (is_a_helper <gimple_statement_asm *>::test): ...this.
142 (is_a_helper <gimple_statement_bind>::test): Convert to...
143 (is_a_helper <gimple_statement_bind *>::test): ...this.
144 (is_a_helper <gimple_statement_call>::test): Convert to...
145 (is_a_helper <gimple_statement_call *>::test): ...this.
146 (is_a_helper <gimple_statement_catch>::test): Convert to...
147 (is_a_helper <gimple_statement_catch *>::test): ...this.
148 (is_a_helper <gimple_statement_resx>::test): Convert to...
149 (is_a_helper <gimple_statement_resx *>::test): ...this.
150 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
151 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
152 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
153 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
154 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
155 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
156 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
157 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
158 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
159 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
160 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
161 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
162 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
163 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
164 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
165 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
166 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
167 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
168 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
169 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
170 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
171 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
172 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
173 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
174 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
175 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
176 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
177 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
178 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
179 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
180 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
181 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
182 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
183 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
184 (is_a_helper <gimple_statement_phi>::test): Convert to...
185 (is_a_helper <gimple_statement_phi *>::test): ...this.
186 (is_a_helper <gimple_statement_transaction>::test): Convert to...
187 (is_a_helper <gimple_statement_transaction *>::test): ...this.
188 (is_a_helper <gimple_statement_try>::test): Convert to...
189 (is_a_helper <gimple_statement_try *>::test): ...this.
190 (is_a_helper <gimple_statement_wce>::test): Convert to...
191 (is_a_helper <gimple_statement_wce *>::test): ...this.
192 (is_a_helper <const gimple_statement_asm>::test): Convert to...
193 (is_a_helper <const gimple_statement_asm *>::test): ...this.
194 (is_a_helper <const gimple_statement_bind>::test): Convert to...
195 (is_a_helper <const gimple_statement_bind *>::test): ...this.
196 (is_a_helper <const gimple_statement_call>::test): Convert to...
197 (is_a_helper <const gimple_statement_call *>::test): ...this.
198 (is_a_helper <const gimple_statement_catch>::test): Convert to...
199 (is_a_helper <const gimple_statement_catch *>::test): ...this.
200 (is_a_helper <const gimple_statement_resx>::test): Convert to...
201 (is_a_helper <const gimple_statement_resx *>::test): ...this.
202 (is_a_helper <const gimple_statement_eh_dispatch>::test):
204 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
206 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
208 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
209 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
211 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
213 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
215 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
217 (is_a_helper <const gimple_statement_omp_return>::test): Convert
219 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
220 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
222 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
223 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
225 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
226 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
227 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
228 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
230 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
231 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
233 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
234 (is_a_helper <const gimple_statement_omp_target>::test): Convert
236 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
237 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
239 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
240 (is_a_helper <const gimple_statement_omp_single>::test): Convert
242 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
243 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
245 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
246 (is_a_helper <const gimple_statement_omp_task>::test): Convert
248 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
249 (is_a_helper <const gimple_statement_phi>::test): Convert to...
250 (is_a_helper <const gimple_statement_phi *>::test): ...this.
251 (is_a_helper <const gimple_statement_transaction>::test): Convert
253 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
254 (is_a_helper <const gimple_statement_with_ops>::test): Convert
256 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
257 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
258 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
259 (is_a_helper <const gimple_statement_with_memory_ops>::test):
261 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
263 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
265 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
267 (gimple_use_ops): Update for removal of implicit pointer from the
269 (gimple_set_use_ops): Likewise.
270 (gimple_vuse): Likewise.
271 (gimple_vdef): Likewise.
272 (gimple_vuse_ptr): Likewise.
273 (gimple_vdef_ptr): Likewise.
274 (gimple_set_vuse): Likewise.
275 (gimple_set_vdef): Likewise.
276 (gimple_omp_return_set_lhs): Likewise.
277 (gimple_omp_return_lhs): Likewise.
278 (gimple_omp_return_lhs_ptr): Likewise.
279 (gimple_call_fntype): Likewise.
280 (gimple_call_set_fntype): Likewise.
281 (gimple_call_set_internal_fn): Likewise.
282 (gimple_call_use_set): Likewise.
283 (gimple_call_clobber_set): Likewise.
284 (gimple_bind_vars): Likewise.
285 (gimple_bind_set_vars): Likewise.
286 (gimple_bind_body_ptr): Likewise.
287 (gimple_bind_set_body): Likewise.
288 (gimple_bind_add_stmt): Likewise.
289 (gimple_bind_block): Likewise.
290 (gimple_bind_set_block): Likewise.
291 (gimple_asm_ninputs): Likewise.
292 (gimple_asm_noutputs): Likewise.
293 (gimple_asm_nclobbers): Likewise.
294 (gimple_asm_nlabels): Likewise.
295 (gimple_asm_input_op): Likewise.
296 (gimple_asm_input_op_ptr): Likewise.
297 (gimple_asm_output_op): Likewise.
298 (gimple_asm_output_op_ptr): Likewise.
299 (gimple_asm_set_output_op): Likewise.
300 (gimple_asm_clobber_op): Likewise.
301 (gimple_asm_set_clobber_op): Likewise.
302 (gimple_asm_label_op): Likewise.
303 (gimple_asm_set_label_op): Likewise.
304 (gimple_asm_string): Likewise.
305 (gimple_catch_types): Likewise.
306 (gimple_catch_types_ptr): Likewise.
307 (gimple_catch_handler_ptr): Likewise.
308 (gimple_catch_set_types): Likewise.
309 (gimple_catch_set_handler): Likewise.
310 (gimple_eh_filter_types): Likewise.
311 (gimple_eh_filter_types_ptr): Likewise.
312 (gimple_eh_filter_failure_ptr): Likewise.
313 (gimple_eh_filter_set_types): Likewise.
314 (gimple_eh_filter_set_failure): Likewise.
315 (gimple_eh_must_not_throw_fndecl): Likewise.
316 (gimple_eh_must_not_throw_set_fndecl): Likewise.
317 (gimple_eh_else_n_body_ptr): Likewise.
318 (gimple_eh_else_e_body_ptr): Likewise.
319 (gimple_eh_else_set_n_body): Likewise.
320 (gimple_eh_else_set_e_body): Likewise.
321 (gimple_try_eval_ptr): Likewise.
322 (gimple_try_cleanup_ptr): Likewise.
323 (gimple_try_set_eval): Likewise.
324 (gimple_try_set_cleanup): Likewise.
325 (gimple_wce_cleanup_ptr): Likewise.
326 (gimple_wce_set_cleanup): Likewise.
327 (gimple_phi_capacity): Likewise.
328 (gimple_phi_num_args): Likewise.
329 (gimple_phi_result): Likewise.
330 (gimple_phi_result_ptr): Likewise.
331 (gimple_phi_set_result): Likewise.
332 (gimple_phi_arg): Likewise.
333 (gimple_phi_set_arg): Likewise.
334 (gimple_resx_region): Likewise.
335 (gimple_resx_set_region): Likewise.
336 (gimple_eh_dispatch_region): Likewise.
337 (gimple_eh_dispatch_set_region): Likewise.
338 (gimple_omp_critical_name): Likewise.
339 (gimple_omp_critical_name_ptr): Likewise.
340 (gimple_omp_critical_set_name): Likewise.
341 (gimple_omp_for_clauses): Likewise.
342 (gimple_omp_for_clauses_ptr): Likewise.
343 (gimple_omp_for_set_clauses): Likewise.
344 (gimple_omp_for_collapse): Likewise.
345 (gimple_omp_for_index): Likewise.
346 (gimple_omp_for_index_ptr): Likewise.
347 (gimple_omp_for_set_index): Likewise.
348 (gimple_omp_for_initial): Likewise.
349 (gimple_omp_for_initial_ptr): Likewise.
350 (gimple_omp_for_set_initial): Likewise.
351 (gimple_omp_for_final): Likewise.
352 (gimple_omp_for_final_ptr): Likewise.
353 (gimple_omp_for_set_final): Likewise.
354 (gimple_omp_for_incr): Likewise.
355 (gimple_omp_for_incr_ptr): Likewise.
356 (gimple_omp_for_set_incr): Likewise.
357 (gimple_omp_for_pre_body_ptr): Likewise.
358 (gimple_omp_for_set_pre_body): Likewise.
359 (gimple_omp_parallel_clauses): Likewise.
360 (gimple_omp_parallel_clauses_ptr): Likewise.
361 (gimple_omp_parallel_set_clauses): Likewise.
362 (gimple_omp_parallel_child_fn): Likewise.
363 (gimple_omp_parallel_child_fn_ptr): Likewise.
364 (gimple_omp_parallel_set_child_fn): Likewise.
365 (gimple_omp_parallel_data_arg): Likewise.
366 (gimple_omp_parallel_data_arg_ptr): Likewise.
367 (gimple_omp_parallel_set_data_arg): Likewise.
368 (gimple_omp_task_clauses): Likewise.
369 (gimple_omp_task_clauses_ptr): Likewise.
370 (gimple_omp_task_set_clauses): Likewise.
371 (gimple_omp_task_child_fn): Likewise.
372 (gimple_omp_task_child_fn_ptr): Likewise.
373 (gimple_omp_task_set_child_fn): Likewise.
374 (gimple_omp_task_data_arg): Likewise.
375 (gimple_omp_task_data_arg_ptr): Likewise.
376 (gimple_omp_task_set_data_arg): Likewise.
377 (gimple_omp_taskreg_clauses): Likewise.
378 (gimple_omp_taskreg_clauses_ptr): Likewise.
379 (gimple_omp_taskreg_set_clauses): Likewise.
380 (gimple_omp_taskreg_child_fn): Likewise.
381 (gimple_omp_taskreg_child_fn_ptr): Likewise.
382 (gimple_omp_taskreg_set_child_fn): Likewise.
383 (gimple_omp_taskreg_data_arg): Likewise.
384 (gimple_omp_taskreg_data_arg_ptr): Likewise.
385 (gimple_omp_taskreg_set_data_arg): Likewise.
386 (gimple_omp_task_copy_fn): Likewise.
387 (gimple_omp_task_copy_fn_ptr): Likewise.
388 (gimple_omp_task_set_copy_fn): Likewise.
389 (gimple_omp_task_arg_size): Likewise.
390 (gimple_omp_task_arg_size_ptr): Likewise.
391 (gimple_omp_task_set_arg_size): Likewise.
392 (gimple_omp_task_arg_align): Likewise.
393 (gimple_omp_task_arg_align_ptr): Likewise.
394 (gimple_omp_task_set_arg_align): Likewise.
395 (gimple_omp_single_clauses): Likewise.
396 (gimple_omp_single_clauses_ptr): Likewise.
397 (gimple_omp_single_set_clauses): Likewise.
398 (gimple_omp_target_clauses): Likewise.
399 (gimple_omp_target_clauses_ptr): Likewise.
400 (gimple_omp_target_set_clauses): Likewise.
401 (gimple_omp_target_child_fn): Likewise.
402 (gimple_omp_target_child_fn_ptr): Likewise.
403 (gimple_omp_target_set_child_fn): Likewise.
404 (gimple_omp_target_data_arg): Likewise.
405 (gimple_omp_target_data_arg_ptr): Likewise.
406 (gimple_omp_target_set_data_arg): Likewise.
407 (gimple_omp_teams_clauses): Likewise.
408 (gimple_omp_teams_clauses_ptr): Likewise.
409 (gimple_omp_teams_set_clauses): Likewise.
410 (gimple_omp_sections_clauses): Likewise.
411 (gimple_omp_sections_clauses_ptr): Likewise.
412 (gimple_omp_sections_set_clauses): Likewise.
413 (gimple_omp_sections_control): Likewise.
414 (gimple_omp_sections_control_ptr): Likewise.
415 (gimple_omp_sections_set_control): Likewise.
416 (gimple_omp_for_set_cond): Likewise.
417 (gimple_omp_for_cond): Likewise.
418 (gimple_omp_atomic_store_set_val): Likewise.
419 (gimple_omp_atomic_store_val): Likewise.
420 (gimple_omp_atomic_store_val_ptr): Likewise.
421 (gimple_omp_atomic_load_set_lhs): Likewise.
422 (gimple_omp_atomic_load_lhs): Likewise.
423 (gimple_omp_atomic_load_lhs_ptr): Likewise.
424 (gimple_omp_atomic_load_set_rhs): Likewise.
425 (gimple_omp_atomic_load_rhs): Likewise.
426 (gimple_omp_atomic_load_rhs_ptr): Likewise.
427 (gimple_omp_continue_control_def): Likewise.
428 (gimple_omp_continue_control_def_ptr): Likewise.
429 (gimple_omp_continue_set_control_def): Likewise.
430 (gimple_omp_continue_control_use): Likewise.
431 (gimple_omp_continue_control_use_ptr): Likewise.
432 (gimple_omp_continue_set_control_use): Likewise.
433 (gimple_transaction_body_ptr): Likewise.
434 (gimple_transaction_label): Likewise.
435 (gimple_transaction_label_ptr): Likewise.
436 (gimple_transaction_set_body): Likewise.
437 (gimple_transaction_set_label): Likewise.
439 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
440 * ipa-inline-analysis.c (inline_write_summary): Likewise.
441 * ipa-ref.c (ipa_record_reference): Likewise.
442 * ipa-reference.c (analyze_function): Likewise.
443 (ipa_reference_write_optimization_summary): Likewise.
444 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
445 (address_taken_from_non_vtable_p): Likewise.
446 (comdat_can_be_unshared_p_1): Likewise.
447 * lto-cgraph.c (lto_output_ref): Likewise.
448 (add_references): Likewise.
449 (compute_ltrans_boundary): Likewise.
450 (output_symtab): Likewise.
451 (input_ref): Likewise.
452 (input_cgraph_1): Likewise.
453 (output_cgraph_opt_summary): Likewise.
454 * lto-streamer-out.c (lto_output): Likewise.
455 (output_symbol_p): Likewise.
456 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
457 (lsei_start_function_in_partition): Likewise.
458 (lsei_next_variable_in_partition): Likewise.
459 (lsei_start_variable_in_partition): Likewise.
460 * symtab.c (insert_to_assembler_name_hash): Likewise.
461 (unlink_from_assembler_name_hash): Likewise.
462 (symtab_unregister_node): Likewise.
463 (symtab_remove_node): Likewise.
464 (dump_symtab_node): Likewise.
465 (verify_symtab_base): Likewise.
466 (verify_symtab_node): Likewise.
467 (symtab_make_decl_local): Likewise.
468 (symtab_alias_ultimate_target): Likewise.
469 (symtab_resolve_alias): Likewise.
470 (symtab_get_symbol_partitioning_class): Likewise.
471 * tree-phinodes.c (allocate_phi_node): Likewise.
472 (reserve_phi_args_for_new_edge): Likewise.
473 (remove_phi_args): Likewise.
474 * varpool.c (varpool_node_for_asm): Likewise.
475 (varpool_remove_unreferenced_decls): Likewise.
477 2014-04-23 Jeff Law <law@redhat.com>
479 PR tree-optimization/60902
480 * tree-ssa-threadedge.c
481 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
482 invalidate outputs from statements that do not produce useful
483 outputs for threading.
485 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
487 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
488 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
489 machine descriptions for Stack Smashing Protector.
491 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
493 * aarch64.md (<optab>_rol<mode>3): New pattern.
494 (<optab>_rolsi3_uxtw): Likewise.
495 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
497 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
499 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
500 (arm_cortex_a12_tune): Likewise.
502 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
504 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
506 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
508 * config/arm/arm.md (arm_rev16si2): New pattern.
509 (arm_rev16si2_alt): Likewise.
510 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
512 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
514 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
515 (rev16<mode>2_alt): Likewise.
516 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
517 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
518 (aarch_rev16_shleft_mask_imm_p): Likewise.
519 (aarch_rev16_p_1): Likewise.
520 (aarch_rev16_p): Likewise.
521 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
522 (aarch_rev16_shright_mask_imm_p): Likewise.
523 (aarch_rev16_shleft_mask_imm_p): Likewise.
525 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
527 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
528 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
530 (cortex_a53_extra_costs): Likewise.
531 (cortex_a57_extra_costs): Likewise.
532 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
533 (cortexa7_extra_costs): Likewise.
534 (cortexa8_extra_costs): Likewise.
535 (cortexa12_extra_costs): Likewise.
536 (cortexa15_extra_costs): Likewise.
537 (v7m_extra_costs): Likewise.
538 (arm_new_rtx_costs): Handle BSWAP.
540 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
542 * config/arm/arm.c (cortexa8_extra_costs): New table.
543 (arm_cortex_a8_tune): New tuning struct.
544 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
546 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
548 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
550 2014-04-23 Richard Biener <rguenther@suse.de>
552 * Makefile.in (OBJS): Remove loop-unswitch.o.
553 * loop-unswitch.c: Delete.
554 * tree-pass.h (make_pass_rtl_unswitch): Remove.
555 * passes.def (pass_rtl_unswitch): Likewise.
556 * loop-init.c (gate_rtl_unswitch): Likewise.
557 (rtl_unswitch): Likewise.
558 (pass_data_rtl_unswitch): Likewise.
559 (pass_rtl_unswitch): Likewise.
560 (make_pass_rtl_unswitch): Likewise.
561 * rtl.h (reversed_condition): Likewise.
562 (compare_and_jump_seq): Likewise.
563 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
565 * loop-unroll.c (compare_and_jump_seq): Likewise.
567 2014-04-23 Richard Biener <rguenther@suse.de>
569 PR tree-optimization/60903
570 * tree-ssa-loop-im.c (analyze_memory_references): Remove
571 commented code block.
572 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
573 loop flags to newly created BBs and edges.
575 2014-04-23 Nick Clifton <nickc@redhat.com>
577 * config/msp430/msp430.c (msp430_handle_option): Move function
579 (msp430_option_override): Simplify mcu and mcpu option handling.
580 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
581 support for -mhwmult command line option.
582 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
583 -mhwmult command line option.
584 (msp430_hwmult_enabled): Delete.
585 (msp43o_output_labelref): Add support for -mhwmult command line
587 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
588 (umulsidi3): Likewise.
589 * config/msp430/msp430.opt (mmcu): Add Report attribute.
590 (mcpu, mlarge, msmall): Likewise.
591 (mhwmult): New option.
592 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
594 (msp430_is_f5_mcu): Remove prototype.
595 (msp430_use_f5_series_hwmult): Add prototype.
596 * config/msp430/msp430-opts.h: New file.
597 * common/config/msp430: New directory.
598 * common/config/msp430/msp430-common.c: New file.
599 * config.gcc (msp430): Remove target_has_targetm_common.
600 * doc/invoke.texi: Document -mhwmult command line option.
602 2014-04-23 Nick Clifton <nickc@redhat.com>
604 * config/i386/cygwin.h (ENDFILE_SPEC): Include
605 default-manifest.o if it can be found in the search path.
606 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
608 2014-04-23 Terry Guo <terry.guo@arm.com>
610 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
612 2014-04-23 Richard Biener <rguenther@suse.de>
615 * tree-inline.c (declare_return_variable): Use mark_addressable.
617 2014-04-23 Richard Biener <rguenther@suse.de>
620 * loop-init.c (loop_optimizer_init): Make sure to apply
621 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
623 2014-04-22 Jakub Jelinek <jakub@redhat.com>
626 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
628 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
629 if flag_sanitize_undefined_trap_on_error.
630 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
631 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
632 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
633 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
634 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
635 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
636 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
637 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
638 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
639 * ubsan.c (ubsan_instrument_unreachable): Return
640 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
641 (ubsan_expand_null_ifn): Emit __builtin_trap ()
642 if flag_sanitize_undefined_trap_on_error and
643 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
644 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
645 instrument_bool_enum_load): Emit __builtin_trap () if
646 flag_sanitize_undefined_trap_on_error and
647 __builtin_handle_*_abort () if !flag_sanitize_recover.
648 * doc/invoke.texi (-fsanitize-recover,
649 -fsanitize-undefined-trap-on-error): Document.
651 2014-04-22 Christian Bruel <christian.bruel@st.com>
653 * config/sh/sh.md (mov<mode>): Replace movQIHI.
654 Force immediates to SImode.
656 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
658 * config/nios2/nios2.md (UNSPEC_ROUND): New.
660 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
661 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
662 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
663 (nios2_fpu_insn): Add entry for round.
664 (N2FPU_NO_ERRNO_P): Define.
665 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
667 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
669 2014-04-22 Richard Henderson <rth@redhat.com>
671 * config/aarch64/aarch64 (addti3, subti3): New expanders.
672 (add<GPI>3_compare0): Remove leading * from name.
673 (add<GPI>3_carryin): Likewise.
674 (sub<GPI>3_compare0): Likewise.
675 (sub<GPI>3_carryin): Likewise.
676 (<su_optab>mulditi3): New expander.
677 (multi3): New expander.
678 (madd<GPI>): Remove leading * from name.
680 2014-04-22 Martin Jambor <mjambor@suse.cz>
682 * cgraphclones.c (cgraph_function_versioning): Copy
683 ipa_transforms_to_apply instead of asserting it is empty.
685 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
688 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
689 on count_exp to get mode.
691 2014-04-22 Andrew Pinski <apinski@cavium.com>
693 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
694 Handle TLS for ILP32.
695 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
696 (tlsie_small_<mode>): this and handle PTR.
697 (tlsie_small_sidi): New pattern.
698 (tlsle_small): Change to an expand to handle ILP32.
699 (tlsle_small_<mode>): New pattern.
700 (tlsdesc_small): Rename to ...
701 (tlsdesc_small_<mode>): this and handle PTR.
703 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
705 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
707 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
709 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
710 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
711 (aarch64_types_signed_poly_qualifiers): Likewise.
712 (aarch64_types_unsigned_signed_qualifiers): Likewise.
713 (aarch64_types_poly_signed_qualifiers): Likewise.
714 (TYPES_REINTERP_SS): Type macro added.
715 (TYPES_REINTERP_SU): Likewise.
716 (TYPES_REINTERP_SP): Likewise.
717 (TYPES_REINTERP_US): Likewise.
718 (TYPES_REINTERP_PS): Likewise.
719 (aarch64_fold_builtin): New expression folding added.
720 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
721 Declarations removed.
722 (REINTERP_SS): Declarations added.
723 (REINTERP_US): Likewise.
724 (REINTERP_PS): Likewise.
725 (REINTERP_SU): Likewise.
726 (REINTERP_SP): Likewise.
727 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
728 (vreinterpretq_p8_f64): Likewise.
729 (vreinterpret_p16_f64): Likewise.
730 (vreinterpretq_p16_f64): Likewise.
731 (vreinterpret_f32_f64): Likewise.
732 (vreinterpretq_f32_f64): Likewise.
733 (vreinterpret_f64_f32): Likewise.
734 (vreinterpret_f64_p8): Likewise.
735 (vreinterpret_f64_p16): Likewise.
736 (vreinterpret_f64_s8): Likewise.
737 (vreinterpret_f64_s16): Likewise.
738 (vreinterpret_f64_s32): Likewise.
739 (vreinterpret_f64_s64): Likewise.
740 (vreinterpret_f64_u8): Likewise.
741 (vreinterpret_f64_u16): Likewise.
742 (vreinterpret_f64_u32): Likewise.
743 (vreinterpret_f64_u64): Likewise.
744 (vreinterpretq_f64_f32): Likewise.
745 (vreinterpretq_f64_p8): Likewise.
746 (vreinterpretq_f64_p16): Likewise.
747 (vreinterpretq_f64_s8): Likewise.
748 (vreinterpretq_f64_s16): Likewise.
749 (vreinterpretq_f64_s32): Likewise.
750 (vreinterpretq_f64_s64): Likewise.
751 (vreinterpretq_f64_u8): Likewise.
752 (vreinterpretq_f64_u16): Likewise.
753 (vreinterpretq_f64_u32): Likewise.
754 (vreinterpretq_f64_u64): Likewise.
755 (vreinterpret_s64_f64): Likewise.
756 (vreinterpretq_s64_f64): Likewise.
757 (vreinterpret_u64_f64): Likewise.
758 (vreinterpretq_u64_f64): Likewise.
759 (vreinterpret_s8_f64): Likewise.
760 (vreinterpretq_s8_f64): Likewise.
761 (vreinterpret_s16_f64): Likewise.
762 (vreinterpretq_s16_f64): Likewise.
763 (vreinterpret_s32_f64): Likewise.
764 (vreinterpretq_s32_f64): Likewise.
765 (vreinterpret_u8_f64): Likewise.
766 (vreinterpretq_u8_f64): Likewise.
767 (vreinterpret_u16_f64): Likewise.
768 (vreinterpretq_u16_f64): Likewise.
769 (vreinterpret_u32_f64): Likewise.
770 (vreinterpretq_u32_f64): Likewise.
772 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
774 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
775 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
776 (vreinterpret_p8_s8): Likewise.
777 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
778 (vreinterpret_p8_s16): Likewise.
779 (vreinterpret_p8_s32): Likewise.
780 (vreinterpret_p8_s64): Likewise.
781 (vreinterpret_p8_f32): Likewise.
782 (vreinterpret_p8_u8): Likewise.
783 (vreinterpret_p8_u16): Likewise.
784 (vreinterpret_p8_u32): Likewise.
785 (vreinterpret_p8_u64): Likewise.
786 (vreinterpret_p8_p16): Likewise.
787 (vreinterpretq_p8_s8): Likewise.
788 (vreinterpretq_p8_s16): Likewise.
789 (vreinterpretq_p8_s32): Likewise.
790 (vreinterpretq_p8_s64): Likewise.
791 (vreinterpretq_p8_f32): Likewise.
792 (vreinterpretq_p8_u8): Likewise.
793 (vreinterpretq_p8_u16): Likewise.
794 (vreinterpretq_p8_u32): Likewise.
795 (vreinterpretq_p8_u64): Likewise.
796 (vreinterpretq_p8_p16): Likewise.
797 (vreinterpret_p16_s8): Likewise.
798 (vreinterpret_p16_s16): Likewise.
799 (vreinterpret_p16_s32): Likewise.
800 (vreinterpret_p16_s64): Likewise.
801 (vreinterpret_p16_f32): Likewise.
802 (vreinterpret_p16_u8): Likewise.
803 (vreinterpret_p16_u16): Likewise.
804 (vreinterpret_p16_u32): Likewise.
805 (vreinterpret_p16_u64): Likewise.
806 (vreinterpret_p16_p8): Likewise.
807 (vreinterpretq_p16_s8): Likewise.
808 (vreinterpretq_p16_s16): Likewise.
809 (vreinterpretq_p16_s32): Likewise.
810 (vreinterpretq_p16_s64): Likewise.
811 (vreinterpretq_p16_f32): Likewise.
812 (vreinterpretq_p16_u8): Likewise.
813 (vreinterpretq_p16_u16): Likewise.
814 (vreinterpretq_p16_u32): Likewise.
815 (vreinterpretq_p16_u64): Likewise.
816 (vreinterpretq_p16_p8): Likewise.
817 (vreinterpret_f32_s8): Likewise.
818 (vreinterpret_f32_s16): Likewise.
819 (vreinterpret_f32_s32): Likewise.
820 (vreinterpret_f32_s64): Likewise.
821 (vreinterpret_f32_u8): Likewise.
822 (vreinterpret_f32_u16): Likewise.
823 (vreinterpret_f32_u32): Likewise.
824 (vreinterpret_f32_u64): Likewise.
825 (vreinterpret_f32_p8): Likewise.
826 (vreinterpret_f32_p16): Likewise.
827 (vreinterpretq_f32_s8): Likewise.
828 (vreinterpretq_f32_s16): Likewise.
829 (vreinterpretq_f32_s32): Likewise.
830 (vreinterpretq_f32_s64): Likewise.
831 (vreinterpretq_f32_u8): Likewise.
832 (vreinterpretq_f32_u16): Likewise.
833 (vreinterpretq_f32_u32): Likewise.
834 (vreinterpretq_f32_u64): Likewise.
835 (vreinterpretq_f32_p8): Likewise.
836 (vreinterpretq_f32_p16): Likewise.
837 (vreinterpret_s64_s8): Likewise.
838 (vreinterpret_s64_s16): Likewise.
839 (vreinterpret_s64_s32): Likewise.
840 (vreinterpret_s64_f32): Likewise.
841 (vreinterpret_s64_u8): Likewise.
842 (vreinterpret_s64_u16): Likewise.
843 (vreinterpret_s64_u32): Likewise.
844 (vreinterpret_s64_u64): Likewise.
845 (vreinterpret_s64_p8): Likewise.
846 (vreinterpret_s64_p16): Likewise.
847 (vreinterpretq_s64_s8): Likewise.
848 (vreinterpretq_s64_s16): Likewise.
849 (vreinterpretq_s64_s32): Likewise.
850 (vreinterpretq_s64_f32): Likewise.
851 (vreinterpretq_s64_u8): Likewise.
852 (vreinterpretq_s64_u16): Likewise.
853 (vreinterpretq_s64_u32): Likewise.
854 (vreinterpretq_s64_u64): Likewise.
855 (vreinterpretq_s64_p8): Likewise.
856 (vreinterpretq_s64_p16): Likewise.
857 (vreinterpret_u64_s8): Likewise.
858 (vreinterpret_u64_s16): Likewise.
859 (vreinterpret_u64_s32): Likewise.
860 (vreinterpret_u64_s64): Likewise.
861 (vreinterpret_u64_f32): Likewise.
862 (vreinterpret_u64_u8): Likewise.
863 (vreinterpret_u64_u16): Likewise.
864 (vreinterpret_u64_u32): Likewise.
865 (vreinterpret_u64_p8): Likewise.
866 (vreinterpret_u64_p16): Likewise.
867 (vreinterpretq_u64_s8): Likewise.
868 (vreinterpretq_u64_s16): Likewise.
869 (vreinterpretq_u64_s32): Likewise.
870 (vreinterpretq_u64_s64): Likewise.
871 (vreinterpretq_u64_f32): Likewise.
872 (vreinterpretq_u64_u8): Likewise.
873 (vreinterpretq_u64_u16): Likewise.
874 (vreinterpretq_u64_u32): Likewise.
875 (vreinterpretq_u64_p8): Likewise.
876 (vreinterpretq_u64_p16): Likewise.
877 (vreinterpret_s8_s16): Likewise.
878 (vreinterpret_s8_s32): Likewise.
879 (vreinterpret_s8_s64): Likewise.
880 (vreinterpret_s8_f32): Likewise.
881 (vreinterpret_s8_u8): Likewise.
882 (vreinterpret_s8_u16): Likewise.
883 (vreinterpret_s8_u32): Likewise.
884 (vreinterpret_s8_u64): Likewise.
885 (vreinterpret_s8_p8): Likewise.
886 (vreinterpret_s8_p16): Likewise.
887 (vreinterpretq_s8_s16): Likewise.
888 (vreinterpretq_s8_s32): Likewise.
889 (vreinterpretq_s8_s64): Likewise.
890 (vreinterpretq_s8_f32): Likewise.
891 (vreinterpretq_s8_u8): Likewise.
892 (vreinterpretq_s8_u16): Likewise.
893 (vreinterpretq_s8_u32): Likewise.
894 (vreinterpretq_s8_u64): Likewise.
895 (vreinterpretq_s8_p8): Likewise.
896 (vreinterpretq_s8_p16): Likewise.
897 (vreinterpret_s16_s8): Likewise.
898 (vreinterpret_s16_s32): Likewise.
899 (vreinterpret_s16_s64): Likewise.
900 (vreinterpret_s16_f32): Likewise.
901 (vreinterpret_s16_u8): Likewise.
902 (vreinterpret_s16_u16): Likewise.
903 (vreinterpret_s16_u32): Likewise.
904 (vreinterpret_s16_u64): Likewise.
905 (vreinterpret_s16_p8): Likewise.
906 (vreinterpret_s16_p16): Likewise.
907 (vreinterpretq_s16_s8): Likewise.
908 (vreinterpretq_s16_s32): Likewise.
909 (vreinterpretq_s16_s64): Likewise.
910 (vreinterpretq_s16_f32): Likewise.
911 (vreinterpretq_s16_u8): Likewise.
912 (vreinterpretq_s16_u16): Likewise.
913 (vreinterpretq_s16_u32): Likewise.
914 (vreinterpretq_s16_u64): Likewise.
915 (vreinterpretq_s16_p8): Likewise.
916 (vreinterpretq_s16_p16): Likewise.
917 (vreinterpret_s32_s8): Likewise.
918 (vreinterpret_s32_s16): Likewise.
919 (vreinterpret_s32_s64): Likewise.
920 (vreinterpret_s32_f32): Likewise.
921 (vreinterpret_s32_u8): Likewise.
922 (vreinterpret_s32_u16): Likewise.
923 (vreinterpret_s32_u32): Likewise.
924 (vreinterpret_s32_u64): Likewise.
925 (vreinterpret_s32_p8): Likewise.
926 (vreinterpret_s32_p16): Likewise.
927 (vreinterpretq_s32_s8): Likewise.
928 (vreinterpretq_s32_s16): Likewise.
929 (vreinterpretq_s32_s64): Likewise.
930 (vreinterpretq_s32_f32): Likewise.
931 (vreinterpretq_s32_u8): Likewise.
932 (vreinterpretq_s32_u16): Likewise.
933 (vreinterpretq_s32_u32): Likewise.
934 (vreinterpretq_s32_u64): Likewise.
935 (vreinterpretq_s32_p8): Likewise.
936 (vreinterpretq_s32_p16): Likewise.
937 (vreinterpret_u8_s8): Likewise.
938 (vreinterpret_u8_s16): Likewise.
939 (vreinterpret_u8_s32): Likewise.
940 (vreinterpret_u8_s64): Likewise.
941 (vreinterpret_u8_f32): Likewise.
942 (vreinterpret_u8_u16): Likewise.
943 (vreinterpret_u8_u32): Likewise.
944 (vreinterpret_u8_u64): Likewise.
945 (vreinterpret_u8_p8): Likewise.
946 (vreinterpret_u8_p16): Likewise.
947 (vreinterpretq_u8_s8): Likewise.
948 (vreinterpretq_u8_s16): Likewise.
949 (vreinterpretq_u8_s32): Likewise.
950 (vreinterpretq_u8_s64): Likewise.
951 (vreinterpretq_u8_f32): Likewise.
952 (vreinterpretq_u8_u16): Likewise.
953 (vreinterpretq_u8_u32): Likewise.
954 (vreinterpretq_u8_u64): Likewise.
955 (vreinterpretq_u8_p8): Likewise.
956 (vreinterpretq_u8_p16): Likewise.
957 (vreinterpret_u16_s8): Likewise.
958 (vreinterpret_u16_s16): Likewise.
959 (vreinterpret_u16_s32): Likewise.
960 (vreinterpret_u16_s64): Likewise.
961 (vreinterpret_u16_f32): Likewise.
962 (vreinterpret_u16_u8): Likewise.
963 (vreinterpret_u16_u32): Likewise.
964 (vreinterpret_u16_u64): Likewise.
965 (vreinterpret_u16_p8): Likewise.
966 (vreinterpret_u16_p16): Likewise.
967 (vreinterpretq_u16_s8): Likewise.
968 (vreinterpretq_u16_s16): Likewise.
969 (vreinterpretq_u16_s32): Likewise.
970 (vreinterpretq_u16_s64): Likewise.
971 (vreinterpretq_u16_f32): Likewise.
972 (vreinterpretq_u16_u8): Likewise.
973 (vreinterpretq_u16_u32): Likewise.
974 (vreinterpretq_u16_u64): Likewise.
975 (vreinterpretq_u16_p8): Likewise.
976 (vreinterpretq_u16_p16): Likewise.
977 (vreinterpret_u32_s8): Likewise.
978 (vreinterpret_u32_s16): Likewise.
979 (vreinterpret_u32_s32): Likewise.
980 (vreinterpret_u32_s64): Likewise.
981 (vreinterpret_u32_f32): Likewise.
982 (vreinterpret_u32_u8): Likewise.
983 (vreinterpret_u32_u16): Likewise.
984 (vreinterpret_u32_u64): Likewise.
985 (vreinterpret_u32_p8): Likewise.
986 (vreinterpret_u32_p16): Likewise.
987 (vreinterpretq_u32_s8): Likewise.
988 (vreinterpretq_u32_s16): Likewise.
989 (vreinterpretq_u32_s32): Likewise.
990 (vreinterpretq_u32_s64): Likewise.
991 (vreinterpretq_u32_f32): Likewise.
992 (vreinterpretq_u32_u8): Likewise.
993 (vreinterpretq_u32_u16): Likewise.
994 (vreinterpretq_u32_u64): Likewise.
995 (vreinterpretq_u32_p8): Likewise.
996 (vreinterpretq_u32_p16): Likewise.
998 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1000 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1002 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1005 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1006 (vqnegd_s64): Likewise.
1007 (vqabs_s64): Likewise.
1008 (vqabsd_s64): Likewise.
1010 2014-04-22 Richard Henderson <rth@redhat.com>
1012 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1013 computation to the top of the loop.
1015 2014-04-22 Renlin <renlin.li@arm.com>
1016 Jiong Wang <jiong.wang@arm.com>
1018 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1019 * config/aarch64/aarch64.c (aarch64_layout_frame)
1020 (aarch64_initial_elimination_offset): Likewise.
1022 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1024 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1027 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1029 * machmode.h (bitwise_mode_for_mode): Declare.
1030 * stor-layout.h (bitwise_type_for_mode): Likewise.
1031 * stor-layout.c (bitwise_mode_for_mode): New function.
1032 (bitwise_type_for_mode): Likewise.
1033 * builtins.c (fold_builtin_memory_op): Use it instead of
1034 int_mode_for_mode and build_nonstandard_integer_type.
1036 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1038 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1039 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1040 (*-*-solaris2*): Simplify.
1041 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1042 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1043 *-*-solaris2.9* handling.
1045 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1047 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1048 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1050 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1051 * configure: Regenerate.
1053 * config/i386/sol2-9.h: Remove.
1055 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1056 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1057 Remove Solaris 9 references.
1059 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1061 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1062 (floatuns<GPI:mode><GPF:mode>2): Remove.
1063 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1064 and floatuns conversions.
1065 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1066 and floatuns conversions.
1067 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1068 (w1,w2): New mode attributes for inequal width conversions.
1070 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1072 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1073 the output asm format.
1075 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1077 * config/aarch64/aarch64-simd.md
1078 (aarch64_cm<optab>di): Always split.
1079 (*aarch64_cm<optab>di): New.
1080 (aarch64_cmtstdi): Always split.
1081 (*aarch64_cmtstdi): New.
1083 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1085 PR tree-optimization/60823
1086 * omp-low.c (ipa_simd_modify_function_body): Go through
1087 all SSA_NAMEs and for those refering to vector arguments
1088 which are going to be replaced adjust SSA_NAME_VAR and,
1089 if it is a default definition, change it into a non-default
1090 definition assigned at the beginning of function from new_decl.
1091 (ipa_simd_modify_stmt_ops): Rewritten.
1092 * tree-dfa.c (set_ssa_default_def): When removing default def,
1093 check for NULL loc instead of NULL *loc.
1095 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1097 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1098 restrictions on core registers for DImode values in Thumb2.
1100 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1102 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1103 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1105 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1107 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1108 (*iordi_notzesidi_di): Likewise.
1109 (*iordi_notsesidi_di): Likewise.
1111 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1113 * config/arm/arm-protos.h (tune_params): New struct members.
1114 * config/arm/arm.c: Initialise tune_params per processor.
1115 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1116 for speed, based on new tune_params.
1118 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1120 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1122 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1124 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1126 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1127 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1128 (vrnda_f64): Likewise.
1129 (vrndi_f64): Likewise.
1130 (vrndm_f64): Likewise.
1131 (vrndn_f64): Likewise.
1132 (vrndp_f64): Likewise.
1133 (vrndx_f64): Likewise.
1135 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1137 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1138 GET_MODE_SIZE argument is enum machine_mode.
1140 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1143 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1144 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1146 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1149 * asan.c (asan_emit_stack_protection): Force the base to align to
1150 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1151 appropriate bits if STRICT_ALIGNMENT.
1152 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1154 (expand_used_vars): Leave a space in the stack frame for alignment
1155 if STRICT_ALIGNMENT.
1157 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1159 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1161 (gimple_store_p): Likewise.
1162 (gimple_assign_load_p): Likewise.
1163 (gimple_assign_cast_p): Likewise.
1164 (gimple_clobber_p): Likewise.
1166 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1167 rather than a gimple.
1168 (gimple_assign_cast_p): Likewise.
1170 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1173 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1174 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1176 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1177 more debug information for E500 if -mdebug=reg.
1179 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1182 * config/i386/i386.c (ix86_expand_builtin)
1183 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1184 register for target RTX.
1185 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1187 2014-04-18 Cong Hou <congh@google.com>
1189 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1190 the widen-mult pattern by handling two operands with different sizes,
1191 and operands whose size is smaller than half of the result type.
1193 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1195 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1196 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1197 (do_estimate_edge_time): Compute it.
1198 * ipa-inline.c (want_inline_small_function_p): Bypass
1199 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1201 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1203 * ipa-inline.c (spec_rem): New static variable.
1204 (dump_overall_stats): New function.
1205 (dump_inline_stats): New function.
1207 2014-04-18 Richard Henderson <rth@redhat.com>
1209 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1210 to GET_MODE_SIZE, not a reg_class_t.
1212 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1214 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1215 (vsx_xxmrglw_<mode>): Likewise.
1217 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1220 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1221 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1222 (rs6000_init_hard_regno_mode_ok): Likewise.
1224 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1226 * ipa-inline.c (inline_small_functions): Account only non-cold
1228 * doc/invoke.texi (inline-unit-growth): Update documentation.
1230 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1232 * config/rs6000/rs6000.md (addti3, subti3): New.
1234 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1237 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1238 comment. Check optimize_insn_for_size_p instead of
1239 optimize_insn_for_speed_p.
1241 2014-04-17 Martin Jambor <mjambor@suse.cz>
1243 * gimple-iterator.c (gsi_start_edge): New function.
1244 * gimple-iterator.h (gsi_start_edge): Declare.
1245 * tree-sra.c (single_non_eh_succ): New function.
1246 (disqualify_ops_if_throwing_stmt): Renamed to
1247 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1248 having one non-EH successor BB.
1249 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1250 generate loads into replacements.
1251 (sra_modify_assign): Likewise and and also use the simple path for
1253 (sra_modify_function_body): Commit statements on edges.
1255 2014-04-17 Richard Biener <rguenther@suse.de>
1258 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1259 comparison results and add clarifying comment.
1261 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1263 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1264 (blank_mode): Initialize it.
1265 (emit_mode_size_inline, emit_mode_nunits_inline,
1266 emit_mode_inner_inline): New functions.
1267 (emit_insn_modes_h): Call them and surround their output with
1268 #if GCC_VERSION >= 4001 ... #endif.
1269 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1270 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1271 mode_* arrays if the argument is __builtin_constant_p.
1272 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1273 is enum machine_mode.
1275 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1277 * passes.c (opt_pass::execute): Adjust.
1278 (pass_manager::execute_pass_mode_switching): Likewise.
1279 (early_local_passes::execute): Likewise.
1280 (execute_one_pass): Pass cfun to the pass's execute method.
1281 * tree-pass.h (opt_pass::execute): Add function * argument.
1282 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1283 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1284 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1285 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1286 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1287 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1288 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1289 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1290 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1291 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1292 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1293 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1294 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1295 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1296 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1297 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1298 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1299 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1300 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1301 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1302 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1303 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1304 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1305 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1306 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1307 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1308 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1309 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1310 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1313 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1315 * passes.c (opt_pass::gate): Take function * argument.
1316 (gate_all_early_local_passes): Merge into
1317 (early_local_passes::gate): this.
1318 (gate_all_early_optimizations): Merge into
1319 (all_early_optimizations::gate): this.
1320 (gate_all_optimizations): Mege into
1321 (all_optimizations::gate): this.
1322 (gate_all_optimizations_g): Merge into
1323 (all_optimizations_g::gate): this.
1324 (gate_rest_of_compilation): Mege into
1325 (rest_of_compilation::gate): this.
1326 (gate_postreload): Merge into
1327 (postreload::gate): this.
1328 (dump_one_pass): Pass cfun to the pass's gate method.
1329 (execute_ipa_summary_passes): Likewise.
1330 (execute_one_pass): Likewise.
1331 (ipa_write_summaries_2): Likewise.
1332 (ipa_write_optimization_summaries_1): Likewise.
1333 (ipa_read_summaries_1): Likewise.
1334 (ipa_read_optimization_summaries_1): Likewise.
1335 (execute_ipa_stmt_fixups): Likewise.
1336 * tree-pass.h (opt_pass::gate): Add function * argument.
1337 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1338 combine-stack-adj.c, combine.c, compare-elim.c,
1339 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1340 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1341 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1342 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1343 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1344 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1345 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1346 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1347 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1348 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1349 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1350 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1351 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1352 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1353 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1354 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1355 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1356 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1357 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1358 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1359 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1360 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1361 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1362 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1363 var-tracking.c, vtable-verify.c, web.c: Adjust.
1365 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1367 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1368 * configure: Regenerate.
1370 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1372 * passes.c (dump_one_pass): don't check pass->has_gate.
1373 (execute_ipa_summary_passes): Likewise.
1374 (execute_one_pass): Likewise.
1375 (ipa_write_summaries_2): Likewise.
1376 (ipa_write_optimization_summaries_1): Likewise.
1377 (ipa_read_optimization_summaries_1): Likewise.
1378 (execute_ipa_stmt_fixups): Likewise.
1379 * tree-pass.h (pass_data::has_gate): Remove.
1380 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1381 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1382 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1383 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1384 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1385 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1386 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1387 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1388 gimple-low.c, gimple-ssa-isolate-paths.c,
1389 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1390 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1391 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1392 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1393 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1394 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1395 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1396 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1397 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1398 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1399 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1400 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1401 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1402 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1403 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1404 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1405 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1406 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1407 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1408 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1409 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1410 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1413 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1415 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1417 * passes.c (pass_manager::register_dump_files_1): Merge into
1418 (pass_manager::register_dump_files): this, and remove its handling of
1419 properties since the pass always has the properties anyway.
1420 (pass_manager::pass_manager): Adjust.
1422 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1424 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1425 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1426 dealing with properties.
1427 (pass_manager::register_dump_files): Adjust.
1429 2014-03-20 Mark Wielaard <mjw@redhat.com>
1431 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1432 then represent the bound as normal constant value.
1434 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1437 Forward port from 4.8 branch
1438 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1440 * config/i386/bmiintrin.h (_blsi_u32): New.
1444 (_blsmsk_u32): Ditto.
1445 (_blsmsk_u64): Ditto.
1446 (_tzcnt_u32): Ditto.
1447 (_tzcnt_u64): Ditto.
1449 2014-04-17 Kito Cheng <kito@0xlab.org>
1451 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1453 2014-04-17 Richard Biener <rguenther@suse.de>
1456 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1457 boolean results for comparisons.
1459 2014-04-17 Richard Biener <rguenther@suse.de>
1461 PR tree-optimization/60836
1462 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1463 initial PHI args to be gimple values.
1465 2014-04-17 Richard Biener <rguenther@suse.de>
1467 PR tree-optimization/60841
1468 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
1469 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
1470 of stmts to SLP build.
1471 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
1472 (vect_analyze_slp): Likewise.
1473 (vect_analyze_slp_instance): Likewise.
1474 (vect_build_slp_tree): Limit overall SLP tree growth.
1475 * tree-vectorizer.h (vect_analyze_data_refs,
1476 vect_analyze_slp): Adjust prototypes.
1478 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1480 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
1483 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1485 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
1486 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
1487 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
1488 for TARGET_SLOW_PSHUFB
1490 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1492 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
1493 * config/i386/i386.c (intel_cost): Ditto.
1495 2014-04-17 Joey Ye <joey.ye@arm.com>
1497 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
1499 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1501 * opts.c (common_handle_option): Disable -fipa-reference coorectly
1504 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1506 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
1507 (type_all_derivations_known_p): New predicate.
1508 (type_all_ctors_visible_p): New predicate.
1509 (type_possibly_instantiated_p): New predicate.
1510 (get_odr_type): Compute all_derivations_known.
1511 (dump_odr_type): Dump the flag.
1512 (maybe_record_type): Cleanup.
1513 (record_target_from_binfo): Add bases_to_consider array;
1514 record bases for types w/o instances and skip CXX destructor.
1515 (possible_polymorphic_call_targets_1): Add bases_to_consider
1516 and consider_construction parameters; check if type may have instance.
1517 (get_polymorphic_call_info): Set maybe_in_construction to true
1518 when we know nothing.
1519 (record_targets_from_bases): Skip CXX destructors; they are
1520 never called for types in construction.
1521 (possible_polymorphic_call_targets): Do not record target when
1522 type may not have instance.
1524 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1527 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
1528 external aliases alive, too.
1530 2014-04-16 Andrew Pinski <apinski@cavium.com>
1532 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
1535 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1537 * final.c (compute_alignments): Do not apply loop alignment to a block
1538 falling through to the exit.
1540 2014-04-16 Catherine Moore <clm@codesourcery.com>
1542 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
1543 Adjust constraints for microMIPS store patterns.
1545 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
1547 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
1549 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1551 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
1552 (append_use): Run at -O0.
1553 (append_vdef): Likewise.
1554 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
1555 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
1557 2014-04-16 Jakub Jelinek <jakub@redhat.com>
1559 PR tree-optimization/60844
1560 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
1561 (propagate_op_to_single_use, remove_visited_stmt_chain,
1562 linearize_expr, repropagate_negates, reassociate_bb): Use it
1563 instead of gsi_remove.
1565 2014-04-16 Martin Jambor <mjambor@suse.cz>
1567 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
1568 ipa_transforms_to_apply.
1569 (cgraph_function_versioning): Assert that old_node has empty
1570 ipa_transforms_to_apply.
1571 * trans-mem.c (ipa_tm_create_version): Likewise.
1572 * tree-inline.c (tree_function_versioning): Do not duplicate
1573 ipa_transforms_to_apply.
1575 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1578 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
1580 Pass necessary as flags on 64-bit Solaris/x86.
1581 Use lowercase relocs for x86_64-*-*.
1582 * configure: Regenerate.
1584 2014-04-15 Jan Hubicka <jh@suse.cz>
1586 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
1587 (maybe_record_node, likely_target_p): Use it.
1589 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1592 Revert following patch
1594 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
1597 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
1598 software floating point or no floating point registers, do not
1599 allow any type in the FPRs. Eliminate a test for SPE SIMD types
1600 in GPRs that occurs after we tested for GPRs that would never be
1603 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
1604 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
1605 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
1606 specifically allow DDmode, since that does not use the SPE SIMD
1609 2014-03-21 Mark Wielaard <mjw@redhat.com>
1611 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
1612 as unsigned or int depending on type and value used.
1614 2014-04-15 Richard Biener <rguenther@suse.de>
1616 PR rtl-optimization/56965
1617 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
1618 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
1620 * alias.c (true_dependence_1): Do not call
1621 nonoverlapping_component_refs_p.
1622 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
1623 nonoverlapping_component_refs_p.
1624 (indirect_refs_may_alias_p): Likewise.
1626 2014-04-15 Teresa Johnson <tejohnson@google.com>
1628 * cfg.c (dump_bb_info): Fix flags check.
1629 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
1631 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1633 PR rtl-optimization/60663
1634 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
1637 2014-04-15 Richard Biener <rguenther@suse.de>
1639 * lto-streamer.h (LTO_major_version): Bump to 4.
1641 2014-04-15 Richard Biener <rguenther@suse.de>
1643 * common.opt (lto_partition_model): New enum.
1644 (flto-partition=): Merge separate options with a single with argument,
1645 add -flto-partition=one support.
1646 * flag-types.h (enum lto_partition_model): Declare.
1647 * opts.c (finish_options): Remove duplicate -flto-partition=
1649 * lto-wrapper.c (run_gcc): Adjust.
1651 2014-04-15 Richard Biener <rguenther@suse.de>
1653 * alias.c (ncr_compar): New function.
1654 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
1656 2014-04-15 Richard Biener <rguenther@suse.de>
1658 * alias.c (record_component_aliases): Do not walk BINFOs.
1660 2014-04-15 Richard Biener <rguenther@suse.de>
1662 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
1663 Add struct function argument and adjust.
1664 (find_func_aliases_for_call): Likewise.
1665 (find_func_aliases): Likewise.
1666 (find_func_clobbers): Likewise.
1667 (intra_create_variable_infos): Likewise.
1668 (compute_points_to_sets): Likewise.
1669 (ipa_pta_execute): Adjust. Do not push/pop cfun.
1671 2014-04-15 Richard Biener <rguenther@suse.de>
1673 * tree.c (iterative_hash_expr): Use enum tree_code_class
1674 to store TREE_CODE_CLASS.
1675 (tree_block): Likewise.
1676 (tree_set_block): Likewise.
1677 * tree.h (fold_build_pointer_plus_loc): Use
1678 convert_to_ptrofftype_loc.
1680 2014-04-15 Jakub Jelinek <jakub@redhat.com>
1683 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
1686 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
1688 * cfgloop.h (struct loop): Move force_vectorize down.
1689 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
1690 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
1691 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
1692 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
1693 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
1694 * tree-core.h (enum annot_expr_kind): Add new kind values.
1695 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
1696 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
1698 * tree.def (ANNOTATE_EXPR): Tweak comment.
1700 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
1702 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
1705 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
1707 * tree.h (TYPE_IDENTIFIER): Declare.
1708 * tree.c (subrange_type_for_debug_p): Use it.
1709 * godump.c (go_format_type): Likewise.
1710 * dwarf2out.c (is_cxx_auto, modified_type_die,
1711 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
1712 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
1714 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
1717 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
1719 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
1721 * config/i386/i386.c (examine_argument): Return bool. Return true if
1722 parameter should be passed in memory.
1723 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
1724 (construct_container): Update calls to examine_argument.
1725 (function_arg_advance_64): Ditto.
1726 (return_in_memory_32): Merge with ix86_return_in_memory.
1727 (return_in_memory_64): Ditto.
1728 (return_in_memory_ms_64): Ditto.
1730 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
1732 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
1733 * coverage.c (coverage_compute_profile_id): Handle externally visible
1736 2014-04-14 Martin Jambor <mjambor@suse.cz>
1738 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
1739 DECL_DISREGARD_INLINE_LIMITS functions.
1741 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
1744 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
1746 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
1749 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
1750 optimize_insn_for_speed_p instead of
1751 optimize_function_for_speed_p.
1753 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
1755 * doc/invoke.texi (free): Document AArch64.
1757 2014-04-14 Richard Biener <rguenther@suse.de>
1759 PR tree-optimization/60042
1760 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
1761 (insert_into_preds_of_block): Do not prevent PHI insertion
1762 for REFERENCE exprs here ...
1763 (eliminate_dom_walker::before_dom_children): ... but prevent
1764 their use here under similar conditions when applied to the
1765 IL after PRE optimizations.
1767 2014-04-14 Richard Biener <rguenther@suse.de>
1769 * passes.def: Move early points-to after early SRA.
1771 2014-04-14 Richard Biener <rguenther@suse.de>
1773 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
1774 check for which sign-changes we allow when forwarding
1775 a converted value into a switch.
1777 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
1779 * stor-layout.c (place_field): Finalize non-constant offset for the
1782 2014-04-14 Richard Biener <rguenther@suse.de>
1784 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
1786 (expand_switch_using_bit_tests_p): Likewise.
1787 (process_switch): Compute and pass on speed_p based on the
1789 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
1790 optimize_bb_for_speed_p.
1792 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
1794 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
1795 * function.h (struct function): Rename has_force_vect_loops into
1796 has_force_vectorize_loops.
1797 * lto-streamer-in.c (input_cfg): Adjust for renaming.
1798 (input_struct_function_base): Likewise.
1799 * lto-streamer-out.c (output_cfg): Likewise.
1800 (output_struct_function_base): Likewise.
1801 * omp-low.c (expand_omp_simd): Likewise.
1802 * tree-cfg.c (move_sese_region_to_fn): Likewise.
1803 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
1804 (version_loop_for_if_conversion): Likewise.
1805 (tree_if_conversion): Likewise.
1806 (main_tree_if_conversion): Likewise.
1807 (gate_tree_if_conversion): Likewise.
1808 * tree-inline.c (copy_loops): Likewise.
1809 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
1810 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
1811 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
1812 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
1813 * tree-vectorizer.c (vectorize_loops): Likewise.
1814 * tree-vectorizer.h (unlimited_cost_model): Likewise.
1816 2014-04-14 Richard Biener <rguenther@suse.de>
1819 * lto-streamer-out.c (wrap_refs): New function.
1820 (lto_output): Wrap symbol references in global initializes in
1821 type-preserving MEM_REFs.
1823 2014-04-14 Christian Bruel <christian.bruel@st.com>
1825 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
1827 2014-04-14 Christian Bruel <christian.bruel@st.com>
1829 * config/sh/sh.md (setmemqi): New expand pattern.
1830 * config/sh/sh.h (CLEAR_RATIO): Define.
1831 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
1832 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
1834 2014-04-14 Richard Biener <rguenther@suse.de>
1837 * fold-const.c (negate_expr_p): Don't negate directional rounding
1839 (fold_negate_expr): Likewise.
1841 2014-04-14 Richard Biener <rguenther@suse.de>
1843 PR tree-optimization/59817
1844 PR tree-optimization/60453
1845 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
1846 recursion to catch all CHRECs in the scalar evolution and restrict
1847 the predicate for the remains appropriately.
1849 2014-04-12 Catherine Moore <clm@codesourcery.com>
1851 * config/mips/constraints.md: Add new register constraint "kb".
1852 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
1853 (*movhi_internal): Likewise.
1854 (*movqi_internal): Likewise.
1855 * config/mips/mips.h (M16_STORE_REGS): New register class.
1856 (REG_CLASS_NAMES): Add M16_STORE_REGS.
1857 (REG_CLASS_CONTENTS): Likewise.
1858 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
1860 2014-04-11 Tobias Burnus <burnus@net-b.de>
1863 * doc/invoke.texi (-Wformat-signedness): Document it.
1864 (Wformat=2): Mention that this enables -Wformat-signedness.
1866 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
1868 * common/config/epiphany/epiphany-common.c
1869 (epiphany_option_optimization_table): Enable section anchors by
1870 default at -O1 or higher.
1871 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
1872 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
1873 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
1874 carries no extra cost.
1875 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
1876 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
1877 * config/epiphany/predicates.md (memclob_operand): New predicate.
1878 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
1879 Use memclob_operand predicate and X constraint for operand 3.
1881 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
1883 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
1884 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
1887 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
1889 PR rtl-optimization/60651
1890 * mode-switching.c (optimize_mode_switching): Make sure to emit
1891 sets of a lower numbered entity before sets of a higher numbered
1892 entity to a mode of the same or lower priority.
1893 When creating a seginfo for a basic block that starts with a code
1894 label, move the insertion point past the code label.
1895 (new_seginfo): Document and enforce requirement that
1896 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
1897 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
1898 * doc/tm.texi: Regenerate.
1900 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
1903 * config/arc/arc.c (arc_save_restore): Fix assert typo.
1905 2013-04-11 Jakub Jelinek <jakub@redhat.com>
1907 * BASE-VER: Set to 4.10.0.
1909 2014-04-11 Tobias Burnus <burnus@net-b.de>
1912 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
1913 * doc/gcc.texi (Service): Update description in the @menu
1914 * doc/invoke.texi (Option Summary): Remove misplaced and
1917 2014-04-11 Steve Ellcey <sellcey@mips.com>
1918 Jakub Jelinek <jakub@redhat.com>
1921 * expr.c (convert_move): Use emit_store_flag_force instead of
1922 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
1925 2014-04-11 Richard Biener <rguenther@suse.de>
1928 * varasm.c (assemble_alias): Avoid endless error reporting
1929 recursion by setting TREE_ASM_WRITTEN.
1931 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1933 * config/s390/s390.md: Add a splitter for NOT rtx.
1935 2014-04-11 Jakub Jelinek <jakub@redhat.com>
1937 PR rtl-optimization/60663
1938 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
1940 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
1941 Jakub Jelinek <jakub@redhat.com>
1944 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
1945 flag from decl_node to node.
1947 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1950 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
1951 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
1952 ameliorating the cases where it can be.
1954 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
1957 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
1959 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1960 (loadsync_<mode>): Change mode.
1961 (load_quadpti, store_quadpti): New.
1962 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1963 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1964 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
1966 2014-04-09 Cong Hou <congh@google.com>
1969 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
1972 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1974 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
1975 instead of vnor to exploit possible fusion opportunity in the
1977 (altivec_expand_vec_perm_const_le): Likewise.
1979 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
1981 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
1982 (loadsync_<mode>): Change mode.
1983 (load_quadpti, store_quadpti): New.
1984 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
1985 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
1987 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
1990 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
1991 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
1992 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
1994 2014-04-08 Richard Biener <rguenther@suse.de>
1997 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
1998 a 64bit widest int print double-int similar to on HWI64 hosts.
2000 2014-04-08 Richard Biener <rguenther@suse.de>
2002 PR tree-optimization/60785
2003 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2004 default defs properly.
2006 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2008 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2009 (Weffc++): Likewise.
2011 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2013 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2014 set completep to false rather than true.
2016 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2019 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2020 ARM_TARGET2_DWARF_FORMAT.
2022 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2025 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2026 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2029 2014-04-07 Richard Biener <rguenther@suse.de>
2031 PR tree-optimization/60766
2032 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2033 (may_eliminate_iv): Convert cand_value_at result to desired type.
2035 2014-04-07 Jason Merrill <jason@redhat.com>
2038 * common.opt (-fno-gnu-unique): Add.
2039 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2041 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2043 * haifa-sched.c: Fix outdated function reference and minor
2044 grammar errors in introductory comment.
2046 2014-04-07 Richard Biener <rguenther@suse.de>
2049 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2051 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2053 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2056 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2057 size accounting for thunks.
2058 (pa_asm_output_mi_thunk): Use final_start_function() and
2059 final_end_function() to output function start and end directives.
2061 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2063 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2064 device specific ISA/ feature information. Remove short_sp and
2065 errata_skip ds. Add avr_device_specific_features enum to have device
2067 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2068 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2069 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2070 updated device specific info.
2071 * config/avr/avr-mcus.def: Merge device specific details to
2072 dev_attribute field.
2073 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2075 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2076 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2077 assembler if RMW isa supported by current device.
2078 * config/avr/genmultilib.awk: Update as device info structure changed.
2079 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2081 2014-04-04 Cong Hou <congh@google.com>
2083 PR tree-optimization/60656
2084 * tree-vect-stmts.c (supportable_widening_operation):
2085 Fix a bug that elements in a vector with vect_used_by_reduction
2086 property are incorrectly reordered when the operation on it is not
2087 consistant with the one in reduction operation.
2089 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2091 PR rtl-optimization/60155
2092 * gcse.c (record_set_data): New function.
2093 (single_set_gcse): New function.
2094 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2095 (hoist_code): Likewise.
2096 (get_pressure_class_and_nregs): Likewise.
2098 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2100 * explow.c (probe_stack_range): Emit a final optimization blockage.
2102 2014-04-04 Anthony Green <green@moxielogic.com>
2104 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2107 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2110 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2111 flags are set only during streaming.
2112 * ipa.c (process_references, walk_polymorphic_call_targets,
2113 symtab_remove_unreachable_nodes): Drop bodies of always inline
2114 after early inlining.
2115 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2117 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2118 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2121 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2124 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2127 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2129 (cortex_a53_fdivd): Likewise.
2131 2014-04-04 Martin Jambor <mjambor@suse.cz>
2134 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2136 * cgraph.c (clone_of_p): Also return true if thunks match.
2137 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2138 cgraph_function_or_thunk_node and an obsolete comment.
2139 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2141 (build_function_decl_skip_args): Likewise.
2142 (set_new_clone_decl_and_node_flags): New function.
2143 (duplicate_thunk_for_node): Likewise.
2144 (redirect_edge_duplicating_thunks): Likewise.
2145 (cgraph_clone_node): New parameter args_to_skip, pass it to
2146 redirect_edge_duplicating_thunks which is called instead of
2147 cgraph_redirect_edge_callee.
2148 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2149 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2151 2014-04-04 Jeff Law <law@redhat.com>
2154 * config/arm/predicates.md (const_int_I_operand): New predicate.
2155 (const_int_M_operand): Similarly.
2156 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2158 (insv_t2, extv_reg, extzv_t2): Likewise.
2159 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2160 (pop_multiple_with_writeback_and_return): Likewise.
2161 (vfp_pop_multiple_with_writeback): Likewise
2163 2014-04-04 Richard Biener <rguenther@suse.de>
2166 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2167 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2169 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2170 * gimplify.c (gimple_add_tmp_var_fn): New function.
2171 * gimple-expr.h (create_tmp_reg_fn): Declare.
2172 * gimple-expr.c (create_tmp_reg_fn): New function.
2173 * gimple-low.c (record_vars_into): Don't change cfun.
2174 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2175 code generation without cfun.
2177 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2180 * Makefile.in (install-driver): Fix shell scripting.
2182 2014-04-03 Cong Hou <congh@google.com>
2184 PR tree-optimization/60505
2185 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2186 threshold of number of iterations below which no vectorization
2188 * tree-vect-loop.c (new_loop_vec_info):
2189 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2190 * tree-vect-loop.c (vect_analyze_loop_operations):
2191 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2192 * tree-vect-loop.c (vect_transform_loop):
2193 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2194 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2195 of iterations of the loop and see if we should build the epilogue.
2197 2014-04-03 Richard Biener <rguenther@suse.de>
2199 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2200 (streamer_tree_cache_create): Adjust.
2201 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2202 to allow optional nodes array.
2203 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2204 (streamer_tree_cache_append): Likewise.
2205 (streamer_tree_cache_create): Create nodes array optionally
2206 as specified by parameter.
2207 * lto-streamer-out.c (create_output_block): Avoid maintaining
2208 the node array in the writer cache.
2209 (DFS_write_tree): Remove assertion.
2210 (produce_asm_for_decls): Free the out decl state hash table early.
2211 * lto-streamer-in.c (lto_data_in_create): Adjust for
2212 streamer_tree_cache_create prototype change.
2214 2014-04-03 Richard Biener <rguenther@suse.de>
2216 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2217 set TREE_CHAIN to NULL_TREE.
2219 2014-04-03 Richard Biener <rguenther@suse.de>
2221 PR tree-optimization/60740
2222 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2223 over all GIMPLE_COND operands.
2225 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2227 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2228 (Weffc++): Remove Scott's numbering, merge lists and reference
2231 2014-04-03 Nick Clifton <nickc@redhat.com>
2233 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2236 2014-04-03 Martin Jambor <mjambor@suse.cz>
2238 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2239 mention gcc_unreachable before failing.
2240 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2243 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2246 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2247 inconsistent code and instead mark the context inconsistent.
2248 (possible_polymorphic_call_targets): For inconsistent contexts
2249 return empty complete list.
2251 2014-04-02 Anthony Green <green@moxielogic.com>
2253 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2254 (extendqisi2, extendhisi2): Define.
2255 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2256 (WCHAR_TYPE): Change to unsigned int.
2258 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2260 PR tree-optimization/60733
2261 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2262 insertion point for PHI candidates to be the end of the feeding
2263 block for the PHI argument.
2265 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2267 PR rtl-optimization/60650
2268 * lra-constraints.c (process_alt_operands): Decrease reject for
2269 earlyclobber matching.
2271 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2273 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2275 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2277 * config/spu/spu.c (pad_bb): Do not crash when the last
2278 insn is CODE_FOR_blockage.
2280 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2282 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2283 lies outside the target mode.
2285 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2288 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2289 software floating point or no floating point registers, do not
2290 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2291 in GPRs that occurs after we tested for GPRs that would never be
2294 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2295 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2296 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2297 specifically allow DDmode, since that does not use the SPE SIMD
2300 2014-04-02 Richard Biener <rguenther@suse.de>
2303 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2304 MODE_INTs. Properly use negv_optab.
2305 (expand_abs): Likewise.
2307 2014-04-02 Richard Biener <rguenther@suse.de>
2310 * Makefile.in (install-driver): Guard extra installs with special
2313 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2315 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2318 2014-04-01 Richard Henderson <rth@redhat.com>
2321 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2322 alternative enabled before register allocation.
2324 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2326 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2327 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2329 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2330 (nios2_got_address): Update nios2_large_got_address call site.
2331 (nios2_delegitimize_address): New function.
2332 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2333 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2334 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2336 2014-04-01 Martin Husemann <martin@duskware.de>
2338 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2341 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2343 PR rtl-optimization/60604
2344 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2345 check from register_operand.
2346 (register_operand): Redefine in terms of general_operand.
2347 (nonmemory_operand): Use register_operand for the non-constant cases.
2349 2014-04-01 Richard Biener <rguenther@suse.de>
2351 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2353 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2355 * doc/invoke.texi (mapp-regs): Clarify.
2357 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2359 * config/i386/avx512fintrin.h (__v32hi): Define type.
2360 (__v64qi): Likewise.
2361 (_mm512_set1_epi8): Define.
2362 (_mm512_set1_epi16): Define.
2363 (_mm512_set4_epi32): Define.
2364 (_mm512_set4_epi64): Define.
2365 (_mm512_set4_pd): Define.
2366 (_mm512_set4_ps): Define.
2367 (_mm512_setr4_epi64): Define.
2368 (_mm512_setr4_epi32): Define.
2369 (_mm512_setr4_pd): Define.
2370 (_mm512_setr4_ps): Define.
2371 (_mm512_setzero_epi32): Define.
2373 2014-03-31 Martin Jambor <mjambor@suse.cz>
2376 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2377 callsite_arguments_match_p. Updated all callers. Also check types of
2378 corresponding formal parameters and actual arguments.
2379 (not_all_callers_have_enough_arguments_p) Renamed to
2380 some_callers_have_mismatched_arguments_p.
2382 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2384 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2386 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2389 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2392 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2394 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2395 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2397 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2398 Use FMAMODE_NOVF512 mode iterator.
2399 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2400 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2401 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2402 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2404 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2405 Use VF_128_256 mode iterator.
2406 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2409 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2411 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2412 static chain if needed.
2414 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2417 * lra-constraints.c (index_part_to_reg): New.
2418 (process_address): Use it.
2420 2014-03-27 Jeff Law <law@redhat.com>
2421 Jakub Jelinek <jakub@redhat.com>
2424 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2425 gen_rtx_{PLUS,MULT} to build up the address expression.
2427 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2428 creating non-canonical RTL.
2430 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2433 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2434 functions; reorganize to make cheap checks first.
2435 (inline_small_functions): Do not estimate growth when dumping;
2437 * ipa-inline.h (inline_summary): Add min_size.
2438 (growth_likely_positive): New function.
2439 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2440 (set_cond_stmt_execution_predicate): Cleanup.
2441 (estimate_edge_size_and_time): Compute min_size.
2442 (estimate_calls_size_and_time): Likewise.
2443 (estimate_node_size_and_time): Likewise.
2444 (inline_update_overall_summary): Update min_size.
2445 (do_estimate_edge_time): Likewise.
2446 (do_estimate_edge_size): Update.
2447 (do_estimate_edge_hints): Update.
2448 (growth_likely_positive): New function.
2450 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2453 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2454 also if addr has VOIDmode.
2456 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2458 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2459 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2461 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2462 instructions as well as AdvancedSIMD loads.
2464 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2466 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2467 Use crypto_aese type.
2468 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
2469 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
2470 crypto_aese, crypto_aesmc. Move to types.md.
2471 * config/arm/types.md (crypto_aes): Split into crypto_aese,
2473 * config/arm/iterators.md (crypto_type): Likewise.
2475 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2477 * cgraph.c: Include expr.h and tree-dfa.h.
2478 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
2481 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2484 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
2485 regs from checking multi-reg pseudos.
2487 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2489 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
2491 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2493 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
2494 if it would clobber the stack pointer, even temporarily.
2496 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
2498 * mode-switching.c: Make small adjustments to the top comment.
2500 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
2502 * config/rs6000/constraints.md (wD constraint): New constraint to
2503 match the constant integer to get the top DImode/DFmode out of a
2504 vector in a VSX register.
2506 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
2507 match the constant integer to get the top DImode/DFmode out of a
2508 vector in a VSX register.
2510 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
2513 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
2516 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
2517 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
2519 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
2520 Optimize vec_extract of 64-bit values, where the value being
2521 extracted is in the top word, where we can use scalar
2522 instructions. Add direct move and store support. Combine the big
2523 endian/little endian vector select load support into a single insn.
2524 (vsx_extract_<mode>_internal1): Likewise.
2525 (vsx_extract_<mode>_internal2): Likewise.
2526 (vsx_extract_<mode>_load): Likewise.
2527 (vsx_extract_<mode>_store): Likewise.
2528 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
2529 combined into vsx_extract_<mode>_load.
2530 (vsx_extract_<mode>_one_le): Likewise.
2532 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
2533 define the top 64-bit vector element.
2535 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
2538 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2539 Document vec_vbpermq builtin.
2542 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
2543 enable use of xxsldwi and xxpermdi builtin functions.
2544 (vec_xxpermdi): Likewise.
2546 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2547 Document use of vec_xxsldwi and vec_xxpermdi builtins.
2549 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
2551 PR rtl-optimization/60650
2552 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
2554 (find_spills_for): New.
2555 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
2556 Spill all pseudos on the second iteration.
2558 2014-03-27 Marek Polacek <polacek@redhat.com>
2561 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
2564 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2566 * config/s390/s390.c (s390_can_use_return_insn): Check for
2567 call-saved FPRs on 31 bit.
2569 2014-03-27 Jakub Jelinek <jakub@redhat.com>
2572 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
2573 if they need regimplification, just drop them instead of
2574 calling gimple_regimplify_operands on them.
2576 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
2579 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
2580 (aarch64_frame_pointer_required): Adjust logic.
2581 (aarch64_can_eliminate): Adjust logic.
2582 (aarch64_override_options_after_change): Adjust logic.
2584 2014-03-27 Dehao Chen <dehao@google.com>
2586 * ipa-inline.c (early_inliner): Update node's inline info.
2588 2014-03-26 Dehao Chen <dehao@google.com>
2590 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
2591 compiler inserted conditional jumps for NAN float check.
2593 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2595 * ubsan.h (ubsan_create_data): Change second argument's type
2596 to const location_t *.
2597 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
2599 (ubsan_create_data): Change second argument to const location_t *PLOC.
2600 Create Loc field whenever PLOC is non-NULL.
2601 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
2602 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
2606 * real.c (real_to_integer2): Change type of low to UHWI.
2608 2014-03-26 Tobias Burnus <burnus@net-b.de>
2610 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
2611 (CILK_SELF_SPECS): New define.
2612 (driver_self_specs): Use it.
2614 2014-03-26 Richard Biener <rguenther@suse.de>
2616 * tree-pretty-print.c (percent_K_format): Implement special
2617 case for LTO and its stripped down BLOCK tree.
2619 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2622 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
2624 * tree-vrp.c (simplify_internal_call_using_ranges): If only
2625 one range is range_int_cst_p, but not both, at least optimize
2626 addition/subtraction of 0 and multiplication by 0 or 1.
2627 * gimple-fold.c (gimple_fold_call): Fold
2628 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
2629 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
2630 INTEGER_CSTs, try to fold at least x * 0 and y - y.
2632 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
2634 PR rtl-optimization/60452
2635 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
2636 <case REG>: Return 1 for invalid offsets from the frame pointer.
2638 2014-03-26 Marek Polacek <polacek@redhat.com>
2641 * doc/extend.texi (C Extensions): Mention variable-length arrays in
2644 2014-03-26 Marek Polacek <polacek@redhat.com>
2647 * doc/extend.texi (Designated Inits): Describe what happens to omitted
2650 2014-03-26 Marek Polacek <polacek@redhat.com>
2653 * ira-color.c (update_conflict_hard_regno_costs): Perform the
2654 multiplication in unsigned type.
2656 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
2658 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
2660 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
2662 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
2664 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
2667 * cif-code.def (UNREACHABLE) New code.
2668 * ipa-inline.c (inline_small_functions): Skip edges to
2669 __builtlin_unreachable.
2670 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
2671 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
2672 predicate to __bulitin_unreachable.
2673 (set_cond_stmt_execution_predicate): Fix issue when
2674 invert_tree_comparison returns ERROR_MARK.
2675 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
2676 propagate to inline clones.
2677 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
2679 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
2680 * cgraphclones.c (cgraph_clone_node): If call destination is already
2681 ureachable, do not redirect it back.
2682 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
2685 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
2687 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
2688 Do not modify inline clones.
2690 2014-03-25 Jakub Jelinek <jakub@redhat.com>
2692 * config/i386/i386.md (general_sext_operand): New mode attr.
2693 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
2694 don't generate (sign_extend (const_int)).
2695 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
2696 operands[2]. Use We constraint instead of <i> and
2697 <general_sext_operand> predicate instead of <general_operand>.
2698 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
2699 * config/i386/constraints.md (We): New constraint.
2700 * config/i386/predicates.md (x86_64_sext_operand,
2701 sext_operand): New predicates.
2703 2014-03-25 Martin Jambor <mjambor@suse.cz>
2706 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
2707 inconsistent devirtualizations to __builtin_unreachable.
2709 2014-03-25 Marek Polacek <polacek@redhat.com>
2712 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
2714 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
2716 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
2717 order of elements for big-endian.
2719 2014-03-25 Richard Biener <rguenther@suse.de>
2722 * gimplify-me.c (gimple_regimplify_operands): Update the
2725 2014-03-25 Martin Jambor <mjambor@suse.cz>
2728 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
2729 (lto_output_varpool_node): Likewise.
2730 (input_overwrite_node): Likewise.
2731 (input_varpool_node): Likewise.
2733 2014-03-25 Richard Biener <rguenther@suse.de>
2735 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
2736 (run_gcc): Likewise.
2738 2014-03-25 Jakub Jelinek <jakub@redhat.com>
2740 * combine.c (simplify_compare_const): Add MODE argument.
2741 Handle mode_width 0 as very large mode_width.
2742 (try_combine, simplify_comparison): Adjust callers.
2744 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
2745 type to avoid signed integer overflow.
2746 * explow.c (plus_constant): Likewise.
2748 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
2750 * doc/generic.texi: Correct typos.
2752 2014-03-24 Tobias Burnus <burnus@net-b.de>
2754 * doc/invoke.texi (-flto): Expand section about
2755 using static libraries with LTO.
2757 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2759 PR rtl-optimization/60501
2760 * optabs.def (addptr3_optab): New optab.
2761 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
2762 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
2763 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
2765 * lra.c (emit_add3_insn): Use the addptr pattern if available.
2767 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
2769 2014-03-24 Ulrich Drepper <drepper@gmail.com>
2771 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
2774 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
2775 (_mm256_undefined_ps): Define.
2776 (_mm256_undefined_pd): Define.
2777 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
2778 (_mm_undefined_pd): Define.
2779 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
2780 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
2781 (_mm512_undefined_ps): Define.
2782 (_mm512_undefined_pd): Define.
2783 Use _mm*_undefined_*.
2784 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
2786 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
2788 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
2789 (lshr_simd): DI mode added.
2790 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
2791 (aarch64_ushr_simddi): Likewise.
2792 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
2793 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
2794 (vshrd_n_u64): Likewise.
2796 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2798 * Makefile.in (s-macro_list): Depend on cc1.
2800 2014-03-23 Teresa Johnson <tejohnson@google.com>
2802 * ipa-utils.c (ipa_print_order): Use specified dump file.
2804 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
2806 PR rtl-optimization/60601
2807 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
2809 * gcc.c (eval_spec_function): Initialize save_growing_value.
2811 2014-03-22 Jakub Jelinek <jakub@redhat.com>
2814 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
2815 code == MINUS_EXPR, never swap op0 with op1.
2817 * toplev.c (init_local_tick): Avoid signed integer multiplication
2819 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
2820 shift by first operand's bitsize.
2822 2014-03-21 Jakub Jelinek <jakub@redhat.com>
2825 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
2827 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
2828 TARGET_ISA_64BIT_P(x).
2830 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2832 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
2833 pattern for vector nor instead of subtract from splat(-1).
2834 (altivec_expand_vec_perm_const_le): Likewise.
2836 2014-03-21 Richard Henderson <rth@twiddle.net>
2839 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
2840 related insns after epilogue_completed.
2842 2014-03-21 Martin Jambor <mjambor@suse.cz>
2845 * cgraph.h (symtab_node): New flag body_removed.
2846 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
2847 when removing bodies.
2848 * symtab.c (dump_symtab_base): Dump body_removed flag.
2849 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
2850 had their bodies removed.
2852 2014-03-21 Martin Jambor <mjambor@suse.cz>
2855 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
2858 2014-03-21 Richard Biener <rguenther@suse.de>
2860 PR tree-optimization/60577
2861 * tree-core.h (struct tree_base): Document nothrow_flag use
2862 in VAR_DECL_NONALIASED.
2863 * tree.h (VAR_DECL_NONALIASED): New.
2864 (may_be_aliased): Adjust.
2865 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
2867 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
2869 * expr.c (expand_expr_real_1): Remove outdated comment.
2871 2014-03-20 Jakub Jelinek <jakub@redhat.com>
2874 * ira.c (adjust_cleared_regs): Call copy_rtx on
2875 *reg_equiv[REGNO (loc)].src_p before passing it to
2876 simplify_replace_fn_rtx.
2879 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
2880 into CONST, put pic register as first operand of PLUS. Use
2881 gen_const_mem for both 32-bit and 64-bit PIC got loads.
2883 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2885 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
2887 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
2889 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
2890 around for store forwarding issue in the FPU on the UT699.
2891 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
2892 loads and operations if -mfix-ut699 is specified.
2893 (divtf3_hq): Tweak attribute.
2894 (sqrttf2_hq): Likewise.
2896 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
2898 * calls.c (store_one_arg): Remove incorrect const qualification on the
2899 type of the temporary.
2900 * cfgexpand.c (expand_return): Likewise.
2901 * expr.c (expand_constructor): Likewise.
2902 (expand_expr_real_1): Likewise.
2904 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2906 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
2909 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
2912 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
2914 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
2916 * config/arm/aarch-common-protos.h
2917 (alu_cost_table): Fix spelling of "extend".
2918 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
2920 2014-03-19 Richard Biener <rguenther@suse.de>
2923 * tree-core.h (tree_type_common): Re-order pointer members
2924 to reduce recursion depth during GC walks.
2926 2014-03-19 Marek Polacek <polacek@redhat.com>
2929 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
2930 before accessing it.
2932 2014-03-19 Richard Biener <rguenther@suse.de>
2935 * lto-streamer-in.c (input_function): In WPA stage do not drop
2938 2014-03-19 Jakub Jelinek <jakub@redhat.com>
2940 PR tree-optimization/60559
2941 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
2942 with build_zero_cst assignment.
2944 2014-03-18 Kai Tietz <ktietz@redhat.com>
2946 PR rtl-optimization/56356
2947 * sdbout.c (sdbout_parms): Verify that parms'
2948 incoming argument is valid.
2949 (sdbout_reg_parms): Likewise.
2951 2014-03-18 Richard Henderson <rth@redhat.com>
2954 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
2955 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
2956 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
2958 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
2960 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
2961 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
2962 Italicize plugin event names in description. Explain that
2963 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
2964 Remind that no GCC functions should be called after PLUGIN_FINISH.
2965 Explain what pragmas with expansion are.
2967 2014-03-18 Martin Liska <mliska@suse.cz>
2969 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
2970 gimple call statement is update.
2971 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
2972 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
2974 2014-03-18 Jakub Jelinek <jakub@redhat.com>
2977 * ubsan.c (ubsan_instrument_unreachable): Call
2978 initialize_sanitizer_builtins.
2979 (ubsan_pass): Likewise.
2982 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
2983 varpool_finalize_decl instead of rest_of_decl_compilation.
2985 2014-03-18 Richard Biener <rguenther@suse.de>
2987 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
2988 by using bitmap_and_compl instead of bitmap_and_compl_into.
2989 (df_rd_transfer_function): Likewise.
2991 2014-03-18 Richard Biener <rguenther@suse.de>
2993 * doc/lto.texi (fresolution): Fix typo.
2995 2014-03-18 Richard Biener <rguenther@suse.de>
2997 * doc/invoke.texi (flto): Update for changes in 4.9.
2999 2014-03-18 Richard Biener <rguenther@suse.de>
3001 * doc/loop.texi: Remove section on the removed lambda framework.
3002 Update loop docs with recent changes in preserving loop structure.
3004 2014-03-18 Richard Biener <rguenther@suse.de>
3006 * doc/lto.texi (-fresolution): Document.
3008 2014-03-18 Richard Biener <rguenther@suse.de>
3010 * doc/contrib.texi: Adjust my name.
3012 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3015 * internal-fn.c: Include diagnostic-core.h.
3016 (expand_BUILTIN_EXPECT): New function.
3017 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3018 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3019 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3020 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3022 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3023 Revert 3 argument __builtin_expect code.
3024 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3025 * gimple-fold.c (gimple_fold_call): Likewise.
3026 * tree.h (fold_builtin_expect): New prototype.
3027 * builtins.c (build_builtin_expect_predicate): Add predictor
3028 argument, if non-NULL, create 3 argument __builtin_expect.
3029 (fold_builtin_expect): No longer static. Add ARG2 argument,
3030 pass it through to build_builtin_expect_predicate.
3031 (fold_builtin_2): Adjust caller.
3032 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3033 * internal-fn.def (BUILTIN_EXPECT): New.
3035 2014-03-18 Tobias Burnus <burnus@net-b.de>
3038 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3039 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3040 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3042 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3045 * predict.c (combine_predictions_for_bb): Fix up formatting.
3046 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3047 fill what it points to if non-NULL.
3048 (tree_predict_by_opcode): Adjust caller, use the predictor.
3049 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3051 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3053 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3054 proper constant for the store mode.
3056 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3058 * symtab.c (change_decl_assembler_name): Fix transparent alias
3061 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3063 * config/aarch64/aarch64.c: Correct the comments about the
3064 aarch64 stack layout.
3066 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3068 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3069 check for GF_OMP_FOR_KIND_FOR.
3071 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3073 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3074 ymm and zmm register names.
3076 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3079 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3080 note creation for the 2010-08-31 changes.
3082 2014-03-17 Marek Polacek <polacek@redhat.com>
3085 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3086 as -fno-tree-loop-vectorize.
3087 (expand_omp_simd): Likewise.
3089 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3091 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3092 (eligible_for_call_delay): New prototype.
3093 * config/sparc/sparc.c (tls_call_delay): Rename into...
3094 (eligible_for_call_delay): ...this. Return false if the instruction
3095 cannot be put in the delay slot of a branch.
3096 (eligible_for_restore_insn): Simplify.
3097 (eligible_for_return_delay): Return false if the instruction cannot be
3098 put in the delay slot of a branch and simplify.
3099 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3100 put in the delay slot of a branch.
3101 * config/sparc/sparc.md (fix_ut699): New attribute.
3102 (tls_call_delay): Delete.
3103 (in_call_delay): Reimplement.
3104 (eligible_for_sibcall_delay): Rename into...
3105 (in_sibcall_delay): ...this.
3106 (eligible_for_return_delay): Rename into...
3107 (in_return_delay): ...this.
3108 (in_branch_delay): Reimplement.
3109 (in_uncond_branch_delay): Delete.
3110 (in_annul_branch_delay): Delete.
3112 2014-03-14 Richard Henderson <rth@redhat.com>
3115 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3116 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3117 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3118 (floathi splitters): Remove.
3119 (float<SWI48x>xf2): New pattern.
3120 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3121 code that tried to handle DImode for 32-bit, but which was excluded
3122 by the pattern's condition. Drop allocation of stack temporary.
3123 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3124 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3125 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3126 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3127 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3128 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3129 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3130 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3131 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3132 (*float<SWI48x><X87MODEF>2_i387): Remove.
3133 (all float _with_temp splitters): Remove.
3134 (*float<SWI48x><MODEF>2_i387): New pattern.
3135 (*float<SWI48><MODEF>2_sse): New pattern.
3136 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3137 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3139 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3140 Marek Polacek <polacek@redhat.com>
3143 * common.opt (dump_base_name_prefixed): New Variable.
3144 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3145 if x_dump_base_name_prefixed is already set, set it at the end.
3147 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3149 PR rtl-optimization/60508
3150 * lra-constraints.c (get_reload_reg): Add new parameter
3152 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3153 Pass the new parameter values.
3155 2014-03-14 Richard Biener <rguenther@suse.de>
3157 * common.opt: Revert unintented changes from r205065.
3160 2014-03-14 Richard Biener <rguenther@suse.de>
3163 * cfghooks.c (split_block): Properly adjust all loops the
3164 block was a latch of.
3166 2014-03-14 Martin Jambor <mjambor@suse.cz>
3169 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3172 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3175 * config/avr/avr.c (avr_set_current_function): Pass function name
3176 through default_strip_name_encoding before sanity checking instead
3177 of skipping the first char of the assembler name.
3179 2014-03-13 Richard Henderson <rth@redhat.com>
3182 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3183 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3184 * config/i386/i386-protos.h: Likewise.
3185 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3186 in the expander instead of a splitter.
3187 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3188 any possibility of requiring a memory.
3189 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3190 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3191 (fp branch splitters): Update for ix86_split_fp_branch.
3192 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3193 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3194 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3195 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3196 (*fop_<MODEF>_3_i387): Likewise.
3197 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3198 (splitters for the fop_* register patterns): Remove.
3199 (fscalexf4_i387): Rename from *fscalexf4_i387.
3200 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3202 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3204 PR tree-optimization/59779
3205 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3206 type for bitsize and maxsize instead of HOST_WIDE_INT.
3208 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3210 PR rtl-optimization/57320
3211 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3212 the CFG after thread_prologue_and_epilogue_insns.
3214 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3216 PR rtl-optimization/57189
3217 * lra-constraints.c (process_alt_operands): Disfavor spilling
3220 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3222 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3224 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3226 PR tree-optimization/59025
3228 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3229 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3231 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3234 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3235 calls of avr_out_plus_1.
3237 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3239 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3240 BB's single pred and update the father loop's latch info later.
3242 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3244 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3248 (VEC_base): Likewise.
3249 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3250 registers, we need to swap double words in little endian mode.
3252 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3253 to be a container mode for 128-bit integer operations added in ISA
3254 2.07. Unlike TImode and PTImode, the preferred register set is
3255 the Altivec/VMX registers for the 128-bit operations.
3257 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3259 (rs6000_split_128bit_ok_p): Likewise.
3261 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3262 macros for creating ISA 2.07 normal and overloaded builtin
3263 functions with 3 arguments.
3264 (BU_P8V_OVERLOAD_3): Likewise.
3265 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3266 for use as overloaded functions.
3267 (VPERM_1TI_UNS): Likewise.
3268 (VSEL_1TI): Likewise.
3269 (VSEL_1TI_UNS): Likewise.
3270 (ST_INTERNAL_1ti): Likewise.
3271 (LD_INTERNAL_1ti): Likewise.
3272 (XXSEL_1TI): Likewise.
3273 (XXSEL_1TI_UNS): Likewise.
3274 (VPERM_1TI): Likewise.
3275 (VPERM_1TI_UNS): Likewise.
3276 (XXPERMDI_1TI): Likewise.
3277 (SET_1TI): Likewise.
3278 (LXVD2X_V1TI): Likewise.
3279 (STXVD2X_V1TI): Likewise.
3280 (VEC_INIT_V1TI): Likewise.
3281 (VEC_SET_V1TI): Likewise.
3282 (VEC_EXT_V1TI): Likewise.
3283 (EQV_V1TI): Likewise.
3284 (NAND_V1TI): Likewise.
3285 (ORC_V1TI): Likewise.
3286 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3287 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3289 (VADDUQM): Likewise.
3290 (VSUBCUQ): Likewise.
3291 (VADDEUQM): Likewise.
3292 (VADDECUQ): Likewise.
3293 (VSUBEUQM): Likewise.
3294 (VSUBECUQ): Likewise.
3296 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3297 __int128_t and __uint128_t types.
3298 (__uint128_type): Likewise.
3299 (altivec_categorize_keyword): Add support for vector __int128_t,
3300 vector __uint128_t, vector __int128, and vector unsigned __int128
3301 as a container type for TImode operations that need to be done in
3302 VSX/Altivec registers.
3303 (rs6000_macro_to_expand): Likewise.
3304 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3305 to support 128-bit integer instructions vaddcuq, vadduqm,
3306 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3307 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3309 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3310 for V1TImode, and set up preferences to use VSX/Altivec registers.
3311 Setup VSX reload handlers.
3312 (rs6000_debug_reg_global): Likewise.
3313 (rs6000_init_hard_regno_mode_ok): Likewise.
3314 (rs6000_preferred_simd_mode): Likewise.
3315 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3316 (easy_altivec_constant): Likewise.
3317 (output_vec_const_move): Likewise.
3318 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3320 (rs6000_expand_vector_extract): Likewise.
3321 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3323 (rs6000_const_vec): Add support for V1TImode.
3324 (rs6000_emit_le_vsx_load): Swap double words when loading or
3325 storing TImode/V1TImode.
3326 (rs6000_emit_le_vsx_store): Likewise.
3327 (rs6000_emit_le_vsx_move): Likewise.
3328 (rs6000_emit_move): Add support for V1TImode.
3329 (altivec_expand_ld_builtin): Likewise.
3330 (altivec_expand_st_builtin): Likewise.
3331 (altivec_expand_vec_init_builtin): Likewise.
3332 (altivec_expand_builtin): Likewise.
3333 (rs6000_init_builtins): Add support for V1TImode type. Add
3334 support for ISA 2.07 128-bit integer builtins. Define type names
3335 for the VSX/Altivec vector types.
3336 (altivec_init_builtins): Add support for overloaded vector
3337 functions with V1TImode type.
3338 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3339 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3341 (rs6000_split_128bit_ok_p): Likewise.
3342 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3343 __int128_t and vector __uint128_t.
3345 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3346 and mode attributes.
3353 (VS_scalar): Likewise.
3354 (VS_double): Likewise.
3355 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3357 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3358 we support the ISA 2.07 128-bit integer arithmetic instructions.
3359 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3360 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3361 and TImode types for use with the builtin functions.
3362 (V1TI_type_node): Likewise.
3363 (unsigned_V1TI_type_node): Likewise.
3364 (intTI_type_internal_node): Likewise.
3365 (uintTI_type_internal_node): Likewise.
3367 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3368 128-bit builtin functions.
3369 (UNSPEC_VADDEUQM): Likewise.
3370 (UNSPEC_VADDECUQ): Likewise.
3371 (UNSPEC_VSUBCUQ): Likewise.
3372 (UNSPEC_VSUBEUQM): Likewise.
3373 (UNSPEC_VSUBECUQ): Likewise.
3374 (VM): Add V1TImode to vector mode iterators.
3376 (VI_unit): Likewise.
3377 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3378 (altivec_vaddcuq): Likewise.
3379 (altivec_vsubuqm): Likewise.
3380 (altivec_vsubcuq): Likewise.
3381 (altivec_vaddeuqm): Likewise.
3382 (altivec_vaddecuq): Likewise.
3383 (altivec_vsubeuqm): Likewise.
3384 (altivec_vsubecuq): Likewise.
3386 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3388 (BOOL_128): Likewise.
3389 (BOOL_REGS_OUTPUT): Likewise.
3390 (BOOL_REGS_OP1): Likewise.
3391 (BOOL_REGS_OP2): Likewise.
3392 (BOOL_REGS_UNARY): Likewise.
3393 (BOOL_REGS_AND_CR0): Likewise.
3395 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3396 128-bit integer builtin support.
3397 (vec_vadduqm): Likewise.
3398 (vec_vaddecuq): Likewise.
3399 (vec_vaddeuqm): Likewise.
3400 (vec_vsubecuq): Likewise.
3401 (vec_vsubeuqm): Likewise.
3402 (vec_vsubcuq): Likewise.
3403 (vec_vsubuqm): Likewise.
3405 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3406 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3407 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3408 128-bit integer add/subtract to ISA 2.07.
3410 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3412 * config/arc/arc.c (arc_predicate_delay_insns):
3413 Fix third argument passed to conditionalize_nonjump.
3415 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3417 * config/aarch64/aarch64-builtins.c
3418 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3419 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3420 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3421 instead of __builtin_lfloor.
3422 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3424 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3426 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3427 (tree_ssa_ifcombine_bb_1): New function.
3428 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3429 is an empty forwarder block to then_bb or vice versa and then_bb
3430 and else_bb are effectively swapped.
3432 2014-03-12 Christian Bruel <christian.bruel@st.com>
3435 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3436 REG_CFA_DEF_CFA note.
3437 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3438 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3440 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3442 PR tree-optimization/60454
3443 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3445 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3447 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3448 Do not define target_cpu_default2 to generic.
3449 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3450 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3451 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3453 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3454 Marc Glisse <marc.glisse@inria.fr>
3456 PR tree-optimization/60502
3457 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3458 instead of build_low_bits_mask.
3460 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3463 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3464 if there are multiple uses, but op doesn't live on E edge.
3465 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3466 clobber stmts before __builtin_unreachable.
3468 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
3470 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
3471 hard_frame_pointer_rtx.
3472 * cse.c (cse_insn): Remove volatile check.
3473 * cselib.c (cselib_process_insn): Likewise.
3474 * dse.c (scan_insn): Likewise.
3476 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3478 * config/arc/arc.c (conditionalize_nonjump): New function,
3480 (arc_ifcvt): ... this.
3481 (arc_predicate_delay_insns): Use it.
3483 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3485 * config/arc/predicates.md (extend_operand): During/after reload,
3486 allow const_int_operand.
3487 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
3488 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
3489 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
3491 (umulsi3_highpart_i): Likewise.
3493 2014-03-11 Richard Biener <rguenther@suse.de>
3495 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
3496 Add asserts to guard possible wrong-code bugs.
3498 2014-03-11 Richard Biener <rguenther@suse.de>
3500 PR tree-optimization/60429
3501 PR tree-optimization/60485
3502 * tree-ssa-structalias.c (set_union_with_increment): Properly
3503 take into account all fields that overlap the shifted vars.
3504 (do_sd_constraint): Likewise.
3505 (do_ds_constraint): Likewise.
3506 (get_constraint_for_ptr_offset): Likewise.
3508 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
3510 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
3511 (nios2_compute_frame_layout):
3512 Add calculation of cfun->machine->fp_save_offset.
3513 (nios2_expand_prologue): Correct setting of frame pointer register
3515 (nios2_expand_epilogue): Update recovery of stack pointer from
3516 frame pointer accordingly.
3517 (nios2_initial_elimination_offset): Update calculation of offset
3518 for eliminating to HARD_FRAME_POINTER_REGNUM.
3520 2014-03-10 Jakub Jelinek <jakub@redhat.com>
3523 * ipa.c (symtab_remove_unreachable_nodes): Don't call
3524 cgraph_get_create_node on VAR_DECLs.
3526 2014-03-10 Richard Biener <rguenther@suse.de>
3529 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
3531 2014-03-08 Douglas B Rupp <rupp@gnat.com>
3533 * config/vms/vms.opt (vms_float_format): New variable.
3535 2014-03-08 Tobias Burnus <burnus@net-b.de>
3537 * doc/invoke.texi (-fcilkplus): Update implementation status.
3539 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
3540 Richard Biener <rguenther@suse.de>
3542 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
3543 consistently accross all TUs.
3544 (run_gcc): Enable -fshort-double automatically at link at link-time
3545 and disallow override.
3547 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
3550 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
3551 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
3552 if they can't be used.
3554 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3556 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
3557 for Solaris 11/x86 ld.
3558 * configure: Regenerate.
3560 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3562 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
3563 (LIB_TLS_SPEC): Save as ld_tls_libs.
3564 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
3565 (HAVE_AS_IX86_TLSLDM): New test.
3566 * configure, config.in: Regenerate.
3567 * config/i386/i386.c (legitimize_tls_address): Fall back to
3568 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
3569 cannot support TLS_MODEL_LOCAL_DYNAMIC.
3570 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
3571 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
3573 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
3575 * common.opt (fira-loop-pressure): Mark as optimization.
3577 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
3579 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
3580 an OpenMP mappable type.
3582 2014-03-06 Matthias Klose <doko@ubuntu.com>
3584 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
3585 MULTILIB_OSDIRNAMES is not defined.
3587 2014-03-06 Jakub Jelinek <jakub@redhat.com>
3588 Meador Inge <meadori@codesourcery.com>
3591 * config/arm/arm.c (arm_tls_symbol_p): Remove.
3592 (arm_legitimize_address): Call legitimize_tls_address for any
3593 arm_tls_referenced_p expression, handle constant addend. Call it
3594 before testing for !TARGET_ARM.
3595 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
3597 2014-03-06 Richard Biener <rguenther@suse.de>
3603 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
3605 * tree-streamer.c (record_common_node): Assert we don't record
3606 nodes with type double.
3607 (preload_common_node): Skip type double, complex double and double
3608 pointer since it is now frontend dependent due to fshort-double option.
3610 2014-03-06 Richard Biener <rguenther@suse.de>
3612 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
3613 or -fno-lto is specified and the linker has full plugin support.
3614 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
3615 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
3616 * lto-wrapper.c (merge_and_complain): Merge compile-time
3617 optimization levels.
3618 (run_gcc): And pass it through to the link options.
3620 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
3624 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
3626 * cselib.c (remove_useless_values): Skip to avoid quadratic
3627 behavior if the condition moved from...
3628 (cselib_process_insn): ... here holds.
3630 2014-03-05 Jakub Jelinek <jakub@redhat.com>
3633 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
3634 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
3637 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
3638 (TM_H): Add x86-tune.def.
3640 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3642 * config/aarch64/aarch64.c (generic_tunings):
3643 Use cortexa57_extra_costs.
3645 2014-03-05 Jakub Jelinek <jakub@redhat.com>
3648 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
3649 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
3650 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
3653 2014-03-04 Heiher <r@hev.cc>
3655 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
3656 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
3658 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
3660 * config/i386/predicates.md (const2356_operand): Change to ...
3661 (const2367_operand): ... this.
3662 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
3664 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
3665 (*avx512pf_scatterpf<mode>sf): Ditto.
3666 (avx512pf_scatterpf<mode>df): Ditto.
3667 (*avx512pf_scatterpf<mode>df_mask): Ditto.
3668 (*avx512pf_scatterpf<mode>df): Ditto.
3669 * config/i386/i386.c (ix86_expand_builtin): Update
3670 incorrect hint operand error message.
3672 2014-03-04 Richard Biener <rguenther@suse.de>
3674 * lto-section-in.c (lto_get_section_data): Fix const cast.
3676 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
3678 * tree-streamer.c (record_common_node): Assert we don't record
3679 nodes with type double.
3680 (preload_common_node): Skip type double, complex double and double
3681 pointer since it is now frontend dependent due to fshort-double option.
3683 2014-03-04 Richard Biener <rguenther@suse.de>
3686 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
3687 (lto_input_toplevel_asms): Likewise.
3688 * lto-section-in.c (lto_get_section_data): Instead do it here
3691 2014-03-04 Richard Biener <rguenther@suse.de>
3693 PR tree-optimization/60382
3694 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
3695 dead PHIs a reduction.
3697 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
3699 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
3701 (_mm_prefetch): Move out of GCC target("sse") pragma.
3702 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
3703 GCC target("prfchw") pragma.
3704 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
3706 * config/i386/i386.c (ix86_option_override_internal): Enable
3707 -mprfchw with -mprefetchwt1.
3709 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
3711 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
3714 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
3716 * opts.h (CL_PCH_IGNORE): Define.
3717 * targhooks.c (option_affects_pch_p):
3718 Return false for options that have CL_PCH_IGNORE set.
3719 * opt-functions.awk: Process PchIgnore.
3720 * doc/options.texi: Document PchIgnore.
3722 * config/arc/arc.opt (misize): Add PchIgnore property.
3724 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3726 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
3727 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
3728 constraint on constants to permit them being loaded into
3729 GENERAL_REGS or BASE_REGS.
3731 2014-03-03 Nick Clifton <nickc@redhat.com>
3733 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
3734 anti-cacnonical alternatives.
3735 (negandhi3_real): New pattern.
3736 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
3738 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
3740 * config/avr/avr-mcus.def: Remove atxmega16x1.
3741 * config/avr/avr-tables.opt: Regenerate.
3742 * config/avr/t-multilib: Regenerate.
3743 * doc/avr-mmcu.texi: Regenerate.
3745 2014-03-03 Tobias Grosser <tobias@grosser.es>
3746 Mircea Namolaru <mircea.namolaru@inria.fr>
3748 PR tree-optimization/58028
3749 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
3752 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3754 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
3755 not handled by recognizers.
3757 2014-03-03 Jakub Jelinek <jakub@redhat.com>
3760 * function.c (expand_function_end): Don't emit
3761 clobber_return_register sequence if clobber_after is a BARRIER.
3762 * cfgexpand.c (construct_exit_block): Append instructions before
3763 return_label to prev_bb.
3765 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3767 * config/rs6000/constraints.md: Document reserved use of "wc".
3769 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
3772 * ipa.c (function_and_variable_visibility): When dissolving comdat
3773 group, also set all symbols to local.
3775 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
3780 2013-12-14 Jan Hubicka <jh@suse.cz>
3782 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
3784 2014-03-02 Jon Beniston <jon@beniston.com>
3790 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
3791 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
3792 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
3793 (simple_return, *simple_return): New patterns
3794 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
3795 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
3797 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
3799 * dwarf2out.c (gen_subprogram_die): Tidy.
3801 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
3804 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
3805 (*mov_t_msb_neg_negc): ... this new insn.
3807 2014-02-28 Jason Merrill <jason@redhat.com>
3810 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
3813 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
3816 * dwarf2out.c (decltype_auto_die): New static.
3817 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
3818 (gen_type_die_with_usage): Handle 'decltype(auto)'.
3819 (is_cxx_auto): Likewise.
3821 2014-02-28 Ian Bolton <ian.bolton@arm.com>
3823 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
3824 we are not using general regs only.
3826 2014-02-28 Richard Biener <rguenther@suse.de>
3829 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
3830 previous fix and only allow to remove trivial pre-headers
3831 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
3832 (remove_forwarder_block): Properly update the latch of a loop.
3834 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
3837 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
3838 (cselib_preserved_hash_table): New.
3839 (preserve_constants_and_equivs): Move preserved vals to it.
3840 (cselib_find_slot): Look it up first.
3841 (cselib_init): Initialize it.
3842 (cselib_finish): Release it.
3843 (dump_cselib_table): Dump it.
3845 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
3848 * cselib.c (remove_useless_values): Skip to avoid quadratic
3849 behavior if the condition moved from...
3850 (cselib_process_insn): ... here holds.
3852 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
3855 * var-tracking.c (vt_initialize): Apply the same condition to
3856 preserve the CFA base value.
3858 2014-02-28 Joey Ye <joey.ye@arm.com>
3861 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
3862 if reload in progress or completed.
3864 2014-02-28 Tobias Burnus <burnus@net-b.de>
3867 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
3870 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
3872 * doc/tm.texi.in (Condition Code Status): Update documention for
3873 relative locations of cc0-setter and cc0-user.
3875 2014-02-27 Jeff Law <law@redhat.com>
3877 PR rtl-optimization/52714
3878 * combine.c (try_combine): When splitting an unrecognized PARALLEL
3879 into two independent simple sets, if I3 is a jump, ensure the
3880 pattern we place into I3 is a (set (pc) ...).
3882 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
3883 Jeff Law <law@redhat.com>
3885 PR rtl-optimization/49847
3886 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
3887 are in different blocks.
3888 * doc/tm.texi (Condition Code Status): Update documention for
3889 relative locations of cc0-setter and cc0-user.
3891 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
3894 * lra.c (lra_emit_add): Check SUBREG too.
3896 2014-02-27 Andreas Schwab <schwab@suse.de>
3898 * config/m68k/m68k.c (m68k_option_override): Disable
3899 -flive-range-shrinkage for classic m68k.
3900 (m68k_override_options_after_change): Likewise.
3902 2014-02-27 Marek Polacek <polacek@redhat.com>
3905 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
3906 -Wmaybe-uninitialized.
3908 2014-02-27 Alan Modra <amodra@gmail.com>
3911 * reload1.c (emit_input_reload_insns): When reload_override_in,
3912 set old to rl->in_reg when rl->in_reg is a subreg.
3914 2014-02-26 Richard Biener <rguenther@suse.de>
3917 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
3919 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
3921 * common/config/i386/predicates.md (const1256_operand): Remove.
3922 (const2356_operand): New.
3923 (const_1_to_2_operand): Remove.
3924 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
3925 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
3926 (*avx512pf_gatherpf<mode>sf): Ditto.
3927 (avx512pf_gatherpf<mode>df): Ditto.
3928 (*avx512pf_gatherpf<mode>df_mask): Ditto.
3929 (*avx512pf_gatherpf<mode>df): Ditto.
3930 (avx512pf_scatterpf<mode>sf): Ditto.
3931 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
3932 (*avx512pf_scatterpf<mode>sf): Ditto.
3933 (avx512pf_scatterpf<mode>df): Ditto.
3934 (*avx512pf_scatterpf<mode>df_mask): Ditto.
3935 (*avx512pf_scatterpf<mode>df): Ditto.
3936 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
3938 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
3940 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
3941 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
3942 (_mm512_mask_testn_epi64_mask): Move to ...
3943 * config/i386/avx512cdintrin.h: Here.
3944 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
3945 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
3946 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
3947 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
3948 TARGET_AVX512F from TARGET_AVX512CD.
3950 2014-02-26 Richard Biener <rguenther@suse.de>
3953 * ipa.c (walk_polymorphic_call_targets): Properly guard
3954 call to inline_update_overall_summary.
3956 2014-02-26 Bin Cheng <bin.cheng@arm.com>
3959 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
3960 and latches only if requested. Fix latch if it is removed.
3961 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
3962 LOOPS_HAVE_PREHEADERS.
3964 2014-02-25 Andrew Pinski <apinski@cavium.com>
3966 * builtins.c (expand_builtin_thread_pointer): Create a new target
3967 when the target is NULL.
3969 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
3971 PR rtl-optimization/60317
3972 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
3973 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
3974 * lra-assigns.c: Include params.h.
3975 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
3976 other reload pseudos considerations.
3978 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3980 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
3981 to use canonical form for nor<mode>3.
3983 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3986 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
3989 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
3991 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
3992 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
3993 (ix86_handle_option): Handle OPT_mprefetchwt1.
3994 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
3995 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
3997 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
3998 OPTION_MASK_ISA_PREFETCHWT1.
3999 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4000 (PTA_PREFETCHWT1): New.
4001 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4002 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4003 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4004 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4005 (*prefetch_avx512pf_<mode>_: Change into ...
4006 (*prefetch_prefetchwt1_<mode>: This.
4007 * config/i386/i386.opt (mprefetchwt1): New.
4008 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4009 (_mm_prefetch): Handle intent to write.
4010 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4012 2014-02-25 Richard Biener <rguenther@suse.de>
4015 * emit-rtl.c (mem_attrs_htab): Remove.
4016 (mem_attrs_htab_hash): Likewise.
4017 (mem_attrs_htab_eq): Likewise.
4018 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4019 (init_emit_once): Do not allocate mem_attrs_htab.
4021 2014-02-25 Richard Biener <rguenther@suse.de>
4024 * lto-opts.c (lto_write_options): Output non-explicit conservative
4025 -fwrapv, -fno-trapv and -fno-strict-overflow.
4026 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4027 (run_gcc): And pass them through.
4029 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4031 * sel-sched.c (calculate_new_fences): New parameter ptime.
4032 Calculate it as a maximum over all fence cycles.
4033 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4034 Print the final schedule timing when sched_verbose.
4036 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4038 PR rtl-optimization/60292
4039 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4040 bit fot the fence instruction.
4042 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4044 * calls.h: Fix typo in comment.
4046 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4048 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4049 adjusting offsetable addresses.
4051 2014-02-24 Guozhi Wei <carrot@google.com>
4053 * sparseset.h (sparseset_pop): Fix the wrong index.
4055 2014-02-24 Walter Lee <walt@tilera.com>
4057 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4058 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4060 * common/config/tilegx/tilegx-common.c
4061 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4062 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4064 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4065 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4066 (tilegx_gimplify_va_arg_expr): Handle big endian.
4067 (tilegx_expand_unaligned_load): Ditto.
4068 (tilegx_expand_unaligned_store): Ditto.
4069 (TARGET_RETURN_IN_MSB): New.
4070 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4071 (TARGET_ENDIAN_DEFAULT): New.
4072 (TARGET_BIG_ENDIAN): Handle big endian.
4073 (BYTES_BIG_ENDIAN): Ditto.
4074 (WORDS_BIG_ENDIAN): Ditto.
4075 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4078 * config/tilegx/tilegx.md (extv): Handle big endian.
4080 (insn_st<n>): Ditto.
4081 (insn_st<n>_add<bitsuffix>): Ditto.
4082 (insn_stnt<n>): Ditto.
4083 (insn_stnt<n>_add<bitsuffix>):Ditto.
4084 (vec_interleave_highv8qi): Handle big endian.
4085 (vec_interleave_highv8qi_be): New.
4086 (vec_interleave_highv8qi_le): New.
4087 (insn_v1int_h): Handle big endian.
4088 (vec_interleave_lowv8qi): Handle big endian.
4089 (vec_interleave_lowv8qi_be): New.
4090 (vec_interleave_lowv8qi_le): New.
4091 (insn_v1int_l): Handle big endian.
4092 (vec_interleave_highv4hi): Handle big endian.
4093 (vec_interleave_highv4hi_be): New.
4094 (vec_interleave_highv4hi_le): New.
4095 (insn_v2int_h): Handle big endian.
4096 (vec_interleave_lowv4hi): Handle big endian.
4097 (vec_interleave_lowv4hi_be): New.
4098 (vec_interleave_lowv4hi_le): New.
4099 (insn_v2int_l): Handle big endian.
4100 (vec_interleave_highv2si): Handle big endian.
4101 (vec_interleave_highv2si_be): New.
4102 (vec_interleave_highv2si_le): New.
4103 (insn_v4int_h): Handle big endian.
4104 (vec_interleave_lowv2si): Handle big endian.
4105 (vec_interleave_lowv2si_be): New.
4106 (vec_interleave_lowv2si_le): New.
4107 (insn_v4int_l): Handle big endian.
4108 * config/tilegx/tilegx.opt (mbig-endian): New option.
4109 (mlittle-endian): New option.
4110 * doc/install.texi: Document tilegxbe-linux.
4111 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4113 2014-02-24 Martin Jambor <mjambor@suse.cz>
4116 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4117 there are no parameter descriptors.
4119 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4121 PR rtl-optimization/60268
4122 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4123 initialization to ...
4124 (sched_rgn_init): ... here.
4125 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4127 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4129 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4132 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4134 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4137 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4139 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4140 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4142 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4144 * config/microblaze/predicates.md: Add cmp_op predicate.
4145 * config/microblaze/microblaze.md: Add branch_compare instruction
4146 which uses cmp_op predicate and emits cmp insn before branch.
4147 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4148 to microblaze_expand_conditional_branch and consolidate logic.
4149 (microblaze_expand_conditional_branch): emit branch_compare
4150 insn instead of handling cmp op separate from branch insn.
4152 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4154 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4157 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4159 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4160 define_insn with define_expand and new define_insn
4161 *altivec_lve<VI_char>x_internal.
4162 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4163 and new define_insn *altivec_stve<VI_char>x_internal.
4164 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4166 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4168 (altivec_expand_stvex_be): New function.
4170 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4172 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4173 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4174 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4175 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4177 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4180 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4181 instead of emit_move_insn.
4183 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4185 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4187 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4188 gen_altivec_vsumsws.
4190 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4192 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4193 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4194 (altivec_lvxl_<mode>): New define_expand incorporating
4195 -maltivec=be semantics where needed.
4196 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4197 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4198 semantics where needed.
4199 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4200 (altivec_stvx_<mode>): New define_expand incorporating
4201 -maltivec=be semantics where needed.
4202 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4203 VM2 iterator instead of V4SI.
4204 (altivec_stvxl_<mode>): New define_expand incorporating
4205 -maltivec=be semantics where needed.
4206 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4207 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4208 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4209 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4210 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4211 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4212 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4213 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4214 ALTIVEC_BUILTIN_STVXL.
4215 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4216 (altivec_expand_stvx_be): Likewise.
4217 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4218 (altivec_expand_lvx_be): Likewise.
4219 (altivec_expand_stvx_be): Likewise.
4220 (altivec_expand_builtin): Add cases for
4221 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4222 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4223 (altivec_init_builtins): Add definitions for
4224 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4225 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4227 2014-02-21 Catherine Moore <clm@codesourcery.com>
4229 * doc/invoke.texi (mvirt, mno-virt): Document.
4230 * config/mips/mips.opt (mvirt): New option.
4231 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4233 2014-02-21 Richard Biener <rguenther@suse.de>
4235 PR tree-optimization/60276
4236 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4237 (STMT_VINFO_MIN_NEG_DIST): New macro.
4238 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4239 STMT_VINFO_MIN_NEG_DIST.
4240 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4241 made for negative dependence distances still hold.
4243 2014-02-21 Richard Biener <rguenther@suse.de>
4246 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4247 DECL_INITIAL for globals not in the current function context.
4249 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4251 PR tree-optimization/56490
4252 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4253 * tree-ssa-uninit.c: Include params.h.
4254 (compute_control_dep_chain): Add num_calls argument, return false
4255 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4256 num_calls to recursive call.
4257 (find_predicates): Change dep_chain into normal array,
4258 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4259 variable and adjust compute_control_dep_chain caller.
4260 (find_def_preds): Likewise.
4262 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4264 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4265 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4267 2014-02-21 Nick Clifton <nickc@redhat.com>
4269 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4270 (pushhi1): Likewise.
4271 (popqi1): Add mode to pre_dec.
4274 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4276 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4277 mode for mask of V8SFmode permutation.
4279 2014-02-20 Richard Henderson <rth@redhat.com>
4282 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4283 a new pseudo for OLDVAL.
4285 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4288 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4289 gen_reg_rtx if d->testing_p.
4290 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4291 if d->testing_p and we will certainly return true.
4292 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4295 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4297 * emit-rtl.c (gen_reg_rtx): Assert that
4298 crtl->emit.regno_pointer_align_length is non-zero.
4300 2014-02-20 Richard Henderson <rth@redhat.com>
4303 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4304 on failure the store back into EXPECT.
4306 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4307 Sandra Loosemore <sandra@codesourcery.com>
4309 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4310 * config/nios2/nios2.c (nios2_function_profiler): Add
4311 -fPIC (flag_pic == 2) support.
4312 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4313 (nios2_large_offset_p): New function.
4314 (nios2_unspec_reloc_p): Move up position, update to use
4315 nios2_large_offset_p.
4316 (nios2_unspec_address): Remove function.
4317 (nios2_unspec_offset): New function.
4318 (nios2_large_got_address): New function.
4319 (nios2_got_address): Add large offset support.
4320 (nios2_legitimize_tls_address): Update usage of removed and new
4322 (nios2_symbol_binds_local_p): New function.
4323 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4324 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4325 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4326 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4327 processing for (const (unspec ...)).
4328 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4330 2014-02-20 Richard Biener <rguenther@suse.de>
4332 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4333 doing the substitution.
4334 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4336 2014-02-20 Martin Jambor <mjambor@suse.cz>
4339 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4340 info when checking whether lattices are bottom.
4342 2014-02-20 Richard Biener <rguenther@suse.de>
4345 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4348 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4351 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4352 parameter specifying the scaling.
4353 (inline_call): Update.
4354 (want_inline_recursively): Guard division by zero.
4355 (recursive_inlining): Update.
4356 * ipa-inline.h (clone_inlined_nodes): Update.
4358 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4361 * config/i386/i386.c (classify_argument): Pass structures of size
4362 64 bytes or less in register.
4364 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4365 Kirill Yukhin <kirill.yukhin@intel.com>
4367 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4368 (_mm_rcp28_round_ss): Ditto.
4369 (_mm_rsqrt28_round_sd): Ditto.
4370 (_mm_rsqrt28_round_ss): Ditto.
4371 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4372 (_mm_rcp14_round_ss): Ditto.
4373 (_mm_rsqrt14_round_sd): Ditto.
4374 (_mm_rsqrt14_round_ss): Ditto.
4375 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4376 the first input operand, get rid of match_dup.
4377 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4379 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4381 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4382 operand as the first input operand, set type attribute.
4383 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4385 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4386 operand as the first input operand, set type attribute.
4388 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4390 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4393 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4396 * config/i386/i386.c (construct_container): Remove TFmode check
4397 for X86_64_INTEGER_CLASS.
4399 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4402 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4403 only when -Wpsabi is enabled.
4405 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4408 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4409 passing arrays in registers are the same as for structs, so remove the
4410 special case for them.
4412 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4414 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4415 destination type, extract only the valid bits if the source type is not
4416 integral and has a different mode.
4418 2014-02-19 Richard Biener <rguenther@suse.de>
4421 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4424 2014-02-19 Richard Biener <rguenther@suse.de>
4427 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4428 (ipa_modify_call_arguments): Emit an argument load explicitely and
4429 preserve virtual SSA form there and for the replacement call.
4430 Do not update SSA form nor free dominance info.
4432 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4434 * ipa.c (function_and_variable_visibility): Also clear WEAK
4435 flag when disolving COMDAT_GROUP.
4437 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4439 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4440 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4442 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4443 do more sanity checks.
4444 (detect_type_change): Return true when giving up early.
4445 (compute_complex_assign_jump_func): Fix type parameter of
4446 ipa_set_ancestor_jf.
4447 (compute_complex_ancestor_jump_func): Likewise.
4448 (update_jump_functions_after_inlining): Fix updating of
4450 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4452 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4454 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4455 inline clones when edge disappears.
4457 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4460 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4461 Split 64-bit moves into 2 patterns. Do not allow the use of
4462 direct move for TDmode in little endian, since the decimal value
4463 has little endian bytes within a word, but the 64-bit pieces are
4464 ordered in a big endian fashion, and normal subreg's of TDmode are
4466 (mov<mode>_64bit_dm): Likewise.
4467 (movtd_64bit_nodm): Likewise.
4469 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4471 PR tree-optimization/60174
4472 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
4473 statement of an SSA_NAME that occurs in an abnormal PHI node.
4475 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4478 * final.c (SEEN_BB): Remove.
4479 (SEEN_NOTE, SEEN_EMITTED): Renumber.
4480 (final_scan_insn): Don't force_source_line on second
4481 NOTE_INSN_BASIC_BLOCK.
4483 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
4486 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
4487 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
4488 (type_natural_mode): Warn ABI change when %zmm register is not
4489 available for AVX512F vector value passing.
4491 2014-02-18 Kai Tietz <ktietz@redhat.com>
4494 * config/i386/i386.c (ix86_expand_prologue): Use value in
4495 rax register as displacement when restoring %r10 or %rax.
4496 Fix wrong offset when restoring both registers.
4498 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4500 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
4501 assertion with conditional return.
4503 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4504 Uros Bizjak <ubizjak@gmail.com>
4507 * config/i386/driver-i386.c (host_detect_local_cpu): If
4508 YMM state is not saved by the OS, also clear has_f16c. Move
4509 CPUID 0x80000001 handling before YMM state saving checking.
4511 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
4513 PR rtl-optimization/58960
4514 * haifa-sched.c (alloc_global_sched_pressure_data): New,
4515 factored out from ...
4516 (sched_init): ... here.
4517 (free_global_sched_pressure_data): New, factored out from ...
4518 (sched_finish): ... here.
4519 * sched-int.h (free_global_sched_pressure_data): Declare.
4520 * sched-rgn.c (nr_regions_initial): New static global.
4521 (haifa_find_rgns): Initialize it.
4522 (schedule_region): Disable sched-pressure for the newly
4525 2014-02-17 Richard Biener <rguenther@suse.de>
4527 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
4528 release SSA defs of pattern stmts.
4530 2014-02-17 Richard Biener <rguenther@suse.de>
4532 * tree-inline.c (expand_call_inline): Release the virtual
4533 operand defined by the call we are about to inline.
4535 2014-02-17 Richard Biener <rguenther@suse.de>
4537 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
4539 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
4540 Ilya Tocar <ilya.tocar@intel.com>
4542 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
4543 arguments order in builtin.
4544 (_mm512_permutexvar_epi64): Ditto.
4545 (_mm512_mask_permutexvar_epi64): Ditto
4546 (_mm512_maskz_permutexvar_epi32): Ditto
4547 (_mm512_permutexvar_epi32): Ditto
4548 (_mm512_mask_permutexvar_epi32): Ditto
4550 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4552 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
4553 (p8_vmrgow): Likewise.
4555 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4557 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
4560 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
4563 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
4564 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
4565 into 64-bit and 32-bit moves. On 64-bit moves, add support for
4566 using direct move instructions on ISA 2.07. Also adjust
4567 instruction length for 64-bit.
4568 (mov<mode>_64bit, TFmode/TDmode): Likewise.
4569 (mov<mode>_32bit, TFmode/TDmode): Likewise.
4571 2014-02-15 Alan Modra <amodra@gmail.com>
4575 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
4576 find_replacement on parts of insn rtl that might be reloaded.
4578 2014-02-15 Richard Biener <rguenther@suse.de>
4580 PR tree-optimization/60183
4581 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
4582 (tree_ssa_phiprop): Calculate and free post-dominators.
4584 2014-02-14 Jeff Law <law@redhat.com>
4586 PR rtl-optimization/60131
4587 * ree.c (get_extended_src_reg): New function.
4588 (combine_reaching_defs): Use it rather than assuming location of REG.
4589 (find_and_remove_re): Verify first operand of extension is
4590 a REG before adding the insns to the copy list.
4592 2014-02-14 Roland McGrath <mcgrathr@google.com>
4594 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
4595 * configure: Regenerated.
4596 * config.in: Regenerated.
4597 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
4598 instead of ASM_SHORT.
4600 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
4601 Richard Earnshaw <rearnsha@arm.com>
4603 PR rtl-optimization/59535
4604 * lra-constraints.c (process_alt_operands): Encourage alternative
4605 when unassigned pseudo class is superset of the alternative class.
4606 (inherit_reload_reg): Don't inherit when optimizing for code size.
4607 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
4608 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
4609 modes not less than 4 for Thumb1.
4611 2014-02-14 Kyle McMartin <kyle@redhat.com>
4614 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
4616 2014-02-14 Richard Biener <rguenther@suse.de>
4618 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
4619 (get_frame_arg): Drop the assert with langhook types_compatible_p.
4620 Do not strip INDIRECT_REFs.
4622 2014-02-14 Richard Biener <rguenther@suse.de>
4625 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
4626 DECL_FUNCTION_SPECIFIC_TARGET.
4627 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
4628 * tree-streamer-out.c (pack_ts_target_option): Remove.
4629 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
4630 (write_ts_function_decl_tree_pointers): Do not stream
4631 DECL_FUNCTION_SPECIFIC_TARGET.
4632 * tree-streamer-in.c (unpack_ts_target_option): Remove.
4633 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
4634 (lto_input_ts_function_decl_tree_pointers): Do not stream
4635 DECL_FUNCTION_SPECIFIC_TARGET.
4637 2014-02-14 Jakub Jelinek <jakub@redhat.com>
4639 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
4640 (get_initial_def_for_induction, vectorizable_induction): Ignore
4641 debug stmts when looking for exit_phi.
4642 (vectorizable_live_operation): Fix up condition.
4644 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4646 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
4647 nreverse() because it changes the content of original tree list.
4649 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4651 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
4652 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
4654 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
4656 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
4657 GNU coding standards.
4659 2014-02-13 Jakub Jelinek <jakub@redhat.com>
4662 * dwarf2out.c (gen_subprogram_die): Don't call
4663 add_calling_convention_attribute if subr_die is old_die.
4665 2014-02-13 Sharad Singhai <singhai@google.com>
4667 * doc/optinfo.texi: Fix order of nodes.
4669 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
4671 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
4672 operands[2], not operands[3].
4674 2014-02-13 Richard Biener <rguenther@suse.de>
4677 * doc/install.texi (ISL): Update recommended version to 0.12.2,
4678 mention the possibility of an in-tree build.
4679 (CLooG): Update recommended version to 0.18.1, mention the
4680 possibility of an in-tree build and clarify that the ISL
4681 bundled with CLooG does not work.
4683 2014-02-13 Jakub Jelinek <jakub@redhat.com>
4686 * expr.c (compress_float_constant): If x is a hard register,
4687 extend into a pseudo and then move to x.
4689 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
4691 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
4692 caused by bad second argument to warning_at() with -mhotpatch and
4693 nested functions (e.g. with gfortran).
4695 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
4697 * opts.c (option_name): Remove "enabled by default" rider.
4699 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
4701 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
4703 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
4704 Uros Bizjak <ubizjak@gmail.com>
4707 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
4708 * configure: Regenerated.
4710 2014-02-12 Richard Biener <rguenther@suse.de>
4712 * vec.c (vec_prefix::calculate_allocation): Move as
4713 inline variant to vec.h.
4714 (vec_prefix::calculate_allocation_1): New out-of-line version.
4715 * vec.h (vec_prefix::calculate_allocation_1): Declare.
4716 (vec_prefix::m_has_auto_buf): Rename to ...
4717 (vec_prefix::m_using_auto_storage): ... this.
4718 (vec_prefix::calculate_allocation): Inline the easy cases
4719 and dispatch to calculate_allocation_1 which doesn't need the
4721 (va_heap::reserve): Use gcc_checking_assert.
4722 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
4723 m_using_auto_storage.
4724 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
4726 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
4727 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
4728 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
4730 2014-02-12 Richard Biener <rguenther@suse.de>
4732 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
4733 when we found a dependence.
4735 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
4737 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
4739 (maybe_fold_stmt): ... into this new function.
4740 * omp-low.c (lower_omp): Update comment.
4742 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
4745 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
4748 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
4750 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
4751 identifiers in comments.
4752 (cortexa53_extra_costs): Likewise.
4753 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
4754 (cortexa7_extra_costs): Likewise.
4755 (cortexa12_extra_costs): Likewise.
4756 (cortexa15_extra_costs): Likewise.
4757 (v7m_extra_costs): Likewise.
4759 2014-02-12 Richard Biener <rguenther@suse.de>
4762 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
4763 of posix_memalign being successful.
4764 (lower_stmt): Restrict lowering of posix_memalign to when
4765 -ftree-bit-ccp is enabled.
4767 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4769 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
4771 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
4773 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
4775 PR rtl-optimization/60116
4776 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
4777 other_insn once the combination has been validated.
4779 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
4782 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
4784 * ipa-devirt.c: Include demangle.h
4785 (odr_violation_reported): New static variable.
4786 (add_type_duplicate): Update odr_violations.
4787 (maybe_record_node): Add completep parameter; update it.
4788 (record_target_from_binfo): Add COMPLETEP parameter;
4789 update it as needed.
4790 (possible_polymorphic_call_targets_1): Likewise.
4791 (struct polymorphic_call_target_d): Add nonconstruction_targets;
4792 rename FINAL to COMPLETE.
4793 (record_targets_from_bases): Sanity check we found the binfo;
4794 fix COMPLETEP updating.
4795 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
4796 parameter, fix computing of COMPLETEP.
4797 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
4798 at LTO time do demangling.
4799 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
4800 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
4802 (gimple_get_virt_method_for_binfo): Likewise.
4803 * gimple-fold.h (gimple_get_virt_method_for_binfo,
4804 gimple_get_virt_method_for_vtable): Update prototypes.
4806 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
4809 * genautomata.c (add_presence_absence): Fix typo with
4810 {final_}presence_list.
4812 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
4815 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
4816 for VSX/Altivec vectors that land in GPR registers.
4818 2014-02-11 Richard Henderson <rth@redhat.com>
4819 Jakub Jelinek <jakub@redhat.com>
4822 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
4823 around drhs if type conversion to lacc->type is not useless.
4825 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4827 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
4829 (cortex-a57.cortex-a53): Likewise.
4830 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
4832 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4834 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
4837 2014-02-11 Renlin Li <Renlin.Li@arm.com>
4839 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
4840 add_options_for_arm_vfp3.
4842 2014-02-11 Jeff Law <law@redhat.com>
4845 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
4846 object with an undesirable mode.
4848 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4851 * config/i386/sol2-9.h: New file.
4852 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
4853 *-*-solaris2.9*): Use it.
4855 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
4857 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
4858 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
4860 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
4862 * config/microblaze/microblaze.c: Extend mcpu version format
4864 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
4866 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
4868 2014-02-10 Richard Henderson <rth@redhat.com>
4871 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
4872 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
4873 ms-abi vs -mno-accumulate-outgoing-args.
4874 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
4875 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
4878 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
4881 * cfgexpand.c (expand_asm_operands): Attach source location to
4882 ASM_INPUT rtx objects.
4883 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
4885 2014-02-10 Nick Clifton <nickc@redhat.com>
4887 * config/mn10300/mn10300.c (popcount): New function.
4888 (mn10300_expand_prologue): Include saved registers in stack usage
4891 2014-02-10 Jeff Law <law@redhat.com>
4894 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
4895 when changing the SET_DEST of a prior insn to avoid an input reload.
4897 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4899 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
4900 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
4901 -mcall-openbsd, or -mcall-linux.
4902 (CC1_ENDIAN_BIG_SPEC): Remove.
4903 (CC1_ENDIAN_LITTLE_SPEC): Remove.
4904 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
4905 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
4906 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
4907 and %cc1_endian_default.
4908 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
4910 2014-02-10 Richard Biener <rguenther@suse.de>
4912 PR tree-optimization/60115
4913 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
4914 MEM_REF handling. Properly verify that the accesses are not
4915 out of the objects bound.
4917 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4919 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
4922 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
4924 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
4925 proper constants and fix formatting.
4926 (possible_polymorphic_call_targets): Fix formatting.
4928 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
4929 Ilya Tocar <ilya.tocar@intel.com>
4931 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
4932 (_mm512_loadu_epi32): Renamed into...
4933 (_mm512_loadu_si512): This.
4934 (_mm512_storeu_epi32): Renamed into...
4935 (_mm512_storeu_si512): This.
4936 (_mm512_maskz_ceil_ps): Removed.
4937 (_mm512_maskz_ceil_pd): Ditto.
4938 (_mm512_maskz_floor_ps): Ditto.
4939 (_mm512_maskz_floor_pd): Ditto.
4940 (_mm512_floor_round_ps): Ditto.
4941 (_mm512_floor_round_pd): Ditto.
4942 (_mm512_ceil_round_ps): Ditto.
4943 (_mm512_ceil_round_pd): Ditto.
4944 (_mm512_mask_floor_round_ps): Ditto.
4945 (_mm512_mask_floor_round_pd): Ditto.
4946 (_mm512_mask_ceil_round_ps): Ditto.
4947 (_mm512_mask_ceil_round_pd): Ditto.
4948 (_mm512_maskz_floor_round_ps): Ditto.
4949 (_mm512_maskz_floor_round_pd): Ditto.
4950 (_mm512_maskz_ceil_round_ps): Ditto.
4951 (_mm512_maskz_ceil_round_pd): Ditto.
4952 (_mm512_expand_pd): Ditto.
4953 (_mm512_expand_ps): Ditto.
4954 * config/i386/i386.c (ix86_builtins): Remove
4955 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
4956 (bdesc_args): Ditto.
4957 * config/i386/predicates.md (const1256_operand): New.
4958 (const_1_to_2_operand): Ditto.
4959 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4960 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4961 (*avx512pf_gatherpf<mode>sf): Ditto.
4962 (avx512pf_gatherpf<mode>df): Ditto.
4963 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4964 (*avx512pf_gatherpf<mode>df): Ditto.
4965 (avx512pf_scatterpf<mode>sf): Ditto.
4966 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4967 (*avx512pf_scatterpf<mode>sf): Ditto.
4968 (avx512pf_scatterpf<mode>df): Ditto.
4969 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4970 (*avx512pf_scatterpf<mode>df): Ditto.
4971 (avx512f_expand<mode>): Removed.
4972 (<shift_insn><mode>3<mask_name>): Change predicate type.
4974 2014-02-08 Jakub Jelinek <jakub@redhat.com>
4976 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
4977 not at the end of datarefs vector use ordered_remove to avoid
4978 reordering datarefs vector.
4981 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
4982 mark local addressable non-static vars as GOVD_PRIVATE
4983 instead of GOVD_LOCAL.
4984 * omp-low.c (lower_omp_for): Move gimple_bind_vars
4985 and BLOCK_VARS of gimple_bind_block to new_stmt rather
4989 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
4990 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
4991 assume_aligned or alloc_align attributes.
4992 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
4993 arguments. Handle also assume_aligned and alloc_align attributes.
4994 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
4995 calls to functions with assume_aligned or alloc_align attributes.
4996 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
4998 2014-02-08 Terry Guo <terry.guo@arm.com>
5000 * doc/invoke.texi: Document ARM -march=armv7e-m.
5002 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5004 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5005 flag on __cilkrts_rethrow builtin.
5008 * ipa-cp.c (determine_versionability): Fail at -O0
5009 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5010 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5013 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5016 * tree-inline.c (copy_forbidden): Fail for
5017 __attribute__((optimize (0))) functions.
5019 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5021 * varpool.c: Include pointer-set.h.
5022 (varpool_remove_unreferenced_decls): Variables in other partitions
5023 will not be output; be however careful to not lose information
5026 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5028 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5029 lookup in the vtable constructor.
5031 2014-02-07 Jeff Law <law@redhat.com>
5034 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5035 define_insn_and_split.
5037 * ipa-inline.c (inline_small_functions): Fix typos.
5039 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5041 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5042 (s390_can_use_return_insn): Declare.
5043 * config/s390/s390.h (EPILOGUE_USES): Define.
5044 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5046 (s390_chunkify_start): Handle return JUMP_LABELs.
5047 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5048 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5049 (s390_can_use_return_insn): New functions.
5050 (s390_fix_long_loop_prediction): Handle conditional returns.
5051 (TARGET_SET_UP_BY_PROLOGUE): Define.
5052 * config/s390/s390.md (ANY_RETURN): New code iterator.
5053 (*creturn, *csimple_return, return, simple_return): New patterns.
5055 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5057 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5058 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5059 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5060 REG_CFA_RESTORE list when deciding not to restore a register.
5062 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5064 * config/s390/s390.c: Include tree-pass.h and context.h.
5065 (s390_early_mach): New function, split out from...
5066 (s390_emit_prologue): ...here.
5067 (pass_data_s390_early_mach): New pass structure.
5068 (pass_s390_early_mach): New class.
5069 (s390_option_override): Create and register early_mach pass.
5070 Move to end of file.
5072 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5074 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5075 to match for the exit block.
5077 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5079 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5080 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5081 Reject misaligned operands.
5083 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5085 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5087 2014-02-07 Richard Biener <rguenther@suse.de>
5090 * gimple-low.c (lower_builtin_posix_memalign): New function.
5091 (lower_stmt): Call it to lower posix_memalign in a way
5092 to make alignment info accessible.
5094 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5097 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5098 __builtin_setjmp_receiver.
5100 2014-02-07 Richard Biener <rguenther@suse.de>
5103 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5104 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5105 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5106 Handle BUILT_IN_POSIX_MEMALIGN.
5107 (find_func_clobbers): Likewise.
5108 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5109 (call_may_clobber_ref_p_1): Likewise.
5111 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5114 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5117 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5120 * lto-cgraph.c (lto_output_node): Use
5121 symtab_get_symbol_partitioning_class.
5122 (lto_output_varpool_node): likewise.
5123 (symtab_get_symbol_partitioning_class): Move here from
5125 * cgraph.h (symbol_partitioning_class): Likewise.
5126 (symtab_get_symbol_partitioning_class): Declare.
5128 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5130 * ggc.h (ggc_internal_cleared_alloc): New macro.
5131 * vec.h (vec_safe_copy): Handle memory stats.
5132 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5133 * target-globals.c (save_target_globals): Likewise.
5135 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5138 * expr.c (emit_move_resolve_push): Export; be bit more selective
5139 on when to clear alias set.
5140 * expr.h (emit_move_resolve_push): Declare.
5141 * function.h (struct function): Add tail_call_marked.
5142 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5143 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5144 * config/i386/i386.md (TImode move expander): De not call
5146 (FP push expanders): Preserve memory attributes.
5147 * config/i386/sse.md (push<mode>1): Remove.
5148 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5149 (ix86_expand_push): Remove.
5150 * config/i386/mmx.md (push<mode>1): Remove.
5152 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5154 PR rtl-optimization/60030
5155 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5156 lopart with paradoxical subreg before shifting it up by hprec.
5158 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5160 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5161 Remove extra newline at end of file.
5162 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5163 (arm_issue_rate): Handle cortexa57.
5164 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5165 (cortex-a57.cortex-a53): Likewise.
5167 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5170 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5171 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5173 (arm_expand_prologue): Adjust all callers.
5174 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5175 info, registers also at the lowest numbered registers side. Use
5176 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5180 * var-tracking.c (adjust_mems): Before adding a SET to
5181 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5183 2014-02-06 Alan Modra <amodra@gmail.com>
5186 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5187 change SDmode to DDmode when lra_in_progress.
5189 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5192 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5193 free_data_ref on the dr first, and before goto again also set dr
5194 to the next dr. For simd_lane_access, free old datarefs[i] before
5195 overwriting it. For get_vectype_for_scalar_type failure, don't
5196 free_data_ref if simd_lane_access.
5198 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5201 * tree.h (opts_for_fn): New inline function.
5202 (opt_for_fn): Define.
5203 * config/i386/i386.c (ix86_function_regparm): Use
5204 opt_for_fn (decl, optimize) instead of optimize.
5206 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5208 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5209 for SYMBOL_REF in large memory model.
5211 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5213 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5215 (cortex-a57): Likewise.
5216 (cortex-a57.cortex-a53): Likewise.
5218 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5219 Kugan Vivekanandarajah <kuganv@linaro.org>
5221 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5223 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5225 2014-02-06 Richard Biener <rguenther@suse.de>
5227 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5228 set_loop_copy and initialize_original_copy_tables.
5230 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5232 * config/aarch64/aarch64-simd.md
5233 (aarch64_ashr_simddi): Change QI to SI.
5235 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5236 Jakub Jelinek <jakub@redhat.com>
5239 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5242 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5244 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5245 CODE_FOR_altivec_vpku[hw]um to
5246 CODE_FOR_altivec_vpku[hw]um_direct.
5247 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5248 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5249 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5250 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5252 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5254 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5255 generation for -maltivec=be.
5256 (altivec_vsumsws): Simplify redundant test.
5258 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5260 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5261 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5262 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5263 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5264 gen_altivec_vpkuwum.
5265 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5267 (altivec_vpks<VI_char>ss): Likewise.
5268 (altivec_vpks<VI_char>us): Likewise.
5269 (altivec_vpku<VI_char>us): Likewise.
5270 (altivec_vpku<VI_char>um): Likewise.
5271 (altivec_vpku<VI_char>um_direct): New (copy of
5272 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5274 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5275 target is little endian and -maltivec=be is not specified.
5276 (*altivec_vupkhs<VU_char>_direct): New (copy of
5277 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5278 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5279 target is little endian and -maltivec=be is not specified.
5280 (*altivec_vupkls<VU_char>_direct): New (copy of
5281 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5282 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5283 little endian and -maltivec=be is not specified.
5284 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5285 little endian and -maltivec=be is not specified.
5287 2014-02-05 Richard Henderson <rth@redhat.com>
5290 * combine-stack-adj.c: Revert r206943.
5291 * sched-int.h (struct deps_desc): Add last_args_size.
5292 * sched-deps.c (init_deps): Initialize it.
5293 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5294 contain REG_ARGS_SIZE notes.
5296 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5298 * lto-cgraph.c (asm_nodes_output): Make global.
5299 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5300 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5301 (driver_handle_option): Handle OPT_fwpa.
5303 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5306 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5307 a comment typo and formatting issue. If odr_hash hasn't been
5308 created, return vNULL and set *completep to false.
5311 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5312 bb with no successors.
5314 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5317 * doc/invoke.texi (-march): Clarify documentation for ARM.
5321 2014-02-05 Richard Biener <rguenther@suse.de>
5323 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5324 when not vectorizing because of too many alias checks.
5325 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5326 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5328 2014-02-05 Nick Clifton <nickc@redhat.com>
5330 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5331 accept extended registers in any mode when compiling for the MN10300.
5333 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5335 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5336 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5337 sanitization attributes.
5338 (can_inline_edge_p): Likewise.
5339 (sanitize_attrs_match_for_inline_p): New function.
5341 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5343 * ipa-prop.c (detect_type_change): Shor circuit testing of
5344 type changes on THIS pointer.
5346 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5349 * config/pa/pa.c (legitimize_tls_address): Return original address
5350 if not passed a SYMBOL_REF rtx.
5351 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5353 (pa_emit_move_sequence): Simplify TLS source operands.
5354 (pa_legitimate_constant_p): Reject all TLS constants.
5355 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5356 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5358 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5360 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5361 groups when we know they are controlled by LTO.
5362 * varasm.c (default_binds_local_p_1): If object is in other partition,
5363 it will be resolved locally.
5365 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5367 * config/host-linux.c (linux_gt_pch_use_address): Don't
5368 use SSIZE_MAX because it is not always defined.
5370 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5373 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5374 threshold for pseudo splitting.
5375 (update_ebb_live_info): Process call argument hard registers and
5376 hard registers from insn definition too.
5377 (max_small_class_regs_num): New constant.
5378 (inherit_in_ebb): Update live hard regs through EBBs. Update
5379 reloads_num only for small register classes. Don't split for
5382 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5385 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5388 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5390 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5393 2014-02-04 Marek Polacek <polacek@redhat.com>
5395 * gdbinit.in (pel): Define.
5397 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5399 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5402 2014-02-04 Richard Biener <rguenther@suse.de>
5405 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5406 in function context local.
5407 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5408 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5409 similar to LTO_imported_decl_ref.
5411 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5413 PR tree-optimization/60002
5414 * cgraphclones.c (build_function_decl_skip_args): Clear
5417 PR tree-optimization/60023
5418 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5419 false to gsi_replace.
5420 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5421 has been in some EH region and vec_stmt could throw, add
5422 vec_stmt into the same EH region.
5423 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5424 has no lhs, ignore it.
5425 * internal-fn.c (expand_MASK_LOAD): Likewise.
5428 * tree-inline.c (copy_forbidden): Fail for
5429 __attribute__((optimize (0))) functions.
5432 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5433 is set, copy one less argument.
5434 (expand_simd_clones): Don't subtract clone_info->inbranch
5435 from simd_clone_struct_alloc argument.
5437 PR rtl-optimization/57915
5438 * recog.c (simplify_while_replacing): If all unary/binary/relational
5439 operation arguments are constant, attempt to simplify those.
5442 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5443 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5445 2014-02-04 Richard Biener <rguenther@suse.de>
5447 PR tree-optimization/60012
5448 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5449 TBAA disambiguation to all DDRs.
5451 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5454 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5455 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5457 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5460 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5462 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5464 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5465 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5467 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5470 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
5471 to figure out targets of polymorphic calls with known decl.
5472 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5473 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
5474 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
5475 (get_polymorphic_call_info): ... here.
5476 (get_polymorphic_call_info_from_invariant): New function.
5478 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5480 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
5481 lookup via vtable pointer; check for type consistency
5482 and turn inconsitent facts into UNREACHABLE.
5483 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5484 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
5485 type inconsistent querries; return UNREACHABLE instead.
5487 2014-02-03 Richard Henderson <rth@twiddle.net>
5490 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
5491 already processed this node.
5492 (normalize_one_pred_1): Pass along mark_set.
5493 (normalize_one_pred): Create and destroy a pointer_set_t.
5494 (normalize_one_pred_chain): Likewise.
5496 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
5498 PR gcov-profile/58602
5499 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
5501 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5504 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
5505 -fno-devirtualize; try to devirtualize by the knowledge of
5506 virtual table pointer given by aggregate propagation.
5507 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5508 (ipa_print_node_jump_functions): Dump also offset that
5509 is relevant for polymorphic calls.
5510 (determine_known_aggregate_parts): Add arg_type parameter; use it
5511 instead of determining the type from pointer type.
5512 (ipa_compute_jump_functions_for_edge): Update call of
5513 determine_known_aggregate_parts.
5514 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
5515 (gimple_get_virt_method_for_binfo): ... here; simplify using
5516 vtable_pointer_value_to_vtable.
5517 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
5518 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
5519 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
5520 (vtable_pointer_value_to_vtable): Break out from ...; handle also
5522 (vtable_pointer_value_to_binfo): ... here.
5523 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
5525 2014-02-03 Teresa Johnson <tejohnson@google.com>
5527 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
5528 redef of outer loop index variable.
5530 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
5534 * doc/extend.texi (Function Attributes): Typo.
5536 2014-02-03 Cong Hou <congh@google.com>
5538 PR tree-optimization/60000
5539 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
5540 if the vectorized statement is a store. A store statement can only
5541 appear at the end of pattern statements.
5543 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5545 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
5546 (ix86_option_override_internal): Default long double to 64-bit for
5547 32-bit Bionic and to 128-bit for 64-bit Bionic.
5549 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
5550 TARGET_LONG_DOUBLE_128 is true.
5551 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
5553 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
5554 (mlong-double-64): Negate -mlong-double-128.
5555 (mlong-double-128): New option.
5557 * config/i386/i386-c.c (ix86_target_macros): Define
5558 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
5560 * doc/invoke.texi: Document -mlong-double-128.
5562 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5564 PR rtl-optimization/60024
5565 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
5567 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
5569 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
5571 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5573 PR rtl-optimization/57662
5574 * sel-sched.c (code_motion_path_driver): Do not mark already not
5575 existing blocks in the visiting bitmap.
5577 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5579 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
5580 on the insn being emitted.
5582 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
5583 Will Deacon <will.deacon@arm.com>
5585 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
5587 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5589 * config/arm/arm-tables.opt: Regenerate.
5591 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5593 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
5594 for vector types other than V16QImode.
5595 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
5596 define_expand, and call altivec_expand_vec_perm_le when producing
5597 code with little endian element order.
5598 (*altivec_vperm_<mode>_internal): New insn having previous
5599 behavior of altivec_vperm_<mode>.
5600 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
5601 altivec_expand_vec_perm_le when producing code with little endian
5603 (*altivec_vperm_<mode>_uns_internal): New insn having previous
5604 behavior of altivec_vperm_<mode>_uns.
5606 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5608 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
5609 (altivec_vsumsws): Add handling for -maltivec=be with a little
5611 (altivec_vsumsws_direct): New.
5612 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
5613 gen_altivec_vsumsws.
5615 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5617 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
5618 vtable_pointer_value_to_binfo): New functions.
5619 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
5620 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
5622 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
5624 * config/nios2/nios2.md (load_got_register): Initialize GOT
5625 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
5626 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
5628 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5630 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
5631 preserverd by passthrough, do not propagate the type.
5633 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5635 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
5636 (mips_atomic_assign_expand_fenv): New function.
5637 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
5639 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5641 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
5642 (__builtin_mips_set_fcsr): Likewise.
5643 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
5644 MIPS_USI_FTYPE_VOID.
5645 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
5646 (mips16_expand_set_fcsr): Likewise.
5647 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
5648 (mips16_set_fcsr_stub): Likewise.
5649 (mips16_get_fcsr_one_only_stub): New class.
5650 (mips16_set_fcsr_one_only_stub): Likewise.
5651 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
5652 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
5653 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
5654 (hard_float): New availability predicate.
5655 (mips_builtins): Add get_fcsr and set_fcsr.
5656 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
5657 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
5658 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
5659 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
5660 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
5663 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5665 * config/mips/mips.c (mips_one_only_stub): New class.
5666 (mips_need_mips16_rdhwr_p): Replace with...
5667 (mips16_rdhwr_stub): ...this new variable.
5668 (mips16_stub_call_address): New function.
5669 (mips16_rdhwr_one_only_stub): New class.
5670 (mips_expand_thread_pointer): Use mips16_stub_call_address.
5671 (mips_output_mips16_rdhwr): Delete.
5672 (mips_finish_stub): New function.
5673 (mips_code_end): Use it to handle rdhwr stubs.
5675 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
5678 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
5679 when calculating size of integer atomic types.
5681 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
5683 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
5685 2014-02-01 Jakub Jelinek <jakub@redhat.com>
5687 PR tree-optimization/60003
5688 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
5689 * profile.c (branch_prob): Use gimple_call_builtin_p
5690 to check for BUILT_IN_SETJMP_RECEIVER.
5691 * tree-inline.c (copy_bb): Call notice_special_calls.
5693 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
5696 * lra-constraints.c (process_alt_operands): Update reload_sum only
5699 2014-01-31 Richard Henderson <rth@redhat.com>
5702 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
5703 until after else_eh is processed.
5705 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
5707 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
5708 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
5709 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
5710 in smmintrin.h, remove them.
5711 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
5712 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
5713 * config/i386/i386.md (ROUND_SAE): Fix value.
5714 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
5715 (const48_operand): New.
5716 * config/i386/subst.md (round), (round_expand): Use
5717 const_4_or_8_to_11_operand.
5718 (round_saeonly), (round_saeonly_expand): Use const48_operand.
5720 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
5722 * config/i386/constraints.md (Yk): Swap meaning with k.
5723 * config/i386/i386.md (movhi_internal): Change Yk to k.
5724 (movqi_internal): Ditto.
5725 (*k<logic><mode>): Ditto.
5728 (kandn<mode>): Ditto.
5729 (*<code>hi_1): Ditto.
5730 (*<code>qi_1): Ditto.
5731 (kxnor<mode>): Ditto.
5732 (kortestzhi): Ditto.
5733 (kortestchi): Ditto.
5735 (*one_cmplhi2_1): Ditto.
5736 (*one_cmplqi2_1): Ditto.
5737 * config/i386/sse.md (): Change k to Yk.
5738 (avx512f_load<mode>_mask): Ditto.
5739 (avx512f_blendm<mode>): Ditto.
5740 (avx512f_store<mode>_mask): Ditto.
5741 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
5742 (avx512f_storedqu<mode>_mask): Ditto.
5743 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
5745 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
5746 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
5747 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
5748 (avx512f_maskcmp<mode>3): Ditto.
5749 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
5750 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
5751 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
5752 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
5753 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
5754 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
5755 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
5756 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
5757 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
5758 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
5759 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
5760 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
5761 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
5762 (vec_extract_lo_<mode>_maskm): Ditto.
5763 (vec_extract_hi_<mode>_maskm): Ditto.
5764 (avx512f_vternlog<mode>_mask): Ditto.
5765 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
5766 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
5767 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
5768 (avx512f_<code>v8div16qi2_mask): Ditto.
5769 (avx512f_<code>v8div16qi2_mask_store): Ditto.
5770 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
5771 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
5772 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
5773 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
5774 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5775 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5776 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5777 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5778 (avx512cd_maskb_vec_dupv8di): Ditto.
5779 (avx512cd_maskw_vec_dupv16si): Ditto.
5780 (avx512f_vpermi2var<mode>3_maskz): Ditto.
5781 (avx512f_vpermi2var<mode>3_mask): Ditto.
5782 (avx512f_vpermi2var<mode>3_mask): Ditto.
5783 (avx512f_vpermt2var<mode>3_maskz): Ditto.
5784 (*avx512f_gathersi<mode>): Ditto.
5785 (*avx512f_gathersi<mode>_2): Ditto.
5786 (*avx512f_gatherdi<mode>): Ditto.
5787 (*avx512f_gatherdi<mode>_2): Ditto.
5788 (*avx512f_scattersi<mode>): Ditto.
5789 (*avx512f_scatterdi<mode>): Ditto.
5790 (avx512f_compress<mode>_mask): Ditto.
5791 (avx512f_compressstore<mode>_mask): Ditto.
5792 (avx512f_expand<mode>_mask): Ditto.
5793 * config/i386/subst.md (mask): Change k to Yk.
5794 (mask_scalar_merge): Ditto.
5797 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
5799 * doc/extend.texi (Vector Extensions): Document ?: in C++.
5801 2014-01-31 Richard Biener <rguenther@suse.de>
5804 * builtins.c (fold_builtin_memory_op): Make sure to not
5805 use a floating-point mode or a boolean or enumeral type for
5808 2014-01-30 DJ Delorie <dj@redhat.com>
5810 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
5811 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
5812 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
5813 whenever main() has an epilogue.
5815 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5817 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
5818 unused variable "field".
5819 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
5820 (vsx_mergeh_<mode>): Likewise.
5821 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
5822 (altivec_vmrghh): Likewise.
5823 (altivec_vmrghw): Likewise.
5824 (altivec_vmrglb): Likewise.
5825 (altivec_vmrglh): Likewise.
5826 (altivec_vmrglw): Likewise.
5827 (altivec_vspltb): Add missing uses.
5828 (altivec_vsplth): Likewise.
5829 (altivec_vspltw): Likewise.
5830 (altivec_vspltsf): Likewise.
5832 2014-01-30 Jakub Jelinek <jakub@redhat.com>
5835 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
5836 frame related instructions.
5838 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
5840 PR rtl-optimization/59959
5841 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
5842 any reload of register whose subreg is invalid.
5844 2014-01-30 Jakub Jelinek <jakub@redhat.com>
5846 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
5847 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
5848 Add missing return type - void.
5850 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5852 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
5853 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
5854 remove element index adjustment for endian (now handled in vsx.md
5856 (altivec_expand_vec_perm_const): Use
5857 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
5858 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
5859 (vsx_xxspltw_<mode>): Adjust element index for little endian.
5860 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
5861 define_expand and a new define_insn *altivec_vspltb_internal;
5862 adjust for -maltivec=be on a little endian target.
5863 (altivec_vspltb_direct): New.
5864 (altivec_vsplth): Divide into a define_expand and a new
5865 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
5866 little endian target.
5867 (altivec_vsplth_direct): New.
5868 (altivec_vspltw): Divide into a define_expand and a new
5869 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
5870 little endian target.
5871 (altivec_vspltw_direct): New.
5872 (altivec_vspltsf): Divide into a define_expand and a new
5873 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
5874 a little endian target.
5876 2014-01-30 Richard Biener <rguenther@suse.de>
5878 PR tree-optimization/59993
5879 * tree-ssa-forwprop.c (associate_pointerplus): Check we
5880 can propagate form the earlier stmt and avoid the transform
5881 when the intermediate result is needed.
5883 2014-01-30 Alangi Derick <alangiderick@gmail.com>
5885 * README.Portability: Fix typo.
5887 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
5889 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
5890 comparison_operator with ordered_comparison_operator.
5892 2014-01-30 Nick Clifton <nickc@redhat.com>
5894 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
5895 Rename to mn10300_store_multiple_regs.
5896 * config/mn10300/mn10300.c: Likewise.
5897 * config/mn10300/mn10300.md (store_movm): Fix typo: call
5898 store_multiple_regs.
5899 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
5900 Call mn10300_store_multiple_regs.
5902 2014-01-30 Nick Clifton <nickc@redhat.com>
5903 DJ Delorie <dj@redhat.com>
5905 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
5906 %fp 2 to keep registers after it properly word-aligned.
5907 (rl78_alloc_physical_registers_umul): Handle the case where both
5908 input operands are the same.
5910 2014-01-30 Richard Biener <rguenther@suse.de>
5912 PR tree-optimization/59903
5913 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
5916 2014-01-30 Jason Merrill <jason@redhat.com>
5919 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
5922 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
5924 2014-01-30 Richard Biener <rguenther@suse.de>
5926 PR tree-optimization/59951
5927 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
5929 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
5932 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
5933 SFmode to DFmode case.
5935 2014-01-29 DJ Delorie <dj@redhat.com>
5937 * config/msp430/msp430.opt (-minrt): New.
5938 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
5940 (ENDFILE_SPEC): Likewise.
5942 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
5944 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
5945 (estimate_function_body_sizes): Use it.
5947 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
5950 * dwarf2out.c (is_cxx_auto): New.
5951 (is_base_type): Use it.
5952 (gen_type_die_with_usage): Likewise.
5954 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5956 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
5957 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
5958 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
5959 -maltivec=be with LE targets.
5960 (vsx_mergeh_<mode>): Likewise.
5961 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
5962 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
5963 (altivec_vmrghb): Replace with define_expand and new
5964 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
5965 (altivec_vmrghb_direct): New define_insn.
5966 (altivec_vmrghh): Replace with define_expand and new
5967 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
5968 (altivec_vmrghh_direct): New define_insn.
5969 (altivec_vmrghw): Replace with define_expand and new
5970 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
5971 (altivec_vmrghw_direct): New define_insn.
5972 (*altivec_vmrghsf): Adjust for endianness.
5973 (altivec_vmrglb): Replace with define_expand and new
5974 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
5975 (altivec_vmrglb_direct): New define_insn.
5976 (altivec_vmrglh): Replace with define_expand and new
5977 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
5978 (altivec_vmrglh_direct): New define_insn.
5979 (altivec_vmrglw): Replace with define_expand and new
5980 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
5981 (altivec_vmrglw_direct): New define_insn.
5982 (*altivec_vmrglsf): Adjust for endianness.
5983 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
5984 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
5985 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
5986 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
5987 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
5988 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
5989 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
5990 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
5992 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
5994 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
5995 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
5998 2014-01-29 Richard Biener <rguenther@suse.de>
6000 PR tree-optimization/58742
6001 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6002 associate_pointerplus_align.
6003 (associate_pointerplus_diff): New function.
6004 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6005 and associate_pointerplus_diff.
6007 2014-01-29 Richard Biener <rguenther@suse.de>
6009 * lto-streamer.h (LTO_major_version): Bump to 3.
6010 (LTO_minor_version): Reset to 0.
6012 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6014 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6015 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6016 (arm_file_start): Generate correct asm header for armv7ve.
6017 * config/arm/bpabi.h: Add multilib support for armv7ve.
6018 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6019 and cortex-a15 to armv7ve.
6020 * config/arm/t-aprofile: Add multilib support for armv7ve.
6021 * doc/invoke.texi: Document -march=armv7ve.
6023 2014-01-29 Richard Biener <rguenther@suse.de>
6025 PR tree-optimization/58742
6026 * tree-ssa-forwprop.c (associate_plusminus): Return true
6027 if we changed sth, defer EH cleanup to ...
6028 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6029 (simplify_mult): New function.
6031 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6034 PR tree-optimization/59920
6035 * tree.c (build_common_builtin_nodes): Remove
6036 __builtin_setjmp_dispatcher initialization.
6037 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6038 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6039 instead of gsi_after_labels + manually skipping debug stmts.
6040 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6041 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6042 * tree-inline.c (copy_edges_for_bb): Remove
6043 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6044 argument. Ignore computed_goto_p stmts. Don't call
6045 make_abnormal_goto_edges. If a call might need abnormal edges
6046 for non-local gotos, see if it already has an edge to
6047 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6048 with true argument, don't do anything then, otherwise add
6049 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6050 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6052 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6053 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6054 (lower_stmt): Don't set data->calls_builtin_setjmp.
6055 (lower_builtin_setjmp): Adjust comment.
6056 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6057 * tree-cfg.c (found_computed_goto): Remove.
6058 (factor_computed_gotos): Remove.
6059 (make_goto_expr_edges): Return bool, true for computed gotos.
6060 Don't call make_abnormal_goto_edges.
6061 (build_gimple_cfg): Don't set found_computed_goto, don't call
6062 factor_computed_gotos.
6063 (computed_goto_p): No longer static.
6064 (make_blocks): Don't set found_computed_goto.
6065 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6066 (make_edges): If make_goto_expr_edges returns true, push bb
6067 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6068 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6069 vector. Record mapping between bbs and OpenMP regions if there
6070 are any, adjust make_gimple_omp_edges caller. Call
6071 handle_abnormal_edges.
6072 (make_abnormal_goto_edges): Remove.
6073 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6074 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6075 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6076 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6077 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6078 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6079 filling *region also set *region_idx to (*region)->entry->index.
6082 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6083 For REGs set ORIGINAL_REGNO.
6085 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6087 * doc/md.texi: Mention that a target shouldn't implement
6088 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6091 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6093 PR tree-optimization/59594
6094 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6095 a copy of the datarefs vector rather than the vector itself.
6097 2014-01-28 Jason Merrill <jason@redhat.com>
6100 * dwarf2out.c (auto_die): New static.
6101 (gen_type_die_with_usage): Handle C++1y 'auto'.
6102 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6105 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6108 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6109 (SPEC_X32): Likewise.
6110 (SPEC_64): Likewise.
6111 * config/i386/i386.c (ix86_option_override_internal): Turn off
6112 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6114 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6115 * config/i386/i386.h (TARGET_16BIT): New macro.
6116 (TARGET_16BIT_P): Likewise.
6117 * config/i386/i386.opt: Add m16.
6118 * doc/invoke.texi: Document -m16.
6120 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6122 PR preprocessor/59935
6123 * input.c (location_get_source_line): Bail out on when line number
6124 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6126 2014-01-28 Richard Biener <rguenther@suse.de>
6128 PR tree-optimization/58742
6129 * tree-ssa-forwprop.c (associate_plusminus): Handle
6130 pointer subtraction of the form (T)(P + A) - (T)P.
6132 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6134 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6137 2014-01-28 Richard Biener <rguenther@suse.de>
6140 2014-01-28 Richard Biener <rguenther@suse.de>
6142 PR rtl-optimization/45364
6143 PR rtl-optimization/59890
6144 * var-tracking.c (local_get_addr_clear_given_value): Handle
6145 already cleared slot.
6146 (val_reset): Handle not allocated local_get_addr_cache.
6147 (vt_find_locations): Use post-order on the inverted CFG.
6149 2014-01-28 Richard Biener <rguenther@suse.de>
6151 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6153 2014-01-28 Richard Biener <rguenther@suse.de>
6155 PR rtl-optimization/45364
6156 PR rtl-optimization/59890
6157 * var-tracking.c (local_get_addr_clear_given_value): Handle
6158 already cleared slot.
6159 (val_reset): Handle not allocated local_get_addr_cache.
6160 (vt_find_locations): Use post-order on the inverted CFG.
6162 2014-01-28 Alan Modra <amodra@gmail.com>
6164 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6165 * configure.ac <recursive call for build != host>: Define
6166 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6167 and LD_FOR_BUILD too.
6168 * configure: Regenerate.
6170 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6172 * config/i386/i386.c (get_builtin_code_for_version): Separate
6173 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6174 Broadwell from Haswell.
6176 2014-01-27 Steve Ellcey <sellcey@mips.com>
6178 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6179 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6180 * config/mips/mips.c (mips_option_override): Change setting
6182 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6183 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6184 Change from Mask to Var.
6186 2014-01-27 Jeff Law <law@redhat.com>
6188 * ipa-inline.c (inline_small_functions): Fix typo.
6190 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6192 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6193 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6194 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6195 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6196 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6197 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6198 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6199 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6200 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6201 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6202 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6203 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6204 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6205 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6206 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6207 (_mm512_storeu_epi64): Ditto.
6208 (_mm512_cmpge_epi32_mask): Ditto.
6209 (_mm512_cmpge_epu32_mask): Ditto.
6210 (_mm512_cmpge_epi64_mask): Ditto.
6211 (_mm512_cmpge_epu64_mask): Ditto.
6212 (_mm512_cmple_epi32_mask): Ditto.
6213 (_mm512_cmple_epu32_mask): Ditto.
6214 (_mm512_cmple_epi64_mask): Ditto.
6215 (_mm512_cmple_epu64_mask): Ditto.
6216 (_mm512_cmplt_epi32_mask): Ditto.
6217 (_mm512_cmplt_epu32_mask): Ditto.
6218 (_mm512_cmplt_epi64_mask): Ditto.
6219 (_mm512_cmplt_epu64_mask): Ditto.
6220 (_mm512_cmpneq_epi32_mask): Ditto.
6221 (_mm512_cmpneq_epu32_mask): Ditto.
6222 (_mm512_cmpneq_epi64_mask): Ditto.
6223 (_mm512_cmpneq_epu64_mask): Ditto.
6224 (_mm512_expand_pd): Ditto.
6225 (_mm512_expand_ps): Ditto.
6226 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6227 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6228 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6229 * config/i386/i386.c (ix86_builtins): Add
6230 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6231 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6232 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6233 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6234 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6235 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6236 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6237 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6238 IX86_BUILTIN_PMOVUSQW512_MEM.
6239 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6240 __builtin_ia32_pmovsqd512mem_mask,
6241 __builtin_ia32_pmovqd512mem_mask,
6242 __builtin_ia32_pmovusqw512mem_mask,
6243 __builtin_ia32_pmovsqw512mem_mask,
6244 __builtin_ia32_pmovqw512mem_mask,
6245 __builtin_ia32_pmovusdw512mem_mask,
6246 __builtin_ia32_pmovsdw512mem_mask,
6247 __builtin_ia32_pmovdw512mem_mask,
6248 __builtin_ia32_pmovqb512mem_mask,
6249 __builtin_ia32_pmovusqb512mem_mask,
6250 __builtin_ia32_pmovsqb512mem_mask,
6251 __builtin_ia32_pmovusdb512mem_mask,
6252 __builtin_ia32_pmovsdb512mem_mask,
6253 __builtin_ia32_pmovdb512mem_mask.
6254 (bdesc_args): Add __builtin_ia32_expanddf512,
6255 __builtin_ia32_expandsf512.
6256 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6257 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6258 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6259 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6260 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6261 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6262 (avx512f_<code>v8div16qi2_mask_store): This.
6263 (avx512f_expand<mode>): New.
6265 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6267 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6269 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6270 (_mm512_prefetch_i32scatter_pd): Ditto.
6271 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6272 (_mm512_prefetch_i64scatter_pd): Ditto.
6273 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6274 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6275 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6276 (_mm512_prefetch_i32scatter_ps): Ditto.
6277 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6278 (_mm512_prefetch_i64scatter_ps): Ditto.
6279 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6280 * config/i386/i386-builtin-types.def: Define
6281 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6282 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6283 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6284 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6285 IX86_BUILTIN_SCATTERPFQPD.
6286 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6287 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6288 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6289 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6290 __builtin_ia32_scatterpfqps.
6291 (ix86_expand_builtin): Expand new built-ins.
6292 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6293 fix memory access data type.
6294 (*avx512pf_gatherpf<mode>_mask): Ditto.
6295 (*avx512pf_gatherpf<mode>): Ditto.
6296 (avx512pf_scatterpf<mode>): Ditto.
6297 (*avx512pf_scatterpf<mode>_mask): Ditto.
6298 (*avx512pf_scatterpf<mode>): Ditto.
6299 (GATHER_SCATTER_SF_MEM_MODE): New.
6300 (avx512pf_gatherpf<mode>df): Ditto.
6301 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6302 (*avx512pf_scatterpf<mode>df): Ditto.
6304 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6307 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6308 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6311 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6313 * common/config/arm/arm-common.c
6314 (arm_rewrite_mcpu): Handle multiple names.
6316 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6318 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6320 * gimple-builder.h (create_gimple_tmp): Delete.
6322 2014-01-27 Christian Bruel <christian.bruel@st.com>
6324 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6327 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6329 * config/pa/pa.md (call): Generate indirect long calls to non-local
6330 functions when outputing 32-bit code.
6331 (call_value): Likewise except for special call to buggy powf function.
6333 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6334 portable runtime and PIC indirect calls.
6335 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6336 and PIC call sequences. Use ldo instead of blr to set return register
6337 in PIC call sequence.
6339 2014-01-25 Walter Lee <walt@tilera.com>
6341 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6342 avoid clobbering a live register.
6344 2014-01-25 Walter Lee <walt@tilera.com>
6346 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6347 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6348 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6349 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6351 2014-01-25 Walter Lee <walt@tilera.com>
6353 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6354 arguments on even registers.
6355 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6357 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6358 (BIGGEST_ALIGNMENT): Ditto.
6359 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6361 2014-01-25 Walter Lee <walt@tilera.com>
6363 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6364 insns before bundling.
6365 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6367 2014-01-25 Walter Lee <walt@tilera.com>
6369 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6370 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6371 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6373 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6375 * config/mips/constraints.md (kl): Delete.
6376 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6377 define expands, using...
6378 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6379 instructions for MIPS16.
6380 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6381 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6383 2014-01-25 Walter Lee <walt@tilera.com>
6385 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6389 2014-01-25 Walter Lee <walt@tilera.com>
6391 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6392 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6394 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6396 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6399 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6401 * print-rtl.c (in_call_function_usage): New var.
6402 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6403 EXPR_LIST mode as mode and not as reg note name.
6406 * cfgloopmanip.c (copy_loop_info): If
6407 loop->warned_aggressive_loop_optimizations, make sure
6408 the flag is set in target loop too.
6410 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6412 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6414 * builtins.def: Likewise.
6415 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6416 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6417 * ira.c (ira_setup_eliminable_regset): Likewise.
6418 * omp-low.c (gate_expand_omp): Likewise.
6419 (execute_lower_omp): Likewise.
6420 (diagnose_sb_0): Likewise.
6421 (gate_diagnose_omp_blocks): Likewise.
6422 (simd_clone_clauses_extract): Likewise.
6425 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6427 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6428 correction for little endian...
6429 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6432 2014-01-24 Jeff Law <law@redhat.com>
6434 PR tree-optimization/59919
6435 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6436 for non-returning calls.
6438 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6440 * common/config/aarch64/aarch64-common.c
6441 (aarch64_rewrite_mcpu): Handle multiple names.
6442 * config/aarch64/aarch64.h
6443 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6445 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6447 * input.c (add_file_to_cache_tab): Handle the case where fopen
6450 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6453 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6454 from push operand if code of push isn't PRE_DEC.
6456 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6459 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6460 -mquad-memory-atomic. Update -mquad-memory documentation to say
6461 it is only used for non-atomic loads/stores.
6463 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6464 -mquad-memory or -mquad-memory-atomic switches.
6466 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6467 -mquad-memory-atomic to ISA 2.07 support.
6469 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
6470 to separate support of normal quad word memory operations (ldq, stq)
6471 from the atomic quad word memory operations.
6473 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
6474 support to separate non-atomic quad word operations from atomic
6475 quad word operations. Disable non-atomic quad word operations in
6476 little endian mode so that we don't have to swap words after the
6477 load and before the store.
6478 (quad_load_store_p): Add comment about atomic quad word support.
6479 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
6480 options printed with -mdebug=reg.
6482 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
6483 -mquad-memory-atomic as the test for whether we have quad word
6484 atomic instructions.
6485 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
6486 or -mp8-vector are used, allow byte/half-word atomic operations.
6488 * config/rs6000/sync.md (load_lockedti): Insure that the address
6489 is a proper indexed or indirect address for the lqarx instruction.
6490 On little endian systems, swap the hi/lo registers after the lqarx
6492 (load_lockedpti): Use indexed_or_indirect_operand predicate to
6493 insure the address is valid for the lqarx instruction.
6494 (store_conditionalti): Insure that the address is a proper indexed
6495 or indirect address for the stqcrx. instruction. On little endian
6496 systems, swap the hi/lo registers before doing the stqcrx.
6498 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
6499 insure the address is valid for the stqcrx. instruction.
6501 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6502 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
6503 type of quad memory support is available.
6505 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
6508 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
6509 there is a danger of looping.
6511 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
6513 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6514 force flag_ira_loop_pressure if set via command line.
6516 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6518 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
6519 (ashr_simd): New builtin handling DI mode.
6520 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
6521 (aarch64_sshr_simddi): New match pattern.
6522 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
6523 (vshrd_n_s64): Likewise.
6524 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
6526 2014-01-23 Nick Clifton <nickc@redhat.com>
6528 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
6529 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
6530 favour of mcu specific scripts.
6531 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
6534 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6535 Alex Velenko <Alex.Velenko@arm.com>
6537 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
6538 (vaddv_s16): Likewise.
6539 (vaddv_s32): Likewise.
6540 (vaddv_u8): Likewise.
6541 (vaddv_u16): Likewise.
6542 (vaddv_u32): Likewise.
6543 (vaddvq_s8): Likewise.
6544 (vaddvq_s16): Likewise.
6545 (vaddvq_s32): Likewise.
6546 (vaddvq_s64): Likewise.
6547 (vaddvq_u8): Likewise.
6548 (vaddvq_u16): Likewise.
6549 (vaddvq_u32): Likewise.
6550 (vaddvq_u64): Likewise.
6551 (vaddv_f32): Likewise.
6552 (vaddvq_f32): Likewise.
6553 (vaddvq_f64): Likewise.
6554 (vmaxv_f32): Likewise.
6555 (vmaxv_s8): Likewise.
6556 (vmaxv_s16): Likewise.
6557 (vmaxv_s32): Likewise.
6558 (vmaxv_u8): Likewise.
6559 (vmaxv_u16): Likewise.
6560 (vmaxv_u32): Likewise.
6561 (vmaxvq_f32): Likewise.
6562 (vmaxvq_f64): Likewise.
6563 (vmaxvq_s8): Likewise.
6564 (vmaxvq_s16): Likewise.
6565 (vmaxvq_s32): Likewise.
6566 (vmaxvq_u8): Likewise.
6567 (vmaxvq_u16): Likewise.
6568 (vmaxvq_u32): Likewise.
6569 (vmaxnmv_f32): Likewise.
6570 (vmaxnmvq_f32): Likewise.
6571 (vmaxnmvq_f64): Likewise.
6572 (vminv_f32): Likewise.
6573 (vminv_s8): Likewise.
6574 (vminv_s16): Likewise.
6575 (vminv_s32): Likewise.
6576 (vminv_u8): Likewise.
6577 (vminv_u16): Likewise.
6578 (vminv_u32): Likewise.
6579 (vminvq_f32): Likewise.
6580 (vminvq_f64): Likewise.
6581 (vminvq_s8): Likewise.
6582 (vminvq_s16): Likewise.
6583 (vminvq_s32): Likewise.
6584 (vminvq_u8): Likewise.
6585 (vminvq_u16): Likewise.
6586 (vminvq_u32): Likewise.
6587 (vminnmv_f32): Likewise.
6588 (vminnmvq_f32): Likewise.
6589 (vminnmvq_f64): Likewise.
6591 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6593 * config/aarch64/aarch64-simd.md
6594 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
6595 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
6596 (*aarch64_mul3_elt<mode>): Likewise.
6597 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
6598 (*aarch64_mul3_elt_to_64v2df): Likewise.
6599 (*aarch64_mla_elt<mode>): Likewise.
6600 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
6601 (*aarch64_mls_elt<mode>): Likewise.
6602 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
6603 (*aarch64_fma4_elt<mode>): Likewise.
6604 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
6605 (*aarch64_fma4_elt_to_64v2df): Likewise.
6606 (*aarch64_fnma4_elt<mode>): Likewise.
6607 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
6608 (*aarch64_fnma4_elt_to_64v2df): Likewise.
6609 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
6610 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
6611 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6612 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6613 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
6614 (aarch64_sqdmull_lane<mode>_internal): Likewise.
6615 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
6617 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
6619 * config/aarch64/aarch64-simd.md
6620 (aarch64_be_checked_get_lane<mode>): New define_expand.
6621 * config/aarch64/aarch64-simd-builtins.def
6622 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
6623 New builtin definition.
6624 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
6625 Use new safe be builtin.
6627 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6629 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
6631 (aarch64_be_st1<mode>): Likewise.
6632 (aarch_ld1<VALL:mode>): Define_expand modified.
6633 (aarch_st1<VALL:mode>): Likewise.
6634 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
6635 (UNSPEC_ST1): Likewise.
6637 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
6639 * config/microblaze/microblaze.md: Add trap insn and attribute
6641 2014-01-23 Dodji Seketeli <dodji@redhat.com>
6643 PR preprocessor/58580
6644 * input.h (location_get_source_line): Take an additional line_size
6646 (void diagnostics_file_cache_fini): Declare new function.
6647 * input.c (struct fcache): New type.
6648 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
6649 New static constants.
6650 (diagnostic_file_cache_init, total_lines_num)
6651 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
6652 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
6653 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
6654 (get_next_line, read_next_line, goto_next_line, read_line_num):
6655 New static function definitions.
6656 (diagnostic_file_cache_fini): New function.
6657 (location_get_source_line): Take an additional output line_len
6658 parameter. Re-write using lookup_or_add_file_to_cache_tab and
6660 * diagnostic.c (diagnostic_finish): Call
6661 diagnostic_file_cache_fini.
6662 (adjust_line): Take an additional input parameter for the length
6663 of the line, rather than calculating it with strlen.
6664 (diagnostic_show_locus): Adjust the use of
6665 location_get_source_line and adjust_line with respect to their new
6666 signature. While displaying a line now, do not stop at the first
6667 null byte. Rather, display the zero byte as a space and keep
6668 going until we reach the size of the line.
6669 * Makefile.in: Add vec.o to OBJS-libcommon
6671 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
6672 Ilya Tocar <ilya.tocar@intel.com>
6674 * config/i386/avx512fintrin.h (_mm512_kmov): New.
6675 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
6676 (__builtin_ia32_kmov16): Ditto.
6677 * config/i386/i386.md (UNSPEC_KMOV): New.
6680 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
6682 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
6683 (_mm512_storeu_si512): Ditto.
6685 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
6688 * rtl.h (get_referenced_operands): Declare.
6689 * recog.c (get_referenced_operands): New function.
6690 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
6691 operands have been referenced when recording LO_SUM references.
6693 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
6695 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
6697 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
6699 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
6700 Enable for generic and recent AMD targets.
6702 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
6704 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
6705 ARG_SIZE note when adjustment was eliminated.
6707 2014-01-22 Jeff Law <law@redhat.com>
6709 PR tree-optimization/59597
6710 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
6711 in file. Accept new argument REGISTERING and use it to modify
6712 dump output appropriately.
6713 (register_jump_thread): Corresponding changes.
6714 (mark_threaded_blocks): Reinstate code to cancel unprofitable
6715 thread paths involving joiner blocks. Add code to dump cancelled
6716 jump threading paths.
6718 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
6720 PR rtl-optimization/59477
6721 * lra-constraints.c (inherit_in_ebb): Process call for living hard
6722 regs. Update reloads_num and potential_reload_hard_regs for all insns.
6724 2014-01-22 Tom Tromey <tromey@redhat.com>
6726 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
6728 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
6730 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
6732 PR rtl-optimization/59896
6733 * lra-constraints.c (process_alt_operands): Check unused note for
6734 matched operands of insn with no output reloads.
6736 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
6738 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
6739 (mips_move_from_gpr_cost): Likewise.
6741 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
6743 PR rtl-optimization/59858
6744 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
6745 ira_class_hard_regs_num.
6746 (process_alt_operands): Increase reject for dying matched operand.
6748 2014-01-21 Jakub Jelinek <jakub@redhat.com>
6751 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
6752 smaller than size, perform several stores or loads and stores
6753 at dst + count - size to store or copy all of size bytes, rather
6754 than just last modesize bytes.
6756 2014-01-20 DJ Delorie <dj@redhat.com>
6758 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
6759 that CLOBBERs are REGs before propogating their values.
6761 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
6764 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
6765 (cgraph_inline_failed_type): New function.
6766 * cgraph.h (DEFCIFCODE): Add type.
6767 (cgraph_inline_failed_type_t): New enum.
6768 (cgraph_inline_failed_type): New prototype.
6769 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
6770 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
6771 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
6772 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
6773 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
6774 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
6775 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
6776 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
6777 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
6778 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
6779 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
6780 OPTIMIZATION_MISMATCH.
6781 * tree-inline.c (expand_call_inline): Emit errors during
6782 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
6784 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
6787 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
6788 mode attribute in insn output.
6790 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
6792 * output.h (output_constant): Delete.
6793 * varasm.c (output_constant): Make private.
6795 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
6797 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
6799 2014-01-20 Jakub Jelinek <jakub@redhat.com>
6802 * tree.h (fold_builtin_strcat): New prototype.
6803 * builtins.c (fold_builtin_strcat): No longer static. Add len
6804 argument, if non-NULL, don't call c_strlen. Optimize
6805 directly into __builtin_memcpy instead of __builtin_strcpy.
6806 (fold_builtin_2): Adjust fold_builtin_strcat caller.
6807 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
6809 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
6811 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
6812 for SImode_address_operand operands, having only a REG argument.
6814 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
6816 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
6817 loader name using mbig-endian.
6818 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
6820 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
6822 * doc/invoke.texi (-march): Clarify documentation for AArch64.
6826 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
6828 * config/aarch64/aarch64-protos.h
6829 (aarch64_cannot_change_mode_class_ptr): Declare.
6830 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
6831 aarch64_cannot_change_mode_class_ptr): New.
6832 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
6833 backend hook aarch64_cannot_change_mode_class.
6835 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
6837 * common/config/aarch64/aarch64-common.c
6838 (aarch64_handle_option): Don't handle any option order logic here.
6839 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
6840 selected_cpu, warn on architecture version mismatch.
6841 (aarch64_override_options): Fix parsing order for option strings.
6843 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6844 Iain Sandoe <iain@codesourcery.com>
6847 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
6848 warning. Amend comment to reflect current functionality.
6850 2014-01-20 Richard Biener <rguenther@suse.de>
6853 * builtins.c (fold_builtin_strcat): Remove case better handled
6854 by tree-ssa-strlen.c.
6856 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
6858 * config/aarch64/aarch64.opt
6859 (mcpu, march, mtune): Make case-insensitive.
6861 2014-01-20 Jakub Jelinek <jakub@redhat.com>
6864 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
6865 if operands[1] is a REG or ZERO_EXTEND of a REG.
6867 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
6869 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
6871 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
6873 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
6874 long non-pic millicode calls.
6876 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6878 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
6880 2014-01-19 Kito Cheng <kito@0xlab.org>
6882 * builtins.c (expand_movstr): Check movstr expand done or fail.
6884 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
6885 H.J. Lu <hongjiu.lu@intel.com>
6888 * config/i386/i386.md (*lea<mode>): Zero-extend return register
6889 to DImode for zero-extended addresses.
6891 2014-01-19 Jakub Jelinek <jakub@redhat.com>
6893 PR rtl-optimization/57763
6894 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
6895 on the new indirect jump_insn and increment LABEL_NUSES (label).
6897 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
6901 * config.gcc (x86_archs): New variable.
6902 (x86_64_archs): Likewise.
6903 (x86_cpus): Likewise.
6904 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
6905 --with-arch/--with-cpu= options.
6906 Support --with-arch=/--with-cpu={nehalem,westmere,
6907 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
6909 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
6911 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
6912 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
6914 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
6916 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
6918 2014-01-18 Jakub Jelinek <jakub@redhat.com>
6921 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
6922 clear cpp_get_options (parse_in)->warn_unused_macros for
6923 ix86_target_macros_internal with cpp_define.
6925 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
6927 * jump.c (delete_related_insns): Keep (use (insn))s.
6928 * reorg.c (redundant_insn): Check for barriers too.
6930 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
6932 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
6934 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
6936 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
6937 call to $$dyncall when TARGET_LONG_CALLS is true.
6939 2014-01-17 Jeff Law <law@redhat.com>
6941 * ree.c (combine_set_extension): Temporarily disable test for
6942 changing number of hard registers.
6944 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
6947 * ipa-inline-analysis.c (inline_free_summary):
6948 Do not free summary of aliases.
6950 2014-01-17 Jakub Jelinek <jakub@redhat.com>
6953 * gimplify.c (gimplify_expr): Use create_tmp_var
6954 instead of create_tmp_var_raw. If cond doesn't have
6955 integral type, don't add the IFN_ANNOTATE builtin at all.
6957 2014-01-17 Martin Jambor <mjambor@suse.cz>
6960 * ipa-cp.c (prev_edge_clone): New variable.
6961 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
6962 Also resize prev_edge_clone vector.
6963 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
6964 (ipcp_edge_removal_hook): New function.
6965 (ipcp_driver): Register ipcp_edge_removal_hook.
6967 2014-01-17 Andrew Pinski <apinski@cavium.com>
6968 Steve Ellcey <sellcey@mips.com>
6971 * config/mips/mips.c (mips_print_operand): Check operand mode instead
6974 2014-01-17 Jeff Law <law@redhat.com>
6977 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
6978 so that pass_ccp runs first.
6980 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
6982 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
6983 (ix86_adjust_cost): Use !TARGET_XXX.
6984 (do_reorder_for_imul): Likewise.
6985 (swap_top_of_ready_list): Likewise.
6986 (ix86_sched_reorder): Likewise.
6988 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
6990 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
6991 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
6992 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
6993 (intel_memset): New. Duplicate slm_memset.
6994 (intel_cost): New. Duplicate slm_cost.
6995 (m_INTEL): New macro.
6996 (processor_target_table): Add "intel".
6997 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
6998 with PROCESSOR_INTEL for "intel".
6999 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7000 PROCESSOR_SILVERMONT.
7001 (ix86_issue_rate): Likewise.
7002 (ix86_adjust_cost): Likewise.
7003 (ia32_multipass_dfa_lookahead): Likewise.
7004 (swap_top_of_ready_list): Likewise.
7005 (ix86_sched_reorder): Likewise.
7006 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7007 instead of TARGET_OPT_AGU.
7008 * config/i386/i386.h (TARGET_INTEL): New.
7009 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7010 (processor_type): Add PROCESSOR_INTEL.
7011 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7012 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7014 2014-01-17 Marek Polacek <polacek@redhat.com>
7017 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7020 2014-01-17 Richard Biener <rguenther@suse.de>
7022 PR tree-optimization/46590
7023 * opts.c (default_options_table): Add entries for
7024 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7025 all enabled at -O1 but not for -Og.
7026 * common.opt (fbranch-count-reg): Remove Init(1).
7027 (fmove-loop-invariants): Likewise.
7028 (ftree-pta): Likewise.
7030 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7032 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7033 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7034 decls to at least the GCC 4.8 used alignments.
7037 * tree-nested.c (convert_nonlocal_reference_stmt,
7038 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7039 of GIMPLE_BIND stmts, adjust associated decls.
7041 2014-01-17 Richard Biener <rguenther@suse.de>
7043 PR tree-optimization/46590
7044 * vec.h (vec<>::bseach): New member function implementing
7045 binary search according to C89 bsearch.
7046 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7047 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7048 bitmap pointer again. Make accesses_in_loop a flat array.
7049 (mem_ref_obstack): New global.
7050 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7051 (mark_ref_stored): Likewise.
7052 (ref_indep_loop_p_2): Likewise.
7053 (set_ref_stored_in_loop): New helper function.
7054 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7055 (memref_free): Adjust.
7056 (record_mem_ref_loc): Simplify.
7057 (gather_mem_refs_stmt): Adjust.
7058 (sort_locs_in_loop_postorder_cmp): New function.
7059 (analyze_memory_references): Sort accesses_in_loop after
7060 loop postorder number.
7061 (find_ref_loc_in_loop_cmp): New function.
7062 (for_all_locs_in_loop): Find relevant cluster of locs in
7063 accesses_in_loop and iterate without recursion.
7064 (execute_sm): Avoid uninit warning.
7065 (struct ref_always_accessed): Simplify.
7066 (ref_always_accessed::operator ()): Likewise.
7067 (ref_always_accessed_p): Likewise.
7068 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7069 loop postorder numbers here.
7070 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7073 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7076 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7077 on decls for which assemble_alias has been called.
7079 2014-01-17 Nick Clifton <nickc@redhat.com>
7081 * config/msp430/msp430.opt: (mcpu): New option.
7082 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7083 (msp430_option_override): Parse target_cpu. If the MCU name
7084 matches a generic string, clear target_mcu.
7085 (msp430_attr): Allow numeric interrupt values up to 63.
7086 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7087 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7089 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7091 * config/msp430/msp430.md (popm): Use %J rather than %I.
7092 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7093 (addhi_cy_i): Use immediate_operand for operand 2.
7094 * doc/invoke.texi: Document -mcpu option.
7096 2014-01-17 Richard Biener <rguenther@suse.de>
7098 PR rtl-optimization/38518
7099 * df.h (df_analyze_loop): Declare.
7100 * df-core.c: Include cfgloop.h.
7101 (df_analyze_1): Split out main part of df_analyze.
7102 (df_analyze): Adjust.
7103 (loop_inverted_post_order_compute): New function.
7104 (loop_post_order_compute): Likewise.
7105 (df_analyze_loop): New function avoiding whole-function
7107 * loop-invariant.c (find_defs): Use df_analyze_loop.
7108 (find_invariants): Adjust.
7109 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7111 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7113 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7114 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7116 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7118 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7119 traversal when removing references.
7121 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7124 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7126 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7129 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7130 pushing a reload for an autoinc when we had previously reloaded an
7131 inner part of the address.
7133 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7135 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7137 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7138 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7139 when not giving up or versioning for alias only because of
7141 (vect_analyze_data_ref_dependences): Set to true.
7142 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7144 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7145 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7149 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7152 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7153 operand 0 predicate for gathers, use a new pseudo as subtarget.
7155 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7158 * lra-constraints.c (process_alt_operands): Add printing debug info.
7159 Check absence of input/output reloads for matched operands too.
7161 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7163 PR rtl-optimization/59835
7164 * ira.c (ira_init_register_move_cost): Increase cost for
7167 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7169 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7171 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7174 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7175 non-register objects. Use gen_(high/low)part more consistently.
7178 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7181 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7182 endian support, remove tests for WORDS_BIG_ENDIAN.
7183 (p8_mfvsrd_3_<mode>): Likewise.
7184 (reload_gpr_from_vsx<mode>): Likewise.
7185 (reload_gpr_from_vsxsf): Likewise.
7186 (p8_mfvsrd_4_disf): Likewise.
7188 2014-01-16 Richard Biener <rguenther@suse.de>
7190 PR rtl-optimization/46590
7191 * lcm.c (compute_antinout_edge): Use postorder iteration.
7192 (compute_laterin): Use inverted postorder iteration.
7194 2014-01-16 Nick Clifton <nickc@redhat.com>
7197 * varasm.c (output_constant): Return the number of bytes actually
7199 (output_constructor_array_range): Update the field size with the
7200 number of bytes emitted by output_constant.
7201 (output_constructor_regular_field): Likewise. Also do not
7202 complain if the total number of bytes emitted is now greater
7203 than the expected fieldpos.
7204 * output.h (output_constant): Update prototype and descriptive comment.
7206 2014-01-16 Marek Polacek <polacek@redhat.com>
7209 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7210 it is error_mark_node.
7212 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7214 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7215 VALID_AVX256_REG_OR_OI_MODE.
7217 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7219 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7220 current procedure should be profiled.
7222 2014-01-15 Andrew Pinski <apinski@cavium.com>
7224 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7225 of moving from/to the STACK_REG register class.
7227 2014-01-15 Richard Henderson <rth@redhat.com>
7230 * reginfo.c (global_regs_decl): Globalize.
7231 * rtl.h (global_regs_decl): Declare.
7232 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7233 reserved via global_regs.
7235 2014-01-15 Teresa Johnson <tejohnson@google.com>
7237 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7239 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7241 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7242 and vmulosh rather than call gen_vec_widen_smult_*.
7243 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7244 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7245 (vec_widen_smult_even_v16qi): Likewise.
7246 (vec_widen_umult_even_v8hi): Likewise.
7247 (vec_widen_smult_even_v8hi): Likewise.
7248 (vec_widen_umult_odd_v16qi): Likewise.
7249 (vec_widen_smult_odd_v16qi): Likewise.
7250 (vec_widen_umult_odd_v8hi): Likewise.
7251 (vec_widen_smult_odd_v8hi): Likewise.
7252 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7253 vmuloub rather than call gen_vec_widen_umult_*.
7254 (vec_widen_umult_lo_v16qi): Likewise.
7255 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7256 vmulosb rather than call gen_vec_widen_smult_*.
7257 (vec_widen_smult_lo_v16qi): Likewise.
7258 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7259 rather than call gen_vec_widen_umult_*.
7260 (vec_widen_umult_lo_v8hi): Likewise.
7261 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7262 rather than call gen_vec_widen_smult_*.
7263 (vec_widen_smult_lo_v8hi): Likewise.
7265 2014-01-15 Jeff Law <law@redhat.com>
7267 PR tree-optimization/59747
7268 * ree.c (find_and_remove_re): Properly handle case where a second
7269 eliminated extension requires widening a copy created for elimination
7270 of a prior extension.
7271 (combine_set_extension): Ensure that the number of hard regs needed
7272 for a destination register does not change when we widen it.
7274 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7276 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7277 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7278 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7279 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7280 (avr-*-rtems*): Likewise.
7281 (bfin*-rtems*): Likewise.
7282 (moxie-*-rtems*): Likewise.
7283 (h8300-*-rtems*): Likewise.
7284 (i[34567]86-*-rtems*): Likewise.
7285 (lm32-*-rtems*): Likewise.
7286 (m32r-*-rtems*): Likewise.
7287 (m68k-*-rtems*): Likewise.
7288 (microblaze*-*-rtems*): Likewise.
7289 (mips*-*-rtems*): Likewise.
7290 (powerpc-*-rtems*): Likewise.
7291 (sh-*-rtems*): Likewise.
7292 (sparc-*-rtems*): Likewise.
7293 (sparc64-*-rtems*): Likewise.
7294 (v850-*-rtems*): Likewise.
7295 (m32c-*-rtems*): Likewise.
7297 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7299 PR rtl-optimization/59511
7300 * ira.c (ira_init_register_move_cost): Use memory costs for some
7301 cases of register move cost calculations.
7302 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7303 instead of BB frequency.
7304 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7305 * lra-assigns.c (find_hard_regno_for): Ditto.
7307 2014-01-15 Richard Biener <rguenther@suse.de>
7309 PR tree-optimization/59822
7310 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7311 (vectorizable_load): Use it to hoist defs of uses of invariant
7312 loads out of the loop.
7314 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7315 Kugan Vivekanandarajah <kuganv@linaro.org>
7318 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7321 2014-01-15 Richard Biener <rguenther@suse.de>
7323 PR rtl-optimization/59802
7324 * lcm.c (compute_available): Use inverted postorder to seed
7325 the initial worklist.
7327 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7330 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7331 ADDR_REGS for invalid symrefs in non-PIC code.
7333 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7336 * builtins.c (determine_block_size): Initialize *probable_max_size
7337 even if len_rtx is CONST_INT.
7339 2014-01-14 Andrew Pinski <apinski@cavium.com>
7341 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7342 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7343 (cortexa53_tunings): Likewise.
7344 (aarch64_sched_issue_rate): New function.
7345 (TARGET_SCHED_ISSUE_RATE): Define.
7347 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7349 * ira-costs.c (find_costs_and_classes): Add missed
7350 ira_init_register_move_cost_if_necessary.
7352 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7355 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7357 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7360 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7361 to indicate if type is used for function return value. Warn ABI
7362 change if the vector mode isn't available for function return value.
7363 (ix86_function_arg_advance): Pass false to type_natural_mode.
7364 (ix86_function_arg): Likewise.
7365 (ix86_gimplify_va_arg): Likewise.
7366 (function_arg_32): Don't warn ABI change.
7367 (ix86_function_value): Pass true to type_natural_mode.
7368 (ix86_return_in_memory): Likewise.
7369 (ix86_struct_value_rtx): Removed.
7370 (TARGET_STRUCT_VALUE_RTX): Likewise.
7372 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7374 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7375 converting a conditional jump into a conditional return.
7377 2014-01-14 Richard Biener <rguenther@suse.de>
7379 PR tree-optimization/58921
7380 PR tree-optimization/59006
7381 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7382 hoisting invariant stmts.
7383 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7384 invariant loads on the preheader edge if possible.
7386 2014-01-14 Joey Ye <joey.ye@arm.com>
7388 * doc/plugin.texi (Building GCC plugins): Update to C++.
7390 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7392 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7393 (_mm_rcp28_round_ss): Ditto.
7394 (_mm_rsqrt28_round_sd): Ditto.
7395 (_mm_rsqrt28_round_ss): Ditto.
7396 (_mm_rcp28_sd): Ditto.
7397 (_mm_rcp28_ss): Ditto.
7398 (_mm_rsqrt28_sd): Ditto.
7399 (_mm_rsqrt28_ss): Ditto.
7400 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7401 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7402 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7403 (IX86_BUILTIN_RCP28SD): Ditto.
7404 (IX86_BUILTIN_RCP28SS): Ditto.
7405 (IX86_BUILTIN_RSQRT28SD): Ditto.
7406 (IX86_BUILTIN_RSQRT28SS): Ditto.
7407 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7408 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7409 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7410 (ix86_expand_special_args_builtin): Expand new FTYPE.
7411 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7412 (srcp14<mode>): Make insn unary.
7413 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7414 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7415 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7416 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7417 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7418 Fix rounding: make it SAE only.
7419 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7421 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7423 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7424 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7425 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7426 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7427 (round_saeonly_mask_scalar_operand4): Ditto.
7428 (round_saeonly_mask_scalar_op3): Ditto.
7429 (round_saeonly_mask_scalar_op4): Ditto.
7431 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7433 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7434 Implement -maltivec=be for vec_insert and vec_extract.
7436 2014-01-10 DJ Delorie <dj@redhat.com>
7438 * config/msp430/msp430.md (call_internal): Don't allow memory
7439 references with SP as the base register.
7440 (call_value_internal): Likewise.
7441 * config/msp430/constraints.md (Yc): New. For memory references
7442 that don't use SP as a base register.
7444 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7445 "an integer without a # prefix"
7446 * config/msp430/msp430.md (epilogue_helper): Use it.
7448 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7451 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7452 AVX512F gather builtins.
7453 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7454 on gather decls with INTEGER_TYPE masktype.
7455 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7456 directly into the builtin rather than hoisting it before loop.
7458 PR tree-optimization/59387
7459 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7460 (scev_const_prop): If folded_casts and type has undefined overflow,
7461 use force_gimple_operand instead of force_gimple_operand_gsi and
7462 for each added stmt if it is assign with
7463 arith_code_with_undefined_signed_overflow, call
7464 rewrite_to_defined_overflow.
7465 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7466 gimple-fold.h instead.
7467 (arith_code_with_undefined_signed_overflow,
7468 rewrite_to_defined_overflow): Moved to ...
7469 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
7470 rewrite_to_defined_overflow): ... here. No longer static.
7471 Include gimplify-me.h.
7472 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
7473 rewrite_to_defined_overflow): New prototypes.
7475 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7477 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
7479 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
7481 * builtins.c (get_object_alignment_2): Minor tweak.
7482 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
7484 2014-01-13 Christian Bruel <christian.bruel@st.com>
7486 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
7487 optimized non constant lengths.
7489 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7492 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
7493 load as __atomic_load_N if possible.
7495 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7497 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
7499 (rs6000_expand_builtin): Adjust call.
7501 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7504 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
7505 * config/rs6000/rs6000.c: Include target-globals.h.
7506 (rs6000_set_current_function): Instead of doing target_reinit
7507 unconditionally, use save_target_globals_default_opts and
7508 restore_target_globals.
7510 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
7512 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
7513 (rs6000_expand_builtin): Handle mffs and mtfsf.
7514 (rs6000_init_builtins): Define mffs and mtfsf.
7515 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
7516 (rs6000_mffs): New pattern.
7517 (rs6000_mtfsf): New pattern.
7519 2014-01-11 Bin Cheng <bin.cheng@arm.com>
7521 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
7522 Start narrowing with START. Apply candidate-use pair
7523 and check overall cost in narrowing.
7524 (iv_ca_prune): Pass new argument.
7526 2014-01-10 Jeff Law <law@redhat.com>
7529 * ree.c (combine_reaching_defs): Ensure the defining statement
7530 occurs before the extension when optimizing extensions with
7531 different source and destination hard registers.
7533 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7536 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
7537 vtables into the type inheritance graph.
7539 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7541 PR rtl-optimization/59754
7542 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
7543 modes in the REGNO != REGNO case.
7545 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7547 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
7549 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7551 PR tree-optimization/59745
7552 * tree-predcom.c (tree_predictive_commoning_loop): Call
7553 free_affine_expand_cache if giving up because components is NULL.
7555 * target-globals.c (save_target_globals): Allocate < 4KB structs using
7556 GC in payload of target_globals struct instead of allocating them on
7557 the heap and the larger structs separately using GC.
7558 * target-globals.h (struct target_globals): Make regs, hard_regs,
7559 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
7560 of GTY((skip)) and change type to void *.
7561 (reset_target_globals): Cast loads from those fields to corresponding
7564 2014-01-10 Steve Ellcey <sellcey@mips.com>
7567 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
7568 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
7569 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
7571 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
7574 * aarch64-modes.def (CC_Zmode): New flags mode.
7575 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
7576 represents an equality.
7577 (aarch64_get_condition_code): Handle CC_Zmode.
7578 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
7580 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7582 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
7583 extraction in good case.
7585 2014-01-10 Richard Biener <rguenther@suse.de>
7587 PR tree-optimization/59374
7588 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
7589 checking after SLP discovery. Mark stmts not participating
7590 in any SLP instance properly.
7592 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7594 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
7595 when handling a SET rtx.
7597 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7599 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
7600 (cortex-a57): Likewise.
7601 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
7603 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7605 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
7606 non-iwmmxt builtins.
7608 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7612 * ipa-devirt.c record_target_from_binfo): Take as argument
7613 stack of binfos and lookup matching one for virtual inheritance.
7614 (possible_polymorphic_call_targets_1): Update.
7616 2014-01-10 Huacai Chen <chenhc@lemote.com>
7618 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
7619 kernel strings for Loongson-2E/2F/3A.
7621 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7624 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
7625 is_gimple_call before calling gimple_call_internal_p.
7627 2014-01-09 Steve Ellcey <sellcey@mips.com>
7629 * Makefile.in (TREE_FLOW_H): Remove.
7630 (TREE_SSA_H): Add file names from tree-flow.h.
7631 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
7632 * tree.h: Remove tree-flow.h reference.
7633 * hash-table.h: Remove tree-flow.h reference.
7634 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
7635 reference with tree-ssa-loop.h.
7637 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7639 * doc/invoke.texi: Add -maltivec={be,le} options, and document
7640 default element-order behavior for -maltivec.
7641 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
7642 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
7643 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
7644 when targeting big endian, at least for now.
7645 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
7647 2014-01-09 Jakub Jelinek <jakub@redhat.com>
7650 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
7651 var satisfies use_register_for_decl, just take into account type
7652 alignment, rather than decl alignment.
7654 PR tree-optimization/59622
7655 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
7656 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
7657 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
7658 Don't devirtualize for inplace at all. For targets.length () == 1,
7659 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
7661 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
7663 * config/i386/i386.md (cpu): Remove the unused btver1.
7665 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
7667 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
7669 2014-01-09 Jakub Jelinek <jakub@redhat.com>
7672 * tree-core.h (struct target_globals): New forward declaration.
7673 (struct tree_target_option): Add globals field.
7674 * tree.h (TREE_TARGET_GLOBALS): Define.
7675 (prepare_target_option_nodes_for_pch): New prototype.
7676 * target-globals.h (struct target_globals): Define even if
7678 * tree.c (prepare_target_option_node_for_pch,
7679 prepare_target_option_nodes_for_pch): New functions.
7680 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
7681 * config/i386/i386.c: Include target-globals.h.
7682 (ix86_set_current_function): Instead of doing target_reinit
7683 unconditionally, use save_target_globals_default_opts and
7684 restore_target_globals.
7686 2014-01-09 Richard Biener <rguenther@suse.de>
7688 PR tree-optimization/59715
7689 * tree-cfg.h (split_critical_edges): Declare.
7690 * tree-cfg.c (split_critical_edges): Export.
7691 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
7693 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
7695 * cfgexpand.c (expand_stack_vars): Optionally disable
7696 asan stack protection.
7697 (expand_used_vars): Likewise.
7698 (partition_stack_vars): Likewise.
7699 * asan.c (asan_emit_stack_protection): Optionally disable
7700 after return stack usage.
7701 (instrument_derefs): Optionally disable memory access instrumentation.
7702 (instrument_builtin_call): Likewise.
7703 (instrument_strlen_call): Likewise.
7704 (asan_protect_global): Optionally disable global variables protection.
7705 * doc/invoke.texi: Added doc for new options.
7706 * params.def: Added new options.
7707 * params.h: Likewise.
7709 2014-01-09 Jakub Jelinek <jakub@redhat.com>
7711 PR rtl-optimization/59724
7712 * ifcvt.c (cond_exec_process_if_block): Don't call
7713 flow_find_head_matching_sequence with 0 longest_match.
7714 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
7715 non-active insns if !stop_after.
7716 (try_head_merge_bb): Revert 2014-01-07 changes.
7718 2014-01-08 Jeff Law <law@redhat.com>
7720 * ree.c (get_sub_rtx): New function, extracted from...
7721 (merge_def_and_ext): Here.
7722 (combine_reaching_defs): Use get_sub_rtx.
7724 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
7726 * cgraph.h (varpool_variable_node): Do not choke on null node.
7728 2014-01-08 Catherine Moore <clm@codesourcery.com>
7730 * config/mips/mips.md (simple_return): Attempt to use JRC
7732 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
7734 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
7736 PR rtl-optimization/59137
7737 * reorg.c (steal_delay_list_from_target): Call update_block for
7739 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
7741 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7743 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
7744 two duplicate entries.
7746 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
7749 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
7751 * config/mips/mips.c (mips_truncated_op_cost): New function.
7752 (mips_rtx_costs): Adjust test for BADDU.
7753 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
7755 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
7757 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
7758 (*baddu_si): ...this new pattern.
7760 2014-01-08 Jakub Jelinek <jakub@redhat.com>
7763 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
7765 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
7768 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
7770 (expand_expr, expand_normal): Adjust.
7771 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
7772 inner_reference_p. Use inner_reference_p to expand inner references.
7773 (store_expr): Adjust.
7774 * cfgexpand.c (expand_call_stmt): Adjust.
7776 2014-01-08 Rong Xu <xur@google.com>
7778 * gcov-io.c (gcov_var): Move from gcov-io.h.
7779 (gcov_position): Ditto.
7780 (gcov_is_error): Ditto.
7781 (gcov_rewrite): Ditto.
7782 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
7783 only part to libgcc/libgcov.h.
7785 2014-01-08 Marek Polacek <polacek@redhat.com>
7788 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
7790 2014-01-08 Marek Polacek <polacek@redhat.com>
7793 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
7795 2014-01-08 Jakub Jelinek <jakub@redhat.com>
7797 PR rtl-optimization/59649
7798 * stor-layout.c (get_mode_bounds): For BImode return
7799 0 and STORE_FLAG_VALUE.
7801 2014-01-08 Richard Biener <rguenther@suse.de>
7804 * gimple.h (is_gimple_builtin_call): Remove.
7805 (gimple_builtin_call_types_compatible_p): New.
7806 (gimple_call_builtin_p): New overload.
7807 * gimple.c (is_gimple_builtin_call): Remove.
7808 (validate_call): Rename to ...
7809 (gimple_builtin_call_types_compatible_p): ... this and export. Also
7811 (validate_type): New static function.
7812 (gimple_call_builtin_p): New overload and adjust.
7813 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
7814 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
7815 (gimple_fold_stmt_to_constant_1): Likewise.
7816 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
7818 2014-01-08 Richard Biener <rguenther@suse.de>
7821 * gimplify.c (gimplify_expr): Gimplify register-register type
7822 VIEW_CONVERT_EXPRs to separate stmts.
7824 2014-01-07 Jeff Law <law@redhat.com>
7827 * ree.c (combine_set_extension): Handle case where source
7828 and destination registers in an extension insn are different.
7829 (combine_reaching_defs): Allow source and destination registers
7830 in extension to be different under limited circumstances.
7831 (add_removable_extension): Remove restriction that the
7832 source and destination registers in the extension are the same.
7833 (find_and_remove_re): Emit a copy from the extension's
7834 destination to its source after the defining insn if
7835 the source and destination registers are different.
7838 * ifcvt.c (merge_if_block): If we are merging a block with more than
7839 one successor with a block with no successors, remove any BARRIER
7840 after the second block.
7842 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
7844 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
7846 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
7849 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
7850 for 14-bit register offsets when INT14_OK_STRICT is false.
7852 2014-01-07 Roland Stigge <stigge@antcom.de>
7853 Michael Meissner <meissner@linux.vnet.ibm.com>
7856 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
7857 Only check TFmode for SPE constants. Don't check TImode or TDmode.
7859 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
7861 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
7864 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
7866 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
7867 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
7868 rtx is const0_rtx or not.
7870 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
7873 * target-globals.c (save_target_globals): Remove this_fn_optab
7875 * toplev.c: Include optabs.h.
7876 (target_reinit): Temporarily restore the global options if another
7877 set of options are in force.
7879 2014-01-07 Jakub Jelinek <jakub@redhat.com>
7881 PR rtl-optimization/58668
7882 * cfgcleanup.c (flow_find_cross_jump): Don't count
7883 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
7884 to determine what is counted.
7885 (flow_find_head_matching_sequence): Use active_insn_p to determine
7887 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
7889 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
7890 determine what is counted.
7892 PR tree-optimization/59643
7893 * tree-predcom.c (split_data_refs_to_components): If one dr is
7894 read and one write, determine_offset fails and the write isn't
7895 in the bad component, just put the read into the bad component.
7897 2014-01-07 Mike Stump <mikestump@comcast.net>
7898 Jakub Jelinek <jakub@redhat.com>
7901 * tree-core.h (struct tree_optimization_option): Change optabs
7902 type from unsigned char * to void *.
7903 * optabs.c (init_tree_optimization_optabs): Adjust
7904 TREE_OPTIMIZATION_OPTABS initialization.
7906 2014-01-06 Jakub Jelinek <jakub@redhat.com>
7909 * config/i386/i386.h (struct machine_function): Add
7910 no_drap_save_restore field.
7911 * config/i386/i386.c (ix86_save_reg): Use
7912 !cfun->machine->no_drap_save_restore instead of
7913 crtl->stack_realign_needed.
7914 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
7915 this function clears frame_pointer_needed. Set
7916 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
7917 and DRAP reg is needed.
7919 2014-01-06 Marek Polacek <polacek@redhat.com>
7922 * doc/implement-c.texi: Mention that other integer types are
7923 permitted as bit-field types in strictly conforming mode.
7925 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
7927 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
7930 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
7932 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
7934 2014-01-06 Martin Jambor <mjambor@suse.cz>
7937 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
7939 * ipa-prop.c (ipa_print_node_params): Fix indentation.
7941 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
7945 * var-tracking.c (add_stores): Preserve the value of the source even if
7946 we don't record the store.
7948 2014-01-06 Terry Guo <terry.guo@arm.com>
7950 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
7952 2014-01-05 Iain Sandoe <iain@codesourcery.com>
7955 * config/darwin.c (darwin_function_section): Adjust return values to
7956 correspond to optimisation changes made in r206070.
7958 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
7960 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
7961 from prefetch_block tune setting.
7962 (nocona_cost): Correct size of prefetch block to 64.
7964 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
7966 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
7967 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
7968 used to save the static chain register in the computation of the offset
7969 from which the FP registers need to be restored.
7971 2014-01-04 Jakub Jelinek <jakub@redhat.com>
7973 PR tree-optimization/59519
7974 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
7975 ICE if get_current_def (current_new_name) is already non-NULL, as long
7976 as it is a phi result of some other phi in *new_exit_bb that has
7979 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
7980 or vmovdqu* for misaligned_operand.
7981 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
7982 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
7983 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
7984 aligned_mem for AVX512F masked aligned load and store builtins and for
7987 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
7989 PR tree-optimization/59651
7990 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
7991 Address range for negative step should be added by TYPE_SIZE_UNIT.
7993 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
7995 * config/m68k/m68k.c (handle_move_double): Handle pushes with
7996 overlapping registers also for registers other than the stack pointer.
7998 2014-01-03 Marek Polacek <polacek@redhat.com>
8001 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8004 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8007 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8010 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8011 (push splitter): Use <P:MODE_SIZE> instead of
8012 GET_MODE_SIZE (<P:MODE>mode).
8013 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8014 (mov -1, reg peephole2): Likewise.
8015 * config/i386/sse.md (*mov<mode>_internal,
8016 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8017 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8018 *<code><mode>3, *andnot<mode>3<mask_name>,
8019 <mask_codefor><code><mode>3<mask_name>): Likewise.
8020 * config/i386/subst.md (mask_mode512bit_condition,
8021 sd_mask_mode512bit_condition): Likewise.
8023 2014-01-02 Xinliang David Li <davidxl@google.com>
8025 PR tree-optimization/59303
8026 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8027 (dump_predicates): Better output format.
8028 (pred_equal_p): New function.
8029 (is_neq_relop_p): Ditto.
8030 (is_neq_zero_form_p): Ditto.
8031 (pred_expr_equal_p): Ditto.
8032 (pred_neg_p): Ditto.
8033 (simplify_pred): Ditto.
8034 (simplify_preds_2): Ditto.
8035 (simplify_preds_3): Ditto.
8036 (simplify_preds_4): Ditto.
8037 (simplify_preds): Ditto.
8039 (push_to_worklist): Ditto.
8040 (get_pred_info_from_cmp): Ditto.
8041 (is_degenerated_phi): Ditto.
8042 (normalize_one_pred_1): Ditto.
8043 (normalize_one_pred): Ditto.
8044 (normalize_one_pred_chain): Ditto.
8045 (normalize_preds): Ditto.
8046 (normalize_cond_1): Remove function.
8047 (normalize_cond): Ditto.
8048 (is_gcond_subset_of): Ditto.
8049 (is_subset_of_any): Ditto.
8050 (is_or_set_subset_of): Ditto.
8051 (is_and_set_subset_of): Ditto.
8052 (is_norm_cond_subset_of): Ditto.
8053 (pred_chain_length_cmp): Ditto.
8054 (convert_control_dep_chain_into_preds): Type change.
8055 (find_predicates): Ditto.
8056 (find_def_preds): Ditto.
8057 (destroy_predicates_vecs): Ditto.
8058 (find_matching_predicates_in_rest_chains): Ditto.
8059 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8060 (is_pred_expr_subset): Ditto.
8061 (is_pred_chain_subset_of): Ditto.
8062 (is_included_in): Ditto.
8063 (is_superset_of): Ditto.
8065 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8067 Update copyright years.
8069 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8071 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8072 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8073 config/arc/arc.md, config/arc/arc.opt,
8074 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8075 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8076 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8077 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8078 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8079 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8080 vtable-verify.c, vtable-verify.h: Use the standard form for the
8083 2014-01-02 Tobias Burnus <burnus@net-b.de>
8085 * gcc.c (process_command): Update copyright notice dates.
8086 * gcov-dump.c: Ditto.
8088 * doc/cpp.texi: Bump @copying's copyright year.
8089 * doc/cppinternals.texi: Ditto.
8090 * doc/gcc.texi: Ditto.
8091 * doc/gccint.texi: Ditto.
8092 * doc/gcov.texi: Ditto.
8093 * doc/install.texi: Ditto.
8094 * doc/invoke.texi: Ditto.
8096 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8098 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8100 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8102 * config/i386/sse.md (*mov<mode>_internal): Guard
8103 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8105 PR rtl-optimization/59647
8106 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8107 new_rtx into UNSIGNED_FLOAT rtxes.
8109 Copyright (C) 2014 Free Software Foundation, Inc.
8111 Copying and distribution of this file, with or without modification,
8112 are permitted in any medium without royalty provided the copyright
8113 notice and this notice are preserved.